Abstract Book 2015-Web

VOLUME 19
PA G E S S 1 S 5 8 8
ISSN 1027 3719
NUMBER 12
DECEMBER 2015
SUPPLEMENT 2
The
International
Journal of Tuberculosis
and Lung Disease
The Official Journal of the International Union Against Tuberculosis and Lung Disease
ABSTRACT BOOK
46th World Conference

on Lung Health of the
International Union Against
Tuberculosis and Lung Disease (The Union)
CAPE TOWN SOUTH AFRICA

26 DECEMBER 2015
SUPPLEMENT 2
VOLUME 19 NUMBER 12
SYMPOSIA
FRIDAY 4 DECEMBER 2015
S1 02. Best practice in management of side effects among
DR-TB patients to improve quality of care
S1 03. Pre-approval of access to new TB drugs: what,
where and how
S2 04. Outstanding issues in HIV/AIDS treatment
S3 05. Zoonotic T8 session I - Trends, diagnostics and
infections models
S5 06. One pilot after another: how can we move toward
a sustainable scale-up of child lung health
interventions?
S6 07. Role of surgery in management of TB: a promise or
redundancy?
S7 08. Maternal and infant TB: advancing our
understanding of pathogenesis, treatment and
prevention
S8 09. The cost of TB and lung health interventions:
methodologies and implications for health systems
and patients
S10 10. Research is needed to increase childrens access to
drug-resistant TB care
S11 11. TB surveillance in health care workers: challenges
and approaches in high-burden, low-resource settings
S12 12. Managing pharmacovigilance systems: optimising
the rational use of new TB medicines and novel
regimens
S13 13. Can mobile technology improve lung health in
poorer countries and communities? Benefits and
opportunities
S14 14. Spelunking for treasure: searching for candidate
biomarkers as prognostic indicators for treatment
outcome
S15 15. TB-HIV epidemiology within the context of national
TB prevalence surveys in Africa
S16 16. Scaling up MDR-TB services in resource-limited
settings: progress and challenges from the field
S18 17. Maximising opportunities for integrating TB
services into maternal and child health programmes
S18 18. Five years of XpertW MTB/RIF implementation:
lessons learnt for increasing the impact of future new
diagnostics
S20 19. Quality laboratory services and impact on patient
care
S20 20. The unaddressed health challenge of TB in prisons:
applying End TB Strategy concepts of patient-centred
care
S21 21. TB contact management in children: closing policyimplementation gaps
S22 22. A community-based approach to reaching the
missing three million
S24 23. Moving towards universal access to drug
susceptibility testing: progress and challenges
S25
S25
S26
DECEMBER 2015
24. Why and how to use tobacco taxes as funding

mechanisms for financing health in low- and middleincome countries
25. Can tuberculosis immunisation prevent initial
infection (not just disease complications)?
26. WHO/ERS initiative on e/mHealth in tuberculosis
and tobacco control
SATURDAY 5 DECEMBER 2015

S29 27. Building on emerging knowledge to develop novel
regimens for paediatric drug-resistant TB
S30 28. End TB, end MDR-TB
S31 29. Pneumonia in children and adults beyond 2015
S32 30. Eliminating the burden of TB in indigenous
populations
S33 31. High burden, neglected patients: addressing the
intersection of tuberculosis and mental disorders
S34 32. Does MPOWER empower women?
S35 33. Modelling to support TB control policy in the era of
the End TB Strategy
S36 34. MDR-TB: a complex disease requiring a
comprehensive health system response
S36 35. Controversies and critical issues in clinical trial
design for TB
S37 36. Innovations to improve the TB/HIV patient
experience: from diagnosis to treatment
S39 37. Inhaled therapies for tuberculosis
S40 38. Affordable solutions to indoor air pollution for onethird of the worlds population
S40 39. Taking stock of TB epidemiological assessments:
experience, lessons learned and the way forward
S41 40. Reinforcing research for TB elimination at the
country level: experiences from path-finding countries
S42 41. The Union/CDC Late-Breaker Session on TB
S46 42. TB research and communities: ethical and practical
considerations
S47 43. Country experiences: enablers for communitybased multi-drug resistant tuberculosis management
programmes
S49 44. Achieving quality Xpert MTB/RIF use for rapid case
detection
S52 46. Strengthening community systems: building the
evidence for impact
S53 47. Getting patients treated for latent TB infection
(LTBI) as TB prevention
S54 48. Beyond accelerated drug development: building
successful partnerships and platforms as a critical path
to TB drug regimens and PreDiCT-TB
S55 49. Sanatoria: back to the future?
S57 50. Smokeless tobacco: a silent killer
S60 51. e/m-Health as a transformational tool in TB control
in high-burden countries
S16 52. Developing a rights-based approach to prevention

and treatment of tuberculosis
SUNDAY 6 DECEMBER 2015
S63 53. Fighting TB in the post-2015 era: paradigm shift in
strategies and plans to put an end to TB
S64 54. Critical updates on the treatment of TB-HIV coinfected children
S66 55. Why are DR-TB patients lost by TB programmes?
Patients perspective on how to address DR-TB
treatment adherence
S67 56. The relationship between gender and specific
tobacco products, tobacco consumption and
inequalities
S68 57. The role of cost-effectiveness in achieving the End
TB Strategy
S70 58. Universal health coverage and insurance: how TB
fits into the health financing picture
S71 59. HIV-TB Late-Breaker Session
S74 60. Moving past clinic-based care: practical steps to
improve TB care through community engagement
S75 61. Progress in clinical trials for TB treatment: 2015
S76 62. Innovative approach to improving access and
quality of care for DR-TB: implementing the ECHO
telehealth model
S78 63. Zoonotic TB session II: What do humans, baboons
and livestock have in common?
S81 64. TB contact investigation in high-risk populations:
innovative approaches and implementation
experiences
S82 65. The XDR experience: clinician and patient
perspectives on treatment and care
ABSTRACT PRESENTATIONS
FRIDAY 4 DECEMBER 2015
e-poster sessions
S84 01. Zoonotic tuberculosis: trends, detections and
interpretation
S88 02. TB in children: perspectives from around the world
S93 03. Drug-resistant TB: outcomes I
S97 04. LTBI: a potpourri
S101 05. Rats, videos and secretions: the landscape of TB
detection
Oral abstract sessions
S106 01. Symptoms, sputum, sequences: TB diagnosis and
resistance
S110 02. Around the world vulnerability takes many forms
S114 03. Placing the patient at the center of Care
S119 04. Law enforcement and awareness
S123 05. Financing TB services
S127 06. MDR-TB: trials, cohorts and predictors
S131 07. Many paths to the same END TB Strategy
S136 08. The molecular basis of PZA resistance: from bench
to bedside
S140 09. HIV and lung health
S144 10. Using media for communication on TB
Poster discussion sessions

S148 01. Extra-pulmonary tuberculosis: detection and
treatment
S151 02. Drug development: something old, something
new
S156 03. Treatment outcomes and ethics: Bedaquiline
S160 04. New approaches to health systems, care and
management for drug-resistant TB
S165 05. Collaborative services for TB and diabetes
S168 06. TB surveillance from Swaziland to Japan
S172 07. Tuberculosis: from the bench to the grave
S176 08. Trends, transmission and tuberculosis case finding
S180 09. Public-private strategies to End TB: the case of India
S184 10. Whats app doc? Use of m-health
S189 11. Basic science: bugs and drugs
S193 12. TB diagnostic challenges: molecular and other
techniques
S197 13. Diverse diagnostics
S202 14. Gene mutation: the engine behind drug resistance
and TB strain diversity
S206 15. GeneXpert: the nuts and bolts of TB diagnosis
S210 16. Retaining our patients in care
S214 17. Including those affected in care
S219 18. Dont forget us: migrants and indigenous
populations
S223 19. TB in children: epidemiology I
S227 20. TB in children: diagnosis
S232 21. COPD, lung function and pulmonary embolism
S236 22. Smoking cessation
S239 23. Social media campaigns
S242 24. Engagement of CSOs, communities and patients in
TB control and management
S246 25. HIV and lung health: something for everyone
SATURDAY 5 DECEMBER 2015
e-poster sessions
S251 06. Preventing infection and transmission of TB in
health facilities and the community
S255 07. The rainbow session: treatment vaccination,
surveillance or tuberculosis and everything in between
S259 08. High-definition insights into drug resistant
tuberculosis
S263 09. Behind bars: TB in correctional settings II
S266 10. The neglected cousins: trends and treatment of
non-tuberculous mycobacteria
11. Improving TB treatment outcomes
12. MDR-TB: impacts of detection and early treatment
13. LTBI in high- and low-burden settings
14. From specimens to outcome: new information for
molecular diagnosis
S287 15. When waves trump paper: the triumph of etechnology
S291 16. From bench to treatment
S269
S274
S278
S282
S295 17. TB & NCD co-morbidity

S299 18. TB in children: diagnosis, prevention and best
practice
S303 19. Tobacco: economics and illicit trade
S307 20. IAP, pneumonia, asthma and chronic lung disease
S311
S316
S319
S324
S328
S332
S336
S340
S343
S348
S354
S358
S362
S366
S370
S373
S377
S382
S387
S392
S397
S400
S405
S409

26. TART: TB risk and impact of ART
27. Drug resistant TB: outcomes
28. Drug-resistant TB: epidemiology and transmission
29. LTBI testing, perceptions and epidemiology
30. Collaborative services for TB and diabetes
31. Spatial epidemiology of tuberculosis
32. The game of numbers: prevalence and incidence of
tuberculosis
33. Molecules, polymorphisms and resistance to TB
34. Thinking broadly to advance the END TB Strategy
35. Financial impact of TB
36. Molecular diagnostics: pitfalls and improvements
37. Rapid diagnosis the world over
38. Phenotypes and genotypes: deciphering resistance
development
39. Tuberculosis drug resistance and susceptibility
testing in a nutshell
40. From safety to supply-chain: lets talk lab
management
41. Stepping up patient access and support
42. Inside the belly of the beast: TB in correctional
settings
43. The urban-rural divide: different places, similar issues
44. Paediatric TB: MDR
45. Paediatric TB: training, BCG and pharmacokinetics
46. Harmful effects of second hand smoke
47. Monitoring the tobacco industry: promotion and
advertisements
48. Enforcement of smoke-free legislation
49. Strategies to improve TB-HIV screening and service
integration
SUNDAY 6 DECEMBER 2015
S414
S418
S421
S426
S430
e-poster sessions
11. m-health solutions
12. Pneumonia and indoor air pollution
13. Case finding and contact tracing
14. Sniffing out TB: from nose to immunochemistry
15. From TB diagnosis to outcomes, and everything in
between

S434 21. TB epidemiology: post and future insights into case
finding, drug resistance and mortality
S438 22. The nature of the beast: emerging molecular

information about mycobacteria
S442 23. TB in children, including MDR
S447 24. The wide swath of vulnerable populations
S451 25. Engagement of CSOs, communities and patients in
TB management and control
S455 26. Lung health: pneumonia, COPD, asthma and cystic
fibrosis
S459 50. TB tidbits: diagnosis, surgery and outcomes
S464 51. M/XDR treatment: toxicity and stigma
S468 52. Community-based and self-administered
treatment of drug-resistant TB
S473 53. Infection control
S476 54. TB and nutrition
S480 55. Molecular epidemiology and modelling: how can
we break the cycle of transmission?
S485 56. Tbilisi to Lima: a journey through TB treatment
outcomes
S489 57. When the tail wags the dog: drug resistance in TB
programmes
S492 58. Advancing the END TB strategy and other policy
issues
S496 59. Basic science: host
S498 60. Imagining and prediction rules for TB diagnosis
S502 61. Alternative diagnostics for drug resistance
S504 62. Insights into drug-resistant tuberculosis from
Ghana to Peru
S509 63. MDR-TB: detection, prevalence and drivers
S513 64. Patient-centered care and methods to improve
treatment adherence
S518 65. Why are you late? Delay in treatment access
S523 66. Gender issues in TB and tobacco
S528 67. The spectrum of life: miners, drug users, the elderly
S533 68. Paediatric TB: IPT, contact tracing and child
perspectives
S537 69. Paediatric TB: epidemiology II
S542 70. The two Ts: linking TB and tobacco
S546 71. From old to new: chicha to e-cigarettes
S549 72. Holding hands: high-level cooperation on tobacco
control
S554 73. IPT: attitudes, application and adherence
Short oral abstract sessions
S558 01. Across the spectrum of TB management
S564 02. How to reach the END of TB
S570 03. HIV and TB: snapshots
S576 04. TB epidemiology: from Malawi to Taiwan
S582 05. Training for the fight against TB
The Official Journal of the International Union Against Tuberculosis and Lung Disease
Editors-in-Chief
Tuberculosis
Martien Borgdorff, Centers for Disease Control Western Kenya, Kisumu, Kenya
Peter Davies, Consultant Chest Physician, University of Liverpool, Liverpool, UK
Lung Disease
Guy Marks, Woolcock Institute of Medical Research, Sydney, NSW, Australia
Associate Editors
MICHAEL ABRAMSON (Australia)
NADIA AIT-KHALED (Algeria)
ISABELLA ANNESI-MAESANO (France)
NILS BILLO (Switzerland)
TOM BOYLES (South Africa)
KEVIN CAIN (USA)
JOSE CAMINERO (Spain)
KEN CASTRO (USA)
PATRICK CHAULK (USA)
CHEN-YUAN CHIANG (Taiwan)
MIA CRAMPIN (UK)
MASOUD DARA (Denmark)
KEVIN M DE COCK (USA)
KEERTAN DHEDA (South Africa)
ANNE FANNING (Canada)
VICTORINO FARGA (Chile)
GIOVANNI FERRARA (Italy)
JEAN-WILLIAM FITTING (Switzerland)
STEPHEN GILLESPIE (UK)
STEVE GRAHAM (Australia)
ROGELIO HERNANDEZ PANDO

(Mexico)
ANNEKE HESSELING (South Africa)
DAVID HUI (China)
MICHAEL IADEMARCO (USA)
WANIS IBRAHIM (Qatar)
S K JINDAL (India)
PETER KAZEMBE (Malawi)
WON-JUNG KOH (Korea)
UMESH LALLOO (South Africa)
CHRISTOPH LANGE (Germany)
STEPHEN LAWN (UK)
CHI-CHIU LEUNG (China)
KEIR LEWIS (UK)
ROBERT LODDENKEMPER (Germany)
CARL LOMBARD (South Africa)
KNUT LONNROTH
(The Netherlands)
DAVID MANNINO (USA)
GIOVANNI MIGLIORI (Italy)
ELLEN MITCHELL (The Netherlands)

CAROLE MITNICK (USA)
JOHN F MURRAY (USA)
ERIC NUERMBERGER (USA)
ANDREW NUNN (UK)
TOM OTTENHOFF (The Netherlands)
MADHUKAR PAI (Canada)
C N PARAMASIVAN (Switzerland)
ROGELIO PEREZ PADILLA (Mexico)
CHRISTIAN PERRONNE (France)
SHAMIM QAZI (Switzerland)
MARY REICHLER (USA)
RENEE RIDZON (USA)
AKIHIRO SEITA (Egypt)
SARITA SHAH (USA)
TOM SHINNICK (USA)
AKOS SOMOSKOVI (Switzerland)
GIOVANNI SOTGIU (Italy)
CATHY STEIN (USA)
TIM STERLING (USA)

JASON STOUT (USA)
WEI-JUIN SU (Taiwan)
WAN CHENG TAN (Canada)
CHARLES THOEN (USA)
ARNAUD TREBUCQ (France)
MUKUND UPLEKAR (India)
SUSAN VAN DEN HOF
(The Netherlands)
ARMAND VAN DEUN (Belgium)
FRANK VAN LETH (The Netherlands)
ANNELIES VAN RIE (USA)
ANDREW VERNON (USA)
ELSA VILLARINO (USA)
ROBERT J WILKINSON (UK)
PAN-CHYR YANG (Taiwan)
JAI-JOON YIM (Korea)
YING ZHANG (USA)
ALI ZUMLA (UK)
Expert statistical review panel

Larry Moulton (USA), Brian Williams (Switzerland)
Ex-officio members (The Union) President of The Union, E. Jane Carter (USA); Past Editors-in-Chief: Michael Iseman (USA),
Nulda Beyers (South Africa), Moira Chan-Yeung (China), Donald Enarson (Canada),
Wing-Wai Yew (China)
Manuscripts and correspondence
MANAGING EDITOR
SUBMISSIONS COORDINATOR
TECHNICAL EDITOR
EDITORIAL OFFICE
CLARE PIERARD
DIRECTOR OF PUBLICATIONS JOSE LUIS CASTRO
RASHA JERANDI
TECHNICAL EDITOR AND ADVERTISING LOUISE TAYLOR
IRENE ROY
MEMBERSHIP/SUBSCRIPTIONS membership@theunion.org
The International Union Against Tuberculosis and Lung Disease (The Union)
68 boulevard Saint Michel, 75006 Paris, France
Tel: (33 1) 44 32 03 60 Fax: (33 1) 43 29 90 83 e-mail: journal@theunion.org website: www.theunion.org
AIMS AND SCOPE. The International Journal of Tuberculosis and Lung Disease is an official journal of The Union. The Journals
main aim is the continuing education of physicians and other health personnel, and the dissemination of the most up-to-date information in the field of tuberculosis and lung health. It publishes original articles and commissioned reviews not only on the clinical
and biological and epidemiological aspects, but alsoand more importantlyon community aspects: fundamental research and
the elaboration, implementation and assessment of field projects and action programmes for tuberculosis control and the promotion of lung health. The Journal welcomes articles submitted on all aspects of lung health, including public health-related issues
such as training programmes, cost-benefit analysis, legislation, epidemiology, intervention studies and health systems research.
DISCLAIMER.
Any opinions expressed or policies advocated do not necessarily reflect those of The Union.
The International Journal of Tuberculosis and Lung Disease is published monthly by The Union.
Volume 19 (2015). Individual membership: 240E. Electronic membership: low- and low-middle-income countries 20E; highmiddle-income and high-income countries 80E. Institutional subscriptions: 300E. All payments to: Membership Services, The
Union, 68 boulevard Saint Michel, 75006 Paris, France. e-mail: membership@theunion.org. Sample copies (libraries), Missing
issues, Address changes: contact Membership Services.
INSTRUCTIONS TO AUTHORS. Instructions on manuscript submission can be obtained from the Union website www.theunion.org.
ADVERTISING SALES. Contact journal@theunion.org.
EXCESS PAGE CHARGE. All articles over required length will be subject to an excess page charge (see Instructions to authors and
website).
FULL TEXT VERSION ONLINE. The full text of the Journal is published online as of Volume 1, 1997. Free access to back issues.
Access for 2015 is free to Union members and subscribers. Address: www.theunion.org (link) or www.ingentaconnect.com
INDEXING AND ABSTRACTING SERVICES. The Journal is indexed and/or abstracted in the following media: CLOCKSS, Current
Contentsw/Clinical Medicine, Excerpta Medica/EMBASE, the Global Health and CAB Abstracts databases, Index Medicus, ISI
Alerting Services, LOCKSS, Medical Documentation Servicew, Medlars, Medline, the Science Citation Indexw, SciSearchw and
the SIIC databases.
ISSN 1027-3719 Copyright q The Union 2015. All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
without the prior permission of The Union.
SUBSCRIPTION INFORMATION.
*
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper)
INT J TUBERC LUNG DIS 19(12s2):S1S587

Q 2015 The Union
46th World Conference on Lung

Health of the International Union Against Tuberculosis
and Lung Disease (The Union)
Cape Town, South Africa, 26 December 2015
SYMPOSIA: FRIDAY
4 DECEMBER 2015
02. Best practice in management of side

effects among DR-TB patients to improve
quality of care
Addressing side effects of MDR-TB treatment
during ambulatory phase in the community
S Islam,1 M Siddiqui,1 S Reza,1 M Basher,1
M Akramul Islam,1 S Sultana,2 F Khanam3 1Health, BRAC,
Dhaka, 2World Health Organization, Dhaka, 3NTP, Dhaka,
Bangladesh.
Background and Challenges: Drug-resistant TB has

become a significant public health problem that threatens progress made in TB care and control worldwide. In
Bangladesh, the rate of MDR-TB among new cases is
relatively low, based upon the last drug resistant survey
(20102011). The NTP began introducing a strategy of
programmatic management of drug-resistant TB
(PMDT) at community level in 2012. As DR-TB patients
are shifting to the community earlier (between 6 to 8
weeks) with the prescribed injectable agent, local
management of side effects is essential.
Intervention: All MDR-TB patients were hospitalized at
the start of treatment in facilities designed for management of DR-TB (central /regional). Sputum microscopy is
done weekly for admitted patients. If two consecutive
sputum samples become negative, condition is clinically
improved and drug is well tolerated, the MDR-TB
patients are shifted to the community for ambulatory
care. At sub-district level, there is an Outpatient DR-TB
Team formed by Upazilla Health and Family Planning
Officer and other health staff. This team selects a DOTS
provider for the patient. If a patient develops any adverse
effects, the assigned DOT provider contacts immediately
with the team for assistance. All health care providers
related to MDR-TB management receive basic orientation before starting of MDR-TB management at community level.
Results and lessons learnt: Due to the scale up of
community based management of MDR-TB in the
country in 2012, the gap between diagnosed and enrolled

MDR-TB patients under treatment reduced and there
was an improvement of treatment success rate. In 2010,
among the diagnosed cases, 70% of MDR-TB patients
were enrolled in treatment which became 95% in 2014.
The lost-to-follow-up rate is also less (around 1%) in the
ambulatory phase than during the hospital stay (14%). A
total of 140 patient records were reviewed from January
2011 to April 2015 for their side effects status during
treatment and 28 (20%) patients were reported with
joint pain, 23 (16%) with vomiting, 16 (11%) with
hearing disturbance, who were treated accordingly in the
community settings.
Conclusion: During the ambulatory phase, MDR-TB
patients could stay with their family and get adequate
psychosocial support. Side effects during this period
could be managed by a local DR-TB team immediately.
This approach may contribute to minimizing the
challenges related to treatment adherence and ensure a
high treatment success rate for Bangladesh.
03. Pre-approval of access to new TB

drugs: what, where and how
Pre-approval access from the industry
perspective: Janssen
C Kambili 1Janssen Pharmaceutical Companies of J&J,
Raritan, NJ, USA.
Pre-approval/expanded access, also called compassionate use (CU), is the use of an investigational medicinal
product outside of a clinical trial. The willingness by
patients with life-threatening conditions to assume
greater risks (than is ordinarily the case in clinical
practice) in exchange for potential benefits, is the
driving force for CU. Several forms of CU exist, some
of which can provide data for analysis. The following
elements are crucial to implementing pre-approval
access for TB drugs: (1) protect patients from unnecessary harm; (2) minimize the risk of treatment failure and
emergence of resistance; (3) exercise fairness; and (4)
comply with prevailing laws and regulations. Although
fulfilling these tenets should be everyones responsibility,
complying with laws and regulations requires special
attention for drug developers. In the USA, the Food,
S2
Symposium abstracts, Friday, 4 December
Drug, and Cosmetic Act sets out the legal framework

that establishes regulations for access to investigational
treatment for serious diseases, if and when certain
conditions are met. Not every country has such a legal
and regulatory framework. The lack of a harmonized
approach creates challenges for a sponsor wishing to
implement a global CU program. Resources must be
mobilized to ensure that due diligence is done and that
local laws and regulations are adhered to. Clear and
simple guidelines and minimally burdensome requirements coupled with the availability of expert consultation to treating physicians appear to streamline the
review process thus reducing delays in the delivery of
the investigational product to eligible patients. Highlevel political commitment and support are critical.
Dialogue with stakeholders helps clear misunderstandings that can erode trust between the sponsor and the
broader community. Janssens bedaquiline CU program,
which started in late 2011 and ended in September
2015, provided access to over 700 patients from over 40
countries. Treatment outcome data were not collected
on all patients, but physicians did report serious adverse
events, including deaths, to the company. Pre-approval
access should not and cannot be a substitute for product
registration. For diseases of public health importance
such as TB, there must exist push-and-pull mechanisms
that ensure speedy but safe and sustained access to
novel products. Supportive policies must be in place so
that those in need are not denied innovative health
solutions.
04. Outstanding issues in HIV/AIDS

treatment
Pros and cons: empiric TB treatment is critical to
reducing TB mortality among PLHIV but would
adversely affect national TB programmes
F Mavhunga 1National Tuberculosis and Leprosy
Programme, Ministry of Health and Social Services,
Windhoek, Namibia.
Background: Diagnosis and confirmation of tuberculosis

(TB) remains a challenge in many settings, particularly in
settings with high TB-HIV coinfection rates. Consequently a number of patients are commenced on TB
treatment late or not at all, resulting in high mortality
and morbidity. A 2015 meta-analysis (Gupta et al., 2015)
based on post-mortem findings among people with HIV
suggests that TB remained undiagnosed at death in
45.8% (95%CI 32.659.1%) of TB cases. A 2011
systematic review found that 17% (median) of people
starting antiretroviral therapy (ART) developed TB
within one year. Among the suggested interventions for
the management of TB-HIV in high TB burden settings is
empiric TB treatment for all people living with HIV in
high TB-HIV settings.
Potential implications for TB programmes: While the
rationale for empiric TB treatment in PLHIV sounds
plausible, there are a number of potential negative
implications for TB programmes. TB programmes in

most high burden settings are already overburdened and
have limited capacity to provide and support treatment
of diagnosed TB patients. Adding empiric treatment to
this burden is likely to stretch the programmes further
with consequent reduction in overall treatment adherence and increased risk of development and spread of
drug resistance. Lessons from the implementation of IPT
also highlight the challenges with implementation of
such a blanket approach when it comes to ensuring
treatment completion under programmatic conditions.
The other risk associated with empiric treatment is the
consequent under-investment in efforts to diagnose TB in
PLHIV, resulting in patients with conditions mimicking
TB being diagnosed late as a result of the TB label, thus
further adversely affecting treatment outcomes. While
the Reducing Early Mortality and Early Morbidity by
Empiric Tuberculosis Treatment Regimens (REMEMBER) trial (Hosseinipour et al., 2015) has shown no
excess toxicity compared to IPT, it also showed no
mortality benefits with empiric treatment. Thus adding
these patients to the TB treatment cohort might
overestimate TB deaths.
Conclusion: While empiric TB treatment for PLHIV may
theoretically improve outcomes in these patients, recent
evidence appears to suggest otherwise. Additionally the
long term system-wide risk on TB control requires
attention in the context of existing treatment support
systems. Investments should rather focus on improving
screening systems and diagnostics for TB in people with
HIV.
Pros and cons: empiric TB treatment is critical to

reducing TB mortality among PLHIV but would
adversely affect national TB
M Hosseinipour 1UNC Project, Lilongwe, Malawi;
2
University of North Carolina, Chapel Hill, NC, USA.
Strategies for reducing tuberculosis (TB) mortality

among TB-HIV co-infected patients in sub-saharan
Africa are needed. Autopsy studies in both the pre and
post antiretroviral therapy era demonstrate TB to be
either a primary or contributing cause of death in the
majority of cases, and often undiagnosed pre-mortem
highlighting missed opportunities for TB treatment
intervention. Prompt, near co-administration of antiretroviral therapy and TB treatment can reduce mortality,
particularly in those with the lowest CD4 counts (,50
cells) in whom microbiologic confirmation of tuberculosis disease may be difficult. While strategies for
improving diagnosis, including systematic screening
and implementation of GeneXpert, have the potential
to identify tuberculosis earlier, the impact on mortality
has been less than expected. Gaps exist in the continuum
of care for suspected TB patients as they fail to navigate
multiple steps within health systems. One such gap,
between presumed TB and confirmed TB diagnoses and
initiation, can be addressed by the use of Empiric TB
treatment. TB treatment initiation based on clinical
acuity and associated signs and symptoms suggestive of
TB before or without microbiologic evidence can result

in same day initiation. The substantial time savings in
initiating life-saving therapy is particularly relevant when
screening and diagnostic services are limited. Empiric
therapy does not add additional value when TB screening
is well performed and ART and isoniazid preventive
therapy are initiated in a timely fashion (A5274
REMEMBER trial). However, in many settings TB
screening and routine availability of GeneXpert, standardized symptom evaluation, culture and time to spend
with each participant are not readily available and
therefore empiric therapy likely still has advantages
when new rapid diagnostics are not readily available.
Arguments for empiric strategies will be highlighted.
05. Zoonotic TB session I - Trends,

diagnostics and infections models
Zoonotic tuberculosis in India: challenges and
opportunities
C P Churamani,1 A K Rastogi,1 A K Tyagi,2
K P Hanumanthappa,3 A Kumar,3 P S Narwal,4
D S Chauhan,5 V M Katoch6 1Central Military Veterinary
Laboratory, Meerut Cantt, 2Department of Biochemistry,
University of Delhi South Campus, New Delhi, 3Department
of Biotechnology, All India Institute of Medical Sciences, New
Delhi, 4RVC Centre and College, Meerut, 5National JALMA
Institute for Leprosy and other Mycobacterial Diseases,
(ICMR), Dr. M. Miyazaki Marg, Taj Ganj, Agara, 6Department
of Health Research, Director General, Indian Council Medical
Research, Ansari Nagar, New Delhi, India.
A quarter of the global burden of human tuberculosis

occurs in India (Global TB Report 2014, WHO).
Implementation of the Stop TB program has reduced
the rates of incidence, prevalence and mortality, (TB
INDIA 2014). However the occurrence of mycobacterial
pathogens in species other than humans has been
ignored. The capability of pathogenic mycobacteria to
cross the species barrier and sustain transmission among
diverse hosts; the proximity of cattle to humans and
human dependence on animal derived meat and dairy
products thwarts the current preventive strategies. The
estimated livestock population of India is 512 million
(19th Livestock Census2012), while the estimated
human population is 1.2 billion. The precise burden of
tuberculosis among non-human hosts is unknown.
Limited studies have demonstrated the presence of
Mycobacterium bovis and M. tuberculosis in human
and bovine samples, denoting that these mycobacterial
pathogens could potentially infect humans as well as
cattle. Besides the need to recognize the enormity of the
problem, it provides a challenging opportunity to
confront tuberculosis wherever it occurs. Vaccination
of cattle is the practical option that would be acceptable
in the Indian context as opposed to the test and slaughter
policy practiced in several countries. Tuberculosis
research in cattle has been beneficial to humans. Hence
utilizing bovines to test and screen vaccine candidates for
protection against tuberculosis would be apt. The
S3
currently used BCG vaccine has not been successful in

imparting protective immunity to humans and cattle.
Five vaccine candidates heat killed M. indicus pranii, and
recombinant BCG (rBCG) over expressing a-crystallin
(Rv2031c), Ag85C (Rv3803) and DNA vaccines (acrystallin; super oxide dismutase, Rv3846), shown to be
efficient in protection against a challenge of virulent M.
tuberculosis in small experimental animals were assessed
(Khera et al., Vaccine 2005; 23: 56555665; Jain et al.,
PlOS ONE 2008. 3: e3869). The induction of central
memory T cells (TCM), has been considered crucial for
protective immunity (Hope et al., Clin Exp Immunol
2005; 139: 4856). PBMC isolated from the immunized
and control cattle, stained with a panel of monoclonal
antibodies specific for TCM cells (CD4/8/cd, CD45RO,
CCR7, CD62L, IL-2, IFN-c) were flow cytometrically
analyzed. The results showed that TCM cells were
sustained up to a period of 30 months in rBCG 85C
immunized cattle compared to BCG immunization.
Figure The nexusplexus of mycobacterial pathogens:
the challenge
Rapid bTB: a novel lateral flow test able to

differentiate Mycobacterium bovis and
Mycobacterium tuberculosis in sputum
L Stewart,1 I Grant,1 M Sander,2 F Egbe2 1Institute for
Global Food Security, Queens University Belfast, Belfast, UK,
2
Tuberculosis Reference Laboratory - Bamenda, Bamenda,
Cameroon.
Background: A novel lateral flow, immunochromatographic assay (LFD) specific for Mycobacterium bovis,
the cause of bovine tuberculosis and zoonotic TB, was
recently developed at Queens University Belfast. The
LFD detects whole M. bovis cells, in contrast to other
commercially available LFD tests (BD MGITTM TBc ID,
SD Bioline TB Ag MPT 64, Capilia TB-Neo kit) which
detect MPT64 antigen secreted during growth. The new
LFD test has been evaluated in the veterinary context,
and its specificity for M. bovis in the broadest sense (i.e.
subsp. bovis, subsp. caprae and BCG) and sensitivity to
detect M. bovis in positive MGITe liquid cultures was
demonstrated comprehensively.
Methods: Preliminary work was carried out by researchers at Queens University Belfast to optimise sputum
sample preparation, estimate the limit of detection
(LOD) of the LFD with M. bovis-spiked sputum samples,
and check LFD specificity by testing a broad range of
non-tuberculous mycobacteria spp. (NTM) and other
bacterial genera commonly encountered in sputum
samples (Haemophilus, Klebsiella, Pseudomonas, Staph-
S4
ylococcus). In the Cameroon laboratory direct detection

of M. bovis in human sputum was attempted, and 50
positive sputum MGITe cultures and 33 cultures of
various Mycobacterium spp. originally isolated from
human sputum were tested.
Results: Sputum sample preparation consisted of digestion with 1% NALC for 30 min, centrifugation at 3000
xg for 20 min, PBS wash, centrifugation again, and pellet
resuspended in KPL blocking buffer before 100 ll was
applied to the LFD. The LOD of the LFD applied to M.
bovis-spiked sputum was estimated to be 104 CFU/ml. A
small number of confirmed Ziehl-Neelsen 3 M. bovis
positive sputum samples were tested directly but no
positive LFD results were obtained. All of the sputum
MGITe cultures and mycobacterial cultures (including
M. tuberculosis, M. africanum, M. bovis, M. intracellulare, M. scrofulaceum, M. fortuitum, M. peregrinum, M. interjectum) tested LFD negative when read
after 15 min except for the M. bovis cultures, thereby
confirming specificity of LFD for M. bovis in the clinical
microbiology context.
Conclusions: Results indicate that the Rapid-bTB LFD
is a very specific test, able to differentiate M. bovis from
M. tuberculosis, M. africanum, and a range of NTM
isolated from human sputum in MGITTM liquid cultures.
However, the LFD lacks sufficient sensitivity to be
applied earlier in the diagnostic process to directly test
human sputum.
focus on detecting active disease and then culling infected

animals or treating infected humans. Yet, unfortunately,
diagnosis is often delayed or methodologies are not
sufficiently sensitive, and infected individuals continue to
transmit Mycobacterium bovis or M. tuberculosis bacilli
to contacts. The overall goal of our program is to define
transmission dynamics of M. tuberculosis and M. bovis
in animal and human populations and implement novel
control strategies that interrupt the processes. The
practical and logistical challenges and overall high costs
associated with appropriate transmission studies in
relevant hosts; the calf or cow for bovine TB and the
macaque for human TB, have limited our understanding
of the process. Numerous investigators over the years
have shown the ferret to be an appropriate model for the
study of influenza virus transmission, and now, preliminary studies show that the ferret can develop acute
disease with pathology similar to humans and cattle and
transmit M. tuberculosis and M. bovis bacilli animal-toanimal. It is widely recognized that improved understanding of the mechanisms of pathogenesis and host
response to natural infection by M. bovis or M.
tuberculosis are needed in order to develop the most
efficacious vaccines, diagnostics and therapeutics. Ongoing studies may show the ferret to be an acceptable
small animal TB model of natural transmission.
How zoonotic tuberculosis is laying new

ground for approaches in mycobacterial
research in Tanzania
S G Mfinanga,1 E Ngadaya,1 R Kazwala2 1National
Institute for Medical Research, Muhimbili Centre, Dar es
Salaam, 2Sokoine University of Agriculture, Dar es Salaam,
Tanzania.
Ferrets: a small animal model for zoonotic

tuberculosis transmission
F Quinn,1 B He,1 T Ross,1 C C Whalen1,2 1Infectious
Diseases, University of Georgia, Athens, GA, 2Epidemiology
and Biostatistics, University of Georgia, Athens, GA, USA.
Tuberculosis epidemics develop when one case is

replaced by one or more cases among contacts. Case
studies, molecular epidemiology and clinical trials all
agree that transmission in the community is the driving
force behind TB epidemics, and blocking transmission is
the most effective way to intervene in an epidemic and
reduce its impact. Current tuberculosis control strategies
are not designed to interrupt transmission, but instead
Background: During the early 1990s, there were limited

studies on bovine tuberculosis (TB) in the country. The
aim of this article is to highlight various research
approaches used in the country for the epidemiology of
M. bovis, and advise on the way forward.
Intervention: The information reported in this article was
obtained by searching eligible articles published in
scientific journals from 1990 to 2015. The search engines
from PubMed (http://www.ncbi.nlm.nih.gov/pubmed),
Google Scholar (http:// scholar.google.com) and Hinari
(http://www.who.int/hinari/en/) were used and key
words included bovine TB and Tanzania.
Results: International consultation lead by World health
Organization in 1993 addressed the potential threat in
view of limited data for M. bovis in Tanzania and
Zambia. During the 1990s, the approaches for investigations of M. bovis were mainly unilateral, whereby the
veterinary institutions were engaged. Since then various
approaches have been used including south-north bovine
networks, which basically involved veterinary discipline
and thus perceived as a unilateral approach. In early
2000s, the medical institutions joined the networks and
hence advanced the networks into one health approach.
As a result of these approaches, research capacity was
built across east and western African Institutions and
scientific contributions have been published. Studies on

the magnitude of M. bovis in cattle and wildlife carried
out across the country from1996 to 2015 have reported
varying prevalences of 0.2% to 13.3%. In humans, a few
small-scale studies, which mostly involved extra pulmonary samples, reported varying prevalence of 10% to
16%. Diverse RFLP and spoligotype patterns including
the predominance of SB1467, SB2190 and SB0133 have
been observed. Studies have identified risk factors and
demonstrated inadequate public knowledge and perception on bovine TB control. The initial impact of these
studies was a policy change by the National TB Control
Program, where in its earlier manuals, M. bovis was
considered rare but in its new manual the statement has
changed and now M. bovis is considered one of
pathogens in human TB in Tanzania.
Conclusions: One health approach achieved more
capacity development and more quality studies in the
investigation of M. bovis. Prevalence of M. bovis in
humans and animals seems to be low and endemic.
Nevertheless, its potential risk remains a public health
concern; underscoring the need for further research.
S5
abattoirs had positive culture. Furthermore, 10.4% (48/

462) and 42.9% (18/42) individual and herd prevalences
were obtained, respectively. Among LW, 69.2% (184/
266) were from southern Nigeria. Overall, 4.6% (7/152)
of butchers and 6.1% (7/114) of marketers, respectively,
had positive cultures. Additionally, molecular characterization identified 86 strains of Mycobacterium bovis (all
from cattle), seven M. tuberculosis (all from humans) and
21 non-tuberculous mycobacteria (from cattle and
humans). Again, age (P 0.001), educational status (P
0.02) and duration in livestock business (P 0.002)
were identified as important risk factors associated with
TB infection among respondents.
Conclusion: Our findings reveal high burden of bTB
among cattle herds (individual prevalence10.4%; herd
prevalence42.9%). Additionally, isolation of 86 strains
of M. bovis from slaughtered cattle in a setting with poor
knowledge/practices towards bTB, portend major health
risks. Consequently, there is an urgent need for implementation of Public Health policies towards addressing
bovine tuberculosis threat in Nigeria.
Prevalence, molecular epidemiology and risk

factors of tuberculosis infection among cattle
and livestock workers in Nigeria
06. One pilot after another: how can we

move toward a sustainable scale-up of
child lung health interventions?
S Cadmus,1 V Akinseye,1 O Adedipe,1 AKwaghe,2

F Ejeh,3 E Cadmus4 1Veterinary Public Health & Preventive
Medicine, University of Ibadan, Ibadan, 2Veterinary Medicine,
University of Maiduguri, Maiduguri, 3Veterinary Microbiology
and Parasitology, University of Maiduguri, Maiduguri,
4
Preventive Medicine and Primary Care, University of Ibadan,
Ibadan, Nigeria.
Viet Nam: TB CARE I project leading to the

country-wide scale-up of symptom-based
contact screening
Background: Tuberculosis (TB) and bovine TB (bTB)

continue to decimate human and cattle populations in
Africa. Recently, TB infrastructure was scaled up to
control human TB in Nigeria (ranked 13th among TB
burdened nations globally); however, nothing was done
to assuage the zoonotic bTB threat. Bovine TB, a
neglected zoonosis, remains a major public health
challenge because those primarily afflicted [livestock
workers (LW)] are neglected. Hence, need to gather
valuable data on prevalence, molecular epidemiology
and risk factors of TB infection among cattle and LW in
Nigeria with a view to influencing Public Health policy.
Methods: We conducted a cross-sectional study among
cattle and LW (butchers and marketers) between 2013
and 2015 in Nigeria. Animal screening involved post
mortem identification of cattle with suspected tuberculous lesions (TLs) at abattoirs and tuberculin testing of
cattle herds. Sputum samples from humans and TLs were
cultured and subjected to a two-step multiplex PCR
technique based on genus typing and genomic regions of
difference, respectively. Interviewer-administered questionnaire was used to assess LW TB related knowledge,
attitude and practices.
Results: Overall, 3595 cattle (abattoirs3133;
herds462) and 266 LW were screened. More than three
quarters (76.3%) of cattle screened were from northern
Nigeria; and 5.0% (156/3133) of those from the
Background: Viet Nam is a high TB burden country but

accurate data on the burden of TB in children are not
available. The annual risk of TB infection (NTP
prevalence survey in 20062007) was 1.7%. Approximately 351 000 children are estimated to be infected with
TB annually and around 13 000 children are estimated to
develop TB annually. However, the NTP reports only
12001300 child TB cases each year. Further, contact
screening and preventive therapy were introduced as
recommendation only in 2011 and are not widely
implemented.
Objectives: To describe and evaluate the establishment
and development of the child contact screening and
management in Viet Nam.
Methods: Community-based contact screening was
introduced in 4 provinces following training of health
workers and establishment of a recording-reporting
system, in collaboration with the NTP. The WHO
recommended symptom-based screening approach was
implemented focusing on: i) Screen and manage children
that are household contacts of a sputum smear positive
TB case in the community; ii) Provide IPT for eligible
child contacts (aged ,5 years and HIV-infected once TB
excluded) at commune (primary care) level; iii) Introduce
a diagnostic algorithm for diagnosis of TB in children at
the district (second care) level; iv) Develop IEC material;
v) ensuring availability of isoniazid 50 mg; and vi) engage
T H Nguyen 1KNCV Tuberculosis Foundation, Hanoi, Viet

Nam.
S6
the wider health care sector by the NTP strengthening

links and collaborating with the child health sector.
Results: The WHO screening strategy has been implemented in 4 provinces covering a total of 38 districts and
704 communes since quarter 4, 2012. 1578 health care
staff (in NTP, general hospitals and paediatricians and
commune health workers) in the health system have been
trained on the new strategy. In the period quarter 4 2012
to quarter 1 2015, 9741 children that were household
contacts of sputum smear-positive TB patients were
screened and registered. 715 (7.3%) were diagnosed with
TB (all forms). Among 3858 eligible children for IPT,
2514 (65%) children accepted IPT. The IPT completion
rate was 84% among 1,154 children. With success of this
new strategy in 4 pilot provinces, NTP has utilised funds
from the Global Fund to scale up this strategy to 6
provinces for the period 20132014 and to all communes
across the country in 2015. Major challenges have been
clinical diagnosis and decentralisation of diagnosis and
management of TB in children, a lack of diagnostic tools,
the establishment of collaboration between paediatricians and NTP, and awareness of the rationale for
preventive therapy among families and health workers.
Conclusion: Viet Nam is among the first countries
globally to implement and scale up the WHO symptom-based strategy in management of child TB contacts.
Lessons learned will inform improving effective implementation as contact screening and managing is rolled
out in Viet Nam.
07. Role of surgery in management of TB:

a promise or redundancy?
Experience from India: main indications and
realities of thoracic surgery in TB patients
for complications and sequel, 1316 patients were

operated upon for hemoptysis; 32 had undergone
bronchial artery embolization before surgery. 15 cases
were taken up as an emergency and 95 cases were taken
up for chest symptoms other than hemoptysis. Pneumonectomy was done in 524 patients, lobectomy in 710,
bilobectomy in 85, segmental resection in 18 and nonanatomical resections were done in 18 patients. There
were 27 early and 52 late deaths. Most patients became
symptom free. Many different procedures were done for
empyema cases with complete closure of space and
healing of chest wall sinus being achieved in almost all
the cases. Overall, the results of surgery in carefully
selected patients for TB surgery are highly gratifying.
Experience from the Russian Federation:

main risks and benefits of surgical
interventions in TB
R K Dewan 1Thoracic Surgery, National Institute of TB and

Respiratory Diseases, New Delhi, India.
D Krasnov,1 I Felker2 1Surgical, Novosibirsk TB Research

Institute, Novosibirsk, 2International, Novosibirsk TB Research
Institute, Novosibirsk, Russian Federation.
Surgery in TB patients: TB Surgery is a high-cost

intervention with moderate to high risks of complications. The main indications for surgery are a persistent
sputum positive state, recurrent hemoptysis or chest
symptoms, empyema and diagnostic.
Experience of TB Surgery at the NITRD, New Delhi:
India is being presented which includes all the cases
operated upon for Pulmonary Tuberculosis at NITRD
from 1995 to 2014. A total of 107 patients were operated
upon as they were persistently sputum positive. MDR
status was demonstrated in 15 and XDR status in 4.
There were 80 male patients and 27 female patients with
an age range of 21 to 52 years. Pneumonectomy was
carried out in 67 cases, lobectomy in 20, bilobectomy in
5, thoracoplasty in 12 and other lung resections in 3.
There were three early deaths; 98 patients became
sputum negative after surgery. 22 patients became
sputum positive again in follow up. Eight of these died
of progressive disease. 76(71%) remained sputum
negative. 18 were lost to follow up later on. BPF
developed in 8 cases. In the group taken up for surgery
The Novosibirsk TB Research Institute has more than 60

years of experience in the surgical treatment of people
with pulmonary TB. Over 400 surgical interventions for
pulmonary TB are carried out every year. Basically, the
following surgical procedures are performed: lung
resections, pneumonectomies, thoracoplasty, operations
on the bronchi, endobronchial valve (EbV) implantation
etc. When identifying indications and contraindications
for surgical treatment, we largely rely on the consensus
paper titled The role of surgery in the treatment of
pulmonary TB and multidrug- and extensively drugresistant TB and released by the WHO Regional Office
for Europe. All patients undergoing surgery keep
receiving drug therapy according to the susceptibility of
MTB. Most patients who have undergone surgery are
still smear positive for M/XDR, even though they had
had proper chemotherapy for 6 months or longer. We
consider that the first priority for surgery in TB,
especially in the regions with high prevalence of M/
XDR, is to deal with smear positiveness. In Russia 56%
of people with cavernous (CV) TB undergo surgery,
while, in our opinion, this figure should be increased to

20%. In our opinion, an extensive use of surgery in
patients who have long been smear-positive and on
whom drug therapy fails is what both the affected people
and the society as a whole will benefit from at large.
However, patients with M/XDR-TB are the most difficult
for surgery, for the course of TB process in them is often
accompanied by progression, distributed CV with an
ongoing infiltration and the occurrence of pathology as
foci in the bulk of lung tissue, the presence of CV in both
lungs, and specific bronchi lesions. This fact reveals the
main risks of the surgical treatment of TB. The other
major risks of the surgical treatment are high levels of
comorbidity, low functional levels, and unsuccessful
operations on the lungs for TB in the past. Resection
interventions are contraindicated in most of these cases.
A novel technique for performing mini-access extrapleural thoracoplasty with an excellent cosmetic result has
been developed in our clinic for patients with cavernous
TB with a high risk of complications following lung
resections. This procedure has been applied to 414
patients. The clinical effect in the form of bacteriological
conversion and the absence of CV at discharge was
attained in 87% of cases. In most of these patients,
thoracoplasty had been carried out in combination with
EbV.
08. Maternal and infant TB: advancing

our understanding of pathogenesis,
treatment and prevention
Challenges in the management of maternal
and perinatal TB: cases from the field
S Naik 1Obstetrics & Gynecology, B J Government Medical
College, Pune, India.
India accounts for 26% of the worlds burden of

tuberculosis (TB). The clinical and laboratory similarities
between TB and pregnancy pose a diagnostic dilemma.
The data on burden of TB among pregnant women in
India are limited but increased obstetric morbidity and
mortality has been reported. Late diagnosis and delayed
treatment are associated with a nine- fold increase in
preterm labour, poor nutritional status, hyperproteinemia and anemia. Intensified search for TB in pregnancy
is the need of the hour. The BJ Government Medical
College (BJGMC) is affiliated with Sassoon General
Hospitals (SGH) in Pune, India, a tertiary care public
teaching hospital that primarily cares for an urban and
semi-urban underserved population. BJGMC-SGH evaluates approximately 4000 suspected TB cases annually
including varied presentations of TB in pregnancy,
ranging from normal to very atypical presentations. This
presentation will:

Review typical and atypical TB cases in the peripartum

period.
Highlight the challenges of managing TB in peripartum women.

S7
Review adverse outcomes of maternal TB in mothers

and infants.
Discuss critical gaps in maternal-infant TB research.
Current research on the pathophysiology,

immunology, prevention and treatment of TB
in pregnant women
J Mathad,1 A Gupta2 1Weill Cornell Medical College, New
York, NY, 2Johns Hopkins University School of Medicine,
Baltimore, MD, USA.
Tuberculosis (TB) incidence in women peaks during

reproductive age with more than 200 000 pregnant
women diagnosed with active TB disease annually. The
period of elevated risk appears to be highest in the early
postpartum periodwith a twofold increased risk
observed in epidemiologic studies. When TB occurs
during pregnancy and postpartum, it is a major cause of
both maternal and infant morbidity and mortality.
However significant scientific gaps in pathogenesis,
treatment and prevention remain. As a result, national
guidelines for screening and management vary widely,
with a lack of data preventing a global consensus.
Available TB diagnostics do not account for the immune
changes that occur during pregnancy and the postpartum
period. Safety data for TB drugs during pregnancy is
minimal, especially for drug-resistant infections. Case
reports are common but offer little useful guidance.
Practitioners in TB-endemic countries have little evidence
by which to guide their management of TB infection and
disease in pregnancy. This presentation will review: 1)
The pathophysiology of tuberculosis in pregnant and
postpartum women; 2) What is known about the
interaction between immune changes of pregnancy and
the host immune response to Mycobacterium tuberculosis; 3) Current recommendations and ongoing research
on the prevention of TB in pregnant and postpartum
women; and 4) Current recommendations and ongoing
research on the treatment of active TB in pregnant and
postpartum women, including HIV-TB co-infection and
drug-resistant TB.
The pharmacokinetics and treatment of first

line TB drugs for infants
A Bekker,1 H S Schaaf,1 H R Draper,1 L Van Der Laan,1
A Hesseling,1 L Wiesner,2 H McIlleron,2 S Murray3
1
Desmond Tutu TB Centre, Department of Paediatrics and
Child Health, Stellenbosch University, Cape Town, 2Division
of Clinical Pharmacology, University of Cape Town, Cape
Town, South Africa; 3Department of Clinical Research, Global
Alliance for TB Drug Development, New York, NY, USA.
Background: There are limited pharmacokinetic data

available for first-line anti-tuberculosis drugs in infancy
(,12 months of age), a period of rapid growth and
maturation, which potentially influences drug disposition.
Methods: Intensive pharmacokinetic sampling was performed in infants on tuberculosis treatment including
rifampicin, isoniazid, pyrazinamide and ethambutol
S8
using 2010 World Health Organization recommended

doses. Regulatory approved single-drug formulations
were used on the day of pharmacokinetic sampling,
including two rifampicin suspensions. Assays were
conducted using liquid chromatography mass spectrometry; pharmacokinetic parameters were generated using
non-compartmental analysis.
Results: Thirty-nine infants (26 male) were included; 14
(36%) with culture-confirmed tuberculosis. Fifteen
(38%) infants were premature (,37 weeks); 5 (13%)
were HIV-infected. The mean corrected age and weight
was 6.6 months and 6.45 kilograms, respectively. The
mean maximum plasma concentration (Cmax) for
rifampicin, isoniazid, pyrazinamide and ethambutol
were 2.9, 7.92, 41.9 and 1.26 lg/ml, and mean area
under the concentration curve (AUC0-8)12.12, 24.68,
239.4 and 5.09 lg.h/ml, respectively. Adjusting for age
and weight, rifampicin formulation differed for the two
formulations used, Cmax and AUC0-8 (GM/Ratio2.55,
95%CI (1.47-4.41), P 0.001; GM/Ratio2.52, 95%CI
(1.34-4.73), P 0.005, respectively). HIV status was
associated with a lower pyrazinamide Cmax and AUC08 (GM/Ratio0.85, 95%CI (0.75-0.96), P 0.013; GM/
Ratio0.79, 95%CI (0.69-0.90), P , 0.001). No other
important differences were observed due to age, weight,
prematurity, ethnicity or gender; 33 (85%) infants had
favorable tuberculosis treatment outcome.
Conclusions: Isoniazid and pyrazinamide concentrations
compared well to proposed adult target concentrations;
ethambutol concentrations were lower, but equivalent to
other paediatric studies. The low rifampicin concentrations observed in infants require further investigation,
especially with consideration of future treatment shortening trials in children.
reviewing outputs in a TB diagnostics dashboard.

Individuals in Tanzania have been trained and software
installed to enable staff to model for themselves the
impact of alternative diagnostic algorithms. The model
has been used to assess eight different diagnostic
algorithms including those using Xpert MTB/RIF in 10
different diagnostic districts and 5 diagnostic district
types. Incremental cost effectiveness ratios have been
calculated.
Results: Implementation of Xpert MTB/RIF in all
diagnostic districts modelled would be cost effective
and have the greatest benefit, measured in DALYs
averted and cures, but with highest additional health
system costs. Depending on the objective of implementation the priority sequence for implementation of new
diagnostic algorithms across the country varies. The
projections show that same day LED is particularly
effective when funds are tight and reduction in DALYs
and increasing cures is top priority. If more funds are
available full roll-out of Xpert is the preferred strategy
starting with the largest centres and those with high HIV.
If MDR-TB detection is the top priority then LED same
day will not assist greatly and early roll-out of Xpert is
preferred. Similarly if time to start treatment is top
priority then reducing high levels of empirically diagnosis
becomes the priority so implementing Xpert in districts
with a high proportion of smear negative TB should be
considered first.
Conclusion: The analysis suggests that an effective policy
for pulmonary TB diagnostics in Tanzania would be full
roll-out of Xpert MTB/RIF in larger districts with high
HIV first and same day LED microscopy elsewhere. The
Virtual Implementation approach is a useful tool that
should be retained in Tanzania to model new implementations of Xpert at district level and other new tools as
they become available.
09. The cost of TB and lung health

interventions: methodologies and
implications for health systems and
patients
Modelling the cost-effectiveness for health
systems of the Xpert MTB/RIF scale-up by
district profile in Tanzania
I Langley,1 B Doulla2 1Liverpool School of Tropical Medicine,
Liverpool, UK; 2National TB and Leprosy Programme, Dar Es
Salaam, Tanzania.
Background: The computer modelling of patient pathways (known as Virtual Implementation) has been used
to assist national policy makers to assess options for the
scale-up of new diagnostics (including Xpert MTB/RIF)
for pulmonary tuberculosis in Tanzania. This presentation will demonstrate how a new version of the model
designed for local staff from the National Tuberculosis
Programme to use themselves has been employed to
model diagnostic algorithms at district level.
Design/Methods: A model using the Virtual Implementation approach has been developed which is easy to use
and contains an Excel user interface for data input and
Costing the practical approach to lung health

strategy in Malawi
M McHenga,1 E Gama2 1Reach Trust, Lilongwe, Malawi;
2
Liverpool School of Tropical Medicine, Liverpool, UK.
Background and challenges to implementation: Chronic

airway diseases (CADs) pose a serious financial challenge
to both health systems and patients in most developing
countries, particularly in sub-Saharan Africa. A diagnosis
for people with chronic or persistent cough is usually
delayed because of individual and health system barriers.
However, delayed diagnosis and treatment facilitates
further transmission, severity of disease with complications and mortality and also increases the costs incurred
by patients and the health system. To reduce the financial
burden of chronic airway diseases, some countries have
successfully used the practical approach to lung health, a
patient-centred approach for diagnosis and treatment of
common respiratory illnesses in primary healthcare
settings.
Aim: The aim of this study is to assess the cost of
implementing the practical approach (PAL) to lung
health strategy in primary healthcare settings in Malawi.
Intervention or response: A retrospective analysis of
health systems and patient costs. Information on
patients socio-economic status, household assets and
chronic airway diseases related costs will be collected by
administering an adaptation of a questionnaire recommended by the Stop TB partnership. The ingredients
approach will be used to estimate information on health
systems cost and the data will be entered directly into a
standard excel program developed for the costing study.
For analysis, descriptive as well as regression techniques
will be used.
Conclusions and key recommendations: We hope to
demonstrate the cost of implementing the practical
approach to lung health from both health systems and
patient perspective. If the cost of the practical approach
to lung health interventions is deemed reasonable, the
approach could be scaled up to all primary healthcare
centres.
The cost of serious adverse events in MDR-TB:

experience from Ethiopia
M Girma,1 E Gama,2 J Madan,3 I Langley,2 S B Squire2
Addis Ababa Science and Technology University, Addis
Ababa, Ethiopia; 2Centre for Applied Health Research and
Delivery (CAHRD), Liverpool School of Tropical Medicine,
Liverpool, 3Warwick Medical School, University of Warwick,
Coventry, UK.
1
Background: Multi-drug-resistant tuberculosis poses a

big economic burden to both the health system and
patients globally. This burden is exacerbated by serious
adverse events which increase the costs incurred by both
the health system and patients, through prolonged
hospitalisation, additional diagnostic tests, auxiliary
drugs and rehabilitation activities. The aim of this study
was to estimate the costs associated with serious adverse
events during multi-drug-resistant tuberculosis treatment
from both health systems and patient perspectives in
Addis Ababa, Ethiopia.
Methods: A retrospective analysis of health systems and
patient costs related to treatment and management of
serious adverse event. To estimate the costs, the
ingredient approach was used alongside case reference
forms, case notes and receipts from diagnostic suppliers
and hospital records. A total of 44 cases were included in
the study.
Results: Preliminary results show that, over 60% of the
cases involved hospitalisation lasting up to 15 days; 80%
required diagnostic tests, auxiliary drugs were adminis-
S9
tered to 75% and 35% underwent rehabilitative

activities.
Conclusion: While multi-drug resistant tuberculosis
treatment is providing a better solution to a majority of
patients, serious adverse events are increasing the costs
for other groups of patient and pushing them into
poverty. Similarly, hospitalisation, auxiliary drugs, diagnostic test and rehabilitation activities entails that health
system costs are also increasing.
Modelling health system costs to support

decision making for MDR-TB diagnostics in
Cape Town
R Dunbar,1 P Naidoo,1 N Beyers,1 I Langley2 1Desmond
Tutu TB Centre, Cape Town, South Africa; 2Liverpool School
of Tropical Medicine, Liverpool, UK.
Background: Improving multidrug-resistant tuberculosis

(MDR-TB) control requires access to accurate and rapid
diagnostics for drug susceptibility testing. New molecular
diagnostic tests for tuberculosis (TB) such as Xpert MTB/
RIF (Xpert) hold the promise of improving TB and
multidrug-resistant (MDR) TB diagnosis. Whilst accurate
data are readily available from laboratory and demonstration studies, many operational questions remain and
are most readily answered through modelling.
Methods: South Africa moved from a smear and culture
based algorithm for diagnosis of TB to an Xpert-based
algorithm. The Witness package was used to model all
aspects of both algorithms to reflect the diagnostic
pathways, including the laboratory procedures, as close
to reality as possible, based on information and data
collected during the PROVE IT (Policy Relevant Outcomes from Validating Evidence on ImpacT) study.
Routine operational data was used to validate the model.
The model was used to assess the effect of operational
changes on yield and laboratory costs.
Results: Routine data showed a declining trend in the
yield of TB cases identified over time in both algorithms.
A binomial regression analysis showed no significant
difference in yield between the algorithms when the time
effect was taken into consideration, with a risk difference
of 0.2% (95%CI 1.8% to 2.3%) (P 0.818) in the
Xpert-based algorithm. Total TB diagnostic costs were
increased by 43% from $440 967 in the smear-culturebased algorithm to $632 262 in the Xpert-based
algorithm. The cost per patient diagnosed with TB
increased by 158% from $49 in the smear/culture-based
algorithm to $125 in the Xpert-based algorithm. The
total cost per MDR-TB case diagnosed was decreased by
3%, from $205 in the smear-culture-based algorithm to
$199 in the Xpert-based algorithm. The model was used
to demonstrate the effect of an optimised algorithm,
increased case-finding and other operational changes on
TB yield and laboratory costs.
Conclusion: This study used a combination of operational data and discrete event simulation to answer
important operational questions that could not be readily
addressed through other methods. Modelling is an
attractive and viable option to guide policy makers in
S10
making decisions about the implemention of new

diagnostic tools and algorithms.
10. Research is needed to increase

childrens access to drug-resistant TB care
Consensus research priorities on children with
drug-resistant TB
B Velayutham,1 D Nair,1 S Ramalingam,1
S Swaminathan,1 C Perez-Velez,2 M Becerra3 1National
Institute for Research in TB, Chennai, India; 2Banner Good
Samaritan Medical Center, University of Arizona College of
Medicine, Phoenix, AZ, 3Department of Global Health and
Social Medicine, Harvard Medical School, Boston, MA, USA.
Background: Numerous knowledge gaps hamper prevention and treatment of childhood drug-resistant
tuberculosis. Identification of research priorities is vital
to fill these gaps and develop strategies to address this
neglected but important problem.
Methods: To identify and rank research priorities in
childhood drug-resistant TB, the Child Health and
Nutrition Research Initiative methodology was adapted.
Research questions related to Epidemiology, Diagnosis,
Treatment and Prevention were framed and the questions
were classified according to research type: Descriptive,
Development and Discovery. Respondents with expertise
in childhood drug-resistant TB were invited to score
these questions using an online survey.
Findings: The top-ranked research question was to
identify the best combination of existing diagnostic tools
for early diagnosis of drug-resistant TB in children.
Highly ranked treatment-related questions centered on
reasons for poor treatment outcomes, adverse effects of
anti-TB drugs, optimal treatment duration and interventions for improving treatment outcomes. The prevalence
of drug-resistant TB was the highest-ranked question in
the epidemiology area. The development type questions
ranked highest focused on interventions for optimal
diagnosis, treatment and modalities for treatment
delivery. The predominant discovery type questions
focused on new drug evaluation and models for
preventive therapy and for preventing new infections.
Interpretation: This was an effort to establish research
priorities for this neglected childhood infectious disease
through a systematic global survey of stakeholders. The
consensus research priority questions identified in this
exercise can serve as a framework that can be updated
periodically and as a shared resource to guide new
research to improve the prevention and treatment of
drug-resistant TB in children.
From estimates to action: setting screening and

treatment targets for children with TB
C Yuen 1Division of Global Health Equity, Brigham and
Womens Hospital, Boston, MA, USA.
Around 1 million children fall sick with tuberculosis each

year, but only 300 000 pediatric cases were notified in
2013. Household contact investigations provide one

method for programs to find children with tuberculosis.
Children living in the households of adult tuberculosis
patients are at high risk for tuberculosis disease and
infection, and they are readily accessible since the adult
patients are already known to the health system. To
facilitate planning and resource allocation, tuberculosis
programs would benefit from estimates of the numbers
of children they should expect to evaluate and treat for
disease or infection if they were to perform household
contact investigations around all newly diagnosed adult
tuberculosis patients. Using an example from Malawi,
we present a method that can be used to generate
national and regional household evaluation and treatment targets based on locally available adult pulmonary
case notifications, census data, and HIV prevalence
estimates, as well as disease and infection risk estimates
from the literature. Nationwide, during a 1-year period,
we predict that around 21 000 child household contacts
will require evaluation, of whom almost 2000 will
require treatment for tuberculosis disease and 7800 will
require isoniazid preventive therapy. We use a generalized version of this method to generate evaluation and
treatment targets for all countries based on case
notifications, World Bank indicators, and Demographic
Health Survey data. We estimate that globally, 7 million
children living in households with known adult pulmonary tuberculosis patients require evaluation each year,
and that over 623 000 cases should be detected among
these children. In 24% of countries in 2013, the number
of notified pediatric tuberculosis cases was less than 20%
of the number predicted. Methods for generating
evaluation and treatment targets for children exposed
at home to tuberculosis can help programs plan and
monitor activities designed to find and treat these
children.
A bitter pill to swallow: developing paediatric

formulations of second-line drugs
J Furin 1Harvard Medical School, Boston, MA, USA.
The large and growing access gap between the number of

children who become sick with drug-resistant tuberculosis (DR-TB) and those who are treated for the disease
each year represents a significant health systems failure.
While there are multiple reasons why children with DRTB are not diagnosed and treated for their disease, one
serious challenge is the medications used to treat the
disease. This presentation will discuss some of the
problems with the existing second-line medications.
Challenges, including perceptions that the drugs are
more dangerous than the disease, lack of proper dosing
recommendations and formulations, and the high costs
of current therapy, all contribute to a perverse situation
in which the most vulnerable pediatric patients are
provided a lower standard of care. This situation can be
reversed with novel partnerships and training models,
pharmacokinetic studies of the relevant drugs, increased
collaboration, and dedicated funding all grounded in a
rights-based approach to DR-TB in children.
11. TB surveillance in health care

workers: challenges and approaches in
high-burden, low-resource settings
Challenges in implementing TB screening for
health care workers in Zimbabwe
A Maruta,1 V Robertson,1 S Balachandra2 1Zimbabwe
Infection Prevention and Control Project, Harare, 2US Centers
for Disease Control and Prevention, Harare, Zimbabwe.
Background and challenges to implementation: Zimbabwe continues to be severely burdened by both TB and
HIV with a prevalence of 0.43% and 14.7%, respectively. The pressure of disease burden on an inadequately
staffed and underfunded health service has resulted in
overcrowded facilities and serious training gaps for
healthcare workers (HCWs). A 2010 health facilities
survey showed that only 21% had an infection prevention and control (IPC) training program and 12% had
infection control plans. Though 96% had an IPC focal
person, only one had formal training in IPC, and two of
27 facilities had an annual TB screening program for
HCWs.
Intervention: ZIPCOP, an IPC project established in
2011, has included TB infection control in its IPC
program. The intervention includes training in TBIC
(including HCW screening) and mentorship. A standardized checklist is used at each visit to gather information
about each facilitys TBIC activities. In the first three
years, the program prioritized 60 sites across Zimbabwe.
Results and lessons learnt: Among the 60 priority sites,
13 implemented HCW TB screening; 10 used a screening
tool and 3 used radiography. Barriers to TB screening
included HCW reluctance based on fear, stigmatization,
and lack of confidentiality. There is no systematic data
collection of TB incidence among HCWs.
Conclusions: Given the higher risk of acquiring TB,
regular HCW screening should be a key component of
facility-based TB IPC programs and could serve as a
proxy measure of effectiveness in reducing nosocomial
transmission. With ZIPCOPs support, HCW screening
has been incorporated in the new National IPC Policy
and Strategic Plan and the National TB Program is
introducing a quarterly TB surveillance tool for HCWs.
However, the introduction of a policy and data collection
tool may not be sufficient without a more comprehensive
approach. Based on a literature review and the above
results, we propose: an accountability framework with
measures improving confidentiality, increased sensitization, strengthening occupational health systems and use
of improved diagnostic tools.
Current state of TB surveillance among health

care workers in China
C Xu 1National Center for Tuberculosis, China CDC, Beijing,
China.
In China, tuberculosis (TB) ranks as the leading cause of

death from infectious disease, with over 1.3 million
people developing active TB each year. TB infection
S11
control, however, is still at an early stage in China.

Several studies have found that healthcare workers
(HCW) in China have higher rates of latent TB infection
than the general population. Even though HCWs are
considered as a high risk group for contracting TB, many
Chinese healthcare facilities lack the policies and
resources for implementing effective TB control programs, such as medical surveillance. There is currently no
national policy for screening or surveillance for TB
among HCWs. Latent TB infection is not treated and,
often, TB is not recognized an as occupationallyacquired disease. Operational research is necessary to
evaluate the costs and benefits of promoting medical
surveillance for TB among Chinese healthcare workers.
Screening and treatment for latent tuberculosis

infection in health workers at Maputo Central
Hospital in Mozambique
E Nunes,1 A Hassane,1 S Amade,1 S Peleve,1 O Augusto,2
S Viegas,2 F Torriani,3 S Graves3 1Maputo Central Hospital,
Maputo, 2National Institute of Health, Maputo,
Mozambique; 3University of California, San Diego, CA, USA.
Background: In Mozambican hospitals, the high burden

of tuberculosis (TB), and limited infection control put
health workers (HW) at risk for TB. We aimed to
evaluate operating characteristics of tuberculin skin
testing (TST) and Quantiferon (QFT), establish the
prevalence latent TB (LTBI) and incidence of active TB
among HW, and assess feasibility and tolerability of
isoniazid preventive therapy (IPT).
Design/Methods: 690 enrolled HW completed symptom
screening, chest X-rays, HIV and tuberculin skin testing
(TST). Those without active TB and TST . 10 mm
(.5mm if HIV) underwent QFT testing. Logistic
multivariate analysis was used and adjusted odds ratios
of LTBI reported. HW with LTBI were risk-stratified by
HIV status and quantitative TST and QFT. Those with
HIV or TST . 15mm and QFT . 1.0 were classified as
high risk and offered IPT. Adherence and adverse
reactions are reported. The symptom screen was repeated
at one year.
Results: Among 690 HW, 4 active TB cases were
diagnosed and 12% were HIV. The positive predictive
value of TST for QFT positivity is 82% in HIV
individuals (95%CI 63-94), not significantly different
from HIV- (79%, 95%CI 74-84). We observed highest
concordance in TST/QFT when TST was . 10 mm. In the
multivariate analysis, duration of service .10 years was
significantly positively associated with LTBI (OR 1.67,
95%CI 1.21-2.30), as was service in the departments of
Surgery, Emergency and Critical Care, Pathology, and
Obstetrics (P , 0.05 for all) when compared with the
Internal Medicine department. 425 (62%) workers were
diagnosed with LTBI of whom 284 (67%) were classified
as high-risk LTBI. Of 333 workers with HIV or high-risk
LTBI counseled to take IPT, 183 completed 6 months, 48
abandoned treatment and five discontinued therapy due
to side effects (rash, gastrointestinal symptoms); no cases
of hepatitis occurred. At one year follow-up, two HW
S12
had developed active TB. Both of these were on IPT at the

time of diagnosis. Due to loss-to-follow-up, only 502.1
person-years were observed. Overall active TB incidence
was 398/100 000 person-years (95%CI 48.2-1438).
Conclusion: Active TB incidence among HW at MCH is
similar to background incidence in Mozambique: 552/
100 000 (WHO 2014). Several departments have higher
LTBI prevalence, suggesting occupational risk and
targets for intervention. QTF and TST results were
moderately concordant. IPT was well-tolerated, however
attrition was high.
URC teams visited these facilities providing information

and education on benefits of taking IPT. The team held
discussions where HCWs voiced their perceptions and
concerns about IPT. Data collection tools were distributed to all the project sites.
Results: Out of the 3107 targeted health care workers
188 (6%) were enrolled on IPT, of the 188, 52 (28%) of
them are HIV-positive, 154 (82%) completed IPT, 3
(2%) transferred out, 9 (5%) defaulted and chose not to
continue due to minor side effects. Three (2%) were
discontinued due to minor persistent side effects and 19
(10%) are still on treatment. HCWs had several concerns
about taking IPT such as side effects caused by INH and
INH stock outs nationally.
Conclusion: TB is an important public health issue,
especially in Swaziland where HCWs are exposed to a
high burden of TB and infection prevention and control
measures are inadequate. The increased uptake of IPT,
intensified case finding through TB screening and
strengthening of infection control measures can reduce
infection among HCWs.
12. Managing pharmacovigilance

systems: optimising the rational use of
new TB medicines and novel regimens
The provision of isoniazid prophylaxis therapy

(IPT) to health care workers: a Swaziland
experience
P Mamba 1Swaziland Wellness Centre, Manzini, Swaziland.
Background: Swaziland has the highest per capita HIV

prevalence in the world as well as a high burden of TB.
The Swaziland Nurses Association established a nursemanaged wellness centre in the country in collaboration
with the International Council of Nurses in 2006 to
provide confidential and comprehensive health service
delivery to health care workers (HCWs) and their
families. From October 2012 to September 2013, the
Swaziland Wellness Centre screened HCWs for TB using
nurse coordinated wellness corners that were established
in 14 sites. A total of 3107 HCWs were targeted to be
screened; 2391 (77%) were screened, 746 (31%) were
presumptive cases and 31 (1%) were confirmed to have
active TB and enrolled on treatment. It was then that a
need to strengthen the 3 Is was noted; this included
strengthening the isoniazid preventive therapy (IPT)
uptake among all HCWs with emphasis to HIV-positive
HCWs.
Intervention: In the 14 sites where TB screening was
conducted, all HCWs were sensitized and trained on IPT
to increase awareness and uptake from January 2014 to
March 2015. Existing structures were used for sensitization; departmental meeting, in-service trainings and
workshops. The wellness Corner nurses were trained
on provision and management of IPT, follow up and
documentation. The Swaziland Wellness Center and
Cohort event monitoring methodology: data

collection, indicators and drug safety profiles
D Falzon 1Global TB Programme, World Health
Organization, Geneva, Switzerland.
Drug-related harms can lead to a tuberculosis (TB)

patient interrupting treatment before completion and can
thus contribute to chronicity, death, reduced quality of
life, treatment failure, transmission of infection, or
emergence of drug-resistance. It is thus important that
adverse events (AEs) especially serious events (SAEs)
be monitored and addressed in TB patients on
treatment. National TB programmes which systematically monitor harms associated with anti-TB medication
are better placed to achieve good outcomes when
introducing more complicated regimens, such as those
for multidrug-resistant TB (MDR-TB) and TB-HIV. The
progressive global scale-up of treatment for MDR-TB
will see more populations with a diverse mix of age,
ethnic profile and comorbidity patterns receiving combinations of second-line TB drugs, some of which are
new or repurposed. This is a welcome development and
brings with it an increased likelihood of cure for many of
these patients. However, vigilance for adverse drug
reactions (ADRs) and drug-drug interactionssome of
which as yet unrecognizedwill be an important
complement. Active pharmacovigilance is currently
recommended by WHO when new drugs such as
bedaquiline and delamanid are used as part of the
treatment of patients with MDR-TB. Cohort-based
approaches to monitor actively for AEs can provide data
for the dual purposes of protecting the individual patient
and informing local and global policy. Dialogue with
different stakeholders, including drug developers, regulators and national drug-safety authorities, will be
needed to ensure appropriate mechanisms are in place
to report safety concerns alongside the monitoring of
response to treatment. This presentation will outline the
approach of the WHO Global TB Programme to active
TB patient drug-safety management and monitoring
based on standard methods for the surveillance of
drug-safety concerns (Figure). These methods and their
adaptations for the context of national TB programmes
reflect discussions held with technical and funding
partners as well as national authorities in recent years
on the practicalities of implementing functional monitoring systems under programmatic conditions. The
presentation will use lessons learnt and best practices
from countries to illustrate how common challenges in
data collection and organisation, generation of indicators and the analysis for causality and signal detection
were addressed.
S13
regimens (e.g., short regimens) for the treatment of

drug-resistant tuberculosis. The purpose of the active
drug safety component is to:
a Ensure TB patient safety through the adequate
monitoring and management of adverse events
b Identify inappropriate medicine use
c Report adverse events for new drugs and novel
regimens from broader clinical practice for more
populations (e.g., with diverse age, sex, and ethnic
profiles)
d Provide drug safety data for analysis to guide National
TB Programs decisions for improving patient safety
e Report adverse events, drug-drug, and drug-disease
interactions for all TB regimens to inform global
policies on their use
Inadequate active safety monitoring and management
could cause unnecessary harm to patients, lead to poor
adherence, and amplify drug resistance. However, few
resource-limited countries have all the structures, systems, or resources necessary to support active drug safety
activities. This talk will focus on approaches for
developing or strengthening systems for active MDRTB drug safety monitoring. It will also provide an
overview on the importance of NTPs coordinating with
National Pharmacovigilance Programs to foster a culture
of adverse events reporting and knowledge sharing at all
levels to ensure the best possible clinical outcomes for all
patients. An update on the bedaquiline donation
program and technical assistance available through
USAID and partners for developing or strengthening
systems for active MDR-TB drug safety monitoring will
also be discussed.
13. Can mobile technology improve lung

health in poorer countries and
communities? Benefits and opportunities
Russias experience in using mobile technology
to educate, monitor and enforce compliance
with tobacco control legislation
I Berezhnova,1 E Zeynalova2 1International Union Against
Tuberculosis and Lung Disease, Moscow, 2Centre for Public
Health Cooperation, Krasnoyarsk, Russian Federation.
Strengthening pharmacovigilance systems:

improving TB patient safety and reducing the
adverse drug reaction burden
A Kwiecien,1 K Kakanda1 1USAIDs Systems for Improved
Access to Pharmaceuticals and Services Program, Arlington,
VA, USA.
Active pharmacovigilance (drug safety) is one of the five

conditions set out in the Companion handbook to the
WHO guidelines for the programmatic management of
drug-resistant tuberculosis for introducing new medicines (e.g., bedaquiline and delamanid) and novel
Background: In 2013, Russia adopted one of the world

best Tobacco Control legislations, which fully bans
smoking in all public places. The legislation came into
force in June 2013, but the government and civil society
have continued developing various mechanisms to
improve the enforcement of the law and to involve the
population in compliance monitoring. Mobile applications are one of the innovation technologies used by the
Ministry of Health and its Tobacco Control allies in
putting the law implementation forward.
Intervention: In October 2014, the Ministry of Health
launched a mobile app Iefs: of lur>t (No smoking
here). Several other agencies developed similar applications, including Tatarstan, which launched its regional
S14
mobile application incorporating tobacco control among

information on other public health policies. One of The
Union grantees developed the Smoke-free Russia
mobile application, which in addition to compliance
monitoring, provided information on the law, penalties,
etc. The Ministry of Healths mobile app is currently
available for the general public, and could be downloaded from the website and by searching in the apps store. It
is an interactive program that accepts comments and
suggestions for the applications further improvement. It
has also been developed on the basis of other existing
applications to create a more efficient for the compliance
monitoring tool.
Results: The app was developed for Android and iOS and
meets the following objectives: to provide an efficient
mechanism to register the law violations, to report to the
enforcement agencies, to cumulate data about number,
locations of the violations. It includes a brief description
of the law provisions, proposes a complaint form, and its
tracking system. The mobile apps allows to quickly and
effectively prepare the complaint, attach a photo of a
case, automatically determine the enforcement agency
which deals with each particular case, call the police if
necessary, call a quit hotline if required, and repeat the
text if needed.
Conclusions: The mobile app No smoking here is
provided to be a useful tool to report about noncompliant cases, help citizens to familiarize themselves
with the law. The Smoke Free Russia application was
used in a few regions to conduct compliance monitoring.
In Orenburg, it allowed volunteers to report and
highlight violations on a digital map in 100 public
places. Overall, the applications do not duplicate each
other, but cover specific areas, which make the applications user friendly and efficient.
Using mobile technology successfully for early

diagnosis of TB in a rural setting
P Banuru Muralidhara,1 S Chadha,1 A Trivedi,1 V Rajput,1
D Tamang,1 R J Singh,1 R Shekhar,2 L Olson2 1International
Union Against Tuberculosis and Lung Diseases, South-East
Asia Office, New Delhi, 2Dimagi Inc., New Delhi, India.
Background: Reaching the Unreached is the principle of

global TB control strategies. Rural areas of India, like
any developing country, have difficulty in accessing the
services for early diagnosis. In these areas, healthcare is
provided by rural health-care providers (RHCPs).
RHCPs, are first point of contact for many illness and
refer TB symptomatics. The outcome of referrals made
by RHCPs to nearest sputum testing center is not known.
We therefore developed a mobile based application with
an aim to track referrals from RHCPs for early diagnosis
of TB.
Intervention: With support from Dimagi, two ComCare
mobile platform applications were developed: one
application exclusively for RHCPs, and the other for
laboratory technicians (LT) at respective sputum testing
centers in the intervention district of Hazaribagh, Jharkhand, India. Hands-on training was provided to RHCPs/
LTs about the use of the application for referrals. In

addition, the RHCPs were oriented on the DOTS
module. Following referral of identified symptomatics
by RHCP to the nearest sputum examination center,
basic information is entered into the CommCare RHCP
application. This information is received on the LT
mobile application through a central server. When the
symptomatic reaches the DMC, the LT conducts a
sputum examination (2 samples) and the results are
entered into the application. The results then show-up on
the RHCP application against the name of the referral. In
case the sputum result is positive, the patient is initiated
on DOTS and the RHCP referred would have the highest
probability of being a DOTS provider. Information at
regular intervals about DOTS is entered until the
completion of treatment. The process is monitored on a
real-time basis.
Results: Out of 50 trained RHCPs (phased manner), 29
RHCPs are regularly using the application for referrals.
Nearly 66 referrals are made by RHCPs over a period of
9 months and 45% were found to be sputum smearpositive. Among the RHCPs, 12 are currently providing
DOTS for 22 TB patients and monitored regularly by
programme and project.
Conclusion: Use of a mobile based application among
RHCPs in rural and remote areas for early referrals and
follow-up highlight positive results. Continued monitoring with RHCPs and LTs is needed to ensure data entry is
regularly made. An incentive based approach for
referrals as per programme guidelines is being reviewed
for increasing referrals and ensuring early diagnosis.
14. Spelunking for treasure: searching for

candidate biomarkers as prognostic
indicators for treatment outcome
Potential and limitations of current candidate
host biomarkers for treatment response
G Walzl,1 K Ronacher,1 S Malherbe,1 D Alland,2 D Zak,3
G Schoolnik,4 G Dolganov,4 J Winter5 1Stellenbosch
University Immunology Research Group, DST/NRF Centre of
Excellence in TB Biomedical Research, MRC Centre for TB
Research, Cape Town, South Africa; 2Rutgers Biomedical and
Health Sciences, Newark, NJ, 3Center for Infectious Disease
Research, Seattle, WA, 4Stanford University, Stanford, CA,
5
Catalysis Foundation for Health, Emeryville, CA, USA.
Biomarkers for TB treatment response and outcome

would facilitate shortening of clinical trials of new antiTB drugs. We have investigated clinical, microbiological,
imaging and immunological parameters and their changes during treatment of pulmonary TB patients to define
biosignatures that are associated with different treatment
phases and outcomes and to explore the feasibility of
predictive models for poor treatment outcomes. Although clinical and microbiological markers represent
important predictors for poor outcomes, they are not
sufficient on their own. Whole blood transcriptomic and
immunological markers in the serum of TB patients
change with distinct kinetics during treatment and cured,

failed and relapse outcomes exhibit different response
patterns. Promising predictive abilities of multi-marker
models for failed treatment and relapse are found in
small cohorts but require prospective testing. [18F]Fluoro-2-deoxy-D-glucose positron emission tomography/computer tomography (FDG-PET/CT) imaging was
evaluated for predictive ability in 100 TB patients from
pre-treatment until one year after completion of therapy.
End of treatment imaging parameters suggest that there
is significant ongoing inflammation even in the absence
of detectable Mycobaterium tuberculosis in sputum and
such ongoing inflammation may indicate subclinical TB
infection and non-sterilizing cure with a high risk for
relapse. The correlation between needs to be investigated
in future studies. Together these studies point towards
the importance of combined bacterial and host-based
biomarker research to predict the outcomes of anti-TB
chemotherapeutic therapy.
Prospective T-cell activation markers for

monitoring treatment and predicting cure
C Geldmacher,1,2 N Heinrich,1,2 A Rachow,1,2
M Hoelscher,1,2 E Ntinginya,3 M J Boeree,3,4 G Kibiki,4
H Semvua4 1Infectious Diseases and Tropical Medicine,
Medical Centre of the University of Munich, Munich,
2
German Centre for Infection Research (DZIF), Partner Site
Munich, Munich, Germany; 3NIMR-Mbeya Medical Research
Center, Mbeya, 4Kilimanjaro Clinical Research Institute,
Moshi, Tanzania.
On behalf of the PanACEA consortium; other institutions involved: Radboud University Medical Center,
Nijmegen, University Centre for Chronic Diseases,
Dekkerswald, Groesbeek, Netherlands. Identification of
predictive diagnostic markers of tuberculosis (TB)
treatment outcome would in theory allow to tailor
treatment duration to individual needs and to identify
patients in need of a longer or different course of
treatment. Recent advances in the field of blood-based
biomarkers demonstrate that novel immunodiagnostic
approaches can improve the diagnosis of active TB in
adults and children. These flow cytometry based
approaches detect and characterize Mycobacterium
tuberculosis -specific Interferon gamma positive T cells
after short term in vitro restimulation of whole blood or
Peripheral Blood Mononuclear Cells with M. tuberculosis-specific antigens. The phenotypic characteristics of
M. tuberculosis-specific T cells can differentiate with
high accuracy between latent and active tuberculosis and
hence mirror at least partially Mycobacterial activity in
vivo. Furthermore, recent research shows that expression
of these T cell activation and differentiation markers on
M. tuberculosis-specific T cells changes after TB
treatment initiation and hence this novel immunodiagnostic approach could potentially help to monitor TB
treatment, to predict TB treatment outcome and as a
consequence to individualize TB therapy. This presentation will provide an overview on the current state of
research in this field and the most recent results from our
own studies.
S15
Use of the simple biomarker suPAR to assess

severity of TB disease and to monitor
treatment
C Wejse1,2 1Department of Public Health, Aarhus University,
Aarhus, 2Department of Infectious Diseases, Aarhus
University Hospital, Aarhus, Denmark.
White blood cells, as well as endothelial cells, express the

urokinase plasminogen activator receptor (uPAR). The
soluble form of the receptor, suPAR, is cleaved from the
cell surface during periods of inflammation and increased
inflammation secondary to activation of the immune
system thereby produces increased concentrations of
suPAR in body fluids. Elevated suPAR levels are
prognostic of a poor outcome in several infectious
diseases including tuberculosis and HIV. It is not specific
for a particular disease, and therefore not a useful
diagnostic marker, but has shown a good prognostic
utility. However, in combination with a clinical score, the
Bandim TB score, high suPAR levels have been associated
with a high probability of a diagnosis of active TB among
presumptive TB patients. The suPAR levels have been
shown to be elevated in patients with active TB
compared with TB-negative individuals and were highest
in smear positive (3.17 ng/ml), followed by smear
negative but culture positive (2.41 ng/ml) and by patients
diagnosed on clinical grounds (2.13) compared with TBnegative individuals (1.73 ng/ml). After treatment,
suPAR levels had decreased to the levels of TB-negative
individuals. Mortality during treatment has been shown
to be higher in TB patients with high suPAR, as well as
among presumptive TB patients not diagnosed with TB.
Increase in suPAR at 1 and 2 months of treatment has
been associated with a Mortality Rate Ratios (MRR) of
4.5 and 2.1, respectively, for the remaining treatment
period. Plasma suPAR levels have also shown to be
higher in MDR-TB patients with relapse and predicted
therapy failure in this group better than ESR. The data on
the use of suPAR to assess severity of disease and monitor
treatment are limited and not an established part of
clinical management, but hold some promise of becoming a candidate marker for treatment outcome.
15. TB-HIV epidemiology within the

context of national TB prevalence
surveys in Africa
Progress of national TB prevalence surveys in
Africa, and the burden of TB-HIV co-infection
among prevalent TB cases from these surveys
I Onozaki 1World Health Organization, Geneva,
Switzerland.
Background: Quite a few Asian countries carried out a

series of nationwide tuberculosis (TB) prevalence surveys
with chest X-ray screening and bacteriological confirmation with culture along the WHO guidelines published in 1958. However, such a survey had never been
carried out on the African continent for five decades until
S16
Ethiopia completed the first national survey in 2011.

Since then, 11 African countries completed a national
survey. In many countries, national TB prevalence
surveys are the only way to reliably measure the burden
of TB disease. These surveys could also provide evidence
about the current performance of TB care and control
and how this could be improved. However, carrying out
HIV testing in the TB prevalence survey to learn about
the TB-HIV situation in the community is still a challenge
in most countries. Out of 11 countries, only Zambia
offered HIV testing to all survey participants, and three
surveys, namely Rwanda, Tanzania and Uganda, covered
TB screening positive, while other countries tried to
determine HIV status of the survey participants and TB
patients through routine HIV and TB services.
Major findings: Observed prevalence of bacteriologically
positive TB is often higher than previous estimates. The
TB prevalence survey always detects bacteriologically
positive TB patients more than annually notified. Smearpositive TB accounts for only 30-40% of bacteriologically positive TB cases when culture is done properly.
Males show significantly higher prevalence than females.
The older the age group, the higher the prevalence is in
most countries. However, the majority of TB patients are
still in the reproductive age group in most countries.
Many countries show higher prevalence in urban clusters
despite their better access to medical services including
hospitals. Although available data are limited, the
prevalence of HIV-positive bacteriologically positive TB
seems to be declining. Most people in community know
their HIV status in sub-saharan African countries. In
these countries, HIV-negative TB is under-estimated.
AFB smear-positive does not mean TB. Introduction of
quality culture, Xpert and chest X-ray have proven that
half or more participants with a smear-positive condition
may not have TB disease in some countries of the TB
prevalence survey. In this symposium, we will learn from
the experiences of three most recent national surveys in
high TB-HIV burden countries in Africa.
16. Scaling up MDR-TB services in

resource-limited settings: progress and
challenges from the field
Rapid scale-up of MDR-TB services through a
decentralised patient management system in
Ethiopia
M Aseresa Melese,1 D Habte,1 Y Molla,1 P G Suarez,2
I Jemal Abdulahi3 1HEAL TB, Management Sciences for
Health, Addis Ababa, Ethiopia; 2Center for Health Services,
Management Sciences for Health, Arlington, VA, USA;
3
Communicable Diseases, Oromia Regional Health Bureau,
Background: Ethiopia, one of the high MDR-TB

countries, started its MDR-TB program in 2009 in one
tertiary care center in Addis Ababa. However, the service
was inaccessible to the majority of Ethiopias population.
As part of the effort to address this challenge, a
decentralized MDR-TB service using ambulatory care

model was designed.
Intervention: Together with the Ministry of Health, the
USAID-funded Help Ethiopia Address the TB Performance (HEAL TB) project started expansion of MDR-TB
services in 2012. The project support included renovation of MDR-TB treatment centers training and orientation of health workers, case finding, and building the
regional capacity for sputum sample culture and drug
sensitivity testing. Also, the regional sample transport
system was strengthened and GeneXpert sites expanded
from none to 51. To ensure the quality of services clinical
mentoring for treatment centers established, organized
MDR-TB clinic days whereby patients come only on a
scheduled day for follow-up, provided continuing
medical education, and instituted a mortality audit.
Besides the expansion of treatment initiating centers,
patients after two months of admission discharged to
health centers to continue the DOT.
Results and lessons learnt: These interventions helped
expand treatment initiating centers (TIC) from 2 to 23
(Figure); testing for presumptive MDR-TB from 662 to
20 200; and patients ever enrolled from 56 to 700. More
than 300 health facilities are providing monthly DOT for
patients discharged from TICs. Six-month culture
negativity rate improved from 41% to 76.4%. The cure
and treatment success rate for the first cohort of 36
patients were 65% and 74.9%, respectively. The
proportion of death, failure of treatment and lost to
follow up for the first cohort were 11.1%, 2.8% and
11.1%, respectively. The main challenges were sample
transport for presumptive cases, as most of the health
centers are far from the Xpert or culture diagnostics
health facilities, frequent failure of the Xpert machines,
expensive to maintain, and lack of ancillary investigations. From the patient side, the daily visit to health
centers for medication for two years is challenging
coupled with poor family income.
Conclusion: The decentralized MDR-TB service model
implemented in the two regions in Ethiopia contributed
to improved case finding and access to quality treatment.
The feasibility of further decentralization of DOT follow
up to community level should be evaluated.
Expanding MDR-TB diagnosis and communitybased MDR-TB management by community

health workers in Bangladesh
M Akramul Islam,1 M Siddiqui,1 M Rana,1 S Islam1
Tuberculosis Control Programme, BRAC, Dhaka,
Bangladesh.
Background and challenges to implementation: Antituberculosis drug resistance is a major public health
problem worldwide. According to the last drug resistance
survey of Bangladesh, the MDR-TB burden is 1.4%
among new cases and 28.5% among retreatment cases.
NTP has been introducing Community based Programmatic Management of DR-TB known as cPMDT since
2012.
Interventions: Initially the community based approach to
MDR-TB services only covered a small percentage of the
country compared to that of the drug-susceptible TB
programme, which has nationwide community based
coverage. That is why to ensure all high risk MDR-TB
symptomatics have access to DR-TB testing and services,
the MDR-TB services have been scaled up throughout
the country by the NTP. Currently there are 39 Xpert
MTB/RIF sites through the country, vs. 28 in 2012.
Diagnosed MDR-TB patients admitted in hospital for a
minimum one month of intensive phase of treatment
initially and ambulatory phase at community level.
Currently sputum microscopy is done weekly for
admitted patients. If two consecutive sputum samples
become negative, then the MDR-TB patients are shifted
to the community for ambulatory care. Monthly sputum
follow up tests at the field level and quarterly culture at
the National TB Reference Laboratory (NTRL) is done
throughout the treatment course. All the health care
providers related to MDR-TB management receive a
basic one day orientation before starting of MDR-TB
management at community level.
Results and lessons learnt: n 2008, a total of 103
confirmed patients were enrolled and received standardized MDR-TB treatment. Out of 103 MDR-TB patients,
the treatment success rate was 66 (64%). Death and
default rate was 8 (8%) and 28 (27%), respectively. In
2012, a total of 286 confirmed patients were enrolled
and the treatment success rate was 204 (72%). Death and
default rate was 34 (12%) and 43 (15%), respectively. In
2008, 22% patients absconded from hospital and 5%
from community during treatment, which decreased
14% and 1%, respectively, in 2012.
Conclusion: Patients adherence during hospital stay
seems to be a difficulty of the hospitalization treatment
system as they had to stay for a long time during intensive
phase. The currently implemented Community based
MDR-TB treatment modality may contribute to minimize this challenge and ensure high treatment success
rate for the National Tuberculosis Control Programme in
Bangladesh.
S17
Achieving rapid scale-up of MDR-TB patient

treatment in Uganda using the mixed model
(hospitalised and ambulatory) MDR-TB
treatment
M Kenneth,1 H Luwaga,1 M Ruhweza,1 D Sama,1
S Stavia,2 F Mugabe,2 S Balcha,3 P G Suarez4
1
Management Sciences for Health (MSH)/Track Tuberculosis
Activity (TRACK TB) Project, Kampala, 2Ministry of Health,
National TB and Leprosy Program, Kampala, 3United States
Agency for International Development, Kampala, Uganda;
4
MSH, Arlington, VA, USA.
Background and challenges to implementation: Uganda

is one of the 22 high burden TB countries in the world
with an annual notification of 47 000 TB cases. A 2010
drug resistance survey for Uganda estimated MDR-TB
prevalence at 1.4% among new bacteriologically confirmed TB cases and 12/1% of previously treated TB
patients. By 2012 an estimated 300 MDR-TB cases had
been notified but not enrolled on treatment due to health
system constraints. Prior to July 2012, MDR-TB
treatment could only be accessed at 2 treatment facilities
and under research or project settings.
Intervention: With support from the USAID funded
TRACK TB project, GFTAMTB and other partners, the
National TB and Leprosy Programme mobilised efforts
to decentralise treatment of drug resistant TB to more
hospitals using the mixed model of care (combination of
ambulation with hospitalisation as required). Multidisciplinary MDR-TB care teams were established at
each hospital through training and mentorship, provision
of logistical and operational support for capacity
building at peripheral DOT facilities, contact screening
and investigation and patient support with transport.
MDR-TB cohort review activities were conducted on a
quarterly basis and the MTB/RIF technology was rolled
out to enhance early detection of rifampicin resistance.
Results: MDR-TB treatment was decentralised to 15
hospitals by the end of 2014. The implementation of the
mixed model improved enrollment from 63 to 490
patients by the end of 2014; 90% MDR-TB patients were
enrolled on the mixed model of MDR-TB management.
There were significant improvements in sputum culture
conversion rates at 12 months of treatment from 23% to
74% (P 0.002) and the proportion of patients with
unknown culture results reduced from 72% to 7% (P
0.021) for patients enrolled before and after introduction
of the mixed model of care. Similarly, sputum culture
conversion improved from 54% to 75% at 6 months and
the proportion of patients with unknown culture results
reduced from 16% to 2% (P 0.706) for patients
enrolled before and during implementation of the mixed
model. However, the improvements in interim outcomes
at 6 months were not statistically significant possibly due
to small patient numbers in the cohorts that were
compared.
Conclusion: The implementation of the mixed model of
MDR-TB care improved access to treatment for MDRTB as well as the interim results at 12 months for patients
in Uganda. More work remains to improve the interim
results at 6 months of treatment.
S18
17. Maximising opportunities for

integrating TB services into maternal and
child health programmes
Global burden of TB in pregnancy: from
epidemiology to programme implications
C Colvin 1United States Agency for International
Development/Global Health Fellows Program, Washington,
DC, USA.
TB is not a condition for which pregnant women

presenting at antenatal care are routinely screened. There
is a need to prioritize settings and potential interventions
such as symptom screening and diagnostic tests to ensure
that those at highest risk have access to the full range of
TB care services. This presentation will include an
overview of recent literature on efforts aimed at
improving TB screening in antenatal care and the current
estimates of TB burden among pregnant women. Future
direction for research and programmatic action based on
these findings will be suggested, along with recommendations for leveraging platforms such as PMTCT
programs and rapidly expanding community based
DOTS efforts.
WHO guidance on integrating TB and TB-HIV

services into maternal, neonatal and child
health programmes
A Kanchar,1 A Baddeley,1 H Getahun1 1World Health
Organization, Geneva, Switzerland.
Background and challenges to implementation: Tuberculosis caused more deaths among women than all causes
of maternal mortality taken together (510 000 vs. 289
000 in 2013). It is among the five leading causes of death
among adult women (2059 years) in low-income
countries. The TB epidemic is intertwined with the HIV
epidemic and it affects women disproportionately
particularly in the African region where about 60% of
all PLHIV are women. Very few national TB or maternal
and child health programmes collect or report pregnancy
specific TB data, however research indicates high rates of

TB particularly in HIV-positive pregnant and postpartum
women ranging between 0.7% to 7.9% and as high as
11% in tuberculin skin test positive women. TB increases
the risk of pregnancy related complications such as preeclampsia and eclampsia, vaginal bleeding, miscarriage
and prematurity, low birth weight, etc. and post-partum
is associated with 2 to 3 fold increase in maternal and
infant mortality.
Results: WHO recommends the integration of tuberculosis services into reproductive, maternal, neonatal and
child health (RMNCH) service delivery platforms,
particularly PMTCT programmes to ensure its early
detection and prompt treatment. National TB and
RMNCH programmes should systematically collaborate
and ensure utilization of all existing service delivery
options to provide integrated TB, RMNCH/PMTCT
services. The WHO Global TB Programme is developing
a toolkit to assist policy makers, national programme
managers and staff in establishing such integrated
services along with its implementation and monitoring.
This toolkit consolidates existing WHO recommendations for the management of TB in women and children,
TB screening and diagnostic algorithms, country best
practices and operational experience and job aids to
facilitate implementation.
Conclusions and key recommendations: National TB and
RMNCH/PMTCT programmes should consider the
integration of tuberculosis services for HIV-positive and
negative, pregnant and post-partum women, new-born,
children under five years and older children along with
integration into the family planning, STI and infertility
services.
18. Five years of XpertW MTB/RIF

implementation: lessons learnt for
increasing the impact of future new
diagnostics
Impact of Xpert MTB/RIF on TB case detection
and treatment: experience of the Brazil NTP
R Ruy De Souza,1 A De Paula Lobo,1 D Pelissari,1
P B Oliveira,1 F Reis,1 M B Ferreira De Freitas Hammerle,1
F. Dockhorn,1 D Barreira1 1National Tuberculosis Program,
Ministry of Health, Brasilia, DF, Brazil.
Brazil began to deploy its network of Xpert MTB/RIF in

May 2014. Until June 2015, 168 535 tests were carried
out, as reported by the laboratories. This monitoring
network is done using Excel spreadsheets, in which are
reported the production of each laboratory, the results in
terms of positivity, resistance, errors and problems of
devices. Initially, is important to inform that data from
May to December 2014 are much variable because of the
growing and scaled installation of the net and data from
January 2015 are still being analyzed to produce an
annual report on the Xpert test network and still cannot
be calculated correctly. Analyzing the data available,
mainly data from 2014, it seems that: i) there was no
S19
increase in the detection of cases, but of laboratory

confirmed cases, as it happened in other countries; ii)
there was also a decrease in the time to onset of
treatment, but because of the faster response of the test
and not because a desirable reorganization of the health
system; iii) the time to start patients treatment with
rifampicin resistance also seems to have diminished in
many states, as the algorithm designed by NTP gives
priority to those patients in referral hospitals. Analyzing
data from July to December 2014, the average positivity
of the tests ranged from 8.6% to 12.3%. The detection of
resistance ranged from 1.4% to 6.3%. These initial
findings seem to support the study of deployment of the
technology that was conducted in 2012 in two municipalities, Rio de Janeiro and Manaus. In this study it was
possible to measure the decrease in time for the
notification of the case, which fell from 19.4 to 11.4
days, and the time to onset of treatment, which fell from
11 to 8 days. The authors also observed a 35% reduction
in TB-related mortality, which may be explained by less
advanced disease among the smear-negative patients
diagnosed by Xpert.
addressed, in Belarus Xpert is used for testing EPTB

samples (pleura, CSF, lymph nodes). The test is also used
for children presumed to have TB and as screening tests
for PLHIV.
Conclusion: Xpert is an important new tool for the
diagnosis of TB, which has the potential to impact not
only individual patient outcomes, but also the course of
the TB epidemic in high burden countries. Perhaps most
importantly, Xpert has established the principle that
nucleic acid amplification tests for TB can be sensitive,
specific and feasible for implementation in poorly
resourced settings. However, there is much work to be
done to better understand how Xpert can be utilized in a
programmatic setting to maximize its potential impact
and cost-effectiveness. Finally, the diagnosis of TB is one
component of the overall control strategy; without
effective linkage into care and strong systems for delivery
of treatment and patient retention, there will be little
impact on the TB epidemic.
Impact of Xpert MTB/RIF on MDR-TB case

detection and treatment: experience of the
Belarus NTP
J Takle 1Global Connectivity, Boston, MA, USA.
A Skrahina,1 A Zalutskaya,1 A Astrauko,1 D Klimuk,1

H Hurevich,1 W Van Gemert,2 M Dara,3 V Rusovich4
1
Ministry of Health, Republican Scientific and Practical Centre
for Pulmonology and Tuberculosis, Minsk, Belarus; 2Global TB
Programme, World Health Organization (WHO), Geneva,
Switzerland; 3Regional Office for Europe, WHO,
Copenhagen, Denmark; 4WHO Country Office, Minsk,
Belarus.
Introduction: Accurate and rapid detection of TB,

including smear-negative TB and drug resistant-TB, is
critical for improving patient outcomes and decreasing
TB transmission. In early 2011, WHO endorsed a novel,
rapid, automated, cartridge-based nucleic acid amplification test, the Xpertw MTB/RIF assay, which can
simultaneously detect TB and rifampicin resistance.
Methods: We describe experience from the rollout of
Xpert in Belarus, a country with a high burden of MDRTB.
Results: Whilst Xpert represents a major advance over
smear microscopy, cost was a major obstacle to
widespread rollout. There has, however, been a rapid
decline in cartridge costs for high burden countries.
However, at present time in Belarus, Xpert is used as
primary test for all TB suspects. With Xpert other tests,
at least microscopy and liquid culture, are always
performed at the same time with the same sample. One
of the most challenging aspects of the implementation
was the implementation of a revised national TB
diagnostics and treatment algorithm. There is a need
for clear, simple, unambiguous and feasible algorithms to
be developed and piloted well in advance of Xpert
implementation so that adequate training of laboratory
and clinical staff can be ensured. Although a number of
issues related to the use of Xpert for EPTB still need to be
Maximising the potential for rapid use of test

results: e/m-tools
Background and challenges to intervention: The recent
proliferation of connectivity for the Cepheid GeneXpert
(e.g. GxAlert, RemoteXpert, CDX and others) evidences
the tremendous potential for the rapid use of test results.
However, like most technology-enabled initiatives in the
developing world, getting the most out of these systems is
not easy. Mobile networks are fragile. Power is intermittent. And lab staff unfamiliar with these systems can
forget to plug in the modems, use them for personal use,
or change the configuration settings after initial setup.
Today, more than 800 labs connect their GeneXperts to a
digital reporting system in order to enhance patient care,
enable EQA programs, and improve laboratory supply
chain. Most programs start out successfully with labs
reporting nearly every day; but over time some of the
programs saw a reduction in use primarily due to the
following: 1. Insufficient network coverage or network
too intermittent; 2. Complicated installation process; 3.
Lack of centralized, MOH support, e.g., being projectled causes issues when the project ends; 4. Lack of a
designated individual responsible for maintaining the
network, IT security, and uptime
Intervention: Introduction of multi-SIM modems that
search for best available network across several
networks; Development of streamlined installation software; Procurement of 1-2 enterprise-class agreements
with in-country mobile network operators (MNO);
MOH-led national-level rollouts, versus project led.
Results and lessons learnt: Programs that utilize one or
more interventions above see more consistent and
sustainable solutions. National planning and stewardship is critical to long-term success. Connectivity
experiences network effects where the more GeneXperts in a country that are connected, the greater the
overall value of the network.
S20
Conclusions: Connectivity for the GeneXpert used to be

initiated by small projects that wanted an easier way to
collect programmatic indicators. Increasingly, connectivity is used as a major tool for MOHs to improve linkage
to care, optimize supply chain, and institute remote QA
on lab diagnostic performance. National-level budgets
and the MOH leadership required for long term success
are becoming the norm. When executed well, the rapid
use of GeneXpert test results will continue to lead
towards better healthcare.
19. Quality laboratory services and

impact on patient care
care. Better quality data delivered faster logically leads to

better healthcare. New algorithms and technologies are
being built to help automate this direct Linkage to Care.
Thoughtful design on effective data exchange will lead to
rapid turnaround of patient results to where they are
needed most.
20. The unaddressed health challenge of

TB in prisons: applying End TB Strategy
concepts of patient-centred care
20 years of TB control in penitentiary systems
in Azerbaijan: achievements and challenges
J Takle 1Global Connectivity, Boston, MA, USA.
R Mekhdiyev,1 E Gurbanova,1 F Huseynov,1 O Baghirov,2

R Tahirli2 1Main Medical Department of the Ministry of
Justice, Baku, 2Specialized Treatment Institution for
Detainees with TB, Baku, Azerbaijan.
Background and challenges to intervention: The growing

coverage of mobile networks is changing how labs can
support clinical care in the field. Rather than relying
solely on paper ledgers and motorcycle couriers to
communicate test results, diagnostic devices like the
Cepheid GeneXpert, Alere PIMA / Q, and others can be
configured to send an electronic copy of their test results
over a mobile network as soon as the test is finished. The
potential benefits span from immediate patient follow
up, to a digitized view of the lab supply chain, to remote
EQA improvements, and more. However, most health
information systems (HIS) are not equipped to receive
this rapid communication. Electronic Medical Records
(EMR), case management tools, and Laboratory Information Systems (LIS) are logical recipients of digital lab
results, but were never designed to exchange data
amongst themselves. As a result, there is no simple way
to match a digital diagnostic result to an existing patient
record automatically, with the necessary degree of
confidence.
Intervention: Perfectionin the form of a truly unique
national ID for each countryis not on the horizon, so
programs must currently administer the same diligence
with digital records that they do with paper records:
results should be matched manually in other HIS and
quality controls put in place. But, new digital systems
must be equipped with the patient demographics
necessary for accurate matching into the other records
systems, e.g. age, sex, HIV status, name, phone number,
etc. And, systems can be designed with integration in
mindbuilt with the mechanisms to accept digital lab
results for effective linkage to care.
Results and lessons learnt: MOHs and their programs
can make tremendous strides by applying the same
criteria they use on paper records as they develop the
matching algorithms and technology bridges necessary to
enable automatic matching.
Conclusions: The rapid turnaround of test results, and
the ability to digitally send them anywhere in the health
system in near-real-time via SMS text, email, or database
connection, holds amazing potential to improve health
Background: The Main Medical Department (MMD) of

Azerbaijan Ministry of Justice provides healthcare
including TB services to 24 penitentiary institutions with
16 000 population.
Design and methods: Following political commitment, a
TB Control Program has been successfully developing in
accordance with WHO recommendations over 20 years.
It includes early TB case finding, qualitative and rapid
diagnostics, standardized TB treatment, strict compliance to DOT and infection control (IC) protocols as well
as uninterrupted supply with certified drugs. Quality
assured TB laboratory is equipped with phenotypic and
genotypic diagnostic tools for precise TB diagnostics.
Both drug-susceptible (DS) and drug-resistant (DR) TB
treatment is centralized at the Specialized Treatment
Institution (STI) with appr. 900 bed-places. IC include
administrative (patient segregation in accordance with
infectiousness, drug sensitivity profile, treatment phase
and detention regimen) environmental (input and extract
ventilation, UV lamps) and individual measures. Patients
are offered permanent support of adherence via additional food and hygienic packages as well as psychological support to overcome stress related to disease and
imprisonment. Uncompleted TB treatment, initiated
within the PS, proceeds after release. Having ensured
appropriate juridical basis MMD has been collaborating
with MoH and NGO to provide support and follow-up
of TB patients after release.
Results: Comparing with 1995 in 2014 mortality from
TB among Azerbaijan prison population decreased by
.65 times (Figure). The Programme achieved significant
treatment success rate both for new smear positive DSTB and MDR-TB cases equal 91% and 80% respectively.
100% of inmates pass through TB screening in pre-trial
isolators and 95% of the prisoners pass through annual
mass screening for TB including questionnaire, X-ray,
and sputum investigation. Special attention is paid to
identification and prevention of TB among HIV-infected.
Training Center functioning since 2012 at the STI was
declared WHO CC on TB prevention and control in
prisons and trained .700 medical and non-medical staff
Rapid turnaround equals better care: designing

county solutions for effective data exchange
of PS. Moreover, delegations from more than 15

countries visited the Program to get acquainted with
the practice. Operational researches are performed on
the Programmes theoretical and practical basis.
Conclusions: Today Azerbaijan TB Control in Prison
Programme is an internationally recognized model for
penitentiaries in and beyond the European Region.
21. TB contact management in children:

closing policy-implementation gaps
Cross-border approaches to IPT: treating
immigrant children from high-prevalence
settings in the USA
A Cruz 1Pediatrics, Baylor College of Medicine, Houston, TX,
USA.
Background: The most effective strategy for preventing

cases of TB disease in low incidence nations is by
identifying and treating patients with TB infection.
However, screening for risk factors and testing for
infection assume access to medical care, which may not
be readily accessible for high-risk patients.
Objectives: In this symposium, we will discuss the
barriers to current prevention strategies and methods to
provide preventive therapy to high-risk individuals, such
as unaccompanied children crossing the Mexico-United
States border into Texas. These approaches include
screening and testing in nontraditional venues (e.g.,
schools), use of interferon gamma release assays to
decrease false-positive results in a BCG-immunized
population, use of shorter-course regimens, and administration of medication(s) under directly-observed preventive therapy.
S21
Is mHealth an effective strategy to improve IPT

adherence in child contacts?
K Du Preez,1 L Du Plessis,1 A Hesseling1 1Paediatrics and
Child Health, Faculty of Medicine and Health Sciences,
Stellenbosch University, Desmond Tutu TB Centre, Cape
Town, South Africa.
Background: Current international and national South

African guidelines recommend 6 months of daily
Isoniazid preventive therapy (IPT) for young and HIV
infected children following documented household TB
exposure. Reported adherence in high-burden TB settings is poor, with the minority of children completing 6
months IPT. Innovative supporting tools are needed to
improve adherence.
Methods: Against the background of a multi-level
intervention to improve IPT delivery in the Khayelitsha
sub-district, Cape Town, we investigated the impact of
monthly text message (sms) reminders for 6 months to
caregivers of children initiated on IPT. As children were
recorded to have started IPT in paper-based registers at
two health clinics during 2014/5, their parents were
included in the reminder sms campaign if their cellular
telephone number had been recorded by routine staff.
Block randomisation was performed for a 14 month
period, to allocate 50% of the months to the sms
intervention (sms-months), and 50% to the control arm
(no sms-months). Children initiated on IPT during smsmonths were included in the campaign, and children
initiated during no sms-months received standard-of
care. Routine health care providers at the clinics were
blinded to the sms status of each month. Outcomes (IPT
completion, and number of months completed) were
compared between the intervention and control arms,
adjusting for household clustering.
Results: 235 children were entered into the registers from
Jan-December 2014; 187 (80%) with recorded cellular
telephone numbers. Of the total 235, 185 (79%) were
from clinic 1 and 50 (21%) from clinic 2. Cellular
telephone numbers were recorded for 142 (77%) and 45
(90%) of children from clinic 1 and 2 respectively.
Analysis is ongoing.
Increasing childrens access to IPT in Western

Kenya
D Szkwarko,1 E J Carter,2,3 L Kamule,3 N Buziba,3,4
P Owiti3 1University of Massachussetts, Worcester, MA,
2
Warren Alpert Medical School of Brown University,
Providence, RI, USA; 3Academic Model Providing Access to
Healthcare (AMPATH), Eldoret, 4Hematology, Moi University,
College of Health Sciences, Eldoret, Kenya.
Background: Isoniazid preventive therapy (IPT) delivery

to child contacts exposed to tuberculosis remains a major
problem in low-resource settings. This is closely intertwined with the operational challenges that tuberculosis
(TB) programs globally continue to face in fulfilling
World Health Organization recommendations for child
contact management and screening. Despite new National Program recommendations in Kenya to initiate
IPT in children under 5 years, major screening barriers
S22
including cost of screening and transportation prevent

child contacts from being brought into care for screening
and IPT initiation. Intervention Funded by TB Reach,
child contacts under five years exposed to smear positive
index cases in 18 clinics throughout western Kenya
became eligible for a child contact screening package
during a 17-month period. The package included
screening chest X-ray, transport fees for index case and
child, and registration fee. Outcomes included the
number of child contacts who were screened and their
diagnoses (active TB or IPT initiation) after program
completion.
Results and lessons learned: 18 sites underwent implementation; 4 failed to complete any tracking, screening,
or treatment initiation. Five sites initiated screening
during quarter 1; the remaining 9 sites were functional by
month 11. 631 child contacts under five years completed
screening. 522 (83%) children initiated IPT; 99 initiated
active disease treatment. Although 1137 smear positive
index cases were identified, only 540 of these were
queried regarding child contacts. Challenges included
limited healthcare staff to conduct interviews, data
collection difficulty, and child contacts transferring
treatment monitoring to other facilities.
Conclusions: IPT delivery to child contacts exposed to
tuberculosis can be improved by eliminating cost to
screening and transportation. However, challenges in
implementation of child contact screening and treatment
initiation and monitoring still exist.
ventions that increased smear microscopy, smear-positive

notifications, treatment outcomes and explored use of
Xpertw MTB/RIF among nomads in Nigeria will be
discussed.
Recommendations: Existing programmes for nomads
afford opportunities for absorption and integration of
TB services using the One-World One-Health approach.
In Nigeria, the National Commission for Nomadic
Education has established an HIV/AIDS desk that
focuses on mobilisation and awareness campaigns, and
TB services could be provided along their migration
routes. Nomadic communities should be provided with
basic TB services, including education, counselling,
screening, detection and treatment. These services should
be taken to nomadic communities using novel outreach
approaches such as mobile units, extension services, case
management, directly observed care, and treatment
supporters linked to neighbouring facilities in a huband-spoke model. Stronger collaborations are recommended between veterinary programmes for nomads
livestocks and TB/public health services. Community
participation should be encouraged to ensure the
sustainability of TB care delivery.
Conclusions: There is inequality, inadequacy and inaccessibility of TB services among nomadic populations
and concerted efforts should be made to improve care.
Nepal: community outreach challenges and

successes in a post-earthquake setting
R S Gopali,1 G Shrestha,1 N Ishikawa,2 K Okada2
Department of Public Health, Japan-Nepal Health and
Tuberculosis Research Association, Kathmandu, Nepal; 2
Department of International Cooperation, Research Institute
of Tuberculosis/JATA, Tokyo, Japan.
1
22. A community-based approach to

reaching the missing three million
Nigeria: walking the walk with nomadic
populations: strategies that work
A Habib,1 M Hamza,1 K Karaye,1 M Bello,1 B Usman,2
I A Abdulkadir2 1College of Health Sciences, Bayero
University, Kano, 2Faculty of Veterinary Medicine, Ahmadu
Bello University, Zaria, Nigeria.
Background: Nomadic societies are marginalized and

vulnerable to infections. Their relative illiteracy, culture,
pastoral lifestyle and living conditions with livestock
predisposes to Tuberculosis (TB), including bovine TB
especially among cattle herders. Community mobility
limits opportunities for detection and access to care.
Methods and Discussions: Lessons from cases and case
studies on TB among the nomadic Fulani in Nigeria and
the Manyatta TB project in Kenya will be discussed.
Clinical and socio-cultural factors among Fulani patients
attending our TB Directly Observed Therapy Short
Course clinic will be presented. Comparison will be
made to native populations in other parts of the world.
Results of a recent large integrated survey of TB
prevalence among 229 nomadic Fulani adult population
and 204 of their cattle herds in Rano, Kano, Nigeria will
be presented (2014). Use of SD Bioline TB Ag MPT64 in
speciation cattle and human mycobacterial isolates to aid
TB detection (9.6%) will be described. Practical inter-
Aim: To improve the health status of internally displaced

people by screening tuberculosis symptomatics for early
case finding.
Method: The earthquakes that struck Nepal in April and
May 2015 caused destruction of houses and massive
displacement of the population. Some of the internally
displaced persons live in overcrowded makeshift settlements or share houses with their relatives, damage of
health care infrastructure and displacement of health
care workers lead to disruption of health care services,
including TB diagnostic and treatment services. The
National Tuberculosis Programme raised concern about
interruption of TB treatment and increased TB transmission due to overcrowding and appealed to partners for
support with early detection of TB among displaced
persons and host communities. JANTRA carried out
active screening of these populations through symptom
screening, mobile microscopy camps at highly affected
districts in Kathmandu valley. The bacteriologically
confirmed TB cases are referred for TB treatment in
nearby DOTS centres or volunteers.
Results: A total of 2783 people (1232 males and 1551
females) residing in the 10 IDP sites in Kathmandu valley
were screened for tuberculosis using a structural questionnaire. Among total screened 42% (1182) were found
with TB signs and symptoms. Among the total presumed
cases, 85% participated in the microscopic camp

providing two sputum samples (morning and spot).
Seven (0.7%) new smear-positive TB patients were
diagnosed from this initiation and all the diagnosed
cases are referred to nearby DOTS centres for treatment.
Total 150 Female Community Health Volunteers and 10
Health Care Providers including laboratory staffs actively participated to implement this approach closely with
health care providers working in local level.
Conclusion: The prevalence of tuberculosis in intervening IDP sites is relatively high (700/100 000 population),
which is roughly four times higher compared than the
national prevalence estimated by the WHO (241/100
000). 56% of participants were female and 44% male,
which is unusual in NTPs laboratory activity, where the
participation of males is always higher than females. On
the other hand, the segregate male and female prevalence
among total suspects examined, it is higher in males
0.78% (3/383) than females 0.64% (4/617). Similar
types of activities must be expanded to other IDPs to
determine the real prevalence of TB.
Greater Mekong sub-region: accessing hard-toreach populations in politically sensitive

settings
D Punpiputt,1,4 A Innes,1,4 K Z Aye,2,4 l Ling3,4 1Asia Pacific
Regional Office, Family Health International, Bangkok,
Thailand; 2Myanmar Country Office, Family Health
International, Yangon, Myanmar; 3China Country Office,
Family Health International, Kunming, China; 4USAID Control
and Prevention Tuberculosis Project, Bangkok, Thailand.
Background: Effective multidrug-resistant tuberculosis

(MDR-TB) control is a challenge for TB platforms and
health systems, as evidenced by the large diagnosis and
treatment gaps reported from global data. Despite
progress since 2009, fewer than 50% of estimated
MDR-TB cases worldwide were diagnosed in 2013,
and the global success rate remains ,50%. The USAID
Control and Prevention Tuberculosis (CAP-TB) Project
has focused on strengthening MDR-TB control in the
Greater Mekong Sub-region of Yunnan Province, China;
Myanmar; and Thailand.
Response: Starting in 2012, the CAP-TB project developed a model for MDR-TB control using a patientcentered perspective to identify gaps in the health
system. The CAP-TB model starts with finding presumptive TB and MDR-TB patients in the community; linking
presumptive patients with the TB system for early
diagnosis and initiation of treatment; and supporting
those diagnosed throughout their 20-24 month regimen.
Community mobilization to support patients used both
standard and innovative methods: by engaging volunteers and outreach workers to provide support and
accountability through a bundle of patient-support
interventions; as well as through social media, creating
virtual communities comprising patients, peers with TB
or MDR-TB, and health care providers. This patientcentered, community-mobilized model was developed
first in urban and centralized levels of the health system,
S23
with the idea to adapt the model to more rural and

inaccessible areas in each country.
Results: In China, the project mobilized community
support for 79 MDR-TB patients diagnosed since May
2012 in Kunming, Yunnan Province, and improved
coordination between the hospital and public health
sectors to maximize continuity of care between inpatient
and community settings. These patients are currently
completing their treatment, with approximately 74%
treatment success rate, 17% loss to follow-up, and 1%
mortality. In 2014, the urban Kunming model was
adapted to the rural setting in Yunnans Zhenxiong
County, which has the highest number of TB patients and
second highest TB prevalence in the province. Scaling up
further, Xinjiang Province, which reports the second
highest TB prevalence in China, requested technical
assistance from CAP-TB to improve treatment success
rates. Political turmoil in Xinjiang likely contributes to
challenges for TB control; the Chest Hospital in Xinjiang
has now adapted CAP-TBs social media strategy to
improve patient-provider communication with the overall goal to improve treatment adherence. In Myanmar,
the CAP-TB project has supported 1101 MDR-TB
patients since January 2013 in Yangon and Mandalay,
which are the centers for MDR-TB control in Lower and
Upper Myanmar. Developing a model based in these
regional centers has laid the foundation for expansion to
more rural and hard-to-reach areas of the country. These
MDR-TB patients in Yangon and Mandalay are completing their treatment with a success rate of approximately 75%. In Thailand, CAP-TB focused on Rayong
Province, which has the third highest MDR-TB case
notification rates among all provinces in Thailand. This
industrial, coastal province attracts internal migrants
from throughout the country as well as cross-border
migrants from Cambodia and Myanmar. Since 2012, the
CAP-TB project has supported 79 MDR-TB patients
currently on treatment in Rayong, focusing on building
multi-disciplinary teams at the provincial level that
coordinate among different levels of the TB system to
improve patient support. Treatment success rates for the
CAP-TB supported Rayong patients are currently 65%,
higher than the national average of 49%.
Conclusions: Developing a patient-centered model that
mobilizes the community, leverages technology and
social media, and strengthens TB platforms can impact
MDR-TB control. Strategic focus on urban, more central
levels of the TB system facilitates development of the
model, which can then be adapted to access marginalized, hard-to-reach populations.
S24
23. Moving towards universal access to

drug susceptibility testing: progress and
challenges
practical alternative for resource limited settings to

decide on initiating treatment earlier while awaiting
DST confirmation.
Poor access to follow-on DST for rifampicinresistant (RR) TB cases: root-cause analysis
done in MSF projects in African & Asian
settings
M Rumaney,1 E Fajardo,1 C Metcalf,1 M Casenghi,2
M De Melo Frietas,3 E Mbofana,4 P Isaakidis,5 E Mohr6
1
`
Southern African Medical Unit, Medecins
Sans Frontieres
(MSF), Operational Centre Brussels, Cape Town, South
Africa; 2Access Campaign, MSF, Geneva, Switzerland;
3
Mozambique Project, MSF, Operational Centre Brussels,
Maputo, Mozambique; 4Zimbabwe Project, MSF, Operational
Centre Brussels, Harare, Zimbabwe; 5India Project, MSF,
Operational Centre Brussels, Mumbai, India; 6South African
Project, MSF, Operational Centre Brussels, Cape Town, South
Africa.
In multidrug-resistant tuberculosis (MDR-TB) prevalent

settings, the World Health Organization (WHO) recommends that Xpert RIF-resistant results be confirmed with
the Line Probe Assay (LPA) or phenotypic drug
susceptibility testing (DST) before initiating patients on
treatment. For low prevalence settings a second Xpert
test should be performed and if RIF-positive, treatment
should be initiated while awaiting confirmation. LPA is
not routinely available in remote settings and is validated
for smear-positive samples only. DST is available as
centralised testing with long turnaround times (TAT) to
results, leading to lengthy transportation, cold-chain
requirements and poor sample quality. This study aims to
describe the challenges associated with confirmatory TB
testing, and the differences in high vs. low prevalent
settings, taking discordant results and its effect into
consideration. Qualitative and quantitative data was
collected from six sites two high prevalent settings,
South Africa (Cape Town); India (Mumbai), and four
low prevalent settings Mozambique (Maputo; Tete);
Zimbabwe (Buhera); Kenya (Kiberia). Qualitative data
was collected via a survey. Quantitative data was
compiled by retrospective patient folder review from
2012-2014. Statistical analysis was performed using
Stata version 12 (StataCorp. Texas, USA). Qualitative
data analysis revealed: 1) a second Xpert test was
performed routinely in 1/6 sites, 2) the distance from
the health facility to the laboratory is between 102000km, depending on the site, 3) 5/6 sites use road as
the medium of transport, 4) first line DST is performed in
all sites, 5) treatment is initiated without a DST result in
2/6, and while awaiting a DST result in 3/6, sites.
Challenges include sub-par transportation, sample contamination, insufficient specimen volume, poor sample
quality, inadequate DST follow-up, delayed entry of
results, and missing data. Quantitative data analysis
presented using statistical analysis. There is an urgency to
strengthen diagnostic capacity to confirm DR-TB,
referral networks for sample transport and rapid resultdelivery. Although repeat testing increases cost, it offers a
Improving access to DST and adherence to

diagnostic algorithms: lessons learnt in Nigeria
through the Xpert MTB/RIF scale-up
P Nwadike,1 M Gidado,1 S Useni,1 J Onazi,1 P Ajiboye,1
C Ehimanre,1 R Eneogu,1 E Elom2 1KNCV/TB CARE 1
Project, Abuja, 2National Tuberculosis & Leprosy Control
Program, Federal Ministry of Health, Abuja, Nigeria.
Introduction/Background: Prior to Xpert MTB/RIF

implementation, Nigeria had only 2 functional culture/
DST laboratories. Conventional DST has a limitation for
access in a huge country like Nigeria with weak referral
mechanism and sample movement limiting the diagnosis
of DR- TB patients. Xpert MTB/RIF was adopted in
2010 with the aim of improving access to DST among the
increasing DR-TB patients. Currently the country has
144 Gene Xpert sites and 5 functional culture/DST
laboratories. The process of scaling up DST was
associated with updating of the diagnostic algorithm
from one priority group (Cat 2 failure) to eight priority
groups (PLHIV, Contact of DR-TB patients; children
with TB symptoms, EPT) among others. However,
because Xpert provided only RIF resistance, more
culture/DST were also scaled up for diagnostic confirmation of PMDT patients.
Objective: This paper describes the additional role of
GeneXpert MTB/RIF machine in providing access to
DST and early enrollment of patients for MDR-TB
management using multistage algorithm development
process
Methodology: This is a retrospective review of routine
quarterly report of both programmatic and supervision
reports to DST and Gene Xpert sites between 2011-2014.
Additionally, desks review of all the existing guidelines
were carried out.
Results: Prior to the adoption of the GeneXpert in 2011,
about 126 cases of DR-TB were diagnosed across the
country between 2007 and 2010 through culture/DST
and LPA in the five functional laboratories across the
country. In 2013, 10 652 presumptive DR-TB cases were
S25
screened using Xpert MTB/RIF; out of which 764 were

MTB detected/RIF resistant. Review and expansion of
the NTBLCP algorithm to include other priority groups
led to the screening of 24 313 using the Xpert machine in
2014, out of which 798 isolates were resistant to
Rifampicin.
Lessons learnt: Rapid expansion of either culture culture/
DST or Xpert MTB/RIF to increase DST access will be
difficult to achieve without proper logistics for referral
and sputum transportation
Conclusion: Access to DST and adherence to algorithms
has improved via Xpert scale up and appropriate
coordination by the NTBLCP with increased and
targeted supportive supervision funded by partners.
a government fund or a separate foundation? In practice,

a wide variety of earmarks and uses are in use, including
health, research, cancer treatment and health promotion.
Fewer countries fund tobacco control directly. The
central issue is the need for governments to commit
sufficient resources for health. Increasing revenues from
tobacco taxes makes it easier for governments to make
that commitment and there is ample space to increase
tobacco taxes in most countries.
24. Why and how to use tobacco taxes as

funding mechanisms for financing health
in low- and middle-income countries
What is the evidence that BCG prevents

infection?
How tobacco taxes can be used to create

funding mechanisms for health financing
The development of effective immunoprophylaxis

against tuberculosis (TB) remains a global priority, but
is hampered by a partially protective Bacillus CalmetteGuerin (BCG) vaccine and an incomplete understanding
of the mechanisms of immunity to Mycobacterium
tuberculosis. Thus far, preventing TB disease, rather
than infection, has been the primary target for vaccine
development. Several areas of research highlight the
importance of including pre-infection vaccines in the
development pipeline. First, modeling data indicate that
a pre-infection vaccine would have high population-level
impact for control of TB disease. Second, targeting host
responses that prevent infection may be more effective
during acute as compared to chronic infection. Third,
epidemiologic data from natural history studies indicate
that resistance to M. tuberculosis infection occurs in a
small percentage of the population. Fourth, retrospective
case-control studies suggest that BCG is associated with
protection from infection. Together, these areas suggest
biologic plausibility and epidemiologic support for
expanding the focus of TB vaccine development efforts
to include prevention of infection as a primary goal along
with interventions that reduce transmission and reactivation.
J Tesche Tobacco Control, International Union Against

Tuberculosis and Lung Disease, Edinburgh, UK.
The lack of sufficient funding for health is a serious issue

in many countries. Increasing tobacco taxes is one way of
increasing government revenues, including funding for
health. There is ample scope to increase tax rates and
cigarette prices in order to increase revenues and funding
for health since tax rates in most countries remain below
rates recommended by the WHO (total taxes of at least
75% of the retail sales price (2015 Report on the Global
Tobacco Epidemic). Comparable retail prices of cigarettes are also significantly lower in middle and lower
income countries than in richer countries. There are
theoretical and practical arguments both for and against
earmarking. Economic theory can support earmarks, but
only in the case of a benefit or user fee. Since smokers use
more health resources than others this is an appropriate
use of dedicated funds. Politically, voters are more likely
to support an increase in taxes if they know that the
money is going toward health. One argument against
earmarks is that they are irrelevant in that funds are not
secure, they can be used for other purposes. Earmarked
funds decrease the flexibility of elected governments. If
the funds go to an off-budget fund, there can be a
decrease in transparency as well. Governments may also
decrease budget allocations for health if a dedicated tax
source goes directly to the health sector or such dedicated
revenues may be insufficient to cover expenditures over
time. If tobacco taxes are to be earmarked for health,
there are a number of decisions that follow: How much
of which taxes will be used and will they be used for
health in general or more specific areas like cancer
research? Will the earmark be hard, passed in legislation or soft, with only a stated intention of where funds
should be used? Are the funds allocated through a direct
line in the budget, and if so, along with other funds or
used only for health? If funds are off budget, is it through
25. Can tuberculosis immunisation

prevent initial infection (not just disease
complications)?
T Hawn School of Medicine, University of Washington,

Seattle, WA, USA.
Defining TB infection
E Nardell Global Health Equity, Brigham & Womens
Hospital, Harvard Medical School, Boston, MA, USA.
The question of whether TB immunization can prevent

infection is heavily dependent on the definition of
infection in the absence of microbiologically proven
disease. Until recently, vaccines like BCG that can
convert the tuberculin skin test (TST) could not be
studied for their ability to prevent infection as defined by
the TST. The advent of interferon gamma release assays
(IGRAs) have already provided observational evidence
that BCG appears to reduce IGRA positivity among
contacts of active disease another common definition of
S26
infection. But both TST and IGRAs are indirect

immunological markers of response to infection, not
infection itself, and recent studies suggest the potential
for transient TST and IGRA positivity, presumably
correlating with transient infection a foreign concept
in traditional TB pathogenesis. Failure to reactivate TB
under the pressure of immunosuppression has been used
in the laboratory to further define the presence or
absence of infection. Less invasively, long-term follow
up of presumed human infection in normal and
immunocompromised hosts who decline chemoprophylaxis provides another window to true infection vs. a
harmless residual immunological footprint of infection.
Potential for disease as a marker of infection, however, is
complicated by differences in strain virulence, host
resistance, and perhaps dose. The definition of infection
in humans is clearly harder to assess than in experimental
animals where organs can be cultured. Moreover, the
question has been raised on whether infections prevented
by immunization are primarily those that might naturally
clear infection anyway, with little or no impact on those
who become infected and progress to disease. Clearly the
question is complex and perhaps not answerable with
todays tools. For now, vaccine impact on infection might
best be studied in animal models where true infection can
be investigated thoroughly. In humans, outcomes should
include both immunological markers of infection and
progression to disease.
Immunisation to prevent TB infection of

medical travelers: the TIPI trial
N Aronson,1 E Nardell2 1Infectious Diseases, Uniformed
Services University of the Health Sciences, Bethesda, MD,
2
Global Equity, Brigham and Womens Hospital, Boston, MA,
USA.
An increasing number of students, physicians, researchers, other healthcare personnel, and humanitarian groups
from areas of low risk travel to work in areas of the
world where the incidence of TB (including MDR- and
XDR-TB) is high. These workers are at risk for infection
and disease from these resistant TB strains. Several
studies and unpublished data indicate that TB infection
risk for travelers to high burden countries is significant
with estimates ranging from 4-8%. The primary objective of this proposed clinical trial is to test the hypothesis
that BCG immunization can reduce the occurrence of
Mycobacterium tuberculosis infection as measured by
interferon gamma release assay (IGRA) conversion in
BCG vaccinated recipients as compared to placebo
recipients. This presentation will introduce the details
of the TB Immunization to Prevent Infection (TIPI)
clinical trial.
Implications for vaccine clinical trials

L Schrager Scientific Affairs, Aeras, Rockville, MD, USA.
Developing a vaccine capable of preventing established

infection with Mycobacterium tuberculosis represents an
intriguing and lofty goal among TB vaccine developers.
In addition to the significant challenges inherent in

developing a vaccine capable of preventing established
M. tuberculosis infection, designing clinical trials to
actually prove vaccine efficacy for this endpoint presents
an additional level of complexity. Currently, a Phase 2
proof of concept clinical trial assessing the potential of
two M. tuberculosis vaccines to prevent M. tuberculosis
infection is ongoing. This trial, however, is utilizing
prevention of M. tuberculosis infection as an indicator of
biological activity, intended to de-risk future trials
geared to assess the ability of vaccines to prevent active
TB disease. A pivotal Phase 3 trial with the object of
establishing prevention of M. tuberculosis infection as a
licensable endpoint will necessitate far more regulatory
scrutiny as compared to the prevention of disease
endpoint. One critical issue that will have to be clarified
before such an approach is likely to be viable from a
regulatory perspective is the reliability of the currently
available interferon gamma release assays (IGRAs) to
accurately identify new M. tuberculosis infections among
an uninfected population. Another likely regulatory
concern is the uncertainty surrounding the actual effect
a vaccine that is less than 100% efficacious in a
prevention of infection study may have in preventing
M. tuberculosis disease. While disease cannot occur
absent infection, there is a theoretical concern that those
in whom a vaccine prevents infection may be the
approximately 90% of persons who would otherwise
control M. tuberculosis for life, rather than the approximately 10% of individuals who will eventually develop
disease, thereby having no real effect on TB disease
outcomes. Unless these issues can be addressed, vaccines
developed to prevent the establishment of M. tuberculosis infection may find their most efficient pathway to
licensure via a Phase 3 clinical trial intended to prevent
TB disease among persons initially uninfected with M.
tuberculosis, considering that more than half of the
persons who develop disease are likely to do so within the
two years following initial infection. Sample size
estimates that will provide insight as to the potential
feasibility of this approach will be presented.
26. WHO/ERS initiative on e/mHealth in

tuberculosis and tobacco control
The WHO/ERS initiative to improve e-/mHealth
for TB and tobacco control: concepts and
progress
D Falzon,1 S Aliberti2 1Global TB Programme, World Health
Organization, Geneva, Switzerland; 2University of Milan
Bicocca, Milan, Italy.
Tuberculosis (TB) remains a major global public health

concern. Information and communication technologies
(ICT) present new possibilities to address this challenge
on different fronts. Various innovative digital health
projects have been initiated by TB programmes and
technical partners worldwide to support their TB care
and prevention efforts. The fast rate with which ICT have
advanced and diversified in recent years creates increasing opportunities: by mid-2015 there were over 7.5
billion mobile phone connections globally, and about
40% of the worlds population had an internet connection. The World Health Organization (WHO) and the
European Respiratory Society (ERS) have joined forces to
promote the broader use of state-of-the-art ICT to the
benefit of TB patients. In line with WHOs post-2015 End
TB Strategy, the two organisations are looking beyond
traditional entry-points to support TB care and prevention. These include the use of e/mHealth interventions to
improve the management of comorbidities (e.g diabetes)
and health risks (e.g. tobacco smoking) which frequently
overlap in the TB patients. In February 2015, the two
organisations jointly organised a technical consultation
in Geneva on digital health for TB and tobacco control.
As a result of this discussion, the participants identified
priority digital health products for future focus and target
product profiles (TPPs) were elaborated for these
concepts (www.who.int/tb/areas-of-work/digital-health/
). WHO appointed a Global Task Force on digital health
for TB and this group is now taking forward the
continued development of the TPPs and country case
studies to study these concepts at large scale. Case studies
are now being mounted in Belarus, the Republic of
Moldova, and elsewhere. It is envisaged that these studies
will cover a diversity of products, including electronic
tools for active TB drug safety management and
monitoring, video observed therapy, connected diagnostics, and eLearning. The presentation will report on the
early outcomes of the WHO/ERS collaboration, the
rationale and content of the TPPs, the anticipated
activities into 2016, and the Digital Health for the End
TB Strategy: an agenda for action which the two
organizations have developed together.
Improving and increasing evidence: from pilot

to clinical trials to scaled mHealth interventions
R Lester Medicine, University of British Columbia, Vancouver,
BC, Canada.
Mobile and electronic health (mHealth and eHealth)

provide exciting new opportunities to support tuberculosis (TB) and tobacco control efforts. Low cost cellular
phones and improving internet access has made these
innovations particularly attractive for overcoming barriers to prevention and care in resource-limited settings and
among vulnerable populations worldwide. The evidence
of effectiveness of these interventions for TB and tobacco
control or related conditions is highly variable, but
continues to rapidly emerge. Many initiatives are
launched as pilot projects, raising fears among some as
causing pilotitis (health system inflammation from too
many disconnected interventions that have been found
successful without being brought to scale). Other projects
are implemented broadly and at high cost with little
evidence of health outcome effectiveness. In between,
some high- (and poor-) quality evidence has emerged for
both prevention (e.g., text2stop smoking cessation) and
for treatment support (e.g., WelTel medication adherence
support via SMS and video observed therapy (VOT)
S27
monitoring). The quality evidence available is not all

positive, but is necessary to inform what works and what
doesnt so that the best investments in health systems can
be made in an impactful and timely way. In this talk, we
will briefly review the current evidence and discuss a
system of developing evidence over the implementation
spectrum - from formative research through pilot studies,
efficacy and effectiveness studies, and ultimately program
monitoring and evaluation at scale. Given the novelty of
such systems in healthcare settings, issues such as
technical capacity (e.g., cellular and wifi availability),
privacy protection, policy and procedures in clinical
environments, and resource / staff availability as well as
stakeholder buy in and sustainable business models are
critical evaluative considerations for adopting these
innovations at scale. As such, implementation science
is increasingly recognized as a necessary research area to
help bridge the evidence to practice gap for best practices
and impact at scale.
Development of target product profiles for

connected diagnostics in TB
D M Cirillo,1 C Denkinger2 1San Raffaele Scientific Institute,
Milano, Italy; 2FIND, Geneva, Switzerland.
eHealth solutions offer a unique opportunity in the

operationalization of TB care to meet the EndTB strategy
put forward by the WHO. Target product profiles (TPPs)
can describe eHealth solutions that target priority needs
of TB programmes in further detail and thus provide a
strategic planning tool as well as a communication tool
for investors and stakeholders. Diagnostic tests are a
cornerstone of TB control in guiding appropriate therapy,
infection control and surveillance. However, often the use
of data generated by diagnostic tests is lost or reaches
decision makers only with substantial delays particularly
in decentralized settings. Decentralization of user-friendly, sophisticated diagnostics platforms has been started
with GeneXpert and has shown that the benefit for
patients and for the public health system in general. Even
more decentralized platforms are on the horizon (e.g.,
Omni or Alere Q). These tests will on the hand facilitate
direct patient management at the point of care; on the
other hand the decentralization will pose challenges to
ensure the appropriate data transfer for stock management, surveillance and reporting. Automated transfer of
data generated by a diagnostic instrument using communication technology directly to the relevant individuals
such as clinicians, public health officials and the patients
themselves can overcome this. This enables shorter time
to diagnosis, error reduction from data copy and transfer,
collection of epidemiological data for analysis and
planning interventions and also direct device and user
management. A future TB test, independently from its
size and location, should aim at having a simple way to
report the results through connectivity-enabled diagnostics alongside appropriate data repository solutions and
data presentation applications.
S28
Diversifying approaches to eLearning for TB

and tobacco control
M Dias,1 C Gratziou2 1World Health Organization, Geneva,
Switzerland; 2European Respiratory Society, Athens, Greece.
Both TB and tobacco are major global public health

concerns, causing millions of deaths each year. There is a
strong association between smoking and TB, with each
fuelling the others impact. Information and communication technology could provide new solutions and
avenues for joint action on both of these diseases. The
World Health Organization (WHO) and the European
Respiratory Society (ERS) have joined forces to support
broader use of state-of-the-art information and communication technologies to the benefit of TB patients and
tobacco users.1 Following a joint WHO/ERS technical
consultation on the role of digital health in TB care and
prevention and tobacco control in February 2015, a
number of e/mHealth eLearning products were identified
as advantageously positioned to make an impact on
global health efforts. These included: i) Comprehensive,
e-Learning web-based courses on TB and smoking
cessation approaches, optimized for mobile devices and
equipped with visual instruction aids aims to help
building capacity and skills of health professionals and
ultimately facilitate the detection of TB among smokers
as well as increase the number of people that stop
smoking; ii) An online portal for patients to retrieve
reliable information about their disease, treatment
options and the risks of smoking; and iii) A clinical
decision-support tool to facilitate the daily work of
practitioners and reduce the number of sub-optimally
treated patients. This presentation will highlight how
new eLearning opportunities via the internet and mobile
computing devices can be integrated into the work of TB
programmes in line with WHOs End TB Strategy and the
evidence-based strategies recommended by WHOs
Framework Convention on Tobacco Control. Novel
approaches such as gamification and peer-education will
be reviewed. The potential impact and benefit of all these
eLearning resources for people with TB and tobacco
users will be presented.
Reference
1 Falzon D, Raviglione M, Bel E, Gratziou C, Bettcher D, Migliori G B.
The role of eHealth and mHealth in tuberculosis and tobacco control: a
WHO/ERS consultation. Eur Respir J 2015; 46: 307311.
E-/mHealth initiatives for TB patients in Belarus

and other European countries
A Skrahina,1 D Klimuk,1 D Falzon,2 M Dara,3
P De Colombani,3 A Dadu,3 V Rusovich4 1Ministry of
Health, Republican Scientific and Practical Centre for
Pulmonology and Tuberculosis, Minsk, Belarus; 2Global TB
Programme, World Health Organization (WHO), Geneva,
Switzerland; 3WHO Regional Office for Europe, Copenhagen,
Denmark; 4WHO Country Office, Minsk, Belarus.
Background: Information and communication technologies (ICT) could strengthen the fight against TB.
However the potential of the ICT to combat TB/MDRTB still remains largely untapped, particularly in high
TB/MDR-TB burden and Former Soviet Union (FSU)

countries. Nevertheless, national TB programmes
(NTPs) in many countries have embarked upon pilot
projects to study how eHealth (electronic health) and
mHealth (mobile health) can be used to help their efforts
in the care and prevention of TB and multidrug-resistant
TB (MDR-TB).
Methods: We describe achievements and challenges in
integrating ICT for different aspects of the work of NTPs
of some European FSU countries (Belarus, Moldova).
Digital health products are considered under four
functions: patient care, surveillance and monitoring,
programme management and eLearning.
Results: In Belarus, the first steps in the development of
electronic TB register were taken in 2006. In 2009, with
the support of the Global Fund (GF) a module for the
electronic registration of drug-susceptible TB cases was
developed and implemented. This work was continued in
2012 when a component for drug-resistant TB was added
and in 2013 with the addition of a TB laboratory
informatics component. Since 2014, a drug management
component has been under implementation. All register
components comply with WHO requirements. An
electronic database to evaluate safety and effectiveness
of MDR-TB regimens which include new and repurposed
drugs is also under development. Projects for the future
include eLearning on PAL and video-observed therapy
(VOT) for MDR-TB patients. TB/MDR-TB patients
always face challenges to comply with the demands of
daily long term treatment and many do not complete
treatment as prescribed. mHealth could facilitate bidirectional exchange between patients and health care
providers. Treatment interruption and loss to follow-up
could be alleviated. Belarus NTP is on its way to the
development of mobile digital health product and
regulatory base for eDOT that are aligned to the
challenges posed by TB to health care providers and
patients. VOT/DOT RCT has been started in Moldova
earlier this year, employing low-cost tablet computers to
clinics.
Conclusion: The ICT shows promise in TB/MDR-TB
care. The process for designing, building and rolling out a
digital application needs to embrace a broad cross
section of representative users and policy makers.
Symposium abstracts, Saturday, 5 December
SYMPOSIA: SATURDAY
5 DECEMBER 2015
27. Building on emerging knowledge to

develop novel regimens for paediatric
drug-resistant TB
Global treatment outcomes in children with
paediatric MDR-TB: systematic review and
meta-analysis
E Harausz,1,2 A Garcia-Prats,1 H S Schaaf,1 A Hesseling,1
J Furin,3 R Menzies,4 S Law,4 J Seddon5 1Department of
Paediatrics and Child Health, Faculty of Medicine and Health
Sciences, Stellenbosch University, Desmond Tutu TB Centre,
Cape Town, South Africa; 2Military HIV Research Program,
Bethesda, MD, 3Harvard Medical School, Harvard Medical
School Department of Global Health and Social Medicine,
Boston, MA, USA; 4Respiratory Epidemiology and Clinical
Research Unit, Montreal Chest Institute, McGill University,
Montreal, QC, Canada; 5Department of Paediatrics, Imperial
College London, London, UK.
Background: There is limited evidence on the optimal

treatment of MDR-TB in children. Although outcomes in
children are good (.80% treatment success), MDR-TB
treatment duration is long, requiring the use of toxic
drugs. We reviewed existing evidence on the treatment of
MDR-TB in children to inform updated global treatment
guidelines on paediatric MDR-TB.
Methods: This was a systematic review and individual
patient data (IPD) meta-analysis of published and
unpublished studies of MDR-TB treatment in children.
We conducted a comprehensive search for articles and
conference abstracts without language restriction. Experts and clinicians in the field were contacted to identify
additional data. Cohorts were eligible regardless of study
design if they included 73 children ,15 years treated for
MDR-TB and had treatment outcomes reported. Authors
were contacted to provide IPD on key factors, including:
HIV status, source case information, culture-confirmed
vs. clinical diagnosis, disease severity, drug susceptibility,
duration and use of drugs, sputum conversion and
treatment outcomes.
Results: The majority of data came from directly
contacting experts in the field; 28 authors and centers
contributed IPD data. The literature search yielded 2673
articles or conference abstracts; 201 full-texts were
reviewed; of these, 55 were eligible and from authors
we had not yet contacted. They were contacted and 3
provided IPD. 997 children from 18 countries were
included: median age 7 years; 40% were male. HIV status
was documented in 850; 40% HIV-infected. 60% of cases
were bacteriologically confirmed and 40% were clinically
diagnosed. 250 children had extrapulmonary TB. Culture-confirmed extensively drug-resistant TB was documented in 35 children. Successful treatment outcomes
(cured or treatment completed) were documented in
77%, while 1.5% failed treatment, 10% died, 6% were
S29
lost to follow-up and 5.5% were not evaluated. Most

regimens included pyrazinamide, prothionamide/ethionamide, a fluoroquinolone and a second-line injectable;
51% used cycloserine/terizidone and few used high-dose
isoniazid, clofazimine and linezolid. Analysis of associations of outcomes with individual drugs and treatment
durations, controlling for key covariates, is ongoing.
Conclusions: This IPD meta-analysis on MDR-TB
treatment outcomes in children promises to be a rich
source of information to help inform improved treatment
of children with MDR-TB.
Emerging data on PK and safety of levofloxacin

and amikacin informs care and design of new
regimens
A Garcia-Prats,1 H R Draper,1 S Thee,1,3 A Hesseling,1
H S Schaaf,1 P Denti,2 H McIlleron2 1Department of
Paediatrics and Child Health, Faculty of Medicine and Health
Sciences, Stellenbosch University, Desmond Tutu TB Centre,
Tygerberg, 2Division of Clinical Pharmacology, University of
Cape Town, Cape Town, South Africa; 3Department of
Paediatric Pneumology and Immunology, Charite
Universitatsmedizin
Berlin, Berlin, Germany.
Levofloxacin is currently the most frequently used

fluoroquinolone for multidrug-resistant tuberculosis
(MDR-TB) treatment and prevention in young children;
the current levofloxacin dose used in practice is 1015mg/kg once daily. The second-line injectable TB
medications are used routinely in MDR-TB treatment
regimens, at a recommended paediatric once daily dose
of 15-30 mg/kg. However the pharmacokinetics of
levofloxacin and the injectables have not been wellcharacterized in children with MDR-TB. The injectables
are associated with permanent sensorineural hearing loss
in up to 25% of children receiving them long-term, likely
related to the total cumulative drug exposure. An
ongoing observational study in Cape Town, South
Africa, is evaluating the pharmacokinetics and safety of
multiple second-line antituberculosis drugs in HIVinfected and uninfected children with MDR-TB. We
have previously reported data from this study on the
pharmacokinetics and short-term safety of levofloxacin
at a dose of 15 mg/kg in a limited number of young
children with MDR-TB (n 22), demonstrating a median
maximum plasma concentration (Cmax) of 6.79 mg/ml
and a median area under the concentration time curve
(AUC0-inf) of 32.9 lg*h/ml, which are well below those
seen in adults following the currently recommended 750
mg once-daily dose (Cmax 8.6 lg/ml, AUC0-24 90.7 lg*h/
ml). Preliminary data on amikacin pharmacokinetics in
28 children with MDR-TB at an intramuscular dose of
20 mg/kg showed a median Cmax of 47.1 lg/ml, which is
above the currently proposed target Cmax of 35-45 lg/ml
in a large proportion of children. Based on this data, we
have now evaluated the pharmacokinetics and safety of a
higher dose of levofloxacin (20 mg/kg) and a lower dose
of amikacin (15 mg/kg). We present here emerging data
from this study on the pharmacokinetics of levofloxacin
in .100 children with MDR-TB at doses of 15 mg/kg
and 20 mg/kg, using state of the art pharmacometric
S30
modeling. We will also present data on the pharmacokinetics and safety of amikacin given at both the 20 mg/
kg once daily and at a reduced dose of 15 mg/kg once
daily. We will discuss the implications of this data for the
design and evaluation of novel, more optimal, safer, and
better-tolerated regimens for the treatment and prevention of MDR-TB in children.
Compassionate use of bedaquiline and

delamanid in children with MDR-and XDR-TB:
early experiences and challenges
M Tadolini,1 L Dambrosio,2,4 G B Migliori,2,3 R Centis,2
S Esposito5 1Department of Medical and Surgical Sciences,
Alma Mater Studiorum University of Bologna, Infectious
Diseases Unit, Bologna, 2World Health Organization
Collaborating Centre for TB and Lung Diseases, Fondazione
S. Maugeri, Care and Research Institute, Tradate, Italy;
3
European Respiratory Society (ERS), Secretary General,
Lausanne, 4Public Health Consulting Group, Lugano,
Switzerland; 5Department of Pathophysiology and
Transplantation, Universita` degli Studi di Milano, Fondazione
IRCCS Ca Granda Ospedale Maggiore Policlinico, Pediatric
Highly Intensive Care Unit, Milan, Italy.
Multidrug-resistant tuberculosis (MDR-TB) is a serious

obstacle to TB control. Delamanid and bedaquiline are
novel anti-TB agents recently approved for the management of MDR-TB patients. Both of them have been
approved for adults only. Our group described a case of
compassionate delamanid use in a 12 years old child
(male, native Italian) who was diagnosed in October
2013 with laryngeal and pulmonary TB, with positive
smear on gastric aspirate. When the drug susceptibility
test (DST) results became available (showing resistance
to all first- and second-line drugs except para-aminosalicylic acid (PAS) and linezolid), the initial regimen
with ethambutol, pyrazinamide, high-dose isoniazid and
moxifloxacin was re-designed (Figure). After an initial
clinical and radiological improvement, the patient
experienced gastric aspirate direct smear reversion to
positive and progressive clinical deterioration. In order to
select the most appropriate treatment regimen, the ERS/
WHO TB Consilium platform was used and four
different world experts in paediatric MDR/XDR-TB
management were consulted. All of them recommended
the inclusion of one of the newly approved TB drugs in a
regimen including also meropenem and clofazimine.
Following the experts recommendations, delamanid was
procured in 10 days via the manufacturer and treatment
initiated, with close monitoring of the patient. After
treatment revision and delamanid introduction (Figure),
a rapid and sustained clinical and radiological improvement was observed, with gastric aspirate bacteriological
conversion. The regimen was tolerated very well with no
major side effects onset. This case shows that there is an
urgent need of expanding the access to new drugs for
children, through the development of a consensus-based
roadmap to ensure new drugs rational use. In our
experience, in the absence of a pre-existing national
authorization process for the compassionate use of an
unregistered drug (formally approved in Europe for
adults only), the TB Consilium experts opinion played a

crucial role in obtaining clearance from Italian regulatory bodies. Building on the newly published data on
pharmacokinetics, safety and efficacy of delamanid in
children, it is essential to stimulate similar clinical trials
on bedaquiline in pediatric population. At the same time
it is important to ensure rapid access to the new drugs
through compassionate programmes in selected cases, to
save patients lives and to break the transmission cycle.
28. End TB, end MDR-TB

Prioritisation: mapping the hotspotsfighting
TB in urban settings
R Orejel 1Programa Nacional Tuberculosis, Mexico,

DF,
Mexico.
Background: It is widely known that there is a higherburden of TB in urban areas than in rural areas. This is
largely because of the living conditions in poor marginal
neighborhoods. Various publications have pointed out
that a 48% of high-burden TB is not attributed to health
services, but rather to the social determinants of health.
The reason for this that the options of intervention should
be out of health services. However, it is necessary to build
and create alliances with the other (The National Crusade
against hunger Cruzada Nacional Contra el Hambre,
Department of water CONAGUA Housing sector, etc).
Design/Methods: 31 municipalities were selected that
have 100 or more cases of tuberculosis in all its forms. In
Mexico these municipalities account for 48% of all cases
in the country. An epidemiological mapping was carried
out for each municipality to select are as to intervene in
considering other epidemiological indicators such as low
detection rates or curing, high percentages of dropouts or
MDR-TB. From the selected municipalities, 31 coordinators in the TB Program were instructed to work on a
different work plan where social determinants can be
addressed and get existing health suppliers involved in
each municipality using the model Big Cities by
proposed PAHO/WHO. It was recommended to work
with private pharmacies, private clinics, schools, textile
factories and assembly plants, as well as Alcoholics
Anonymous, midwives, churches and others. The National Program is monitoring that the work plans are
being carried out. In the following September there will
be results of this initiative.
Results: In a period of 4 months, 50 respiratory
symptomatic have been detected along with 2 confirmed
cases.
Conclusion: The alliances with other sectors can

contribute to increasing case detection and bettering
the quality of life for those affected by TB.
Making the case: TIME and One Health

modelling the epidemic and the impact of
interventions
M Lalli,1 R Houben,1 A Gebhard,2 B H Nguyen,3
N V Nhung3 1London School of Hygiene & Tropical Medicine,
London, UK; 2KNCV Tuberculosis Foundation, The Hague,
Netherlands; 3National Tuberculosis Control Programme,
Hanoi, Viet Nam.
Mathematical modelling is increasingly being used in the

policy development cycle to generate evidence that can
inform policy decisions. Modelling provides a rational
framework in which assumptions can be made explicit
and combined with data, which can be used to support an
Investment Case for national and international funders.
However, mathematical models are often not accessible
for policy makers in low- and middle-income countries
due to the complexity and required expertise. The TIME
modelling suite was developed by The London School of
Hygiene & Tropical Medicine in partnership with Avenir
Health as a user-friendly modelling tool that can be used
in-country. TIME works in conjunction with One
Health, a costing and financing tool for strategic health
planning. TIME and OneHealth have been used in Viet
Nam to inform policy discussions and the development of
the Concept Note to the Global Fund to Fight AIDS,
Tuberculosis and Malaria. Through discussions with the
Viet Nam National TB Control Programme, KNCV and
LSHTM, local epidemiological and costing data were
collected and used to populate the TIME and OneHealth
models. The models were then used to generate
projections of the TB epidemic and associated costs,
with a focus on the potential impact of new MDR-TB
policies. We will present the methods and results of this
process, highlight achievements and lessons learned.
This work was initiated by the KNCV Tuberculosis Foundation and
funded by WHO.
29. Pneumonia in children and adults

beyond 2015
Pneumonia in adults in resource-limited
settings: an update
C Feldman Department of Internal Medicine, University of
the Witwatersrand, Johannesburg, South Africa.
A number of challenges exist with regard to the diagnosis

and management of community-acquired pneumonia
(CAP) in adults in resource-limited settings, not least of
which are a lack of adequate local data on which to base
important clinical decisions. Appropriate treatment
recommendations and guidelines are therefore not
usually available in these settings and patient management is often not standardized. In resource limited
regions, human immunodeficiency virus (HIV) infections
S31
and tuberculosis tend to be more common, so that CAP

occurring in these patients is a more complex clinical
problem, with a broader spectrum of possible pathogens,
including unusual and/or opportunistic pathogens. There
also tends to be a higher mortality, particularly in
severely ill cases, and especially among those with
advanced immunocompromise. Early diagnosis, assessment of disease severity and institution of appropriate
empiric antibiotic therapy are all important means of
decreasing the mortality from CAP; unfortunately each
of these face potential challenges in resource-limited
regions. With regard to likely microbial aetiology, a full
range of diagnostic tests is often not available. While
severity of illness scoring systems have been thoroughly
studied in well-resourced areas, few have been validated
in resource-limited settings. Furthermore a broad range
of antibiotics are often not available in resource-limited
situations. In contrast, studies have clearly indicated that
collection of local epidemiological data and the development of locally appropriate guidelines based on the
available information can go a long way to improving the
outcome from CAP in resource-limited settings.
HIV-related pneumonia: towards elimination

beyond 2015
M Kelly 1Department of General Pediatrics, Duke University
Medical Center, Durham, NC, USA.
More than 150 million episodes of pneumonia occur

each year among children. Most child deaths from
pneumonia occur in sub-Saharan Africa and South-East
Asia. In these regions, HIV remains a major obstacle to
reducing pneumonia-related mortality in children. Pneumonia is the leading cause of death among HIV-infected
children, and the aetiologies and treatment outcomes in
these children differ from HIV-negative children. HIVinfected children have a higher incidence of bacterial
pneumonia, particularly caused by Streptococcus pneumoniae, and are at risk for pneumonia caused by the
opportunistic pathogens Pneumocystis jirovecii and
cytomegalovirus. HIV-infected children with pneumonia
also have worse short-term outcomes and a higher casefatality than HIV-negative children. In some highly
endemic countries, the incorporation of antiretroviral
therapy into antenatal services has markedly reduced
mother-to-child HIV transmission and, in turn, led to
rising numbers of infants who are exposed to HIV in
utero but do not acquire the virus. Despite the absence of
HIV infection, HIV-exposed, uninfected children have
immune abnormalities and are also at higher risk of
pneumonia and death during early childhood. The
presenter will discuss the current global burden and
diagnostic and treatment strategies for pneumonia
among HIV-infected children as well as new data on
the aetiologies and outcomes of pneumonia among HIVexposed, uninfected children. Finally, strategies aimed at
eliminating HIV-related pneumonia beyond 2015 will be
reviewed, including the prevention or early identification
of infant HIV infection, cotrimoxazole prophylaxis, and
vaccination against respiratory pathogens.
S32
30. Eliminating the burden of TB in

indigenous populations
Latin America region disaggregates TB data on
indigenous populations: progress in Mexico
and Colombia
M Del Granado,1 S Guessoum,1 S Del Pino1
Communicable Disease and Health Analysis Department;
Gender and Life Course Department, Pan American Health
Organization, Washington, DC, USA.
Despite progress in terms of economic and social

development, Latin America continues to be one of the
most inequitable and unequal regions in the world. The
indigenous peoples of Latin America continue to face
discrimination and are left to live in poverty. In Latin
America there are 826 different indigenous groups with a
total population of 45 million. The health conditions of
the indigenous population indicate high levels of
morbidity and mortality with regards to diseases of
inequality such as tuberculosis (TB). TB remains a
serious public health problem in Latin America. In
2013, WHO estimated that there were 270 600 new TB
cases and 16 080 TB deaths. The Latin American
countries reported 208 400 new TB cases (77% of
estimated) for the same year. This specific analysis was
possible only in Brazil, Colombia, Mexico and Venezuela. These countries possess TB information systems with
nominal records, as well as the inclusion of the variable
of ethnicity. The analysis was done for 2010-2014. By
2010, Brazils indigenous population represented 0.5%
of the total population and 1% of new TB cases with RR
of 2.3; in Colombia 7% of all TB cases in the country are
concentrated in 3.4% in the indigenous population with
RR of 2; the TB incidence in Mexico in indigenous is
lower than the national incidence with a RR of less than
1. In Venezuela the analysis was carried out for 2010 to
2014, during which we observed a decrease of the TB
incidence in the country by 1.1% per year and 15.7% per
year for the indigenous population. This analysis was
limited, however by the lack of population projections of
indigenous and the bias linked to their self-identification.
This is possibly attributed to the double stigma of being
identified as indigenous and a TB patient.
Explaining trends in TB rates in North American

K Dehghani,1,7 E Robinson,1,7 B Soborg,2 J Picard,3
G Alvarez,4 M Cooper,5 R Long,6 Z Lan7 1 Public Health
Department, Cree Board of Health & Social Service of James
Bay, Montreal, QC, Canada; 2Statens Serum Institut,
Copenhagen, Denmark; 3Public Health Department, Nunavik
Regional Board of Health & Social Services, Kuujjuaq, QC,
4
Divisions of Respirology & Infectious Diseases, University of
Ottawa, Ottawa, ON, Canada; 5Department of Health &
Social Services, Government of Alaska, Anchorage, AK, USA;
6
Department of Medicine, University of Alberta, Edmonton,
AL, 7McGill University, Montreal, QC, Canada.
Introduction: The continued high incidence of TB in

indigenous populations of North America and Greenland
represents a major public health challenge. It is unclear

whether this should be addressed by intensifying TB
control and prevention interventions, or by addressing
the underlying social and economic disadvantages faced
by these indigenous populations.
Methods: This ecological study examined the trends of
tuberculosis notification rates from 1960 to 2014 in
indigenous populations in Nunavut, Nunavik, Eeyou
Istchee (Cree QC), Alberta, Alaska, and Greenland. The
association of these trends with major historical landmarks, mass TB control and prevention strategies (mass
radiographic screening, INH treatment and BCG vaccination of infants) other major health indicators (e.g. life
expectancy) was examined. Furthermore, the trends of
other health (e.g. diabetes), health system (e.g. health
budget), lifestyle (e.g. smoking) and socioeconomic (e.g.
crowded housing) indicators were examined from 1980s
on, as these indicators became more available. The TB
and other health and socioeconomic trends in these years
were compared within and between these populations,
and with the Canadian general population. This study
used population data already in the public domain. The
data collection included search of epidemiologic sources
(for example, census and surveillance data), systematic
search of scientific and medical literature (for example
through PubMed, and Emb-Base), as well as publicly
available grey or non-indexed literature (such as regional
public health reports and conference proceedings).
Regression analysis was used to estimate the change in
slope of notification trends associated with changes in
social and economic indicators as well as changes in TB
control activities.
Results: All indigenous populations had very high TB
notification rates in the 1950s and 1960s, which declined
dramatically to a nadir between 1985 to 1995. In three
Inuit populations - in Nunavut, Nunavik, and Greenland, there has been a substantial resurgence since 1995.
Possible reasons for this parallel resurgence, not seen in
the other three indigenous populations analyzed, will be
presented.
31. High burden, neglected patients:

addressing the intersection of
tuberculosis and mental disorders
Understanding and addressing the link
between depression and TB treatment
A Pasha,1 O Qureshi,1 S Khowaja,1 Z Hasnain,1
N Baig-ansari,1 A Khan1 1Interactive Research and
Development, Karachi, Pakistan.
Background: This study has added to the dearth of

existing literature on the relationship between a TB
patients mental health and their treatment. The study
aimed to measure the pattern of depression and anxiety
in TB patients during the course of their treatment and to
understand whether depression and anxiety increased,
decreased or stayed constant over the course of
treatment.
Design/Methods: This randomized, cross-sectional study
was conducted at the TB clinic at Indus Hospital in
Karachi. The treatment period for CAT 1 TB patients (6
months) was divided into five stages: Stage 1: Prediagnosis (before test results were attained); Stage 2:
Month 0 (i.e. day of enrollment/first day of TB
treatment); Stage 3: Month 2 of treatment; Stage 4:
Month 4 of treatment and Stage 5: Month 6 of treatment.
Fifty patients from each stage were screened using the
Aga Khan University Anxiety and Depression Scale
(AKUADS), resulting in a total sample size of 250
patients. All TB patients who screened positive for
depression and/or anxiety were offered free mental
health counseling immediately following their screening
as well as on-going weekly sessions, as required.
Results: Our results show that 48% of patients in Stage 1
screened positive for depression and anxiety, 52% in
Stage 2, 30% in Stage 3, 20% in Stage 4 and 24% in
Stage 5. In other words, the prevalence of depression and
anxiety in the pre-diagnosis and earlier stages of TB
treatment are significantly higher than those in later
stages. In all 5 stages, females were more likely to screen
positive for depression and anxiety, particularly in Stage
2, with 76% of females screening positive for depression
and anxiety as compared to only 28% of males.
Conclusion: Through this study, the impact of TB
treatment on mental health and vice versa was explored,
adding depth to the implications of possible interventions
that can be developed to target the maladaptive effects
and outcomes of this relationship. The interventions that
can be developed from these results have the potential of
strengthening mental health coping strategies for a TB
patients personal and environmental stressors during the
length of the treatment. We hope the results from this
study are utilized for the positive development and
implementation of psychosocial support interventions
that will improve the mental health of TB patients, and
that will ultimately increase adherence to treatment and
improve TB treatment outcomes.
S33
Depression and other mental illnesses: silent

drivers of the TB epidemic
S Annika,1 M Wainberg,1 D Boccia,2 A Karpati3
1
Columbia University, New York, NY, USA; 2London School of
Hygiene & Tropical Medicine, London, UK; 3International
Union Against Tuberculosis and Lung Disease, North America
Office, New York, NY, USA.
TB has long been considered as much a social as

biological disease; the same socioeconomic conditions
that increase risk for exposure contribute to both disease
reactivation and poor treatment outcomes, posing a
significant challenge to global TB control. Mental
illnesses such as depression, substance addiction, and
severe mental disorders, are also strongly associated with
poverty, and have been linked to all negative TB
outcomes including increased morbidity, mortality,
drug-resistance, and community transmission. As with
poverty, there may be a bi-directional association
between TB and mental disorders due to inflammatory
processes, psychiatric side-effects, and social stigma,
which may partially mediate poor TB outcomes. Treating
comorbid mental disorders may help to interrupt the
vicious cycle of poverty, mental illness, and TB by
increasing functional status and enhancing health behaviors that favor positive TB outcomes. This conceptual
framework offers new insights into the complex interaction between social, psychological, and biological factors
that challenge global TB control, with significant policy,
research, and clinical implications.
Sustainability of motivational counseling:

Integrated Management of Physician-delivered
Alcohol Care for TB patients (IMPACT) trial,
Tomsk, Russia
V Livchits,1 A Solovyeva,1 S Shin2 1Department of TB
Programs in Russia, Partners in Health, Boston, MA, 2Project
COPE, Partners in Health, Albuquerque, New Mexico, USA.
Objective: To perform an assessment identifying the

sustainability of brief counseling intervention (BCI) for
excessive alcohol use among tuberculosis patients within
the Tomsk TB service after the completion of the RCT
Integrated Management of Physician-delivered Alcohol
Care for Tuberculosis Patients (IMPACT) research
phase.
Methods: We conducted audio-taped individual interviews using semi-structured questionnaires after obtaining verbal informed consent with four TB practitioners
from Tomsk TB Services out of six who had participated
in the RCT. Interviews were transcribed, coded with a
standardized framework, and analyzed to identify
emergent domains. We used domain analysis to describe
the TB doctors perceptions of the multi-level factors that
influence BCI practice in post-RCT phase and illustrate
current situation in providing care to TB patients with
alcohol use disorders (AUD).
Results: The study found that all participants discontinued BCI at some point and the motivational interview
(MI) skills were not sustained after the completion of the
research. TB doctors were unanimous in describing how
S34
difficult it was to maintain alcohol interventions since the

completion of the RCT working with TB patients with
AUDs and to ensure compliance with the intervention.
Although the doctors considered the doctor-patient
relationship strong enough to sustain targeting of alcohol
use, they often failed to follow up on an initial patient
AUD assessment. They identified key barriers providing
BCI systematically, including: high workload, low staff
motivation, lack of booster trainings, and lack of interest
from the administration. Doctors suggestions for strategies that could address these barriers include: assign a
counseling coordinator for: setting up each patients
management map at the hospital entry point, monitoring
accomplishments, and coaching new medical staff;
establish recognition and incentives for good counseling
practices; provide intensive booster series of trainings.
Discussion: The improvement in the MI skills achieved
during the research period is ceased after the research is
phased out. Hence, the patients are barred from the
beneficial effect of the intervention. Though the TB
doctors believe that BCI is an effective and convenient
strategy to work with TB patients with ADU, the
intervention could not be continued, as it has not been
incorporated in the formal standard of care.
32. Does MPOWER empower women?

Why women should be targeted for tobacco
control
M Aghi 1Healis Sekhsaria Institute for Public Health, New
Delhi, India.
Background: The WHO estimates 80 000100 000

people initiate tobacco use on daily basis; a significant
number among them are females. In India, the increasing
trend of tobacco use among females is indeed being seen.
A secondary analysis of the Global Adult Tobacco Survey
(GATS, India 20092010) data reported initiation of
tobacco use among males and females at age of 12 years.
More females (5%) started smoking at age 12 than men
(3%) at the same age. Twelve percent of females started
using smokeless tobacco products at age of 12 years, as
opposed to 4% of males. Also, data shows that 44%
females start dual tobacco use at same age compared to
32% of males. Additionally, on average, an Indian
woman starts to smoke at 17.5 earlier than her male
counterpart. A recent journal article shows tobacco
prevalence among Indian women doubled during the
period 1995 to 2009 from 10 to 20%. Smokeless tobacco
use has also increased over time. This significant rise in
tobacco use is to a great extent due to the relentless and
unchecked work of the tobacco industry whose sole
objective is making money. This is happening despite the
Indian tobacco control legislation, which has provisions
for smoke-free public places, bans on tobacco advertising, and prohibition on sale to minors. Women leaders in
tobacco control have traced this worrying trend to the
fact that tobacco control policy makers have not paid
sufficient attention to the unique problems of girls and

women.
Conclusion: FCTC could be a critical tool for achieving
the goal of reducing tobacco use among women,
provided it keeps a clear commitment to gender with a
focus on womens and girls specific issues and makes it a
basis for action in the 21st century.
What gender-specific research is needed to

inform and evaluate the elements of
MPOWER?
A Amos Usher Institute of Population Health Sciences and
Informatics, University of Edinburgh, Edinburgh, UK.
Global female tobacco use is increasingly complex,

involving diverse products and factors including tobacco
marketing, globalisation, urbanisation and changes in
womens social and economic status. In high income
countries (HICs) female smoking is declining but
increasingly concentrated among disadvantaged women.
In low and middle income countries (LMICs) the pattern
is more complex; in several regions the gap between girls
and boys smoking is narrowing. The past twenty-five
years have seen numerous calls for action on women and
tobacco, and for considering the impact of tobacco on
womens health and economic well-being. However,
tobacco control research, policy and programmes have
remained largely gender blind. There has been little
recognition of the importance of understanding the
context and challenges of girls and womens smoking,
their exposure to secondhand smoke, and the increase in
womens smoking in LMICs. Thus many opportunities
for knowledge generation and effective responses remain
lost or overlooked. This presentation will consider what
is meant by gender-specific research and why such
research is needed to inform both the development and
evaluation of the different elements of MPOWER. It will
argue that tobacco control strategies aimed at addressing
the female smoking epidemic need to be grounded in a
multi-faceted understanding of womens smoking, in
order to create a more robust research base for
developing interventions for girls and women. This will
help animate the WHO-FCTCs concern for gender
specificity and womens leadership, and reduce the
impact of tobacco on women.
Trends in global tobacco prevalence and use

among women
J Tonsing,1 A Pandey1 1International Union Against
Tuberculosis and Lung Disease, South-East Asia Office, New
Delhi, India.
Background: Globally, about 40% of men smoke

compared with nearly 9% of women. However, the
epidemic of tobacco use among women is increasing and
there is growing concern among policy makers with
respect to alarming smoking prevalence among women.
The WHO Framework Convention on Tobacco Control
has emphasized the need for gender specific tobacco
control strategies in their tobacco control programs.
About 250 million women in the world are daily

smokers. In developed countries 22 percent and in
developing countries 9 percent of women smoke tobacco.
Of the more than 5 million people who die every year
from tobacco use, approximately 1.5 million are women.
Most (75%) of these women live in low- and middleincome countries.
Methodology: Using secondary data available in public
domain including survey reports, publications and peer
reviewed papers trend among global tobacco prevalence
and use among women has been analysed.
Results: Cigarette smoking among women is declining in
many developed countries, notably in Australia, Canada,
UK and USA. But this trend is not similar across all
developed countries or in the developing countries. In
several southern, central and eastern European countries
cigarette smoking is either still increasing among women
or has not shown any decline. In India female smoking
has increased more than double (1.4 2.9%) during the
period 20052009. Further survey of smoking trends in
youths showed that in half of the 151 countries surveyed,
similar numbers of girls and boys smoked. Evidence
suggests that most of these girls and boys will continue to
smoke into adulthood. Unless urgent action is taken,
tobacco use could kill up to 8 million people every year
by 2030, of which 2.5 million would be women.
Recommendation: There is an urgent need to institute
gender based tobacco control policies to stop the rise in
womens smoking rates. Tobacco prevention and cessation programmes should be integrated into maternal,
child and reproductive health services.
33. Modelling to support TB control

policy in the era of the End TB Strategy
TIME TB Model: a new publicly available TB
model to support country-level TB control
policy
R Houben,1 D Pedrazzoli,1 M Lalli,1 R White,1
M Hamilton,2 C Pretorius2 1London School of Hygiene &
Tropical Medicine, London, UK; 2Avenir Health, Glastonbury,
CT, USA.
Mathematical modelling can generate key inputs into

developing rational TB policy decisions on the global,
country or sub-national level. In addition, by providing a
logical framework that brings together stakeholders
(NTP, funder, community representatives, research community), the available data and necessary assumptions,
modelling can improve communication and shared
understanding of the local epidemiological situation.
Through these discussions, gaps in data can be highlighted, driving local data collection and operational
research. The TIME modelling suite was developed to
provide this framework and support decision making at
the local level. In addition, it has a user-friendly interface
which allows for capacity building and increased local
ownership of the model and result. TIME has been used
as part of Global Fund applications for a number of
S35
countries, and is also part of a number of long-term

collaborations with countries, illustrating the value of
TIME and modelling throughout the grant application
and implementation cycle. The presentation will highlight the structure of the TIME models, the importance of
data to inform the model, and illustrate the use of TIME
through a number of case-studies.
Experience using models to support TB control

policies in Viet Nam
R White,1 M Lalli,1 R Houben,1 A Gebhard,2 H T T Thuy,3
N V Nhung,3 B H Nguyen3 1London School of Hygiene &
Tropical Medicine, London, UK; 2KNCV Tuberculosis
Foundation, The Hague, Netherlands; 3National Tuberculosis
Control Programme, Hanoi, Viet Nam.
Previously, the Viet Nam National TB control programme (VN NTP) set targets for the next years based on
the historical case notification data and the amount of
budget used in previous years only. Ambitious targets
were needed to attract national and international
funders. In order to develop a more comprehensive
strategic plan to mobilize funding and achieve the set
goals, Viet Nam needed a scientific sound assessment and
projection for the TB control programme objectives,
with clearly identified and justified activities. Without a
model to create projection of TB epidemic in Viet Nam, it
was not possible to anticipate what activities Viet Nam
could achieve the set goal to decrease the TB prevalence
to 131 per 100 000 population by 2020. In 2014, the
TIME Impact and One Health tools were applied to
inform the VN NTP strategy and funding applications,
with support from KNCV. Epidemiological and costing
data of the VN NTP were used to populate the TIME
Impact and OneHealth models, generate projections of
the TB epidemic and associated costs for the future plans.
Two areas were focussed on: case finding and new MDRTB policies. The baseline projections indicated that after
a continuous decline in trend between 1990-2010, the TB
burden was due to start increasing after 2020 if NTP
activity levels remained stable. Furthermore, the model
showed that, even with strongly increased efforts,
notifications will continue to decline due to the
background decrease in prevalence, and pushing with
additional case finding is key to continue decline of TB
burden. Accordingly, the VN NTP to increase case
finding to find more TB cases in community with the set
objective to increase case detection rate by 8% by 2017,
to continue the ongoing decline in TB incidence,
mortality and prevalence by 2%, 4% and 4% respectively. In addition, a scale-up of Drug Sensitivity testing
and introduction of a new 9 month second line drug
regimen could become cost-saving. These results have
been used to develop the Concept Note to the Global
Fund to Fight AIDS, Tuberculosis and Malaria. The
Concept Note was highly appreciated and approved by
the funders to provide additional support beyond the
country allocation to scale up the programmatic management of drug resistant tuberculosis (PMDT).
S36
Modelling to explore the cost-effectiveness

and resource implications of reaching the post2015 targets
N Menzies,1 G Gomez,2 A Vassall3 1Department of Global
Health and Population, Harvard T.H. Chan School of Public
Health, Boston, MA, USA; 2Amsterdam Institute for Global
Health and Development, University of Amsterdam,
Amsterdam, Netherlands; 3Department of Global Health,
London School of Hygiene & Tropical Medicine, London, UK.
The post-2015 End TB Strategy formalizes goals for

aggressive TB control. One outcome anticipated by the
strategy is reduction in global TB incidence and
prevalence by 50% and 75% respectively by 2025. A
study organized by the TB Modelling and Analysis
Consortium (TB-MAC) tested the feasibility of these
goals when applied to three major high-burden countries
China, India, and South Africa. This study was
implemented using computer models of TB epidemiology
to project long-term policy outcomes. To address
concerns that individual models may provide biased
results we invited participation from multiple independent research groups. We developed cost projections to
provide information on resource implications and costeffectiveness. Eleven research groups participated in the
study. We examined six discrete intervention approaches,
assuming interventions would be scaled up to high (but
feasible) coverage by 2025. Intervention scenarios were
developed by country stakeholders, based on current
technology. Implementation differed by country, considering local policy preferences / capacity. The costing
approach, using country-specific cost models, was
designed to accommodate differing capabilities of
participating models and estimate results for multiple
perspectives, using a synthesis of empirical cost data.
Health impact (DALYs averted) and resource implications (total costs) were estimated for 2015-2035. We
estimated resource needs, over time and by payer, as well
as the cost-effectiveness of competing policy scenarios.
Policies were compared against each other and against a
status quo policy that held TB control activities at
current levels. Compared to the status quo, most
interventions appeared cost-effective when evaluated
against conventional willingness-to-pay thresholds.
However, many interventions required substantially
increased budgets, and the opportunity cost of increased
spending (or reallocation within the TB budget) is likely
higher than implied by conventional thresholds. Incremental costs were generally front-loaded, with rapid cost
increases during scale-up. In contrast incremental health
gains accumulated more slowly, and cost-savings due to
interruption of transmission followed a similar pattern.
Interventions had important implications for financial
protection: while lower TB burden benefits households
over the long-term, results suggested an initial increase in
costs borne by TB patients and households as interventions were scaled up.
34. MDR-TB: a complex disease requiring

a comprehensive health system response
A typical patients journey: how the health
system fails the patient
M Loveday,1 A Voce,2 N Padayatchi,3 J Brust,4
K Wallengren5 Health Systems Research Unit, South African
Medical Research Council, Cape Town, 2Department of
Public Health, University of KwaZulu-Natal, Durban,
3
Department of CAPRISA, University of KwaZulu-Natal,
Durban, South Africa; 4Department of Medicine, Montefiore
Medical Center & Albert Einstein College of Medicine, Bronx,
NY, USA; 5TB and HIV Investigative Network (Think), Durban,
South Africa..
We describe hospital performance from the patient

perspective by the use of a graphic which visually
represents a patients treatment journey. The graphic
was used to report study findings to study sites and as a
catalyst for a quality improvement process.
Patient and health care worker experiences of

MDR-TB management
M P Mbiko1 1Department of Health, Bizana, South Africa.
Although there has been an attempt to decentralise

MDR-TB services by the South African Department of
Health, access to services and the costs incurred in
accessing services remain challenges for MDR-TB
patients. And, even if a patient gets to the hospital, one
of the many different components needed to provide an
optimal service may be unavailable, compromising
treatment. We share patient and healthcare worker
experiences of MDR-TB services.
35. Controversies and critical issues in

clinical trial design for TB
What are the meaningful endpoints in phase 2
and phase 3 trials?
S Dorman1 1Johns Hopkins University School of Medicine,
Baltimore, MD, USA.
The approaches used in recent years to assess new

tuberculosis drug regimens are lengthy, inefficient, and
expensive. This talk will set the stage by providing an
overview of commonly-used approaches, including

potential strengths and weaknesses as well as gaps in
scientific knowledge and implementation that limit
progress. This talk will then focus in on Phase 2 and
Phase 3 clinical trials and explore strategies that might be
helpful in overcoming some of the challenges.
Pragmatic trials and other alternatives to

standard RCTs
A Nunn1 Medical Research Council Clinical Trials Unit, UCL,
London, UK.
Late stage clinical trials are increasingly expensive and

time consuming. This is particularly true when small
differences in efficacy are being compared or noninferiority trials are designed with relatively small
margins of non-inferiority. If we do conduct a large scale
trial how relevant are the results? Experience has shown
that results in clinical trials may not be readily
demonstrated in real life situations. The patient population in a clinical trial is invariably restricted and the
settings under which trials are conducted are often far
removed from every day experience. Is it possible for our
trials to more closely reflect real life conditions? Can we
be more efficient in our designs? Do we always need
trials? Why not well conducted cohort studies? What are
the limitations of alternative study designs? The opportunities and challenges presented by alternative approaches will be discussed.
Building clinical trial capacity in low-resource,

high-burden settings
I D Rusen1 1Department of Research, International Union
Against Tuberculosis and Lung Disease, North America
Office, New York, NY, USA.
High quality, programmatically-relevant research is

increasingly recognized as an essential component to
guide tuberculosis policy and program decisions in highburden settings. Clinical trials, in particular, require a
specialized site capacity to undertake scientifically sound
research. The available sites with existing capacity for
clinical trial implementation within the tuberculosis field
is limited. As a result, STREAM Clinical Trial partners
have worked with sites not previously engaged in MDRTB clinical trials - or in some cases - any previous clinical
trials. Through routine protocol training, additional
GCP training and intensive monitoring and support,
STREAM trial sites have performed well to date and will
serve as a growing and necessary pool of clinical trial
sites in low-resource, high-burden settings. Regulatory
clinical trials may bring particular challenges in these
settings compared to pragmatic clinical trials and require
additional capacity building efforts. Attention to all
aspects of trial implementation, including the laboratory
component, data management and supply procurement
have been important features of building clinical trial
capacity within the STREAM trial.
S37
Ethical challenges in TB research

J Denholm1,2 1Victorian Tuberculosis Program, Melbourne,
VIC, 2Department of Microbiology and Immunology at the
Peter Doherty Institute for Infection and Immunity, University
of Melbourne, Melbourne, VIC, Australia.
From the very beginning, ethical controversies have

abounded in medical research. Likewise, the history of
TB research includes a number of examples of ethically
problematic episodes. Emerging clinical and public
health challenges around TB demand new research and
practice innovations, but rapid development can also
bring with it unexpected ethical tensions. Rather than
simply waiting for ethical problems to arise, the
development of new research methodologies should
include proactive consideration of potential moral
tensions, and monitoring and evaluation of research
should be alert for emerging ethical themes.
36. Innovations to improve the TB/HIV

patient experience: from diagnosis to
treatment
The potential and pitfalls of using Xpert MTB/
RIF to detect TB in people living with HIV
(PLHIV): what have we learned from the
African experience?
M Casenghi,1 C Jennifer (check),1 H Huerga,2 E Fajardo,3
`
E Goemere3 Access Campaign, Medecins
Sans Frontieres
(MSF), Geneva, Switzerland; 2Department of Epicentre, MSF,
Paris, France; Southern Africa Medical Unit, MSF, Cape Town,
South Africa.
In 2013, an estimated 1.1 millian HIV-positive new

tuberculosis (TB) cases were reported globally, and subSaharan Africa carries about 78% of this dual epidemic.
TB remains the leading cause of death among PLHIV but
TB diagnosis is often missed or delayed in this patient
population due to limitations of diagnostic tools
available in resource limited settings. In December
2010, WHO recommended Xpert MTB/RIF (Xpert) as
the initial test for people at risk of HIV-associated TB and
MDR-TB. Xpert simultaneously detects TB and resistance to rifampicin and can be placed at intermediary
level of laboratory networks. Primarily designed to work
on sputum samples, Xpert pooled sensitivity for detection of TB amongst PLHIV and smear negative cases is
79% and 67%, respectively. Clinical trials assessing
impact of Xpert used as initial test compared to routine
diagnostic algorithms showed that use of this test
increased the proportion of bacteriologically confirmed
TB cases among people initiated on treatment and
reduced the proportion of people with confirmed TB
who did not start therapy. Xpert also reduced time to
treatment initiation. However, no impact on the number
of people initiated on treatment (suggesting that Xpert
replaced rather than complemented empirical treatment)
and on TB-related mortality was observed. Availability
of chest X-Ray and high rate of empirical treatment in
S38
most study settings might have undermined Xpert

impact. More significant effect might be observed in
settings where initiation of TB treatment on empirical
basis is not common practice. In addition, the availability
of HIV assays to be performed on GeneXpert platform
offers potential to catalyze better integration of HIV and
TB services. Despite these advantages, suboptimal
sensitivity on smear negative cases and weak performance on extra-pulmonary TB (EPTB) samples remain
key Xpert limitations and management of test negative
cases is challenging. In conclusion, Xpert offers key
advantages for the diagnosis of HIV-associated TB.
Increased rates of bacteriological confirmation and
upfront testing for drug resistance can facilitate rapid
treatment initiation and selection of most effective
regimen, with potential to reduce transmission. However, diagnostic tests with improved sensitivity on smear
negative cases, better performance on EPTB and suitable
for further decentralization are needed in order to further
improve diagnosis and timely, effective treatment of TB
in PLHIV.
Strengthening specimen referral and transport

networks in resource-limited settings: a
Nigerian pilot to improve diagnosis of TB in
PLHIV
S Useni,1 M Gidado,1 J Onazi,1 S Gande,1 P Nwadike,1
S Massaut,2 S Kik,2 G Akang3 1KNCV Nigeria/Challenge TB,
Abuja, Nigeria; 2KNCV, The Hague, Netherlands; 3National
TBL Program, Abuja, Nigeria.
Background: The use of Xpert MTB/RIF has proven to

be effective in rapid diagnosis of M. tuberculosis and
rifampicin resistance (RR); however the programmatic
use, especially among PLHIV, requires specific programmatic/referral arrangements.
Objective: This pilot project aimed to develop a practical
model for specimen referral and transportation thereby
increasing the number of PLHIV screened, tested for TB
and put on treatment.
Methodology: A 20-month project at two sites was
implemented in six phases; rapid assessment and site
selection; pilot model design; stakeholders sensitization;
pilot implementation; supervision; M&E, data collection, analysis and reporting. A hub and spoke approach
was used to link a tertiary or secondary facility with an
Xpert machine to at least 10 peripheral facilities. A
specimen transportation system was supported to ensure
sputum specimens were brought twice a week to the
Xpert machine and hardcopies of the results returned to
the referring facility.
Results: At baseline, only 9 PLHIVs were tested with
Xpert and in pilot period 782. There was a 1.9 fold
increase of notified bacteriological confirmed TB cases;
at baseline 180 cases compared to pilot of 347 cases. In
the pilot period 1232 TB cases were registered for
treatment (including 497 PLHIV), 824 of these were
bacteriologically confirmed cases (including 354
PLHIV). Overall, 60% (1066 out of 2665) of all sputum
specimens that were tested with Xpert were from referral

sites.
Conclusion: A well-coordinated sputum transportation
system to Xpert testing sites minimizes the movement of
patients and increases access to diagnostic and care
services; not only for TB but also HIV.
Are TB-HIV one-stop shops effective for rapid

TB diagnosis and treatment of PLHIV?
R Ncube,1 C Zishiri,1 M Nqobile,1 K Charambira,1
R Dlodlo,2 S Hove,3 P Shiri,4 C Sandy4 1Department of
TB-HIV, International Union Against Tuberculosis and Lung
Disease, Harare, Zimbabwe; 2Department of TB-HIV,
International Union Against Tuberculosis and Lung Disease,
Paris, France; Health Department, Bulawayo City Council,
Bulawayo, 4Department of AIDS and TB, Ministry of Health
and Child Care, Harare, Zimbabwe.
Background: Tuberculosis (TB) remains a disease of

public health importance in Zimbabwe. Despite gains in
curtailing the scourge, the country still ranks among the
top 22 high burdened countries. Estimated incidence in
2013 was 552 per 100 000, HIV predominantly driving
the epidemic with co-infection rates as high as 69%.
Integrated TB-HIV patient care has in the past remained
constrained by centralized TB diagnostic services, with
presumed TB patients accessing services from district and
provincial referral facilities. In addition, provision of
anti-retroviral services has predominantly been doctor
led.
Intervention: In 2007, the Ministry of Health and Child
Care, with funding from European Union and technical
support from The Union piloted a nurse-led decentralized One Stop Shop TB-HIV care model in three high
volume urban clinics in Zimbabwe. Between 2012 and
2014, with additional funding from the United States
Agency for International Development, this model was
rolled out to 23 additional facilities in 17 urban
communities. The scope of support included minor
refurbishments to improve patient flow and infection
control; training and post training mentorship of health
care workers on Provider Initiated HIV Testing and
counselling, co-trimoxazole preventive therapy; antiretroviral treatment (ART) initiation and follow up for
co-infected patients. A dedicated specimen transport
system was established through a network of motorcycle
transporting sputum and other HIV related specimens.
All supported sites were equipped with Point of Care
CD4 machines to improve timely treatment monitoring
and recording and reporting were strengthened.
Results: There was a sustained increase in TB-HIV
service uptake indicators across all sites supported.
Notably, ART uptake among co-infected TB patients
increased by 14% point from 70% in the January-March
2013 cohort to 84% in July-September 2014. Turnaround time for TB diagnostic results plummeted from
more than a week before the support to the current 24
48 hours.
Conclusions: A Nurse-led One Stop Shop TB-HIV
diagnostic and treatment service is feasible in a resource
limited primary health care setting. The country has since
mobilized additional funding for decentralization beyond the urban primary care setting.
Automated reporting of TB diagnostic results:

progress, challenges and country experiences
A Habib 1Interactive Health Solutions, Karachi, Pakistan.
Background: A key factor in successful treatment of TB is

the early start of treatment. Not only does early
treatment start improve chances of treatment success,
but in the case of pulmonary TB, it also means that that
the index patient is likely to infect fewer people around
them. Rapid diagnostic devices like the GeneXpert have
reduced the time to diagnosis, but given that the results
seen by a lab technician can still take a long time to be
seen by someone who can take action; this rapid
diagnosis has not always been accompanied by a rapid
start of treatment
Intervention: Software to automatically transport GeneXpert results was developed as three independent open
source initiatives. These applications were able to receive
results from the GeneXpert as soon as the test was
completed and transmit them to a central server over the
Internet or SMS. They were also integrated with
commonly used medical record systems to create a
linkage to care. Alerts could be generated from the server
and patients, caregivers, and program staff could be
notified about test completion and the results. Additionally, the applications also captured cartridge information
and error information allowing inventory and best
practices to be remotely monitored.
Results: These applications have been deployed in more
than 20 countries with several others in preparatory
phases. Over the past three years these systems have been
used to collect some 400 000 results. While a systematic
impact assessment has not been conducted to date, those
using the systems have claimed that their access to data
and the ability to monitor all of the machines centrally
and remotely are a huge advantage for TB programs. The
independent developers of the three systems have
converged onto a single open source platform to make
it simpler for countries to deploy these systems. Other
similar systems have also been developed by commercial
vendors.
Conclusion: Systems to automatically GeneXpert results
have been widely deployed and generally well-received.
These systems speed the rate at which test results are
made available thereby allowing treatment to be started
faster and allow remote monitoring of a countrys
network of GeneXpert machines.
37. Inhaled therapies for tuberculosis

Ways and means for inhaled drug delivery in TB
S Giovagnoli,1 A Schoubben,1 M Ricci1 1Department of
Pharmaceutical Sciences, University of Perugia, Perugia, Italy.
Inhalation is unarguably the best way to achieve local

drug accumulation in the lungs. Such strategy is suitable
S39
to treat lung infections, especially TB, being systemic TB

therapy affected by high dosages and lengthy treatments
with high risks of toxic effects and low patient
adherence. Here we discuss some of the strategies to
achieve antiTB drug (ATD) pulmonary delivery useful for
vaccination as well with a focus on their potential impact
on TB therapy and market. Increased evidences support
inhalation as the only way to obtain high lung
concentrations of multiple ATD at the desired pharmacological ratio, otherwise impossible to achieve through
systemic administration, with limited increased costs
compared to the oral route, in light even of the
availability of cheap disposable devices. Based on the
evidence that the search for novel and more potent, but
often more toxic, ATD cannot be the only solution to the
TB problem, we believe that a shift in the approaches to
TB treatment is needed and that such change relies on the
development of effective inhalation protocols. In particular, we propose the hydrophobic strategy based on drug
modification through ion pairing and complexation to
improve inhalable drug penetration into tissues and
intracellular activity. A few examples include metal
complexes and ion pairs of second line peptide and
aminoglycoside ATD (capreomycin, amikacin, kanamycin) obtained with palladium, dehoxycholic acid (DCA)
and oleic acid. Moreover, vaccines and the microbiome
strategy aimed at enhancing prevention and control of
the infection and to limit the progression of the disease
show some potential when delivered through inhalation
in animal models. In addition, a few clinical trials are
exploring safety and efficacy of such procedures. The
proposed approaches provide multiple advantages. On
the one side, the physical modification of the drugs is
intrinsically advantageous compared to chemical synthesis of new compounds as it may grant shorter bench-tomarket time and repurposing of old actives. On the other
side, delivery of microbiomes and vaccines through
inhalation can grant better control of TB infections and
resistance development. The technology required to
develop TB inhalation therapy is already available and
the higher cost of production compared to oral systems is
balanced by higher efficacy. A global and concerted
effort is however required to speed up innovation in this
field.
Clinical, epidemiological and patient

compliance factors affecting deployment of
inhaled therapies for TB in the field
B Fourie,1 R Nettey2 1University of Pretoria, Pretoria, South
Africa; 2Northwestern University, Chicago, IL, USA.
Background: Patient and system delays in low income

countries are important determinants of success in TB
control programmes. Curing patients after diagnosis is
subject to rigid compliance with a treatment regimen that
may last as long as 24 months. Simplification of regimens
and treatment-shortening are high priorities. Inhaled
therapies offer major opportunities for lower dose
adjunctive therapy to existing regimens. Recent research
provides strong evidence for enhanced efficacy at lower
S40
dose of several anti-TB agents formulated in dry powder

form. Our aim in this study was to gain some degree of
understanding as to whether inhaled drugs delivered by a
simple breath-activated inhaler instead of injections
would be favoured by health care workers overseeing
treatment of MDR-TB patients.
Methods: We developed and applied a short structured
questionnaire to 48 health care service providers
(doctors, nurses, or treatment supporters) who had
current experience in managing TB patients and were
familiar with administering injections to patients with
multi-drug resistant TB in four South African provincial
hospitals/clinics. This opinion-based survey focused
largely on the providers experience with administering
injections for TB treatment. We introduced to them, in a
descriptive manner using a Plastiape RS01 monodose dry
powder inhaler, the concept of inhalation therapy. The
questionnaire also solicited opinions about ease of use,
product acceptability, predictors of obstacles to adoption, and suggestions for product improvement.
Results: Of the population of service providers surveyed,
almost all respondents (92%) felt patients would be
comfortable using the inhaler and most (84%) found the
inhaler easy to use. Most of the respondents (88%)
indicated that they would probably use inhaled therapies
for TB in patients if it was available. There were some
concerns about implementing the inhaler, with 22%
finding its durability on repeated use somewhat questionable and 26% querying whether the device might
pose a risk for cross-infection or recurrent self-infection
if used repeatedly.
Conclusion: Dry powder inhaled agents delivered by
simple devices, instead of injectable agents, might be
feasible and acceptable to clinical staff managing TB
patients. The educational gap that distinguishes the array
of patients normally requiring injectable anti-TB agents
in low-resource settings would need to be taken into
consideration when such technologies are introduced.
38. Affordable solutions to indoor air

pollution for one-third of the worlds
population
Indoor air pollution and the health of children
under five in Bangladesh
M Khalequzzaman,1 M Kamijima,2 K Sakai,3 B A Hoque,4
N A Chowdhury,5 T Nakajima6 1Department of Public
Health and Informatics, Bangabandhu Sheikh Mujib Medical
University, Dhaka, Bangladesh; 2Department of Occupational
and Environmental Health, Nagoya City University Graduate
School of Medical Sciences, Nagoya, 3Department of
Environmental Health, Nagoya City Public Health Research
Institute, Nagoya, Japan; 4Environment and Population
Research Centre, Dhaka, ; 5DFID, British High Commission,
Dhaka, Bangladesh; 6Department of Occupational and
Environmental Health, Nagoya University Graduate School of
Medicine, Nagoya, Japan.
Indoor air pollution levels are reported to be higher with

biomass fuel, and a number of respiratory diseases in
children are associated with pollution from burning such

fuel. For their domestic source of energy, 92% of
Bangladeshs people depend on biomass fuel. Indoor air
pollution is responsible for an estimated 3.6% of the
overall disease burden in the country. About 25% of
deaths among children under 5 years old in Bangladesh
are associated with acute respiratory tract infection
(ARTI). Switching from biomass to fossil fuel is thus
encouraged in Bangladesh, as pollution levels are
believed to be higher with biomass fuel. However, little
is known about the situation in developing countries. In
our studies we checked the indoor air concentrations of
volatile organic compounds (VOCs), carbon monoxide
(CO), carbon dioxide (CO2), nitrogen dioxide (NO2),
and dust particles. The measured pollution concentrations shows that CO were found to be significantly
higher in biomass fuel users, while concentrations of
benzene, xylene, toluene, hexane, total VOCs, and NO2
were significantly higher in the fossil fuel users.
Considering the seasonal variation, levels of VOCs such
as benzene, toluene, and xylene, were higher in winter
than summer. On the other hand, levels of CO, CO2 were
found to be higher in summer than winter. We also
checked the geographical distribution of the indoor air
pollutants. The pollutant concentrations were higher in
urban kitchen than rural ones. The health impacts of the
mentioned indoor air pollutants were checked among the
children under five years old in Bangladesh but could not
attribute the occurrence of respiratory symptoms among
children. Other factors may be involved.
39. Taking stock of TB epidemiological

assessments: experience, lessons learned
and the way forward
The NTP perspective: from recommendation to
implementation and policy change
C Widaningrum Sub Directorate of TB Control Program,
Ministry of Health Republic of Indonesia, Jakarta, Indonesia.
Background: Indonesia is one of the countries with the

highest TB burden in the world. TB burden has been
estimated from different sources: surveys, modeling
estimates, and notification. Series of activities and
discussion were conducted to evaluate TB epidemiological situation in Indonesia and to integrate them into TB
control program policies
Methods: Joint External Monitoring Mission (JEMM)
was conducted in February, 2013 to assess TB control
situation in Indonesia. Assessment on the surveillance
system and available data to measure the trend of
incidence, prevalence, and mortality was part of the
JEMM activities. TB NPS was conducted on 2013-2014.
In 2014, National TB Control Program invited WHO
technical advisors and National Institute of Health
Research and Development (NIHRD) to update TB
burden estimates based on NPS results, notification data,
and available data on TB mortality from sample
registration survey of cause of death. Indonesia was
developing National Strategy for TB Control at the end

2014.
Results: TB incidence, prevalence, and mortality and
their trends were estimated in 2013 and updated with the
analysis of TB NPS in 2014. TB prevalence, TB
incidence, and TB mortality was estimated to be 660,
403, and 42 per 100 000 persons respectively. It was
estimated that 1 600 000 TB cases existed in Indonesia in
2013. The updated estimates were much higher than the
previous estimates. Gaps on case detection and case
notification were identified. The updated estimates were
utilized to develop the strategic plans of TB control. As it
was higher than previously estimated, it posed important
strategic and budget implication. Intensified case finding
and TB mandatory notification were put as important
strategies. Case based electronic registration was improved. Challenges remains that TB surveillance system
has not covered most health facilities which manage TB
cases, cause of death reporting was not routinely done,
and under notification of TB-HIV, drug resistant TB,
pediatrics TB is high.
Conclusion: TB epidemiological assessment in Indonesia
is a dynamic process. Recommendations resulted from
the assessment were taken into account for policy
development. Efforts are needed to close the surveillance
gaps so TB burden can be estimated from routine
surveillance.
40. Reinforcing research for TB

elimination at the country level:
experiences from path-finding countries
Developing a national TB research strategy: an
example from Kenya
M Muhwa Kenya Association for the Prevention of
Tuberculosis and Lung Diseases, Nairobi, Kenya.
Tuberculosis remains a major public health concern in

Kenya. Although Kenya has made tremendous progress
in its war against TB many challenges remain. These
include a persistent inability to find an estimated 15-20%
incident cases of TB; delays in TB diagnosis; low
treatment success rates in some parts of the country;
inadequate uptake of interventions to identify and
prevent TB in persons living with HIV and prevention,
care and control of drug resistant TB. The solutions for
many of these challenges demand the design and
implementation of program based operations and
implementation research to identify the most affordable,
sustainable, effective and efficient approaches. To
achieve this end Kenya formed a multi stakeholder
national TB research task force in 2008 which was
charged with several responsibilities including identifying the TB programs research priorities, guide the
development of research proposals, organize forums for
dissemination of research findings to all stakeholders and
the integration of research findings into routine program
activities, support the development of appropriate
training programmes for operations research, maintain
S41
an inventory of planned, ongoing and completed TB

research projects, coordinate research activities to ensure
conformity with identified national research priorities
and to avoid unnecessary and potentially costly duplicative research efforts and develop and implement a costed
strategy for multi year TB operations research. The task
force has organized several meetings in which Kenyas TB
research priorities were identified and results of completed research projects disseminated. Efforts have also
been made to develop a mentorship program to support
young TB researchers to start and complete the research
cycle. This presentation will highlight the achievements
that have been made so far, the persisting challenges and
constraints and the opportunities that could be used to
enhance TB research to support the full implementation
of the End TB Strategy in Kenya.
Undertaking the full spectrum of research in a

TB-endemic country: an example from South
Africa
V Mizrahi1 1Institute of Infectious Disease and Molecular
Medicine, University of Cape Town, Cape Town, South
Africa.
South Africa is a TB-endemic country with significant

scientific, technological and clinical research capabilities
and infrastructure. Working at the interface of the
laboratory, the clinic and the community, researchers in
South Africa are playing an important role the discovery,
development, evaluation and implementation new tools
for the diagnosis, prevention and treatment of TB
through research that spans the entire spectrum from
basic through to clinical, operational, implementation,
health services and health systems research. Within this
context, the peculiarities of the TB epidemic in South
Africa have driven the development of highly specialised
research capacity and expertise in the areas of HIVassociated TB, paediatric TB and MDR/XDR-TB. These
developments have, in turn, provided a platform for
South African researchers to engage in collaborative,
interdisciplinary TB research at the national, regional
and international level, supported by novel funding
instruments. They have also provided a mechanism for
using research as a vehicle for training the next
generation of scientists, clinicians and public health
specialists for the country. In this talk, I will describe the
key strengths of the TB research enterprise in South
Africa, and use specific examples to illustrate how
research is being used to address critical questions in
TB control.
S42
41. The Union/CDC Late-Breaker Session

on TB
The mortality impact of point-of-care urine
lipoarabinomannan testing to guide
tuberculosis treatment initiation in HIVinfected hospitalised patients: a multi-country
randomised controlled trial
J G Peter,14 L S Zijenah,5 D Chanda,1,6,7 P Clowes,8,9 M
Lesosky,1 C J Lombard,10 G Kadzirange,11 T Bandason,12
A Chansa,6,7 N Liusha,6,7 C Mangu,8 B Mtafya,8 H Msila,8
B Mtafya,12 A Rachow,8,9,13 M Hoelscher,8,9,13 P Mwaba,6
G Theron,1,14 K Dheda14 1Lung Infection and Immunity
Unit, Division of Pulmonology & UCT Lung Institute,
Department of Medicine, University of Cape Town, Cape
Town, 2Institute of Infectious Diseases and Molecular
Medicine, University of Cape Town, Cape Town, 3University
of Cape Town Lung Institute, Cape Town, 4Division of Clinical
Immunology and Allergology, Department of Medicine,
University of Cape Town, Cape Town, South Africa;
5
University of Zimbabwe College of Health Sciences,
Department of Immunology Harare, Zimbabwe; 6University
Teaching Hospital, Lusaka, 7Institute for Medical Research &
Training (IMReT), Lusaka, Zambia; 8National Institute for
Medical Research, Mbeya Medical Research Centre, Mbeya,
Tanzania; 9Division of Infectious Diseases and Tropical
Medicine, Medical Centre of the University of Munich,
Munich, Germany; 10Biostatistics Unit, South African Medical
Research Council & School of Public Health and Family
Medicine, University of Cape Town Cape Town, South Africa;
11
University of Zimbabwe College of Health Sciences,
Department of Medicine, Harare, 12Biomedical Research and
Training Institute, Harare, Zimbabwe; 13German Centre for
Infection Research, Munich, Germany; 14DST/NRF of
Excellence for Biomedical Tuberculosis Research, and MRC
Centre for Molecular and Cellular Biology, Division of
Molecular Biology and Human Genetics, Faculty of Medicine
and Health Sciences, Stellenbosch University, Tygerberg,
South Africa. Fax: (27) 214 047 651. e-mail:
keertan.dheda@uct.ac.za
Background: HIV-associated tuberculosis (TB) is difficult to diagnose and results in high mortality. Frequent
extra-pulmonary presentation, inability to obtain sputum, and paucibacillary samples limits the usefulness of
nucleic acid amplification tests and smear-microscopy.
We assessed the diagnostic impact of a urine-based
lateral flow point-of-care lipoarabinomannan assay
(LAM) on the mortality of TB patients.
Methods: 2659 hospitalised adult HIV-infected patients
with TB symptoms from South Africa, Tanzania, Zambia
and Zimbabwe (median CD4 count 86 cells/mm3) were
randomly allocated to either routine diagnostics (smearmicroscopy, Xpert MTB/RIF, and culture) or routine
diagnostics plus adjunctive urine LAM testing. The
primary endpoint was all-cause mortality at 8 weeks.
Findings: Urine LAM testing resulted in more patients
starting anti-TB treatment (52% versus 47%, P 0.024),
a decrease in the median (IQR) days to treatment
initiation [0(0-2) versus 1(0-3), P , 0.001], and less
empiric treatment (48% versus 70%, P , 0.001).
Overall, patients in the LAM study arm had reduced
mortality [adjusted HR (95%CI) 0.81 (0.69-0.95), P

0.009]. The effect size varied by clinical characteristics
with the greatest mortality reduction in those with CD4
650cells/mm3 [adjusted HR (95%CI) 0.68 (0.54-0.85)],
sputum scarce or GeneXpert MTB/RIF/ smear negativity
[adjusted HR (95%CI) 0.72 (0.55-0.95)], and at the
Zambian site [adjusted HR (95%CI) 0.67 (0.49-0.90)]
where patients were more likely to have poor prognostic
features [stage 4 disease, low weight, haemoglobin and
Karnofsky performance status, and high TBScore].
Overall 8 week TB-related morbidity (change in TBScore
and Karnofsky Performance Scale) was similar in both
arms.
Interpretation: Bedside LAM-guided rapid treatment
initiation in hospitalised HIV-infected patients with
suspected TB was associated with reduced 8-week
mortality. The extent of the mortality benefit is likely
to be greatest in those with poor prognostic features or
who are rapid sputum test negative/ sputum scarce.
Funding: European Developing Clinical Trials Partnership (EDCTP; TB NEAT), and the South African MRC
and NRF. ClinicalTrials.gov number: NCT01770730;
https://clinicaltrials.gov/ct2/show/NCT01770730
Yield and impact of intensified case

finding and isoniazid preventive therapy
for tuberculosis among people living
with HIV in Viet Nam
T Cowger,1 L H Thai,1 B D Duong,2 N V Nhung,3
D T Nhan,2 C K Thoa,2 V T Khanh,4 T Thinh,5 N H
Dung,6 N T B Yen,6 D V Ngoc,5 M McConnell,1 S
Whitehead,1 E S Pevzner1
1
US Centers for Disease Control and Prevention, Atlanta, GA,
USA; 2Viet Nam Authority for HIV/AIDS Control, Ministry of
Health, Hanoi, 3National Lung Hospital, Hanoi, 4VAAC-U.S.
CDC Co-agreement Project, Ministry of Health, Hanoi, 5Ho
Chi Minh City Provincial AIDS Committee, Ho Chi Minh,
6
Pham Ngoc Thach Hospital, Ho Chi Minh, Viet Nam
Introduction: In 2011, Viet Nam piloted an evidencebased algorithm to detect tuberculosis (TB) and provide
isoniazid preventive therapy (IPT) among people living
with HIV (PLHIV). We assessed the impact of antiretroviral therapy (ART) and IPT on TB incidence, and
whether regular screening for TB reduced mortality.
Methods: PLHIV aged .15 years at four HIV outpatient
clinics in Viet Nam were screened for any cough, fever, or
night sweats for .3 weeks and if positive, received
sputum smear examination, chest radiography, and/or
liquid culture per the diagnostic algorithm at each
clinical encounter within 1 year of follow-up. We
recorded timing of screening, ART and IPT uptake, TB
diagnoses and deaths. We conducted two multivariable
Cox regression analyses with ART and IPT as timevarying covariates and incident TB and death as
endpoints.
Results: Of 789 participants enrolled, 64 (8.1%) were
diagnosed with TB at enrollment and 36 (4.6%) were
excluded for missing data. Among the remaining 689

participants, TB incidence was 4966/100 000 personyears. TB incidence was 903 vs. 10 331 per 100 000
person-years, respectively, among participants prescribed
and not prescribed IPT, adjusted hazard ratio (aHR) 0.13
(0.030.50); incidence was 4128/100 000 person-years
among participants not yet on ART, aHR 12.1 (2.559.0)
and 19 550/100 000 person-years among participants on
ART ,3 months duration, aHR 16.9 (4.268.3)
compared to 1256/100 000 person-years among those
on ART 73 months. Survival was greater among
participants screened regularly for TB (i.e., 71 screening
every 4 months) (Figure). The impact of regular screening
was greatest after 120 days of follow-up: mortality was
634/100 000 person-years among participants with
regular screening versus 9439/100 000 person-years
among those without regular screening, aHR: 0.07
(0.010.45).
Conclusions: Regular TB screening provided health care
workers multiple opportunities to diagnose and treat TB,
and likely contributed to reduced mortality among
participants in this high TB prevalence population.
Figure Survival for patients with regular screening for TB
compared with survival for those without regular
screening for TB
S43
A novel socioeconomic intervention

improves TB cure and chemoprophylaxis
completion in Peruvian shantytowns: a
randomized controlled evaluation
T Wingfield,13 M A Tovar,2,4 D Huff,2,5 D
Boccia,2,6 R Montoya,2 E Ramos,4 S Datta,1,3 M J
Saunders,1,2 J J Lewis,2,6 R H Gilman,7 C A
Evans1,2,4
1
IFHAD Innovation For Health And Development, Infectious

Diseases & Immunity and Wellcome Trust Centre for Global
Health Research, Imperial College London, UK; 2Innovacion

Benefica
Por la Salud Y Desarrollo (IPSYD), Asociacion

PRISMA, Lima, Peru; 3Monsall Infectious Diseases Unit, North
Manchester General Hospital, Manchester, UK; 4IFHAD
Innovation For Health And Development, Universidad
Peruana Cayetano Heredia, Lima, Peru; 5Tulane University
School of Public Health and Tropical Medicine, New Orleans,
LA, USA; 6Infectious Disease Epidemiology, London School of
Hygiene & Tropical Medicine, London, UK; 7Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, USA
Background: Socioeconomic support is part of the

recently ratified post-2015 global End TB Strategy.
However, evidence evaluating impact of TB-specific
socioeconomic interventions is extremely limited.
Objective: To evaluate the impact of a socioeconomic
intervention on TB care and prevention.
Methods: Design: A household-randomized controlled
study. Setting: 32 shantytown-communities, Callao,
Peru. Participants: All TB patients treated by the
Peruvian TB Program. Randomisation: Patient households were randomly assigned 1:1 to controls that
received Peruvian TB program standard of care or
intervention households that additionally received the
socioeconomic intervention. Intervention consisted of
socioeconomic support throughout TB treatment. Economic support constituted conditional cash transfers up
to US$230 (4.7% of average TB-affected household
annual-income) to mitigate TB-related costs, incentivise
and enable care. Social support constituted householdvisits and TB club community-meetings aiming to inform
and empower, reduce stigma, and facilitate mutualsupport. Outcomes compared intervention with control
households. The primary outcome was TB chemoprophylaxis completion in contacts aged ,20 years. A priori
analyses of contacts aged ,5 and 5-19 years were also
performed. Secondary outcome was confirmed TB
patient cure.
Results: Seven months of recruitment achieved sample
size to test for a 33% effect on the primary outcome.
90% (282/312) patients were recruited who had 518
eligible contacts. Patients were randomized to the
intervention (n 135) and control (n 147) arms.
Outcome (Figure): Intervention contacts were more
likely to start (43% [95% confidence intervals, CI3649] versus 25% [95%CI19-31], adjusted odds ratio,
aOR 2.0 [95%CI 1.3-3.8], P 0.0001) and more
likely to complete (19%[95%CI 14-25] versus
11%[95%CI 7.0-15], aOR 1.9 [95%CI1.1-3.3, P
S44
0.002) TB chemoprophylaxis than control contacts.

This impact was also significant in sub-analyses of ages
,5 and 5-19 years. Confirmed cure was more likely in
intervention than control patients (47% [95%CI39-56]
versus (34% [95%CI26-42], aOR-1.8 [95%CI 1.12.9], P 0.02).
Conclusions: The socioeconomic intervention improved
TB cure rates and TB chemoprophylaxis uptake and
completion in an impoverished setting. This evidence
supports implementation of recent global policy changes.
Figure Uptake, adherence to, and completion of six
months of TB chemoprophylaxis in household contacts
(n 518) and confirmed TB cure in TB patients (n 282).
Error bars show 95% confidence intervals. P values on
the lower lines are those derived from univariate logistic
regression against the binary outcome variable of
chemoprophylaxis completion or TB cure. P values on
the upper lines are those derived from multiple logistic
regression against the binary outcome variable of
chemoprophylaxis completion or TB cure adjusting for
food insecurity, age, gender, education level, employment
status, crowding level, and monthly household income.
Methodology: Adherent monocytes from healthy individuals were infected with BCG or virulent mycobacteria: H37Rv (laboratory strain) or CDC1551 (Clinical
strain). Monocytes were co-cultured with E-cig liquid, Ecig vapour, cigarette smoke extract (CSE) or nicotine. Ecig liquid and vapour were obtained from nicotine
containing Twispw brand E-cigarettes. E-cigarette vapour was collected through RPMI during a 3 & 5 minute
vape in a similar published method used to generate
tobacco smoke extract. Cell viability (trypan blue
staining) and TNF responses (ELISA) were measured at
18 h.
Results: Toxicity experiments demonstrated a dose
dependent toxicity of CSE, nicotine and E-Liquid. Using
low concentrations of exposures, we demonstrated a
consistent reduction in TNF-alpha production for all
exposures: E-cig vapour (using a 50% concentration of 3
and 5 minute vape sessions) reduced TNF-a production
by a mean 38% (SD 26) P , 0.001 and 78% (SD 20) P ,
0.001, respectively. Nicotine 100 lg/ml alone reduced
TNF-a production by a mean 43% (SD 24) P , 0.001
and 1% E-cig liquid by 78% (SD 23) P , 0.001. 10%
CSE reduced TNF-a by 67% (SD 23) P , 0.001 in
response to virulent Mycobacterium tuberculosis infection.
Conclusions: Although electronic cigarettes are reported
to be less harmful than tobacco, these data demonstrate
impairment to a key mycobacterial immune response.
Caution should be advocated especially in TB endemic
regions about using electronic cigarettes, until the full
effect on mycobacterial immunity is clarified.
Early bactericidal activity of AZD5847 in

pulmonary tuberculosis
J J Furin,1 J Du Bois,2 E van Brakel,2 P Chheng,1 A
Venter,3 C A Peloquin,4 B Thiel,1 S M Debanne,5
W H Boom,1, A H Diacon,2,6 J L Johnson1
1
Division of Pulmonology & UCT Lung Institute, Department

of Medicine, University of Cape Town, Cape Town, 2Institute
of Infectious Diseases and Molecular Medicine, University of
Cape Town, Cape Town, South Africa
Tuberculosis Research Unit, Department of Medicine, Case

Western Reserve University School of Medicine and
University Hospitals Case Medical Center, Cleveland, OH,
USA; 2TASK Applied Science, Bellville, 3Centre for Molecular
and Cellular Biology, Faculty of Medicine and Health
Sciences, Stellenbosch University, Tygerberg, South Africa;
4
Infectious Disease Pharmacokinetics Laboratory, University
of Florida College of Pharmacy and Emerging Pathogens
Institute, Gainesville, FL, 5Department of Epidemiology and
Biostatistics, Case Western Reserve University School of
Medicine, Cleveland, OH, USA; 6Division of Physiology,
Department of Biomedical Sciences, Faculty of Medicine and
Health Sciences, Stellenbosch University, Tygerberg, South
Africa
Background: Electronic cigarettes (E-cig) are gaining

widespread popularity as a safer alternative to tobacco.
There is a paucity of data evaluating their short-term
impact on respiratory infections. Tobacco cigarettes
impair human immune responses to mycobacterial
infection and double the risk of developing tuberculosis.
Background: New therapies are needed to treat patients

with drug-resistant TB. We conducted a randomized,
open label, phase 2a dose ranging trial to evaluate the 14day bactericidal activity, safety and pharmacokinetics of
AZD5847 (AstraZeneca, Wilmington, DE, USA), a new
oxazolidinone antibiotic, in patients with pulmonary TB.
Electronic-cigarettes, nicotine and

tobacco smoke impair human immune
responses to virulent Mycobacterium
tuberculosis infection
R Chang,1 M Davids,1 K Dheda,1,2 E Bateman,1 R
van Zyl-Smit1
1
Methods: Seventy-five adults with newly-diagnosed,

sputum smear-positive TB were randomly assigned to
receive AZD5847 at different doses (500 mg once daily;
500 mg twice daily; 1200 mg once daily; and 800 mg
twice daily) or standard treatment (Rifafour; HRZE) The
primary efficacy endpoint was the mean daily fall in
MTB log10 CFU per ml of sputum, analyzed with linear
mixed modeling techniques for longitudinal data.
Change in time to positivity (TTP) in MGIT culture
over 14 days was a secondary endpoint. SAS 9.3 was
used for statistical calculations. The trial was registered
at www.clinicaltrials.gov (NCT01516203).
Results: The mean 14-day bactericidal activity of HRZE
was consistent with previous studies (0.163 log10 CFU/
day). AZD5847 at a dose of 500 mg twice daily showed a
decline in CFU over time that excluded zero (average
change of 0.039 log 10 CFU/day, P , 0.0048).
AZD5847 at doses of both 500 mg and 800 mg twice
daily also showed an increase in TTP in MGIT. The
90%CI of the effects on TTP of both doses excluded
zero, with an increase of 1.24 hours per day and 1.21
hours per day respectively (P 0.013 and 0.018,
respectively). Over the exposure range tested,
AZD5847 appeared to be generally well tolerated. The
most frequent adverse events seen were hematologic,
hepatic and gastrointestinal. The drug exhibited loglinear elimination with a half-life of 5 to 7 hours.
Conclusions: AZD5847 at the 500 mg and 800 mg twice
daily doses showed an EBA that was different from zero.
The drug merits further evaluation in patients with
pulmonary TB.
Support: US NIAID/NIH HHSN266200700022C/NO1AI-70022. Clinical trial registration number: NCT
01516203.
Unexpected high frequency of Rv0678

mutations in MDR-TB patients without
documented prior use of clofazimine or
bedaquiline
N Coeck,1 C Villellas,2 C J Meehan,1 N Lounis,2 S
Niemann,3 L Rigouts,1 B de Jong,1 K Andries2
1
Institute of Tropical Medicine, Antwerp, 2Janssen Research

and Development, Beerse, Belgium; 3National Reference
Centre for Mycobacteria, Borstel, Germany. e-mail:
kandries@its.jnj.com
Background: Mutations in Rv0678, a gene regulating the

expression of the MmpS5-MmpL5 efflux pump, have
been shown to lead to 2- to 8- fold increases in
bedaquiline (BDQ) MIC, and 2- to 4-fold increases in
clofazimine (CFZ) MIC.1 After the unexpected discovery
of an Rv0678 mutation in a baseline isolate from a
patient without documented prior use of CFZ or BDQ,
the prevalence of these mutations in clinical isolates was
assessed.
Methods: Baseline isolates from two bedaquiline MDRTB clinical trials were assessed for Rv0678 mutations by
Sanger sequencing; corresponding BDQ MICs were
S45
determined on 7H11 agar. Rv0678 mutations were also

investigated in DS TB sequences of the Hamburg cohort,2
using a modified version of the KvarQ software.3
Results: Rv0678 mutations were identified in 21/345
(6.1%) of MDR-TB baseline isolates from the two
clinical trials. BDQ MICs of these isolates were either
normal (0.03 to 0.24 lg/ml, n 10), high (0.48 to 1, n
8), or low (0.004 to 0.015, n 4). Mutations were also
often identified in the intergenic region between Rv0678
and Rv0677c. A mutation at position 11 in the
intergenic region appeared to increase the sensitivity for
BDQ (median MIC 0.015). In DS TB isolates, the
frequency of Rv0678 mutations was lower (4/817 or
0.5%), but still higher than expected based on the
theoretical mutation rate.
Conclusions: Rv0678 mutations occur more frequently
in MDR-TB patients than previously anticipated, and are
not consistently associated with increased MICs for
BDQ. Given the variety of mutations in Rv0678 and
their variable effects on the MIC, phenotypic DST
methods are preferred to molecular methods to assess
BDQ susceptibility. The origin of Rv0678 mutations is
unknown, but cannot be linked to prior use of BDQ or
CFZ.
References
1 Andries K, Villellas C, Coeck N, et al. Acquired resistance of
Mycobacterium tuberculosis to bedaquiline. PLOS ONE 2014; 9:
e102135.
2 Kohl T A, Diel R, Harmsen D, et al. Whole-genome-based
Mycobacterium tuberculosis surveillance: a standardized, portable,
and expandable approach. J Clin Microbiol 2014; 52: 24792486.
3 Steiner A, Stucki D, Coscolla M, Borrell S, Gagneux S. KvarQ:
targeted and direct variant calling from fastq reads of bacterial genomes.
BMC Genomics 2014; 15: 881.
TB type
MDR-TB
DS TB
Rv0678
mutation
intergenic
region position
8 to 9
Intergenic
region
position 11
n (%)
n (%)
n (%)
345
817
21 (6.1%)
4 (0.5%)
2 (0.6%)
0
37 (10.7%)
0
TB tuberculosis; MDR-TB multi-drug resistant tuberculosis; DS TB

drug susceptible tuberculosis.
S46
First results with a 9-month regimen for

multidrug-resistant tuberculosis (MDRTB) in francophone Africa
42. TB research and communities: ethical

and practical considerations
Community engagement in descriptive studies
C Kuaban,1 Z Kashongwe,2 A Bakayoko,3 T

Ndikumana,4 S Hassane,5 S Bassirou,5 V
Fikouma,6 M Gninafon,7 M Ouedraogo,8 V
9
Schwoebel,9 K G Koura,9 H Rieder,9 A Trebucq

1
Bamenda University, Bamenda, Cameroon; 2Bukavu

University, Bukavu, Democratic Republic of Congo; 3CHU
Treichville, Abidjan, Ivory Coast; 4National Tuberculosis and
Leprosy Program, Bujumbura, Burundi; 5Action Damien,
Niamey, Niger; 6Bangui University, Bangui, Central African
Republic; 7National Tuberculosis Program, Cotonou, Benin;
8
CHU Yo, Ouagadougou, Burkina-Faso; 9International Union
Against Tuberculosis and Lung Disease, Paris, France. e-mail:
vschwoebel@theunion.org
Background: A 9-month regimen (4KmGfxPtoHCfzEZ/

5GfxCfzEZ) for multidrug-resistant tuberculosis (MDRTB) has been more than 80% effective in Bangladesh. We
report results for the first cohort of patients included in
an observational study coordinated by the International
Union Against Tuberculosis and Lung Disease aiming to
evaluate the effectiveness and tolerance of this regimen
(moxifloxacin replacing gatifloxacin) in 9 countries of
francophone Africa.
Methods: From January 2013 to March 2015, 1022
adult patients were recruited in Benin, Burkina Faso,
Burundi, Cameroon, Ivory Coast, Niger, Central African
Republic, Democratic Republic of Congo and Rwanda.
The patients receive treatment under strict daily observation after standard clinical and laboratory examinations. Adverse drug events and clinical and bacteriological response to treatment are followed up monthly until
treatment completion, and bacteriological status is
monitored 6-monthly thereafter for 2 years. We describe
the characteristics and analyzed the clinical and bacteriological response to treatment and the rate of adverse
events of the first cohort of patients included before 1
July 2014.
Results: Among this cohort of 507 patients37%
female, median age 33 years, 88% previously treated
for TB, 21% HIV-positive80.9% had treatment
success, 7.7% died, 6.5% were lost to follow-up and
4.9% were failures. The case fatality rate was higher in
HIV-positive than in HIV-negative patients (17.1% vs.
5.0%, P , 0.001). The most frequent adverse event was
mild gastric pain and vomiting during the 2 first months
of the treatment. Hearing loss .70dB was reported in
9.2% of the cohort. Severe (grade 4) adverse events
occurred in 2.7%, and were significantly more frequent
among HIV-positive than among HIV-negative patients
(6.9% vs. 1.6%, P 0.003).
Conclusions: Preliminary treatment results on the first
enrolled patients are excellent. Implementing this short
MDR-TB treatment regimen has proven feasible in the
participating countries, with limited rates of adverse
events except for hearing loss.
A Choko,1 R Sambakunsi,1 M Nliwasa,1,2

M Kumwenda,1,2 G Sinjani,1 L Chiume,1 L Corbett,1,3
N Desmond1,4 1Malawi Liverpool Wellcome Trust Clinical
Research Programme, Blantyre, 2University of Malawi,
College of Medicine, Blantyre, Malawi; 3London School of
Hygiene & Tropical Medicine, London, 4Liverpool School of
Tropical Medicine, Liverpool, UK.
Background: Ethical research requires reflective engagement between researchers and communities to maximise
mutual benefits. Community engagement can be utilised
to inform, consult or to enable local decision-making.
Here we describe a community liaison system (CLS)
established to facilitate consultation and dialogue during
an HIV self-testing (HIVST) and TB screening trial in
urban Malawi.
Methods: The trial setting was 28 small neighbourhoods,
defined by community-health worker (CHW) catchment
populations. Four cluster representatives (CRs) per
neighbourhood (128 in total) were recruited using
participatory methods (nominated and voted at residential meetings). Their primary mandate was to meet
weekly as a neighbourhood team, to report all discussions (both positive and negative) relating to the trial,
submitting their reports at monthly facilitated CR
meetings where concerns were discussed and appropriate
responses developed. CRs performed a secondary role
reporting on all deaths for verbal autopsy (dual
representation). Separately, a standard community
advisory board (CAB) included 15 community members/leaders.
Results: Of 883 issues raised over the first 6 months from
416 residents, concerns most often related to lack of
HIVST (41%) in control neighbourhoods. Additional
common concerns included misunderstandings of HIV
care pathways for people who self-tested positive.
Women (49%) raised more concerns than men (27%)
and mixed groups (24%). Benefits to the trial included
CLS motivation to HIVST clients to fully participate
with TB screening (sputum submission) and rapid tracing
of participants with newly diagnosed TB. By sharing
their collective lack of major adverse events (suicides,
and gender-based violence) in real time, the CLS also
supported rapid gains in community confidence around
the safety of HIVST: a previously untried intervention.
Furthermore, CRs assisted in QA activities and maintaining HIVST availability within defined neighbourhood boundaries despite very high demand (~84%
population uptake).
Conclusions: A linked community liaison system was
highly successful in facilitating dialogue between the
research team and community, and dealing promptly
with harmful rumours. Dual representation was highly
beneficial to the trial team, although with potential to
compromise the true community voice. Initial feedback
on the community demand for HIVST was validated by

subsequent update data.
Community engagement in TB vaccine studies

M Tameris1 1South African Tuberculosis Vaccine Initiative,
University of Cape Town, Cape Town, South Africa.
Background: The South African Tuberculosis Vaccine

Initiative (SATVI) has been conducting clinical trials of
novel tuberculosis (TB) vaccines in Worcester, Western
Cape since 2005. Novel TB vaccines are targeted at
infants as BCG replacements or at adolescents and adults
as BCG boosters. Clinical researchers have a responsibility to the communities in which they work to engage
with them in research activities and to build their
capacity to understand and give input into the research.
SATVI works closely with our community stakeholders
in a variety of ways.
Methods: SATVI has established an active community
advisory board (CAB) which has an adolescent CAB
component. A comic book describing a young mothers
decision to enrol her infant into a TB vaccine trial has
been produced, distributed and dramatised. Two Wellcome Trust International Engagement grants have
funded projects using drama to introduce the concepts
of clinical research including participant rights, and TB
vaccine research, to high school pupils and to the general
population respectively.
Results: SATVI continues to have an excellent relationship with the community in which we work. stakeholders
include the local Departments of Health and Education,
local business forum members, and most importantly the
man in the street who volunteers to participate in our
trials.
Conclusion: Active and ongoing community engagement
is essential when conducting clinical research.
43. Country experiences: enablers for

community-based multi-drug resistant
tuberculosis management programmes
The effects of Directly Observed Therapy (DOT)
in MDR-TB treatment, a systematic review and
field experience from China
X Wei,1 J Yin1 1JC School of Public Health and Primary Care,
Chinese University of Hong Kong, Hong Kong, China.
Background: Directly observed therapy (DOT) was

recommended by WHO to improve adherence of
Multidrug-resistant tuberculosis (MDR-TB). However,
the effectiveness of DOT on medication adherence and
treatment outcomes of MDR-TB was mixed in previous
studies. Thus, with the aims of assessing medication
adherence and treatment outcome among MDR-TB
patients and their associations with provision of DOT,
we systematically reviewed previous studies, and conducted a retrospective cohort study in two provinces of
China.
S47
Methods: We searched studies published in English

between January 1970 and January 2015 in the major
electronic databases. Two reviewers independently
screened articles and extracted relevant information for
the systematic review. The cohort study was conducted in
Zhejiang (ZJ) and Shandong (SD) provinces. Culture
confirmed MDR-TB patients who had begun their
treatment between January 1 2009 and June 30 2012
in ZJ and between January 1 2009 and April 30 2012 in
SD were included into the study. We reviewed clinical
charts and conducted a questionnaire survey with the
patients. Treatment interruption was defined as missing a
dose for at least a day but less than eight consecutive
weeks, while severe interruption was defined as missing
doses between two to eight consecutive weeks.
Results: A total of 29 articles were included in the review.
Studies adopting full DOT (68%, 95%CI 63-73%) had a
significantly higher pooled treatment success rate than
studies using self-administration treatment (47%,
95%CI 41-52%). Among the 330 patients (220 in ZJ
and 110 in SD), 175 (53%) had a treatment interruption,
while 62 (19%) reported severe interruption; 251 (76%)
reported a culture conversion after six months treatment;
154 (47%) patients received DOT with 111 (72%)
received from family members. Patients who were order
(OR2.14, 95%CI 1.12-4.10), did not receive DOT
(OR0.47, 95%CI 0.25-0.92), injected by health care
workers (OR0.28, 95%CI 0.08-0.98) had more severe
interruptions than those who were younger, received
DOT and injected by family members. Patients who had
higher level of education (OR1.95, 95%CI 1.09-3.50)
and had no interruption (OR0.54, 95%CI 0.31-0.95)
were more likely to have six-month culture conversion.
Conclusions: Treatment interruptions are common in
study sites. DOT was effective in improving adherence
and successful treatment outcome among MDR-TB
patients. Family members may act as treatment supporters after necessary training.
The role of health insurance schemes in

managing community-based MDR-TB patients
via designated hospitals in China
Q Sun Center for Health Management and Policy, Shandong
University, Jinan, China.
Background: Poor MDR-TB patients detection, low

standard treatment rate of MDR-TB patients, lack of
effective MDR-TB patients management and higher
economic burden of MDR-TB treatment are the main
challenges in the MDR-TB control of China. Under the
support of the Bill and Melinda Gates Foundation, China
imitated a comprehensive MDR-TB control project in
four cities since the year 2011. One of the main
interventions is increasing the reimbursement of health
insurance to MDR-TB treatment for alleviating the
economic burden of MDR-TB patients.
Methods: A field survey was conducted in the four pilot
cities. The health insurance policies were collected;
MDR-TB patients were survey using structured questionnaires to evaluate their economic disease burden
S48
under the existed health insurance programs. Descriptive

statistical methods were used to analyze the effects of
health insurance program on decreasing the economic
diseases burden of MDR-TB patients.
Results: The pilot projects regulated the health insurance
schemes will cover at least 50% of MDR-TB treatment
expenditure with the standard treatment protocol, the
patients will cover 10% of medical expense, the left are
covered by the projects. In the four pilot cities, we use the
MDR-TB inpatient expenditure of the New Rural
Cooperative Medical Scheme (NCMS) as example, it
was found the proportion of MDR-TB inpatient reimbursement to total inpatients expenditure of NCMS in
the four cities was very tiny varying from 0.02% to
0.05%. A total of 73 MDR-TB patients were investigated. Before the project reimbursement to MDR-TB
patients, it was estimated that the total treatment cost
of MDR-TB patients in the two years treatment period
was 23 430 Yuan (18 966~30 894 Yuan) the total
medical expenditure was 15 166 Yuan (12 811~17 898).
Among the surveyed MDR-TB patients, 84% of them got
the catastrophic medical expenditure in treating MDRTB. After the reimbursement by the project, the total
treatment costs decreased 67%, but 65.75% of MDR-TB
patients still suffered the catastrophic medical expenditure.
Conclusions: The pilot projects decreased the economic
burden of MDR-TB patients, but the effects and the
sustainability of MDR-TB control is uncertainly without
the subsidy of the project. So, a case-based payment by
health insurance programs should be designed and
implemented.
Expanding MDR-TB management through

community health workers in Bangladesh and
Nepal: NGO experiences
S Islam,1 M Rana,1 M Siddiqui,1 M Akramul Islam,1
M Husain2 1Department of Health, BRAC, Dhaka,
2
Department of Health, NTP, Dhaka, Bangladesh.
Background and challenges: Bangladesh ranks in 10th on

the list of 27 high priority countries for MDR-TB.
According to a recent drug resistance survey, MDR-TB
burden is 1.4% among new cases and 28.5% among
retreatment cases. However, considering the size of the
population the absolute number of MDR-TB patients
would be higher. National Tuberculosis Control Programme initiated Programmatic Management of DR-TB
cases since 2008.
Intervention: Patients diagnosed in the National Institute
of Diseases of the Chest and Hospital admitted for 6 to 8
months of intensive phase of treatment initially and 18
months ambulatory phase at community. Currently,
sputum microscopy is done weekly for admitted patients.
If two consecutive sputa become negative then the MDRTB patients are shifted to the community for ambulatory
care. Sputum follow up test like microscopy is done
monthly at the field level and culture is done quarterly at
TB Reference Laboratory throughout the treatment
course. All the health care providers related to MDR-
TB management receive basic one day orientation before

starting of MDR-TB management at community level.
The community health workers play important role in
educating MDR-TB patients and their family members
for continuation the treatment and provide psychosocial
support. They ensures daily intake of medicine during
ambulatory phase of treatment and they also encourage
patient to continue regular follow-up as per NTP
guideline.
Results and lesson learnt: In 2008, a total of 70
confirmed patients were enrolled in BRAC covered TB
areas and received standardized MDR-TB treatment.
Out of 70 MDR-TB patients, 48 patients completed
treatment successfully; the treatment success rate was
69%. The number of patients who were death and
default was 6 (9%) and 16 (23%) respectively. In 2012, a
total of 333 confirmed patients were enrolled and 238
patients completed treatment successfully, the treatment
success rate was 71%. The number of patients who died
or defaulted was 36 (11%) and 5 2 (16%), respectively.
Conclusion: Patients adherence during hospital stay
seems to be a difficulty in hospitalization treatment
system as they had to stay for long time during intensive
phase. The currently implemented community based
MDR-TB treatment modality may contribute to minimize this challenge and ensure high treatment success
rate for National Tuberculosis Control Programme in
Bangladesh.
Expanding MDR-TB management through

community health workers in Bangladesh and
Nepal: NGO experiences
S C Baral,1 S Khanal,1 B Lamichhane,2 K D Wagle2
Department of Research, Health Research and Social
Development Forum (HERD), Kathmandu, 2Department of
Programme, National Tuberculosis Centre, Bhaktapur, Nepal.
1
Background: TB is one of the major health problems in

Nepal. More recently, the emergence of drug resistance
(DR-TB) poses even a greater challenge to the NTP. In
2014, total of 349 DR-TB and 25 XDR-TB cases were
enrolled for treatment. Treatment Success Rate of DR
patients was 76.6%, however it was 11% for XDR.
Methods: Qualitative interviews held with DR-TB
patients and their family members and group discussions
with frontline health workers and stakeholders. We also
reviewed NTP annual data.
Findings: Nepal initiated DR-TB management in 2005
with 5 treatment centres and 11 sub treatment centres.
Since then there has been remarkable progress in the field
of DR-TB control. The key achievements include
expansion of DR-TB centres from 5 to 13 and DR-TB
sub centres from 11 to 71 from 2005 to 2013,
respectively, covering 42 of 75 districts, with increased
involvement of community health workers. Ambulatory
DR-TB treatment is now managed at the sub centre once
patients have started their treatment at the centres.
Similarly, 11 DR-TB hostels provide facility based
treatment. Nepals move to decentralised DR-TB treatment has led to improved DR-TB management. The
trend of DR case finding has been on rise from 94

(18.6%) in 2009 to 349 (22.3%) in 2013. The number of
DR-TB patients registered for treatment was 87 in 2005
to 201 in 2010. Similarly, the treatment success rate has
increased to 76.6% in 2010 as compared to 64.7% in
2007 with significant decrease in loss to follow up from
18.7% in 2007 to 7.3% in 2010. Major challenges
include: poor adherence to the treatment, lack of
community awareness, limited skilled human resources,
limited provision of infection control measures in health
facilities, relocation of patients for treatment resulting
socioeconomic burden to patients.
Conclusion: Though the decentralised treatment services
have been a major a step forward in DR-TB management
in Nepal, challenges remain: access to early diagnosis and
treatment, quality care at the point of delivery, socioeconomic and psychological barriers to DR-TB patients
and their family, skilled human resources at referral and
periphery level facilities, continuation of DR-TB hostels,
among others. Therefore, in order to achieve better
outcomes with less burden on DR-TB patients and
families, there is a need for further expansion and
strengthening of patient centered decentralised DR-TB
care tailored to the local context.
Developing and implementing an Internetbased information system, SITETB, to integrate

services among health providers in Brazil for
MDR-TB care
P Bartholomay Oliveira,1 D Barreira,1
S Barbosa Codenotti1 1National Tuberculosis Programme,
BRASILIA, Distrito Federal, Brazil.
According to the World Health Organization (WHO), in

2014 480 000 new cases of multidrug-resistant tuberculosis (MDR-TB) were estimated worldwide and approximately 210 000 deaths from the disease. When there is
diagnosis of MDR-TB alternative treatment regimens are
needed to control it, which usually have a lower cure
rate, less favorable prognosis, more side effects and
higher cost. The Brazil monitors cases of MDR-TB by
Information System Tuberculosis Special Treatments
(SITE-TB), is an online system available in all the federal
units, and provides information for epidemiological
surveillance, assists decision making, and the management of drugs. The SITE-TB allows the secondary and
tertiary reference units notify, track and see the outcome
the cases that require special treatment, including cases
of mycobacteria non tuberculosis (MNT). The system
was developed in partnership with Management Sciences
for Health (MSH) aimed at to provide for the monitoring
of cases with special schemes, clinical monitoring and
rational use of medicines, essential strategies for the
proper management and control of drug resistant
tuberculosis (DR-TB). In 2012 a pilot study was
conducted in three Federal Units, with different profiles,
and after some adjustments, all the states were trained to
use the system. Currently there are about 600 health
professionals using the system for surveillance, clinical
care, and the management of drugs. In addition, a group
S49
of ten TB experts in DR-TB validate all cases to monitor

the proper use of the national protocol, to do technical
support and to provide better management of medicines.
Currently, the SITE-TB is used by all the 27 states, for
about 250 units that report and accompanying special
cases. In 2014, 1738 cases of special tuberculosis
treatments were followed in SITE-TB with indication
for use of special schemes due to liver disease, severe
intolerance, drug allergy or comorbidities, as well as
2416 cases of DR-TB and 548 cases of MNT.
44. Achieving quality Xpert MTB/RIF use

for rapid case detection
The role of Xpert MTB/RIF testing in improving
the clinical diagnosis of patients in Pakistan
E Qadeer,1 S Tahseen,2 F M Khanzada2 1National
Tuberculosis Control Program, Islamabad, 2Department of
National Tuberculosis Reference Laboratory, National
Tuberculosis Control Program, Islamabad, Pakistan.
Background: Pakistan is fourth among high burden

countries for TB and DR-TB. 82% of cases notified are
pulmonary TB and 50% of these are bacteriologically
diagnosed; 9.9% of all incident cases notified are
children. 75% of estimated MDR-TB burden is among
new and only 25% among retreatment PTB cases. The
capacity for conventional bacteriological diagnosis of TB
cases among smear-negative PTB, childhood TB and
EPTB is limited.
Intervention: GeneXpert was introduced in Pakistan in
2011 soon after its endorsement by WHO. It was
implemented initially with focus to enhance DRTB case
diagnosis but its use was soon expanded to improve
bacteriological diagnosis of TB in vulnerable, at risk
population (children and HIV ve) and those having
difficulty in diagnosis. The number of Xpert machines
increased from eleven in 2011 to 67 in 2014 and patients
tested increased from 482 In year 2011 to 56 252 in year
2014 and by mid-2015 a total of 101 242 patients had
been tested across the country. We present the role of
Xpert in clinical diagnosis of TB in 20 252 patients (
19.8%) tested at the National TB Reference laboratory
(NRL) during this period .
Results: Among patients tested (July 2011-2015), 50.8%
had no history of previous treatment, 12.9% were under
15 yrs of age and 6.2% of patient had extra -pulmonary
lesions. Number of adults and children tested with PTB
or EPTB has increased from 116 (2011) to 7081 (2014)
but proportion of paediatric cases has declined from
21.6% to 12.9% over time. In patients with no history of
previous treatment, Mycobacterium tuberculosis was
detected in 19.7% of patient s (22.6% in adults and
9.1% in ,15yrs) and Rifampicin resistant (RR) was
detected in 11.4% (11.6% in adults and 9.6% in ,
15Yrs). In patients with history of previous treatment M.
tuberculosis was detected in 37.4% (37.9% in adults and
27.7% in ,15yrs) and RR was detected in 22.7%
(22.5%in adults and 28.9% in ,15yrs). Among extra-
S50
pulmonary specimen, M. tuberculosis was detected in

20.0% and Rifampicin resistant in 12.8% of cases.
Conclusion: Introduction of Xpert along with enhancing
diagnosis of DR-TB cases has played an important role in
bacteriological diagnosis of tuberculosis in children,
adults with difficulty in diagnosis as well as extrapulmonary cases. There is need for sensitization of
pediatricians and surgeons on role of diagnostic assay in
bacteriological diagnosis of these difficult cases.
Yield of TB lymphadenitis with cytology vs.

GeneXpert and culture in Ethiopia
T Gemechu,1 M Aseresa Melese,2 N Demmelash,2
D Habte,2 B Tessema,3 P G Suarez4 1Department of
Pathology, College of Health Sciences, Addis Ababa
University, Addis Ababa, 2Center for Health Services, HEAL TB
Project, Management Sciences for Health, Addis Ababa,
3
Department of Medical Microbiology, School of Biomedical
and Laboratory Sciences, College of Medicine and Health
Sciences, University of Gondar, Gondar, Ethiopia; 4Center for
Health Services, Management Sciences for Health, Arlington,
VA, USA.
Background: GeneXpert MTB/RIF assay has been

deployed close to the point of patient care in the fight
against TB. The Ethiopian MOH has endorsed the test
and there has been a wider installation of this technology
into the existing TB diagnostic system. On the other
hand, cytology microscopic reading, which has been the
first line diagnostic tool for TB lymphadenitis is only
conducted in centers staffed by Pathologists. The
objective of this study was to compare the level of
agreement between cytology readings and GeneXpert
results for diagnosis of TB lymphadinitis.
Methods: A comparison was made between the cytology
readings of two senior pathologists versus Xpert results
of fine needle lymph node aspirates. A senior pathologist
took a split sample of FNA from patients clinically
suspected of TB lymphadenitis. The cytology smears
were stained with Wrights stain and read by two senior
pathologists. The cytology reading is considered TB if the
two pathologists reported concordant TB diagnosis. A
senior microbiologist runs the Xpert test following the
standard procedure. Both the pathologist and the
microbiologist were blinded for the results of each other.

We calculated the concordance rate of the two test results
using Kappa statistics.
Results: A total of 150 TB lymphadenitis suspected
patients were included in this study. From the total of 150
TB lymphadenitis suspects, 114 (76.0%) and 86 (57.3%)
were diagnosed as TB by cytology reading and Xpert,
respectively. Four (4.7%) of the TB patients diagnosed by
Xpert were found to have Rifampicin Resistant Tuberculosis. Out of 114 patients diagnosed as TB lymphadenitis by cytology, 84 (73.7%) were also diagnosed as TB
by Xpert. The cytology labeled 36 of the suspects as no
TB of which 34 (94.4%) were also negative for TB by
Xpert. The overall agreement level between the two tests
was moderate (Kappa 0.54, P , 0.001). The agreement
levels of Cytology and GeneXpert results in the caseous
and pus aspirates were 0.58 (P , 0.001) and 0.49 (P ,
0.001) respectively. Caseous aspirates yielded less TB
positive result than pus aspirates: 69.2% versus 84.6%
by cytology and 47.7% versus 67.9% by GeneXpert,
respectively.
Conclusion: There is a good agreement between cytology
reading and GeneXpert. GeneXpert is very helpful
diagnostic tool for TB lymphadenitis in a country where
there are few pathologists. We cannot conclude which
method is superior because we do not have the gold
standard, culture to compare with in this study.
Contribution of GeneXpert in MDR-TB case

finding in Ethiopia
J Degu,1 G Gudeto,1 J Seid,1 D Habte,1 G Ayana2 1HEAL TB
Project, Management Sciences for Health, Addis Ababa,
2
Regional Laboratories Capacity Building, Ethiopian Public
Health Institute, Addis Ababa, Ethiopia.
Background: Ensuring adequate availability of rapid

tuberculosis (TB) diagnostic tools such as GeneXpert is
essential for early detection and prompt treatment of TB.
Toward this end, the USAID-funded Help Ethiopia
Address the Low TB Performance (HEAL TB) project
supported the roll out of GeneXpert MTB/RIF in two
large regions of Ethiopia. We present the experiences of
GeneXpert implementation and its contribution in multidrug resistant TB (MDR-TB) case finding in the two
regions.
Interventions: In collaboration with the RHBs and the
Ethiopian Public Health Institution (EPHI), HEAL TB
trained laboratory personnel, developed standard operating procedures and reporting tools, provided site level
mentoring, deployed short-term technical assistance
through an international expert, and monitored progress
through monthly reporting and feedback. Moreover
HEAL TB oriented clinicians on the eligibility criteria
for GeneXpert, facilitated sample transportation, and
ensured adequate supply of cartridges. The trends in
GeneXpert utilization and its contribution in the
diagnosis of MDR-TB cases were evaluated on a
quarterly basis, and the contribution was compared with
that of conventional culture methods.
Results and lessons learnt: GeneXpert service started in

the two regions as of January 2014 whereas the diagnosis
of MDR-TB before January 2014 was only made by LPA
culture. Of 108 MDR-TB cases reported in JanuaryMarch 2014 when the project introduced Xpert use,
58% were diagnosed by Xpert while 42% were
diagnosed by LPA culture. By the end of June 2015, a
total of 41 GeneXpert machines were functional in the
two regions. GeneXpert has progressively replaced
culture in the diagnosis of MDR TB; it contributed for
74 (93%) of the 80 MDR-TB by June 2015. The number
of new cases of MDR-TB enrolled to Second Line Drugs
per quarter has doubled since the introduction of
GeneXpert service in the HEAL TB supported regions.
Conclusion: GeneXpert contributed to doubling of
MDR-TB case finding and enrollment in care. More
effort is needed to further strengthen the site level service
utilization capacity.
The use of verification and external quality

assessment panels to monitor the accuracy of
Xpert MTB/RIF testing across 18 countries
1
1,2
A David, V Magida, W Stevens, L Scott, V Deyde,

A Adelekan,3 C Gilpin,4 R Coombs5 1National Priority
Program, NHLS, Johannesburg, 2Department of Molecular
Medicine and Haematology, University of the Witwatersrand,
Johannesburg, 3Centres for Disease Control, Johannesburg,
South Africa; 4Global TB Programme, World Health
Organization, Geneva, Switzerland; 5Department of AIDS
Clinical Trial Group, University of Washington, Seattle, WA,
USA.
Background: More than 10 million Xpertw MTB/RIF

(Cepheid) cartridges and 17 883 GeneXpert (Cepheid)
instrument modules, used for diagnosing Mycobacterium
tuberculosis have been procured globally (up to Q4
2014). Quality management of this molecular test
encompasses verification, external quality assessment
(EQA) and real time instrument/assay performance
monitoring. Verification (determining an instrument
[GeneXpert module] is fit for purpose) is recommended
upon instrument installation or relocation, module
replacement and module calibration. EQA (or proficiency testing) is performed to determine pre-analytical,
analytical and post analytical user performance and real
time monitoring through cloud based software (such as
RemoteXpert, Cepheid) can assist with daily instrument/
site/country performance. The Dried Culture Spot (DCS)
proficiency testing matrix endorsed by the WHO to
verify GeneXpert instruments is also available as a global
EQA program and both can be used to monitor
performance.
Method: DCS material consists of inactivated M.
tuberculosis (susceptible to Rifampicin (RIF) for verification). DCS for EQA is tested 3 times per year
comprising 4 DCS (MTB with RIF susceptible, RIF
resistant and non-TB mycobacteria). Submission and
reporting of DCS results is managed through an in-house
software program www.tbgxmonitor.com.
Results: A total of 18 countries currently participate in
the DCS EQA program. The Global participants (.1site)
S51
include: South Africas NHLS (207 sites), AIDS Clinical

Trial Group (11 countries world-wide), Walter Reed
Army Institute of Research (3 countries), and Namibia
Institute of Pathology. Instruments verified with DCS
material and registered on www.tbgxmonitor.com have
fluctuated from over the years since 2011 to 2015. EQA
uptake increased from 17 participants in 2012 to 295 by
second panel in 2015. By 30 April 2015, 4.377
verification DCS were performed globally. DCS used
for EQA reported an overall 98.2% correct result.
Incorrect results accounted for 0.7%, 0.6% and 0.8%
for device error rates in 2013, 2014 and 2015 (up to Q1),
respectively. Conclusion: The DCS quality program
highlights the successful, rapid implementation of quality
assured Xpertw MTB/RIF testing globally. The standardization of testing material and minimal variation of
testing sites illustrates stability and robustness of
instruments. High quality assay performance over time
as well as consistency among sites and users was
demonstrated.
Quality indicators and maintenance experience

of Xpert MTB/RIF in South Africa
P Marokane,1 S Molapo,1 V Magida,1 L Berrie,1
W Stevens,1,2 L Scott2 1NPP, National Health Laboratory
Service, Johannesburg, 2Department of Haematology and
Molecular Medicine, University of Witwatersrand,
Johannesburg, South Africa.
Background: South Africa initiated the roll out of the

Xpertw MTB/RIF assay across 221 laboratories performing smear microscopy. To ensure ongoing quality of
services, the National Priority Programme (NPP) of
National Health Laboratory Services (NHLS) developed
and monitored quality measures for molecular testing
that are consistent with good laboratory practices.
Method: The performance of the 221 laboratories is
reviewed on a monthly basis. All analysers (n 309)
deployed within the NHLS are connected and interfaced
to the Laboratory information System (LIS) allowing for
central collection of patient results. Data extracted from
the Central Data Warehouse (CDW) is analysed to
determine the national Mycobacterium tuberculosis
complex (MTBC), Rifampicin (RIF) resistance and error
rates. In addition, 211 devices were connected to the
GeneXpert operational dashboard which allowed the
NPP to receive data in real time and respond to potential
laboratory, assay and analyser problems occurring in a
more efficient, timely and targeted approach. The use
and analysis of the real time data enhanced the ability to
respond to instrument failures that might results in
module replacements and maintenance, operator training needs, potential laboratory contamination and assay
related issues followed by root cause analysis.
Results: A total of 6 538 967 specimens have been
processed as at 31 July 2015. The national TB positivity
rate among individuals presumed to have MDR-TB or
HIV-TB was found to be up to 16-18% in the first year
and 13-14% in the second year, 10-11% in the third and
fourth years and has reduced to 9% in the 5th year, after
introduction of Xpertw MTB/RIF assay. Average Rifam-
S52
picin resistance detection rates have remained around

7% since project inception with some geographical
variations. Average error rate has ranged consistently
below 3%. Over 3 million specimens were uploaded on
the Xpert operational dashboard as at 31 July 2015.
Comparisons of data from the RemoteXpert against LIS
showed a 1% bias due to removal and repeat on error
testing not extracted by the LIS. Since March 2011 to
December 2014 South Africa replaced (33%) 1272 out
of 4168 modules. Conclusion: The performance indicators have never been reported in real time and the
implementation and utilization of these systems improved the overall programmatic monitoring and evaluation, which ultimately improved the quality of the
laboratory processes and results.
46. Strengthening community systems:

building the evidence for impact
Community-based monitoring and evaluation
in Uganda: ensuring TB services delivered by
public and private sectors are available and
accessible
P Masembe1 1African Young Positives Network, Kampala,
Uganda.
One third of people estimated to have TB are either not

reached for diagnosis and treatment by the current health
systems. Even in patients who are identified, TB or
MDR-TB is often diagnosed and treated late. In order to
reach the unreached and to find TB patients early in the
course of their illness, a wider range of stakeholders
already involved in community-based activities needs to
be engaged. Civil society organizations (CSOs) that are
active in community-based development, particularly in
primary health care, human immunodeficiency virus
(HIV) infection and maternal and child health, but have
not yet included TB in their priorities and activities. CBO
membership consists entirely of the community members
themselves, so these organizations can be considered to
represent the community most directly. NGOs and other
CSOs engage in activities that range from community
mobilization, service delivery and technical assistance to
research and advocacy. The strengths of CSOs active in
health care interventions at the community level include
their reach and spread and their ability to engage
marginalized or remote groups. These organizations
have a comparative advantage because of their understanding of the local context. Greater collaboration
between CSOs and local and national governments could
greatly enhance development outcomes. Reducing the
complexity of transactions and increasing the speed of
facilitation are key factors that improve the operating
environment for community actors. NTPs or their
equivalents have the responsibility of creating enabling
national or local legal, policy and administrative
environments to support the effective engagement of
NGOs and other CSOs in TB activities. CSOs should also
stimulate and support the development of an enabling
legal and policy environment through constructive

dialogue and engagement with the NTPs or relevant
legislative structures. The coalition should meet regularly
and have mechanism for sharing information of common
interest to improve contributions to national TB plans,
discuss challenges and opportunities. The NTP, with
inputs from the coalition and other stakeholders, should
prepare a national policy to facilitate effective engagement of CSOs in TB prevention, diagnosis, treatment,
patient care and support, and in research activities
aligned with the national health strategy.
Contributions of communities in improving

case detection in rural Ethiopia through TB
REACH
M A Yassin,1 D Gemechu,2,3 S Theobald,3 L Cuevas3 1The
Global Fund, Geneva, Switzerland; 2REACH Ethiopia,
Hawassa, Ethiopia; 3Liverpool School of Tropical Medicine,
Liverpool, UK.
Introduction: In 2013, WHO estimated that three million

cases of tuberculosis (TB) were missed. Ethiopia contributed to 3% of missed cases mainly due to health
system and community factors. Different interventions
were tried but with limited gain in case finding. We
report successful implementation of TB REACH project
in improving case finding in remote and rural communities in southern Ethiopia.
Methods: This is an innovative community-based collaborative project with the Ministry of Health of
Ethiopia implemented in southern Ethiopia reaching a
population of 7.5 million. The key interventions included
familiarization, capacity strengthening, communitybased intervention by engaging HEWs in TB diagnosis
and treatment, mobile communication, slide referral and
ensured sustained continuum of care including contact
tracing, provision of Isoniazid Preventive Therapy (IPT)
for asymptomatic under five years children. District
supervisors were hired; motor bikes were provided
training and contributed to diagnostic capacity building
over four years.
Results: During this period, more than 180 000
presumptive TB cases were identified; sputum collected
and slide fixed by HEWs in the community and slides
were collected by district field supervisors for examination. Of these, more than 9000 smear-positive TB cases
were diagnosed from the slides prepared and fixed by
HEWs. Over all in the project area, more than 13 000
smear-positive TB cases and more than 29 000 all forms
of TB were diagnosed and treated. The community-based
intervention contributed to about 70% cases detected in
the project area. More than 9000 household contact
tracing was done and screened more than 35 000
contacts, identified more than 4000 symptomatic contacts, examined more than 3000 symptomatic and
detected 243 smear-positive cases, about 2500 under
five years asymptomatic children were put on IPT. About
6000 presumptive cases were examined using GeneXpert
machine and detected 634 MTB cases.
Discussion: Our approach improved cases finding and

treatment outcome in remote and rural communities as
the interventions addressed key barriers: physical and
socio-economic factors affecting the health service
utilization. Such approaches that increase access to
services and ensure equity could be done in high burden
resource-constrained settings. However, it requires proper planning, monitoring and evaluation for its success in
addition to government commitment.
S53
during the course of treatment to see the prognosis.

Shasthya Sebika (SS) receives 6.42$ per successfully
treated TB patient as incentive from BRAC.
Results and lessons learnt: Among the infectious (new
smear positive) TB cases over the decade from 2003 to
2013, treatment success rate of the yearly cohort
increased from 88.88% to 94.82%. The proportion of
patients who were lost to follow-up decreased from
1.62% to 0.89%. The death rate also decreased from 7%
to 3.5%.
Conclusions: Involvement of community people as
DOTS provider and door step efforts in the community
regarding TB control found to be effective in better
treatment outcome among infectious cases over the
years. Early diagnosis and prompt treatment initiation
of TB cases from the community will reduce the spread of
transmission.
47. Getting patients treated for latent TB

infection (LTBI) as TB prevention
Implementing shorter LTBI treatment regimens
in low-incidence settings: lessons learned from
3HP post-marketing surveillance
Expanding TB diagnosis and community-based
TB management through community systems
in Bangladesh
C Ho,1 N Nwana,1 A Sandul,1 S Morris,1 R Moro,1

B Stewart1 1Division of TB Elimination, Centers for Disease
Control and Prevention, Atlanta, GA, USA.
S Islam,1 M Rana,1 M Siddiqui,1 K Uddin,1

M Akramul Islam,1 M Husain2 1Department of Health,
BRAC, Dhaka, 2NTP, Dhaka, Bangladesh.
Background: Published studies document that shorter

treatment regimens for latent TB Infection (LTBI) are
associated with higher completion rates. A large clinical
trial showed the 12-dose weekly isoniazid-rifapentine
(3HP) regimen to be non-inferior to the standard 9month daily isoniazid regimen; it was associated with
higher completion rates and had a distinct profile of
adverse events. The aim of the 3HP post-marketing
surveillance project was to evaluate treatment completion rates and adverse events in non-research program
settings.
Methods: An observational prospective cohort of patients from sixteen project sites who started 3HP between
June 2011 and December 2013 were followed to
treatment completion. Before each dose was dispensed,
patients were asked about symptoms since the previous
visit. Three different tiers of data collection were
designed to accommodate differences in information
routinely collected by sites. Uniform tracking forms for
each tier were used to collect patient demographic
information, symptom screening results, and final
treatment outcome.
Results: Data from 3307 persons receiving 39 530 doses
were available for analysis. In this project, 2884/3307
(87%) persons completed 3HP; completion rates ranged
from 81-100% across sites. Discontinuing treatment was
associated with homelessness (ARR1.65; 95%CI: 1.202.27), age .65 years (ARR1.61; 95%CI: 1.21-2.13),
and being non-Hispanic White (ARR1.38; 95%CI:
1.14-1.68). Reporting at least one symptom before the
11th dose was associated with treatment discontinuation
Background and challenges to implementation: TB is a

major public health concern in Bangladesh. BRAC, a
development organization initiated community based
tuberculosis control programme in 1984 in one subdistrict to control TB. This model was gradually
expanded since 1994 in collaboration with National
Tuberculosis Control Programme (NTP) and currently
covering two-third of the country with 93 million
populations. Considering the long treatment duration,
community based interventions intensified throughout
BRAC implementing areas to increase TB case notification, to improve treatment success rate and to minimize
unfavorable treatment outcome. Interventions: BRAC
expanded peripheral laboratory facilities with solar
system in very remote areas to increase the identification
of smear positive TB cases. Outreach sputum collection
centers are conducted to improve access to diagnosis.
BRACs frontline health workers known as Shasthya
Sebika (SS) plays an important role in implementing
community based DOT. They disseminate TB messages
in the community during household visits and refer TB
symptomatic for sputum examination at NTP designated
laboratories. Treatment of all TB patients is initiated by
graduate medical doctors and daily DOT is ensured by
Shasthya Sebika (SS). BRACs health workers give proper
counseling and motivation to the TB patients for daily
DOT and ensure follow up sputum examination timely
S54
(OR 4.5, 95%CI: 3.6-5.6). Shortness of breath (OR 16.4;

95%CI 10.2-26.3) and fever/chills (OR 10.3; 95%CI 9.8
10.7) were most associated with treatment discontinuation, while nausea, the most commonly reported
symptom, was least associated with treatment discontinuation if it was the only symptom reported (OR 0.4;
95%CI 0.2-0.7). In analyses of 2400 patients with
medical information available, the additional risk factor
of having ever smoked was associated with discontinuing
3HP (ARR1.50; 95%CI: 1.21-1.85; P , 0.001). There
were no deaths or permanent medical sequelae reported
during treatment.
Conclusion: Completion of 3HP treatment in U.S.
program settings is high, and similar to that seen in
published trials. Isolated nausea, while commonly
reported in patients, was not associated with treatment
discontinuation. Persons who were homeless, age .65
years, non-Hispanic White, or smokers were more likely
to discontinue treatment. The 3HP regimen is generally
safe and well tolerated in these US program settings.
48. Beyond accelerated drug

development: building successful
partnerships and platforms as a critical
path to TB drug regimens and PreDiCT-TB
Collaborative approaches to defining
contribution of animal models
1
H Pertinez, L Bonnett, G R Davies

Liverpool, Liverpool, UK.
1 1
University of
Background: Mouse model data has played a central role

in preclinical development of anti-tuberculosis drugs
despite ongoing debate about representativeness and
reproducibility. Within the PreDiCT-TB consortium,
close collaboration between experimentalists and statistical modellers has been central to optimising designs and
applying a common modelling approach to analysis of
mouse experiments performed for a common set of drug
combinations. Collaborating with external scientific
partners, PreDiCT-TB has begun to assemble a more
comprehensive database of mouse studies than previously, facilitating comparisons between such models and
also with clinical outcomes on the basis of ranking of
treatments and their effect sizes. This meta-analytic
approach has been informative in other therapeutic areas
and promises to lead to a deeper understanding of the
strengths and weaknesses of this key step in preclinical
development.
Methods: Timecourse profile datasets of M. tuberculosis
load in mice under therapy, typically CFU plated on solid
media from lung homogenate, were fitted with mathematical models describing bacterial elimination by
nonlinear regression methods. Empirical models describing a monoexponential elimination of M. tuberculosis
were generally applied and exponential elimination rate
constants (k_net) estimated this parameter is equivalent
to the slope of elimination seen with log-transformed
CFU data. The relative efficacy of therapies was assessed

by ranking them in order of their values for this
parameter faster elimination rates reflected in a more
negative k_net parameter value.
Results: Ranking tables, according to k_net for various
therapies in 3 mouse models are presented with
correlation coefficients to compare parameter values
and rankings in the separate models against each other,
and a meta-analysis with accompanying Forrest plot was
carried out to derive a general ranking of therapies in
mice. Finally all mouse rankings were compared to
rankings of therapies in a clinical setting, based on
several endpoint measures of efficacy (e.g. EBA 0-x, TTP
at 8 weeks etc.) derived by a separate meta-analysis also
carried out under Predict-TB.
Conclusion: Of the models analysed so far, the Janssen
mouse dataset showed the closest match to the ranking of
therapies in human, with a general combined mouse
model ranking showing some correlation with clinical
rankings of EBA0-2 and 8 week TTP data. Ongoing
work will analyse and incorporate further data from
other mouse models.
Learning from clinical trial data

L Bonnett,1,2 G R Davies1 1Department of Clinical Infection,
Microbiology, and Immunology, University of Liverpool,
Liverpool, 2Department of Biostatistics, University of
Liverpool, Liverpool, UK.
Background: To understand and calibrate preclinical

development of anti-tuberculosis (TB) drugs it is
necessary to define their performance in real-life clinical
trials. A systematic review and summary meta-analysis
were undertaken to examine the existing evidence base,
as well as historical and current development practices,
particularly the progression of compounds from phase II
to phase III studies.
Methods: Relevant studies evaluating key drugs in firstline treatment for drug-sensitive and isoniazid monoresistant individuals with pulmonary TB were included.
Early clinical outcomes of interest were EBA over 2, 7
and 14 days, and the proportion of patients with negative
culture results at 8 weeks. Death, treatment failure, and
relapse were the main later phase outcomes of interest.
Results: The review included more than 300 published
trials comprising over 80 000 participants. 66 and 91
distinct drug combinations were identified in Phase II
and III trials respectively, though few were common to
both phases - only 9 distinct regimens were considered
across multiple studies and early phase endpoints. In
addition to the lack of consistency in the investigated
treatments across studies, trial characteristics and
reported outcomes were also highly variable. This
suggests the need for a core outcome set for reporting
in TB treatment trials a minimum set of outcomes to be
reported in all future TB trials. Individual participant
data has been requested for all included studies with
mixed success. Without this data it is currently difficult
to evaluate outcomes consistently across trials, including
analyses not performed by the original investigators.
Conclusions: The sparseness of the existing evidence base

in trials of treatment for pulmonary TB is concerning it
makes comparison of results at different phases of
development quite limited. It also suggests that historical
and current development practices should be critically
reviewed. In fact, the existing dataset which includes
over 60 years worth of data is too small to provide
convincing support to the effectiveness of the current
typical drug development pathway, particularly the
progression of compounds from phase II to phase III
studies. The lack of consistency in reporting of outcomes,
and the lack of individual participant data are additional
challenges which need addressing if the evidence base for
treatments of TB is to be improved.
S55
Conclusion: The CPTR modeling and simulation efforts

propose to develop quantitative tools to provide answers
regarding how the interaction between pathogen dynamics, drug exposure and immune system interactions can
be quantified and applied to inform TB drug development decisions.
Improving TB clinical trial design through

modeling and simulation
K Romero 1Department of Clinical Pharmacology, Critical
Path Institute, Tucson, AZ, USA.
Background: The Critical Path to TB Drug Regimens

(CPTR) Regulatory Sciences programs are focused on derisking the decision-making in the TB regimen development process. These efforts include the development of
quantitative tools to optimize TB trial design through
simulations. Each of these programs is enabled by the
CPTR TB clinical trial data platform that includes critical
data contributions from industry and academia. These
clinical trials datasets have been mapped to a unified TB
Data Standard, allowing the data to be integrated for
more informed quantitative analyses.
Methods: A Physiologically-based pharmacokinetic
(PBPK) model that captures the complex mechanistic
drug distribution processes in the TB-infected, inflamed,
and damaged lung has been developed, with the goal of
improving treatment and regimen selection designed to
optimize drug exposure across the heterogeneous sites of
action in the lung. This model will optimize clinical trial
design for first-in-human studies of anti-TB regimens.
Additionally, there is a need for quantitative tools to
more accurately evaluate efficacy in phase II clinical
trials for combination regimens for TB and to reliably
predict clinically-relevant endpoints for phase III clinical
trials, based on early efficacy evaluations during phase II.
Based on phase II and III patient-level data, a model of
the longitudinal changes in time-to-positivity (TTP) is
being developed, capturing relevant sources of variability. A subsequent model linking changes in TTP from
liquid culture to durable cure will be developed.
Simulations of longitudinal TTP data will be performed
in sub-populations of interest and the resulting effect on
durable cure will be determined.
Results: A PBPK model for the TB-infected lung, a virtual
South African population, and compound files for
standard-of-care drugs has been developed and implemented on the Simcyp platform, and is fully available for
drug developers. Preliminary results of the TTP model
show that baseline pathogen burden is a significant
predictor of the rate of TTP change over time, allowing
the identification of distinct sub-populations.
49. Sanatoria: back to the future?

The sanatoria movement: past and future
A Story 1FIND & TREAT, London, UK.
Despite Kochs brilliant exposition of the mode by which

tuberculosis spreads, its complex etiology continued to
mask the facts of transmission for many years. The
sanatorium movement was largely founded on the belief
that it offered a promise of cure and much needed care as
opposed to a means of limiting spread of infection. From
the outset the movement was contentious and hotly
debated. Insufficiency of beds to accommodate all
sufferers created a class divide, with the poor largely
reliant on home care; segregation fuelled stigma; beds
diverted scarce resources from ambulatory care; and even
under the best conditions, over half of those admitted
were dead within five years. With the advent of effective
treatment and declining cases the movement was largely
disbanded. The reality of untreatable tuberculosis has
come full circle. Today there is an urgent necessity to reexamine the rationale that underpinned the sanatoria
movement and its potential contribution to tuberculosis
control. The need for facilities to offer long term
hospitalisation and care for patients with M/XDR
disease and for patients whose complex social circumstances render ambulatory care a near impossibility is a
global public health priority. While the ethical, moral
and public health arguments remain complex and
contentious, there is now a need to move beyond debate
to action. This first talk will set the scene for a
symposium which is long overdue.
S56
Netherlands: high-income, low-incidence, high

programme budget
W De Lange 1Department of Pulmonary Diseases and
Tuberculosis, Tuberculosis Center Beatrixoord University
Medical Center Groningen, Groningen, Netherlands.
The Netherlands has two national tertiary TB referral

and expertise centers. Apart from 10 beds for noninfectious TB patients, e.g. with compliance, addiction or
psychiatric problems, modern sanatorium Beatrixoord is
equipped with state-of-the-art isolation rooms for twenty
persons. These are suitable for admitting infectious XDR
patients as well as severe immunocompromised, normal
sensitive, TB-patients, in separate rooms on the same
ward, for a price competitive with a general hospital. TB
in the Netherlands is predominantly an imported disease.
In 2014, thirty-four different nationalities stayed on the
TB ward. The team is specialized in the transcultural
approach of individual patient providing medical,
nursing, nutritional, physiotherapy, psychiatric, social,
rehabilitation and activity related support. Beatrixoord is
also the designated centre in the Netherlands for
mandatory isolation, when infectious TB patients ignore
all advice for voluntary isolation. It is also located close
to the national reception centre for asylum seekers in the
Netherlands, where radiographic TB screening is performed within 24 hours of arrival, allowing for
immediate admission and isolation or further examination, e.g., by bronchoscopy. The Netherlands is a country
in the TB pre-elimination phase, with only 823 TB
patients diagnosed in 2014. Maintaining TB expertise is
difficult among physicians, even among pulmonologists
and and infectiologists. The treatment of MDR/XDR-TB
patients is the responsibility of the national TB consultants for clinical TB working in the tertiary TB referral
centers, with a successful outcome in 86% of the cases.
Therapeutic drug monitoring, especially for second- and
third line TB drugs, is a focus of research at Beatrixoord.
The two tertiary pulmonologists of Beatrixoord annually
perform approximately 800 consultations by telephone
and work very close with the public health TBprofessionals. In the Netherlands, a low-incidence
country, two highly specialized tertiary TB referral
centers compliment primary and secondary medically
and socially complicated TB care, preserving TB
expertise and promoting research.
Barcelona: high-income, low-incidence, low

programme budget
I Monedero 1International Union Against Tuberculosis and
Lung Disease, Paris, France.
Barcelona is one of the most touristic and cosmopolitan

cities in Europe. Like in any other big city, TB thrives,
hitting the most vulnerable populations hard. This
inconvenient truth is frequently denied by politicians
and institutions, leading to low investment in outreach
and socially excluded populations. Barcelona was badly
hit in the 1980s by TB at the hand of two other increasing
epidemics: HIV and intravenous drug use. Mortality
rates were so high that there was no excuse not to build a
specific system to manage TB in the city. Following the

lessons learned from New York MDR-TB-HIV epidemic,
a multi-stakeholder system was created. It was fine-tuned
during years, lessons and errors. It would finally be
known as the Barcelona model for TB control. One of
the pillars of this model is the existence of what we could
call a modern sanatorium where TB patients are
admitted when presenting difficult clinical and also
social conditions. The use of such a facility has proved
to increase the cure rates among those more difficult to
be cured: TB-HIV, MDR and XDR-TB and also TB
patients presenting alcoholism, multidrug addictions and
other mental or social conditions that jeopardize patient
adherence. The overall Barcelona model and the relevance of a sanatorium in Barcelona city center will be
reviewed during the lecture.
Dominican Republic: middle-income, highincidence, low programme budget

M Rodriguez 1MDR-TB Unit, National Tuberculosis Control
Program, Santo Domingo, Dominican Republic.
The Dominican Republic (DR) is a Caribbean country

with an estimated tuberculosis incidence of 59.6. The
only surveillance study performed in 1994-1995, showed
initial MDR of 6.6% and 19.7% among previously
treated patient. In 2006 with financial aid from the
Global Fund and the technical assistance of The Union
and PAHO/WHO, 24 MDR patients started second line
drugs (SLD) treatment at a special unit in Juan Pablo Pina
Hospital (HJPP), San Cristobal,
where 18 beds were
assigned distributed in 3 rooms. The DOT-plus project

protocol was accepted with a mixed model of care: a
hospitalized phase, initially thought to be the intensive
phase, and an ambulatory phase at centers near the
patients home. To increase the accessed to treatment, the
hospitalization stage was reduced from 6 months in the
original project to 90 days. It has been reduced to 45-60
days. In 2008, 9 more beds where assigned for patients
from the north-central part of the country in Luis Morillo
King Hospital, La Vega. Rooms in the hospital units are
not individualized nor have negative pressure. There are
located in the second and third level of old hospitals with
a strong natural ventilation. By 2010, a complete
ambulatory treatment model started in selected health
centers, including primary care centers. Actually 22
centers have started SLD ambulatory treatments and
there are 20 more under evaluation. Three rooms for
separation were accommodated in HJPP in 2013. Until
2014, a total of 845 cases have initiated SLD treatment:
719 MDR/RR, 6 XDR, 48 poly-resistant, 72 presumed
MDR cases. Eighty-six percent (86%) initiated with the
mix model of care. All XDR patients started treatment
hospitalized. The hospitalization period lasted between 6
to 8 months. Reviewing all cases from 2010 until 2012
(331 cases): 73% of those who started in mixed model
had a successful outcome and 71% of the complete
ambulatory care. The default rate was high in both but
greater among those in the complete ambulatory model
(26% versus 19%). Death was reported in 6% of the mix
model. The 2012 MDR cohort success rate was 72%, Of

2 XDR patients that initiated at 2011: 1 died and 1
failure. After having access to Group 5 SLDs, of the 4
patients in 2013, 3 are cured and 1 had failure, after
irregular adherence. Patients with XDR had a hospitalization period of at least 6 months. For the Dominican
Republic, the mixed model of care has showed good
results in the management of drug-resistant TB.
South Africa: low-income, high-incidence, high

programme budget
J Ferreira 1North West Department of Health, MDR/XDR-TB
Unit, Tshepong Hospital, Klerksdorp, South Africa.
In South Africa, tuberculosis is still the number one cause

of morbidity and mortality, with 350 000 cases
diagnosed annually. In 2013 cases diagnosed as Rifampicin resistant were 21 000, and confirmed MDR just less
than 7000. There are 2500 beds available for DR-TB in
the country, so we have decentralised care to a large
extent. Well patients are managed as out-patients, with
injection teams travelling to their homes to administer
injections, and only sick patients are admitted. Patients
with susceptible TB are managed at primary health care
level. We have a population of 50 000 people in 9
provinces. The prevalence of HIV is estimated to be 1520%, but among DR-TB patients it is as high as 75-80%.
Susceptible TB has an 80% treatment success rate, but
for MDR it drops to a country average of 40%, with a
20% defaulter rate, and a 20% mortality rate, for both
HIV co-infected and non-infected patients. I work in the
MDR/XDR-TB Unit at Tshepong Hospital in Klerksdorp, in the North West Province of South Africa. We
provide services to 4000 people, as the Centre of
Excellence in the province. The unit has 76 MDR and
22 XDR beds. Since the introduction of GeneXpert in
2011, the number of patients diagnosed with drugresistant TB in our province has more than doubled (720
in 2014). Patients who are admitted in the unit have
varied demographics: a) All smear positive patients for
isolation purposes, until smear negative; b) Smear
negative patients who are weak and sick; c) Patients
who cannot get to the clinic for injections/no injection
team is available; d) Patients with poor social circumstances, and lack of support at home; e) Children with
insufficient adult supervision at home; f) Chronic
defaulters, who are repeatedly admitted in a poor
condition; g) XDR and Pre-XDR patients until culture
conversion. Since the opening of our unit in 2000, we
have had succesful outcomes for MDR-TB in 60% of
cases, but since the necessary advent of decentralisation,
we have noticed a significant increase in the number of
defaulters in our province, as we have always strived to
create a culture of adherence while patients are admitted
in the unit. Conclusion: In South Africa there is a definite
place for In-patient Units for management of DR TB,
when taking into consideration the length of treatment,
against the background of poor social circumstances.
S57
50. Smokeless tobacco: a silent killer

Smokeless tobacco, cancers and cardiovascular
diseases: a systematic review and metaanalysis
S Shah,1 A L Vidyasagaran,1 K Siddiqi,1 M Jawad2
Department of Health Sciences, University of York, York,
2
Imperial College, London, UK.
Background: Smokeless tobacco (ST) has long been

recognised as a carcinogen in humans. Other potential
health effects associated with its use include adverse
cardiovascular outcomes such as myocardial infarction
and stroke. A previous meta-analysis reporting the risk of
cancers and cardiovascular diseases (CVD) among users
of ST was restricted to studies from Sweden and North
America. Recognising that ST consists of diverse
products, and the products consumed in Asia are
different from those consumed in Europe and North
America, we aimed to update the existing review and
expand on it by including studies from all geographical
areas. We aimed to determine whether the ST products in
Asia showed differences in their cancer and CVD risks as
compared to products consumed in Sweden and North
America.
Design/Methods: A systematic search of the literature
was conducted by combining an exhaustive list of terms
for ST with terms for specific cancers and CVD
outcomes. The search was not restricted by language,
search period, geographical areas or population characteristics. Studies included in the meta-analysis were
identified after initial screening of abstracts, followed
by full-text screening of short-listed papers. Both
screenings were carried out independently by two
reviewers. The meta-analysis was carried out using
RevMan version 5.3.
Results: 39 studies reporting risk estimates for cancers of
mouth, pharynx, oesophagus, larynx, pancreas and lungs
were included. The pooled relative risk (RR) for mouth
cancers was 3.43 (95%CI 2.265.19), pharynx 2.23
(95%CI 1.553.20), and oesophagus 2.17 (95%CI 1.70
2.78). 14 studies were included in the meta-analysis for
CVD outcomes. While the pooled estimates did not show
a significantly increased risk for CVDs, the RR for MI
from studies in Asia was 1.40 (95%CI 1.011.95), and
1.57 (95%CI 1.241.99) from one large case-control
study conducted in 52 countries.
Conclusion: ST is widely consumed worldwide and its
use results in substantial morbidity and mortality.
However, there are marked differences in the type of
ST products available, patterns of consumption and
associated risks in different regions and countries.
Studies from Asia indicate an increased risk of cancers
and CVDs for ST use, whereas results of studies from
Europe and Americas do not support such an association.
S58
Global burden of disease due to smokeless

tobacco consumption in adults: analysis of data
from 113 countries
K Siddiqi,1 S Shah,1 A Vidyasagaran1 1University of York,
York, UK.
Background: Smokeless tobacco is consumed in most

countries in the world. In view of its widespread use and
increasing awareness of the associated risks, there is a
need for a detailed assessment of its impact on health. We
present the first global estimates of the burden of disease
due to consumption of smokeless tobacco by adults.
Methods The burden attributable to smokeless tobacco
use in adults was estimated as a proportion of the
disability-adjusted life-years (DALYs) lost and deaths
reported in the 2010 Global Burden of Disease study. We
used the comparative risk assessment method, which
evaluates changes in population health that result from
modifying a populations exposure to a risk factor.
Population exposure was extrapolated from countryspecific prevalence of smokeless tobacco consumption,
and changes in population health were estimated using
disease-specific risk estimates (relative risks/odds ratios)
associated with it. Country-specific prevalence estimates
were obtained through systematically searching for all
relevant studies. Disease-specific risks were estimated by
conducting systematic reviews and meta-analyses based
on epidemiological studies. Results We found adult
smokeless tobacco consumption figures for 115 countries
and estimated burden of disease figures for 113 of these
countries. Our estimates indicate that in 2010, smokeless
tobacco use led to 1.7 million DALYs lost and 62 283
deaths due to cancers of mouth, pharynx and oesophagus
and, based on data from the benchmark 52 country
INTERHEART study, 4.7 million DALYs lost and 204
309 deaths from ischaemic heart disease. Over 85 % of
this burden was in South-East Asia.
Conclusions: Smokeless tobacco results in considerable,
potentially preventable, global morbidity and mortality
from cancer; estimates in relation to ischaemic heart
disease need to be interpreted with more caution, but
nonetheless suggest that the likely burden of disease is
also substantial. The World Health Organization needs
to consider incorporating regulation of smokeless tobacco into its Framework Convention for Tobacco Control.
Smokeless tobacco consumption among

adolescents: a global epidemiological
perspective
I Agaku,1 O Ayo-Yusuf2 1Department of Office on Smoking
and Health, Epidemiology branch, National Center for
Chronic Disease Prevention and Health Promotion, U.S.
Centers for Disease Control and Prevention, Atlanta, GA,
USA; 2Department of Office of the Director, School of Oral
Health Sciences, University of Limpopo, Medunsa Campus,
Pretoria, South Africa.
Background: Smokeless tobacco (ST) products can lead

to tobacco addiction, disease, and death. The cultural
differences in global adult tobacco use suggest that
adolescents tobacco use may also vary across cultures.
We assessed ST use among adolescents in 57 countries
during 2007-2010 using school-based surveys.
Methods: Data were analyzed for 57 countries for which
nationally representative data were available, using the
20072010 Global Youth Tobacco Surveys (for 55
countries); the 2009 US National Youth Tobacco Survey,
and the 2008-2009 Canada Youth Smoking Survey.
Number of countries by World Health Organization
(WHO) region were: Africa, AFRO (n 11); Eastern
Mediterranean, EMRO (n 9); Europe, EURO (n 11);
the Americas, AMRO (n 14); South-East Asia, SEAR (n
8), and Western Pacific, WPRO (n 4). Analyses were
restricted to persons aged 13-15 years in each country.
Current ST use was defined as past 30-day use of chewing
tobacco, pinch, snuff, dip, or snus. Data were weighted

and analyzed by sex, WHO region and World Bank
national income categories. Cross-country dispersion in
ST use was assessed with the Coefficient of variation
(CV).
Results: ST use varied widely among countries
(CV0.62); overall prevalence ranged from 1.1% in
Montenegro to 16.4% in Congo. Prevalence by WHO
region was: AFRO (median 7.4%; range5.4% to
16.4%); EMRO (median 5.1%; range1.6% to
12.6%); EURO (median 2.0%; range1.1% to
6.9%); AMRO (median 5.3%; range1.8% to
9.8%); SEAR (median 6.3%; range2.8%, to 9.4%),
and WPRO (median 5.1%; range2.1% to 8.7%).
Prevalence by national income category was: low-income
countries (median 6.2%; range2.5% to 16.4%),
lower-middle-income countries (median 6.2%;
range2.0% to 14.4%), upper-middle-income countries
(median 4.7%; range1.1% to 11.3%), and highincome countries (median 3.4%; range1.8% to
9.8%). Overall ST prevalence exceeded 10% in 5
countries 4 in AFRO and 1 in EMRO regions. Girls
ST use equaled or exceeded that of boys in 10 low and
middle income countries.
Conclusions: These findings suggest that region-specific
patterns of ST use exist, with higher ST use observed in
low and middle income countries, particularly in the
AFRO and SEAR regions. Evidence-based interventions
outlined in the WHO MPOWER package, such as raising
tobacco taxes, and warning about the dangers of tobacco
use, may prevent and reduce ST use among adolescents.
Smokeless tobacco addiction: an evaluation of

ST dependence measures in Bangladeshi
consumers
R Huque,1 K Siddiqi,2 S Shah2 1University of Dhaka, Dhaka,
Bangladesh; 2University of York, York, UK;
Background: Smokeless tobacco (ST) is an emerging

public health issue in Bangladesh where adult ST use
prevalence is more than 20% and in certain subgroups of
population its prevalence is more than 40%. ST products
contain high nicotine content which is highly addictive.
With scarcity in research related to ST dependence, we
have conducted a validation study to assess whether and
to what extent the psychometric properties of existing
tobacco dependency measures are a reliable and valid
measure of tobacco dependence among Bangladeshi ST
users. We also explored the underlying constructs and
processes that explain dependence among ST users in
Bangladesh.
Design/Methods: We recruited 200 ST users (50 male
and 50 female) from Dhaka city through convenience
sampling techniques. Target respondents for this study
were adult (age greater than 18 years at the time of study
enrolment) smokeless tobacco users with a history of .1
tin/pouch ST use per week for at least one year, and non
smokers. An existing questionnaire to measure dependence had been used to study participants and its
psychometric properties were assessed using quantitative
S59
methods. Saliva samples were collected from participants

and tested to measure their cotinine concentration.
Results: The findings suggest that majority of the users
used Zarda (68.5%) followed by betel quid with tobacco
(6%). On an average, a user takes ST 11.7 times per day.
Around 9% users take ST daily and 42% users felt the
urge to use ST all the time while 27% had the urge almost
all the time. Only 19% users attempted to quit ST use in
last 12 months and only 21.5% users tried to cut down
the ST use. The findings show that 50% of the users had
very high dependency on ST and 41% had high
dependency. The mean cotinine concentration among
ST users was 462.7 (SD 270.9). Beta coefficient presents
increase in cotinine concentration with decrease in time
to place first dip after waking up.
Conclusion: Bangladeshi ST consumers are highly
dependent on ST and their cotinine concentration is also
very high, which needs attention of policy makers.
Developing and evaluating a behavioural

support intervention for smokeless tobacco
cessation in Pakistan and the UK
O Dogar,1 K Siddiqi,1 C Jackson,1 R Bibi,1 H Thomson,2
I Kellar3 1Department of Health Sciences, University of York,
York, 2Department of Health, Leeds City Council, Leeds,
3
Institute of Psychological Sciences, University of Leeds,
Leeds, UK.
Background: Around 300 million people worldwide

consume smokeless tobacco (SLT). Considered to be a
part of the culture, its use is particularly common in
people of South Asian-origin. A recent Cochrane review
suggests that behavioural support interventions (BSIs)
are likely to be effective in SLT cessation. However,
studies in this review were mainly conducted in the US
and European populations who, compared to South
Asians, use less addictive and less hazardous products
(e.g. moist snuff). Moreover, its consumption among
South Asians has strong socio-cultural dimensions,
which need to be considered in BSIs. We aim to evaluate
the acceptability, feasibility and fidelity of a new BSI for
South Asians trying to quit using SLT.
Methods: This is a two-country, multi-method pilot
study, using qualitative and quantitative methods develop a solid theoretical framework for development of a
SLT cessation intervention that is culturally appropriate
for use with the South-Asian population. The BSI was
developed, tested and refined in three iterative phases:
Development phase (6 months): Two expert panel
workshops were conducted to, first, select the behavioural determinants of SLT use and related behaviour
change techniques (BCTs) based on their importance and
relevance and second, to inform the structure and key
messages for the intervention. Feasibility assessment
phase (8 months): The BSI was tested at five sites in the
UK and Pakistan, delivered by trained cessation advisors
to 32 SLT users. All BSI sessions (pre-quit, quit and postquit) will be audiotaped and analysed for fidelity, 16
participants and four advisors will be interviewed in
depth to assess the feasibility and acceptability. All study
S60
participants will be tested for SLT use status via urine/

salivary cotinine testing at 6 months. Refinement phase
(4 months): The BSI will be refined in light of the results
from feasibility testing.
Results: Findings from the qualitative interviews, the
audio-taped sessions and participant SLT use status at 6
months will inform any further modifications needed in
the content, structure or format of the BSI.
Conclusion: The refined BSI will be the first behavioural
intervention of its kind developed specifically for SLT
users of the South Asian-origin, with a comprehensive
methodology involving the use of BCTs proven effective
for tobacco cessation, ready for effectiveness testing in a
randomised controlled trial.
WHOs Global TB Programme, the European Respiratory Society and other partners in 2015, will be used to
discuss potential digital health approaches for TB care
and prevention under four rubrics: patient care, surveillance and monitoring, programme management and
eLearning. Published literature and results from selected
projects - including experiences from outside the TB field
but which are reasonably applicable - will be used to
illustrate issues relevant to the scalability of the digital
health products prioritised by the strategic agenda for
TB.
51. e/m-Health as a transformational tool

in TB control in high-burden countries
T Gabunia 1University Research Co. LLC Branch Office,

Tbilisi, Georgia.
Setting up eHealth for TB at scale: lessons

learned
D Falzon Global TB Programme, World Health Organization,
Geneva, Switzerland.
In the last decades, advances in information and

communication technologies (ICT) have stimulated
many innovative e/mHealth interventions targeting
various components of TB care and prevention in both
high and low resource settings. The population-level
impact of these efforts remains nonetheless unclear to a
large degree and the evidence incomplete or at times
conflicting. This is a result of two basic weaknesses:
firstly, that only a relative minority of digital health
interventions has succeeded in moving beyond the
piloting phase; and secondly because implementation
research on scalability has been relatively rare. The
presentation will touch upon i) the definitional issues of
what scale signifies to different users and which
elements could be scaled (project, strategy or technology); ii) the implementation concerns such as the
importance of identifying and aligning the motives of
different constituencies to increase long-term project
viability; what resources and components are required to
grow a project from pilot to large scale; and handling the
rapidly-evolving landscape of ICT, and iii) the evaluation
of effect, the documentation of unintended consequences
and how measurement may change according to the
technology and as the intervention moves to scale. The
logistics of scalability differs substantially across the
broad array of digital products which exist or which are
expected in a near future. For instance, multilingual
eLearning material could swiftly reach across huge
geographic space when it takes advantage of an everexpanding internet penetration. Other approaches, such
as SMS-mediated interventions to improve patient
treatment adherence; video observed therapy; and the
management of surveillance data through online networks, are also relatively well positioned for rapid scaleup. The contents of the Digital Health for the End TB
Strategy : an agenda for action, which was elaborated by
Challenges and opportunities for

implementing e/m-Health for TB in resourcelimited settings: experience from Georgia
Background: Tuberculosis and especially drug resistant

TB are critical public health threats in Georgia. In 2014,
MDR-TB was found in 11.2% of new and 38.6% of
previously treated cases. In response to TB challenges the
Government of Georgia (GoG) has identified effective
governance and monitoring of TB response as strategic
priorities, including development of a well-designed
health management information system (HMIS) which
is critical for effective program monitoring, and evidence-based policy making. Intervention: The USAID
Georgia Tuberculosis Prevention Project has assisted the
National TB program in introducing innovative m/
eHealth tools using mobile phones, tablets, and webbased systems to enable optimization of communication
and the exchange of information, and data among
healthcare professionals and with patients. The e-TB
system is composed of TB case registration, laboratory,
prescription and treatment monitoring components. All
indicators and data collection tools are aligned to the
latest WHO standards. The e-TB module works by
allowing providers to upload data to the system quickly
for each patient with TB and presumptive TB. The
Ministry of Health uses the module for generating casebased financial reports and simplifying transactions via
electronic reporting. The mHealth component is also
functional that facilitates patient counselling using a
tablet based education module (which time- and geo-tags
sessions) and allows electronic recording of patient
attendance to directly observed therapy sessions via SMS.
Results: After a successful pilot, a Government Decree
has mandated electronic recording and reporting of TB
related data from May 1st, 2015 onward. Within the first
quarter of the implementation more than 80% of TB
providers countrywide started using the eTB module on a
daily basis. There were 2947 TB cases currently on
treatment registered and daily DOT attendance reported
for 92% of cases.
Conclusion: Availability of real time data on TB is of
critical importance for timely risk analysis and providing
adequate resources to high-risk TB patients. The highlevel political commitment, donor support and private
sector involvement have been the critical factors ensuring
the implementation of the new HMIS module. The

political will and collaboration of many stakeholders
promoted the systems sustainability by overcoming the
initial barriers such as low computer literacy and lack of
IT infrastructure and limited access to the Internet.
S61
to remain engaged in care such as Diabetes Mellitus and

HIV.
Using geographic information systems and

patient mapping tools to improve follow up
and support for MDR-TB patients in Swaziland
M Calnan,1 T Zanamwe1 1University Research Co. LLC,
Mbabane, Swaziland.
Background: Swaziland is challenged by dual epidemics

of Multi-Drug Resistant TB and HIV. A TB drug
resistance survey conducted in 2010 reported MDR-TB
prevalence of 7.7% in new patients and 33.8% in
previously treated patients, with an estimated 200 cases
occurring annually. Mobile populations, reluctance to
take long and difficult treatment for MDR-TB and
limited country resources makes it difficult to keep
patients engaged in care. Systems to track and follow up
patients have not been effective unless patient can be
traced physically and treatment provided nearer their
homes.
Intervention: Geographical information systems (GIS)
introduced in 2010 have been used to locate MDR-TB
patients in the community and contribute to the national
surveillance. Once registered in care, MDR-TB patients
are visited at home: geo-coordinates of the homestead are
obtained and uploaded. Active cell phone numbers for
the patient and treatment supporter are also obtained
and clinic contacts shared should patient have any
concerns. Once patient misses a clinic appointment, the
follow up tracking system is effected; patients receive a
phone call to enquire the reason for missing the
appointment and a home visit is scheduled. Location of
the home is guided by the coordinates earlier captured.
Patient coordinates are transposed into a map which is
used to: determine location and distribution of the
patients, distance to nearest health facility, monitor for
hotspots or clustering of the disease and plan for
decentralization of MDR-TB treatment services.
Results: Between 2010 and 2015, 1050 MDR-TB
patients have been mapped; majority of the patients
being situated along the Manzini-Mbabane corridor and
the border with South Africa. Four hotspots have been
identified over this period and outbreak investigation
reports reveal that slums, seasonal workers and highly
industrialized areas characterize these hotspots. With
phone call follow up and home visits, the lost to follow
up has declined from 35% in 2012 to 18% in 2015. The
distribution of the patients has informed the decentralization of MDR-TB services to 2 more facilities. Client
satisfaction surveys reveal that patients feel connected to
their care givers and are satisfied with the services
provided.
Conclusion: M-Health has informed the service delivery
for MDR-TB patients in Swaziland and has the potential
to be used for other chronic diseases that require patients
52. Developing a rights-based approach

to prevention and treatment of
tuberculosis
Empowering women and children through a
rights-based approach to TB
P Mahesh 1ALERT-INDIA, Mumbai, India.
A human rights-based approach to TB that ensures

fundamental human entitlements social, economic and
political promote the well-being and empowerment of
women and girls. However, a human rights-based
approach to TB prevention and care, while aiming to
realize equal human rights for women and girls.
Empowerment lies in developing their awareness, as
human rights cannot be given, but must be actively
claimed by those who hold them. Barriers to the
realization of human rights for women and children
increase their vulnerability to TB and restrict their access
to diagnostic and treatment services. The need of the
hour is to address the socioeconomic and other health
determinants. My first-hand experience as a TB patient
of how devastating TB is, not only threatens peoples
bodies, but also economic livelihoods and closest
relationships. The primary barriers women and girls
face is a lack of autonomy, financial and physical
dependency and no or relatively low TB-related literacy,
S62
which impacts negatively and perpetuates stigma.

Interventions to address these barriers include legal
literacy campaigns, inclusion of women with TB in
program development, advocacy, lobbying and implementation of laws and policies that prohibit discrimination. Interventions to expand access to quality TB
prevention and care should aim to establish an enabling
legal and policy environment. Empowerment of women
and girls through right health education about TB and
their rights during intensive case detection enable them to
fight the disease without discrimination. A rights-based
approach to TB will compel governments to take
responsibility to ensure women with TB are empowered
enough to engage in policy making through representation in public institutions and other bodies so that they
can contribute to decisions that affect them and their
communities as this recognizes the need for women and
girls to actively claim rights in order to achieve better
prevention and treatment outcomes. It also promotes the
knowledge and understanding of legal instruments and
institutional machinery to seek protection and redress for
rights violations. And crucially, it demands the; participation of women in standard setting and in formulating,
implementing, and monitoring laws, policies, and
programmes to combat TB by promoting a gender
response approach to TB prevention and care.
Symposium abstracts, Sunday, 6 December
S63
SYMPOSIA: SUNDAY
6 DECEMBER 2015
2
Global Financing Facility in Support of Every Woman Every Child.
Concept Note. September 2014.
53. Fighting TB in the post-2015 era:

paradigm shift in strategies and plans to
put an end to TB
Community-led planning and implementation

of TB initiatives in the post-2015 era
New approaches in financing health and

development: implications for the fight
against TB
P L Osewe 1World Bank Group, Washington, DC, USA
In the last 25 years, countries have made great strides in

reducing maternal and child mortality: since 1990 there
has been a 47% reduction in maternal mortality and a
49% reduction in child mortality.1 Despite these tremendous gains, much more needs to be done. The child
mortality rate in low-income countries is more than 15
times higher than in high-income countries and maternal
mortality is nearly 30 times higher. Yet, there is an
opportunity for a grand convergence in mortality
between low-income countries and the best-performing
middle-income countries. To achieve this convergence,
however, countries will need to significantly ramp up
investments in reproductive maternal, newborn, child
and adolescent health. Such increased investment has the
potential to help prevent a total of 4 million maternal
deaths, 101 million child deaths, and 22 million births
between 2015 and 2030 in 74 of the highest burden
countries.2 On July 13, 2015, the United Nations, the
World Bank Group, and the Governments of Canada,
Norway and the United States joined country and global
health leaders to launch the groundbreaking Global
Financing Facility (GFF) in support of Every Woman
Every Child. The GFF is a key financing platform of the
United Nations Secretary-Generals Global Strategy for
Womens, Childrens and Adolescents Health. It aims to
mobilize private sector resources that, in addition to
public sector resources, help close the $33.3 billion
annual funding gap in the financing of essential
interventions required to improve the health of women,
children and adolescents.3 The GFF is an essential part of
the paradigm shift in development financing, emphasizing the essential but changing role of official development assistance in unlocking domestic resources and
private flows and focusing on results. It also has the
potential to act as a pathfinder for financing the SDGs in
the post-2015 era. This session will discuss the creation
of the GFF, key lessons learnt for innovative financing of
health sector priorities, and the implications for the fight
against TB. The target audience includes policy makers;
TB programme managers and health services providers;
civil society and non-governmental organisation representatives; development partners.
References
1
The Global Strategy for Womens, Childrens and Adolescents Health
3
World Bank Press Release. July 13, 2015. Global Financing Facility
Launched with Billions Already Mobilized to End Maternal and Child
Mortality by 2030
T Abdullaev 1Stop TB Partnership, Geneva, Switzerland
Communities of people affected by TB may be instrumental in various activities, from raising awareness and
counseling to active case finding, psychosocial support
and service provision. The voices of patient community,
their unique experiences are crucial in programme
planning, implementation, monitoring and evaluation.
However, the principle Nothing For Us Without Us, so
common in HIV world, continues to be largely neglected
in the TB community. Patients are still seen by many as
recipients of services rather than a helpful resource. The
change in the attitude towards patients has been
consistently promoted by the Stop TB Partnership. In
its new Global Plan to Stop TB, the Stop TB Partnership
stresses that, If countries are to move away from this
passive approach to one where they seek to find and treat
all the TB existing in their population, a radical shift in
mindset is needed. Innovative changes to health delivery
systems are required to navigate that roadmap. This is
where civil society and communities have a critical role
to play. These stakeholders [. . .] have already been
identified as fundamental in the drive towards better
access to health and universal health coverage. But even
with recognition of the importance of engagement of the
patient community, little can be changed without
consistent support to the process of community mobilization and strengthening. The development of communities of PLHIV often became possible due to the
community mobilization, capacity building, institutional, technical and financial support efforts by UNAIDS,
the Global Fund and other partners. If similar efforts are
taken to strengthen the community of people affected by
TB, they will be able to become important partners in
ending TB epidemic. For the post-2015 era to be a
turning point in the fight against TB, patient community
has to be engaged meaningfully in programme development, implementation and assessment. The Global
Funds funding model envisages involvement of community representatives in the country dialogue. This
participatory approach should be replicated beyond the
Global Fund. It is also important for existing community
networks, such as the Global Coalition of TB Activists,
regional and local TB coalitions, to continue their work
aimed at supporting the development of capacity and
leadership potential of patient activists, networking of
patients and their groups, inclusion of patient representatives in policy and decision making at all levels.
S64
54. Critical updates on the treatment of

TB-HIV co-infected children
Morbidity and mortality in TB-HIV co-infected
children on TB treatment initiated on ART in
the PAANTHER cohort in Asia and West Africa
O Marcy,1 L Borand,1 M Tejiokem,2 F Ateba Ndongo,3
B Nacro,4 P Msellati,5 K Truong Huu,6 V Ung7 1Institut
Pasteur du Cambodge, Epidemiology and Public Health Unit,
Phnom Penh, Cambodia; 2Centre Pasteur du Cameroun,
Service dEpidemiologie
et de Sante Publique, Yaounde,
3
` et Enfant de la Fondation Chantal Biya,
Centre Mere
Yaounde, Cameroon; 4Centre Hospitalier Universitaire Souro
Sanou, Bobo Dioulasso, Burkina Faso; 5UMI 233- U1175
TransVIHMI, IRD , Universite de Montpellier, Montpellier,
France; 6Pediatric Hospital Number One, Ho Chi Minh City,
Viet Nam; 7University of Health Sciences, Phnom Penh,
Cambodia,
Background:: Despite expanded access to antiretroviral

treatment (ART) which has modified HIV-TB coinfection epidemic and had a great impact on TB disease
incidence, TB remains a frequent opportunistic infection
in HIV-infected children with major treatment challenges. Data is scarce on pediatric mortality and morbidity
under joint TB and ART.
Methods: HIV-infected children aged 613 years with a
suspicion of intra-thoracic TB were enrolled in the
PAANTHER cohort after parental informed consent in
pediatric HIV clinics in Burkina Faso, Cambodia,
Cameroon, and Viet Nam. Children underwent TB
diagnostic procedures and were initiated on ART and
TB following national recommendations. We assessed
mortality and morbidity over a follow-up period of 6
months.
Results: A total of 438 children were enrolled in the
cohort, including 171 (39%) on ART at inclusion. In the
267 children without ART at inclusion, 138 (51.7%)
were male, the median age was 6.2 (IQR 2.3-9.1) years,
the median CD4% was 9.0% (IQR 2.0 18.0), and the
median viral load was 5.9 (IQR 5.4 6.4) log10 copies/
ml. Of these 267 children, 139 (52.1%) were initiated on
TB treatment within a median time of 7 days (IQR: 4
10) with 119 (85.6%) subsequently initiating ART, and
128 children were not treated as TB with 97 (76.7%)
initiating ART. Overall, ART was started at a median
time of 0.8 (IQR: 0.5 1.4) months after inclusion and
the median duration of follow-up was 5.9 months
(95%CI: 5.5-6.3). Mortality rates dropped from 92.5
per 100 PY (95%CI: 62.8 - 131.7) to 29.6 (53.4 - 118.8)
after ART introduction: 28 children died before ART
initiation and 24 while on ART, with 16 deaths occurring
in children on TB drugs and ART. Of 119 children
treated for TB who secondarily initiated ART, 12
(10.0%) died, 5 (4.2%) had a paradoxical TB-associated
IRIS reported, and 16 (13.4%) had non-tuberculosis
mycobacteria (NTM) diagnosed secondarily. NTM
infection was not associated with mortality (P 1.0).
In the 97 children without initial TB diagnosis who
initiated ART, 13 (13.4%). were subsequently treated for
TB, which was associated with a higher risk of mortality

(adjusted OR 6.0, 95%CI 1.3-28.1).
Conclusion: Mortality remains high in the first 6 months
following suspicion of TB in ART-nave children, despite
prompt TB treatment and access to ART. Shorter time to
ART initiation, improved
Understanding the contribution of HIV to the

risk of developing tuberculosis in children: a
systematic review and meta-analysis
P Dodd,1 J Seddon2 1University of Sheffield, Sheffield,
2
Imperial College London, London, UK
Following infection, children are at elevated risk of

progression to active tuberculosis (TB), and are more
prone to severe manifestations of disease. In adults,
infection with the human immunodeficiency virus (HIV)
is well established as a potent risk factor for TB progression the protective effect of anti-retroviral therapy
(ART) has been quantified. A similar quantitative
understanding of the effects of HIV and ART on TB
progression in children is lacking. We sought to address
this gap by a systematic review of the literature. Two
classes of study were identified: 1) TB cohorts - children
with TB where HIV prevalence had been measured; 2)
HIV cohorts - children with HIV infection where the
incidence of TB had been measured. Our literature search
identified 3197 potentially relevant records. After
screening, full text articles were assessed for 271 of
these. This resulted in 38 TB cohorts and 31 HIV
cohorts. Data on HIV infection were extracted from TB
cohorts, a subset of which included data on HIV
infection prevalence in children without TB. A random
effects meta-analysis was applied to synthesise an odds
ratio for HIV infection given TB disease, which can be
transformed into an incidence rate ratio for TB disease
given HIV infection. A confirmatory analysis was
performed on all TB cohorts, using UNAIDS HIV
prevalence estimates as controls. We found that, as with
adults, HIV in children is a strong risk factor for
progression to TB disease, and ART strongly protective.
This emphasizes the important of prevention of motherto-child HIV transmission in averting paediatric TB in
high HIV-prevalence settings.
TB screening and IPT delivery in HIV-infected

children: programmatic uptake and challenges
in Mozambique
WC Buck,1 P Gudo,1 H N Raman,1 K Jobarteh,1 F Bata,2
A Couto,2 K Alberto,3 I Manhica3 1Centers for Disease
Control and Prevention, Maputo, Mozambique,
2
Department of HIV and STIs, Mozambique Ministry of
Health, Maputo, Mozambique, 3National TB Control
Program, Mozambique Ministry of Health, Maputo,
Mozambique
Background: TB is a leading cause of morbidity and

mortality in HIV-infected children. Early diagnosis of
both TB and HIV can improve outcomes, and Mozambique guidelines recommend active case-finding through
routine opt-out testing (ROOT) for HIV in children with

TB and routine screening for TB at all HIV consultations.
The burden of TB in HIV-infected children can be
reduced with isoniazid preventive therapy (IPT), and
national guidelines call for its use in newly enrolled
children without active TB.
Methods: The Centers for Disease Control and Prevention (CDC) supports the Mozambique Ministry of
Health HIV and TB programs and PEPFAR-funded
clinical partners report routine data from both programs.
Facilities serving 72.4% of children in HIV care in the
seven CDC-supported provinces use electronic patient
tracking systems (EPTS). This was a retrospective review
of site-level TB-HIV indicators for children 0-14 years
from EPTS sites over three program years (2012-2014) to
assess compliance with guidelines and identify areas for
improvement.
Results: From the TB sector, ROOT for patients of all
ages (disaggregated data not available) improved from
92% to 98% (2012, 2014). From the HIV sector, TB
screening for HIV-infected children at the last visit
increased from 50% to 77% (2012, 2014). This
correlated with an increase in the percentage of HIVinfected children in care receiving TB treatment each year
from 1.4% to 2.9% (2012, 2014). And for newly
enrolled children, the percentage receiving either IPT or
TB treatment increased from 8.4% to 24.0% (2012,
2014), with provincial results in the Figure.
Discussion: Active HIV case-finding in the TB sector is
programmatically strong with ROOT approaching
100% coverage. Results for intensified TB case finding
in the HIV sector are below targets, but improving.
While it is not possible to definitively link the increase in
pediatric TB treatment to more frequent screening, it is
encouraging to see these indicators trend together, and
monitoring them as a pair may be a way to assess the
quality of TB screening. IPT for newly enrolled children
remains a significant challenge with only about 25% of
eligible patients being reached in 2014. Further research
is needed to identify barriers which may include provider
training and comfort ruling out active TB in settings with
limited diagnostics, medication availability, or patient
factors, so that targeted activities can be designed to
accelerate scale-up this evidence-based intervention.
S65
Pharmacokinetics of the first-line TB drugs in

West African children with and without HIV coinfection
A Kwara,1,2 F Gallani,1,2 A Enimil,3,4 A Sarfo,3 A Orstin,3
S Antwi,3,4 H Yang,5 L Wiesner6 1Medicine, Warren Alpert
Medical School of Brown University, Providence, RI,
2
Medicine, The Miriam Hospital, Providence, RI, USA; 3Child
Health, Komfo Anokye Teaching Hospital, Kumasi, 4Child
Health, School of Medical Sciences, Kwame Nkrumah
University of Science and Technology, Kumasi, Ghana;
5
Biostatistics and Computational Biology, University of
Rochester School of Medicine and Dentistry, Rochester, NY,
USA; 6Medicine, University of Cape Town, Cape Town, South
Africa.
Background: Pharmacokinetic data on the first-line

antituberculosis drugs in children is limited. We investigated the pharmacokinetics of these drugs in children
who were mostly treated with World Health Organization (WHO) revised dosages for children.
Methods: Ghanaian children with tuberculosis on firstline therapy for at least 4 weeks had blood samples
collected at pre-dose, 1, 2, 4, and 8 hours post-dose.
Drug concentrations were determined by validated liquid
chromatography mass spectrometry methods and pharmacokinetic parameters calculated using noncompartmental analysis. Factors associated with plasma peak
concentration (Cmax) and the area under the timeconcentration curve 0-8h (AUC0-8hr) of each drug were
examined using univariate and multivariate analysis.
Results: Of the 62 children, 32 (51.6%) were male, 29
(46.8%) were younger than 5 years old and 28 (45.2%)
had HIV coinfection. Three patients had undetectable
pyrazinamide and ethambutol concentrations. The median (IQR) AUC0-8hr for isoniazid was 17.7 (10.2
23.4) lg.hr/mL, rifampin was 26.0 (15.3 36.1) lg.hr/
mL, pyrazinamide was 144.6 (111.5 201.2) lg.hr/mL,
ethambutol was 6.7 (3.8 10.4) lg.hr/mL. Of the
children who received recommended weight-band dosages, 44/51 (86.3%), 46/56 (82.1%), 27/56 (48.2%) and
21/51 (41.2%) achieved target Cmax for isoniazid,
pyrazinamide, ethambutol and rifampin, respectively.
In multivariate analysis, age, Sex, HIV coinfection status,
and drug dosage in mg/kg were associated with the drugs
plasma drug Cmax or AUC0-8hr.
Conclusions: The WHO revised dosages appeared to be
adequate for isoniazid and pyrazinamide, but not for
rifampin or ethambutol in this population. Higher
dosages of rifampin and ethambutol than currently
recommended may be required in most children.
Children with HIV coinfection had lower plasma
exposure of the first-line antituberculosis drugs.
S66
55. Why are DR-TB patients lost by TB

programmes? Patients perspective on
how to address DR-TB treatment
adherence
Why are patients not adherent to DR-TB
treatment? Patients testimony
J Furin 1Global Health and Social Medicine, Harvard Medical
School, Boston, MA, USA,
Discussions on adherence tend to exclude the voices of

individuals with DR-TB and those of survivors. These
leads to adherence strategies that may not be of greatest
benefit to the affected community. This presentation will
focus on actual experiences of persons who have survived
DR-TB or who are currently undergoing therapy. Such
insights and experiences are the only way to truly build
patient-centered care.
Addressing LTFU in high-burden MDR-TB

countries: example from Belarus
A Skrahina,1 G Gurevich,1 D Klimuk,1 L Zhilevich,2
A Tkacheva,3 P De Colombani,4 M Dara,4 D Falzon5
1
Ministry of Health, Republican Scientific and Practical Centre
for Pulmonology and Tuberculosis, Minsk, 2Ministry of
Health, General Directorate of Medical Care, Minsk, 3Ministry
of Health, Directorate of Economic Analysis and
Development of Health, Minsk, Belarus; 4World Health
Organization (WHO), Regional Office for Europe,
Copenhagen, Denmark; 5Global TB Programme, WHO,
Geneva, Switzerland.
Background: Notification of TB new cases and deaths in

Belarus has been progressively falling respectively from
54 to 41 and from 12.1 to 6.8 per 100 000 population
during the last 7 years. Nevertheless, MDR-TB was
found in 32% of the newly-diagnosed and in 76% of the
previously-treated TB cases (2011 nationwide drug
resistance survey). Treatment outcomes of MDR-TB
are far from satisfactory (51% treatment success, 11%
deaths, 24% failures and 10% loss to follow up).
Outpatient care can reduce the risk of cross contaminations in health facilities, the cost of MDR-TB treatment
and disruption of their social and working life. Nevertheless, evidence shows that strengthening outpatient
care is very important to prevent treatment failure and
loss to follow up, particularly among those patients at
higher risk such as the unemployed, homeless, alcohol
and drug users, former prisoners.
Method: A pilot project started in January 2014 in
Mogilev region with WHO support. In this project, the
Ministry of Health provided additional remuneration to
primary health care providers them with for every TB
patient visited from cost savings by reducing the number
of TB hospital beds.
Results: Two operational research studies were conducted during the implementation of the project. One showed
that MDR-TB treatment cost can be reduced by more
than 5200 US$ per patient by decreasing inpatient care
with outpatient care and social support. The other study
showed that MDR-TB patients with social support had

better treatment outcomes compared with those without
it: treatment success 70% vs. 53%; mortality 2% vs. 6%;
treatment failure 27% vs. 40%.
Conclusion: Reallocation of funds from TB hospitals to
primary health care is an essential element for improving
MDR-TB treatment outcomes and effective and sustainable patient support.
Reaching zero LTFU through community-based

care for DR-TB and HIV co-infected patients:
example from South Africa and Lesotho
L Ramangoaela-Molatlhoe 1Health Department, Dr J S
Moroka, Free State, South Africa
Background: South Africa is one of the countries in the

world highly burdened with DR TB with poor overall
treatment outcomes especially notable for high rates of
loss to follow-up. The Free State province has approximately 480 DR TB patients and faces serious challenges
keeping patients in therapy. Free State is located at a
National and International crossroads in Southern Africa
and it is home to a large migrant population. This
population is faced with serious challenges when it comes
to successfully completing therapy. Of note different
rates of lost to follow up occur at different stages of
treatment 6, 12, 18, 24 months of treatment and different
barriers have been noticed and established at these
different stages. Based on the statistics of these lost to
follow up patients the plan is to look retrospectively at
the records to identify these factors and to plan for
interventions and to improve retention in care.
Methodology: Retrospective review of records and
trends of patients seen at the Drug resistant Tuberculosis
units in the Free State as well as information gathered
from TEBA Free State and Lesotho.
Results: Preliminary review of the information shows
that more than 25% are lost to follow up after enrolling
in care and treatment. These would be patients who
interrupt, default, relocate, die, or leave the facility after
initiation on treatment. The review analysis also shows
that these patients face multiple barriers in completing
therapy including, socio-economic, transportation and
adverse events to treatment. Some of these patients are
traced back to treatment, others were not found, others
had died.
Interpretation and conclusion: New strategies are needed. Preliminary data suggests that community based
models of care that bring services closer to the patients
and provide ongoing monitoring and early interventions
to adverse events and other social problems facing the
patient population are urgently needed. These strategies
will also help retain patients in care. Also it is important
to look at strategies that that will assist the households
and implement preventive therapy for those households
exposed.
Study results: LTFU among MDR-TB patients in

the Philippines and what other countries can
learn
E Kurbatova,1 T Tupasi2 1The US Centers for Disease
Control and Prevention (CDC), Atlanta, GA, USA; 2Tropical
Disease Foundation, Inc, Makati City, Philippines.
Background: In the Philippines, the number and proportion of patients lost to follow-up (LTFU) in the
Programmatic Management of Drug-Resistant Tuberculosis (PMDT) has steadily increased each year since
2007. This study was undertaken to identify factors
associated with LTFU from multidrug-resistant (MDR)
tuberculosis (TB) treatment.
Methods: A case-control study included adult (718
years old) patients with rifampicin-resistant TB who
initiated MDR TB treatment between July 1-December
31, 2012 in the Philippines. Cases were patients lost to
follow-up from MDR TB treatment. Controls were
patients who were continuing MDR TB treatment or
had a treatment outcome of cured, completed, or failed.
In-depth interviews were conducted with both cases and
controls, and their medical records were reviewed.
Independent predictors of LTFU were identified using
multivariable logistic regression analysis.
Results: A total of 91 cases and 182 controls were
enrolled. Among cases, the mean age was 4113 years,
66% were males, and 1.7% were HIV positive. Among
controls, the mean age was 3712 years, 57% were
males, and 1.7% were HIV positive. Independent
predictors of LTFU included: patients rating of their
vomiting as more severe following medication doses
(OR1.10 per one point in cumulative score, 95% CI
1.01-1.21), and alcohol abuse (OR2.84, 95% CI 1.395.80), while protective factors included: receiving any
type of assistance from the TB Program (OR0.08, 95%
CI 0.03-0.25), better general TB knowledge (OR0.97
per point in cumulative score, 95% CI 0.95-0.99), and
higher levels of trust in, rapport with, and support from
physicians and nursing staff (OR0.93 per one point in
cumulative score, 95% CI 0.90-0.96). The most common
patient-reported reason for LTFU was adverse drug
reaction (ADR) or the fear of ADR (58%).
Conclusion: Interventions aimed at reducing treatment
LTFU should include prevention and effective management of ADRs, patient education, enhancing providerpatient relationships, and improving program assistance
for patients.
56. The relationship between gender and

specific tobacco products, tobacco
consumption and inequalities
Inequalities, gender and tobacco consumption
A Amos 1Usher Institute of Population Health Sciences and
Informatics, University of Edinburgh, Edinburgh, UK.
Smoking is the leading preventable cause of premature

mortality and socioeconomic inequalities in health in
women in many high income countries (HICs) that are at
S67
Stage 4 of the tobacco epidemic. This includes countries

such as the US, Canada and several countries in the
European Union including the UK. While female
smoking prevalence in these countries is declining, the
social gradient in smoking is not. Indeed, female smoking
is increasingly concentrated among disadvantaged
groups. In HICs girls from disadvantaged backgrounds
are more likely to start smoking and less likely to quit
when adults. In low and middle income countries
(LMICs) the pattern is more complex, but there is also
evidence that smoking by girls and women is linked to
disadvantage. This presentation will give an overview of
inequalities in smoking in girls and women around the
world. It will then draw on the findings of three recent
systematic reviews on social inequalities and smoking to
identify those policies which have a positive equity
impact ie reduce inequalities in smoking. The reviews
covered: population and individual level interventions/
policies on youth smoking prevention; population level
interventions/policies on adult smoking cessation; and
individual level adult cessation support interventions.
The presentation will highlight the lack of studies which
have assessed the equity impact of tobacco control
policies, and the even greater lack of studies that have
looked at their equity impact from a gender perspective.
Finally, it will consider the implications of these findings
for developing effective gender-sensitive equity-orientated tobacco control strategies.
Use of water-pipe and oral tobacco among

women
D Aslan 1Public Health, Hacettepe University, Faculty of
Medicine, Ankara, Turkey
Currently there are 200 million women using tobacco

products globally and there is an increasing trend in use
of non-cigarette tobacco products among women due to
the recent marketing strategies of the tobacco industry.
Culture, geography, gender-based roles, price policies,
advertisements, etc are among influencing factors.
Scientific reports highlight different tactics used by
tobacco industry on women and younger generations
since decades. Without gender sensitive approaches, it
may not be possible to overcome the problem as the
problem includes many different dynamics. World
Health Organization urges on the rising tobacco
consumption trend in women and alerted its partner
countries twice via 31st of May, World No Tobacco Day
messages since 1988 (Women and tobacco: the female
smoker: at added risk;1989 and Gender and tobacco
with an emphasis on marketing to women; 2010). Since
2005, with the strong contribution of the World Health
Organization-Framework Convention on Tobacco Control (WHO-FCTC), comprehensive tobacco control
policies started to be implemented all over the world.
Although WHO-FCTC addresses all tobacco products,
all partner countries expectedly starts the struggle with
cigarette smoking as it is the most frequently used
tobacco products in the country. The industry saves time
to extend its tactics on its new products while the state
S68
bodies organizes additional legal regulations/inspections

on other tobacco products like water pipe, oral tobacco,
etc. Although there are a number of non-cigarette
tobacco products, this speechs content is limited to
water pipe and oral tobacco among women. Water pipe
is a frequently used tobacco product in the Middle East
and Eastern Mediterranean countries. Many health risks
of water pipe use like cancer, infectious diseases,
respiratory threats are not well known influences in the
community. Oral tobacco use is another type of noncigarette tobacco consumption like chewing, etc. Oral
squamous cell carcinoma is the main threat of such
consumption which is also not a well-known effect in the
society. Such information lacking may mislead the people
to use non-cigarette tobacco products. In order to
achieve the goal of TOBACCO FREE WORLD, struggle
methods should be strengthened on all tobacco products
without any distinction based on the WHO-FCTC
including the M-POWER strategies in gender sensitive
perspective.
E-cigarettes: tobacco industry and non-tobacco

industry strategies
M De Andrade 1School of Health in Social Science,
University of Edinburgh, Edinburgh, UK
The need to understand tobacco industry (TI) tactics

with regard to e-cigarettes is well documented. To date,
however, our understanding of the TIs business strategies in this area has largely been based on speculation
informed by the TIs deceptive past, ad hoc inquiries and
isolated examples reported in commentaries, editorials
and opinion pieces. This work does suggest cause for
concern. All the tobacco multinational companies
(MNCs) have now heavily invested in the e-cigarette
market raising uneasiness for tobacco control that the
focus for these companies will be on their core tobacco
business, with e-cigarettes being used primarily to bolster
this. There are clear indicators, for example, that for
industry this is about business not public health.
Similarly, the TIs low tar lie and failed attempts at
promoting a safe cigarette have added weight to
concerns that e-cigarettes may be industrys latest Trojan
horse. E-cigarettes, however, are distinctly different
from traditional cigarettes and could significantly
improve public health because of their extensive appeal
as smoking cessation devices. In another echo of past
misdemeanours, the promotion of these products as
lifestyle accessories using a combination of evocative
advertising on television and in traditional, digital and
social media outlets, sponsorship and celebrity endorsements has an appeal to young people and women in
particular. E-cigarette marketing has also been identified
as a source of misleading product information for
consumers. This presentation will systematically examine the business strategies being deployed in the ecigarette market by both tobacco MNCs and independents. Specifically, it will compare and contrast their
respective: (i) business approaches. (ii) targeting strategies. (iii) positions on harm reduction and health claims.
(iv) marketing plans (including product development,

pricing, distribution, packaging and promotion). The
implications for tobacco control will then be discussed,
with particular reference to existing regulation and
whether or not this needs to be strengthened to protect
young people and non-smokers. Specifically, it will
consider potential threats for inequalities and women.
How to counteract the changing industry

tactics
M Haglund 1ThinkTank Tobaksfakta, Stockholm, Sweden
Tobacco use among women is a 21st century global

epidemic. The smoking trends of women and men change
over time and across populations as do gender norms and
roles. Today men is moving out of the tobacco pandemic
and women moving in, a trend that has been clearly
understood by the tobacco industry. Currently 250
million women are daily smokers. In industrialized
countries, 22% women are smokers and in developing
countries, 12%.The number of female tobacco users is
predicted to continue to rise over the next several decades
as a result of increased prevalence, and population
growth in particular in low and middle income countries.
Although the roots of tobacco uptake for women and
girls often include cultural, psychosocial, and socioeconomic factors, including body image and peer pressure
the most important factor influencing women to start
smoking is the tobacco industrys marketing campaigns
aggressively target women. Since the 1920s tobacco
companies have been imposing women with their
seducing messages in country after country with
increased smoking among women as the result. This
presentation will discuss different actions on how to
counteract the tobacco epidemic among women incl. the
tobacco industrys tactics e.g. the need to implement
comprehensive bans on tobacco promotion and marketing and to use a gender approach whereby sex-and agedifferentiated data are systematically taking into consideration. It is also vital to acknowledge womens tobacco
use as a major health problem and to build international
as well as national consensus around this issue.
57. The role of cost-effectiveness in

achieving the End TB Strategy
Overview of TB control costs and financing
issues
C Owunna,1 D Collins2 1SIAPS Program, Management
Sciences for Health, Arlington, VA, 2Management Sciences
for Health, Medford, MA, USA.
An analysis of the TB treatment targets needed to achieve

the goal of eradicating TB in many high-burden countries
indicates that significant increases in resources will be
needed to achieve those targets. Much of these additional
resources are needed urgently over the next few years to
get the incidence rate to fall sufficiently to eradicate TB
by 2035. This resources will need to come from increased
domestic spending since donor funding is likely to

decrease, especially in countries whose economies are
improving, albeit slowly. This increased domestic funding will have to come from government budgets and from
national health insurance. How much of the government
financing should and can come from national, provincial
or district budgets is open to question, but that should
probably be based on the role of each level of government
in TB control. In terms of national health insurance,
there is already pressure on premiums, on the expected
contents of the benefits coverage, and affordability will
remain an issue. What is clear is that substantial
evidence-based advocacy will be needed to persuade the
government and the national health insurance scheme to
increase and sustain the investment in TB control to
achieve the goal of TB eradication. What is also clear is
that cost-effectiveness and efficiency will be crucial to
maximize detection, especially as regards reaching the
poor and vulnerable. Improved screening will be
essential to ensure that no opportunities are missed to
identify cases. This presentation draws on the analysis
and lessons learned in Indonesia, where considerable
assessment has been conducted on the cost of scaling-up
TB control services, on financing options and on costeffectiveness.
The cost-effectiveness of different screening

and diagnostic methods: lessons from several
countries
D Dowdy 1John Hopkins University School of Medicine,
Baltimore, MD, USA
The cost-effectiveness of diagnostic testing for tuberculosis (TB) depends on a number of factors beyond just the
cost of the diagnostic test. These factors include the
intended role of the test (symptomatic testing, systematic
screening, drug susceptibility testing, etc.), the population in which the test is employed (probability of TB,
HIV prevalence, drug resistance profiles), existing
diagnostic and empiric treatment patterns, and integration of the test into current healthcare systems. In areas
of high prevalence of HIV and drug-resistant TB, cost
and cost-effectiveness considerations of TB diagnosis are
overshadowed by the cost of treating these conditions. A
determination of cost-effectiveness may not equate to
affordability or an appropriate decision to scale up a
diagnostic test. Major areas of data uncertainty also exist
- including the timing of TB transmission and the
accuracy of diagnostic tests across settings and study
designs. While most published cost-effectiveness analyses
of TB screening and diagnosis have found those activities
to be highly cost-effective, the assumptions underlying
those analyses may not reflect actual implementation.
Case studies from Africa and Southeast Asia are
illustrative of these principles. For example, Xpert
MTB/RIF was found to be highly cost-effective in
southern Africa, but early evaluations considered levels
of empiric treatment to be low, and did not account for
costs of antiretroviral therapy. Incorporating costs of
HIV care and the high probability of empiric therapy (as
S69
later demonstrated in randomized trials) dramatically

attenuates the cost-effectiveness of Xpert. Evidence from
Uganda and Tanzania also suggests that the costeffectiveness of Xpert as implemented in the field may
be substantially lower than estimates from idealized
settings. In Southeast Asia, the cost-effectiveness of novel
diagnostic tools has been shown to reflect tradeoffs
between accuracy and ability to implement each test, in
both diagnostic and screening settings. These case studies
demonstrate that the cost-effectiveness of TB diagnosis
represents a complex interplay between diagnostic test
characteristics, healthcare systems, patients, providers,
ancillary conditions (e.g., HIV), and implementation.
Economic evaluation of new TB diagnostics in

South Africa: an analysis of differences found
in early trial and real world cost-effectiveness
N Foster,1 S Cleary,1 L Cunnama,1 A Vassall,2
G Churchyard3 1University of Cape Town, Cape Town, South
Africa; 2London School of Hygiene and Tropical Medicine,
London, UK; 3Aurum Institute, Johannesburg, South Africa
Background: Given limited resources, policy makers

have to make decisions regarding where investments in
health systems are likely to have the greatest impact on
reducing the burden of TB. Since social protection
against the cost of illness is a key policy objective of
the post-2015 Global strategy for TB, the costs incurred
by patients and their households is an important
consideration in targeting TB interventions. However,
the evidence base on these costs is limited, as is their
application to prioritisation within TB control.
Methods: A longitudinal patient cost survey was
conducted, sampling patients from 10 health facilities
in four provinces in South Africa. A total of 351 patients
with symptoms suggestive of TB were interviewed at
time of receiving a TB diagnostic and again 6 months
later. A separate cohort of 168 TB patients was
interviewed at 2 and 5 months on treatment.
Results and discussion: Those who started TB treatment
reported a mean monthly individual income (prior to
symptoms) of $76.05 per month. During the TB episode,
the greatest cost was incurred in the pre-treatment period
($85.72 out of a total of $209.97 per TB episode). Direct
out-of-pocket costs were approximately 12% of annual
individual income. The primary economic loss was
income loss prior to starting treatment, which accounted
for almost half pre-treatment costs. Interventions focussed on shortening the pre-treatment period and
relieving the cost of illness, such as increased case finding
efforts, have a relatively high potential in terms of
averting the economic burden of TB on patients and
households. These interventions should be considered as
part of the package, alongside social support to those
diagnosed with TB, to address the poverty impact of TB
in South Africa.
S70
Modeling the economic impact of TB treatment

interruption
D Collins,1 E Rutta1 1Management Sciences for Health,
Medford, MA, USA
The interruption of treatment by TB patients has a

significant impact on their health and recovery. Whether
interruption is caused by patient challenges or by stockouts of accessible medicines, it results in increased
transmission and can result in drug-resistance and can
cause additional costs for health providers and governments who are often already underfunded. Interruption
can also increase patient costs which can further
contribute to loss to follow-up. This presentation
describes a costing tool that has been developed to
calculate the costs of treatment interruption and the
savings from preventing them. The results are intended
for advocating with managers, policy makers and donors
for greater investment in preventing treatment interruption and for budgeting for the resources needed.
The economic costs of TB medicine stock-outs

in the Philippines
H Lam,1 T Sy,1 A Rivera,1 K Cheng,1 P Chua,1 M Lamayra,1
E Rutta,2 D Collins3 1Institute of Health Policy and
Development Studies, National Institute of Health, University
of the Philippines, Manila, Philippines; 2Systems for Improved
Access for Pharmaceuticals and Services, Management
Sciences for Health, Arlington, VA, 3Health Programs Group,
Management sciences for Health, Medford, MA, USA
Having a reliable supply of TB medicines is necessary

element of a TB control program. Stock-outs of
medicines result in delays and interruptions in treatment
which can impact outcomes, increase transmission and
can result in drug-resistance. They can also increase
patient costs which can further contribute to loss to
follow-up. Stock-outs of TB drugs is an important reason
for treatment interruption but is often not sufficiently
taken into account. More attention is generally paid to
patient issues (demand) than the availability of medicines
(supply issues). There are many sunk costs in TB control
that are wasted if there are no medicines for the patients.
Case finding, diagnosis, contact tracing, dedicated TB
staff and facilities are all worthless if there are no
medicines. Without medicines the patient will continue
to infect other people and will not be cured. With
intermitted supplies of medicines the results may be the
same but there is also the risk of developing MDR-TB or
XDR-TB which is more costly to treat and more difficult
to cure. The SIAPS project in The Philippines, through a
national survey of 200 facilities with the IHPDS,
determined that 27.1% of sampled rural health units
and 54.2% of sampled warehouses experienced stockouts with a mean durations of between 0 and 95 days for
drugs and between 0 and 181 days for laboratory
supplies. An intervention mix that would cost USD 13
per day of reduced stockout was identified. This
presentation describes the steps that were followed to
reduce stock-outs and the costs of those steps. It also
describes that economic costs that were saved for
providers, patients and society of reducing stock-outs

and makes a compelling argument for further investment
in improving supply chain.
58. Universal health coverage and

insurance: how TB fits into the health
financing picture
A framework for analysing TB and insurance,
and assigning TB financing to either insurance
or government budgets
D Collins,1 F Hafidz2 1Management Sciences for Health,
Medford, MA, USA; 2Gadja Madah University, Yogyakarta,
Indonesia
In order to end the TB epidemic by 2035 in the highburden countries the vast majority of TB cases will have
to be detected and successfully treated by 2019 so that
incidence rates are to start to fall in time. This will be a
major challenge, especially for Multi-Drug Resistant
Tuberculosis (MDR-TB) where the numbers of cases
detected and treated are often low. Increased case
detection and treatment will require significant additional resources and with reducing donor funding in some
countries, more and more of the costs will have to be
funded from domestic sources. To ensure equitable
access to services, the bulk of the financing will have to
come from government budgets and social health
insurance. And in some countries government funding
can come from local, provincial or national levels. One of
the challenges for a country where the financing comes
from these different sources is to ensure that there is
adequate funding from each source and that none of the
inter-related areas of a TB program are neglected. In
Indonesia, a financing framework was developed based
on the expected role of each level of government in TB
control as well as the role and likely coverage under the
social health insurance system. Detailed costs were
projected for each program element (e.g., detection, case
finding, diagnosis and treatment) and for each resource
(e.g., salaries, laboratory reagents and medicines). Using
a new costing and financing model, the projected costs
were then allocated to the various funding sources. This
provided a valuable basis for advocacy to the government and the insurance agency and for developing cost
and financing expectations for both public and private
sector providers. This presentation will describe the
framework used to project the costs of the TB program in
Indonesia and the assignation of roles and funding to the
three levels of government and to social health insurance.
Challenges and solutions for reaching more TB

providers with insurance schemes in the
Philippines
AMC Garfin 1Department of Health, Manila, Philippines
Social Health Insurance in the Philippines started from

the Governments desire to provide its people with access
to effective medical care services at an affordable price.

This started in 1972 when the law creating the Medicare
Program was enacted. That was the countrys first attempt at universal social health insurance. After twentyfive years, the program covered only the employed sector
and paid for a very small fraction of the countrys health
expenditures and this paved the way for PhilHealth. In
1995, Republic Act 7875 was signed and established the
National Health Insurance Program. With this, PhilHealth was created to administer the program with the
mandate of providing universal social health insurance
coverage. For tuberculosis, an out-patient benefit package was formulated as a response to the need to provide
accessible and quality health care services to its members
and their qualified dependents who are suffering form
TB. For the TB program this is a mechanism to ensure
quality of TB services and an additional financing source.
The package covers all drug susceptible TB patients. All
certified and accredited health care providers can
participate. A reimbursement of about 90U$ will be
provided to the facility for each TB patient that has
completed treatment. A referring physician will also be
provided with about 23U$ for each referred patient. This
initiative started in 2003 and there are several challenges
and actions to be taken to improve the system. Here are
some of the challenges and actions to be made. 1. Low
utilization rate due to non-disclosure of patients of
PhilHealth membership. Enhanced health care institution portal will be provided to address this issue. 2.
Problems in the process of claims and reimbursement for
file cases. A trust fund will be crated and a circular will be
provided to include the allocation of the package as a
guide to the facility. 3. Low number of certified and
accredited DOTS facilities. Funds will be provided to for
facility improvement and review of the certification and
accreditation process will be done. 4. Limited number of
staff to do certification due to the rationalization at the
regional level. Trainings for regional staff will be done
supported by partners.
Maintaining NTP functions in the context of

national health insurance
S Jittimanee,1 S Somthong,1 C Nampeng,1 J Vorasingha,1
C Namwat1 1Disease Control, MOPH, National TB Program,
Bangkok, Thailand
The burden of tuberculosis (TB) (mortality, prevalence

and incidence) in Thailand is gradually reducing,
according to the World Health Organization (WHO).
An important contribution to this reduction is likely to be
free access to TB diagnosis, treatment and care due to the
vertical program before 2002 and the decentralized
program through the universal health care scheme since
2002 and the high standard of clinical care available in
hospitals, both public and private. The health system in
Thailand is the link with three insurance schemes, which
cover 99% of the Thai population. Most important is the
Universal Coverage Scheme (UCS), established in 2001
and run through the National Health Security Office
(NHSO). The others are the Civil Servants Medical
S71
Benefits Scheme (CSMBS) and the Social Security Scheme

(SSS), which go back to the 1980s and 1990s, respectively. However, these insurance packages are not
harmonized, leading to some inequity a problem that
is recognized and being addressed. The CSMBS and UCS
are financed by general taxation whereas the SSS is
financed by a payroll tax with a tripartite contribution
shared by employer, employee and the government, with
1.5 % of the salary as premium. These schemes cover the
costs of basic care for almost all the population, and have
nearly eliminated the risk of catastrophic health expenditure for families. Bureau of TB(BTB), the center unit of
NTP is responsible for policy development, supervision,
monitoring and evaluation, training, surveillance, laboratory reference issues and research. The BTB has only a
technical support function towards the NHSO and provincial health offices. Hospital practice is driven by the
reimbursement amount agreed upon with the insurance
schemes and, as a consequence, adherence to national
technical standards, such as those set by the BTB, is
sometimes compromised. However, the NHSO is open to
discussion on the package of care that should be
remunerated. Private providers and academic centres
are not accountable to the BTB and generally do not
report to them. This limits the coordination between
program staff to provide support for treatment practice,
recording, reporting, supervision and monitoring. However there are multiple health care providers. The private
sector is large and not fully regulated. Migrants and
prisoners are challenging groups when it comes to
providing health services. Both these groups are particularly vulnerable to TB.
59. HIV-TB Late-Breaker Session

Sulfamethoxazole/trimethoprim/isoniazid/
pyridoxine scored tablets are bioequivalent to
individual products and are acceptable to
patients with advanced HIV infection in the
REALITY trial
D M Gibb,1 M Bwakura-Dangarembizi,2 D Abhyankar,3
A J Szubert,1 C Agutu,4 A Lugemwa,5 J Abach,6
M Kemigisa,7 J Mallewa,8 A Siika,9 A Mukuye,10
V Mehta,3 G Malhotra,3 E Nambi,9 M Thomason,1
S Pett,1 J Berkley,4 B Meli,9 C Kityo,10 K Nathoo,2
V Musiime,11
A S Walker,1 J Gogtay,3 REALITY Trial Team 1MRC Clinical
Trials Unit at University College London, London, UK;
2
University of Zimbabwe, Harare, Zimbabwe; 3Cipla Ltd.,
Mumbai, India; 4KEMRI Wellcome Trust Research
Programme, Kilifi, Kenya; 5Joint Clinical Research Centre,
Mbarara, 6Joint Clinical Research Centre, Gulu, 7Joint Clinical
Research Centre, Fort Portal, Uganda; 8University of Malawi,
Blantyre, Malawi; 9Moi University Clinical Research Centre,
Eldoret, Kenya; 10Joint Clinical Research Centre, Mbale,
11
Joint Clinical Research Centre, Kampala, and Makerere
University College of Health Sciences, Kampala, Uganda.
Background: Human immunodeficiency virus (HIV)

infected patients initiating antiretroviral therapy (ART)
S72
with low CD4 counts particularly benefit from opportunistic infection (OI) and anti-tuberculosis prophylaxis.
However, individual drug formulations of isoniazid and
cotrimoxazole add to the pill burden of ART and
adherence difficulties early in treatment when mortality
is highest. Stockouts are also frequent in low-resource
settings.
Methods: Sulfamethoxazole/trimethoprim/isoniazid/pyridoxine (800/160/300/25 mg) fixed-dose combination
(FDC) scored tablets, made by Cipla, India, were
evaluated for bioequivalence vs. separate tablets of
SeptrinwForte (sulfamethoxazole 800 mg/trimethoprim
160 mg [GlaxoWellcome]) and isoniazid 300 mg
(Sandoz) in an open-label, randomized, single-dose,
two-treatment, two-period, 26-sample crossover pharmacokinetic study. Acceptability and adherence questionnaire data were collected in the ongoing, 2x2x2
factorial, REALITY randomised trial evaluating 12-week
enhanced OI prophylaxis, 4-drug ART and enhanced
nutrition in 1805 African adults/children (75 years)
with CD4 ,100 cells/mm3. Within-individual data on
FDC (weeks 12-24) vs. cotrimoxazole alone (weeks 0
12) were compared in 319 patients; and in weeks 012
among those receiving FDCfluconazole (for prophylaxis against cryptococcal disease) (n 543) vs. cotrimoxazole (n 604).
Results: Among 28 fasting healthy subjects (18 male; 10
female; mean (range) age 31 (1845) years; BMI 25.1
2.9), geometric mean test-to-reference ratios for sulfamethoxazole (AUC 99.8%, 90%CI 96.2103.5 and
C max 103.2%, 90%CI 99.5107.0), trimethoprim
(97.2%, 90%CI 93.7100.9; 98.2%, 90%CI 93.4
103.3) and isoniazid 103.8% (90%CI 99.5108.3;
104.3%, 90%CI 95.1114.4) were well within required
80125% range. In REALITY, no within-individual
differences in acceptability were reported between FDC
(1224 weeks) vs. cotrimoxazole (012 weeks): 99.7%
vs 99.4% reported none/not much interference with
everyday life and 95.6% vs. 96.6% reported that drugs
were very easy/easy to take. Comparing groups over 0
12 weeks gave similar results: 500/543 (92.1%) vs. 565/
604 (93.5%) (P 0.8) reported taking medication was
very easy/easy; 4.0% vs. 3.9% reported missing 71
dose.
Conclusions: Sulfamethoxazole/trimethoprim/isoniazid/
pyridoxine scored FDC tablets are bioequivalent to
individual drugs; data have been submitted for World
Health Organization prequalification. They are acceptable, reduce pill burden and could improve adherence for
adults/children 75 years, in addition to simplifying drug
distribution for HIV programmes.
Household point-of-care CD4 testing and IPT

initiation in a TB contact tracing programme:
feasibility, uptake and linkage to care
L Page-Shipp,1 J J Lewis,2 V Kavindhran,1,3 S Senoge,1
E Zishiri,1 F Popane,1 V N Chihota,1,3 D Clark,1
G J Churchyard,1-3 S Charalambous1,3 1Aurum Institute,
Johannesburg, South Africa; 2Department of Infectious
Disease Epidemiology, London School of Hygiene & Tropical
Medicine, London, UK; 3School of Public Health, University of
Witwatersrand, Johannesburg, South Africa.
Background: Household TB contact tracing provides an

opportunity for HIV and Point-of-Care CD4 (PoCCD4)
testing and home-based IPT initiation. Feasibility, HIV
uptake, linkage-to-care, IPT initiation and completion
were assessed.
Methods: A pragmatic cluster-randomised trial among
household contacts of index TB patients in two districts
in South Africa. Households were randomised into three
arms: 1) Standard of Care (SoC): TB symptom screening,
TB testing if symptomatic and HCT offer; 2) SoC plus
PoCCD4 for newly diagnosed HIV (PoCCD4 arm); 3)
SoC, PoCCD4 and household-initiated IPT (HH-IPT
arm) if CD4 .350 and clinically eligible. Follow-up was
at 10 weeks and 6 months to determine linkage-to-care
(SoC/PoCCD4 arms) and IPT initiation and continuation
(PoCCD4/ HH-IPT).
Results: 2243 index TB patients enrolled between
November 2012 and August 2014; 63.2% female,
median age 30 years. HCT uptake was 34.7% in SOC;
increasing to 40.2% in the PoCCD4 arm (RR 1.16;
95%CI: 0.99, 1.36) and to 39.9% in the HH-IPT arm
(RR 1.15; 95%CI 0.99, 1.35), (P 0.06 and 0.08
respectively). 98 new HIV were evaluated for linkageto-care within 3 months. Linkage-to-care was recorded
for 30.8% in the SoC arm and 42.1% in the PoCCD4
arm (RR1.37; 95%CI 0.68, 2.76, P 0.38). In the
PoCCD4 arm, 3/20 (15.0%) reported initiating IPT. In
the HH-IPT arm, 20/21 (95.2%) initiated IPT in the
household (RR 8.20; 95%CI 1.98, 33.95, P 0.004).
Receipt of at least four months of IPT was reported for 3/
20 (15.0%) in the PoCCD4 arm, and 12/21 contacts
(57.1%) in the PoCCD4HH-IPT arm (P 0.009).
Conclusions: Household PoCCD4 testing and IPT
initiation is feasible. PoCCD4 increased HCT uptake
but did not significantly increase linkage-to-care. IPT
initiation and completion was significantly increased in
the household intervention but numbers were small due
to low numbers of new HIV household contacts.
Although feasible, these interventions had low impact
due to the low uptake of HIV testing in households.
Cell phone ringtone intervention to improve

adherence to antiretroviral therapy in
Yaounde, Cameroon: a randomized controlled
trial
E W Pefura-Yone,1,2 Z Nzina-Toupendi,3 A D Balkissou,1,2
A P Kengne,4 E Afane-Ze1,2 1Faculty of Medicine and
Biomedical Sciences, University of Yaounde 1, Yaounde,
2
Yaounde Jamot Hospital, Yaounde, 3Institut Superieur

de
Technologie Medicale,
Yaounde, Cameroon; 4South Africa
Medical Research Council, Cape Town, South Africa.
Background: Despite continuing education of people

living with HIV infection (PLWHIV), up to one third still
have poor adherence to antiretroviral therapy (ART). We
assessed the effect of cell phone-based reminders (ringtones) on adherence to ART among PLWHIV receiving
care at the Approved Treatment Centre for HIV infection
(ATC) of Yaounde Jamot Hospital (YJH), Cameroon.
Methods: This single-center, single-blinded randomized
controlled trial was conducted from December 2014 to
May 2015 (6 months). PLWHIV aged .18 years were
randomly assigned to intervention (ringtones group) or
control group (routine care group) on a 1:1 allocation
ratio. Adherence to ART was assessed using the Center
Adherence Support for Evaluation self-reported adherence index (CASE), with an index score .10 indicating
high adherence. The primary endpoint was the proportion of patients achieving high adherence to ART in both
groups after 4 months of follow-up (M4). Secondary
endpoints included adherence rate at month 2 and
retention in care.
Results: A total of 199 participants were allocated to the
ringtones group and 201 to the usual care group. The
positive history of tuberculosis was found in 48.2% and
37.3% of the patients in the ringtones and control groups
respectively. The adherence rate at M4 was 68.8% in the
ringtones group and 54.2% in the control group,
corresponding to a relative 27% (95%CI 8-49%; P
0.003) higher adherence in the intervention group.
Adherence rate to ART after 2 months of follow-up
intervention vs. control group was 80.4% vs. 54.7%, P
, 0.001. No statistical significant difference was found
between the two groups regarding the retention in care
during follow-up.
Conclusions: Programming ringtone alarms on mobile
phones of HIV-infected patients improved the short and
medium terms adherence to ART. Extended follow-up of
a larger cohort is needed to assess the effects on the
outcomes of care.
S73
Adherence to antiretroviral therapy is

associated with retention in care for
extensively drug-resistant tuberculosis HIV coinfected patients undergoing treatment in
South Africa
M R ODonnell,1-3 A Daftary,3,4 K R Amico,5 A Wolf,1
G Friedland,6 D Bangsberg,7 N Padayatchi3 1Division of
Pulmonary, Allergy, and Critical Care Medicine, Columbia
University Medical Center, New York, NY, 2Department of
Epidemiology, Mailman School of Public Health, Columbia
University Medical Center, New York, NY, USA; 3Centre for
AIDS Programme of Research in South Africa (CAPRISA),
Durban, South Africa; 4Dalla Lana School of Public Health,
University of Toronto, Toronto, ON, Canada; 5School of Public
Health, University of Michigan, Ann Arbor, MI, 6Yale
University School of Medicine, New Haven, CT, 7Harvard
Medical School, Boston, MA, USA.
Background: There are critical gaps in our understanding

of the linkage between medication adherence and
retention in care for patients with drug-resistant tuberculosis (TB) and HIV, which may contribute to low
treatment success.
Methods: XDR-TB patients in KwaZulu-Natal, South
Africa, were prospectively enrolled at start of treatment
and followed monthly for .2 years (PROX study).
Adherence was assessed monthly by study staff by 7-day
recall. Missing patients were followed up in person or by
telephone. Optimal adherence was defined as self-report
of taking all pills in the previous 7 days; missing any pills
was defined as suboptimal adherence. Loss of retention
in care was defined as missing .2 consecutive monthly
clinic visits without known transfer or death.
Results: From October 2009 through July 2011, 104
XDR-TB patients (79% HIV co-infected, 82.5% on
antiretroviral therapy [ART]) were enrolled. At 24
months 16% of patients defaulted (loss of retention in
care), while 41% died, 26% cured, 6% completed
treatment and 11% failed treatment. HIV co-infection
did not affect retention (Panel A) nor did ART (data not
shown). For XDR-TB HIV patients on ART (n 64),
cumulative optimal adherence to ART at 6 months
predicted higher levels of retention in care at 24 months
compared to suboptimal ART adherence (86% vs. 57%,
P , 0.01) (Panel B). This difference was significant on
univariate and on multivariate analysis (OR 6.08 1.05
43.92, P 0.04).
Conclusion: Although default rates were high, XDR-TB
HIV patients reporting optimal ART adherence were
significantly more likely to be retained in care through 24
months. This finding may be due to enhanced ART
counseling, behavioral links between optimal adherence
and retention, or other factors. We speculate that
interventions to improve adherence may also improve
retention in care. Further studies to understand and
improve adherence and retention in care for drugresistant TB-HIV are critically needed.
S74
Acquired drug resistance in rifampicin

susceptible pulmonary tuberculosis a
programmatic setting with a high prevalence
of HIV co-infection
N Rockwood,1,2 F Sirgel,3 G Meintjes,1,2,4
R J Wilkinson,1,2,5 1Department of Medicine, Imperial
College, London, UK; 2Clinical Infectious Diseases Research
Initiative, Institute of Infectious Disease and Molecular
Medicine, University of Cape Town, Cape Town, 3DST/NRF
Centre of Excellence for Biomedical Tuberculosis Research/
SAMRC Centre for Tuberculosis Research, Division of
Molecular Biology and Human Genetics, Faculty of Health
Sciences, Stellenbosch University, Tygerberg, 4Department of
Medicine, University of Cape Town, Cape Town, South
Africa; 5MRC Francis Crick Institute, Mill Hill Laboratory,
London, UK.
Background: The incidence of acquired drug resistance

(ADR) in patients treated for rifampicin susceptible (RS)
tuberculosis (TB) has been reported to be as high as 510%. We prospectively studied the incidence and risk
factors of ADR in Khayelitsha township, Western Cape,
South Africa, where HIV co-infection is present in 5070% of TB patients.
Methods: Consecutive cases of Xpert MTB/RIF confirmed RS pulmonary TB were recruited at diagnosis and
followed up at 2 and 5 months when sputum TB cultures
(MGIT) and adherence assessment (urine acetylisoniazid;
pill counts; self-report) were done. Treatment followed
local TB program guidelines. MTBDRplus was performed on positive cultures at 2 and 5 months to screen
for ADR, with spoligotyping carried out on suspected
cases of ADR. We retrospectively tested for isoniazid
monoresistance (IMR) in stored baseline isolates. Minimum inhibitory concentrations (MIC) were determined
in a sub-cohort at baseline (n 109) and 2 months (n
25) using BACTEC MGIT 960 system.
Results: 308 participants (33% retreatment) were
recruited and followed up between March 2013 and
December 2014. 62% were HIV co-infected [median
CD4 count 233cells/mm3, (IQR 101-389), 39% on
antiretroviral therapy at baseline]. Poor adherence was
documented in 8-12% of patients. 32% of male
participants were ex-prisoners. 2 cases of ADR were
detected at 5 months, incidence 0.7% (95%CI 0.09 to
2.5%). MIC at baseline was similar when analysed
according to HIV, retreatment and culture converter
status. In paired isolates, there was no significant change
in MIC by 2 months for any patient for rifampicin/
isoniazid/pyrazinamide (defined as x4 change). 6%
(95%CI 3.6-9.6%) of baseline isolates had IMR (13/17
had low level IMR-inhA mutation or MIC60.4lg/ml).
Despite being on the standard first-line regimen, no
participants with IMR amplified resistance.
Conclusions: After roll out of Xpert in this wellfunctioning programmatic setting, ADR arising in
baseline RS TB is rare (0.7%) with acquired resistance
to isoniazid and rifampicin each in 1 patient, both with
evidence of poor adherence.
60. Moving past clinic-based care:

practical steps to improve TB care
through community engagement
Mobilising communities to fight TB in remote
islands of Indonesia
E Siahaan,1 S Ruschak,2 J Ake2 1Program Dept, Yayasan
Menara Agung Pengharapan Internasional, Medan Johor,
Indonesia; 2Global Program, MAP International, Brunswick,
GA, USA
Background: South Nias Regency is one of the most

under-developed areas in Sumatera Province. It contains
over 100 islands, many of which are inhabited by
indigenous people. TB prevalence in the Regency is
higher than the revised national estimate, but due to the
lack of human resources and the high transportation
costs between these remote islands, many TB patients do
not access care. Interventions: Indigenous community
volunteers were recruited and trained to screen for
symptoms of TB. Individuals suspected of having TB
were then referred to a health facility to smear
microscopy. Community volunteers facilitated transportation to ensure testing occurred. When a patient was
detected, the volunteers helped to initiate and follow up
treatment on the remote islands. In addition to these
community-based activities, laboratory technicians were
trained and monitored to improve the quality of
diagnostic services offered through government health
facilities.
Results: In a span of just 9 months, 1536 community

volunteers screened over 50 000 people for symptoms of
TB across 112 villages. These activities resulted in the
detection of 579 suspected TB patients who were referred
for testing. Over 225 smear-positive TB patients were
diagnosed and 44 of these newly detected patients were
among children aged less than 14 years. These active case
finding activities resulted in 86% increase in new

smear-positive case notifications compared with the
same quarters in the prior year, indicating that the TB
patients detected among people living on remote islands
would likely not have been treated in the absence of the
project.
Conclusion: This project demonstrates that community
health volunteers can be a highly effective solution for
expanding TB care services into remote and underserved
populations in conjunction with strengthening existing
government laboratory services. In light of the significant
upward revision Indonesias prevalence estimate and
subsequent decline in its case detection rate, initiatives
which are able to expand services and demonstrate
increases in anti-TB treatment uptake should be prioritized for scale-up.
Improving state-wide TB case notifications

through engaging leaders in nomadic
pastoralist communities in Nigeria
S John,1 M Gidado,2 T Dahiru,3 A Fanning,4 A Codlin,5
J Creswell,5 AD Belel6 (check S John???) 1Adamawa State
Tuberculosis and Leprosy Control Programme, Yola, 2KNCV
Nigeria/Challenge TB, Abuja, 3Netherlands Leprosy Relief,
Jos, Nigeria; 4Health, University of Alberta, Edmonton,
Alberta, Canada; 5STOP TB Partnership, Geneva, Switzerland;
6
Adamawa State Primary Health Care Development Agency,
Yola, Nigeria,
Background: With an estimated population of 4.1

million, Adamawa, one of the 36 States of Nigeria has
an estimated 450 000 Nomadic pastoralists. Poor access
to health services, low immunization coverage, poor
housing and overcrowding, high rates of bovine TB
infection and consumption of unpasteurized milk are
known to contribute to high rates of TB among Nomads.
Although Nigeria has one of the highest tuberculosis (TB)
burdens in Africa, case detection rates remain relatively
low. With support from Nomadic Community Leaders
(CL) and the Stop TB Partnership through its Wave 2 TB
Reach grant in 2011, TB control service was launched
within Nomadic Communities of Adamawa State.
Objective: To demonstrate results of 2 years Active TB
Case Finding (ACF) among Nomadic Pastoralists
through engagement of Nomadic Community Leaders
in Adamawa State of Nigeria.
Methods: ACF among nomadic populations was implemented from1st January 2012 to 31st December 2013 in
Adamawa State. The process involved mapping of cattle
routes, Nomadic settlements and Health Facilities,
advocacy and engagement of identified Nomadic CLs,
identification of Community Volunteers (CVs), reorientation of General Health Workers (GHWs) and training
of the CVs on sputum collection, transportation and
documentation. The CLs took responsibility for identification of CVs and provided leadership in the organization and conduct of the training & screening days
among community settlers. The CLs also accepted all
processes involved. Sputa were tested by AFB microscopy
and smear negative samples examined by GeneXpert
S75
MTB/Rif (Gx). A total of 378 community screening days

was organised.
Results: 680 CVs, 435 GHWs, 85 laboratory staff and 21
TB supervisors were trained on TB screening, diagnosis
and treatment. 96 376 nomads were verbally screened,
sputa from 9890 (10.3%) presumptive TB cases were
collected and examined while 1150 (11.6%) smear positive TB cases were detected. Of 8740 nomads with smearnegative microscopy results, 1685 (19.3%) were tested
using Gx, yielding an additional 160 TB cases. An increase
of 112% in number of patients submitting sputum for
smear microscopy State-wide compared with the 2 years
before the intervention was observed. New smear-positive
notifications increased by 49.5%, while notifications of
all forms of TB increased by 24.5% compared with
expected notifications based on historical trends.
Conclusion: Engaging Nomadic CLs to be the face of the
project was key to its success.
61. Progress in clinical trials for TB

treatment: 2015
High-dose rifapentine: preparation and design
of a four-month regimen trial
S Dorman 1Johns Hopkins University School of Medicine,
Baltimore, MD, USA
Optimizing the activity of individual components of a

tuberculosis regimen is part of a strategy to shorten
overall treatment duration. Rifamycins have inimitable
anti-tuberculosis activity, and evidence from animal and
clinical studies shows that the currently-recommended
10 mg/kg dosage of rifampicin is at the low end of the
dose-response curve. For drug-susceptible tuberculosis,
we and others have performed a sequence of research
studies with the goal of understanding how to optimize
rifamycin activity. Rifapentine is a cyclopentyl ringsubstituted rifamycin that, compared with rifampicin,
has a longer half-life and lower minimum inhibitory
concentration against M. tuberculosis. This talk will
provide an overview of pre-clinical, phase 1, and phase 2
data using daily rifapentine as a component of tuberculosis treatment. This talk will focus on the design and
implementation of a phase 3 multicenter, randomized,
controlled clinical trial to determine whether 4-month
regimens that contain daily rifapentine can cure drugsusceptible smear-positive pulmonary tuberculosis. The
investigational regimens in this phase 3 trial are A) 2
months of daily RPT/INH/PZA/EMB followed by 2
months of daily RPT/INH and B) 2 months of daily RPT/
INH/PZA/MOX followed by 2 months of daily RPT/
INH/MOX.
S76
Tuberculosis trials from the AIDS Clinical Trials

Group
S Swindells 1University of Nebraska Medical Center, Omaha,
NE, USA
The AIDS Clinical Trials Group (ACTG) was initially

established in 1987 to support the research efforts of the
National Institute of Allergy and Infectious Diseases
(NIAID) of the US National Institutes of Health. The
ACTG established and supports the largest Network of
expert clinical and translational investigators and therapeutic clinical trials units in the world, including sites in
resource-limited countries. These investigators and units
serve as the major resource for HIV/AIDS research,
treatment, care, and training/education in their communities. The Tuberculosis Transformative Science Group
of the ACTG is responsible for development, implementation, and oversight of the ACTG research agenda
related to the treatment and prevention of tuberculosis
with and without HIV co-infection. We collaborate and
coordinate with multiple partners internationally, and
contribute to global policy. Mentoring and training of
new investigators, particularly those in high burden
countries, is an important priority for the group. The
ACTG scientific agenda for TB includes the following
priorities: TB TREATMENT SHORTENING: To identify regimens to shorten drug-susceptible TB treatment to
63 months in patients with and without HIV MDR-TB
TREATMENT: To identify regimens for MDR TB
treatment in 6 6 months in patients with and without
HIV PREVENTIVE THERAPY: To treat latent TB in 1
month and MDR-TB infection in 3-6 months TB-HIV
CO-TREATMENT: To optimize the treatment of TBHIV co-infection, and evaluate and minimize drug-drug
interactions TRANSFORMATIVE SCIENCES: Pharmacology, biomarkers, laboratory monitoring and diagnostics, host-directed therapies This presentation will
include recent accomplishments of the group, and review
of ongoing studies and those in development for TB
prevention and treatment shortening for drug sensitive
and drug resistant TB. Developmental work with hostdirected therapies, biomarker discovery and validation
will be presented.
62. Innovative approach to improving

access and quality of care for DR-TB:
implementing the ECHO telehealth
model
knowledge and practice of clinical care teams in rural

and underserved communities. The ECHO model
based on a combination of multi-point videoconferencing, case-based learning, disease management and
outcomes monitoringwas initially developed and
shown to be effective in improving access to care and
treatment outcomes for Hepatitis C in rural and
underserved areas of New Mexico. The ECHO model
of clinical mentoring through guided practice is being
adapted in several global settings to increase access to
high quality care for tuberculosis.
Objectives: To improve access to high quality TB care
and treatment in both high and low burden TB settings
by leveraging regional and national expertise to support
a community of practice and learning. Interventions The
ECHO model has been adapted to support case
management of all TB cases in the low incidence setting
of New Mexico. In the high burden setting of Viet Nam,
it has been adapted to support case management of
complex DR-TB cases. TeleECHO sessions use a cloudbased videoconferencing platform to connect multidisciplinary regional and national experts to review and
provide consultation for TB cases presented over the
video network by local teams. Each teleECHO clinic
includes standardized case presentations, Q&A discussions and a brief clinical didactic talk.
Results: In 2015 successful TB ECHO clinics were
launched in both high [Viet Nam, 2/2015] and low
burden settings [New Mexico, 4/2015]. Participant
feedback has been very positive. Planning is in progress
for adaptation of the ECHO model to TB control
programs in Namibia and Mexico, the US state of
Washington, as a collaborative effort of US Regional
Training and Medical Consultation Centers (Curry
International TB Center, Mayo Clinic Center for TB)
with Denver Metro TB Control for low-incidence states,
and along the US-Mexico border for binational cases.
Conclusion: The ECHO telehealth model is being
applied to high burden TB settings to maintain and
improve MDR-TB control and low burden, low incidence TB settings to maintain and improve TB case
management and program quality. The ECHO model is
being adapted to TB control programs globally to
address quality of care and to strengthen the knowledge
and clinical practice of multi-disciplinary teams of TB
providers.
Adaptation of the ECHO model to strengthen

US-Mexico border DR-TB control
MART Castellanos,1 M A Garcia,2 B Struminger3
Tuberculosis, Ministry oh Health Mexico, Mexico,
2
Tuberculosis, Ministry of Health Mexico, Mexico DF, Mexico;
3
School of Medicine, University of New Mexico,
Albuquerque, NM, USA
1
Overview of the ECHO model and a framework

for ECHO implementation for DR-TB globally
L Chen,1 A Raftery,1 B Struminger2 1UCSF Curry
International Tuberculosis Center, San Francisco, CA,
2
University of New Mexico Health Sciences Center,
Background: The ECHO [Extension for Community

Healthcare Outcomes] telehealth model is an innovative
evidence-based intervention designed to strengthen the
Background: Tuberculosis is a very important public

health problem in Mexico, with around 20 000 new cases
per year; 33% are attributed to the six states bordering
the USA. 38% of the DR TB cases in Mexico are in these
border states. The Extension for Community Healthcare
Outcomes [ECHO] clinical mentoring and education
model is being adapted for use along the US-MX border

to promote improved patient outcomes of binational DR
TB cases; the model uses video conferencing technology
to facilitate discussion of binational and US-MX border
patient cases among MX and US experts on DR TB, with
a primary focus on Mexican TB patients living in the US
or who will be referred to Mexico.
Objectives: The ECHO Model is being implemented to
help ensure timely clinical review and case management
of binational and US-MX border DR TB patients and to
empower public health nurses and other health workers
in direct contact with the patients. This clinical
mentoring and education model aims to enhance TB
program quality improvement through positive changes
in the annual performance targets and DR TB cure
indicators.
Methods: In April 2015 the New Mexico DOH and
Project ECHO at the University of New Mexico
launched a TB ECHO program for the State of New
Mexico [NM], and in July 2015 began inviting TB
program managers and expert clinicians from existing
regional consultation and referral programs in the US
and Mexico to participate in the monthly case management discussions about the care and treatment of
binational TB cases linked to the State of NM.
Results: Beginning in July 2015 there have been monthly
TB ECHO sessions which have included 4-5 binational
patient case discussions each month, with participation
of clinical and public health experts from Mexico and the
US, as well as national and state TB program managers
from both sides of the border collaborating to develop
and coordinate treatment and case management recommendations.
Next Steps: The New Mexico binational case discussion
experience has served as a pilot for the development of a
US-MX binational DR TB ECHO program that will be
supported by the US-MX Border Health Commission
with plans to launch in late 2015 or early 2016.
Conclusion: Adaptation of the ECHO model for
collaborative binational DR TB case management and
clinical mentoring promises to strengthen binational and
border state TB programs in Mexico and USA to
standardize comprehensive case review and coordination
of case management to improve treatment outcomes of
binational DR and complex TB cases.
Integration of ECHO for DR TB with an HIV

ECHO program in Namibia
N Ruswa,1 N Hamunime,1 M Tadesse,1 F Mavhunga,1
F Tjituka1 1Directorate of Special Programmes, Ministry of
Health and Social Services, Windhoek, Namibia
Background: Namibia is a country in Southern Africa

with a high incidence rate of TB (651/100 000) and high
prevalence of HIV (13.3%). The country is sparsely
populated at with 2.6 persons per square kilometre, the
2nd lowest population density in the world. There are 34
district hospitals, spread over the 825 615 km2 and one
central hospital in the capital. In 2014, 367 cases with
drug resistant TB were registered. A clinical review
S77
consilium was established in 2009 to assist the districts

with the management of drug resistant TB cases. In
addition, the country has a clinical mentorship programme for HIV care.
Objectives: To pilot a clinical case discussion system for
drug resistant TB and TB-HIV using the ECHO telehealth model in Namibia, in order to improve interaction
between key health care workers, experts/mentors and
programme officers. Intervention All district TB offices
had been equipped with laptop computers. With support
from CDC and the Project ECHO, district TB and
leprosy coordinators and other key members of the
regional teams were granted access to the cloud-based
teleconferencing system. Presentation of cases had been
normally done by sending a fax and/or e-mails to a
national coordinator, who would then present to the
consilium. Attendance would be restricted to the persons
who could be physically present. With the ECHO model,
several district teams with cases to present would log on
at the same time, present and listen to each others
presentations.
Results: Requiring minimal add-on resources to what
was already available, the ECHO telehealth model
allowed regional and district health-care workers with
cases to present an opportunity to virtually attend the
consilium meetings, as well as use their experience to
respond to other teams presentations. This allowed for
real-time interaction, including prompt responses to
queries and prompt feedback, as well as learning between
teams. Teams with unstable internet connections, however, were often left out.
Conclusion: In a high TB-HIV setting with limited
human resources and experts like Namibia, a telehealth
conferencing system has a great potential. Because
electronic gadgets (computers, cellphones, tablets) are
often ubiquitous at least in the district health-care
setting, yet underutilised, this presents yet another
opportunity to integrate quality healthcare with available technology, maximise the benefits for the patients.
Implementation of the ECHO model for

comprehensive TB case management in New
Mexico, a low-incidence setting
D Fortune,1 D Isaacks,1 M Burgos,2 A Armistad,3
B Struminger3 1TB Program, New Mexico Department of
Health, Santa Fe, NM, 2Department of Internal Medicine,
University of New Mexico School of Medicine, Albuquerque,
NM, 3University of New Mexico Health Sciences Center,
Background: Project ECHO is an educational initiative

started 12 years ago at the University of New Mexico
Health Sciences Center with a mission to expand access
to high quality medical care in rural and underserved
communities across NM and globally. The ECHO model
uses a combination of video conferencing technology,
case-based learning, promotion of best practices and
monitoring of outcomes to support professional communities of learning and practice.
Objectives: The NM TB program (TBP) wants to ensure
effective nurse case management (NCM) of all persons
S78
with active TB by empowering the public health nurses

(PHN) with the knowledge and skills needed to ensure
effective completion of treatment (COT) for each
patient. NCM is an essential link for effective COT for
all persons with active TB Disease (TBD). The TBP aims
to enhance program quality improvement evidenced by
changes in the annual performance targets measured by
the 2020 National TB Indicators Project (NTIP).
Methods: NM averages 50 cases of active TB per year.
The NM TB TeleECHO Clinic, a collaborative between
the NM DOH and Project ECHO, convenes PHNs on a
monthly basis across a video network to review records
of all persons with active TB in NM. The focus is
strengthening of NCM skills, professional learning and
program quality improvement. Two trainings were held
during March/April in Albuquerque to equip the NCM
with information needed to initiate TB ECHO. A selfefficacy evaluation was completed to determine their
baseline perceived level of competence in providing
NCM. The TBP and ECHO are partnering to develop
an evaluation tool. Each clinic includes a brief didactic.
Results: NM TBP launched its first TB ECHO on April
20, 2015. In the first five months there have been 130
participants from across the United States and Mexico.
Twenty-seven unique persons with active TB (for a total
of 85 presentations) were presented by 17 NCM. Patient
ages ranged from six months to 88 years with TB from a
variety of sites. A total of 478 CMEs have been awarded.
Next Steps: Launch the Navajo Nation (NN) TB ECHO
Clinic and continue to provide mentoring for other TB
ECHO programs regionally and nationally.
Conclusion: This collaborative effort will enhance NCM
for persons diagnosed with active TB through standardized comprehensive case review and case-based learning.
The TB ECHO clinics have offered accessible TB nursing
educational opportunities for PHNs.
63. Zoonotic TB session II: What do

humans, baboons and livestock have in
common?
Human tuberculosis due to Mycobacterium
bovis in the United States 20062013
C Scott,1 J Cavanaugh,1 R Pratt,1 BJ Silk,1 P Lobue,1
P Moonan1 1Centers for Disease Control and Prevention,
Atlanta, Georgia, USA.
Background: Mycobacterium bovis is one of several

closely related mycobacteria of the M. tuberculosis
complex. Since 2005, nearly all culture-confirmed
tuberculosis (TB) cases reported in the United States
have been characterized by spoligotyping and MIRUVNTR, genotyping techniques that can differentiate M.
bovis from M. tuberculosis. Previous studies describing
factors associated with M. bovis disease in humans have
been limited by small cohorts, geographical scope, and
availability of robust data.
Methods: We analyzed data from CDCs Tuberculosis
Genotyping Information Management System to distin-
guish TB disease caused by M. bovis from M. tuberculosis among cases reported in the United States during
20062013. To determine demographic and clinical
factors independently associated with M. bovis disease,
as compared to persons with M. tuberculosis, we
estimated adjusted prevalence ratios (aPR) and 95%
confidence intervals (CI) using generalized linear modeling.
Results: A total of 70 312 culture-confirmed cases of TB
were reported. Of those 59 366 (84.4%) included
spoligotyping and MIRU-VNTR results; 862 (1.5%)
cases were identified as M. bovis. The annual percentage
of TB cases attributable to M. bovis remained consistent
nationally during 20062013 (range: 1.31.6%). Geographically, the incidence rate of M. bovis was higher
than the national average (0.29 per 100 000 persons) in
the US-Mexico border region (Figure). Among TB cases,
the prevalence of M. bovis was higher among females
(aPR 1.4, 95%CI 1.31.6), children (compared to 25-44
year olds) (0-4 years: aPR 1.9, 95%CI 1.42.8 and 5-14
years: aPR 4.0, 95%CI 3.15.3), foreign-born persons
(aPR 1.4, 95%CI 1.21.7), Hispanics (compared to nonHispanic Whites) (aPR 3.9, 95%CI 3.05.0), and
residents of US-Mexico border counties (aPR 2.0,
95%CI 1.72.4). Cases with M. bovis were more likely
to have exclusively extrapulmonary disease (aPR 3.7,
95%CI 3.34.2) or both pulmonary and extrapulmonary
disease (aPR 2.4, 95%CI 2.12.9).
Conclusions: This is the first comprehensive analysis of
M. bovis disease since genotype surveillance was
implemented routinely in the United States. Women,
children, foreign-born persons, and Hispanics are disproportionately affected by M. bovis, which is independently associated with extrapulmonary disease. Targeted
prevention among Hispanic and foreign-born persons
and residents of the US-Mexico border region is
warranted.
Figure: Cumulative incidence of human Mycobacterium
bovis tuberculosis disease, by county, United States,
2006-2013.
Longevity of M. bovis in fresh and souring milk

and the screening of milk from communal
cattle with a high prevalence of BTB for M.
bovis
A Michel,1 C Geoghegan,2 T Hlokwe,3 K Raseleka,3
W Getz,4 T Marcotty5 1Faculty of Veterinary Science, Dept.
Veterinary Tropical Diseases, University of Pretoria,
Onderstepoort, 2Mammal Research Institute, University of
Pretoria, Pretoria, 3Zoonotic Diseases Section,
ARC-Onderstepoort Veterinary Institute, Onderstepoort,
South Africa; 4Environmental Science, Policy and
Management, University of California, Berkely, Berkeley, CA,
USA; 5Institute of Tropical Medicine, Antwerp, Belgium
Background: Unpasteurised fresh and souring dairy

products form an essential component of household
diets throughout many rural communities in southern
Africa. The presence of milk-borne zoonotic pathogens
such as Mycobacterium bovis, the causative agent of
bovine tuberculosis and zoonotic tuberculosis in humans,
constitute a public health threat, especially in areas with
poor disease surveillance in livestock and highly compromised human health due to HIV/AIDS.
Methods: In this study we used culture to determine the
longevity of M. bovis in experimentally inoculated fresh
and naturally souring milk obtained from communal
cattle in the KwaZulu-Natal province of South Africa.
The effect of bacterial load and storage temperature on
the survival of M. bovis was evaluated by spiking
mixtures of fresh milk and starter soured milk (aMasi)
culture with three concentrations of bacteria (102, 104,
107 colony forming units/ml), followed by incubation
under controlled laboratory conditions that mimicked
ambient indoor (208C) and outdoor (338C) temperatures
and periodic sampling and testing over time (0-56 days).
Results: M. bovis cultured from samples of the fresh and
souring milk was identified by PCR analysis. At the
highest spiking concentration (107cfu/ml), M. bovis
survived for at least 2 weeks at 208C, but, at all
concentrations in the 338C treatment, M. bovis was
absent by 3 days after inoculation. Logistic regression
analysis was used to assess the effects of bacterial
concentration and time since inoculation, as well as
determine the potential half-life of M. bovis in raw
souring milk. Given the most favourable tested conditions for bacterial survival (208C), approximately 25%
of mycobacteria were alive after one day of storage (95%
CI: 9-53%), giving an estimated half-life of M. bovis in
raw souring milk of approximately 12 hours (95% CI: 727 hours).
Conclusions: This study demonstrates that M. bovis may
survive in fresh and souring milk for periods of time that
represent a risk of exposure to people consuming these
products, as well as domestic or wild animal populations
that have reported opportunities to consume homemade
unpasteurised dairy products. The temperature at which
the milk is soured and stored substantially affects the
survival time of M. bovis.
S79
Zoonotic tuberculosis in Brazil: a very low

prevalence or a neglected problem?
M Silva,1 FR Araujo,2 RR Da Costa,3 P Suffys,4 AS Rocha,4
MDC Guimaraes5 1Embrapa Dairy Cattle, Embrapa, Juiz de
Fora, MG, 2Embrapa Beef Cattle, Embrapa, Campo Grande,
MS, 3Joao Penido Reginal Hospital, FHEMIG, Juiz de Fora,
Graduaca
MG, 4IOC, Fiocruz, Rio de Janeiro, RJ, 5Pos
o em
Saude
Publica,
Universidade Federal de Minas Gerais, Belo
Horizonte, MG, Brazil
Background: Sputum acid-fast bacilli (AFB) microscopy

and histopathology are the standard criteria for tuberculosis (TB) diagnosis in Latin America (LA), including
Brazil. However, these procedures may potentially
overlook zoonotic TB cases since these methods are not
able to detect Mycobacterium bovis nor differentiate it
from the rest of the Mycobacterium tuberculosis complex
(MTC).
Results and Discussion: To our knowledge, systematic
surveys on zoonotic TB in LA have only occurred in
Argentina, linking health and agriculture research
centers (National Agricultural Technology Institute). In
that country, in the period 1982-1984, which involved 15
human laboratories, M. bovis accounted for 0.47% of
the samples, ranging from 0.04% to 1.85%, depending
on the degree of involvement with livestock in each
region. The estimated proportion of zoonotic TB due to
M. bovis in LA is 2% and 8% of pulmonary (PTB) and
extrapulmonary (EPTB) TB cases, respectively. It is
believed that in Mexico this proportion is greater. There
is a lack of official data regarding the current prevalence
of bovine TB in Brazil. According to the Ministry of
Agriculture, Livestock and Food Supply there was a
national average prevalence of 1.3% of the cattle from
1989 to 1998. It is also known that the incidence of
zoonotic TB in Brazil was high in the last century, causing
3.5% and 13.2% of all TB and TB meningitis cases found
by two small surveys in 1974 and 1955, respectively.
However, more recent studies in urban centers in Brazil,
also of small sizes, indicated zoonotic TB prevalence of ,
1%, , 1%, zero, zero and 1.6% for, respectively, Sao
Paulo (2001-2005), Rio de Janeiro (1987-2006), Juiz de
Fora (2011), Rio de Janeiro (2011), Juiz de Fora (2014).
More systematic studies on zoonotic TB, using media
with pyruvate, are necessary in LA and Brazil. One
alternative includes nested surveys within other studies
such as the multiple drug resistance (MDR) of MTC.
This could potentially save resources and construct better
prevalence estimates. Brazil could also create a collaboration between The Ministry of Health and The
Agricultural Research Corporation (EMBRAPA) for
these surveys, as in Argentina.
Conclusion: The last published studies in Brazil found
that the prevalence of zoonotic TB was very low or
marginal. However, all results were based on small
surveys in restricted locations or referral laboratories
results, overlooking the real prevalence of zoonotic TB.
S80
Domestic Mycobacterium bovis circulating in

Chacma baboons and their public health
implication
1
M Musso, B Hangombe, C Nakajima, Y Suzuki

1
School of Veterinary Medicine, University of Zambia, Lusaka,
Zambia; 2Hokkaido University Research Center for Zoonosis
Control, Hokkaido, Japan
Bovine tuberculosis (BTB), caused by Mycobacterium

bovis has insidiously and mutely devastated livestock and
wildlife of the Kafue Basin in Zambia. Until recently, the
disease was diagnosed only in cattle and Kafue lechwe
antelopes (Kobus leche kafuensis). However, a recent
screening of Chacma baboons (Papio ursinus) revealed a
spillover effect in infection. Field necropsy findings had
revealed typical granulomatous lesions in the entire chest
cavity affecting the lungs and all the ancillary draining
lymph nodes of the affected baboon. In order to elucidate
the exact strain of circulating Mycobacterium tuberculosis complex (MTC) bacteria, a logical and chronological step wise genotyping analysis was applied. Firstly, a
MTC-discrimination multiplex PCR (MTCD-MPCR)
targeting genetic regions cfp32, RD9 and RD 12 was
used for species differentiation followed with the
application of deletion analysis using RD4, TbD1 and
3 0 cfp32 probes. Then genetic fingerprinting of the
identified M. bovis strains was done using the direct
repeat (DR) based spacer oligonucleotide typing (spoligotyping) in association with mycobacterial interspersed
repetitive unitsvariable-number tandem repeats
(MIRU-VNTR). Spoligotyping and MIRU-VNTR analyses revealed that tuberculosis in baboons of Lochinvar
National Park shared one distinct spoligo pattern; SB0
120, with cattle and Kafue Lechwe antelopes. This
finding has significant public health implications given
the close proximity with which baboons roam and
forcibly enter human dwellings. This scenario reflects the
baboons nature with regards to their omnivory eating
habits, spatial distribution in close relation to human
settlements, habitat and resource utilization patterns,
disease susceptibility, transmission modes and their
social interaction patterns with access to defensible
common pool resources (CPRs) available to humans.
The full public health implication, impact and extent of
this current finding are likely to be seen in the long-term.
Multiplicative factors responsible for its occurrence in
this particular specie of social animals will partly depend
on the pathogenicity of this BTB strain (SB 120).
Therefore future research direction in this area should
determine the possible implications of this finding on
human health at the wildlifelivestockhuman interface.
Tuberculosis diagnostic challenges in wildlife

and public health implications
M Miller,1 SDC Parsons,1 P Van Helden,1 P Buss,2
F Olea-Popelka3 1DST/NRF Centre of Excellence for
Biomedical Tuberculosis Research, MRC Centre for TB
Research, Division of Molecular Biology & Human Genetics,
Faculty of Medicine & Health Sciences, Stellenbosch
University, Cape Town, 2Kruger National Park, Veterinary
Wildlife Services, South African National Parks, Skukuza,
South Africa; 3College of Veterinary Medicine and Biomedical
Science, Department of Clinical Sciences, Colorado State
University, Fort Collins, CO, USA
Background: Bovine tuberculosis (bTB) affects wildlife

populations worldwide. Over 21 wildlife species have
been reported with bTB in South Africa alone. There is a
broad diversity of species affected including carnivores,
primates, and bovids. The presence of multiple susceptible hosts complicates detection and disease control.
Diagnostic challenges: There are no validated tests for
bTB in most wildlife species. Some of the challenges to
diagnosing bTB in wildlife are differences in epidemiology, pathogenesis, and immunologic responses between
hosts; variable or absent clinical signs in infected animals
until disease is advanced; frequent environmental exposure to nontuberculous mycobacteria; requirement for
species-specific reagents for immunodiagnostic tests; and
logistical difficulties associated with testing wildlife. To
overcome these issues, the focus has been on developing
blood-based tests using Mycobacterium bovis-specific
antigens that may be used across a range of hosts.
Identification of novel biomarkers, development of
species cross-reactive reagents, and improved techniques
for sample processing and transport are currently
underway.
Public health implications: With diminishing habitats,
there are increased wildlife-livestock-human interfaces
and a growing threat of disease transmission. With a lack
of tools for detection and control of bTB in wildlife, there
is a threat of spillback to livestock, as well as
transmission to both animals and humans directly
through contact or ingestion of infected animals products
or indirectly, through contamination of the environment.
Limited bTB control in wildlife presents a public health
concern for zoonotic TB.
64. TB contact investigation in high-risk

populations: innovative approaches and
implementation experiences
The yield of contact investigation in a rural
setting in Ethiopia
DJ Dare,1 Z Dememew,1 Dr Habte,1 M Ensermu,1
G Negussie 2 1HEAL TB Project, Management Sciences for
Health, Addis Ababa, 2Disease Prevention and Control,
Amhara Regional Health Bureau, Bahir Dar, Ethiopia.
Background: Screening close contacts of index tuberculosis (TB) cases is one of the key strategies for case
finding in high TB burden settings. According to current
recommendations, contact investigation is done prospectively. Here we present results from a retrospective
contact investigation approach implemented in six zones
in two regions of Ethiopia.
Methods: Between June-September 2014, we recruited
and trained lay providers to do active case search through
community based contact tracing. They first identified
the list of smear positive pulmonary TB (SS) cases
treated in the catchment health centers in the previous 3
years. Then they visited the houses of each SS index case
and enumerated their household contacts. Using a
symptom-based screening checklist, they screened each
household contact and referred those with symptoms to
catchment health centers for diagnosis and management.
In the routine system, Health Extension Workers screen
household contacts of SS TB index cases currently on
treatment. We computed the yield of this new approach
and compared with the yield from the routine community TB referral.
Results: Of 77 626 HH contacts of 20 805 index SS TB
cases screened, 9142 (11.8%) had presumptive TB. Over
a half (54%) of these were further evaluated at the health
centers within 2 months of referral. The yield of all forms
of TB among presumptive cases through the retrospective
screening approach was 5%. On the other hand, the yield
among 22 426 presumptive TB cases referred through the
routine community TB activity by health extension
workers during the same time period was 2.9%.
Conclusions: Targeting household contacts of SS index
cases through a retrospective screening approach can
serve as additional high yield strategy in TB case finding.
Further studies are needed to determine the optimal
timing for the retrospective tracing. Mechanisms should
be in place to ensure timely linkage of the presumptive
TB cases to diagnostic centers.
Contact screening and expanding childhood TB

diagnosis and treatment of TB in Bangladesh
S Islam,1 M Siddiqui,1 M Rana,1 M Akramul Islam,1
M Husain1 1Health, BRAC, Dhaka, Bangladesh
Background and Challenges: Bangladesh is one of the

high TB burden countries in the world. BRAC, a nongovernment organization, is jointly working with national TB control programme to combat against tuberculosis since 1994. Due to unavailability of diagnostic
S81
facilities and scarcity of pediatricians, diagnosis and

management of child tuberculosis is still a major
challenge for the country especially in rural setting. It is
necessary to raise awareness in the community and
increase screening through contact tracing of household
cases within the existing programme for identifying more
child TB cases.
Interventions: Considering the different challenges,
BRAC started conducting orientation to pediatrician at
tertiary hospitals and district levels. A basic training is
given to health care providers on sign symptoms of child
TB to increase awareness in the community during
mobilization. A contact tracing register was introduced
in the existing programme of BRAC in late 2013 and
contact tracing initiated among the family members who
are in close contact with pulmonary TB cases and were
sent for screening. During their households visit, BRACs
frontline health workers disseminate basic messages,
identify presumptive cases and refer them to the specific
centres. Poor child TB symptomatic gets financial
support from BRAC for investigation and transportation
purpose.
Result and lessons learnt: In 2014, a total of 929
pediatricians were reached through 64 networking
meetings on child TB diagnosis and management at
tertiary and district level hospitals. In addition, 1029
health care providers were trained on child tuberculosis
including basic messages, identification and referral.
There was an increase in child case identification from
2865 in 2013 to 3554 in 2014. In 2014, 9.7% of the total
child TB cases were identified through contact tracing.
Conclusion: Evaluation of symptomatic contacts may
increase identification of child cases. An orientation
session could be planned for health care providers to
improve the contact tracing skills. The children who are
in close contact of pulmonary tuberculosis patient if even
asymptomatic should be referred for screening and
should do follow-up.
TB contact investigation: experiences from

prisons in Uganda
E Kizito,1 M Kenneth,1 M Ruhweza,1 S Stavia,2
P Donggo,3 B Woldemariam,4 P G Suarez4 1Project
(TRACK TB), Management Sciences for Health (MSH),
Kampala, 2National Tuberculosis and Leprosy Program,
Ministry of health, Kampala, 3Lira regional Referral Hospital,
Ministry of Health, Lira, Uganda; 4Track Tuberculosis Activity,
Management Sciences for Health, Arlington, TX, USA
Introduction: One of the components of the WHO Stop

TB strategy is to address the needs of TB contacts, and of
poor and vulnerable populations. Inadequate ventilation,
overcrowding and inadequate TB infection control
practices in congregate settings increase the risk of
transmission of all forms of tuberculosis. During 2014,
Lira regional referral hospital in Northern Uganda
diagnosed 12 patients with drug sensitive TB and 2
patients with multi-drug resistant TB all of whom were
serving jail sentences in nearby prisons. Over the years,
such patients would be transferred to the hospital for
S82
treatment but there was no documented contact screening and investigation for all inmates.
Intervention: Lira Regional referral hospital with support from the USAID funded TRACK TB Project
conducted education sessions about TB for the prisons
authorities and inmates with a focus on TB transmission
and infection control including the need to promptly
identify and evaluate presumptive TB cases. Multidisciplinary teams comprising nurses, clinicians and
laboratory technologists visited each of the prisons to
conduct the sensitization session and screening of all
inmates using the intensified case finding tool (ICF).
Sputum samples were collected from all inmates that
were found to have TB symptoms (presumptive TB cases)
and subjected to an MTB/RIF test.
Results and lessons: A total of 300 inmates were screened

for TB, all were males just like the index cases and the
average age was 32 years of age. All these contacts were
HIV negative. Fifty three symptomatic contacts were
identified and had samples subjected to the MTB/RIF
test; 3 contacts (1%) were diagnosed with were
diagnosed with TB as shown in the cascade in the Figure.
Conclusion and next steps: Contact investigation in Lira
prison led to identification of inmates with undiagnosed
TB which increased the risk of transmission within the
congregate setting. We recommend TB contact screening
for both household and close contacts of inmates
diagnosed with TB in addition to the periodic and
routine TB screening at entry into the prison setting.
TB clusters in villages: community screening

and GIS mapping
1
S Hamusse Daba, M Aseresa Melese, DJ Dare,

B Lindtjorn3 1Oromia Regional Health Bureau, Addis Ababa,
2
Management Sciences for Health, HEAL TB Project, Addis
Ababa, Ethiopia; 3University of Bergen, Bergen, Norway
Background: Identifying TB hot spots using special

analysis may contribute for targeted TB control measure
and can be used as advocacy tool. We used GIS and
Spatial Scan Statistics to describe the spatial distribution
of TB and to identify TB hot spots in a district in Central
Ethiopia.
Methods: A population-based cross sectional study
conducted in the Hetosa District of Arsi Zone of central
Ethiopia from May 2013 to July. We used symptom

checklists and sputum microscopy for TB diagnosis.
Demographic data and geographical location of the
identified smear positive TB cases and the residential
addresses of all study participants in the study area were
also collected using handheld Global Positioning System
(GPS) receivers. We used spatial scan statistic used to
identify purely spatial and spacetime clusters of tuberculosis among permanent residents. The statistical significance of identified clusters was based on comparing the
likelihood ratio test statistics against a null distribution
obtained from Monte Carlo simulation. The number of
variation was set at 999 and the significance level was set
at 0.05.
Results: A total of 106 smear positive pulmonary TB
cases were identified with the overall smear-positive TB
incidence rate of 323.2/100 000 (range, 0-608) population per year in the study area. The most likely statically
significant cluster for high rates and secondary cluster of
TB were identified in three kebeles (neighborhoods),
Gonde Finchema,Tado Lamen and Shaki Sherara
(RR:3.10, P0.028) and in three kebeles Oda Jila,Seru
Ankato and Boru Lencha (RR:2.89, P0.052) respectively. Furthermore, the most likely statistically significant cluster for low rate of TB was identified in one
kebekle Dawe Guticha Kebele (RR:1.68, P0.019).
Areas with high rate of TB in the purely spatial analysis
were also found to have statistically significant high rate
of TB in the spacetime analysis. Without including
purely spatial clusters, the most likely statistically
significant cluster for high rates of TB was also identified
in Gonde Finchema, Tado Lamen Kebeles and Shaki
Sherara kebeles during the dry season between December
2013 and May 2014 (RR:6.10, P 0.016).
Conclusion: In this study, we have demonstrated the
presence of clusters of PTB in Hetosa district of Arsi
Zone of Central Ethiopia. Using spatial and spatialtemporal analysis may be useful for designing areaspecific interventions for TB control.
65. The XDR experience: clinician and

patient perspectives on treatment and
care
Knowledge, attitudes and behaviours of XDRTB patients and family members in the Western
Cape, South Africa
M Senthilingam,1,2 E Pietersen,1 J Te Riele,1 P Sedres,1
K Dheda,1 R McNerney2 1University of Cape Town, Cape
Town, South Africa; 2London School of Hygiene and Tropical
Medicine, London, UK
Objective: Patient-level data are required to inform

strategies interrupting transmission and default in patients with extensively drug-resistant TB (XDR-TB) to
improve models of care and identify potential routes of
transmission. We therefore explored the experiences,
lifestyle, attitudes and needs of uncured XDR-TB
patients, who failed or interrupted therapy, living without

treatment in the community.
Methods: We conducted in-depth interviews with 12
community-based patients from South Africa. Family
members were interviewed when patients were unavailable. Interviews were analysed using inductive thematic
analysis.
Results: The thematic experiences identified from the
interviews were: i) living with but not being cured of
XDR-TB, ii) altered lifestyle in the community, iii)
experiences with community healthcare and iv) local
community members, and v) wants and needs. Patients
identified mistrust in healthcare, futility of treatment
regimens, a need for a purpose in life, and subsistence as
major concerns. Restriction of living in the community for
patients whose treatment had failed resulted in selfimposed isolation. Defaulters focused more on the neverending drug regimen and bad experiences with healthcare
contributing to non-adherence. Family members emphasized an under-recognized experience of unforeseen burden, obligation, worry, and discomfort. Lack of knowledge and lack of concern about transmission was evident.
Conclusion: Current models of care are not adequately
meeting the needs of uncured XDR-TB patients and
relatives. These data inform the need for communitybased palliative care, vocational facilities to improve
economic opportunities, home-based infection control,
and improved psychosocial support to increase patient
adherence, reduce transmission, provide income, and
relieve burden on family members.
Social constraints of DR-TB care: the

construction and management of stigma
A Daftary 1University of Toronto, Toronto, ON, Canada
The social inequality and marginalisation experienced by

patients living with multi drug-resistant tuberculosis
(MDR-TB)and extensively drug-resistant tuberculosis
(XDR-TB) is seldom characterised. This presentation
draws upon contemporary understandings of stigma
theory, and qualitative evidence from studies involving
patients and health care workers in KwaZulu Natal,
South Africa, to develop insight into the construction and
management of stigma in drug-resistant TB. Felt and
enacted TB in MDR- and XDR-TB, as compared to drugsusceptible TB, may reflect the distinct ways in which each
form of TB is experienced and managed by patients, and
by health care providers. The coincident risk of HIV
infection may add an additional layer of social complexity
into the lived experience of drug-resistant TB, patients
consequent health-seeking behaviours and treatment
decisions may reflect their need to mitigate multiple constraints and resist dual stigmatisation. A better understanding of how stigma is produced, managed and/or
resisted in drug-resistant TB can inform the development
of health care policies and practices that prevent stigma
reification, with specific implications in TB service
delivery, adherence support, prevention of acquired
drug-resistance, and the care of HIV coinfected patients
S83
The patient journey: treatment for XDR-TB in

South Africa
M Loveday,1 L Vanleeuw,1 V Mehlomakhulu1 1Health
Systems Research Unit, South African Medical Research
Council, Cape Town, South Africa.
To better understand patients with DR-TB and inform

interventions on treatment adherence, morbidity and
mortality, we collected socio demographic information
as well as data on patients treatment experiences from
79 DR-TB patients. We found that many DR-TB
patients have low levels of education and depend on
disability and child support grants for survival. Experiences of health services varied, but travel to the
specialised TB hospital for monthly management and
repeat medications was time consuming, costly and at
times unsuccessful. Treatment adherence together with
treatment outcomes can be improved by making access
to treatment easier and improving health service
functioning.
A view from the other side: perspectives on

clinical management of XDR-TB
I Master 1MDR TB Unit, King Dinuzulu Hospital Complex,
Durban, South Africa
A practical look from the coal face from the clinicians

perspective of the XDR program in South Africa. The
situation is looked at in 3 periods. The period prior to
2006, between 2006 and 2013 and from 2013 onwards.
A look at the difficulties, problems facing clinicians and
the patient and what changed over the years. A short
look as well at what is facing clinicians in the current
period and going forward.
S84
Abstract presentations, Friday, 4 December
FRIDAY
4 DECEMBER 2015
e-POSTER SESSIONS
01. Zoonotic tuberculosis: trends,
detections and interpretation
EP-100-04 Development of a single-tube triplex
real-time PCR assay for differential diagnosis of
Mycobacterium tuberculosis complex
WL Huang,1 R Jou1,2 1Centers for Disease Control, Taipei,
2
National Yang-Ming University, Taipei, Taiwan. Fax: (886) 2
2653 1387. e-mail: rwj@cdc.gov.tw
Background: To set up a laboratory-based surveillance

system for Mycobacterium bovis infection in high-risk
areas and to monitor adverse reactions of BCG vaccination, we need to have a simply and rapid test for
simultaneous detection of M. bovis, M. bovis-BCG and
other Mycobacterium tuberculosis complex (MTBC).
Methods: A single-tube triplex real-time PCR using three
fluorescence probes was developed. For validation study,
we tested 72 reference Mycobacterium species including
12 M. tuberculosis, 1 M. microti, 1 M. africanum, 3 M.
bovis, 17 M. bovis-BCG and 38 nontuberculous mycobacteria (NTM), and 625 clinical MTBC isolates. In
addition, for clinical evaluation, 226 smear-positive
sputum specimens were tested. Results were compared
to that of spoligotyping, a line-probe assay and/or
bacterial culture.
Results: The detection limit of the triplex real-time PCR
is 10 fg DNA. Of the 72 reference isolates tested, no
discordant result was observed. Specificity of the triplex
real-time PCR is 100%. Of the 625 clinical isolates
tested, 14 M. bovis were identified using the triple realtime PCR and confirmed by both commercial tests.
Furthermore, of the 226 sputum specimens, 64 (28.3%)
were culture-negative, 69 (30.5%) were non-MTBC and
75 (33.2%) were MTBC culture-positive, 10 (4.4%)
were contaminated, and 8 (3.5%) had no culture results.
Of the 75 MTBC positive isolates, the sensitivities of the
triple real-time PCR and the line-probe assay were
97.3% (73/75) and 93.3% (70/75), respectively. Besides,
of the 226 sputum specimens, we further differentially
identified 2 M. bovis cases later confirmed by bacterial
cultures.
Conclusion: The performance of the triple real-time PCR
for detecting MTBC was comparable to that of the lineprobe assay. In addition, the triple real-time PCR can be
applied for M. bovis surveillance.
EP-101-04 Risk factors associated with bovine

tuberculosis in Ecuadorian workers from farms
and slaughterhouses
F Proano-Perez,1 R Benitez-Capistros,2 W Benitez-Ortiz,2
L Ron-Garrido,2 F Portaels,3 A Linden,4 L Rigouts3
1
Universidad De Las Fuerzas Armadas Espe, Sangolqui,
2
International Centre for Zoonoses, Universidad Central Del
Ecuador, Quito, Ecuador; 3Biomedical Sciences, Institute of
Tropical Medicine, Antwerp, 4University of Liege, Liege,
Belgium. Fax: (593) 223 4952. e-mail:
freddyproanoperez@yahoo.com
Background: Bovine tuberculosis (BTB) is a chronic

zoonotic disease caused by Mycobacterium bovis, which
has a worldwide distribution in domestic and wildlife
animals. In Ecuador, there is no national BTB control
program.
Design/Methods: The aim of this study was to evaluate
the state of the disease in cattle and humans, and evaluate
the risk factors associated in the transmission of BTB.
Twenty dairy herds comprising 2022 cattle were selected
from medium and large size herds located in Mejia
canton. In humans, the study was carried out in 157
people, conformed by farm workers and slaughterhouse
workers, with the purpose to evaluate the prevalence of
Mycobacterium spp. in risk populations by tuberculin
skin test (TST) and analyse risk factors.
Results: The true annual incidence was 1.70%, and the
true prevalence was 7.41% and 7.13%, in 2 years. The
number of skin testpositive cases also increased
significantly with age (P 0.03), contacts with other
species of animals (P , 0.01), and introduction of new
cattle (P 0.04). Herd prevalence was 55% in the first
year and 65% in the follow up study. Mycobacterium
spp. was estimated in 29% [C.I. 22 - 36% at 95%] of
the farm and slaughterhouse workers. The results
establish a significant association among the positivity
of TST to masculine gender (P 0,026), consumption of
milk coming from farms (P 0,030), consuming raw milk
(P 0.000), and consumption of elaborated cheeses
made by hand (P 0.003). Consumption of raw milk gave
an odds ratio of 4.65 [ C.I. 2.0110.75 at 95%].
Conclusion: This study shows the necessity for a national
BTB control program in Ecuador which should be
supported by sanitary policies to allow the identification
of the endemic areas nationwide, establishing animal
movement restrictions, culling, sanitary slaughtering,
compensation strategy for culling and laboratory tests to
identify the Mycobacterium species and strains involved.
In addition, due to the lack of knowledge in cattle
farmers about this zoonoses, education of farm and
abattoir workers about the disease could help prevent
new cases in animals and humans, and help improve the
use of preventive measures.
EP-102-04 Zoonotic tuberculosis: the case of

mahouts and captive Asian elephants in
southern India
1
2 1
V Kummannoor P Pillai, D Abraham Government

Medical College Hospital, Kottayam, 2Kerala State Animal
Husbandry Department, Kozhikode, India. e-mail:
kpvda@hotmail.co.in
Background and challenges to implementation: Cases of

tuberculosis in captive elephants are mostly caused by
Mycobacterium tuberculosis. In the hands-on and open
systems of captive elephant management in southern
India, infected captive elephants present the risk of
zoonotic tuberculosis to their mahouts (traditional
elephant keepers). There are nearly a thousand captive
Asian elephants and not less than 3,000 mahouts in
southern India. Post-mortem examinations of captive
elephants have shown caseous nodules in lung parenchyma from which M. tuberculosis was isolated by
culture on LJ media. Systematic and regular tuberculosis
screening of mahouts and captive elephants still remains
a challenge.
Intervention or response: Over a three year period we
piloted systematic tuberculosis screening of both mahouts and captive elephants at several different locations
in southern India. One time screening of nearly 800
elephants and their mahouts was achieved. Screening of
mahouts was done by clinical examination and tuberculin skin testing, followed by chest X-ray and sputum
collection as required. Screening of elephants was done
with the USDA licensed rapid serum test DPP Vet Assayw
(Chembio Diagnostics Inc. Medford NY), followed by
trunk wash culture for isolation and identification of
mycobacteria. Past mahouts of eight elephants from
which M. tuberculosis was isolated at post-mortem
examination were screened for tuberculosis infection.
Results and lessons learnt: One case of cutaneous
tuberculosis in the palm was identified in a mahout
who attended an elephant with active tuberculosis.
Prevalence of latent tuberculosis infection among mahouts is 53% and the seroprevalence of tuberculosis
infection among captive elephants in southern India is
nearly 15%. Mahouts and captive elephants in southern
India are highly migrant and locating the subjects for
contact tracing and follow-up testing is challenging.
Mahouts, mainly because of their nomadic lifestyles, are
least aware of their basic health care needs as well the
need for regular tuberculosis screening. Awareness of the
owners of captive elephants for tuberculosis screening of
elephants is also minimal.
Conclusions and key recommendations: For early diagnosis and treatment, regular and systematic tuberculosis
screening of mahouts and captive elephants are essential.
Provisions need to be made for paid leave for mahouts
identified with active tuberculosis to complete full course
treatment.
S85
EP-103-04 Detection of potential reservoirs of

M. tuberculosis and M. bovis
K P Hanumanthappa1 1All India Institute of Medical
Sciences, New Delhi, India. Fax: (91) 112 658 8663. e-mail:
hk_prasad@hotmail.com
Background: Several pathogenic mycobacteria such as

M. tuberculosis, M. bovis, and M. avium subspecies para
tuberculosis are capable of causing clinical disease in
humans as well as animals. For e.g., M. tuberculosis and
M. bovis are responsible for human and bovine
tuberculosis; whereas M. avium subspecies paratuberculosis is the causative agent of Johnes disease among
bovines and has been implicated in Crohns disease
among humans. The presence of pathogenic mycobacteria in diverse species weaves a complex web perpetuating
its survival in the environment. However, the consequence of the prevalence of pathogenic mycobacteria in
diverse species has not been systematically assessed;
owing to the technical difficulties in terms of: primary
isolation, demanding culture conditions, time taken, cost
and laborious nature of the classical methods of
identification. Identification of mycobacteria has been
restricted to a limited number of referral laboratories.
Further, it has not been considered essential to distinguish M. bovis and M. tuberculosis in clinical samples as
would be the case in other microbial diseases, since
patients respond to a common panel of anti-tubercular
therapy.
Methods: In this regard we have developed Nested-PCR
and visual format assays using molecular beacons for
direct identification of M. tuberculosis and M. bovis in
clinical samples. Identical techniques would go a long
way in assessing the presence of the various mycobacterial pathogens in the clinical samples & its prevalence in
the environment.
Results: Our own in house studies have demonstrated the
presence of M. bovis apart from M. tuberculosis in
clinical samples. Similarly besides M. bovis, M. tuberculosis has been detected and isolated from bovine samples,
(Reverse Zoonoses), demonstrating that the intimate
association of patients and cattle facilitates the transmission of these pathogens from humans-cattle-humans.
Conclusion: Therefore, there is a dire need to develop
technology that reliably facilitates identification of
mycobacterial pathogens in circulation. This would help
in the generation of data for the systematic identification
of the prevailing & potential reservoirs of infection. The
impact such facts would be significant in developing
versatile strategies for the prevention of the spread of
mycobacterial pathogens among humans and other
susceptible hosts.
S86
EP-104-04 Isolation of Mycobacterium

tuberculosis in livestock workers and
implications for zooanthroponotic
transmission in Ibadan, south-western Nigeria
EP-105-04 Is there a difference in time to

sputum culture conversion between
Mycobacterium bovis and Mycobacterium
tuberculosis disease?
S Cadmus,1 V Akinseye,1 A Adegbulu,1 N Ovwighose,1

M Ayoola,1 J Ogugua,1 H Adesokan,1 E Cadmus1
1
University of Ibadan, Ibadan, Nigeria. e-mail:
sibcadmus@yahoo.com
C Scott,1 J Ershova,1 J Cavanaugh,1 G M Mazurek,1

B J Silk,1 P Lobue,1 P Moonan1 1U.S. Centers for Disease
Control and Prevention, Atlanta, GA, USA. Fax: (1) 404 718
8308. e-mail: ibk9@cdc.gov
Background: Tuberculosis (TB) remains a disease of

global health importance particularly in countries of subsaharan Africa where knowledge of the disease is poor.
Again in Africa, little is also known about the epidemiology of TB among livestock workers who play key roles
in the transmission of the disease along the humananimal-ecosystem interface. We determined the prevalence of pulmonary mycobacterial infection, risk factors
associated with its occurrence among livestock workers
in Ibadan, south-western Nigeria.
Methods: We conducted a cross-sectional study among
livestock workers between August, 2014 and March
2015. Sputum samples were cultured and subjected to a
two-step multiplex PCR technique based on genus typing
and genomic regions of difference respectively. An
interviewer-administered questionnaire was also used to
assess their TB related knowledge, attitude and practices.
Results: Overall, 184 livestock workers (traders 114;
butchers 70) were screened. Majority (81.5%) were
males, aged between 21- 49 years, while over a third
(38.6%) had no formal education and 46.7% did not
have more than 10 years work experience. Though
majority (80%) were knowledgeable about TB, most
(95%) had poor TB related attitude and practices. In all,
6.1% (7/114) of the livestock traders and 7.1% (5/70) of
the butchers had positive cultures. Molecular characterization techniques identified six of the seven isolates from
the traders as non-tuberculous mycobacteria (NTM)
whereas from the butchers, four Mycobacterium tuberculosis and one NTM were identified. Importantly, age
and educational status were significantly associated with
TB infection among the workers (P , 0.05).
Conclusion: We isolated and confirmed M. tuberculosis
(and not M. bovis) as mycobacterial species responsible
for pulmonary TB infections in livestock workers
(though M. bovis has earlier been reported among this
group in the area). Given poor attitude and practices of
these workers towards TB, coupled with their direct
contact with cattle in farms and markets, health threats
at this interface may promote zooanthroponotic transmission of M. tuberculosis and its emerging multi-drug
resistant strains to cattle in the study area. We therefore
reiterate the importance of molecular typing methods in
the identification of mycobacteria in infected humans
and cattle and also advocate multidisciplinary approaches of One Health to mitigate public health problems
associated with TB in Africa.
Background: Mycobacterium bovis is intrinsically resistant to pyrazinamide (PZA) due to a single nucleotide
polymorphism of the pncA gene. PZA offers sterilizing
activity on semidormant bacilli and, when used in
combination with rifampin and isoniazid, shortens the
duration of tuberculosis treatment from 9 months to 6
months. Guidelines recommend extending treatment of
tuberculosis due to PZA resistant organisms (including
M. bovis) from 6 to 9 months; however, the evidence base
used to develop treatment guidelines have not distinguished M. bovis from M. tuberculosis disease. We
compare the time to sputum culture conversion between
M. bovis and M. tuberculosis cases.
Design/Methods: This retrospective cohort analysis
included all culture-positive tuberculosis cases with
spoligotyping and 24-locus MIRU-VNTR results among
patients who received standard 4-drug treatment at the
time of diagnosis in the USA during 20062011. We
excluded from analysis cases with initial or acquired
resistance to rifampin or isoniazid and cases who were
dead at diagnosis or died during the course of treatment.
The time to sputum culture conversion was estimated
with Kaplan-Meier curves using treatment start date and
date of first consistently culture-negative sputum. We
used Cox proportional hazard modeling to calculate
adjusted hazard ratios (aHR) and 95% confidence
intervals (95%CI) and identify factors associated with
time to culture conversion.
Results: A total of 26 419 cases, 195 (0.7%) M. bovis and

26 224 (99.3%) M. tuberculosis cases were included in
the study. At two months of treatment (the point at which
the decision to extend treatment is made), 78% of M.
bovis and 65% of M. tuberculosis cases converted
sputum culture to negative (Figure). M. bovis cases
converted sputum-cultures to negative 40% faster than
M. tuberculosis cases (aHR 1.4, 95%CI 1.21.6).

Directly observed therapy (aHR 1.1, 95%CI 1.0-1.1)
and provider type (aHR 1.1, 95%CI 1.06-1.13) did not
substantially impact conversion rates but were statistically significant.
Conclusion: In this population-based, cohort study, M.
bovis cases converted sputum-cultures to negative sooner
than M. tuberculosis cases. The loss of PZA among M.
bovis cases does not appear to lengthen time to culture
conversion among those who received standard 4-drug
treatment at diagnosis. Randomized controlled trials
may be needed to confirm that M. bovis disease can be
treated in 6 months.
EP-106-04 New diagnostic tests for bovine

tuberculosis
S-H Wang,1 J M Balada-Llasat,1 W G Hunt,2
G Gebreyes,1 J Torrelles,1 P Jeff1 1Ohio State University,
Columbus, OH, 2Nationwide Childrens Hospital, Columbus,
OH, USA. Fax: (1) 614 293 4556. e-mail:
shu-hua.wang@osumc.edu
Background: Bovine Tuberculosis (TB) is one of the

World Health Organizations seven neglected endemic
zoonotic diseases. Currently, there is no a rapid-point-ofcare, easy-to-use diagnostic test for bovine TB. We
proposed to improve and implement a lipoarabinomannan antigen-antibody (LAM-Ag-Ab) test and develop a pyrazinamide (PZA) drug susceptibility culture test,
where PZA resistance is a surrogate marker for
Mycobacterium bovis infection.
Design/Methods: In collaboration with Michigan Department of Agriculture and Rural Development, we
collected urine from cattle suspected of bovine-TB
disease in Michigan (US), and conducted the LAM-AgAb test in urine and the PZA culture.
Results: Our results showed that 29 of 35 cows suspected
of bovine TB tested positive by the urine LAM-Ag-Ab
test. These results were obtained in only 25 minutes,
without the requirement of specialized personnel or
equipment. Using the US Department of Agriculture
bovine TB detection procedures, 12 of 21 animals (57%)
were positive by interferon gamma release assay (IFN-c)
blood test, 18 of 21 animals (86%) were positive by gross
pathology and histopathology, and 19 of 21 animals
(90%) were positive for M. bovis culture. When
compared to IFN-c, the LAM-Ag-Ab test showed higher
sensitivity (33.3% more accurate). In some cases the
LAM-Ag-Ab vs. visual pathology inspection or culture
results were discordant. That is, the LAM-Ag-Ab was
positive in 14.3 % (3 of 21 animals) more cases than
pathology but this test gave LAM negative (infection
indicator) in comparison to culture in 9.5% of the
specimens (2 of 21 animals). The evaluation of the PZA
culture is ongoing, however, using M. bovis spiked urine
in the lab setting, this test was optimized. Specifically, we
validated detection and differentiation of M. tuberculosis
complex infections and determination of PZA drugresistance in 14 days.
Conclusion: Urine LAM-Ag-Ab and PZA culture tests
offer the potential to diagnose M. bovis infection in
S87
cattle in a more rapid and less expensive fashion than

current methods. We anticipate that processing the
urine using quick and easy biochemical procedures
will further increase the sensitivity of the LAM-Ag-Ab
test.
EP-107-04 Identification of environmental nontuberculous mycobacteria of public health

importance in cattle in Zimbabwe
N Chinombe,1 L Padya,2 M Magwenzi,1 J Mbanga,2
P Nziramasanga1 1University of Zimbabwe, Harare,
2
National University of Science and Technology, Bulawayo,
Zimbabwe. e-mail: nyasha.chinombe@gmail.com
Background: The genus Mycobacterium is made up of

highly heterogeneous bacteria that are found ubiquitously in the environment and in animals such as cattle.
To date, more than 100 Mycobacterium species have
been identified. Some of the Mycobacterium species are
now known to be opportunistic pathogens in humans.
Rapid laboratory identification of these bacteria at
species level is therefore critical if infections are to be
controlled. Since cattle live in close proximity with
people, we set out to identify nontuberculous Mycobacterium species of public health importance in cattle
in Zimbabwe.
Design/Methods: Cattle cowdung samples were collected throughout Zimbabwe and mycobacteria were
isolated using culture method. DNA was isolated from
the cultures and 16S ribosomal RNA gene of Mycobacterium was amplified by polymerase reaction. The
amplicons were sequenced. Bioinformatics analyses of
the sequences were used to identify the mycobacteria at
species level.
Results: A total of 134 cowdung samples were collected
throughout Zimbabwe. Out of these, 49 samples were
positive for suspected mycobacteria by acid fast test.
PCR products were sequenced in 26 samples that
amplified well. Analysis of the sequences showed that 7
(27%) belonged to Mycobacterium neoaurum, 6 (23%)
to Mycobacterium fortuitum, 3 (12%) to Mycobacterium goodii, 2 (1%) to Mycobacterium arupense, 2 (1%)
to Mycobacterium septicum/peregrinum and 1 isolate
(0.04%) to Mycobacterium elephantis, There were 5
(19%) isolates that were non-mycobacteria.
Conclusion: Several species of genus Mycobacterium that
may cause human diseases were isolated and identified
by 16S rDNA sequencing in Zimbabwe. The study
therefore provides a molecular basis for species identification of NTM species in animals and humans.
S88
EP-108-04 Human tuberculosis due to

Mycobacterium bovis in England, Wales and
Northern Ireland: a 12-year national cohort
analysis of the epidemiology
J Davidson,1 M Lalor,1 L Thomas,1 I Abubakar,1,2
D Zenner1 1Public Health England, London, 2University
College London, London, UK. e-mail:
jennifer.davidson@phe.gov.uk
Background: Public and political interest in the control

of bovine tuberculosis (TB) in animals is high in the UK,
with high profile interventions including badger and
infected cattle culls aiming to prevent transmission to
humans. We described the epidemiology of human M.
bovis in the UK, identified factors associated with human
M. bovis disease compared to M. tuberculosis, and
quantified the occurrence of animal contact and unpasteurised milk consumption in M. bovis cases in England,
Wales and Northern Ireland (EW&NI) to inform public
health policy.
Design/Methods: We conducted a retrospective cohort
study including all culture confirmed TB cases speciated
as M.bovis between 2002 and 2013 in EW & NI.
Additionally we carried out a multivariable case-case
analysis comparing M. bovis cases to M. tuberculosis
cases to identify factors that distinguish M.bovis cases.
Results: Between 2002 and 2013, there was 318 M. bovis
TB cases, of which 90.6% (288/318) were clinically
notified. 50.6% (161/318) of cases were aged over 65
years. Of notified cases, 65.6% (189/288) were UK born,
with 60.8% (155) aged over 65 years. The majority
(66.2%; 45/68) of non-UK born cases were aged between
15 and 44 years. On average, 27 M. bovis cases were
identified annually compared to 4940 M. tuberculosis
cases. Being 65 years or older, UK born or long-term UK
resident, of white ethnicity and not having a social risk
factor were independently associated with M. bovis in
multivariable analysis when compared to M. tuberculosis
cases. From additional exposure information collected
for M. bovis cases, the most frequent (41.7%; 86/206)
exposure was consumption of unpasteurised milk, with
the majority (79.6; 43/54) having consumed more than
10 years before diagnosis. 20.4% (42/206) of cases had
work related animal exposures, 14 of which had contact
with M. bovis infected animals.
Conclusion: There is a low number of M. bovis cases in
EW&NI, accounting for a low proportion of TB cases.
Compared to M. tuberculosis cases, M. bovis cases were
more likely to occur amongst older white UK born
individuals, likely to be attributed to reactivation of
infection acquired before widespread milk pasteurisation. A low number of new infections occur mainly
amongst individuals with M. bovis related exposures and
those likely to have acquired the infection abroad. The
findings suggest there is a low risk of acquisition of M.
bovis in humans in EW&NI, with effective control
measures to prevent the zoonotic spread in place.
02. TB in children: perspectives from

around the world
EP-109-04 Accuracy of diagnostic tests for
childhood pulmonary tuberculosis: estimation
via Bayesian latent class analysis
S Schumacher,1 M Van Smeden,2 N Dendukuri,1
M Nicol,3 M Pai,1 H Zar4 1McGill University, Montreal, QC,
Canada; 2Julius Center for Health Sciences and Primary Care,
University Medical Center Utrecht, Utrecht, Netherlands;
3
University of Cape Town and National Health Laboratory
Service, Cape Town, 4Dept of Paediatrics & Child Health, Red
Cross Childrens Hospital and MRC Unit on Child &
Adolescent Health, University of Cape Town, Cape Town,
South Africa. e-mail: samuel.schumacher@mail.mcgill.ca
Background: The absence of a gold standard diagnostic

test for childhood PTB complicates the assessment of
novel tests and also makes diagnosing individual patients
and estimating population prevalence difficult. Latent
Class Analysis is a statistical modeling technique that
allows simultaneous estimation of test accuracy and
disease prevalence in the absence of a gold standard.
Design/Methods: We analyzed data from a prospective
cohort study of 910 hospitalized children with suspected
PTB in South Africa in whom MGIT, smear microscopy
and Xpert were done on respiratory specimens and TST
and Chest X-ray had been performed. We fitted a
hierarchical Bayesian latent class model based on how
observed test results and patient characteristics relate to
the unobserved true disease status (TB or not TB). We
incorporated both predictors of the pre-test probability
and covariates modifying test accuracy in our model.
Conditional dependence between microbiological tests
was modeled via a random effect, representing unobserved bacillary burden.
Results: Median age was 25 months (IQR 13-59), 22%
were HIV infected, 28% were malnourished (MN, based
on weight for age Z-score ,-2) and 57% were started on
tuberculosis treatment. We estimated sensitivity of
MGIT, Xpert and smear respectively as 52% (95%CrI
40-68), 44% (95%CrI 34-57) and 18% (95%CrI 13-25)
with specificities above 98% for all three tests. X-ray was
estimated to have sensitivity of 61% (95%CrI 52-70)
and specificity of 80% (95%CrI 75-85). TST sensitivity
decreased with increasing immunosuppression from
80% (HIV-MN-) to 69% (HIV-MN) to 47%
(HIVMN-) to 34% (HIVMN). TB prevalence was
estimated to be 32% (95%CrI 24-41) overall, 48%
(95%CrI 38-59) among the treated and 10% (95%CrI 418) among the untreated. In comparison, using classification based on a clinical case definition, 18% of
children had definite TB (microbiologically confirmed),
34% had no TB (not clinically diagnosed and improved
without treatment on follow up) and 48% had possible
TB (all others).
Conclusion: Bayesian Latent Class Analysis makes
optimal use of the imperfect information from available
test results and is a valuable modeling framework for
childhood PTB and other diseases when no gold standard
exists. Increased use of Latent Class Analysis should
facilitate development and evaluation of improved novel

tests but also refine our understanding and optimal use of
technologies that are currently available.
Table
Test
Culture
Xpert
Smear
CXR
TST (HIVMaln)
TST (HIVMaln)
TST (HIVMaln)
TST (HIVMaln)
Sensitivity %
(95% CrI)
52
44
18
61
80
69
47
34
(40-68)
(34-57)
(13-25)
(52-70)
(71-87)
(54-81)
(28-68)
(17-55)
Specificity %
(95% CrI)
100
99
100
80
70
(99-100)
(98-100)
(99-100)
(75-85)
(64-77)
CrI credible interval (Bayesian analog of frequentist confidence interval);

CXR chest X-ray; TST tuberculin skin test; HIV human immunodeficiency virus; Maln Malnutrition based on weight for age Z-score ,-2.
EP-110-04 Characteristics and outcomes of

drug-resistant tuberculosis children: a
multicenter retrospective study of Medecins

`
Sans Frontieres
programs
M Bastard,1 A Hayrapetyan,2 N Sapaev,3 T Dlamini,4
S Khurkhumal,5 C Hewison,6 F Varaine,6 M Bonnet1,7
1
Epicentre, Paris, France; 2National Tuberculosis Control
Office, Yerevan, Armenia; 3Deputy Chief Doctor of the
Republican TB Dispensary, Nukus, Uzbekistan; 4Ministry of
Health-TB National Control Program National Manager,
Mbanane, Swaziland; 5Director of the National TB Program,
`
Sukhumi, Abkhazia, Georgia; 6Medecins
Sans Frontieres,
Paris, 7Unite Mixte Internationale UMI233-U1175, Institute of
Research for Development, Montpellier, France. e-mail:
mathieu.bastard@geneva.msf.org
Background: Paediatric drug-resistant tuberculosis (DRTB) remains highly neglected and very little data is
available on the characteristics and outcomes of children
with DR-TB. We present results from a retrospective
study of seven Medecins sans Frontières DR-TB programs.
Methods: Demographic, clinical and bacteriological
characteristics of children (,15 years) at treatment
initiation were described. We also presented preliminary
DR-TB treatment outcomes for patients starting treatment at least two years before censoring date of the
database.
Results: Between 2006 and 2013, 259 children (53.8%
female) started on treatment. Median age was 8 [IQR 412], 72.5% were new cases, 19.1% were contacts of a
multidrug-resistant tuberculosis (MDR-TB) cases and
13% (27/208) were HIV-positive. The main symptoms at
admission were fever (36%) and weight loss (21%).
Majority had pulmonary TB (66.7%) and 13.4% had
cavities on chest X-ray. Of 259 children who started a
DR-TB treatment, only 52 (20.1%) were bacteriologically confirmed (43 were confirmed by culture and 9 by
smear). Drug susceptibility testing among culture positive showed that 25.9% had poly-drug resistant TB and
74.1% had MDR-TB. Among MDR-TB, majority had
no resistance to second line drugs (61.3%), 35.8% had
resistance to injectables, and 3.2% to fluoroquinolones.
Main side-effects experienced during treatment were
S89
gastrointestinal and hepatic. Of 42 children with a final

outcome: 16 cured (38.1%), 21 completed treatment
(50%), 3 died (7.1%), 1 failed treatment and 1 defaulted
from treatment (2.4%).
Conclusions: Our study, showing that 80% of the
children start a DR-TB treatment without confirmed
resistance, reinforces the urgent need of TB diagnostic
and resistance test adapted to children in order to avoid
unnecessarily exposing children to long and potentially
toxic treatment. Preliminary outcome results are encouraging.
EP-111-04 Role of chest X-ray in the diagnosis

of intra-thoracic childhood tuberculosis
S H Syed Usman,1 R Nathavitharana,2 B Velayutham,1
V Chandrasekaran,1 G Sharma,1 D Baskaran,1
S Swaminathan1 1National Institute for Research in
Tuberculosis, Chennai, India; 2Beth Israel Deaconess Medical
Center, Boston, MA, USA. e-mail: drsyed@rediffmail.com
Background and aims: Children with TB are frequently

misdiagnosed (both under and over-diagnosed). In the
absence of bacteriologic confirmation, chest radiograph
plays an important role in the diagnosis of pediatric
tuberculosis (TB). We evaluated the role of chest X-ray in
the initial diagnostic work up.
Patients and methods: Children ,15 years with clinical
signs and symptoms suggestive of TB not responding to
antibiotics were enrolled. Children underwent chestXray, sputum or gastric lavage for AFB smear and
M.tuberculosis culture. The X-rays were reviewed by
two independent radiologists in blinded manner. Further
a panel, consisting of a paediatric pulmonologist,
pulmonologist and general physician, reviewed any
abnormal X-rays and classified asAbnormal suggestive
of TB, Abnormal non-TB and normal.
Results: 813 children were enrolled and given an initial
treatment decision. 9/813 (1.1%) patients had positive
smears, and 26/813 (3.2%) cultures were positive and of
these, 4, 11 and 11 had abnormal X-ray suggestive of TB,
abnormal X-ray and normal X-rays respectively. Of 813
X-rays assessed, 125 (15%) were read as abnormal and
523(64%) as normal by both readers. There was
discordance in 165 (20%) X-rays and kappa was 0.47
indicating moderate level of inter-reader agreement. Of
160 patients with abnormal X-rays determined by panel,
57 (36%) were started on anti-tuberculosis therapy
(ATT), of these 16(28%) were based on bacteriologic
confirmation and 41 (72%) were based on radiologic
and clinical features.
Conclusion: The role of X-ray for the diagnosis of TB in a
paediatric clinic setting is unclear given the range of
abnormalities and lack of specificity. Further research is
warranted to determine the use of other imaging
modalities for childhood TB diagnosis and a comprehensive diagnostic strategy.
S90
EP-112-04 Increasing childhood TB notifications

through systematic screening in public-sector
hospitals in rural Pakistan
A Malik,1,2 F Amanullah,1 S Saleem,1 M Jaswal,1
H Hussain1,2 1The Indus Hospital, Karachi, 2Interactive
Research and Development, Karachi, Pakistan. e-mail:
amyn.malik@irdresearch.org
Background: An estimated 0.5 million children develop

TB each year and account for 6-10% of all TB cases of all
ages. This is likely an underestimate of the true child TB
burden, owing to poor access to diagnostics and
underreporting. Active screening is known to increase
adult TB case notification urban settings.
Design/Methods: Lay screeners were recruited and
trained on using interactive algorithms (Decision Support System) on handheld android devices to screen
children at the pediatric outpatient departments of three
large secondary and tertiary care public hospitals in
Jamshoro district of Sindh, from Quarter 4 of 2014.
Children with one or more symptoms of TB (cough 7 2
weeks or contact history of TB in the last 2 years or a
combination of two of the following: fever 7 2 weeks,
night sweats or weight loss) were referred to a medical
officer who clinically evaluated and ordered diagnostic
tests including chest X-ray, AFB smear, and Xpert MTB/
RIF assay. All tests were offered free of cost to the
children. Diagnosed patients were started on TB treatment and caregivers were asked to bring other children in
the household for screening. Medical officers at participating hospitals were trained in pediatric TB, and
management of complex cases was supervised by a
pediatric TB specialist at The Indus Hospital using free
applications like Skype and WhatsApp.
Results: Screeners assessed 33 505 children between

October 2014 and March 2015, of which 859 were
classified as children with presumptive TB. Of these, 221
children were diagnosed with TB disease and were
started on treatment. This resulted in an over two fold
increase in pediatric TB case detection and notification
across 12 BMUs in the intervention district. In the
control district of Hyderabad there was no change in
pediatric TB case notification. Interestingly of the 221
children started on treatment, 57% were in the 0-4 year
age group. This is contrary to the pattern seen at the

Indus Hospital TB program where the majority of
children with TB belong to the .5 year age group.
Conclusion: Children with TB disease often remain
undiagnosed as symptoms are nonspecific. Higher case
detection in the 0-4 year age group in our rural setting
likely occurred because of the integration of TB screening
with the EPI program, resulting in earlier TB detection in
this vulnerable group. Active screening at pediatric
outpatient departments of rural hospitals can increase
child TB detection and notification many fold.
EP-113-04 Effect of empirical tuberculosis

treatment on mortality of children with clinical
suspicion of intrathoracic tuberculosis
M Bonnet,1,2 M Nansumba,2,3 M Bastard,4 P Orikiriza,2
P De Beaudrap,1,2 Y BoumIi,2,3 J Kiwanuka,3,5
E Kumbakumba3,5 1Institut de Recherche pour le
Developpement
(IRD), Montpellier, France; 2Epicentre,
Mbarara, 3Mbarara University of Science and Technology,
Mbarara, Uganda; 4Epicentre, Paris, France; 5Mbarara
Regional Reference Hospital, Mbarara, Uganda. e-mail:
maryline.bonnet@geneva.msf.org
Background: Use of empirical tuberculosis (TB) treatment is common in high burden countries due to lack of
effective TB diagnostic tests. We assessed the effect of
empirical TB treatment on mortality of children with
clinical suspicion of intrathoracic TB in Uganda.
Methods: Children ,14 years old with at least one of the
following signs (unexplained weight loss, persistent
respiratory symptoms, fever, fatigue or reduced playfulness) underwent clinical investigation, chest X-ray,
tuberculin skin test (TST) and sputum mycobacterial tests
(microscopy, XpertMTB/RIF, culture) and were followed
for 3 months. Decision to initiate treatment was based on
mycobacterial, clinical and radiological findings. The
effect of empirical treatment on mortality was adjusted on
age, sex, malnutrition (weight for height z-score), HIV
infection, TB contact history, TST results (5mm cut-off).
Results: 392 children were included: median (interquartile range, IQR) age of 4 years [1.4, 7.5], 45% female,
31.2% HIV infected and 20.5% with moderate to severe
malnutrition. 144 (36.7%) children were started on
treatment (18 with confirmed TB and 126 empirically).
At 3 months of follow-up, 25/392 (6.4%) children died:
15/144 (10.4%) on TB treatment vs. 10/248 (4.0%) not
on treatment (P 0.005). Median (IQR) time to death
was 20 days (8, 75): 13 for children on TB treatment vs.
21 for non-treated ones. There was significantly more
deaths in the following groups: ,2 years old (12%, 15/
125) vs. older ones (3.7%, 10/267), moderately to
severely malnourished (15.2%, 12/79) vs. others
(3.6%, 11/306), empirical treatment (11.9%, 15/126)
vs. confirmed TB (0%, 0/18), negative (6.7%, 19/284)
vs. positive TST result (1.0%, 1/99) and no TB contact
history (7.6%, 24/315) vs. exposure (1.3%, 1/76). After
multivariate analysis, empirical TB treatment (aOR 4.1,
95%CI 1.6-10.5) and negative TST result (aOR 12.5,
95%CI 4.7-26.3) remained associated with death.
S91
Conclusion: The results suggest that children on empirical TB treatment are more likely to die than confirmed
TB cases or non-treated cases. This could be explained by
the difficulty to diagnose TB in children, the increased
likelihood of starting empirical treatment in children
with severe clinical condition or comorbidities and the
possibility of misdiagnosis of other severe infections.
They highlight the urgent needs for more effective TB
diagnosis tests for children presenting with severe clinical
conditions.
include high rates of empirical treatment in children, as

well high severity of illness leading to lower treatment
initiation thresholds. In such settings, GXP availability
may not have a large impact on clinicians decision
making. More sensitive pediatric POC testing with lower
limits of detection may eventually influence clinician
behaviour, leading to less empiric treatment and improved resource allocation.
Table Characteristics of children diagnosed with TB disease
based on GXP testing/results
Patients diagnosed with TB disease
EP-114-04 Impact of XpertW sputum

diagnostics in children with TB disease on
treatment decision in a high-burden setting
1
J Bacha, J Benjamin, L Campbell, P Clowes,

C Mangu,2 A Dinardo,3 K Ngo,3 A Mandalakas3 1Baylor
College of Medicine Childrens Foundation - Tanzania,
Mbeya, 2National Institute of Medical Research - Mbeya
Medical Research Centre (NIMR-MMRC), Mbeya, Tanzania;
3
Baylor College of Medicine, Houston, TX, USA. e-mail:
bacha@bcm.edu
Background: 2014 WHO policy update recommends

GeneXpert (GXP) as the initial diagnostic test in children
suspected of having TB, leading to countries working on
improving access to GXP. However, GXP may have
limited impact in children due to pauci-bacilliary, culture
negative nature of childhood TB and use of clinical
diagnosis in children. The added diagnostic value of GXP
in children in resource limited settings remains unclear
and needs investigation.
Design/Methods: Retrospective chart review between
March 2013 and December 2014 of patients seen at the
Baylor Tanzania Centre of Excellence (COE) in Mbeya,
Tanzania. Data on children referred for presumptive TB
and diagnosed with TB were captured using a standardized data collection tool and unique data base.
Results: Of 504 children (57% HIV-infected) referred for
presumptive TB, 35% (176/504) were diagnosed with
TB. 70% (355/504) completed GXP testing as part of
their evaluation. GXP was completed by 69% (226/328)
of children where TB was excluded and 73% (129/176)
of children diagnosed with TB. Among children not
diagnosed with TB, GXP was negative in .99% (327/
328) and not resulted in 1 child. Of those diagnosed with
TB disease, GXP was positive in 9% (12/129), negative
in 90% (116/129) and not resulted in 1% (1/129).
Median time to antituberculosis treatment (ATT) among
children with TB disease was 3 days (0-142), and did not
vary widely between those with GXP not done (2 days, 025d), GXP positive (3 days, 0-28d), or GXP negative (4
days, 0-142d). Older children were more likely to have a
positive GXP while younger children were more likely to
not complete GXP (ANOVA, P 0.001; Table). Hospitalized children were less likely to have done GXP than
outpatient referrals (P 0.003). There was no association
between HIV status or presence of severe acute malnutrition and GXP result/use (P 0.431 and P 0.306).
Conclusion: In our setting, clinicians were less likely to
use GXP in hospitalized or younger children and use of
GXP testing did not reduce time to ATT. Likely reasons
Characteristics
GXP
GXP
GXP
not
positive negative
done
(n 12) (n 116) (n 47)
Median age
(years, range)
12.0
6.2
3.1
years
years
years
(2.3
(0.6
(0.3
19.0)
18.3)
18.3)
HIV-positive (%) 6
69 (60%) 25
(50%)
(53%)
HIV-exposed
0 (0%) 9 (8%)
2 (4%)
(%)
Inpatient status
5
41 (35%) 31
(%)
(42%)
(66%)
11.68
0.003
(2)
SAM (%)
6
31 (27%) 16
(50%)
(34%)
2.37
0.306
(2)
Mean time
4.5
13.5 (0
3.2 (0
to ATT
(028)
142)
25)
(days, range)
X2
(df)
9.37
(2)
P value
*
0.001
0.431
5.09*
(2)
0.007
* (df) ANOVA F-value);
Fishers exact test was used for analyses involving cell sizes of 5 or fewer
individuals.
TB tuberculosis; GXP GeneXpert; HIV human immunodeficiency virus;
SAM severe acute malnutrition; df degrees of freedom; ATT antituberculosis treatment.
EP-115-04 Long-term safety, tolerability, and

pharmacokinetics of delamanid in children
aged 1217 years
J Hafkin,1 M Gler,2 M Frias,3 A De Leon,4 N Hittel,5
L Geiter,1 C Wells,1 S Mallikaarjun1 1Otsuka
Pharmaceutical Development and Commercialization,
Rockville, MD, USA; 2Otsuka Manila Research Center, Makati
City, 3De La Salle Health Sciences Institute, Cavite, 4Lung
Center of the Philippines, Quezon City, Philippines; 5Otsuka
Novel Products Germany, Munich, Germany. e-mail:
jeffrey.hafkin@otsuka-us.com
Background: Limited data exists with which to guide the

treatment of multidrug-resistant tuberculosis (MDR-TB)
in children. Delamanid, a novel anti-TB agent with
bactericidal activity, has demonstrated efficacy in previous clinical trials of MDR-TB in adults. The aim of this
current trial is to assess the long-term safety, tolerability,
and pharmacokinetics of delamanid in children (ages 1217 years) with MDR-TB.
Design/Methods: A Phase 2, open-label, uncontrolled,
trial of delamanid in children with MDR-TB aged 12-17
S92
was performed at De La Salle Health Sciences Institute

and Lung Center of the Philippines. Delamanid 100mg
BID was administered with an optimized background
regimen (OBR) for six months followed by OBR alone in
accordance with WHO guidelines. Clinical and safety
assessments were performed approximately every 2-4
weeks for up to one year; sparse PK sampling for
concentrations of delamanid and DM-6705 (a key
delamanid metabolite) was performed on days 1, 14,
56, 98, 154, 182, 189, 196, 203, 210, and 238.
Results: Nine patients were screened and seven enrolled
in the trial. Six (86%) had confirmed MDR-TB, and one
(14%) had probable MDR-TB. All seven (100%) had
pulmonary disease, and six (86%) had previously
received treatment for TB. The median age was 15 years
(13-17); four were male, three female; the median weight
and BMI were 38.5 (26-45) kg and 15.9 kg/m2 (15-20),
respectively. No patients discontinued trial participation
or follow-up prior to the end of the trial; however, one
patient had delamanid withdrawn permanently due to
nonadherence and subsequent pregnancy. No serious
adverse events were reported; the most common adverse
event reported was headache, occurring in five patients
(71.4%). No patient experienced an absolute QTcF
.500ms. Over the course of delamanid treatment, the
median Cmax ranged from 354 to 657ng/mL. Compared
to adults, median delamanid concentrations were higher
in childrenthough within the range seen in adult
clinical trials. Median DM-6705 levels, however, remained lower in children than in adults.
Conclusion: Long-term exposure to delamanid 100 mg
BID was safe and well-tolerated by this cohort of
children (ages 12-17 years) with MDR-TB. Delamanid
plasma concentrations in this age group were within the
range of those seen in adult clinical trials, suggesting that
the current standard adult dosing is adequate for this age
group. Additional data collection activities are underway
to assess the same parameters in children ,12 years with
MDR-TB.
EP-116-04 Poor bacteriological yield of

alternative sampling strategies for paediatric
tuberculosis at a primary care clinic in
Johannesburg, South Africa
C Hanrahan,1 L Mutunga,2 J Bassett,2 H France,2 S Omar,3
N Ismail,3 H Dansey,2 A Van Rie4,5 1Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, USA;
2
Witkoppen Health and Welfare Centre, Johannesburg,
3
National Institute for Communicable Diseases,
Johannesburg, South Africa; 4University of Antwerp,
Antwerp, Belgium; 5University of North Carolina School of
Public Health, Chapel Hill, NC, USA. e-mail:
chanraha@jhsph.edu
Background: Diagnosis of paediatric tuberculosis (TB) is

challenging due to paucibacillary disease and difficulty in
collecting sputum. Hospital studies have demonstrated
the utility of alternative sampling strategies such as
nasopharyngeal aspirate (NPA) and new diagnostics such
as Xpert MTB/RIF (Xpert). We aimed to determine the
optimal sampling strategy and diagnostics at primary

care, where children present with less severe disease.
Design/Methods: We enrolled a prospective cohort of
children (age 6 weeks - 10 years) with signs and
symptoms of TB at Witkoppen Health and Welfare
Centre, a primary care clinic in Johannesburg, South
Africa. Tuberculin skin testing (TST) and chest X-ray
(CXR) were performed in all. In those unable to
expectorate, one induced sputum (IS), one gastric
aspirate (GA) and two NPAs were collected. Stool was
collected from all children. Samples were processed at
the National TB Reference Laboratory for smear
microscopy, liquid culture and Xpert. Determine TB
LAM Ag (urine LAM) was done on-site for all HIVinfected children. All children were followed up clinically
for 2 months. TB disease status was classified using
consensus clinical case definitions.
Results: From July 2013-December 2014 we enrolled 119
children. Median age was 21 months (IQR: 12-42), and
18% were HIV-infected. About half (n 69, 58%) were
contacts of known TB cases. TST was positive in 24%,
and 55% had a CXR suggestive of TB. Spontaneous
sputum was successfully collected in 12%. Among the
105 unable to expectorate, one IS was successfully
collected in 98%, NPA in 96%, and GA in 58%. Stool
was collected in 96% of all children, and urine in 62% of
HIV-infected children. A physician initiated TB treatment
in 49 (41%) children; only 2 (4%) had confirmed TB, 46
(94%) probable or possible TB and 1 (2%) unlikely TB.
Among the 70 children not treated, 36 (51%) had
probable or possible TB, and 34 (49%) unlikely or not
TB. Of all 469 samples collected, smear microscopy was
positive in 0%, Xpert in ,1% (n 4) and culture in ,1%
(n 2). NPA had the highest yield (3/101 patients, 3%).
Among HIV-infected children with probable or possible
TB, 2/11 (18%) were positive by urine LAM, while 1/2
(50%) were positive in those with unlikely or not TB.
Conclusion: At primary care, alternative sampling
strategies proved feasible but resulted in a low diagnostic
yield with available tools. Extensive efforts to bacteriologically diagnose children did not contribute to clinical
management, as almost all children started treatment
based on clinical findings.
EP-117-04 Experiences of care givers for

children with tuberculosis in and around
Gaborone, Botswana
C Stillson,1 H Okatch,1,2 R Frasso,1,3 L Mazhani,4 T David,4
T Arscott-Mills,1,4,5,6 L Ntshimane,5 A Steenhoff1,4,5,6
1
Perelman School of Medicine, University of Pennsylvania,
Philadelphia, PA, USA; 2University of Botswana, Gaborone,
Botswana; 3Center for Public Health Initiatives, University of
Pennsylvania, Philadelphia, PA, USA; 4Department of
Paediatric & Adolescent Medicine, Faculty of Health Sciences,
University of Botswana School of Medicine, Gaborone,
5
Botswana-U Penn Partnership, Gaborone, Botswana;
6
Childrens Hospital of Philadelphia, Philadelphia, PA, USA.
e-mail: christian.stillson@gmail.com
Background: The length and complexity of treatment for

tuberculosis (TB) poses challenges for children who are
too young to manage their own care. This leaves the

responsibility for treatment delivery on their caregivers.
The knowledge, attitudes and beliefs of these caregivers
towards TB likely have a direct effect on disease
outcome. Previous studies have noted low health literacy
in caregivers for children with TB. We sought to
investigate the lived experience of caregivers for children
on treatment for TB, and to explore caregivers
experiences with health care staff, perceived challenges,
unmet needs and barriers to successful treatment in
Gaborone, Botswana and its surrounding districts.
Methods: Semi-structured interviews were conducted
with 28 caregivers of children currently on anti TB
treatment. Interviews were guided by a piloted interview
instrument and conducted in Setswana or English at the
interviewees discretion. Audiotapes of interviews were
transcribed and translated into English where necessary.
Transcripts were then uploaded into the qualitative
analysis software NVivo 10 (QSR International) to
facilitate a systematic thematic analysis. A coding
schema was developed from the transcripts by two
investigators versed in local culture and context. All
transcripts were then double coded and analyzed for
inter-coder reliability.
Results: Respondents reported both positive and negative
experiences as caregivers. Many reported barriers to
timely diagnosis and effective TB treatment for their
children. These included misdiagnoses before beginning
TB treatment, low levels of TB health education at the
time of diagnosis, inconsistent access to medication at
their point of care, and difficulty administering medication to the child. Respondents also reported significant
relief in seeing improvements in their childs condition as
TB treatment progressed.
Conclusions: Caregivers identified multiple barriers in
timely diagnosis and treatment of pediatric TB and in
their knowledge of TB. The barriers identified by
caregivers in administering tablets highlight the need
for child-friendly drug formulations. Efforts to improve
TB outcomes in this and similar settings should be
focused on improving diagnosis, consistent provision of
quality education, and ensuring the availability of childfriendly formulations of TB drugs.
S93
03. Drug-resistant TB: outcomes - I

EP-118-04 Effectiveness of individualized
regimen for treatment of isoniazidmonoresistant tuberculosis: a multicenter
`
study of Medecins
Sans Frontieres
programmes
M Bastard,1 E Sanchez,1 P Du Cros,2 A Telnov,3
C Hewison,4 H Khamraev Atadjan,5 M Bonnet,1,6
F Varaine4 1Epicentre, Paris, France; 2Medecins

Sans
`
`
Sans Frontieres
(MSF),
Frontieres,
London, UK; 3Medecins
Geneva, Switzerland; 4MSF, Paris, France; 5Teaching Assistant
of the Department of Public Health Administration, Nukus
Branch of Tashkent Pediatric Medical Institute, Nukus,
Uzbekistan; 6Unite Mixte Internationale UMI233-U1175,
Institute of Research for Development, Montpellier, France.
e-mail: mathieu.bastard@geneva.msf.org
Background: Although data exist for multi-drug resistant

TB (MDR-TB) patients, very little is available on other
resistant forms of TB, such as poly-drug resistant TB, and
particularly those resistant only to isoniazid or to
isoniazid and streptomycin (H(S) resistant).
Methods: We retrospectively analysed data from seven
DR-TB programs. Confirmed H(S) resistant patients
starting treatment before July 2011 were included in the
study. Patients received an adapted treatment regimen
with first or second line drugs according to their clinical
status and previous DRTB exposure. Characteristics at
treatment initiation, extension to MDR-TB during
treatment and treatment outcome were described. Coxproportional hazards model was fitted to explore risk
factors of unsuccessful treatment outcome.
Results: A total of 310 H(S) resistant patients started on
treatment: 65.8% female, median age of 36 [IQR 26-48],
43.1% were new cases, 74.7% had cavities on chest Xray and 25.9% had a BMI , 18.5 kg/m2 and 9.5% were
HIV-positive. Smear was positive in 68.4% and streptomycin (S)-resistance was found in 76.2% of patients. At
initiation, 30.6% of patients received first line drugs only
and 69.4% received second line drugs (injectables and
fluoroquinolones). Extension to MDR-TB during treatment was found in 37/310 (11.9%) patients and did not
differ by regimen received. Among them, 10/37 (27.0%)
extended also resistance to injectables and 5/37 (13.5%)
to fluoroquinolones. Treatment outcomes were: 30.3%
cured, 41.6% completed treatment, 2.6% died, 13.9%
failed treatment and 11.6% defaulted from treatment.
Successful outcome did not differ according to Sresistance (P 0.314) and prescription of second line
drugs compared to first line drugs (P 0.15). Among
treatment failures, 86.1% became MDR-TB, 23.3%
became resistance to injectables and 11.6% became
resistant to fluoroquinolones. In multivariate analysis,
being 735 years (aHR2.21, 95%CI 1.26-3.87) and a
BMI,18.5 kg/m2 (aHR2.84, 95%CI 1.70-4.75) were
independently associated with an unsuccessful treatment
outcome.
Conclusions: Despite treatment adaptation, the proportion of patients who failed treatment is particularly high
compared to what is generally observed among full
S94
susceptible TB. Development of resistance to rifampicin,

injectables and fluoroquinolones are of particular concern and underlines the need for close monitoring and for
revising the treatment recommendations for H(S) resistant patients.
EP-119-04 Interim treatment outcomes in drugresistant TB patients who were offered secondline drug susceptibility testing at point of entry
in Delhi, India
started on treatment in April-June 2014 (when compared

with same cohort of patients previous year).
Conclusions and key recommendations: Early second
line DST with added benefit of an individualized DST
guided DR-TB regimen, engenders successful treatment
outcomes among DR-TB and XDR-TB patients. This
emphasizes the need to fast track the scale up of SL DST
facilities and implement guidelines for DST guided
individualized regimen within the standard program
framework across high burden settings.
A Khanna,1 S Chandra,2 A Bhatnagar,3 H Mahmud,4

N Singla,5 B Neeti1 1Directorate of Health Services, State TB
Office, New Delhi, 2World Health Organization, Country
Office for India, New Delhi, 3Rajan Babu Institute of
Pulmonary Medicine and Tuberculosis, New Delhi, 4New
Delhi TB Centre, New Delhi, 5National Institute of TB and
Respiratory Diseases, New Delhi, India. e-mail:
chandras@rntcp.org
D Mibei,1 J Kiarie,1 M Kamene,1 A Wairia2 1Kenya Ministry

of Health, Nairobi, 2TB ARC, Center for Health Solutions,
Nairobi, Kenya. e-mail: dorothymibei@gmail.com
Background and challenges to implementation: Early

detection of second line drug-resistant TB (DR-TB) with
subsequent individualized treatment regimen is the
mainstay for successful treatment outcomes in DR-TB
patients. Delhi State with its load of over 6000 DR-TB
patients on treatment since 2009, took the initiative of
second line drug sensitivity test (SL DST) for all the
multidrug resistant/Rifampicin resistant patients diagnosed at point of entry. Accordingly, framework for
laboratory and treatment center scale up was planned for
a State wide roll out of base line SL DST to all the
diagnosed DR-TB patients from April 2014. Our aim is
to assess the interim treatment outcomes of those
patients who were enrolled for early SL DST since April
2014-June 2014 in Delhi.
Intervention or response: Facilities for second line drug
sensitivity test (SL DST) were scaled up in the city across
the three culture DST labs. Sample of all patients who
were diagnosed as DR-TB starting April 2014 was tested
for SL DST at the three certified labs. This is in contrast
to the earlier guideline of testing only follow up positives
of a DR TB treatment regimen for SL DST. Based on
CDST results obtained for Ofloxacin and Kannamycin
drugs, the treatment regimen was individualized in
accordance with the program guidelines.
Results and lessons learnt: With the introduction of SL
DST for all the diagnosed MDR-TB cases, the first cohort
of patients registered in April-June 2014, had their
interim culture conversions reported after 9 months of
standardized treatment with DR-TB and XDR-TB
regimes. In the study period, 22% baseline Ofloxacin
resistant TB cases were diagnosed among the samples
processed for culture. Of the 406 DR TB patients
enrolled for treatment, there were 264 culture conversions (65%) and 24 (7%) deaths (as compared to 55%
culture conversions and 12% deaths with the earlier
protocol having SLDST for follow up positives of DR TB
treatment). Similarly, out of 26 XDR-TB initiated on
treatment in the study period, 8 (31%) have culture
converted during the study period.Significant improvement (P , 0.01) in culture conversions and decline in
death rates was observed for DR-TB and XDR-TB
Background: Successful treatment outcomes among drug

resistant (DR-TB) patients are of public health importance to prevent further spread of drug resistant strains to
the general population. In addition, this will prevent the
development of more resistance which would further
pose more challenges in its management. Treatment of
DR-TB is long and involves the use of more toxic drugs.
There are factors that have been documented to affect
treatment outcomes; nutritional status, resistance pattern
and model of care. This study analyzed treatment
outcomes of DR-TB patients in Kenya and the factors
associated with successful treatment outcomes.
Design/Methods: A retrospective analysis of secondary
data from the National Tuberculosis, Leprosy and Lung
disease program was done. DR-TB data from the
national database (TIBU) for the year 2012 was
downloaded and exported to Excel. Data cleaning was
done and the data was then exported to SPSS version 21
for analysis. Univariate, Bivariate and Multivariate
analysis was done.
Results: A total of 205 DR-TB patients; 185 MDR-TB, 9
monoresistant and 11 PDR, were included in the
analysis. Of these, 169 (82.4%) had a successful
treatment outcome, 18 (9%) died and 18 (9%) were lost
to follow up. Among MDR-TB patients, 154 (83.2%)
had a successful treatment outcome. Factors analyzed to
determine association with treatment outcomes included;
age, gender, facility type, resistance pattern, nutritional
status and support, model of care and HIV status. Only
gender (P 0.04) and HIV status (P 0.02) were found to
be predictors of successful treatment outcomes. Females
were more likely to have successful treatment outcomes
than males (OR 3.86, 95%CI: 1.47, 10.12) and HIV
negative status increased the likelihood of a successful
treatment outcome (OR 3.53, 95%CI: 1.4, 8.9). More
males (120-58.5%) were reported than females. Treatment success rates were higher among females with 76
(89.4%) being treated successfully. All the DRTB
patients were tested for HIV and 152 (74.1%) of them
were HIV negative with 131 (86.2%) registering
successful treatment outcomes. Among the HIV positive
EP-120-04 An analysis of factors associated

with successful treatment outcomes among
drug-resistant tuberculosis patients in Kenya
patients, 38 (71.7%) were successfully treated of whom

35 (71.7%) were on ART.
Conclusion: Successful treatment outcomes were higher
than the WHO target. There is need however to assess
and address the issues among HIV positive patients and
males so as to improve treatment outcomes in these
groups.
EP-121-04 9-month short-course MDR-TB

treatment in HIV- and non-HIV-co-infected
patients in Uzbekistan and Swaziland: interim
outcomes of two prospective studie
E Casas,1 T Gashu,2 J Greig,3 K Keus,2 T Ndlamini,4
H Khamraev Atadjan,5 C Berry,6 P Du Cros3 1Medecins

`
Sans Frontieres
(MSF), Operational Center Amsterdam,
Amsterdam, Netherlands; 2MSF, Operational Center
Amsterdam, Matsapha, Swaziland; 3MSF, Operational Center
Amsterdam, London, UK; 4Ministry of Health, Mbabane,
Swaziland; 5Public Health Administration, Tashkent,
Uzbekistan; 6MSF, Operational Center Amsterdam, Nukus,
Uzbekistan. e-mail: esther.casas@amsterdam.msf.org
Background: The WHO-recommended treatment regimen for multidrug resistant tuberculosis (MDR-TB) is
lengthy, toxic, and has only a 54% success rate. A success
rate of 84% has been reported for a 9-11 month regimen;
evidence for this regimen is lacking in high MDR-TB and
HIV co-infection settings. MSF, with the respective
ministries of health, is using the short-course regimen
in Uzbekistan and Swaziland. We present interim
outcomes of two prospective, observational studies of
the safety and effectiveness of short-course MDR-TB
treatment.
Design/Methods: We analysed outcomes from September 2013-December 2014. All consenting MDR-TB
patients diagnosed using molecular or culture/drug
susceptibility testing were included. Outcomes are
defined according to pre-2013 WHO definitions. Toxicity was documented with Division of AIDS (DAIDS)
grading. Ethics approval: MSF Ethics Review Board
(ERB) and the ERBs of Swaziland and Uzbekistan.
Results: Characteristics of the 105 Uzbekistan patients:
median age 30.1 years (IQR 24.0-43.2), 51 (48.6%)
male, none HIV-positive, and 74 (70.5%) new cases.
Outcomes at analysis: 66 (62.9%) on treatment, 20
(19.0%) cure, 4 (3.8%) treatment complete, 2 (1.9%)
died, 3 (2.9%) treatment failure, and 10 (9.5%) lost to
follow-up. Culture conversion after 4 months of treatment was 89.3% (95%CI 61.6-84.8). Characteristics of
the 57 Swaziland patients: median age 35.0 years (IQR
28.3-43.1), 23 (40.4%) male, 42 (73.7%) HIV-positive
(15 [35.7%] male, median age 35 years [IQR 30-38]),
and 42 (73.7%) new cases. Outcomes at analysis: 39
(68.4%) on treatment, 10 (17.5%) cured, 6 (10.5%)
died, 2 (3.5%) treatment failure, and none lost to followup. Outcomes among HIV patients: 29 (69%) on
treatment, 6 (14%) cured, 6 (14%) died, and 1 (2.4%)
treatment failure. Culture conversion after 4 months of
treatment was 90%. Ssevere adverse events grade 3 and 4
occurred in 5 (9.4%) and 13 (24.5%) patients in
Swaziland and in 8 (7.6%) and 2 (1.9%) patients in
S95
Uzbekistan, respectively, 3 (5.7%) experienced severe

ototoxicity. ECG at 4 weeks: median QTc increase of
16.5 ms (Swaziland) and 21 ms (Uzbekistan); acquired
prolonged QT syndrome in one patient resulted in
treatment failure (Uzbekistan).
Conclusion: The short-course regimen had satisfactory
culture conversion and interim outcomes in high HIV coinfection and drug-resistant populations. Safety, including the first ECG data reported for this regimen was also
satisfactory.
EP-122-04 Why did some patients fail secondline therapy for drug-resistant pulmonary
tuberculosis? A case series in Karachi, Pakistan
N Salahuddin,1 S Butt,1 A Mashhadi,1 S Adnan,1
M Y Memon,1 M Basir,1 H Hussain1 1Indus Hospital,
Karachi, Pakistan. e-mail: naseemsal@hotmail.com
Objective: To define causes of failure of treatment of

Drug Resistant Pulmonary Tuberculosis (DR-PTB) in a
tertiary care hospital in a high burden tuberculosis
country.
Intervention: A retrospective case series study was
conducted at our hospital. All cases diagnosed and
treated as DR-PTB over 57 months between June 2010
and March 2015 were included. Failure of treatment was
defined as: persistent positive cultures beyond 10 months
of treatment; culture conversion from negative to
positive; progressive clinical or radiological deterioration. Demographics, clinical, laboratory, and outcome
data were retrieved from medical records.Progress was
determined clinically and by monthly sputum cultures. A
radiologist interpreted pre treatment chest radiograph
severity.
Results: 722 patients were diagnosed with DR-PTB and
treated under Programmatic Management of Drug
Resistant TB (PMDT). 7 had XDR, 24 MDR and 5
PDR. All patients received quality assured second line
drugs (SLDs) through direct supervision. Group 5 drugs
were added to original SLDs on a failing regimen in 27
(75%); 36 (5%) failed treatment. 50% presented within
1-10 months of their illness, while others presented from
11-48 months before diagnosis and treatment. 44%
waited 1-3 months before SLDs became available. Mean
age was 33 years (5-65), and 41.7% were female. At
enrolment 4 (11%) were treatment nave, 32 (90%) had
received prior treatment as drug sensitive TB (21 Cat1,
11 Cat 2). 4 patients had coexisting diabetes, 8% had
Hepatitis B, 11% Hepatitis C and none had HIV and 3
were smokers. Pre treatment mean hemoglobin was 10.9
gm and average BMI was 17.3. Severe disease was
evident in 12 patients (36%), defined as loss of 25-50%
lung volume, and moderate disease in 15 (46%) defined
as 5-10% loss of lung volume. None discontinued
treatment because of adverse events. Discontinuationof
treatment was by consensus among clinicians. Surgery
was not an option for lack of proper facility. 19% have
died, 41% are alive, and 19% could not be traced by
phone calls.
S96
Conclusion: 5% of all DR-PTB cases failed treatment.

National failure level was 5% in 2011. Presumptive
causes are: low BMI, anemia, pre treatment delay and
baseline severe parenchymal disease. Since 2012 time to
treatment at our PMDT site is considerably shortened
due to early availability of SLDs. Also, since 2013
introduction of GeneXpert has considerably reduced
time to diagnosis and treatment. This will greatly
improve outcome of MDR treatment.
Conclusion: Our data suggest that for MDR-TB patients

there is a trend toward delayed culture conversion in
patients who received more than one month of inappropriate DSTB therapy. Future treatment algorithms will
need to consider whether starting patients on empiric
therapies with delays until DST is available may worsen
short and long term outcomes, and future studies need to
examine whether there is a time threshold before these
inappropriate empiric therapies have an impact.
EP-123-04 Inappropriate treatment of

multidrug-resistant tuberculosis patients:
effects on culture conversion and final
outcomes
EP-124-04 High survival and treatment success

with concurrent MDR-TB and HIV treatment in
KwaZulu-Natal, South Africa
P Chitneni, M Buckley, J Rohr, D Theron, E Kendall,

R M Warren,4 K Jacobson1,4 1Boston University / Boston
Medical Center, Boston, MA, USA; 2Brewelskloof Hospital,
Worcester, South Africa; 3Johns Hopkins Hospital, Baltimore,
MD, USA; 4Stellenbosch University, Tygerberg, South Africa.
e-mail: PoojaChitneni@gmail.com
Background: In many high burden countries, recognition

of multidrug resistant tuberculosis (MDR-TB) occurs
weeks after initial tuberculosis (TB) diagnosis because
patients do not receive drug susceptibility testing (DST)
until they fail drug susceptible (DS) TB therapy and
because culture-based DSTs take weeks to produce
results. While awaiting DST results, many MDR-TB
patients are begun on DSTB therapy. We aimed to assess
whether time on DSTB treatment was associated with
delayed sputum culture conversion and worse final
outcomes in patients with MDR-TB.
Design/Methods: We analyzed a retrospective cohort of
223 patients who began MDR-TB treatment between
2007 and 2010 in a rural TB hospital in the Western
Cape Province, South Africa. We used Cox proportional
hazards modeling from time of starting appropriate
MDR treatment to culture conversion, stratifying patients by whether they received DSTB treatment for less
than vs. greater than one month after having an MDRTB positive sputum culture collected. DSTB treatment
was defined as being on isoniazid, rifampicin, ethambutol, and pyrazinamide. We performed logistic regression
analysis to evaluate the association between time on
DSTB therapy and final treatment outcomes; good
outcomes included cure, treatment completion and
transfer out in good condition, and poor outcomes
included treatment failure, death, default, and transfer in
poor condition.
Results: While on MDR-TB therapy, 202 patients
(90.6%) converted their cultures to negative. Mean time
to culture conversion was 1.57 months. In univariate
analysis, being on inappropriate DSTB treatment for
greater than one month was associated with longer
culture conversion times (adjusted HR0.76 [0.57-1.01]
P 0.056). There was no significant difference in
likelihood of poor treatment outcomes at 24 months
between patients on inappropriate first line treatment for
more than one month compared to those on inappropriate treatment for less than 1 month (unadjusted
OR1.39 [0.82-2.36] P 0.230).
J Brust,1 S Shah,2,3 S Allana,2 T Mthiyane,4 K Mlisana,4,5

P Moodley,4,5 I Master,6 N Gandhi2 1Montefiore Medical
Center & Albert Einstein College of Medicine, Bronx, NY,
2
Rollins School of Public Health, Emory University, Atlanta,
GA, 3U.S. Centers for Disease Control and Prevention,
Atlanta, GA, USA; 4University of KwaZulu-Natal, Durban,
5
National Health Laboratory Services, Durban, 6King Dinuzulu
(George V) Hospital, Durban, South Africa. Fax: (1) 718 561
5165. e-mail: jcb26@columbia.edu
Background: More than 80% of patients with MDR TB

in South Africa are co-infected with HIV. Treatment
outcomes for patients with MDR TB-HIV co-infection
have historically been poor, but most studies either
predated the availability of antiretroviral therapy (ART)
or were performed retrospectively. We prospectively
compared treatment outcomes in MDR TB-HIV coinfected patients on ART with MDR-TB patients without
HIV infection.
Methods: In this observational cohort study, cultureconfirmed MDR TB with and without HIV co-infection
were treated with a standardized South African MDR TB
regimen, including kanamycin, moxifloxacin, pyrazinamide, ethambutol, ethionamide and terizidone. Subjects
were followed until one year after the completion of their
MDR TB treatment. All HIV-infected patients received
ART, using the standard regimens available in South
Africa. Patients were seen monthly to monitor response
to therapy and obtain sputum cultures. CD4 counts and
HIV viral loads were performed every 3 months.
Treatment outcomes were determined using standard
consensus definitions for MDR-TB.
Results: Among the 206 enrolled subjects, 151 were HIVinfected, 131 (64%) were female, and the median age
was 33 years (IQR 26-41). Median time to sputum
culture conversion was 59 (IQR 51-113) days and was
similar among those with and without HIV co-infection.
At MDR-TB treatment initiation, median CD4 count
was 225 (IQR 132-408) cells/mm3 and 68% had an
undetectable viral load. At the time of censoring, subjects
had been followed for a median of 725 days (IQR 497904) and 22 (11%) had died. 110 (53%) had been
declared cured, 20 (10%) had defaulted, 2 (1%) had
failed therapy, and 53 (26%) were still receiving therapy.
Among those who died, 18 were HIV-infected (12% of
HIVve) and 4 were HIV-negative (7% of HIV-ve; P
0.34). Median time to death was 263 days (IQR 74-616).
The median CD4 count at 24-months was 372 cells/mm3
(IQR 239-536), with a median increase of 78 (IQR 29177) cells/mm3 from baseline. 81% had an undetectable
VL after 1 year of concurrent MDR-TB HIV therapy.
Conclusion: Subjects with MDR-TB HIV co-infection
receiving concurrent ART experienced favorable survival
and treatment outcomes similar to a comparable cohort
of HIV-uninfected MDR-TB patients. Although taking
both MDR TB therapy and ART is challenging for
patients due to pill burden and adverse events, the
recommendation to provide concurrent therapy is
supported by prospectively collected data.
EP-125-04 Outcomes for adolescents

undergoing multidrug-resistant tuberculosis
treatment in Lima, Peru
J Furin,1 M Milstein,2 D Tierney,3 C Mitnick,2 P Isaakidis4
1
Case Western Reserve University, Cleveland, OH, 2Harvard
Medical School, Boston, MA, 3Brigham and Womens
`
Hospital, Boston, MA, USA; 4Medecins
Sans Frontieres,
Mumbai, India. e-mail: meredith_milstein@hms.harvard.edu
Background: Multidrug-resistant tuberculosis (MDRTB) is understudied among adolescents, a unique

subpopulation undergoing emotional maturity, physical
development, and transitioning into roles with greater
responsibility and time constraints. This adolescent
period may have social implications, leading to different
health outcomes than adults. We aim to identify whether
MDR-TB treatment outcomes in adolescence differ from
adults and the predictors for these outcomes.
Design/Methods: We conducted a retrospective cohort
study to assess whether adolescence was associated with
MDR-TB treatment outcome. Predictors included demographics, disease severity indicators and comorbidities.
Outcomes were separately assessed for adolescents (1019) and adults (.20). v2 tests, Fishers exact tests, or ttests were used to identify significant differences between
adolescents and adults. Cox proportional hazards
models were used to assess the effects of baseline
covariates on time-to-death.
Results: We found identified 90 adolescents and 577
adults in the cohort. Successful outcomes were observed
in 76% adolescents and 64% adults (P 0.0362).
Further, 11% adolescents and 22% adults died (P
0.0159). Significant differences in baseline characteristics between adolescents and adults included previous
receipt of 2 or less regimens (59% vs. 25%, respectively;
P , 0.0001), bilateral, cavitary findings on chest X-ray
(44% vs. 57%, respectively; P 0.0252), low hematocrit
(57% vs. 86%, respectively; P ,0.0001), and average
number of resistant agents (4.99 [SD: 1.58], 5.50 [SD:
1.71], respectively; P 0.0051). In assessing whether
characteristics differed within adolescents, as stratified
by positive and negative outcomes, significant differences
included being enrolled in Northern Lima (52% vs. 27%,
respectively; P 0.0476) and having extra-pulmonary
TB (0% vs. 14%; P 0.0131). Univariate cox
proportional hazards analysis in adolescents found that
having at least one comorbidity (cardiovascular disease,
diabetes, hepatitis, seizures, renal failure, psychiatric
S97
disorder, smoking, alcohol or substance abuse) was a

significant predictor of death P 0.0178).
Conclusion: Adolescents showed more successful outcomes, less death, and less presentation of severity
indicators and comorbidities than adults. To better
understand specific risk factors that impact treatment
outcomes for MDR-TB in adolescents, further analysis
should be completed. Next steps will include: 1)
multivariable cox proportional hazards analysis to
identify all predictors of negative outcomes.
04. LTBI: a potpourri

EP-126-04 Prevalence of LTBI infection among
household contacts of multidrug-resistant and
new TB patients in Ho Chi Minh City, Viet Nam
G Fox,1 N Nguyen,2 T A Nguyen,1 T L Nguyen,1 N A Le,1
B Tran Ngoc,3 R Menzies,4 G Marks1 1Woolcock Institute of
Medical Research, Sydney, NSW, 2National Lung Hospital,
Sydney, NSW, Australia; 3Pham Ngoc Thach Hospital, Ho Chi
Minh City, Viet Nam; 4McGill University, Montreal, QC,
Canada. Fax: (61) 291 140 010. e-mail:
foxsimile@gmail.com
Background: Household contacts of patients with multidrug resistant (MDR)-TB are a high-priority group for
screening, according to recent World Health Organization (WHO) guidelines. However, the difference in
transmission associated with exposure to MDR-TB,
compared to drug susceptible TB, is poorly understood.
This study aimed to compare the prevalence of latent
tuberculosis infection (LTBI) among household contacts
of patients with smear positive MDR-TB and new TB in
Viet Nam, and assess demographic characteristics
associated with these outcomes.
Design/Methods: 136 household contacts of smear
positive MDR-TB patients and 182 household contacts
of newly diagnosed TB were screened in nine District TB
units in Ho Chi Minh City, Viet Nam. Tuberculin skin
testing (TST) was performed on all contacts with a
10mm cut-off was used to define LTBI. Chest X-ray,
symptom screening and GeneXpert were used to exclude
active disease. Multivariable logistic regression was used
to calculate the associations between demographic
characteristics and LTBI.
Results: The LTBI prevalence was 39% (30.8-47.2 %)
among contacts of MDR-TB patients and 25.8% (19.432.2%) among contacts of new patients. Contacts of
MDR-TB patients had a higher prevalence of infection in
comparison to the contacts of new AFB () patients
(adjusted OR 1.8; 95%CI: 1.1 3.1). TST positivity
was more common among contacts who were those
identifying themselves as a minority ethnic group
(Chinese or Khmer) (adjusted OR: 2.6; 95%CI: 1.16.2). Infection increased with age. No other social or
demographic risk factors for infection were identified.
Conclusion: Household contacts of MDR-TB have a high
risk of becoming infected, compared to contacts of
patients with new TB. Our findings support WHO
S98
recommendations that household contacts of MDR-TB

patients should be a high-priority group for routine
contact investigation.
EP-127-04 Evaluation of TB contact tracing in

Brazil: operational research
N H Orfao,1 A Wysocki,2 M A Ponce,2 T Arakawa,1
A Beraldo,1 L M Lopes,1 M E Brunello,1 A Kritski,3
T C Scatena Villa1 1Ribeirao Preto School of Nursing,
Ribeirao Preto, SP, 2Ceres University, Sao Jose Do Rio Preto,
SP, 3Federal University of Rio de Janeiro, Rio De Janeiro, RJ,
Brazil. e-mail: lilisew@yahoo.com.br
Background: Tuberculosis (TB) control actions has been

decentralized to Primary Health Care (PHC) in Brazil
since 2004 and TB contact tracing is considered a
cornerstone of the national TB program (NTP). PHC
arrangements related to the process of care influences the
performance of TB control strategy, requiring researches
to evaluate the incorporation of TB contact tracing in
PHC settings. Aim: to analyze PHC services performance
to PTB contact tracing in a decentralized priority setting
for TB control in Brazil.
Methods: Operational research, conducted in Sao Jose
do Rio Preto (SJRP) SP - Brazil. The study population
included 336 pulmonary TB (PTB) contacts (study
population) of 213 PTB patients followed by PHC in
all five health districts of SJRP among 2012-2013.
Contacts of PTB patients that lived in SJRP but was
followed in other city had been excluded from the study.
Data about PTB contact tracing were collected from
System Notification and Monitoring of Cases of Tuberculosis (TB-WEB), System of TB Chemoprophylaxis and
documental analysis of Medical records. PTB contacts
were categorized in 2 groups: under and above 10 years
old, according to NTP recommendations. Descriptive
statistics and Chi square test for categorical variables
were carried out as statistical analysis.
Results: 69 (20.5%) PTB contacts had no evaluation but
were incorrectly registered as no indication for isoniazid
preventive therapy (IPT). Regarding the 211 (78.7%)
PTB contacts under 10 years old: 87 (41.2%) met the
indication criteria for IPT, 39 (39.1%) did not do IPT and
of these, 10 (25.6%) was incorrectly registered as no
indication for IPT. Of the 56 (82.4%) PTB contacts
above10 years old: 19 (33.9%) met the indication criteria
for IPT and of these, 9 (47.4%) did not do IPT. PTB
contacts under 10 years old were more evaluated by Xray (Only X-ray P 0.0324; X-rayTuberculin Test P ,
0.001). There was no information in medical records
about: PTB convocation to Health assessment (54.2%),
respiratory symptoms, vaccination with BCG and
morbidities (100%) and active finding when TB contact
did not go to PHC (58.3%).
Conclusion: PHC is performing incomplete TB contacts
evaluation, incorrect registers of IPT indications and
absence of notes in medical records regarding the
identification, search and evaluation of contacts. There
is need to systematize an active case finding and
surveillance for PTB contacts in PHC.
EP-128-04 Predictive value of the QuantiFERON

test in the Danish population
T Hermansen,1,2 T Lillebaek,1 K Langholz,2 P Andersen,1
P Ravn2 1Statens Serum Institut, Copenhagen,
2
Nordsjaelland Hospital, Hillerd, Denmark. e-mail:
thomasstighermansen@hotmail.com
Background: Interferon-c release assays (IGRA) are used

for the diagnosis of latent infection with M. tuberculosis
(LTBI). The QuantiFERON-TB Gold In-Tube Test (QFT)
has been used in Denmark since 2005. The aim of this
study was to determine the predictive value of the QFT
among individuals who had a QFT performed between
2005 and 2011 and subsequently developed (incident
TB).
Methods: Results of QFT and information on TB
diagnosis were obtained. Incident TB was defined as
development of TB more than 3 months after a QFT
result. We determined the risk of incident TB depending
on the initial QFT result, and calculated the positive and
negative predictive values, and assessed individual risk
factors such as age.
Results: 15 745 persons had an available QFT result
representing 46 258 person-years of follow-up, with a
median follow-up time of 2.86 years (IQR 1.70). Of all
tests performed, 85.2% (13,422) were negative, 9.6%
(1519) positive and 5.1% (804) indeterminate. Nineteen
patients with a positive test and 17 patients with a
negative test developed incident TB. The incidence rate
for QFT-positive individuals was 4.14/1000/year and for
QFT-negative individuals 0.44/1000/year. The positive
predictive value (PPV) of the test was 1.25% (19/1519)
and the negative predictive value (NPV) was 99.87%
(13,405/13,422). The test was affected by patient age;
among patients ,35 years the PPV was 2.22% and NPV
99.95%, and among patients 735 years the PPV was
0.90% and NPV 99.84%. Only 5.3% (1/19) of the
incident TB cases, who had an initial positive QFT,
received prophylactic isoniazid treatment.
Conclusion: In a Danish cohort of persons screened with
the QFT, we found a positive QFT to be a predictor for
progression to TB, with an approximately ten times
higher incidence rate compared to QFT-negative individuals. The predictive value of the QFT was influenced
by age, with both a higher PPV and NPV among children
and adults ,35 years.
EP-129-04 Performance of QuantiFERON-TB

Gold In-Tube vs. the tuberculin skin test in
screening for latent tuberculous infection in
health care workers in Georgia
V Mirtskhulava,1,2 R Kempker,3 M Kipiani,1 N Tabagari,2
J Whitaker4 1National Center for Tuberculosis and Lung
Diseases, Tbilisi, 2David Tvildiani Medical University, Tbilisi,
Georgia; 3Emory University, Atlanta, GA, 4Mayo Clinic,
Rochester, MN, USA. e-mail: verikomir@gmail.com
Background: Limited data exists on the performance of

QuantiFERON-TB Gold In-Tube (QFT-GIT) vs. Tuberculin Skin Test (TST) for serial screening of latent
tuberculosis infection (LTBI) among Healthcare workers
(HCWs) in low and middle income TB endemic

countries.
Design/Methods: We conducted a retrospective analysis
of QFT-GIT and TST results of 163 HCWs. Our
participants were HCWs who had TST and QFT-GIT
performed at baseline and had repeated testing between
March 2009 and September 2010 as part of a prior
study.1 Proportions of concordant results of the two
diagnostic tests were compared between HCWs with a
bacillus Calmette-Guerin (BCG) scar vs. no BCG scar
and between HCWs with self-reported daily occupational TB exposure vs. those who did not see TB patients on a
daily basis. The proportions were compared by twoproportion z-test. Logistic regression was used to assess
risk factors for discordant results at baseline and
repeated testing.
Results: Among 163 HCWs, 79% (129) of HCWs were
found to have a BCG Scar by visual inspection. Forty-five
percent (73) reported daily occupational TB exposure.
Mean time between baseline and follow-up LTBI tests
was 69 weeks (standard deviation 25 weeks). Proportion
of concordant results of QFT-GIT and TST was higher
among those HCWs with no BCG scar compared to the
group with a BCG scar both at baseline (79% vs. 67%, P
, 0.18) and repeated testing (79% vs. 70%, P , 0.27).
Furthermore, proportion of concordant results of the
two diagnostic tests for LTBI was higher among HCWs
with daily occupational TB exposure compared to those
HCWs who did not see TB patients on a daily basis at
baseline (78% vs. 63%, P , 0.04) and repeated testing
(75% vs. 69%, P , 0.36). In multivariate analysis, we
found that the HCWs with discordant LTBI test results
TST-positive / QFT-GIT negative group, were less likely
to report daily occupational TB exposure (adjusted odds
ratio (aOR): 0.4, P , 0.02) compared to the HCWs with
concordant LTBI test results at baseline LTBI screening.
The association was not significant at repeated testing
(aOR: 0.7, P , 0.31).
Conclusions: Concordance between the two tests for
LTBI was higher among HCWs with high occupational
TB exposure at baseline screening and the difference
between the two tests performance vanished at repeated
testing. 1. Whitaker J A, Mirtskhulava V, Kipiani M, et
al. Prevalence and incidence of latent tuberculosis
infection in Georgian healthcare workers. PLOS ONE
2013; 8: e58202.
EP-130-04 Occupational screening for latent

tuberculous infection in German health care
workers
A Nienhaus,1 A Schablon2 1Institute for Statutory Accident
Insurance and Prevention in the Health and Welfare Services,
Hamburg, 2University Clinics Hamburg Eppendorf, Hamburg,
Germany. e-mail: albert.nienhaus@bgw-online.de
Background: Healthcare workers (HCWs) in countries

with low incidence of tuberculosis are still considered a
high-risk group for latent TB infection (LTBI) and are
sometimes routinely screened for LTBI. In this study the
German Occupational TB Network data is analysed in
S99
order to estimate the prevalence and incidence of LTBI in

German HCWs.
Design/Methods: 4123 HCWs were screened with the
QuantiFERON Gold in tube (QFT) at least once, a
second QFT was performed on 1117 HCWs either in the
course of contact tracing or serial examination following
German OSH regulations. TB history and occupational
exposure was assessed with a standardized questionnaire. Active TB was excluded by X-ray when the QFT
was positive.
Results: 9.2% of the HCWs were QFT positive. No
active TB was diagnosed. Positive in at least two
consecutive QFT were 5.6%. Instable QFT results were
observed in 3.6%. Reversions occurred about 10 times
more often than conversions. Risk factors for a stable
positive QFT result (positive in at least two QFTs) were
age .55 years (OR 3.8), foreign country of birth (OR
2.4), personal history of TB (OR 6.5) and workplace,
e.g., internal medicine (OR 1.4), infection ward (OR 1.8)
or geriatric care (OR 1.8). Using a borderline zone from
0.2 to ,0.7 IU/ml reduced the number of instable QFT
results by 41.3%.
Conclusion: In countries with a low incidence of TB and
high hygiene standards the LTBI infection risk for HCWs
seems to be low. Introducing a borderline zone from 0.2
to ,0.7 IU/ml may help to avoid unnecessary X-rays and
preventive chemotherapy. No case of active TB was
observed during the study period of eight years.
Therefore, it seems reasonable to restrict TB screening
to HCWs who had known contact with infectious
patients or materials.
EP-131-04 Cost-effectiveness of the testing and

treatment of latent tuberculous infection: a
systematic literature review
C Angeletti,1 M Sane Schepisi,1 D Goletti,1 R Mancini,1
R Mancini,1 G Sotgiu,2 A Matteelli,3 H Getahun,3 E
Girardi1 1National Institute for Infectious Diseases IRCCS L.
Spallanzani, Rome, 2University of Sassari, Sassari, Italy;
3
World Health Organization, Geneva, Switzerland. e-mail:
monica.saneschepisi@inmi.it
Background: We conducted a systematic literature

review to summarise current evidence on the cost-benefit
and cost-effectiveness of latent TB infection (LTBI)
testing and treatment.
Methods: We searched studies on costs and outcomes of
any screening strategies and any drug regimen for LTBI
compared to no intervention in any setting and population group using the following databases: MEDLINE,
EMBASE, CINHAL, The Cochrane Central Register of
Controlled Trials, NHS Economic Evaluations Database
and Health Economics Evaluations Database. The search
was updated on 15 March 2015 with no time limit. For
the purpose of the analysis countries were categorized
using income and estimated TB incidence as follows: A.
upper-middle income countries with TB incidence less
than 100/100 000; B. low income and/or high TB
incidence countries.
Results: Our literature search yielded 5997 unduplicated
records and after considering their references and the
S100
bibliography of relevant reviews on the topic, we finally

included 45 original articles published between 1981 and
2014. Nine studies analyzing testing and treatment of
LTBI in persons migrating to catergory A countries, show
that this intervention may determine savings for the
health care system or have a favorable incremental costeffectiveness ratio, when screened persons originate from
countries with a TB incidence above 120-150/100 000.
Similarly savings or a favorable incremental costeffectiveness ratio are reported in 7 studies (all conducted
in category A countries) on contacts of patients with
active TB, and in 7 studies on persons living with HIV/
AIDS both in category A and category B countries.
Limited evidence is available for other population
groups.
Conclusion: The available evidence suggest that screening and treatment for latent TB infection may be a costeffective intervention for persons migrating form high TB
incidence countries, contacts of active TB cases and
persons living with HIV. However, extrapolation of costeffectiveness measures from one setting to another is
difficult due to marked variability in economic inputs, in
epidemiologic and natural history parameters, as well as
in assumptions on preventive treatment effectiveness.
Results: A common complaint was that long queues in

the clinics presented a challenge to access HIV-TB
services and that going to the clinic was an all-day
event. However, the majority of patients rated the
quality of care and services highly. There were also
concerns expressed about the variability in TB screening
and diagnosis procedures. In addition to the lack of staff,
there was confusion reported by providers regarding the
current state of TB-related guidelines for PLHIV. For
example, only a few clinics had started using the TST per
national guidelines. Additionally, many patients had
never heard of or received IPT. When patients and
providers were asked about a blood test to screen for TB
nearly all were supportive of the idea and felt it would
help overcome logistical barriers to current screening
practices. Structural factors such as HIV-TB-related
stigma, poverty and food insecurity were raised as
barriers to retention in care and adherence to TB-HIV
medications.
Conclusion: Findings underscore the need for institutional capacity building in clinics, including training of
providers on the implementation of national guidelines
on TST and IPT for PLHIV; as well as community
awareness programs to push demand for these services.
EP-133-04 Institutional and structural barriers

to TB screening in South Africa: qualitative
insights from health care providers and
patients from the TEKO trial
EP-134-04 Impact of rifapentine price reduction

on utilization for TB treatment in the USA
C Tudor,1 D Kerrigan,2 E Variava,3 J Golub,2,5

K Motlhaoleng,4 L Lebina,4 N Martinson4 1International
Council of Nurses, Geneva, Switzerland; 2Johns Hopkins
University Bloomberg School of Public Health, Baltimore, MD,
USA; 3Department of Health Tshepong Hospital, Klerksdorp,
4
Perinatal HIV Research Unit (PHRU), Soweto, South Africa;
5
Johns Hopkins University School of Medicine, Baltimore,
MD, USA. e-mail: tudor@icn.ch
Background: Tuberculosis (TB) remains the leading

cause of mortality among people living with HIV
(PLHIV) globally. The TEKO study in South Africa is
comparing the impact of latent tuberculosis screening for
newly diagnosed HIV patients via QuantiFERONwTB
Gold In-Tube test (QGIT) at time of blood drawn for
CD4 count compared to a tuberculin skin test (TST).
Prior to trial implementation a baseline qualitative study
was conducted to assess patient and provider experiences, attitudes and operational issues related to TB-HIV
services in all clinics in the study.
Design/Methods: Fifty-six in-depth interviews were
conducted: 28 with PLHIV and 28 with care providers
from 14 public clinics in North West Province, South
Africa. Key domains explored included TB screening,
diagnosis, isoniazid preventive therapy (IPT), and
treatment processes. All participants were asked their
thoughts about the use of a blood test to screen for latent
TB. All interviews were audio recorded and transcribed.
Narrative analysis was used to document key elements of
the TB screening and management process from each
interview. Narratives were synthesized across participants looking for patterns of consensus and diversity.
A Deluca,1 M Macaraig,2 K Mcginnis,2,3 D Wegener,4

J Kanouse,4 M Frick,5 L Mckenna5 1Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, 2New
York City Department of Mental Health and Hygiene, Long
Island City, NY, 3U.S. Centers for Disease Control and
Prevention, Atlanta, GA, 4National TB Controllers
Association, Smyrna, GA, 5Treatment Action Group, New
York, NY, USA. e-mail: andeluca@jhsph.edu
Background: USA (US) tuberculosis (TB) programs

reported that the high cost of rifapentine (RPT) was a
deterrent to adoption of a new 12 dose once-weekly
regimen of isoniazid and RPT (3HP) for the treatment of
TB infection. In 2013, in response to a community-driven
advocacy campaign, RPTs manufacturer, Sanofi Aventis,
reduced the price per 32-tablet blister pack by 56%, from
US$52 to US$32 for those eligible for federally discounted pricing. A survey of US TB programs was conducted
to assess whether the price reduction had an impact on
RPT use.
Design/Methods: In December 2014, a web-based survey
was sent to the 62 TB programs receiving federal
funding. One knowledgeable respondent per TB program
completed the survey. Descriptive analysis was performed on survey responses.
Results: Of the 62 US TB programs, 61 (98%) responded
to the survey. A total of 46/61 (77%) respondents were
aware of the RPT price drop, and 45/60 (75%) knew
their TB programs have access to RPT at the reduced
price. Among TB programs that were both aware of the
price reduction and able to access it, 24/36 (67%)
reported increased use of RPT after the price decrease.
Among all TB programs, the majority (62%) were not
using RPT as often as they would like. Over 50% of TB
programs reported barriers to use, including cost to the

program (72%), directly observed therapy requirement
for 3HP (72%), quantity of pills for 3HP (56%), and
concerns about safety (52%). Over 60% of TB programs
want more information about use of RPT in HIV-infected
individuals, those with elevated liver function tests, and
pregnant women.
Conclusion: Use of RPT increased following the 2013
price reduction. Pricing advocacy can remove an
important barrier to programmatic uptake of TB
innovations. There is a need for further education to
increase awareness of the reduced cost of RPT and to
address concerns about use of RPT in special populations. Dissemination of research findings, communication about discounted drug availability, and considerations of patient acceptability (e.g., daily pill quantity)
must occur in parallel with efforts to achieve affordable
drug pricing. As improved drug regimens become
available domestically and internationally, these lessons
could help communities planning for the launch of new
treatment options. Advocates, TB-affected communities,
and public health professionals can work proactively
with pharmaceutical companies to achieve affordable
pricing that facilitates drug access.
EP-135-04 Integrating existing programs to

address barriers to treatment of latent
tuberculous infection
J Burzynski,1 K Mcginnis,1,2 C Chuck,1 M Macaraig1
Department of Health and Mental Hygiene, New York, NY,
2
U.S. Centers for Disease Control and Prevention, Atlanta,
GA, USA. e-mail: mmacarai@health.nyc.gov
Background: Adherence to treatment for tuberculosis

(TB) infection with 9 months of isoniazid is low. In New
York City (NYC) a 3-month, once-weekly regimen of
isoniazid and rifapentine (3HP) under directly observed
therapy (DOT) increased treatment completion, but inperson clinic DOT was a barrier. A second initiative using
video conferencing for remote DOT observations
(VDOT) was effective with active TB patients. This
describes how NYC addressed clinic DOT barriers for
patients on 3HP by offering observations by VDOT.
Intervention: NYC integrated 3HP with VDOT (V3HP)
by updating staff responsibilities and policies, including
adapting current 3HP and VDOT protocols, patient
follow-up procedures, medication issuance, and documentation procedures for electronic medical records.
Patients eligible for V3HP must meet current 3HP and
VDOT criteria and be willing to use a personal videoenabled device during business hours. To prevent
interruption of program services, resources were obtained in advance, such as a license for the video
software, hardware equipment, language services, and
a reference guide to help patients use the software. Buy-in
and training was accomplished by presentations to
supervisors and physicians, in-servicing nurses, and
training three centralized staff members to conduct
V3HP observations and record outcomes. Excel databases were created to monitor staff workload and
S101
technology issues. DOT observations and adverse reactions were documented in an existing electronic medical
record system. The pilot will enroll patients from
February 2015 to August 2015. Treatment completion
will be compared with historical data.
Results: In the pilots first nine weeks 18 patients
accepted V3HP. No observations were done on five
patients: three preferred clinic DOT, one had technical
issues and returned to clinic DOT, and one stopped
treatment due to side effects from the first clinic-based
dose. Among the 12 remaining patients, 41 VDOT
observations have been conducted (range: 18 doses);
median time was 5.5 minutes (range: 332). All patients
have been adherent to VDOT. No major side effects have
been reported. Language services are provided for one
patient. Technological issues stemming from connection
problems have been the most prevalent issue to date.
Conclusions: Combining existing programs can provide
treatment delivery options that overcome barriers.
Further analysis will determine if these efforts improve
treatment completion for TB infection.
05. Rats, videos and secretions: the

landsape of TB detection
EP-136-04 Differences in secretion of MPT64
antigen between Mycobacterium africanum
West African type 2 and Mycobacterium
tuberculosis sensu stricto
A Ofori-Anyinam,1,2 F Kanuteh,2 I Adetifa,3 S Agbla,2
B De Jong,1 M Antonio,2 F Gehre1,2 1Institute of Tropical
Medicine, Antwerp, Belgium; 2Medical Research Council
Unit, The Gambia, Fajara, Gambia; 3KEMRI-Wellcome Trust
Research Programme, Kilifi, Kenya. e-mail: nofori@mrc.gm
Background: High Mycobacterium tuberculosis strain

diversity in West Africa, with all lineages within the M.
tuberculosis complex (MTBC) present, makes this region
ideal for assessing the performance of diagnostics and
vaccines. The 24kD secreted protein, MPT64, associated
with virulence and produced solely by members of the M.
tuberculosis complex is a target of widely used rapid
speciation tests, yet the impact of strain diversity on the
performance of these tests in West Africa has not been
clearly shown. Previous reports indicated the inability of
these tests to detect strains which had lost mpt64 or
acquired mutations. In the Gambia, up to half of
Tuberculosis (TB) infections are caused by Mycobacterium africanum West African type 2 (MAF2), a strain
genetically distinct from other M. tuberculosis strains .
Preliminary sputum RNA expression data showed a
down regulation of mpt64 in MAF2 compared to M.
tuberculosis. We hypothesised that differences might
exist in the rate of secretion of the MPT64 antigen
between M. tuberculosis and MAF2 which might
influence detection by MPT64 rapid tests.
Design/Methods: Sputum from patients with suspected
TB within the Greater Banjul area, where 80% of all
individuals with TB in the Gambia are found, were
S102
decontaminated, confirmed by smear microscopy for

Acid-Fast Bacilli (AFB) and cultured in the Bactec
Mycobacteria Growth Indicator Tube (MGIT) 960
machine. 173 AFB positive and blood agar negative
cultures were tested with the Becton Dickinson (BD
TBcID) and Standard Diagnostics (SD Bioline) rapid
speciation kits on the day of MGIT positivity. Cultures
which tested negative were re-incubated at 378C and
subjected to follow up rapid tests at 3, 10, 15 and 90
days. All cultures were spoligotyped to determine strain
lineages. Survival analysis and regression modelling of
the data was performed using STATA version 12.1.
Results: We observed time and lineage specific differences in the detection of M. tuberculosis and MAF2 by the
rapid tests. MTB strains converted to positivity approximately three times faster than MAF2 within the 10 day
window proposed for conducting the test described in the
BD TBcID manual (Hazard Ratio2.85, 95%CI1.884.32, P ,0.001). Although the SD Bioline kit did not
specify a time frame for testing, similar observations
were made.
Conclusion: Our findings underscore the need to
incorporate molecular tests in routine identification
and confirmation of all members of the MTBC.
Ultimately, future development of diagnostics and
vaccines should be guided by implications of strain
diversity on product success.
tuberculosis genome were 73% (range: 4.03-99.3%)

and 0.0185X coverage (range: 0.002-0.693X), respectively. This provided enough coverage for phylogenetic
placement of 13 of the 16 samples at lineage level shown
in the maximum likelihood tree (figure), as confirmed by
spoligotyping. Microbiome analysis showed no additional obligate pathogens and analysis of microbiome
ecology was deemed inappropriate for the small sample
size. Future studies will evaluate the clinical and
epidemiological impact of metagenomic diagnosis by
comparison to traditionally used methods and analyse
the sputum-association microbiome with better resolution than currently used 16S rRNA genotyping methods.
Conclusion: The preliminary data show promise for
diagnosis of tuberculosis and other respiratory pathogens
by direct sputum metagenomics. The purposes of
numerous culture-dependent methods are integrated into
a single rapid approach for any respiratory pathogen;
including detection of mixed strains and co-infections,
high-resolution epidemiological characterisation and
identification of drug-resistance mutations, with increased pathogen genomic coverage. Further technology
developments may make this feasible at point-of-care,
from other clinical sample types and scalable to local
demand.
EP-137-04 Rapid diagnosis and epidemiology of

Mycobacterium tuberculosis complex strains in
The Gambia by metagenomic DNA sequencing
directly from sputum
E Doughty,1,2 M Sergeant,2 I Adetifa,3 M Antonio,1
M Pallen2 1Medical Research Council Unit, The Gambia,
Fajara, Gambia; 2University of Warwick, Coventry, 3London
School of Hygiene & Tropical Medicine, London, UK. e-mail:
emma.doughty@warwick.ac.uk
Background: Absence of timely and accessible diagnosis

is widely considered a barrier to patient treatment and
control of the tuberculosis (TB) epidemic. Currently used
methods are slow, infeasible in many locations and
provide limited information including drug sensitivity.
Our initial data show that metagenomics is a promising
method for diagnosis to circumvent these issues.
Design/Methods: For preliminary studies, DNA was
extracted from 16 AFB smear positive sputum after
attempts to deplete human DNA and sequenced using the
Illumina MiSeq. The MTBC- and human-derived DNA
contents were determined by mapping against the
respective reference genomes. TB Phylogenetic lineages
were assigned from SNPs and the positions of IS6110
elements identified in tuberculous reads, then verified by
spoligotyping of cultured isolates. The microbiome was
analysed using microbial identification software. Optimisation of laboratory methods is on going before
evaluation of this method for direct diagnosis, epidemiology and in vivo research of organisms in sputum.
Results: Using this method, the medians for human DNA
content in each sample and coverage of the M.
EP-138-04 Diagnosis of pulmonary and extrapulmonary tuberculosis in Paraguay using the

electronic nose (Paranose study)
C Magis-Escurra,1 R Coronel Teixeira,2 M Rodriguez,2
J Gerritsen,3 G Chaparro Abente,2 D Perez,2 J Yntema4
1
Radboud University Medical Centre, Nijmegen, Netherlands;
2
Instituto Nacional de Enfermedades Respiratrios y
Ambientales, Asuncion, Paraguay; 3ENose Company,
Zutphen, 4Radboud University Medical Centre, Nijmegen,
Netherlands. e-mail: cecile.magis-escurra@radboudumc.nl
Background: Diagnostic techniques for TB still remain

largely unattainable in developing countries due to the
requirement of extensive laboratory facilities and the
need for an increasing number of trained staff. A pilot
study in Bangladesh, with a prototype of an electronic
nose device, found a sensitivity of 95.9% and specificity
of 98.5% to detect pulmonary TB. The company

developed is a simplified, easy-to-use, point-of-care
device that can be produced at low cost. We tested the
diagnostic accuracy of this electronic nose device in a TB
reference centre in Asuncion.
Design/Methods: Observational study in adult patients

with pulmonary TB, pleural TB, and matched healthy
controls, after signing informed consent, from May until
December 2014. TB diagnosis was established by gold
standard (culture of Mycobacterium tuberculosis complex) or other supporting evidence. Clinical data
included age, gender, symptoms, body weight, comorbidity, onset of disease, co-medication and factors
influencing breath. Chest X-ray, blood chemistry and
HIV testing were done in TB patients, and sputum
samples were taken for microscopy and culture for M.
tuberculosis if expectorating. Measurements of the air
composition was done breathing normally during 5
minutes.
Results: 86 participants were evaluated with the device.
Thirty eight patients had pulmonary TB, 3 pleural TB,
and 45 were healthy controls. The results of breath
samples showed 41 positive and 45 negative signals, a
result above or below the threshold value calculated by
the model. We subsequently compared the outcomes
with our database and concluded that 40 were proven TB
negatives and 38 participants had proven TB. That
resulted in 5 possible false positives and 3 possible false
negatives. Sensitivity in this first analysis is 93% and
specificity 89% which is shown in the ROC curve.
Factors influencing the diagnostic accuracy are not yet
completely evaluated.
Conclusion: It is a simple, non invasive test, and showed
a high accuracy to detect TB in Paraguay. The
demonstrated sensitivity is much higher than any other
currently available test to diagnose active TB. These
results implicate that the device may be of great value in
variable future settings. Of course, more research needs
to be done in patients with paucibacillar TB, in different
regions in the world, and to study the influence of
consumption of food or either toxic substances shortly
before the test is performed.
S103
EP-139-04 GxAlert monitors and reduces high

testing error rates in Nigerias GeneXpertW
machine
K Jimoh Agbaiyero,1,2 L Ekbladh,1,2 M Benezet,2 J Takle,3
C Macek3 1Abt Associates, Abuja, Nigeria; 2Abt Associates,
Bethesda, MD, 3SystemOne, Boston, MA, USA. e-mail:
kehindeagbaiyero@yahoo.com
Background: Nigeria is a WHO high burden country for

drug-resistant tuberculosis (DR-TB). However, national
diagnostic and surveillance capacities are compromised
due to high error rates on GeneXpert MTB/RIF machines
testing for DR-TB. The Nigerian Tuberculosis and
Leprosy Control Program (NTBLCP) pursued development of a cost-effective technology, GxAlert, to improve
DR-TB reporting speed and accuracy and to monitor
individual GeneXpert machine error rates to strengthen
both healthcare system responses to DR-TB and management of geneXpert investments.
Design/Methods: Abt Associates and SystemOne developed GxAlert for the NTBLCP. GxAlert is an innovative,
cost-effective solution that networks geneXpert machines, sending average error rates by individually
identified machine in real time to an online database.
This cuts reporting time from months to seconds and
enables more timely interventions such as recalibration
of machines and staff supervision. GxAlert tracks service
warranty thresholds and module replacement for each
device. GxAlert is configured on geneXpert systems by
installing a modem from a local telecom to access the
internet and send encrypted data to the online GxAlert
database. Piloted in 8 laboratories, GxAlert has since
expanded to 46 GeneXpert sites in Nigeria.
Results: The GeneXpert testing error rate in 35 facilities
decreased to 5% and below in March 2015 from 9% and
above in April, 2014. Data on machines were producing
higher than normal error rates were uploaded automatically from GeneXpert systems to a secure online
dashboard display to improve national level program
management, while SMS alerts were sent to Facility
Managers, Local Government supervisors and State
program managers for immediate action. Faster, better
quality data on MDR-TB portion of total TB indicators
are now reported with very minimal error rates.
Conclusion: The use of GxAlert to network GeneXpert
test results and operational status is a cost-effective
model for national sustainability of rapid molecular
diagnostic systems. GxAlert enables quick remote
technical support, troubleshooting, and maintenance of
the geneXpert machines as well as ongoing disease
surveillance by collecting and displaying compiled test
results and incidence rates by facility area. Based on
Nigeria, a pilot of the GxAlert approach in other highburden DR-TB countries is recommended.
S104
EP-140-04 Loss-of-function mutations in pepQ

confer cross-resistance to bedaquiline and
clofazimine in Mycobacterium tuberculosis
EP-141-04 Characterization of Mycobacterium

tuberculosis complex isolates with discordant
rifampicin susceptibility test results
D Almeida,1 T Ioerger,2 S-Y Li,1 S Tyagi,1 K Andries,3

K Mdluli,4 J Grosset,1 J Sacchettini,5 E Nuermberger1,6
1
Johns Hopkins University Center for TB Research, Baltimore,
MD, 2Texas A&M University, College Station, TX, USA;
3
Tibotec BVBA, Beerse, Belgium; 4Global Alliance for TB Drug
Development, New York, NY, 5Texas A&M University, College
Station, TX, 6Johns Hopkins Bloomberg School of Public
Health, Baltimore, MD, USA. e-mail:
deepak.almeida@gmail.com
C N Beylis,1,2 Y Ghebrekristos,1,2 J Wojno,1,2 M Nicol,1,2

J Simpson,3 P Da Silva4 1National Health Laboratory Service,
Cape Town, 2Universtity of Cape Town, Cape Town, 3National
Health Laboratory Services, Cape Town, 4University of
Witwaterstrand, Johannesburg, South Africa. e-mail:
yonas.ghebrekristos@nhls.ac.za
Background: Bedaquiline (BDQ), a new mycobacterial

ATP synthase inhibitor active against M. tuberculosis has
recently been approved by the FDA for treatment of
multidrug-resistant tuberculosis. A new non-target-based
mechanism of cross-resistance to BDQ and clofazimine
(CFZ, C) mediated by mutation of Rv0678 was recently
described and shown to be responsible for failure to
convert sputum cultures. For new TB drugs, it is essential
to define and catalog the mutations conferring resistance
in order to carry out proper follow-up of patients who
fail and relapse after receiving the drug, to design
appropriate diagnostic tests (including rapid molecular
tests), and to conduct population level surveillance for
changes in susceptibility. Here we present our findings on
mice efficacy studies where mutants with reduced
susceptibility to BDQ were selected and found to be also
cross-resistant to CFZ.
Methods: Efficacy studies in mice evaluating BDQ with
or without CFZ selected mutants that grew on BDQ- and
CFZ-containing plates. Minimum inhibitory concentrations (MICs) were determined using doubling concentrations in 7H11 agar. Mutation analysis was carried out
by whole genome sequencing and confirmed by PCR. To
confirm cross-resistance in vivo, BALB/c mice were
infected with either the H37Rv parent strain or one
resistant mutant (J5) and then randomized to receive no
treatment, isoniazid (10 mg/kg) alone, BDQ (12.5, 25, or
50 mg/kg) alone, CFZ (20 mg/kg) alone, or BDQ 25 mg/
kg plus CFZ for 4 weeks before determining lung CFU
counts. Complementation of the J5 mutant with the wild
type gene was performed and the complemented strain
was tested for BDQ and CFZ susceptibility in vitro and in
vivo alongside the parent and J5 strains.
Results: A total of 5 mice, 2 from BDQ alone and 3 from
BDQ-CFZ treated group harboured mutants with higher
MICs. Their MIC for BDQ and CFZ were 0.12-0.25 and
0.5-1 lg/ml, in contrast to that of H37Rv wild type
which were 0.03 and 0.25 lg/ml, respectively. Whole
genome sequencing studies revealed that 4 out of 5
respective isolates had mutations in the pep Q (Rv2535c)
gene. In-vivo efficacy studies in mice with one of the
resistant mutant (J5) showed reduced killing by BDQ and
CFZ as compared with wild type H37Rv. Complementation with wild type pepQ restored the susceptibility to
BDQ and CFZ in vitro and in vivo.
Conclusion: Mutations in pepQ gene reduce susceptibility to BDQ and CFZ.
Background: The Xpert MTB/RIF is the initial screening

test for all TB suspects in the South African National TB
Control programme, with rifampicin (RIF) resistance
confirmed by subsequent lineprobe assay (LPA) (GenoType MTBDRplus) and/or phenotypic (culture based)
drug susceptibility testing. In the majority of cases, there
is concordance between the 2 genotypic tests, as well as
between the overall genotypic and phenotypic results.
Disputed rpoB mutations have been described where
genotypic tests detect an rpoB alteration but phenotypic
tests miss the RIF resistance (low level RIF resistance).
We aim to further characterize the discordant rifampicin
results by rpoB sequencing and rifampicin MIC testing.
Design/Methods: Consecutive isolates with discordant
Xpert: LPA results as well as isolates with possible
disputed rpoB alterations detected by the LPA were
selected. Sanger sequencing of the rpoB gene from
codons 462591 was performed. MIC determination
was performed using the Mycobacterial Growth Indicator Tube (MGIT) 960 system and Epicenter Software
with RIF concentrations 0.0625, 0.125, 0.25, 0.5, 1, 5ug/
ml. Analysis of results was performed on MS Excel.
Results: Ninety one isolates had Xpert RIF resistant, LPA
RIF susceptible results, of which 80 had rpoB sequencing
performed. Xpert was false resistant in 55 (68.7%) and
LPA was false susceptible in 25 (31.3%). A total of 205
isolates with possible disputed rpoB alterations was
detected by the LPA, of which sequencing results were
available for 140. Four mutations made up 70.6% of the
total detected, of which L511P, D516Y, H526N and
L533P were detected in 44 (31.4%), 22 (15.7%), 17
(12.1%) and 16 (11.4%) respectively. All of these
isolates had a RIF MIC lower than 1ug/ml
Conclusion: The majority of the discordant Xpert and
LPA results was false Xpert resistance, likely due to
technical errors and / or delays during processing of
specimens. False LPA susceptible results also occurred,
likely due to errors in processing and / or interpretation.
The most common disputed rpoB alterations detected in
Cape Town echo those detected worldwide. The LPA is
useful to screen for disputed rpoB alterations. rpoB
sequencing and MIC testing are useful tools to troubleshoot discordant results, as well as to inform the TB
clinician about which disputed rpoB alterations confer
low level rifampicin resistance.
EP-142-04 Rat detection technology improves

TB case finding in Tanzanian prisons
G Mgode,1,2 C Cox,2 T Edwards,2 C Mulder,2 E Valverde,3
G Mwesiga,2 J Maeda,2 J Malewa4 1Sokoine University of
Agriculture, Pest Management Centre, Morogoro,
2
Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling APOPO, Sokoine University of Agriculture, Morogoro,
Tanzania; 3Anti-Persoonsmijnen Ontmijnende Product
Ontwikkeling - APOPO, Veterinary School, University
Eduardo Mondlane, Maputo, Mozambique; 4Ministry of
Home Affairs, Tanzania Prisons Service, Dar Es Salaam,
Tanzania. Fax: (255) 232 601 485. e-mail:
gfmgode@hotmail.com
Background: The prevalence of tuberculosis (TB) in

prisons in sub-Saharan Africa is higher than in the
general population and prisons are considered an
important source of TB transmission. TB control in
prisons through proper screening and diagnosis is
challenging in most low-income countries with a high
burden of TB. The use of APOPO TB detection rats has
recently demonstrated an increase in case detection of
over 40% in DOTS centres in Tanzania and Mozambique. We report findings of an intervention implementing
TB detection rats for improving TB case finding in
selected prisons in Tanzania.
Design/Methods: Sputum samples (n 8979) were
collected from 4561 individuals of a prison population
consisting of inmates, prison staff and their families
seeking TB diagnosis at Ukonga prison dispensary
(January 2013 to April 2015) and Keko prison dispensary (February 2015 April 2015) in Dar es Salaam,
Tanzania. Samples were routinely examined by conventional Ziehl Neelsen (ZN) direct smear microscopy at the
two dispensaries and were subsequently securely transported to APOPO TB detection laboratory at Sokoine
University of Agriculture, in Morogoro, Tanzania.
Samples were heat inactivated before being tested by
the detection rats. Samples indicated as TB positive by
detection rats were concentrated and confirmed by light
emitting diode fluorescence microscopy (LED-FM) or
ZN microscopy.
Results: Detection rats tested 4,561 individuals from the
prison population. A total of 505 individuals (11.1%)
were diagnosed as smear-positive TB by the prison
dispensaries, whereas rats detected an additional 170 TB
patients. Thus, detection rats increased the case detection
in the prison population by 34%, and increased the
prevalence of smear-positive TB to 14.8% (n 675
patients).
Conclusion: Using detection rats in combination with
smear-microscopy has been dramatically improving TB
case finding in prisons in Tanzania. This finding is
consistent with those found in general populations tested
by rats in Tanzania and Mozambique. The high sample
throughput due to high speed of rats which screens up to
140 sputum in 20 minutes, reasonable accuracy, in
combination with their low operational costs make the
detection rat technology very suitable for mass screening
of high-risk populations in different geographic regions
with high TB burden. Plans are underway to deploy
S105
detection rats in active TB case finding in selected prisons

in Tanzania and Mozambique.
EP-143-04 Do instructional videos on sputum

submission result in increased tuberculosis case
detection?
G Mhalu,1 J Hella,1 F Mhimbira,1 B Doulla,2
B Mutayoba,2 T Seimon,3 M Weiss,4 L Fenner4 1Ifakara
Health Institute, Pwani, 2Ministry of Heath, Dar Es Salaam,
Tanzania; 3Interactive Research and Development, Karachi,
Pakistan; 4Swiss Tropical and Public Health Institute, Basel,
Switzerland. e-mail: gmhalu@ihi.or.tz
Background: Sputum smear microscopy remains the

cornerstone of diagnostic algorithms in national tuberculosis (TB) control programs in low-income settings.
Instructions on how to produce a good sputum sample
are often included in manuals, but to formulate this
comprehensibly proves exceedingly difficult. Objective:
We examined the effect of an instructional video about
the production of diagnostic sputum on case detection of
TB.
Methods: We prepared a culturally adapted instructional
video for sputum submission. We analyzed 200 presumptive TB cases coughing for more than two weeks
who attended the outpatient department of the governmental hospital in Mwananyamala (Dar es Salaam,
Tanzania). They were randomly assigned to either receive
instructions on sputum submission using the video before
submission (intervention group, n 100) or using routine
verbal instructions (control group, n 100). Sputum
samples were examined for volume, quality, and presence
of acid-fast bacilli by experienced laboratory technicians
blinded to study groups.
Results: Median age was 39.1 years (interquartile range
37.0-50.0); 94 (47%) were females, 106 (53%) were
males, and 49 (24.5%) were HIV-infected. We found that
the instructional video on intervention was associated
with detection of a higher proportion of microscopically
confirmed cases (56%, 95% confidence interval
[95%CI] 45.7-65.9%, sputum smear positive patients
in the intervention group vs. 23%, 95%CI 15.2-32.5%,
in the control group, P , 0.0001); an increase in volume
of specimen defined as a volume 73ml (78%, 95%CI
68.6-85.7%, vs. 45%, 95%CI 35.0-55.3%, P , 0.0001);
and specimens less likely to be salivary (14%, 95%CI
7.9-22.4%, vs. 39%, 95%CI 29.4-49.3%, P 0.0001,
Figure). Older age, but not the HIV status or sex,
modified the effectiveness of the intervention by improving it positively. When asked how well the video
instructions were understood (n 100), the majority of
patients in the intervention group reported to have
understood the video instructions well (97%). Most of
the patients thought the video would be useful (92%).
Conclusion: Sputum submission instructional videos
increased the yield of TB cases through better quality
of sputum samples. If confirmed in larger studies,
instructional videos may have a substantial effect on
the case yield using sputum microscopy and also
molecular tests. This low-cost strategy should be
S106
Oral abstract sessions, Friday, 4 December
considered as part of the efforts to control TB in

resource-limited settings.
ORAL ABSTRACT SESSIONS

01. Symptoms, sputum, sequences: TB
diagnosis and resistance
OA-300-04 Clinical and sputum characteristics
as predictors of Mycobacterium tuberculosis
positive sputum in community-wide active case
finding for tuberculosis
J Ho,1,2,3 P Nguyen,3 T A Nguyen,3 K H Tran,4
S V Nguyen,4 N V Nhung,5 G Fox,2,3 G Marks1,2,3
1
University of New South Wales, Sydney, NSW, 2Woolcock
Institute of Medical Research, Sydney, NSW, Australia;
3
Woolcock Institute of Medical Research, Hanoi, 4Ca Mau
Centre for Social Disease Prevention, Ca Mau, 5National
Tuberculosis Program, Hanoi, Viet Nam. e-mail:
jho@gmp.usyd.edu.au
Background: We are currently conducting a randomized

controlled trial of community-wide active case finding in
which we conduct a house-to-house survey and collect a
single, spontaneously expectorated sputum specimen
from all persons who can produce this. The specimen is
tested using Xpert MTB/RIF. Confirmation is by chest Xray and TB culture. This approach has the advantage of
ensuring a high participation rate in difficult-to-access
communities but scalability may be limited by the cost of
Xpert testing. In this study we seek to establish whether
the absence of reported symptoms or the macroscopic
appearance of the sputum could be used to exclude TB
and hence allow some sputum specimens to be discarded
without testing.
Methods: This study was performed in Ca Mau Province,
Vietnam. All participants were screened for the presence
of TB symptoms. Those able to produce sputum 71 ml
were tested using Xpert, regardless of reported symp-
toms. Sputum quality was assessed using a color chart,

representing salivary, mucoid, muco-purulent, purulent
and blood stained sputum. Sputum volume was estimated visually. Likelihood ratios (LR) were calculated as the
prevalence of a given result in Xpert positive participants
divided by prevalence of that same result in Xpert
negative participants.
Results: Among 33 337 sputum specimens submitted for
testing, 21 624 (65%) were 71 ml and were tested by
Xpert MTB/RIF. 93% of those tested were 62 ml. 159
participants had positive Xpert MTB test results (0.75%
of those with valid Xpert tests). The prevalence of
current cough, daily cough 72 weeks, current sputum
production, daily sputum 72 weeks and hemoptysis
were 33%, 15%, 28%, 13% and 1%, respectively.
Nearly all sputum specimens were mucoid (93%) or
muco-purulent (6%) in appearance. The LR for mucoid
sputum was 0.90 and for muco-purulent sputum it was
2.37. Only 50 (0.24%) sputum specimens were classified
as salivary and one (2%) of these was Xpert MTB
positive. The highest LR for any symptom being present
was 3.67 (hemoptysis) and the lowest LR for any
symptom being absent was 0.70 (current sputum and
current cough). The false negative rate for each of these
symptoms was .0.4%.
Conclusion: In the setting of community wide active case
finding for TB, neither the absence of symptoms nor
macroscopic sputum quality has strong negative predictive value for the Xpert MTB result. Hence these
attributes will not be useful in excluding some specimens
from testing. Support: NHMRC Australia.
OA-301-04 A multicenter study on the new

version of the GenoType MTBDRsl assay for
detection of resistance to fluoroquinolones
and second-line injectable drugs
E Tagliani,1 P Miotto,1 A Cabibbe,1 M Mansjo,2
S Hoffner,2 D Hillemann,3 A Zalutskaya,8,9 D M Cirillo,1,2
A Skrahina4 11Emerging Bacterial Pathogens Unit, San
Raffaele Scientific Institute, Milan, Italy; 2Unit for Highly
Pathogenic Bacteria, Public Health Agency of Sweden, Solna,
Sweden; 3Diagnostic Mycobacteriology, Research Center
Borstel, Borstel, Germany; 4Republican Research and Practical
Centre for Pulmonology and TB, Minsk, Belarus. e-mail:
tagliani.elisa@hsr.it
Availability of rapid molecular tests for the detection of

extensively drug-resistance (XDR) TB is critical in areas
with high rates of MDR and XDR-TB and in settings with
limited conventional drug susceptibility test (DST)
capacity. We conducted a multicenter study to evaluate
the performance of the new GenoType MTBDRsl V2.0
(Hain Lifescience) assay for detection of mutations
leading to fluoroquinolones (FQs) and second-line
injectable drugs (SLIDs) resistance. The assay targets
mutations in gyrA, gyrB genes for FQ-R, and in rrs and eis
genes for SLID-R. Performance was evaluated on 228 M.
tuberculosis complex isolates (73 FQ-R, 147 SLID-R) and
231 smear positive specimens (57 FQ-R, 63 SLID-R)
comparing results to phenotypic DSTand sequencing. For
indirect testing, the concordance of MTBDRsl V2.0 with
phenotypic DST for FQ-R was 94.7% (95%CI 91.097.0%), the test sensitivity and specificity were 83.6%
(95%CI 73.4- 90.3%) and 100% (95%CI 97.6-100%)
respectively. For SLIDs, the concordance of the assay with
phenotypic DSTwas 87.7%, the sensitivity and specificity
were 86.4% (95%CI 79.9-91.0%) and 90.1% (95%CI
81.7-94.9%) respectively. The test correctly identified
86.4% of SLID-R isolates, including 27/127 (21.3%)
strains carrying mutations in eis promoter only. The
concordance between MTBDRsl V2.0 and gene sequencing was .97% for all the genomic regions analyzed. For
direct testing, the overall diagnostic accuracy for FQ-R
was 97.0% (95%CI 93.9-98.5%) and the sensitivity and
specificity were 93.0% (95%CI 83.3-97.2%) and 98.3%
(95%CI 95.1-99.4%) respectively. The test correctly
identified 53 (92.9%) FQ-R specimens. The diagnostic
accuracy for SLID-R detection in specimens was 90.9%
(95%CI 86.5-94.0%) with sensitivity and specificity of
88.9% (95%CI 78.8-94.5%) and 91.7% (95%CI 86.595.0%) respectively. The test correctly identified 56
(88.9%) SLID-R cases. Compared to V1.0, the sensitivity
for FQ-R remained unvaried, whereas MTBDRsl V2.0
showed an improvement for the detection of XDR-TB
with specificity and sensitivity of 81.8% and 80.4% for
direct and indirect testing respectively. The inclusion of
probes to detect mutations in eis lead to higher sensitivity
for detection of KAN-R for both direct and indirect testing
(96% and 95.4% respectively) as compared to V1.0
(66.9%). MTBDRsl V2.0 thus represents a reliable test
for the rapid detection of resistance to second-line drugs
and a useful screening tool to guide appropriate MDR-TB
treatment.
OA-302-04 A six-marker serum biosignature

shows promise in the diagnosis of TB disease in
African primary health care clinic attendees
presumed to have TB
N Chegou,1 J Sutherland,2 A Crampin,3
M Van Der Vyver,4 R Howe,5 H Mayanja-Kizza,6 G Walzl1
1
Biomedical Sciences, DST/NRF Centre of Excellence for
Biomedical Tuberculosis Research & SAMRC Centre for
Tuberculosis Research, Division of Molecular Biology and
Human Genetics, Stellenbosch University, Tygerberg
Campus, Cape Town, South Africa; 2Medical Research
Council Unit, Fajara, Gambia; 3Karonga Prevention Study,
Chilumba, Malawi; 4Faculty of Health Sciences, University of
Namibia, Windhoek, Namibia; 5Armauer Hansen Research
Institute, Addis Ababa, Ethiopia; 6Makerere University,
Kampala, Uganda. Fax: (27) 219 389 863. e-mail:
novel@sun.ac.za
Background: There is an urgent need for simple, rapid,

inexpensive and yet accurate tools for the diagnosis of TB
disease at points-of-care in resource-limited settings. In
the present study, we investigated the accuracy of host
inflammatory biomarkers detected in serum samples
obtained from individuals suspected of having pulmonary TB disease at primary health care clinics situated in
six African countries (Ethiopia, Uganda, The Gambia,
Malawi, Namibia and South Africa), for the diagnosis of
TB disease.
S107
Design/Methods: We collected serum samples from

individuals presenting with symptoms requiring investigation for pulmonary TB at the respective clinics, prior to
microbiological diagnosis of TB disease. We then
evaluated the levels of .40 host inflammatory biomarkers in the stored serum samples using the Luminex
platform. On the basis of laboratory, clinical and
radiological findings, participants were classified into
the following groups using a pre-established diagnostic
algorithm: definite TB, probable TB, possible TB,
questionable disease status or no-TB.
Results: In a pilot study conducted on 148 individuals
recruited from the Ethiopian and South African study
sites, a 5-marker serum biosignature diagnosed TB
disease with a sensitivity of 86% and specificity of
90%. We then evaluated the accuracy of this biosignature in a validation cohort including 717 individuals who
were recruited from field sites situated in Uganda, The
Gambia, Malawi, Namibia and South Africa. Of the 702
individuals who were finally analyzed in the validation
study, 175 were definite TB cases, 11 were probable and
29 were possible TB cases whereas 487 of the study
participants did not have TB. In six individuals the
diagnosis of TB remained uncertain. A 6-marker
biosignature comprising of CRP, prealbumin, IFN-c,
Complement factor H, apolipoprotein-A1 and IP-10
ascertained TB disease with a sensitivity of 89% and
specificity of 76.0%, with positive and negative predictive values of 61.1% and 94.0% respectively, in the test
sample set, regardless of HIV infection status or study
site.
Conclusion: We have identified a host serum biosignature with promise in the diagnosis of TB disease
regardless of HIV infection status or ethnicity in Africa.
These results hold promise for the development of a fieldfriendly point-of-care screening test for TB.
OA-303-04 XpertW MTB/RIF molecular probe

binding characteristics associated with
discordant confirmatory rifampicin resistance
testing
R Berhanu,1 P Da Silva,2 K Schnippel,3 R Kularatne,2
L Scott,4 W Stevens,2,4 C Firnhaber,3,5 C Lippincott1 1Right
To Care, Helen Joseph Hospital, Johannesburg, 2National
Health Laboratory Services, Department of Clinical
Microbiology and Infectious Diseases, University of the
Witwatersrand, Johannesburg, 3Right to Care Research
Department, Centurion South Africa, Centurion,
4
Department of Molecular Medicine and Haematology,
School of Pathology, University of the Witwatersrand,
Johannesburg, 5Clinical HIV Research Unit, Department of
Internal Medicine, Faculty of Health Science, University of
Witwatersrand, Johannesburg, South Africa. e-mail:
rebecca.berhanu@righttocare.org
Background: Xpert MTB/RIF (Xpert) is the first-line

diagnostic test for Mycobacterium tuberculosis (MTB)
infection and rifampicin resistance in South Africa. Xpert
rifampicin resistance is treated presumptively as multidrug resistant tuberculosis (MDR) while awaiting
confirmatory line probe assay (LPA) and phenotypic
S108
drug-sensitivity testing (DST). No guidance exists

regarding management of patients with discordant
results. We aim to describe factors associated with
discordant confirmatory rifampicin resistance.
Table Patients with rifampicin resistance diagnosed by Xpert
MTB/RIF and subsequent discordant rifampicin resistance
compared to patients with confirmed rifampicin resistance.
RIF
resistantConTB
firmed
patients
RIF
initiating
resistreatment tance*
n 270 n 225
n (%)
n (%)
Discordant RIF
resistance
n 44
n (%) P value
Male
129
(47.8)
105
(46.5)
24
(54.6)
0.3
HIV infection
218
(81.6)
185
(82.6)
33
(76.7)
0.4
Median CD4 (IQR)

Xpert probe results
available
Xpert quantitative value
Very low
Low, medium, high
Detection method
Probe drop-out
Probe delay DCT 74
Median CT (IQR)
Xpert probe
Probe B
Probe E
Probe A, C, or D
DCT#
DCT 44.9
DCT 75
Median DCT (IQR)
82
86
67
(25173) (25169) (16195)
0.7
178
(65.9)
142
(62.8)
36
(81.8)
0.02
28
(15.7)
150
(84.3)
17
(12.0)
125
(88.0)
11
(30.6)
25
(69.4)
0.006
139
(78.1)
39
(21.9)
18.75
(14.7
22.4)
131
(92.2)
11
(7.8)
18.35
(14.4
21.8)
8
(22.2)
28
(77.8)
22.15
(17.7
28.95)
,0.001
35
(19.7)
66
(37.1)
77
(43.3)
17
(12.0)
58
(40.8)
67
(47.2)
18
(50)
8
(22.2)
10
(27.8)
,0.001
19
(67.9)
9
(32.1)
4.45
(4.25
5.35)
0.002
20
1
(52.6)
(10.0)
18
9
(47.4)
(90.0)
4.8
8.0
(4.47.6) (6.09.1)
,0.001
,0.001
*Any test confirming RIF resistance, i.e., second Xpert, line probe assay,
phenotypic DST.
Discordance defined as two or more RIF sensitivity tests with discordant

results (Xpert to Xpert discordance, Xpert to LPA discordance or Xpert to
phenotypic DST discordance).
Difference of proportions calculated using v2, difference of continuous

variables calculated using Wilcoxon rank-sum (Mann-Whitney test), P value
presented.
P-value ,0.05 (two-sided, .95% confidence).

#
Delayed binding of a Xpert MTB/RIF probe with DCT . 4 (cut-off for
resistance in the Generation 4 cartridge).
RIF rifampicin; TB tuberculosis; HIV human immunodeficiency virus;
IQRinterquartile range; CT cycle threshold; DCTdifference between the
maximum and minimum probe CT value; DST drug susceptibility testing.
Design/Methods: We retrospectively reviewed a cohort

of patients with Xpert rifampicin resistance who were
referred to an outpatient, decentralized drug-resistant TB
treatment facility in Johannesburg, South Africa between
June 2012 and January 2015. LPA, DST or repeat Xpert
were considered confirmatory tests of rifampicin resistance. Patients without a confirmatory test were excluded. The burden of MTB is reflected by Xpert version G4s
cycle threshold (CT). The hybridization pattern of the
five molecular beacon probes (A-E) was classified as
dropout (no hybridization) or delayed (DCT 7 4).
Results: Among the 270 patients included in the analysis,
129 (47.5%) were male, and 218 (81.6%) were HIVinfected with a median CD4 count of 82 cells/mm3.
Rifampicin resistance was confirmed in 226 (83.7%)
while discordance, defined as one or more confirmatory
tests indicating rifampicin sensitive TB, occurred in 44
(16.3%). Xpert probe dropout (139, 78.1%) occurred
more commonly than probe delay (39, 21.9%). Discordance was associated with Xpert probe delay (P , 0.001)
and very low MTB burden (CT .28) (P 0.006)
Discordance was 12.74 times as likely to occur (95%CI
6.18 - 25.19) in patients with probe delay compared to
probe dropout. The median CT value in patients with
discordance was higher (22.15, IQR 17.728.95) than in
those with confirmed rifampicin resistance (18.35, IQR
14.421.8) (P , 0.001). Of the 39 patients with probe
delay, DCT 4-4.9 had a greater association with
discordance than DCT 7 5 (P 0.02). Probe B (50%)
and Probe E (22.2%) accounted for the majority of
discordant results.
Conclusion: Clinicians and national TB programs using
Xpert need to be aware of the association of delayed
Xpert probe binding with discordant rifampicin resistance. Further research is needed to determine if delayed
probe binding with discordance is a marker for false
positive Xpert rifampicin resistance, infection with
mixed strains of MTB or another cause.
OA-304-04 Evaluation of Genotype MTBDRsl v2

line probe assay for the detection of secondline Mycobacterium tuberculosis resistance in
culture isolates
N Ismail,1 Y Gardee,1,2 A W Dreyer1,2 1National Institute for
Communicable Diseases, Johannesburg, 2University of
Witwatersrand, Johannesburg, South Africa. Fax: (27) 118
855 339. e-mail: yasming@nicd.ac.za
Background: South Africa is listed by WHO as 4th in

Africa amongst the 22 high burden TB nations globally.
Studies have indicated that rapidity in diagnosis of Drug
Resistant (DR)-TB would lead to early appropriate and
effective treatment. Phenotypic drug susceptibility testing (DST) is labour intensive with a long turn-aroundtime (TAT) to results. Molecular assays offer the
advantage of speed. WHO Expert Group 2013 conclusions on Genotype MTBDRsl v1 were that the assay
cannot replace phenotypic DST but it may be used as a
supplementary test for DST for fluoroquinolones (FLQ)
and second-line injectable anti-TB drugs (SLID) as a
rule-in test for XDR-TB. Genotype MTBDRsl v2 LPA
has been redesigned for the detection of FLQ (gyr A &
gyrB) and SLID (rrs & eis) resistance mutations.
Design/Methods: Objective: Evaluate the diagnostic

performance of MTBDRsl v2 for the dectection of
secondline drug resistance in M. tuberculosis Isolates.
Methods: A total of 134 M. tuberculosis Isolates were
analysed by MTBDRsl v2 using manufacturers instructions. The collection comprised of 78 well-characterized
M. tuberculosis Quality Control (QC) strains used in the
WHO Supra National Laboratory Network (SRLN) and
58 Rifampicin-resistant laboratory isolates. Sensitivity
and specificity of MTBDRsl v2 was determined with
BACTEC MGIT 960 DST as a reference method, WHO
2013 recommended critical concentrations were used for
DST.
Results: LPA and DST had (134/134) (100%) results
available. The sensitivity and specificity for MTBDRsl v2
was calculated by means of a 2X2 table. FLQ sensitivity
was 98.96% and specificity 97.37%, SLID sensitivity
was 100% and specificity 100%. Individual SLID results
were respectively, Amikacin sensitivity 100% and
specificity 100%, Kanamycin sensitivity 100% and
specificity 100% and Capreomycin sensitivity was
96.33% and specificity 100%. 36 isolates had gyr A
mutations, 6 isolates gyr A MUT3C/D94G mutations.
No gyr B mutations were found. 10 isolates were FLQ
hetero-resistant. 21 isolates had rrs MUT1/ A1401G
mutation with 3 isolates showing low-level Kanamycin
resistance, eis WT2 absent.
Conclusion: MTBDRsl v2 perfomed well with high
sensitvity and specificity for FLQ and SLID. We
recommend that if mutations are present true resistance
is shown. The rest would require confirmation. Including
this test in the National TB Program would aid in faster
diagnosis and appropriate treatment for DR-TB.
OA-305-04 Pyrazinamide drug susceptibility

testing predicts smear and culture conversion
in MDR-TB
G E Velasquez,1,2 R I Calderon,3 M Becerra,3,4,5
C C Contreras,3 R Yataco,3 J Galea,3 L Lecca,3
M B Murray3,4,5,6 1Brigham and Womens Hospital, Division
of Infectious Diseases, Boston, MA, 2Massachusetts General
Hospital, Division of Infectious Diseases, Boston, MA,
3
Partners In Health / Socios En Salud, Lima, 4Harvard Medical
School, Boston, MA, 5Brigham and Womens Hospital,
Division of Global Health Equity, Boston, MA, 6Harvard T.H.
Chan School of Public Health, Boston, MA, USA. e-mail:
gvelasqu@post.harvard.edu
Background: The relevance of phenotypic pyrazinamide

drug susceptibility testing on MDR-TB treatment outcomes is not well understood. Here we evaluate the
association between pyrazinamide resistance and sputum
smear and culture conversion among multidrug-resistant
tuberculosis patients.
Design/Methods: We enrolled patients with active
tuberculosis in two health regions of Lima, Peru from
September 2009 through August 2012. We obtained
sputum smears and cultures at baseline and at two
months of therapy, and tested for pyrazinamide resistance using the modified Wayne method (pyrazinamidase
assay). We restricted our analysis to patients with
S109
multidrug-resistant tuberculosis who were treated with

first-line drugs during the first two months of treatment.
We constructed a multivariate logistic regression model
to test the association between baseline pyrazinamide
resistance and two-month smear and culture positivity,
adjusting for age, sex, and baseline ethambutol resistance.
Results: Three hundred and eleven multidrug-resistant
tuberculosis patients were eligible for the study, of whom
157 (50.5%) were resistant to pyrazinamide at baseline.
Pyrazinamide resistance was associated with higher odds
of two-month smear positivity (OR 2.31; 95%CI,
1.40-3.81) and two-month culture positivity (OR 4.27;
95%CI, 2.36-7.72) after we adjusted for age, sex, and
baseline ethambutol resistance.
Conclusion: We demonstrate that pyrazinamide resistance as measured by the modified Wayne method
(pyrazinamidase assay) is predictive of lower two-month
sputum smear and culture conversion among multidrugresistant tuberculosis patients. Our findings reinforce the
clinical value of including routine pyrazinamide drug
susceptibility testing in baseline investigations for patients with active tuberculosis.
OA-306-04 An amplicon-based next generation

sequencing platform for the Illumina MiSeq for
10 genes involved in first- and second-line drug
resistance
D Operario,1,2 S Heysell,1,2 A Koeppel,1,2 E Houpt,1,2
S Turner,1,2 S Pholwat,1,2 S Banu,3 G Kibiki4 1University of
Virginia, Charlottesville, VA, 2University of Virginia,
Charlottesville, VA, USA; 3International Centre for Diarrhoeal
Disease Research, Bangladesh, Dhaka, Bangladesh;
4
Kilimanjaro Clinical Research Institute, Moshi, Tanzania.
e-mail: scott.heysell@gmail.com
Background: Commercially available line probe assays

for detection of mutations in M. tuberculosis drug
resistance genes are of insufficient sensitivity. Emerging
reports describe specific amino acid changes within
resistance determining regions (RDRs) correlating with
low/high level phenotypic resistance. Conventional
sequencing (Sanger) provides more precise amino acid
data, but is expensive, labor intensive, and may not
provide a precise indication of a mixed genotype. Thus,
we developed a simplified, high throughput protocol for
sequencing on the Illumina MiSeq for rapid characterization of RDRs impacting key first and second-line antiTB drugs.
Methods: An amplicon-based sequencing assay was
designed for 12 RDRs in 10 genes: inhA, katG, rpoB,
embB, pncA, gyrA, gyrB, rpsL, rrs, eis. 17 amplicons
were employed to cover a read length of 150 to 300 base
pairs, over a total of 2,924 bases of the M. tuberculosis
genome. After quality filtering, sequencing trimming,
and alignment to a custom reference sequence (H37Rv),
potential mutation loci were called using a haplotypebased variant detection algorithm (freebayes). We then
tested this platform on DNA extracts from phenotypically diverse M. tuberculosis isolates across our collab-
S110
orative study sites (Bangladesh, Siberia, Tanzania, Thailand) and compared these results to Sanger sequencing.
Results: 229 M. tuberculosis isolates, the majority MDR
by phenotypic drug-susceptibility testing, were sequenced on the Illumina MiSeq. Turnaround time was
approximately 60 hours. A minimum coverage of 100x
was achieved in 91%, yielding over 5 million highquality (7Q30) reads for multiple samples within a
single run. In a subset of isolates (n 109) with complete
Sanger sequencing, accuracy for the MiSeq was .95%
for the majority of gene RDRs (inhA, katG, rpoB, rpsL,
gyrA, gyrB, rrs, eis) but less so for pncA (89.8%) and
embB (83.3%). The depth of sequencing result on the
Illumina MiSeq revealed mixed genotype not detected by
Sanger.
Conclusions: Next generation sequencing platforms,
such as this developed for the first time on the Illumina
MiSeq, have the potential to more specifically ascribe
variants in drug-resistance present at unprecedented
levels of detection, but must ultimately be coupled with
studies of quantitative phenotypic resistance and sound
clinical epidemiology. Currently, this high throughput
protocol may be envisioned for reference laboratories to
track drug-resistance trends at the population level.
constitutional symptoms had higher TQVM, quantitative culture, and AFB concentrations (all P 6 0.006).
Adverse treatment outcomes occurred in 9.4% (3/32)
patients, all of whom had multi-drug resistant tuberculosis (MDR-TB) but outcome was not predicted by pretreatment results for TQVM, quantitative culture or
AFB. 6.4% (13/209) of household contacts developed TB
disease during follow-up for median 6.3 years (IQR 6.56.5). Contacts of patients with lower than median
TQVM results in sputum were more likely to develop
TB disease (univariable cox regression hazard ratio
HR3.8 P 0.03). This association persisted after
adjustment for disease severity, chemoprophylaxis,
drug-resistance and social determinants (adjusted
HR4.0, P 0.03, Figure).
Conclusions: A minority of TB in pre-treatment sputum
were viable, whether assessed with TQVM or culture.
Although, TQVM was highest in patients with more
severe constitutional symptoms, this was not associated
with treatment outcome. Paradoxically, patients with
more TQVM-positive cells in their sputum were less
infectious to their household contacts. TQVM-negative
cells may be slowly metabolising TB that is better
adapted to survive airborne transmission and cause
infection.
OA-307-04 The role of sputum TB quantitative

viability microscopy to predict patient
infectiousness
S Datta,1,2,3 J Sherman,1,2 T Valencia,1,2 M Tovar,1,2
E Ramos,1,2 R Gilman,4 C Evans1,2,3 1IFHAD: Innovation for
Health and Development, Lima, 2Innovacion Por la Salud Y el
Desarollo (IPSYD), Lima, Peru; 3Wellcome Trust Centre for
Global Health Research, London, UK; 4Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, USA.
e-mail: sumo.datta31@gmail.com
Background: Close contacts of people with smearpositive pulmonary TB are recommended to undergo
contact investigation. However, acid-fast sputum microscopy cannot accurately assess infectiousness, and
does not differentiate live from dead TB. TB quantitative
viability microscopy (TQVM) with fluorescein diacetate
predicts the amount of culturable Mycobacterium TB in
sputum and rapidly assesses early response to TB
treatment but implications for infectiousness are unknown.
Objective: To evaluate sputum TQVM for predicting
treatment outcome and infectiousness.
Methods: 35 newly diagnosed, sputum smear-positive
patients in Lima, Peru, and 209 of their household
contacts were recruited. All patients gave pre-treatment
sputum samples that underwent laboratory decontamination with sodium hydroxide and were tested with
TQVM, quantitative culture and auramine-microscopy
to quantify acid-fast bacilli (AFB). Patients and contacts
were interviewed for demographic and clinical data at
recruitment, and followed-up for 6 years.
Findings: TQVM concentrations were median 5.1%
(interquartile range IQR 2.4-11%) of AFB concentrations. Quantitative culture results were median 1.8%
(IQR 0.42-2.2%) of AFB concentrations. Patients with
02. Around the world vulnerability takes

many forms
OA-308-04 Effectiveness of tuberculosis
integration in reproductive health on case
notification in 13 provinces of Afghanistan
A Momand,1 M Rashidi,1 G Qader,1 M Seddiq,2
K Ayoubi,2 S M Sayedi,1 M Shefa,1 P G Suarez3
1
Management Sciences for Health, Kabul, 2Ministry of Public
Health, Kabul, Afghanistan; 3Management Sciences for
Health, Boston, MD; USA. e-mail: amomand@msh.org
Background and challenges to implementation: In

Afghanistan, tuberculosis (TB) disease affects mostly
women especially during their reproductive years that
affect their children and families that leave women
marginalized and vulnerable for TB. Chronological data

from 2009-2014 shows that around 62% of TB cases
notified were among women. To overcome the challenges, USAID supported TB CAP and TB CARE projects
assisted NTP to integrate TB control activities in
reproductive health services. The aim of this assessment
was to determine the role of female health care workers
in case notification in Afghanistan
Intervention or response: TB CAP and TB CARE l
assisted NTP to develop and revised Standard Operational Procedures (SOPs) on case detection and diagnosis, Treatment, TB IC, drug management manual and TB
guideline, develop and disseminate 62 000 information,
education and communication (IEC) material, trained
4600 front line female health workers from 2010 2014
in 380 health facilities in 13 provinces to increase the case
detection among women. In January 2015, Challenge TB
and NTP team reviewed and analyzed the data from
2009-2014 , using standardized NTP recording and
reporting questioner, documented all form TB cases and
new sputum smear positive (NSS) cases.
Results and lessons learnt: The study revealed that
number of all form TB cases among women primarily
in reproductive age increased by 45.4% and number of
new sputum smear positive cases increased by 50%
respectively from 2009 till 2014.
Conclusions and key recommendations: The study
determine that the integration of TB in reproductive
health services contributed in increased case finding by
providing institutional capacity and technical leadership
to implement TB control activities in their working areas
therefore we recommend this integrated approach to all
similar settings
OA-309-04 Engagement of workplaces in TB

care: support and achievements in Bangladesh
K Khatun,1 M Siddiqui,1 M Akramul Islam,1
M Chowdhury,1 S Alam,1 S Islam,1 M M Rahman,2
M Quader3 1BRAC, Dhaka, 2National Tuberculosis
Programme, Dhaka, 3Bangladesh Garment Manufacturers
and Exporters Association, Dhaka, Bangladesh. Fax: (88)
029 888 026. e-mail: shayla.dhaka@gmail.com
Background and challenges to implementation: The

garment industry of Bangladesh is a main earning source
from abroad for the last 25 years which provides
employment of about three million people. As employees
spend most of their working hours here, the workplace is
the ideal place to tackle TB. TB interrupts workflow,
lowers productivity and raises the direct and indirect
costs such as the replacement and retraining of workers.
To address this problem, a partnership programme was
initiated in 2010 between BRAC and BGMEA (Bangladesh Garment Manufacturers and Exporters Association) with the support of NTP. BGMEA posseses more
than 4000 factories with 2.4 million workers.
Intervention or response: A total of 11 DOTS centers are
established in health centres of BGMEA to diagnose and
treat TB cases among workers under DOTS strategy.
Presumptive TB cases are tested for sputum microscopy;
negative cases that have other TB symptoms get
S111
investigation support for other test (such as X-ray,

FNAC, Biopsy). All TB cases are then registered under
NTP and treat accordingly. Orientation on TB has also
been conducted to raise awareness among factory owners
and workers.
Results and lessons learnt: In the BGMEA working area
through 208 orientation events, 10 377 and 10 351
workers and managers/owners of different garment
factory were orientated on TB in 2013 and 2014
respectively. In 2013 and 2014, 619 and 838 TB cases
were diagnosed and registered. Among them 283 and
362 were new smear positive, 133 and 168 smear
negative and 176 and 255 were extra pulmonary TB
case. Treatment success rate was 93.05 % and 95.05%
among new smears positive case in 2012 and 2013
respectively. In 2013, 38 presumptive cases are given
social support and 34 persons are diagnosed as TB
whereas in 2014, 98 persons were diagnosed as TB
against 161 social support. Regarding patient referral, in
2013, 593 patients are referred by BGMEA staff, 24
patients by Government Hospital and 2 patients by
Government field staff. In 2014, 663 patients are referred
by BGMEA staff, 159 patients by Government Hospital,
10 by private practitioners and 3 by cured TB patients.
Conclusions and key recommendations: Establishment
of DOTS centres in workplaces, different orientation to
factory workers and to the management authority, and
social support to presumptive TB cases increased case
notification, treatment adherence and treatment success
rate.
OA-310-04 Improving case notification in the

elderly and other vulnerable communities in
Cambodia using chest X-ray and XpertW MTB/
RIF testing
M Chry,1 K Mom,1 C Andrew,2 C Jacob,2 L Gerstel3
Cambodia Anti-Tuberculosis Association, Phnom Penh,
Cambodia; 2Stop TB Partnership, Geneva, Switzerland;
3
Royal Tropical Institute, Amsterdam, Netherlands. e-mail:
rath@thecata.org.kh
Background: Since 2000, the TB burden in Cambodia

has been reduced by 50%. Despite this impressive
achievement, the country still has one of the highest TB
prevalence rates in the world (748 cases per 100 000
population in 2013). The elderly (aged 755 years)
account for an estimated 50% of the TB burden, yet this
group has low case notification rates owing to poor
access to health services.
Methods: Active case finding (ACF) using mobile teams
equipped with a chest X-ray (CXR) and GeneXpert
machine was conducted in 4 operational districts. Staff
set up ACF days at existing health facilities and screened
people using a TB symptom questionnaire and CXR.
Those with a positive screen were then tested on the
Xpert MTB/RIF assay and detected patients were
initiated on appropriate treatment. Village health support groups (VHSGs) mobilized the communities around
each health facility to ensure high screening volumes
during ACF days. Though the elderly were targeted, all
S112
community members were eligible for screening. Project

yield and TB notification data were analysed to assess the
impact of activities on treatment initiation.
Results: Across four districts, 76 health facilities were
visited by the mobile teams on ACF days, resulting in the
screening of 9,260 individuals for TB symptoms and by
CXR. 1756 (19.0%) individuals were then tested using
the Xpert MTB/RIF assay, resulting in the detection of
324 (18.5%) MTB-positive patients. New bacteriologically-positive notifications increased 119.3% for all
ages and 266.2% for those 755 years during the ACF
quarters compared to trend expected notifications. The
temporary ACF activities had a more lasting effect than
anticipated, as notifications were above trend three full
quarters post-intervention. When extending the postACF period to 12 months to capture this effect, new
bacteriologically-positive notifications were 54.0% and
109.2% higher than expected for all ages and 755
years respectively. This translates into an estimated 361
TB patients of all ages and 263 patients aged 755 years
being treated who would not have received care in the
absence of the ACF intervention.
Conclusion: ACF using mobile teams was able to
increase access to health facilities in vulnerable communities, particularly among the elderly, and to detect and
treat many previously missed TB patients. In settings
with low notification rates, targeted ACF interventions
such as this should be prioritized to reduce the pool of
prevalent TB patients.
Interventions: Regarding health care services for floating

population, free chest radiograph and sputum-smear
microscopy examinations were provided in TB dispensaries, while health examination screening for TB in
designated spots. Patients were given free first-line drugs,
basic follow-up examinations, as well as transportation
and nutrition subsidies during treatment course. Medical
service time was prolonged. Group and individual
psychological counseling were provided to target cases.
With enhanced health service and TB surveillance system
for floating population, TB dispensary monitored epidemic, traced cases referred by other health institutions,
and provided health care services. DOT was conducted
by primary health care center. Specific measures included
establishing mechanism of inter-sector coordination and
technical exchange, cross-area management procedure,
hiring full-time health care staff, increasing health stuffs
subsidies of DOT, conducting training and health
promotion, as well as cooperating with NGOs.
Results and lessons learned: 165 529 TB patients were
detected and treated in total, of which 43.7% were smear
positive cases. The case registered rate was increasing
from 26.2/100 000 in 2007 to 71.4/100 000 in 2011,
higher than the national average level. The rate of cases
arrived in TB dispensaries was increased form 65.5% to
87.1%. The successful treatment rate was increased from
79.2% to 91.5%, while the default rate decreased from
16.1% to 2.9%. The total satisfaction rate of patients
and acceptability rate of health stuff regarding the
control mode was respectively 78% and 89%. Critical
to those effective outcomes were government commitment, sufficient financial support, well-constructed
health system and quality control mechanism.
Conclusions: Significant achievements of TB control
mode for floating population were observed in China:
the case detection and treatment adherence had been well
improved. This control mode should be enhanced and
expanded nationwide with sustainable government
commitment.
OA-312-04 Risk factors associated with

acquisition of tuberculosis in two Colombian
prisons
OA-311-04 Exploration of methods of TB for
internal migrants: the experience of China
J Li,1 X Liu,1 S Jiang,1 H Zhang,1 L-X Wang1 1National
Center for TB Control and Prevention, China CDC, Beijing,
China. e-mail: lijun@chinatb.org
Background: Floating population in China is increasingly

one of major challenges in TB control, with the number
increased to 2360 million in 2012. Considering the low
case detection, poor treatment adherence and high
contagious risk among them, China had explored TB
control mode for floating population from October 2006
to June 2012, totally covered 198 counties/districts and
75 cities in 29 provinces, supported by the Globe Fund
and local government.
L Arroyave,1 L Lopez,
D Marin,1 M P Arbelaez,1
3
2 1
Y Keynan, Z Rueda Universidad de Antioquia UDEA,
Medellin, 2Universidad Pontificia Bolivariana, Medellin,
Colombia; 3University of Manitoba, Winnipeg, MB, Canada.
e-mail: zulmaruedav@gmail.com
Background: Incarceration is associated with high

incidence of tuberculosis (TB) and latent tuberculosis
infection (LTBI). We wanted to measure the incidence of
latent tuberculosis infection (LTBI) and identify risk
factors associated with it.
Design/Methods: In a previous study, 829 male inmates
from two prisons in Medellin and Itagui, Colombia were
evaluated for prevalence of LTBI, and 186 prisoners were
negative to the first two step TST administration. We
followed people with negative TST for two years. TST
were administered according to CDC guidelines and
were not applied in persons with previous or current

active TB, severe adverse event with prior TST administration, immunosuppressive treatment and administration of live vaccines (MMR, varicella or the live
attenuated influenza vaccine) in the 4 weeks before
TST application. A TST result of 10 mm induration and
an increase of 6 mm between baseline and follow-up
application was considered as a converter. The adjusted
relative risk and their 95% confidence intervals were
estimated using the binomial regression adjusting by
standard errors.
Results: Among 120 people that we could follow, 100
remained incarcerated and 20 were released. The overall
percentage of TST conversion over two years was 23.5%
(28/119). The incidences of conversion were: in Prison
One-34% (22/65), there was no conversion in Prison
Two, and 30% (6/20) among people that were released
(those converters were incarcerated in Prison One). The
risk factors for LTBI conversion were history of contact
with a TB case (RR: 1.56; 95%CI 1.51-1.61), inhaled
drugs (RR: 1.31; 95%CI 1.26-1.37), BMI .25kg/m2
(RR: 1.49; 95%CI 1.48-1.50), and being incarcerated in
Prison One (RR: 1.11; 95%CI 1.08-1.14). During
follow-up, 7 inmates developed active TB (six had a
positive TST by two-step method and one had a previous
negative TST).
Conclusion: Active TB transmission in prisoners is high,
and it differed between prisons. The data suggests that
some LTBI identified in the initial screening are actually
active TB thus contributing to transmission. It is
important to apply TST by two-step method and
follow-up prisoners because of the high incidence of TB
infection, and they benefit of the administration of
isoniazid after ruling out active TB.
OA-313-04 Effectiveness of an alcohol

intervention strategy among TB patients: a
study from South India
S113
a community participatory approach was prepared to

guide the sessions. This involved one to one counseling
sessions and visual aids were used to explain facts about
TB and the effects of alcohol use on TB treatment and
management
Results: Of the total 872 registered TB patients, 298
(31%) were found to have AUD (alcohol use disorders).
The number of TB patients in the experimental and
control arms were 113 and 185 respectively. Percentage
with favorable treatment outcomes (cured/completed
treatment) was significantly higher (v2 20.8, 1 d.f., P ,
0.001) in intervention (87%) compared to the control
group (62%). It was also observed that the overall
adherence to TB treatment was significantly higher (v2
35.1, 2 d.f. P , 0.001) among the intervention group
when compared to the control group. Among the
intervention group, 88% completed more than 2
intervention sessions. Percentage of patients with favorable treatment outcomes increased with the number of
interventions attended (v2 43.3, 4 d.f. P , 0.001). It
was seen that the proportion with regular consumption
of doses till the end of treatment increased with the
number of interventions attended (v2 44.6, 4 d.f. P ,
0.001). Participants expressed interest towards the
effectiveness of the interventions especially with regard
to the visual aids that were used and the time taken for
the sessions
Conclusion: Alcohol intervention among TB patients is
effective in ensuring favorable treatment outcomes and
TB treatment adherence. This intervention was feasible
and acceptable to patients. This study suggests the need
for trained counselors in TB care settings to ensure
effective alcohol intervention strategies to ensure TB
treatment compliance among TB patients with AUDs
OA-314-04 TB screening of hard-to-reach

populations at household level with volunteer
field workers in Umzinyathi, Kwa-Zulu Natal
B Thomas,1 B Watson,1,2 E Senthil Kumar,1 S Chandra,1

A Deepalakshmi,1 A Dhanalakshmi,1 C Manogaran1
1
Social and Behavioral Research, National Institute for
Research in Tuberculosis, Chennai, 2Statistics, National
Institute for Research in Tuberculosis, Chennai, India. Fax:
(91) 442 836 2528. e-mail: beenaelli09@gmail.com
D Turner,1 S Shenoi,2 R P Brooks,2 M Ma,2 L Ngubane1

1
Umvoti AIDS Centre, Greytown, South Africa; 2Yale School
of Medicine, New Haven, CT, USA. Fax: (27) 033 413 4726.
e-mail: derek@umvoti.co.za
Background: The influence of alcohol with regard to

delays in seeking care and non-compliance for treatment
has been well documented. There is an urgent need to
develop a feasible, acceptable alcohol intervention
strategy among TB patients to ensure successful treatment completion and a better quality of life. It is against
this background that this study was conducted.
Methodology: This is a two-arm randomized controlled
study in which 4 zones of the Chennai Corporation were
randomly selected and allotted (2 each) to the experimental and control arms. TB patients (.18 years)
registered from August 2013 to January 2014, under
the RNTCP, were assessed using AUDIT scale (.8 to
,20). The intervention consisted of 4 individual
counseling sessions (0, 2, 4 and 6th month) conducted
by trained interventionists. An intervention manual using
Background: South Africa is facing multiple issues in

reaching deep rural populations with targeted TB
investigation at a household level for those with recent
TB contacts. Due to the nature of clinic based care
options, many households are unable to access services,
particularly men and non-ambulant household members,
either due to costs involved, distance to facilities or
stigma.
Intervention: We trained semi-literate community volunteers to screen for TB, HIV and Diabetes at a household
level using a 10 question tick sheet covering major signs
and symptoms as well as asking for known household TB
contact in the previous year. The aim was to screen only
households where the care givers knew family members
to be ill or to have TB signs and symptoms. In all 37 care
givers screened 338 people in 145 households over a six
S114
week period, with the intention of reaching those who do

not travel to centralised clinics, in particular men.
Findings: Of 219 people with no active TB, 31% of
whom were male, 44 had had contact with a TB patient
at household level in the previous year and 66 were
screened positive as TB suspects. Known HIV prevalence
was 25.8% with a further 9.3% showing multiple signs
of HIV and Diabetes prevalence was 11.7% with a
further 33% screened as suspects. Of 47 people screened
who are known active TB cases, 44.7% were men, 85%
had a TB contact in the previous year and 39 screened
positive for all major signs and symptoms of TB
treatment (cough, night sweats and/or fever and persistent weight loss).
Conclusion: The care givers were able to identify and
screen households with active and suspected TB whilst
gaining access to traditionally hard to reach population
groups, with care givers screening high percentages of
household members with multiple signs and symptoms of
TB, HIV and Diabetes. For those with active TB,
indicators suggest that the majority is likely to be still
infectious, with high percentages of major symptoms in
evidence, indicating adherence and conversion issues
which are reflected ib DoH statistics for the District.
Results indicate that disclosure of status and willingness
to screen is significantly higher than usually found in
community case finding initiatives due to the nonthreatening and informal screening process as well as
the familiarity of the field worker with their community
households.
indicated if fluorography shows abnormalities. Followup care algorithms for detainees with presumptive TB are
poorly developed and initiation of treatment is often
significantly delayed.
Intervention or response: PATH in partnership with the
All-Ukrainian Penitentiary Network and in coordination
with the NTP and the State Penitentiary Service
implemented a project Increasing TB Case Detection
in the Ukrainian Detention System with support from
TB REACH. The goal of the project was to improve the
early TB diagnosis and increase the yield of bacteriologically confirmed TB patients among pre-trial detainees in
3 SIZOs in Ukraine, in the regions of Kherson, Mykolaiv
and Lviv through systematic verbal screening for
symptoms and risk factors of TB, followed by the
sputum smear microscopy and further culture/DST, if
indicated and establish prompt referral to timely
treatment.
Results and lessons learnt: Across 12 months, 8361
detainees were verbally screened upon entry and then
quarterly, resulting in the identification of 3280 (39.2%)
individuals with suspected TB who were referred for
further evaluation (smear microscopy, culture/DST). 58
bacteriologically confirmed TB patients were detected
and a further 120 TB patients were clinically diagnosed.
This translates into a TB prevalence of 1435 patients per
100 000 population, more than 10-fold higher than the
national prevalence estimate. All TB patients were
initiated on treatment in SIZO or in civil TB hospitals.
Conclusions and key recommendations: Early TB case
detection among detainees, based on the verbal screening
algorithm, resulted in the significant increase (178 TB
cases comparing 40 during previous 12 months) of the
number of TB cases detected and referred for immediate
treatment. Active interventions to rapidly identify
infectious patients and treat them are essential to
interrupt TB transmission inside and outside the detention settings.
03. Placing the patient at the center of

care
OA-316-04 Patient-centered TB treatment
approach: a cluster randomized controlled trial
in Armenia, 2015
OA-315-04 Early TB case detection in pre-trial

detention centers (SIZOs) in Ukraine
1
S Leontyeva, K Gamazina, O Bogdanov, M Gagarkin

1
Organization for Appropriate Technology in Health, Kyiv,
2
PATH, Kyiv, 3All-Ukranian Penitentiary Network, Kyiv,
Ukraine. Fax: (380) 444 962 627. e-mail:
sleontyeva@oath.org.ua
Background and challenges to implementation: Over the

past decade, Ukraine has improved TB care services in
both the civil and prison systems. However, overcrowding in pre-trial detention centers (SIZOs) has rarely been
prioritized, these facilities fall between the prison and
civil TB care systems. In SIZOs, TB diagnosis is limited to
digital CXR screening upon entry. Smear microscopy is
N Truzyan,1 V Petrosyan,1 A Harutyunyan,1

V Khachadourian,1 M Thompson2 1American University of
Armenia, Yerevan, Armenia, 2University of North Carolina at
Charlotte, Charlotte, NC, USA. Fax: (374) 1051 2566.
e-mail: tnune@aua.am
Background: Tuberculosis (TB) remains a leading public

health concern and an infectious cause for high rates of
morbidity and mortality worldwide. Various approaches
including patient empowerment, education and counseling sessions, and involvement of family members and
community workers have been suggested for improving
treatment adherence and outcome. This study, supported
by Grand Challenges Canada, evaluated if an innovative
multi-component people-centered TB care strategy is as

good as usual care and to measure the impact of
education/counseling on TB patients and their family
members knowledge.
Design/Methods: This open-label, stratified cluster
randomized controlled trial with two parallel arms
involved 395 drug-sensitive TB patients in the continuation phase of treatment from 52 outpatient-TB-centers
(clusters) in Armenia. TB patients in the intervention arm
(195) and their family members (132) participated in a
90-minutes educational/counseling session to raise their
knowledge, provide physiological support and enhance
treatment adherence. Patients received required medication for one week during the weekly visits to TB-centers
and daily SMS reminders and phone calls to take their
medications, to assure adherence and monitoring for
potential side effects. Control arm patients (200)
followed the WHO-recommended directly-observedtreatment (DOT) in TB-centers. Both groups participated
in baseline and 4-5 months follow-up survey. The
National TB Control Center provided patients treatment
outcomes information.
Results: The initial assessment revealed no difference in
TB treatment outcomes between the intervention and
control groups (P . 0.05), data collection will finish in
May 2015. Having similar mean knowledge score
(patients 22.5, family members 22.1, P . 0.05) at
baseline, both TB patients and their family members in
both arms reported substantial improvement in knowledge at follow-up. However, patients and their family
members in the intervention arm after adjusting for
baseline knowledge, demonstrated statistically significantly better knowledge improvement than those in the
control arm (24.0 vs. 23.1; P 0.02 for patients and 24.4
vs. 23.0 ; P 0.002 for family members).
Conclusion: This new strategy can be applied in resource
poor countries to reduce healthcare providers workload,
save time and financial resources that could be potentially used for other TB related services.
OA-317-04 Tuberculosis knowledge, attitudes

and practices survey in an urban poor area with
high TB prevalence rates in Rio de Janeiro,
Brazil
A C Carvalho,1 L M P Oliveira,1 L Isidoro,1 V Trajano,1
M L B C Mello,1 T C Araujo-Jorge,1 J S Garcia,1,2
S C Cavalcante1,2 1Oswaldo Cruz Institute (IOC), FioCruz, Rio
de Janeiro, RJ, 2Municipal Health Secretariat, Rio de Janeiro,
RJ, Brazil. Fax: (55) 21 2562 1432. e-mail:
a.carvalho@libero.it
Background: We designed a study involving tuberculosis

(TB) patients and their families to evaluate the impact of
educational activities in the reduction of treatment
default. Knowledge, attitudes and practices (KAP) survey
was initially performed to collect data to support
educational intervention.
Design/Methods: From Nov 14, patients starting TB
treatment at a public health clinic that assists Rocinhas
community were invited to participate with their
relatives. After written informed consent, they answered
S115
a KAP questionnaire based on WHO KAP model. Two

trained health workers performed the interviews, which
consisted of 31 open and multiple-choice questions.
Results: 54 drug-sensitive, HIV negative, TB patients
were diagnosed and 34 (63%) accepted to participate.
Nine family members were recruited for a total of 43
KAP interviews. Participants were in majority males
(53%), had , 40 years of age (63%) and , 8 years of
school (81%). Family and friends were the source of first
information on TB for 43% of participants, followed by
television (20%), health professionals (7%) and printed
materials (7%). 85% of participants considered TB a
serious disease (very/somewhat serious). Cough was the
symptom more frequently associated with TB (88% of
responses). When asked how a person can get TB, 93%
answered through the air when a person with TB coughs
or sneezes, 65% sharing glasses, dishes and cutlery, and
30% touching items in public places. All knew that TB
can be cured and 88% reported that it happens if you use
specific drugs given by the clinic. 77% considered HIV
infection a condition that increases the risk of TB, 79%
had already met a person with TB and for 80% the
statement closest to their feeling about people with TB
disease was I feel compassion and desire to help. Most
participants (56%) considered themselves well informed
about TB, but almost all (98%) would like to receive
more information on it. We finally asked TB patients
what worries them most when they think about TB: the
fear of transmitting TB to others and the wish to
complete TB treatment properly were the most frequent
answers (Figure).
Conclusion: These results indicate that TB is part of the
communitys daily life, is perceived as a serious illness,
and although knowledge about appropriate treatment is
good, modes of transmission are still confused. The
interest of patients and families in learning more about
the disease makes us hopeful about good acceptance and
participation in educational activities that are being
implemented. Figure: Word cloud of TB patients
response to the question What worries you the most
when you think about TB?(www.wordle.net)
S116
OA-318-04 Counselling is associated with

improved adherence to early phase treatment
for multidrug-resistant tuberculosis
B Sengupta,1 K Roy,2 A Dutta,3 S Sahoo3 1CARE India, New
Delhi, India; 2CARE USA, Atlanta, GA, USA; 3Asian Institute
of Public Health, Bhubaneswar, India. e-mail:
bsengupta@careindia.org
Background and challenges to implementation: Compliance to Multidrug-resistant Tuberculosis (MDR-TB)

treatment is a challenge to the patient and the health
system for its duration and complexity. Programmatic
Management of Drug-resistant TB (PMDT), India is
implemented to address the MDR-TB problem. CARE
India with support from Lilly MDR-TB partnership with
PMDT provided psychosocial counselling support to the
MDR-TB patients from 2011in the DOTS Plus site of KS
Roy TB Hospital in West Bengal to increase treatment
adherence, which was evaluated through this study.
Intervention or response: The study included 265
patients from KS Roy TB Hospital as the Formally
Counselled (FC) group and 223 from SCB Medical
College, Orissa as the Not Formally Counselled (NFC)
group; those initiated on treatment between January
2011 and June 2013. Data included age, gender, weight,
MDR-TB suspect criteria , number of doses missed
within first six months and days taken to undergo first
sputum follow-up culture and start continuation phase
from initiation.
Results and lessons learnt: The FCs were mostly
diagnosed using suspect criterion B compared to A in
NFCs; also had less males and had lower average age.
Fifty-eight percent FCs (n 153) did not miss a single
dose of treatment to 32% (n 71) NFCs; those missing
more than 5 doses were 13% and 34% respectively. The
adjusted risk of missing doses was 1.84 times more in
absence of counseling (1.40 to 1.66, P , 0.001). The
median number of days for first follow-up sputum
culture was 103 days and 112.5 days in the FCs and
NFCs respectively ; the adjusted mean difference being
13.81 (6.40-21.39, P , 0.001). The median number of
days to start continuation phase was 205 in FCs vs. 260
in NFCs, the adjusted mean difference being 41.32 (17.
91-64.69, P , 0.001).
Conclusions and key recommendations: The results
demonstrated the beneficial association between counseling and treatment compliance to Drug-resistant TB
patients; hence the overarching recommendation is to
mainstream counselling in the programme.
OA-319-04 Ambulatory vs. hospital-based

treatment for MDR-TB under programmatic
conditions
M Kaela,1 N Ruswa,1,2 F Mavhunga,1 A Zezai2 1Ministry of
Health and Social Services, Windhoek, 2KNCV Tuberculosis
Foundation, Windhoek, Namibia. Fax: (264) 61 252 740.
e-mail: ncruswa@yahoo.com
Background and challenges to implementation: Namibia

is Southern African country with a high incidence rate of
TB. Patients with MDR TB in Namibia undergo
hospitalization in the intensive phase of treatment

followed by one or more forms of ambulatory treatment.
The WHO guidelines advocate for ambulatory care for
MDR TB, but this has sparked local debate on the risks
and feasibility. Grootfontein District in Namibia is home
to the San bushman tribes who are disproportionately
affected by MDR TB. Due to their nomadic lifestyles and
social structures, hospital admission is viewed very
negatively. The prolonged treatment for MDR TB was
associated with high rates of loss to follow-up (45%).
Intervention or response: In 2010, after local consultations with patients and ACSM interventions, a decision
was made to introduce a community based ambulatory
care model for DR-TB in the Tsumkwe Constituency. No
additional health workers were recruited, but the local
clinic nurses and community health workers were
responsible for patient treatment and monitoring, with
technical support from district doctors and the national
clinical committee. Final treatment outcomes for patients
treated in this district between 2009 and 2012 were
analysed and compared with those from other districts
implementing a hospital-based intensive phase of treatment.
Results and lessons learnt: In 2010 and 2011, 62 patients
with DR TB received regimens in line with WHO
recommendations from the local clinic, but with no
hospital admission. This included an injectable phase of
at least 8 months, for which patients had to present daily
at the clinic. In the same period 553 patients were treated
nationally. Treatment success rates between the 2 groups
were 55% and 57% respectively
Conclusions and key recommendations: There was no
significant difference in MDR TB treatment outcomes
between ambulatory care settings and traditional hospital-based treatment settings. This demonstrates that
ambulatory treatment indeed has a place in countries
like Namibia, with the additional benefit of cutting costs
and maintaining stable families, especially among
vulnerable populations.
OA-320-04 Introduction of out-patient care for

DS-/MDR-TB patients in Tajikistan
I Leimane,1 O Bobokhojaev,2 M Makhmudova,3
Z Abdulloeva,3 S Verver1 1KNCV Tuberculosis Foundation,
The Hague, Netherlands; 2Republican Center of Protection of
Population from TB of Republic of Tajikistan, Dushanbe,
3
KNCV Tuberculosis Foundation, Dushanbe, Tajikistan. Fax:
(31) 70 358 4004. e-mail: ieva.leimane@kncvtbc.org
Background and challenges to implementation: Tajikistan is one of the 27 high burden MDR-TB countries
with 108 new notified cases per 100 000 population.
10% of MDR-TB patients have XDR-TB. The number of
(X) MDR-TB patients continues to rise. The majority of
DS/ MDR-TB patients are poor. Loss-to-follow-up of TB
treatment is common, with a principal reason to migrate
for employment. Patients are hospitalized for long period
of time. Therefore, NTP decided to introduce patients
centered out-patient care model and piloted it in nine
districts.
Intervention or response: In 2013, TB CARE I project
provided TA to NTP to facilitate a shift from in-patient
to out-patient treatment in PHC facilities. Local governments developed implementation plans for psycho-social
support (PSS) and provided quarterly social packages to
DS-/MDR-TB patients. A patients support team was
established at the NTP to provide PSS. Clinical staff,
community activists and religious leaders were trained
on outpatient care protocol, communication skills, and
stigma reduction. A supportive supervision system was
introduced during implementation including on-the-job
training. The DS-/MDR-TB data of 9 districts were
recorded in the Central TB Register and analyzed.
Results and lessons learnt: The proportion of registered
DS-TB patients in out-patient care increased from 20%
(115 out of 579) in 2013 to 58% (311 out of 536) in
2014. The proportion of DS-TB in-/out-patients who
received PSS increased from 24% (45 out of 184 eligible)
in 2013 to 74% (120 out of 162 eligible) in 2014. The
proportion of registered MDR-TB out-patients almost
tripled from 21% (11 out of 52) in 2013 to 58% (38 out
of 66) in 2014. The number of MDR-TB in-/out-patients,
who received PSS increased from 15 in 2013 to 118 in
2014. In 2014, among all registered MDR-TB patients
98% (65) received PSS. In 2013, 57 MDR-TB patients
have been enrolled on out-patient treatment. No patients
were lost-to-follow-up since then. We learned that, since
little local government funding is available, PSS focus
should be on MDR-TB patients.
Conclusions and key recommendations: This approach
gave positive results and contributed to the sustainability
of out-patients care and better adherence to treatment.
The coordination of PSS from local government,
collaboration among medical workers from the PHC
and TB facilities, community activists and religious
leaders are main factors of success. It is recommended
to apply this model elsewhere.
S117
OA-321-04 Patients perceptions of

interventions to improve MDR-TB treatment
completion in the Philippines
T Tupasi,1 A Garfin,2 J Mangan,3 L Naval,1 J Pancho,4
M Mantala,5 E Kurbatova,3 A Golubkov6 1Tropical Disease
Foundation, Inc., Makati City, 2The National Tuberculosis
Control Program, Department of Health, Manila, Philippines;
3
U.S. Centers for Disease Control and Prevention, Division of
Tuberculosis Elimination, Atlanta, GA, USA; 4National Center
for Pulmonary Research, Lung Center of the Philippines,
Quezon City, 5Technical Assistance to the Countries (TASC)
Project, Manila, Philippines; 6U.S. Agency for International
Development, Washington, DC, USA. Fax: (352) 265 7683.
e-mail: bpy4@cdc.gov
Background: As part of an effort to improve treatment

completion rates at programmatic management of drugresistant tuberculosis (PMDT) treatment centers in the
Philippines, a case-control study was conducted. One
aim of this study was to describe patients views on
interventions that would affect adherence to treatment
for multidrug-resistant (MDR) tuberculosis (TB).
Methods: In-depth interviews were conducted with
patients who were lost to follow-up (cases) and patients
who were continuing or had completed treatment
(controls) between April and July of 2014 among
patients with rifampicin-resistant TB who had initiated
MDR TB treatment between July and December of 2012
at 15 PMDT treatment centers throughout the Philippines. Responses to open-ended questions were thematically analyzed.
Results: Interviews were conducted with 91 cases and
182 controls. Suggested interventions could be categorized as either approaches to improve adherence or
approaches to remove barriers. The top three themes that
emerged were: (1) the need for transportation assistance
or improvements to the current transportation assistance
program, (2) food assistance, and (3) difficulties patients
encountered related to their anti-TB medications. These
themes were identified by 63%, 60%, and 32% of
participants (both cases and controls) respectively.
Proposed improvements to the transportation assistance
program included: providing funds to cover the full cost
of transportation to and from treatment centers,
providing funds for companions to accompany patients
who are too weak to travel alone, and eliminating delays
or interruptions in the National Tuberculosis Control
Program-organized travel assistance program. Participants from both groups who advocated for food
assistance suggested that it be provided at centers when
patients receive treatment. Some noted food assistance
would help lessen medication side effects and aid
recovery. Participants from both groups who commented
about medication difficulties expressed their hopes for
medications with fewer side effects, medication that
could achieve a cure in a shorter time, a smaller quantity
of pills to ingest per dose, and treatment that would not
inhibit activities of daily living.
Conclusions: Study findings point to the need for a more
patient-centered approach within the PMDT strategy,
specifically improvements in transportation assistance
S118
and food support, and a need for simpler, less toxic

regimens.
OA-322-04 Patient-centred M/XDR-TB care:

what do patients in urban and rural Peru say?
S Mookherji,1 K Alegria-Flores2 1George Washington
University, Washington, DC, 2University of North Carolina,
Chapel Hill, NC, USA. Fax: (1) 202 994 9193. e-mail:
smookher@gwu.edu
Background: Controlling multidrug and extremely drugresistant tuberculosis (M/XDR-TB) poses a grave challenge to public health, globally. Studies repeatedly show
that effective M/XDR-TB management goes beyond
strategies recommended in global TB control plans,
uniformly pointing to patient-centred care as critical to
progress in controlling the epidemic. We aim to describe
patient-centred care in M/XDR-TB management in Peru,
which presented a context where M/XDR-TB remains a
persistent problem, even with increased capacity for
diagnosis, decreased costs of treatment, and increased
financial and human resources.
Design/Methods: We used a case-based study design,
purposively selecting urban and rural sites with some of
the highest M/XDR-TB prevalence in Peru, employing
multiple qualitative data sources: 58 patient interviews, 5
provider focus groups, and observations in 8 facilities,
which we compared for triangulation and verification.
Current M/XDR-TB strategy recommendations, health
systems factors, and social determinants were the
primary parameters in the analytical framework. Our
approach emphasizes patient voices to identify the
current status and challenges in providing patientcentered care for M/XDR-TB.
Results: Overwhelmingly, patients were demoralized
during the M/XDR-TB diagnosis process. Several factors
contributed to fear, frustration, and depression, highlighting the crucial need for psychosocial support.
Because CHWs and community outreach were available
in one site and not the other, we could identify how these
services made critical positive contributions to patient
experiences. CHWs as part of M/XDR-TB management
and care contributed to patient-centredness and to
disease control by: 1) supporting patients to complete
treatment by offering more convenient DOT and
providing ongoing psychosocial support; 2) reducing
primary transmission of M/XDR-TB and diagnosis
delays with active contact tracing, case finding, and
patient education.
Conclusion: Our findings reinforce that it is not direct
observation of drug ingestion that is important for TB
control, but the continuous support from the health
system and community that can be conveyed through
regular DOT contacts. Patients need interaction, not
observation. If TB programs can provide the right kind of
patient-centred DOT flexible, personal, and convenient
patients say that psychosocial, economic, and motivational barriers to successful M/XDR-TB management are
diminished.
OA-323-04 Adverse drug reactions and

resultant health-related quality of life during
multidrug-resistant tuberculosis treatment in
South Africa
A Kelly,1 B Smith,1 Z Luo,2 J Farley3 1Michigan State
University, East Lansing, MI, 2Michigan State University, East
Lansing, MI, 3Johns Hopkins University, Baltimore, MD, USA.
e-mail: ana.kelly@hc.msu.edu
Background: Multidrug-resistant tuberculosis (MDRTB) treatment has been described as worse than the
illness itself due in large part to the high incidence of
adverse drug reactions (ADRs) from treatment, yet little
is known about the impact of such effects on patients
lives. This cross-sectional, observational study examined
the effect of patient-reported ADRs from MDR-TB
treatment on health-related quality-of-life (HRQOL).
Design/Methods: Patients in the intensive phase of
MDR-TB treatment were recruited from May July
2014 from an outpatient clinic in Durban, South Africa.
Patients were interviewed about the presence of 18
possible ADRs experienced during the last 30 days of
treatment and the degree of bother (rated 1 4) of each.
The EuroQOL group five-dimension questionnaire (EQ5D) was used to produce a HRQOL utility score.
Multiple linear regressions and partial correlations were
used to determine the detriment in utility attributable to
ADRs from MDR-TB treatment.
Results: The sample included 121 MDR-TB patients, 62
(51%) female, and 90 (74%) co-infected with HIV.
Patients were on a standardized MDR-TB regimen with a
median time-on-treatment of 4 months. Insomnia was
the most common ADR, experienced by 83 (69%) of the
patients with 57 (69%) reporting the highest degree of
bother. Patients reported an average of 8.6 6 4.1 ADRs.
An increased number of total ADRs was associated with
a reduction in HRQOL utility (P , 0.001). Six ADRs
were significantly associated with a decrement in utility
and explained 45% of the variance. Gastrointestinal
symptomsnausea/vomiting and bothersome (level 4)
diarrheahad the greatest effect size (0.33), followed
by ADRs affecting movement: bothersome peripheral
neuropathy, arthralgia, and myalgia (0.2 to 0.25) and
lastly, tinnitus (0.2).
Conclusion: In this study, one of the first to quantify the
effect of ADRs on HRQOL during MDR-TB treatment,
ADRs were common and demonstrated a substantial
decrement in utility. These findings will enable clinicians
to provide more patient-centered care as they are tasked
with prioritizing ADR management in low-resource
settings.
04. Law enforcement and awareness

OA-324-04 Mobile court operation for tobacco
control in Bangladesh: a success story of
enforcement
M A U Ahsan,1 A M Alamgir Sikder,1
S M Mahbubus Sobhan1 1Ministry of Health & Family
Welfare, Dhaka, Bangladesh. Fax: (880) 2958 0255. e-mail:
ahsan5965@gmail.com
Background: Mobile court is one of the popular court

which moves to the place of occurrence and the
Executive Magistrates, on the spot, impose penalty for
the violation of provision of laws. Mobile court
operation for the enforcement of tobacco control law is
even popular in Bangladesh. However, the Executive
Magistrates conduct mobile courts on almost 100 laws in
the country, and also most of the time they are busy with
heinous crime control laws, so it is challenging to move
the court for tobacco control many times. Moreover the
Ministry of Health does not have any Executive
Magistrate specific for the enforcement of tobacco
control law.
Intervention: National Tobacco Control Cell (NTCC),
Ministry of Health & Family Welfare has undertaken to
conduct mobile court on tobacco control law throughout
the country during October 2014-March 2015. In the
district level the Executive Magistrates work under the
supervision of Deputy Commissioner (DC) of the
respective districts. As the DC is the chairperson of the
taskforce committee for tobacco control in the district
level, NTCC took the opportunity to use the platform of
the committee and managed to convince the DC to
undertake the program. NTCC has already trained 187
executive magistrates on the tobacco control law. Civil
Surgeon of the respected districts organize at least two
the mobile court during the period and send a report to
NTCC. Mobile court has been conducted in all 64
districts of Bangladesh.
Results and lesson learned: Within these six months
executive magistrates conducted 252 mobile courts
throughout the country. They have fined 1279 persons
and 68 organizations for violation of different provisions
of the tobacco control law. The total amount collected as
fine was 1 515 940 BDT (19 690 USD). Almost all
mobile court could able to earn a number of media
1coverage, which has a multiple effect of the program. A
vast public awareness has done through the activity.
Conclusion: Mobile court is a proven instrument for
enforcement of the provisions of the tobacco control law
in Bangladesh. One mobile court can earn a multiple
effect as it attracts huge media coverage.
S119
OA-325-04 Evaluation of the first national

tobacco control mass media campaign in China
X Jin,1 Y Zhang,1 W Chen,2 W Zhao,1 J Song1 1Chinese
Center for Health Education, Beijing, 2World Lung
Foundation, Beijing, China. e-mail: yaya22nd@hotmail.com
Background: Tobacco use is the leading preventable

cause of morbidity and mortality worldwide, yet the
usage rate continues to increase in low- and middleincome countries (LMICs) such as China. Of over 1.3
billion smokers in the world, about one-third (360
million) lives in China, which accounts for 28% of the
general population.In 2009, it was estimated that
approximate 72% of nonsmokers in China were exposed
to secondhand smoking (SHS), which was much higher
than the worldwide average level (40% of children, 35%
of women, and 33% of men). As a result, tobacco control
in China has recently has garnered domestic and
international interests for the diffusion of evidence-based
policies and programs into the country to forestall the
impending tobacco-induced epidemic. In 2013, the WLF
collaborated with the Chinese Center for Health
Education (CCHE), a national branch of the Chinese
Ministry of Health (MOH), to implement the first
national tobacco control mass media campaign in China.
The campaign was expected to improve knowledge
about health effects of smoking and SHS exposure and
shape attitudes and behaviors toward tobacco control
policies, particularly SFPs. Objective: To evaluate the
effectiveness of the first national tobacco control mass
media campaign piloted by Bloomberg Global Initiative,
World Lung Foundation and Chinese Center for Health
Education in 2013.
Methods: The campaign was adapted from materials
originally produced in the UK, which depicts a hospital
environment with secondhand smoke (SHS) to show how
it affects everyone. It was broadcast on two national
Television channels and seven provincial satellite networks for a 4-week period. A total of 3000 participants
were selected to complete interviews through a stratified
random sampling from ten provinces.The participants
consisted of 1509 (50.3%) males and 1491 (49.7)
females, 1161 (38.7%) smokers, and 1839 (61.3%)
non-smokers.
Results: The campaign reached about 24% of the
Chinese population and increased smokers knowledge
about the harmful effects of SHS exposure for nonsmokers (57.4% vs. 68.5%, P , 0.05). Both smokers and
non-smokers increased their knowledge that smoking
leads to diseases other than lung cancer (Table).
Although smokers exposed to the campaign were more
likely to take actions to reduce smoking than those not
exposed (17% vs. 8%, P , 0.05), the percentage was still
low. The support for bans on smoking in public places
increased after the campaign, especially among nonsmokers and females. However, the support for bans on
smoking in bars (43%) and restaurants (56%) were
much lower than for bans in hospitals (97%), schools
(96%), and public transport (89%).
Conclusions: The mass media campaign reinforced
peoples knowledge and attitudes about harmful health
S120
effects of smoking and SHS exposure, increased peoples

desire to quit, and improved peoples support for
smoking bans in public places. Therefore, policymakers
and the public health communities should dedicate more
resources for health related mass media campaigns
(Table).
Table Effects of mass media campaign exposure on knowledge,
behaviors, and attitudes among smokers and non-smokers (n
3000)
Smoker*
SHS leads to serious illnesses of
important vital organ in nonsmokers
Smoking leads to lung diseases other
than cancer
Non-smoker
Smoking leads to lung diseases other
than cancer
SHS leads to lung cancer in nonsmokers
No exposure
(n 500)
%
Exposure
(n 165)
%
57.4
68.5
74.2
81.2
74.2
81.2
74.6
80.6
* Model adjusted for TVC exposure, number of cigarettes smoked, intention

to quit, age, sex, education, weekly TV viewing, weekly surfing on computer,
weekly surfing on mobile phone.
Significant at P , 0.05.
provide better understanding of its association with

tobacco use. The inclusion criteria has not been
expanded to include institutionalized people (i.e. residential, school and college students, army personnel etc),
internally displaced persons, refugees and migrant
population as they are more vulnerable to tobacco use.
The reasons for initiation and continuation use, knowledge of people regarding provisions of COTPA, 2003,
place for storage/disposal of tobacco products is missing.
Conclusion: As India prepares for next GATS in 2015-16
it is pertinent to remove flaws identified in the prior
round so as to make more efficient and effective tool for
policy makers. As prevalence is not an appropriate
indicator to measure tobacco burden due to influence of
factors such as increase in life span/better cure/ diagnostics etc. Therefore incidence of tobacco use needs to be
incorporated to reveal the desired impact of policy
implementation.
OA-327-04 Making a tertiary health care

institution a model tobacco-free educational
institution: a case study from the Indira Gandhi
Medical College in Shimla, India
R Chand1 1Indira Gandhi Medical College, Shimla, India.
e-mail: dr.rameshigmc@gmail.com
Model adjusted for TVC exposure, age, sex, education, weekly TV viewing,
weekly surfing on computer, weekly surfing on mobile phone.
OA-326-04 Review of the Global Adult Tobacco

Survey, India, 2010: recommendations for the
next round 20152016
G Chauhan1 1Postgraduate Institute of Medical Education &
Research, Chandigarh, India. e-mail:
dr.ankursangwan@gmail.com
Background: The tobacco epidemic is currently a major

cause of preventable disease and deaths globally,
claiming the lives of nearly six million people a year,
killing nearly half of the users. GATS (Global Adult
Tobacco Survey) are monitoring process in response to
WHO FCTC (Framework Convention on Tobacco
Control) to track Tobacco Control indicators like
tobacco prevalence (smoking and smokeless), exposure
to second hand smoke, cessation, knowledge attitude and
perceptions that are critical for control and policy
development. As second round of GATS is proposed in
2015-16 it is high time to address the limitations and take
suitable remedial actions for further improvements to
quality of data.
Design/Methods: GATS 2009-10 has been analyzed and
discussed with the researchers of PGIMER Chandigarh
and implementers within the research framework including the objectives, indicators chosen, methodology,
population and sample size, sampling method, study
tool used, quality assurance and data analysis. The study
tool was discussed with respect to questions which
should be added, deleted or modified in order to improve
response and data quality.
Results: Stratification of the population on basis of socioeconomic status has not been included in GATS to
Background and challenges to implementation: Indira

Gandhi Medical College and Hospital (IGMC) Shimla in
Himachal Pradesh is one of the largest and oldest
teaching institutions in Northern India having 872 bed
strengths and more than 1200 daily OPD registration.
One fifth of adults in the state are tobacco users. Indian
tobacco control legislation (COTPA) bans smoking in
public places and sale of tobacco products in and around
an educational institution; however the implementation
has been sub-optimal. The present case study enumerates
the steps taken by the department of hospital administration to make IGMC- a model tobacco free educational
institution.
Intervention or response: A high level institutional core
committee for tobacco control was constituted. A
strategic and comprehensive action plan was formulated.
Department wise nodal officers were notified to ensure
enforcement in respective departments and buildings.
Circular was shared to all departments, students and
employees association to comply with the law. Signages
as specified under the law were displayed at all
prominent places in campus. Several boards having
anti-tobacco messages were displayed across the hospital
campus, walls and also on OPD siip. Strict law
enforcement was ensured, till date more than 300
violators has been punished and fine amount of Rs
54630/- has been collected.
Results and lessons learnt: There is an overwhelmingly
positive response from all sections of peoples, employees,
students, patients and visitors. The smoking has been
literally banned throughout the campus. Department of
hospital administration is doing regular monitoring of
enforcement activities. The number of people seeking
tobacco cessation services in the TCC is increased; from
January 1, 2014 till date 277 persons has seeked help

from TCC to quit tobacco. There is no tobacco selling
shop inside and within 100 yards of the premise. The
institute qualifies for the model tobacco free educational
institution.
Conclusions and key recommendations: IGMC and
associated hospitals becomes a model for tobacco free
educational institution in India demonstrating successful
implementation of tobacco control policies. Other
tertiary care health institutions and teaching hospitals
in India or in similar settings in other countries can also
be made tobacco free educational institutions.
OA-328-04 Building capacity for tobacco-free

schools in Rural Maharashtra: the Salaam
Mumbai model
D Patil,1 D Chadha1 1Salaam Bombay Foundation, Mumbai,
India. e-mail: deepak.patil@salaambombay.org
Background and challenges to implementation: Tobacco

is a significant public health risk faced by youth in rural
India. In rural Maharashtra, as many as 45.4% of youth
use tobacco. Tobacco use is commonly seen on school
grounds across the state. Indias tobacco control law, the
Cigarettes and Other Tobacco Products Act (2003),
imposes restrictions on the use of tobacco on school
campuses and prohibits the sale of tobacco within 100
yards of school grounds.
Intervention or response: In 2008 the Salaam Bombay
Foundation began a tobacco-free schools initiative in
rural Maharashtra. Awareness meetings were held with
the Direct of State Education and District Education
Officers to build motivation to promote the tobacco-free
schools programme. District-level workshops were held
with Master Trainers for the Education Ministry. Master
Trainers included training on tobacco-free schools as
part of their regular training curriculum. Once trained,
teachers conducted tobacco control initiatives with
students. Steps were taken by teachers, students and
administrators to meet the CBSC criteria.Salaam Bombay Foundation rewards schools for reaching levels of
compliance with the criteria created by SBF in accordance with a education board. Bronze-level awards are
given to schools reaching 4 out of 11 criteria. Silver-level
awards are given to schools reaching 8 criteria. Goldlevel awards are provided to school reaching all 11
criteria.
Results and lessons learnt: To date, 124 schools have
reached the Bronze level of compliance, 56 schools have
reached the Silver level, and 27 schools have reached the
Gold level of compliance (full compliance with CBSC
criteria).
Conclusions and key recommendations: Using the
existing infrastructure of Maharashtras rural teacher
training programme has been a successful and efficient
strategy for increasing compliance with tobacco-free
schools criteria. Using a graded system to reward schools
provides an incentive for schools to expand compliance
and serves as a useful tool for measuring increases in
compliance over time.
S121
OA-329-04 Using litigation to protect and

promote lung health
K Hashmi1 1Society for Alternative Media and Research,
Islamabad, Pakistan. e-mail: hashmi@ctcpak.org
Background and challenges: Tobacco use is chronic to

lung health and Shisha affects direly. Shisha came to
Pakistan in 2000 and in a decades period, it became a
profitable industry. Shisha cafes owners spent millions on
their cafes and ambiance attracted even non-smoker
including youth to use shisha for fun. This perception
attracted all age groups and families. In March 2011, the
Government of Pakistan realizing the threat ordered
crackdowns against Shisha Cafes and take legal action in
the light of Tobacco Control Law in Pakistan. Shisha
cafes owners challenged the actions of authorities to
court, filing different cases and alleging that law
enforcement agencies action are malafide and not be
in the interest of public health.
Intervention or response: Assessing the environment,
Society for Alternative Media and Research (SAMAR) to
support Ministry of Health through a writ petition
became a party in the cases.Over the period of four years,
provided the court with facts on shisha as a tobacco
product and relevant laws. SAMAR also did a mapping
on shisha cafes, gathered evidence on breach of laws by
visiting different cafes in different cities, taking photographic and other evidences to present in the court. The
court dismissed this Petition and others and accepted
SAMARs Writ Petition No. 25011/2011 in a single order
enforcement authorities to strictly enforce that Ordinance which is being be monitored by Provincial Home
Department. Furthermore, the Court vide a reported
Judgment ordered that smoking, including shisha, was
ban in any public place, also declaring open spaces of
restaurants a part of restaurant covered by the definition,
hence, none allowed to smoke in any public place
collectively.
Results and lessons learnt: Litigation is an important tool
to promote lung health. Civil society is an important
component of civil cases. During the 4 years period since
2010, all eight cases related to shisha were were
dismissed/disposed in favor of tobacco control marking
a 100% success rate. The litigation precedent in one
province, set an example for other provinces to follow
(Pakistan has four provinces). One province, Balochistan, also banned shisha products while other, Sindh
province, is in process of enforcing tobacco control laws
on shisha as well where SAMAR will be utilizing this
experience.
Conclusions and key recommendations: Litigation can
be a successful tool for promoting lung health by making
other forms of tobacco more coherent with the laws.
S122
OA-330-04 Population-level tobacco control

communication program best buys to support
the changing public health landscape in lowand middle-income countries
T Turk,1 N Murukutla,1 S Mullin1 1World Lung Foundation,
New York, NY, USA. Fax: (1) 212 542 8871. e-mail:
tturk@worldlungfoundation.org
Background: WHO has identified the need for policy

makers making investments in order to tackle risk factors
for non-communicable diseases (NCDs), to assess
interventions based on their: i) health impact; ii) costeffectiveness; iii) cost of implementation; and iv)
feasibility of scale-up, particularly in resource constrained settings. One of the four interventions identified
by WHO as a best buy is a component of the
MPOWER package designed to Warn about the dangers
of tobacco use. These include hard-hitting, population
level, tobacco control campaigns, delivered through mass
media delivery channels. The concept of best buys for
tobacco control is explored through operational outcome
surveys recently conducted in a number of LMIC
settings.
Design/Methods: National, cross sectional post-intervention knowledge, attitude and behavioural (KAB)
surveys of tobacco users were implemented in a number
of LMIC settings following tobacco control communication campaigns. Key performance indicators included
audience recall of campaign messages, and changes in
knowledge and attitudes toward tobacco use including;
personal risk perceptions, self-efficacy perceptions and
behavioural intentions. Additionally, tobacco users were
assessed according to the number of cessation attempts
of campaign aware vs. campaign unaware audience
segments as well as assessment of average cost per
cessation attempt.
Results: Findings across LMIC settings identified generally high recall of the campaigns by both urban and rural
audiences, with significant differences (,.95%) by
campaign aware respondents in relation to a number of
KAB indicators. Tens of millions of tobacco users made a
quit attempt as a result of the campaigns at very low cost
(less than 1 cent) per cessation attempt. Furthermore,
campaign aware audience attitudes supported strengthened tobacco control policy initiatives. Following advocacy of findings, governments were more fully engaged in
campaign implementation through the provision of
matched funding and institutionalisation of approaches
for tobacco and other NCD risk factors.
Conclusion: Best practice, population level, communication campaigns for tobacco control demonstrate through
compelling evidence that they are interventions that can
have significant public health impact and be highly costeffective, inexpensive and feasible to implement in
resource constrained settings of LMICs. Hard hitting
communication campaigns warning about the harms of
tobacco can be particularly cost-effective and provide
best buys in highly populated LMICs, as a result of
rapidly emerging media networks, relatively low-cost
media rates and more fully engaged viewing audiences.
Findings point to the need to urgently scale up mass
communication campaign activities to reduce the burden

of tobacco related diseases in LMICs.
OA-331-04 Cost-effectiveness of tobacco

control mass media campaigns in LMICs
N Murukutla,1 T Turk,1 N Negi,1 S Mullin1 1World Lung
Foundation, New York, NY, USA. Fax: (1) 212 542 8871.
e-mail: nmurukutla@worldlungfoundation.org
Background: Tobacco consumption is a significant risk

factor for non-communicable diseases, including lung
diseases, that together are estimated to cost low- and
middle-income countries (LMICs) close to 7 trillion USD
in health care costs. Scalable and cost-effective strategies
for tobacco control are therefore a critical public health
priority. Population level mass media campaigns for
tobacco control have been proven to reduce consumption
and prevent tobacco uptake. To date however the costeffectiveness and the return on investment of this
approach has not been established in LMICs. Four recent
tobacco control mass media campaigns in China, India,
Senegal and Vietnam were studied for their costeffectiveness.
Design/Methods: Household survey data measuring
campaign impact was combined with campaign expenditure data to identify costs associated with campaignattributable changes in knowledge, attitudes and behaviors. Logistic regression models, controlling for confounders, calculated differences between campaign
aware and unaware respondents, and through extrapolation, the numbers of individuals in the populations
reflecting campaign-attributable changes in knowledge,
attitudes and behaviors. Sensitivity analyses with success
rates ranging from 10 - 100% were performed for each
outcome. Cost-effectiveness ratios were then calculated.
Results: Campaign awareness was associated with
improved campaign-related knowledge in 7 million
Chinese, 16 million Indians, 247 000 Senegalese, and 1
million Vietnamese. The cost per person to improve
knowledge (at 50% campaign attributable) was lowest in
India at US$0.12 and highest in Senegal at US$1.7.
Campaign awareness was associated with increased quit
attempts among 29 million Indian SLT tobacco users, 4
million Chinese, and 680 000 Vietnamese smokers, with
costs per person (50% attributable) at US$0.07, $0.21,
and US$0.56 respectively, testifying to the relatively good
returns on campaign budget expenditures. Return on
investment in terms of life years saved will be presented.
Conclusion: This study highlights the large reach and
cost-effectiveness of mass media campaigns in LMICs.
Opportunities for future programme practice indicate
that the rapid uptake of mass media such as television
and social media, coupled with the more active viewing
of television content by new adopters, provides an
important and cost-effective channel of communication
for behaviour change in LMICs.
05. Financing TB services
Table Cost per person identified as having TB in Project Axshya

Sl no
OA-332-04 Value for money: it costs US$ 90 to

identify a TB patient in a community setting,
India
M Rout,1 P Banuru Muralidhara,1 T Palorkar,1 A Das,1
B Entoor Ramachandran,1 B Thapa1 1The International
Union Against Tuberculosis and Lung Disease, South-East
Asia Office, New Delhi, India. Fax: (91) 114 605 4430.
e-mail: bmprasad@theunion.org
Background: END-TB strategy is focused at community

taking onus of Tuberculosis, prevention and care.
Globally, many programmes are designed towards
sensitizing and engaging community members. The
Global Fund supported grant in Project Axshya in India
adopted a strategy to engage community members
through locally available Non-Government Organizations (NGOs) who are known in community for their
other social-development including health. The activities
of project were so designed that the organizations
utilized the available resources for a maximum output
in line with Global Fund Value for Money.
Intervention or response: NGOs (4 to 5) were identified
in each of 300 districts of Project Axshya through subrecipient partners. Members and volunteer of these
organizations were trained on Tuberculosis to implement
project activities. The activities/package included, (a)
community meetings (b) mid-media activities (c) active
case finding (d) tracking of referrals and sputum
collection and transportation (SCT) and (e) retrieval of
patients who lost to follow-up of the treatment. Rural
Health Care Providers (RHCP) were also trained and
engaged within community to identify presumptive TB
person and provide diagnosis and treatment through
DOTS.
Results and lessons learnt: Over 1200 NGOs, 10 000
volunteers and 19 000 RHCPs were identified and
trained through the project. With the involvement of
community, the project could offer package of TB
services to 23 million populations during April 2013 to
December 2014. Average cost for reaching one person is
only US$ 0.16. Through community intervention a total
of 496 083 presumptive TB persons were identified and
undergone diagnosis for TB and out of that 41 257
persons identified as having TB and .90% of them are
on DOTS. 76% of the presumptive TB persons were
provided sputum collection and transportation service.
Around 3.8 million USD was incurred as expenditure for
these activities, at an average cost US$ 7.62 per
presumptive TB patient identified and US$ 91.6 per
person identified as having TB.
Conclusions and key recommendations: This intervention has demonstrated that it costs 90 US$ per person
identified as having TB. It is worthwhile to interpret the
amount spent not only at identifying TB patients but also
efforts that were put forward towards generating
awareness about TB in community. This community
driven model could be expanded to reach the goal of
END-TB strategy.
S123
1
2
3
4
5
6
7
Indicator
Number of TB presumptive referred for
sputum examination in the reporting
quarter
Sputum collected and Transported of
presumptive TB person identified
Total presumptive TB person identified
Among presumptive TB person identified,
number person identified as having TB
Overall Expenditures
Cost per presumptive TB person identified
Cost per TB patient identified
Total
118,826
377,257
496,083
41,257
US$ 3,772,523
US$ 8
US$ 91
OA-333-04 Assessment of costs related to

community-based DOTS services: the FHI360/
TB CARE I experience in Mozambique
A Mataruse,1 P Posse,2 I Manhica,3
S Machevo Chilundo,1 Z Cuna,1 A Abdula,1 C Juliana,1
E Oliveras1 1FHI360 Mozambique, Maputo, 2Independent
Consultant, Maputo, 3Mozambique National TB Control
Program, Maputo, Mozambique. e-mail:
amataruse@fhi360.org
Background: Community Based Direct Observed Treatment (CB-DOTs) is effective at improving presumptive
case identification, referral to health services, diagnosis
and treatment support. In Mozambique, CB-DOTs has
been implemented since 2006, with more than 60 of 128
districts covered by 2012. A variety of CB-DOTS models
have been used by the local NGOs implementing the
strategy. Globally, evidence shows that CB-DOTs is more
cost effective than health facility-administered TB
control services with studies in Uganda, South Africa
and Brazil showing cost savings. Evidence on costeffectiveness in Mozambique is needed to ensure support
for CB-DOTs and information on the relative cost
effectiveness of different models is needed to determine
how best to implement CB-DOTs. A cost-effectiveness
analysis of NGOs implementing CB-DOTs was conducted in 2014
Design/Methods: Retrospective cost and results data
were collected from 9 CB-DOTs NGOs for the period
October 2011 to September 2013. An ingredients
approach was used. Costs of personnel, supplies and
equipment, training, and transport were captured.
FHI360 and health facility costs were excluded. Costs
were analyzed by component of care: identification,
referral and follow-up of treatment. Costs were adjusted
for inflation to 2013 and compared using the Incremental Cost Effectiveness Ratio (ICER). The ICER calculations were based on differences between costs per NGO
divided by results. ICER was then compared to a gross
domestic product. Cost-effectiveness was compared
using the treatment success rate which is the sum of the
completed treatment and cure rates
Results: The total average cost per patient to successfully
complete TB treatment was US$701, ranging from $204
to $1833 per partner. Both costs and efficacy varied
widely across the NGOs. The costs per component of
care were roughly equally distributed. The leading
S124
contributors to costs were personnel (37%) and equipment (26%), though these too varied widely by NGO.
With the exception of 2 partners who did not supply nonmonetary incentives to volunteers, the cost per volunteer
averaged $134
Conclusion: The average cost of $701 to successfully
treat a TB patient is consistent with findings from other
countries although it is twice the $350/case threshold set
by TB REACH. Given the range of cost-effectiveness,
further analysis is needed to fully understand the factors
that affect cost-effectiveness in the Mozambican context
in order to guide CB-DOTs implementation.
OA-334-04 MDR-TB cost minimization study in

Nigeria
B Musa,1 A Kuznik,2 A Habib1 1Bayero University Kano,
Kano, Nigeria; 2CELGENE, Warren, NJ, USA. e-mail:
babamaiyaki2000@yahoo.co.uk
Background: In Nigeria, patients with multi drug

resistant tuberculosis (MDR-TB) are treated in hospital
facilities. However, the World Health Organization
(WHO) also recommends home based treatment as a
viable alternative. We compared the cost of facility (F)
based MDR-TB care to home (H) based care from the
perspective of the Nigerian national health system.
Design/Methods: We assessed the expected costs of the
two MDR-TB treatment approaches using a decisionanalytic model with a follow-up of 2 years. MDR-TB
treatment outcomes were obtained from a systematic
review of randomized clinical trials. The outcomes of
interest included treatment success, treatment failure,
treatment default, and mortality, and did not vary
significantly between the two alternatives. Treatment
costs included the cost of: drug therapy (F, H), hospital
stay (F), nurse care (F, H), physician care (F), nursing
facility (F), and transport to the health care provider (H).
Finally, we estimate the potential cost savings associated
with home based treatment for all patients starting
MDR-TB treatment in Nigeria.
Results: The average expected total treatment cost for a
Nigerian patient treated for MDR TB was estimated at
US$ 2957 based on the facility based model and $1480
based on the home based care model, a potential saving
of 50%. One of the major drivers of this difference is the
significantly more intensive, and therefore more costly,
nurse care in the hospital facility. In the year 2013, a total
of 426 patients were initiated on facility based MDR TB
treatment in Nigeria, thus, the potential savings associated with implementation of home based care is
estimated at US$ 629 202 per year.
Conclusion: In Nigeria, treatment of MDR TB using
home based care is expected to result in similar patient
outcomes at markedly reduced public health costs
compared with facility based care.
OA-335-04 Post-diagnosis costs for patients

living with TB and HIV and their households in
South Africa
D Mudzengi,1 S Sweeney,2 T Kufa,1,3 P Hippner,1
S Charalambous,1,2,3 A Grant,2 G Churchyard,1,3
A Vassall2 1The Aurum Institute, Johannesburg, South Africa;
2
London School of Hygiene & Tropical Medicine, London, UK;
3
University of the Witwatersrand, Johannesburg, South
Africa. e-mail: dmudzengi@auruminstitute.org
Background: The World Health Organization (WHO)

has a new post-2015 strategy which aims to eliminate
catastrophic costs for TB patients. In South Africa, TB
affects the poorest; and many of those already suffering
the economic hardship due to HIV. Previous studies,
from other settings, suggest that the experience of the TB
patient immediately post-diagnosis can cause considerable economic distress, particularly for those who are
HIV positive. This study investigates the costs incurred
by people being treated for TB, HIV and both conditions
combined in South Africa.
Methods: This study focused on the post-diagnostic costs
incurred by TB and HIV patients attending 18 primary
health care clinics in the Ekurhuleni North sub-district,
Gauteng, South Africa. The clinics were participating in
a cluster randomised trial to evaluate a TB-HIV
integration intervention. Data on direct and indirect
patients post-diagnosis costs were collected using
structured questionnaires administered to trial participants within 6 months after HIV diagnosis or starting TB
treatment. Data were captured in an electronic database
and analysed in Microsoft Excel.
Results: The study enrolled 463 participants; 25% with
TB and HIV combined, 11% with only TB and 65% with
only HIV. Mean monthly post-diagnosis costs were
highest for TB-HIV patients ($118.23) when compared
with TB only ($83.58) and HIV only ($45.82). While
out-of-pocket expenses were high, income lost through
job changes or missed work was also substantial, as
patients felt unwell and/or sought care. Indirect costs
accounted for 41% of TB-HIV, 46% of TB and 34% of
HIV only; total costs.
Conclusion: Costs incurred by TB patients after diagnosis are often catastrophic, and are substantially higher for
those co-infected with HIV, given the complex postdiagnosis trajectory. Better integration of care and social
support to address both out-of pocket expenditure and
income loss to this, most vulnerable group, is urgently
required in South Africa; and elsewhere if global targets
are to be met.
OA-336-04 Economic cost of patient centred

tuberculosis treatment in Tanzania
A Mkopi,1 F Van Leth,2 N Range,3 F Lwilla,1
S M Egwaga,4 A Schulze,5 E. Geubbels1 1Ifakara Health
Institute, Dar Es Salaam, Tanzania; 2University of Amsterdam,
Amsterdam Institute for Global Health and Development,
Amsterdam, Netherlands; 3National Institute for Medical
Research, Dar Es Salaam, 4National Tuberculosis and Leprosy
Programme, Ministry of Health and Social Welfare, Dar Es
Salaam, Tanzania; 5Swiss Agency for Development and
Cooperation, Berne, Switzerland. Fax: (255) 23 262 5312.
e-mail: abdallah.mkopi@gmail.com
Background: In 2006, the National Tuberculosis and

Leprosy Programme (NTLP) of Tanzania introduced a
strategy known as Patient Centred Tuberculosis Treatment (PCT). We estimated the economic costs per HBDOT patient treated under the PCT strategy in Tanzania
from a societal perspective and compared this to a
hypothetical scenario of health facility-based DOT (HFDOT).
Design/Methods: This cross sectional study in one urban
and two rural districts measured TB service delivery costs
and both direct and indirect patient / treatment supporter
costs using WHO guidelines. We used structured
questionnaires to collect the requred information.
Intervention: Patient Centred Tuberculosis Treatment
(PCT); newly diagnosed patients are given the opportunity to choose either home-based DOT (HB-DOT)
supervised by a non-medical person of their preference,
or health facility-based DOT (HF-DOT) observed by a
health worker.
Results: Average direct costs amounted to US$4.8 and
US$5.2 per visit for a patient and a treatment supporter
respectively. Indirect costs related to treatment or
support were US$61.0 and US$57.5 per month for a
patient and a treatment supporter respectively. Total
economic costs for treating a new HB-DOT patient
under the PCT strategy was US$901.7, compared with
US$1211.1 for treating a new HF-DOT patient. Total
costs incurred by patients and treatment supporters were
higher than service delivery costs for the health system
and NTLP.
Conclusion: The present study shows that compared to
conventional DOT strategies, the PCT strategy is
relatively less costly in terms of both service delivery
costs and personal costs incurred by patients and
treatment supporters. However, even under the PCT
strategy the bulk of the economic burden is carried by TB
patients and their treatment supporters. Therefore,
intervention strategies to further reduce the costs to
patients and treatment supporters are highly needed
OA-337-04 Cost of TB diagnosis and treatment

packages in Viet Nam: estimates and
projections, 20152020
K Vu Duy,1 M Hoang Van,1 H K Pham2 1Hanoi Medical
University, Hanoi, 2World Health Organization in Viet Nam,
Hanoi, Viet Nam. e-mail: vuduykien@gmail.com
Background: Moving towards a sustainable financing

mechanism for TB is reflected in the National TB
S125
Strategy 2014-2020, vision 2030 approved by the

Government in March 2014. In order to inform the
policy to support the TB program, we conducted this
costing study for TB diagnosis and treatment packages in
Viet Nam.
Design/Methods: The costs of providing tuberculosis
diagnosis and treatment packages in this study were
estimated from the provider perspective. We considered
different factors that impacted on the costs such as age
group, types of TB, type of TB services and health facility
levels. 1) To estimate the costs of providing TB diagnosis
and treatment packages in Viet Nam, and 2) To project a
needed annual budget for the TB programme in Vietnam
during 2015-2020.
Results: At national level, the unit costs of diagnosis and
treatment for a pulmonary TB AFB() ranged 97.2-171.1
US$, a pulmonary AFB(-) ranged 158.7-272.7 US$ and a
extra-pulmonary TB case ranged 107.3-148.3 US$. The
unit costs for diagnosis and treatment of a MDR-TB case
were very high for both children (ranged 1211.2-1580.2
US$) and adults (ranged 1474.1-1849.6 US$) at national
level. At provincial level, the costs of diagnosis and
treatment for a pulmonary TB AFB(-) ranged 92.8-157.9
US$, for a pulmonary AFB() ranged 152.9-237.2 US$,
for a extra-pulmonary TB case ranged 98.7-130.7 US$,
and for a MDR TB case ranged 1202.8-1819.6 US$. At
district level, the costs of diagnosis and treatment for a
pulmonary TB AFB(-) ranged 122.1-162.3 US$, for a
pulmonary AFB() ranged 51.2-101.7 US$, for a extrapulmonary TB case ranged 53.1-118.2 US$, and for a
MDR TB case ranged 1568.2-2391.2 US$. The annual
budget needed for TB program was about 15-19.8
million US$ in 2015, then increased about 21-29.8
million US$ in 2020. The annual revenue of the Health
Insurance (HI) Fund in Vietnam is about 1.85 billion
US$. This budget is equal to about 1 % of the annual
revenue of the HI fund.
Conclusion: This study provides the cost estimation for
TB diagnosis and treatment in Viet Nam 2015-2020. As
the donor exit policies in the coming time, TB diagnosis
and treatment should be covered under Health Insurance
programme in order to maintain the crucial TB services
for Vietnamese.
OA-338-04 TB-related dissaving was common

and correlated with incurring catastrophic
costs in TB-affected households in Peruvian
shantytowns
T Wingfield,1,2,3 M Tovar,1,4 D Huff,1,5 J Lewis,1,6
Benefica
R Montoya,1 D Boccia,1,6 C Evans1,2,4 1Asociacion

PRISMA, Lima, Peru; 2Imperial College London, London,
3
North Manchester General Hospital, Manchester, UK;
4
Universidad Peruana Cayetano Heredia, Lima, Peru; 5Tulane
University, New Orleans, LA, USA; 6London School of
Hygiene & Tropical Medicine, London, UK. e-mail:
tom.wingfield@ifhad.org
Background: The World Health Organizations (WHO)

post-2015 global TB control policy states that no TBaffected family should incur catastrophic costs. However,
measuring catastrophic costs may be operationally
S126
difficult for National TB Programs. Dissaving is a proxy

measure of financial shock but its correlation with
catastrophic costs is not yet known and is a priority
research question for the WHO Global TB Programme
Task Force. We aimed to evaluate the association of TBrelated dissaving with catastrophic costs to inform post2015 TB control policy.
Methods: Consecutive TB patients diagnosed by the
Peruvian National TB Program in 32 contiguous
shantytown communities of north Lima, Peru, were
invited to participate in the study. Patient households
total TB-related costs (lost income plus direct out-ofpocket expenditure) and dissaving variables (Figure)
were recorded throughout treatment. A composite
dissaving index in arbitrary units was derived by
principal component analysis (PCA) from all dissaving
variables. Catastrophic costs were defined as total costs
of 720% of the annual income of the same household,
as previously defined in the study setting.
Results: 312 TB patients were invited to participate, of
whom 282 were recruited and had costs and dissaving
data available for analysis. The specific dissaving
variables that best explained the PCA-derived dissaving
score in this setting were: missing scheduled payments,
starting a new job, selling or pawning household items,
undertaking small scale fundraising activities, and being
asked to eat elsewhere. Patients with greater than average
dissaving were significantly more likely to: incur
catastrophic costs, be poorer, have a lower education
level, to have greater food insecurity, and have longer
symptom duration prior to diagnosis (Figure). There was
a significant positive correlation between increasing total
costs as a proportion of annual income and dissaving
score (Spearmans correlation coefficient 0.26, P ,
0.0001).
Conclusions: In this setting, dissaving was associated

with incurring greater and catastrophic costs, and
belonging to more vulnerable patient households. This
evidence supports the use of dissaving variables and/or a

dissaving score as a proxy for catastrophic costs that may
be adapted to local settings.
OA-339-04 Are TB patients in Republic of

Macedonia at risk for catastrophic health
expenditures? Results of a cross sectionalstudy
D Gudeva Nikovska,1 F Tozija1,2 1Faculty of Medicine,
University SS Cyril and Methodius, Skopje, 2National Institute
of Public Health, Skopje, Macedonia, Yugoslav Republic. Fax:
(389) 2317 7983. e-mail: dgnikovs.ka@gmail.com
Background: Republic of Macedonia (RM) has developed a comprehensive TB control program over the past
decade, funded by both Governmental funds and more
comprehensively with the generous funding from Global
Fund to Fight HIV/AIDS, Tuberculosis and Malaria
(GFATM) since 2006. In turn, this is reflected in the
incidence numbers, showing that TB is on constant
decrease and reaching the MDG#6 (with incidence of
25.9/100 000 in 2006 to 15.2/100 000 in 2014, that
classifies RM as low incidence country). TB diagnostics,
treatment and follow-up is free of charge for all
population, regardless of the insurance status. However,
not all the population benefitted from implementation of
activities, the highest rates still being notified in the
north-west part of the country and in certain ethnic
groups, mostly affecting the age group 25-54.
Design/Methods: Nested case-control study was conducted on a sample of 605 households (HH); face-to-face
interviews were conducted, using selected modules from
World Health Survey questionnaire. Cases are TB
patients registered Jul, 2012 Jun, 2013 (n 315) and
controls HH in their neighborhood (n 290). Data were
analyzed with SPSS 19.0, utilizing logistic regression to
measure predictive value of most important SDH.
Results: Analyzed by groups of respondents, the mean
value of costs for treatment were higher in controls, with
exception of costs for transport that are significantly
higher in TB cases (t 6.548, df 4 83, P , 0.01), which
is most likely associated with the regional distribution of
TB dispensaries around the country where TB patients
seek care. Statistically significant differences are noted by
place of residence, costs being higher among TB patients
living in rural areas (t-test 2.145, df 125, P 0.034)
and in the North-West region (t-test 1.212, df 7, P ,
0.001). Of particular importance is the share of costs for
health care in TB patients in the total HH expenditures,
amounting to 55-70%, a percentage that certainly
implies catastrophic health expenditures for these HH
that were already identified as of lower socio-economic
status.
Conclusions: High share of expenditure on healthcare in
the total HH expenditures, particularly among TB
patients residing in rural areas implies not only catastrophic health expenditures, but also existence of certain
degree of inequity in access for this socially disadvantaged population stratum. The findings needs further
exploration of the underlying causes, particularly due to
the fact that treatment of TB is free of charge under

provisions of the National TB control program, especially notified regional differences in expenditures that
imply inequities in access to health care of certain
vulnerable groups.
06. MDR-TB: trials, cohorts and predictors

OA-340-04 Bedaquiline prevents acquired
resistance to second-line drugs
1
S127
mutation in Rv0678 gene occurred in 1 (failing patient)

out of 10 with post-baseline isolates. None of the
patients in the BDQ group progressed from MDR to
pre-XDR-TB or from pre-XDR-TB to XDR-TB.
Conclusion: Despite the limited number of paired
samples, addition of BDQ to the preferred MDR-TB
background regimen decreased development of resistance and halted the progression from MDR-TB to preXDR-TB and from pre-XDR-TB to XDR-TB.
OA-341-04 Efflux-pump-mediated acquired

resistance to bedaquiline in a clinical trial
N Lounis, A Diacon, A Pym, B Van Baelen,

R Van Heeswijk,1 M Haxaire-Theeuwes,1 B Dannemann,4
K Andries1 1Janssen Research and Development, Beerse,
Belgium; 2Faculty of Medicine and Health Sciences,
Stellenbosch University, Cape Town, 3Medical Research
Council and Kwazulu Research Institute for Tuberculosis and
HIV, Durban, South Africa; 4Janssen-R&D, LLC, Titusville, NJ,
USA. e-mail: nlounis@its.jnj.com
K Andries,1 N Coeck,2 C Villellas,1 B Van Baelen,1

N Lounis,1 L Rigouts,2 B Dannemann,3 P Alexander4
1
Janssen Research and Development, Beerse, 2Institute of
Tropical Medicine, Antwerp, Belgium; 3Janssen-R&D, LLC,
Titusville, NJ, USA; 4Medical Research Council and
KwaZulu-Natal Research Institute for Tuberculosis and HIV,
Durban, South Africa. e-mail: kandries@its.jnj.com
Background: Study C208/2 (NCT00449644) was a

randomized, double-blind, placebo-controlled Phase II
study where patients diagnosed with MDR-TB or preXDR-TB were treated for 24 weeks with bedaquiline
(BDQ) or placebo, combined with a preferred MDR-TB
background regimen (BR) consisting of ofloxacin (OFX),
kanamycin (KM), pyrazinamide (PZA), ethionamide
(ETO) and cycloserine followed by BR only for up to
18 more months. Treatment response defined as sustained culture conversion was assessed over 120 weeks.
The role of BDQ in preventing acquired resistance to
second-line drugs was assessed.
Methods: Drug susceptibility tests (DST) for OFX, KM,
ETO, EMB, streptomycin (SM), capreomycin (CM), and
para-aminosalicylic acid (PAS) were performed on 7H11
agar (proportion method) and for PZA in the MGIT960
system. The minimal inhibitory concentration (MIC) for
BDQ was determined on 7H11 agar. Acquired resistance
to a specific drug was defined as a change from
susceptible at baseline to resistance at any post-baseline
visit.
Results: In the final analysis, baseline and post-baseline
DST results were available for 11 of the 66 mITT
patients in the BDQ group and 30 of the 66 mITT
patients in the placebo group. In the placebo group, 16/
30 (53%) acquired resistance to at least 1 additional
TB drug. Acquired resistance developed to OFX (7 out
of 26 patients with a susceptible isolate at baseline),
EMB (6/15), SM (1/5), PZA (2/11), ETO (2/27), KM
(1/24), and CM (1/24). The response rate was 56% (7/
16) in patients who acquired resistance and 79 % (11/
14) in patients who did not. Evolution towards preXDR or XDR-TB was observed in 8 of 66 (12%)
patients: 6 progressed from MDR-TB to pre-XDR-TB
and 2 from pre-XDR-TB to XDR-TB. In the BDQ
group, 2/11 patients (18%, 1 responder, 1 nonresponder) acquired resistance to at least 1 additional
TB drug. Acquired resistance to SM was found in 1
patient out of 2. Acquired resistance to EMB and CM
was identified in one additional patient (out of 4 and 8,
respectively). A 4-fold increase in BDQ MIC due to a
Background: Study C209 was an open-label Phase II

study where patients diagnosed with MDRH&R-TB,
pre-XDR-TB or XDR-TB were treated with bedaquiline
(BDQ) for 24 weeks combined with an individualised
MDR-TB background regimen for up to 120 weeks. The
majority had confirmed infection with an MDRH&RTB strain (93/174 or 53.4%), with 44 (25.3%) and 37
(21.3%) infected with a pre-XDR-TB strain or XDR-TB
strain, respectively.
Methods: Acquired resistance was defined as an at least
4-fold increase in BDQ MIC on 7H11 agar in any postbaseline isolate (week 24 or later), compared to the MIC
of the corresponding baseline isolate.
Results: Twenty-four patients included in the efficacy
analysis who received BDQ had paired isolates from
baseline and from week 24 or later. Nineteen paired
isolates were genotypically identical (by MIRU-VNTR)
and 12 of these had 74-fold increases in BDQ MIC. In
all these 12 post-baseline isolates, mutations were
identified in Rv0678. Mutations in this regulatory gene
lead to overexpression of the MmpS5-MmpL5 efflux
pumP and 2- to 8- fold increases in BDQ MIC. When
analysing the effect of MIC increases on treatment
response, 5 of the 12 patients (41.7%) with 74-fold
increases in BDQ MIC were responders at endpoint,
while 3 of the 7 patients (42.9%) with ,4-fold increases
in BDQ MIC were responders at endpoint (final
analysis). When considering absolute values, 9 of the
12 patients with increased MICs had a post-baseline
BDQ MIC of 70.25 lg/ml, and 3 had an MIC ,0.25 lg/
ml. Nine of 13 patients (69.2%) with post-baseline BDQ
MICs 70.25 lg/ml were responders at endpoint, while 4
of 11 patients (36.4%) with post-baseline BDQ MICs
,0.25 lg/mlL were responders at endpoint. BDQ MIC
increases occurred more frequently in XDR-TB patients
(7/37, 19%) and pre-XDR-TB patients (4/44, 4%) than
in MDRH&R-TB patients (1/93, 1%), and more
frequently in patients with a background regimen
interruption of at least 2 weeks (4/34, 12%) than in
those without such interruption (8/140, 6%).
S128
Conclusions: Having pre-XDR-TB or XDR-TB and

treatment non-adherence increased the risk of acquired
resistance to BDQ. Treatment of MDRH&R-TB patients
involved less risk. However, no clear relationship was
observed between this low-level, efflux-based acquired
resistance to BDQ and treatment response (sputum
culture conversion) at endpoint. Based on these data,
increases in BDQ MICs due to mutations in Rv0678 may
not necessarily translate into clinical resistance.
OA-342-04 Comparison of 2- and 6-month

culture conversion as early predictors of
multidrug-resistant TB treatment success:
multicentre study of MSF programmes
M Bastard,1 E Sanchez,1 H Khamraev Atadjan,2 Z Tigay,3
A Hayrapetyan,4 M Kamene Kimenya Mariita,5
S Khurkhumal,6,8 T Dlamini,7 F Varaine,9 M Bonnet1,10
1
Epicentre, Paris, France; 2Tashkent Pediatric Medical
Institute, Nukus, 3Chief Doctor of Republican TB Hospital,
Nukus, Uzbekistan; 4National Tuberculosis Control Office,
Yerevan, Armenia; 5Programmatic Management of Drug
Resistant Tuberculosis, Nairobi, Kenya; 6National Tuberculosis
Program, Soukumi, Georgia; 7Ministry of Health TB National
Control Program National Manager, Mbanane, Swaziland;
8
National Tuberculosis Program, Soukumi, Abkhazia,
`
Sans Frontieres,
Paris, 10Unite Mixte
Georgia; 9Medecins
Internationale UMI233-U1175, Institute of Research for
Development, Montpellier, France. e-mail:
mathieu.bastard@geneva.msf.org
Background: An early treatment efficacy indicator for

multi-drug resistant TB (MDR-TB) could lead to better
treatment success through earlier treatment adaptation
and potentially be in the evaluation of treatment
efficacy. We evaluated the values of 2 and 6-month
culture conversion to predict final treatment outcomes
using data from seven MSF drug-resistant TB programs.
Methods: Confirmed MDR-TB patients starting treatment from 2001 to 2011 with available results on followup culture and a treatment outcome assigned were
included in the study. Proportion of patients converted
at month 2 and 6 of treatment was assessed and
performances of these indicators to predict treatment
success were assessed. We also fitted a random-intercept
logistic model to adjust the estimates for potential
confounders.
Results: A total of 1445 MDR-TB patients were
included in this study: 57.1% male, median age of 32
years [IQR 24-42], 82.2% previously treated cases.
Profiles of baseline resistance were as follow: 66.6%
were resistant to H and R only, 31% had additional
resistance to injectables or fluoroquinolones and 2.4%
had both resistance to injectables and fluoroquinolones.
A total of 319 (22.1%) and 709 (49.1%) patients
converted at 2 and 6 months, respectively. Two-month
culture conversion had a positive predicted value (PPV)
of 75.5% (95%CI 70.5-80.2), sensitivity of 27.7%
(95%CI 24.8-30.8)specificity of 86.5% (95%CI 83.489.1) and negative predicted value (NPV) of 44.2%
(95%CI 41.3-47.2) against final treatment outcome. At
6 months, culture conversion had a PPV of 74.9%

(95%CI 72.0-77.6), 81.6% (95%CI 78.8-84.1) sensitivity, 58.7% (95%CI 54.5-62.7) specificity and 67.9%
(95%CI 63.6 72.0) NPV. These performances did not
differ strongly by baseline resistance profiles. After
adjustment on gender, age, body mass index, baseline
resistance profile and being previously treated, culture
conversion at month 2 and 6 remained a strong
predictor of treatment success (aOR2.22, 95%CI
1.65-2.98 and aOR6.18, 95%CI 4.78-7.98, respectively). The association was even stronger when excluding
defaulters.
Conclusions: Although 2 and 6 months culture conversion had similar PPV for treatment success, only 28% of
patients who succeed treatment converted by 2 months
compared to 82% at 6 months. The 6-month culture
conversion is probably a better predictor of the final
treatment outcome for MDR-TB. However, performances found in this study are substantially lower than those
reported in literature.
OA-344-04 Subgroup analysis of patients

receiving bedaquiline as part of a multidrugresistant tuberculosis treatment regimen
A Pym,1 A Diacon,2 F Conradie,3 S Tang,4
B Van Baelen,5 N Lounis,5 M Haxaire-Theeuwes,5
B Dannemann6 1Medical Research Council and Kwazulu
Research Institute for TB and HIV (K-RITH), Durban,
2
Division of Medical Physiology and the Department of
Science and Technology/National Research Foundation
Centre of Excellence for Biomedical Tuberculosis Research
and Medical Research Council Centre for Tuberculosis
Research, Cape Town, 3Clinical HIV Research Unit,
Department of Medicine, Faculty of Health Sciences,
Africa; 4Tuberculosis Department, Beijing Chest Hospital,
Capital Medical University, Beijing Tuberculosis and Thoracic
Tumor Research Institute, Beijing, China; 5Janssen Infectious
Diseases BVBA, Beerse, Belgium; 6Janssen Research &
Development LLC, Titusville, NJ, USA. Fax: (27) 31 203
4702. e-mail: alexanderpym@hotmail.com
Background: C209 was a Phase 2, open-label, single-arm

trial (NCT00910871) of bedaquiline (BDQ) as part of an
individualized background regimen (BR). The 120-wk
sputum culture conversion rate (missingfailure) was
72% in 205 newly diagnosed or pre-treated adults with
confirmed sputum-smear positive pulmonary MDR-TB,
pre-extensively or extensively drug resistant (pre-XDR or
XDR) TB. All patients (pts) received BDQ for 24 wks
(400mg qd for 2 wks, 200mg tiw for 22 wks) with a BR
according to WHO guidelines continued for up to 120
wks (final assessment).
Methods: Treatment response (sustained culture conversion at 120 wks) was investigated for subgroups
according to baseline characteristics of individual patients (gender, age, race), disease severity (radiology,
body mass index, serum albumin, time to culture
positivity, smear grade), drug resistance (previous drug
use, extent of resistance) and the BR received.
Table Sputum culture conversion rates at 120 wks

Subgroup
Responders*
n/N (%)
Sex
Male
Female
94/132 (71)
54/73 (74)
Age (years)
1845
.4565
.65
Race
American Indian
Asian
Black/African American
White
Lung cavitations
None or ,2cm
72cm in one lung only
72cm in both lungs
Extent of resistance to M. tuberculosis strain
MDR-TB
Pre-XDR-TB
Fluoroquinolone resistant
Second-line injectable resistant
XDR-TB
115/160 (72)
31/43 (72)
2/2 (100)
5/6
70/84
39/67
34/48
(83)
(83)
(58)
(71)
57/70 (81)
73/108 (68)
18/27 (67)
68/93 (73)
31/44 (71)
23/31 (74)
8/13 (62)
23/37 (62)
Previous use of second-line drugs

Newly diagnosed
Non-newly diagnosed
23/28 (82)
125/177 (71)
Background regimen at baseline

Fluoroquinolones
Pyrazinamide
Second-line injectable
Linezolid
127/180(71)
108/152 (71)
137/191 (72)
9/12 (75)
Baseline pyrazinamide susceptibility

Susceptible
Resistant
Baseline albumin grade
Grade 0
Grade 1
Grade 2
Grade 3
Baseline body mass index (kg/m2)
,18
718,20
720,25
725
Median time to positive signal in MGIT
at baseline (days)
610
.10
Baseline AFB score
0
Scanty
1
2
3
25/38 (66)
96/135 (71)
126/167 (75)
10/15 (67)
12/22 (55)
0/1 (0)
35/49
40/59
57/77
16/20
(71)
(68)
(74)
(80)
49/72 (68)
99/133 (74)
34/38
44/60
32/42
22/34
16/31
(90)
(73)
(76)
(65)
(52)
*Responder culture conversion.
Race based on FDA template.
Excluding pre-XDR-TB and XDR-TB;
Amikacin, kanamycin and capreomycin;
MDR-TB multidrug-resistant TB; XDR-T extensively drug-resistant TB;

AFB acid-fast bacilli.
Results: Overall, 45% of pts had MDR-TB, 22% preXDR-TB, 18% XDR-TB and in 15% with at least MDRTB the extent of further resistance could not be
determined. At baseline; 88% used fluoroquinolones
S129
(49% ofloxacin; 32% levofloxacin), 78% ethionamide

or prothionamide, 74% pyrazinamide, 74% aminoglycosides and 12 pts used linezolid (5 were pre-XDR and 4
XDR). Conversion rates were 73%, 71% and 62% in pts
with MDR-TB, pre-XDR-TB and XDR-TB, respectively
(Table). All pts with XDR-TB at baseline who sputum
culture converted at some point and who completed the
trial (n 23) remained culture negative at treatment end
or after treatment-free follow-up (16/23, median duration 5.4 months). Of the linezolid users (n 12), 9 were
responders (75%), 2 were non responders (17%) and 1
pt withdrew consent but had culture converted at
discontinuation. This conversion rate was comparable
to that of the overall study. In a multivariate analysis of
the subgroup data, the conversion rate was significantly
lower in pts who previously used second-line drugs and
in pts with higher baseline AFB score.
Conclusion: Addition of BDQ to an MDR-TB regimen in
heavily pre-treated pts resulted in durable conversion
rates at 120 wks, irrespective of infection sub-type. In the
majority of pts with XDR-TB, 6 months of BDQ
administration followed by continuation of an individualized BR was sufficient both to convert the sputum
culture and prevent short-term relapse.
OA-345-04 Twelve-month outcomes of patients

initiating drug-resistant tuberculosis treatment
at a decentralized outpatient clinic in
Johannesburg, South Africa
K Hirasen,1 R Berhanu,2 D Evans,1 K Schnippel,2 L Long,1
S Rosen,1,3 I Sanne1,2,3,4 1Health Economics and
Epidemiology Research Office, Johannesburg, 2Right to Care,
Johannesburg, South Africa; 3Center for Global Health &
Development, Boston, MA, USA; 4Clinical HIV Research Unit,
Johannesburg, South Africa. e-mail: khirasen@heroza.org
Background: Treatment success rates for drug-resistant

(DR) TB in South Africa are poor, due in part to high
rates of loss to follow-up (LTFU). In 2011, South African
guidelines expanded to include decentralized, outpatient
DR TB care to improve treatment access; however, the
concern was that outpatient treatment would lead to
increased LTFU. Sufficient data have now accrued to
allow analysis of rates of LTFU and death and to identify
factors associated with LFTU during the first 12 months
after treatment initiation at a decentralized, public sector
DR-TB treatment site in Johannesburg.
Methods: Prospective observational cohort analysis of
adults (718) who initiated treatment for rifampicin
resistance from March 2013 to March 2014. LTFU was
defined as a treatment interruption 72 months. Factors
associated with LTFU were identified using Cox hazards
models to produce hazard ratios (HR). Person time was
calculated from DR-TB initiation until either a final
outcome reported or 12 months of follow-up occurred.
Results: 134 patients enrolled in the study; 32 were
excluded from analysis (28 transferred; 4 drug sensitive).
Of the 102 remaining, 49 (48%) were male, median age
37 (IQR 2944). 84 (82%) were HIV-infected, with a
median CD4 count of 113 cells/mm3 (IQR 35-262), and
40/84 (48%) were on antiretroviral therapy (ART) at
S130
baseline. Pulmonary and extrapulmonary TB occurred in

90 (88%) and 12 (12%), respectively; 33/69 (48%) were
smear positive at baseline. At 12 months, 10 patients
(11%) had died, 25 (28%) were LTFU, and 55 (61%)
remained on treatment. Females had a significantly
higher risk of LTFU than males (aHR 3.82; 95%CI
1.06-13.83) but lower risk of mortality (HR 0.78; 95%CI
0.22-2.78). There was no association between LTFU and
drug resistance pattern, employment status, or HIV
status. All deaths occurred before 6 months of treatment
whereas LFTU persisted throughout the 12 months.
Conclusion: At this decentralized DR-TB treatment site,
12-month LTFU and mortality were high. Mortality
occurred early in treatment which is consistent with
previous findings. Being female was associated with a
higher risk of LTFU but a lower risk of mortality. The
high risk of LTFU among females is a finding that differs
from results of other DR-TB and HIV treatment
programs and warrants further study.
male. Two (of 9) provinces, KwaZulu Natal and Eastern

Cape, accounted for 64% of cases. HIV status was
reported for 79.6%; of patients with known status,
61.6% (n 266) were HIV-positive and patient
outcomes had not been evaluated or reported for 128
(23.6%); these patients were excluded from the outcome
analysis. 241 of the 415 with outcomes died during
treatment (58.1%), 10.4% were lost to follow-up,
14.7% on treatment or failed treatment, and 11.3 %
were successfully treated. Positive baseline smear microscopy was a predictor of mortality relative risk (RR)
1.58 (95%CI: 1.25 to 2.00); having negative sputum
culture at 6 months was a negative predictor of mortality
(RR: 0.37, 95%CI: 0.25 to 0.53). HIV-positive patients
were also at higher RR for mortality (1.24, 95%CI: 1.02
to 1.52), even when controlling for smear positive status
(aRR:1.29, 95%CI: 1.02 to 1.62). Antiretroviral therapy
for HIV coverage was high (91.6%) and therefore was
not statistically significant (results not shown). Treatment with capreomycin, PAS, or moxifloxacin did not
significantly affect RR of mortality.
Table Relative risk*^ of mortality during XDR-TB treatment by
patient and treatment characteristics, XDR-TB patients with
treatment outcomes registered in South Africa from 2007 to
2009
Adjusted
Count / sample Relative risk relative risk
(95%CI)
(%)
(95%CI)
All patients
Male
HIV-positive
OA-346-04 Mortality during XDR-TB treatment

in South Africa, 20072009: a retrospective
cohort analysis
N Ndjeka,1 S S Dlamini,1 N Bantubani,2 K Schnippel3 1TB
Control and Management Cluster, Pretoria, 2TB HIV
Investigative Network of KwaZulu Natal (THINK), Durban,
3
Witswatersrand University, Johannesburg, South Africa. Fax:
(27) 866 046 792. e-mail: ndjeka@esnet.co.za
Background: Between 2007 and 2014 approximately

7000 extensively drug resistant (XDR-) TB cases were
diagnosed in South Africa of which 4567 received
treatment. Treatment success is low (15-20%) and
mortality high (50-60%) among XDR-TB patients.
Design/Methods: In 2011-2012, a retrospective review
of patient files, clinic and electronic DR TB registers for
XDR-TB patients registered for treatment in South
Africa between 2007 and 2009 was conducted.During
the study period, XDR-TB treatment was provided on an
inpatient basis until sputum culture conversion in
specialized, centralized hospitals. XDR-TB treatment
regimens were individualized based upon resistance test
results but often included capreomycin and para-aminosalicylic acid (PAS); moxifloxacin was not commonly
used until late in the period.
Results: 543 XDR-TB patients were registered during the
period. Median age was 35 (IQR: 27-43), 49.4% were
Age category
024 years
2539 years
4054 years
7 55 years
TB disease
Previous TB
reported
Smear positive at
initiation
Culture negative
at 6 months
Treatment regimen
Moxifloxacin
Capreomycin
PAS
241/415
(58.1%)
206/413
(49.9%)
206/346
(59.4%)
0.88
(0.751.04)
1.24
(1.021.52)
0.90
(0.721.12)
1.29
(1.021.62)
68/408
(16.7%)
187/408
(45.8%)
126/408
(30.9%)
27/408
(6.6%)
0.84
(0.651.10)
0.91
(0.761.10)
REF
1.08
(0.791.46)
0.86
(0.681.10)
REF
0.89
(0.621.27)
0.96
(0.561.65)
287/415
(69.2%)
209/344
(60.8%)
133/239
(55.7%)
0.90
(0.761.06)
1.58
(1.252.00)
0.37
(0.250.53)
1.01
(0.791.29)
1.61
(1.252.07)
0.36
(0.240.55)
125/403
(31.0%)
341/403
(84.6%)
342/403
(84.6%)
0.86
(0.711.05)
0.97
(0.771.23)
0.87
(0.701.07)
1.06
(0.851.33)
0.94
(0.711.24)
0.87
(0.671.13)
* Relative risk calculated using Poisson regression with robust standard

errors.
Adjusted for baseline smear status (positive or negative) and HIV status
(positive or negative).
XDR-TB extensively drug-resistant tuberculosis; CI - confidence interval;

HIV human immunodeficiency virus; PAS para-aminosalicylic acid.
Conclusion: XDR-TB continues to claim a lot of lives.

Low treatment successful completion rate and high
mortality are still key features of this condition.

Introduction of new medicines which have high rates of
early smear and culture conversion is imperative to
reduce mortality from XDR TB.
OA-347-04 Mortality among patients treated

for multidrug-resistant tuberculosis in nine
countries, 20052010
H Kirking,1 M Yagui Moscoso,2 M Van Der Walt,3
T Tupasi,4 O Demikhova,5 T Dalton,1 E Kurbatova,1
The Global Petts Investigators1 1U.S. Centers for Disease
Control and Prevention, Atlanta, GA, USA; 2Peru Ministry of
Health, Lima, Peru; 3South Africa Medical Research Council,
Pretoria, South Africa; 4Tropical Disease Foundation, Manila,
Philippines; 5Central Tuberculosis Research Institute of the
Russian Academy of Medical Sciences, Moscow, Russian
Federation. e-mail: hrj7@cdc.gov
Background: Worldwide, an estimated 5% of tuberculosis cases are multidrug-resistant tuberculosis (MDRTB), which is associated with increased mortality and can
be treated only with prolonged, poorly tolerated
regimens. To identify opportunities to reduce MDR TBrelated mortality, we investigated predictors of mortality
in a multinational prospective observational cohort study
of patients undergoing treatment.
Design/Methods: During 20052008, adults (718
years) were enrolled at the start of treatment for MDRTB in Estonia, Latvia, Peru, Philippines, Russia, South
Africa, South Korea, Thailand, and Taiwan. Patients
were followed up until end of treatment or end of
observation on June 30, 2010. Patients who did not
follow study protocol and those without data on
outcome or time until outcome were excluded. The
survival distribution was estimated using the Kaplan
Meier method. Predictors of death were ascertained by
using multivariate Cox proportional hazards regression
modeling.
Results: In total, 1550/1761 (88%) patients were
included. An estimated 84.7% of patients survived past
two years (95%CI 82.6%86.9%). Independent predictors of mortality included HIV infection (adjusted hazard
ratio [aHR]: 1.8; 95% confidence interval [CI]: 1.12.8),
body mass index of ,18.5 kg/m2 (aHR: 2.7; 95%CI:
2.03.8), and baseline resistance to 71 second-line
injectable drugs (aHR: 2.2; 95%CI 1.53.2) or fluoroquinolones (aHR 2.2, 95%CI 1.43.5).
Conclusion: Among patients with MDR-TB, 2 treatable
conditionsHIV and low BMIwere associated with
mortality. Data from future investigations of interventions for these conditions may help to increase survival
rates.
S131
07. Many paths to the same END TB

strategy
OA-348-04 How did we perform? An
assessment of global investment effectiveness
in sub-Saharan Africa against key tuberculosis
control performance indicators
K G Koura,1 F Boillot,1 V Schwoebel,1 A Trebucq1
1
Paris, France. e-mail: kgkoura@theunion.org
Background: While the DOTS strategy provides the

technical framework for tuberculosis control since TB
was declared a public-health emergency in 1990, Global
Fund (GF) grants eventually provided low-income
countries in sub-saharan Africa (SSA) with substantial
additional resources to accelerate progress. Additional
components like Performance Based Funding (PBF) and
partnership were expected to provide powerful additional incentives to accelerate progress. As the horizon for the
Millennium Development Goals comes close, we examined the trends in TB notifications and treatment success
in SSA, looking whether the implementation of GF grants
accelerated increases in these indicators.
Design/Methods: We consolidated country data for the
WHO-Afro region, obtained from the Global Health
Observatory from 1995, when countries regularly
reported to WHO. We compared data from the 19952003 periodbefore countries accessed to GF grantsto
those from the 2004-2012 period when GF support
eventually matured. We compared time series for notified
new cases (total and smear-positive) and treatment
success rates, with the null hypothesis that GF grants
did not accelerate trends. We tested statistical significance of change by segmented regression analysis after
adjusting for first order auto-correlation using STATA.
Results: We observed no progress acceleration for the 3
selected indicators after GF grants were implemented (P
. 0.1). Surprisingly, case notification of new smearpositive cases slowed down after 2003. Progress in
notification of all new cases and treatment success
slowed down from 2008.
Conclusion: As prevalence surveys increasingly suggest
that the TB burden in SSA is underestimated, the failure
of the GF to boost progress in TB control is alarming.
Documentation to assess the magnitude of GF disbursements is not easily accessible but the hypothesis they
substantially increased funding is reasonable. In countries with limited health care access, the StopTB strategy
focused on care supply may have reached its limits. A
scope for the End TB strategy broadened to the demand
side, with emphasis on research to find how to transform
resources into results is a logical development. Substantial resources that only the GF has proved able to
mobilize and sustain will remain necessary. To face the
challenge, the GF may need to adjust its approaches
beyond its current funding model and to clarify the
impact on results of its partnership model and the
incentives it delivers.
S132
OA-349-04 Reaching missed tuberculosis cases:

lessons learnt using screening tools used in
Ghanas 2013 National Tuberculosis Prevalence
Survey
1
F Bonsu, I Law, R Gockah, Z Wagaw,

N N Hanson-Nortey,1 M Tadolini,2 E Owusu-Dabo,3
I Onozaki2 1National TB Control Programme, Ghana Health
Service, Accra, Ghana; 2World Health Organization, Geneva,
Switzerland; 3RESEARCH, Kumasi Centre for Collaborative
Reserach (KCCR), Accra, Ghana. e-mail:
zalebachew@gmail.com
Background: Low Tuberculosis case notification rate is a

challenge for Ghana Tuberculosis control. It is believed
existing screening tools may partly be an important
underlying reason. We examined the various screening
tools used during a prevalence survey and determine
proportion of survey tuberculosis cases that would have
been detected using routine symptom screening (coughing two weeks or more) and Acid-Fast Bacilli smear
microscopy as first diagnosis.The aim of this study is to
determine proportion of cases that are missed under
routine programmatic conditions and evaluate the
sensitivity of screening tools used during the survey.
Design/Methods: A nationwide, clustered-sampled,
stratified (urban/rural), cross-sectional survey, was conducted in 2013. All consenting participants were
screened for tuberculosis using: chest X-ray examination
and a symptom based screening of cough of two-week
duration. Individuals identified with an abnormal chest
X-ray or cough symptoms of more than two weeks
duration were eligible for sputum examination. All spot
and morning sputum specimens collected from suspects
were examined by Acid-Fast Bacilli smear microscopy
and liquid culture (MIGIT).
Results: Among 67 757 eligible individuals, 61 726 697
(91.1%) participated in the survey. Of which 8298
(13.4%) participants were eligible for sputum examination. A total of 202 tuberculosis cases: 64 smear positive
and 138 smear negative culture positive were identified.
Smear microscopy detected 31.7% (25.3, 38.1) of cases
while 68.3 % (61.9%, 74.7%) of survey tuberculosis
case were smear negative culture positive. Based on the
survey finding sensitivity of symptom screening was 40.6
%( 33.83, 47.37). From the total 167 surveys tuberculosis cases who had X-ray examination 92.1 %( 89, 96.2)
had abnormal Chest X-ray. Among the total survey cases
only 20.3 %(%( 14.8, 25.9) were symptom positive and
smear positive pulmonary tuberculosis.
Conclusion: Close to 80% of survey cases would have
been missed if survey was conducted using routine
programmatic screening tools and diagnosis. The study
underscores the importance of digital X-ray as a
screening tool for improved and early tuberculosis case
detection as found in other studies. GeneXpert and
culture diagnosis will exclude Non-Tuberculosis Mycobacteria (NTM) among notified tuberculosis cases.
OA-350-04 Impact of Ebola virus disease

outbreak on tuberculosis case management in
Guinea
H Traore,1 A M Bangoura,2 D Adjoua,1 C Diallo,1
T Kazadi3 1Population Services International, Conakry,
2
Programme National de Lutte Antituberculeuse, Conakry,
Guinea; 3Population Services International, Washington, DC,
USA. e-mail: htraore@psiguinee.org
Background: Guinea is the epicenter of the Ebola Virus

Disease (EVD) outbreak since March 2014. This
impacted the control of many diseases including tuberculosis (TB). The Global Fund asked PSI and The
National TB Program to look at what is EVDs impact
on TB screening/treatment and address the issue
Design/Methods: The project conducted a meta-analysis
of 8 activity reports (2013 and 2014) from The National
TB Program and Partners against TB in Guinea. Data
was analyzed in Epi InfoTM7.
Results: The impact of EVD affected the continuity of
care of TB in Guinea. Compared to 2013, PSI found there
was a 16% decrease (30% in Middle Guinea) in the
average number of patients screened for TB and 6% in
the average number of TB patients on care at the
Conakry TB Center (CAT la Carrière). This caused a
reduction in case management activities in 22% (n 68)
of TB Screening Centers (TSC), mainly in the forest
region due to fear of health workers contacting EVD,
EVD deaths among health workers, and reassigning
health care providers to EVD treatment activities.
Monitoring activities were difficult to implement, and
the supervision of 12% of all prefectures (n 33) were
reached in Guinea. Moreover, a third (n 15) of TSC in
the forest region experienced stock outs of TB supplies
and drugs, and the effectiveness of implemented education activities and tracing of TB defaulter cases were
compromised.
Conclusion: The results show that EVD had a major
impact on TB screening and treatment in Guinea. To
address the problem, PSI/Guinea, the Global Fund TB
Principal Recipient, in collaboration with the National
TB Program, developed a plan to mitigate the effects of
EVD by focusing on: strengthening capacity of TSC
providers by integrating EVD prevention modules into
TB training curricula; the implementation of prevention
and EVD infection control measures within TSC; and
sensitization of EVD awareness coupled with TB
defaulter case tracing at the community level. This
mitigation plan should strengthen the capacity of 247
health and community workers, and equip 68 TB TSCs
with EVD prevention kits to ensure continuity of care
and follow-up TB activities. The selected interventions
will increase TSCs patient attendance, and better equip
health care providers to deal with the EVD epidemic
while ensuring quality care for TB patients.
S133
OA-351-04 From road map to road:

comprehensive tuberculosis management and
identification of missing private patients in
urban Karnataka, India
O George,1 H Gururaj,2 R Ganesan,1 R Swamickan,3
M Benezet4 1Abt Associates Inc., New Delhi, 2Karnataka
Health Promotion Trust, Bengaluru, 3USA Agency for
International Development, New Delhi, India; 4Abt Associates
Inc., Bethesda, MD, USA. Fax: (91) 112 653 5172. e-mail:
oommen.geo@gmail.com
Background: The need to identify and manage missing

tuberculosis (TB) patients is urgent. Global strategies
promote comprehensive urban TB interventions which
include engagement of private health care providers
(PHCP). While the why and what are alluded to, the
how has largely eluded resource-strapped TB programs.
India needs cost-effective, innovative ideas to engage its
dominant private sector in TB prevention and care.
Intervention: The USAID-funded Strengthening Health
Outcomes through the Private Sector (SHOPS) project in
India developed a private provider interface agency
(PPIA) model to demonstrate impact in 42 towns of
Karnataka. The PPIA engaged the Revised National TB
Control Program (RNTCP), PHCP and communities to
synergistically effect stakeholder collaboration and result
in comprehensive care for TB patients. The model
covered a population of 16.8 million, focusing on 1.1
million living in 663 urban slums. SHOPS mapped and
engaged urban slum communities, and qualified and
unqualified health providers. Access to certified diagnostic tests was enabled through sputum sample transportation and dialogue with RNTCP. TB patients in
slums were offered in-person care and support; others
had access to a telephone Careline counseling and
adherence support service. Community-level patient
support groups aided patients and families live with TB
and fostered treatment compliance.
Results: The intervention halved delays in TB diagnosis
and treatment initiation among new smear positive
patients from intervention slums. Treatment compliance
among 1409 RNTCP and 371 private patients receiving
in-person support by the PPIA was 95% and 97%
respectively. Completed cohorts showed that 94% of 128
privately treated patients had successful outcomes. In the
18 months of intervention, PHCP notified 6342 TB
patients to SHOPS, while PPIA referrals from slums
added 1499 more patients. Children made up 18% of
patients known to SHOPS (11% were less than 5 years of
age). SHOPS reported more than 40% additional private
TB patients by the end of the project.
Conclusions: A comprehensive urban-TB model demonstrated the potential to identify and effectively manage
missing TB patients through inter-sectoral coordination, especially through engagement of private providers
and by respecting patient choices. The model established
ways to improve TB detection under RNTCP, identify
additional private patients, reduce delays and improve
treatment outcomes.
OA-352-04 Modelling health system costs to

support decision making for MDR-TB
diagnostics in Cape Town
R Dunbar,1 P Naidoo,1 I Langley,2 N Beyers1 1Desmond
Tutu TB Centre, Cape Town, South Africa; 2Liverpool School
of Tropical Medicine, Liverpool, UK. Fax: (27) 219 389 719.
e-mail: rdun@sun.ac.za
Background: Improving multidrug-resistant tuberculosis

(MDR-TB) control requires access to accurate and rapid
diagnostics for drug susceptibility testing. New molecular diagnostic tests for tuberculosis (TB) such as Xpert
MTB/RIF (Xpert) hold the promise of improving TB and
multidrug-resistant (MDR) TB diagnosis. Whilst accurate data are readily available from laboratory and
demonstration studies, many operational questions
remain and are most readily answered through modelling.
Methods: South Africa moved from a smear and culture
based algorithm for diagnosis of TB to an Xpert-based
algorithm. The Witness package was used to model all
aspects of both algorithms to reflect the diagnostic
pathways, including the laboratory procedures, as close
to reality as possible, based on information and data
collected during the PROVE IT (Policy Relevant Outcomes from Validating Evidence on ImpacT) study.
Routine operational data was used to validate the model.
The model was used to assess the effect of operational
changes on yield and laboratory costs.
Results: Routine data showed a declining trend in the
yield of TB cases identified over time in both algorithms.
A binomial regression analysis showed no significant
difference in yield between the algorithms when the time
effect was taken into consideration, with a risk difference
of 0.2% (95%CI 1.8% to 2.3%) (P 0.818) in the
Xpert-based algorithm. Total TB diagnostic costs were
increased by 43% from $440,967 in the smear-culturebased algorithm to $632 262 in the Xpert-based
algorithm. The cost per patient diagnosed with TB
increased by 158% from $49 in the smear/culture-based
algorithm to $125 in the Xpert-based algorithm. The
total cost per MDR-TB case diagnosed was decreased by
3%, from $205 in the smear-culture-based algorithm to
S134
$199 in the Xpert-based algorithm. The model was used

to demonstrate the effect of an optimised algorithm,
increased case-finding and other operational changes on
TB yield and laboratory costs.
Conclusion: This study used a combination of operational data and discrete event simulation to answer
important operational questions that could not be readily
addressed through other methods. Modelling is an
attractive and viable option to guide policy makers in
making decisions about the implemention of new
diagnostic tools and algorithms.
Conclusion: Due to the ecological nature of this study we

cannot show a causal relationship between social
protection levels and TB rates. This work could influence
funding agencies, such as the Global Fund to Fight AIDS,
Tuberculosis, and Malaria, as well as national treatment
programs to increase their funding of programs outside
of the medical sphere.
OA-353-04 Social protection and TB: an

ecological analysis
A Siroka,1 K Lonnroth,2 N Ponce1 1University of California,
Los Angeles, Los Angeles, CA, USA; 2World Health
Organization, Geneva, Switzerland. e-mail:
asiroka@gmail.com
Background: Pillar Two of the End TB Strategy stresses

the need for social protection programs that aim to
alleviate poverty. Without progressive policies in areas
outside of the medical sphere it will be impossible to
achieve the ambitious goals set forth in the End TB
Strategy. There is a limited amount of research on the
relationship between social protection and TB. Reeves et
al. show the inverse relationship between social protection and TB in European nations with sizeable social
protection systems and secure welfare mechanisms. This
paper builds upon this work by analyzing this association
with a global purview.
Design/Methods: We aim to show the association
between levels of social protection, measured as the
percentage of GDP spent on social protection programs
and national TB prevalence, incidence, and mortality
rates. Social protection data, which covers the years 2000
2012, were obtained from the International Labor
Organizations (ILO) Social Protection Department. TB
burden rates are from the World Health Organizations
Global TB database. The full model includes nationallevel factors that could also influence TB rates. These
include GDP per capita, HIV rates, and the strength of
the health system. Further control variables are population density, percentage of foreign-born residents, and
TB treatment success rate among new cases. This last
control variable is meant to represent the strength of the
national TB treatment program. To account for the
clustering of data points within countries we used a
country-level fixed effects model.
Results: In multivariate models, social protection levels
were significantly inversely associated with all three
measures of TB burden. This was formally tested using a
joint significance test of social protection and its squared
term. The association is especially strong in setting with
low levels of social protection spending (as a percentage
of GDP). For example, moving from 0% to 1% of GDP
in social protection spending within a country is
associated with an 1833 person per 100 000 decline in
the TB prevalence rate.
OA-354-04 Four degrees of separation: the

influence of social and provider networks in
the steps to diagnosis of active tuberculosis in
Urban Uganda
J Sekandi,1,2 S Zalwango,1 R Kakaire,2 L Martinez,2
A Kizza,2 A Ezeamama,2 N Kiwanuka,1 CC Whalen2
1
Makerere University School of Public Health, Kampala,
Uganda; 2University of Georgia, Athens, GA, USA. e-mail:
jsekandi@musph.ac.ug
Background: Delay in tuberculosis (TB) diagnosis

adversely affects patients health outcomes and prolongs
transmission in the community. The steps taken by active
TB patients to seek a diagnosis are not well understood.
Design/Methods: Using retrospective cohort design, TB
patients on treatment for three months or less and aged
718 years were enrolled from three public clinics in
Kampala, Uganda, between March and July 2014.
Structured face-to-face interviews and network analysis
were used to evaluate information about patients social
and health provider networks. Steps, or degrees of
separation, were defined as the number of contacts a
patient made to seek help since recognition of symptoms
to the time of final diagnosis. Social contacts included
family members, friends and coworkers. Health providers included public and private hospitals, private clinics,
public health centers, pharmacies, village health workers
and herbal healers. Descriptive and step sequence
analyses were performed, followed by Cox regression
analyses to determine factors associated with steps and
time to final TB diagnosis. Hazard ratios and 95%
confidence intervals were calculated to represent the
likelihood of timely TB diagnosis.
Results: Of 294 TB patients, 58% were male and median
age was 30 (IQR; 24-38) years. The median number of
steps was 4 (IQR: 3, 7) corresponding to 70 (IQR:

28,140) days to diagnosis. New patients had more steps
and time to diagnosis compared retreatment patients (5
vs. 3, P , 0.0001; 84 vs. 46 days P , 0.0001). Fifty-eight
percent of patients first contacted persons in their social
network. The first step to seeking care accounted for
41% of the patients time to diagnosis while visits to nonTB providers and TB providers (without a TB diagnosis)
accounted for 34% and 11% respectively. New TB
patients vs. retreatment (HR: 0.66, 95%CI; 1.11, 1.99),
those who first contacted a non-TB provider vs.
contacting social network (HR: 0.72 95%CI; 0.55,
0.95) and HIV seronegative vs. seropositive patients
(HR: 0.70, 95%CI; 0.53, 0.92) had a significantly lower
likelihood of a timely final diagnosis.
Conclusion: On average, there were four steps or degrees
of separation between the onset of symptoms in a TB
patient and a final diagnosis. The patients social and
provider networks influenced their respective diagnostic
pathways. Most delays occurred in the first step,
representing the patients initial decisions to seek help,
or through interactions with non-TB providers. TB
control programs should strengthen community education and TB screening as well as continuing medical
education to health providers to ensure timely diagnosis
of TB.
OA-355-04 Households receiving a TB-specific

social protection intervention in Peruvian
shantytowns placed more importance on social
than economic support
R Montoya,1 M Tovar,1,2 D Huff,1,3,6 T Wingfield,1,4,5
E Ramos,2 J Lewis,1,3 D Boccia,1,3 C Evans1,2,4 1Asociacion
Benefica
PRISMA, Lima, 2Universidad Peruana Cayetano
Heredia, Lima, Peru; 3London School of Hygiene & Tropical
Medicine, London, 4Imperial College London, London,
5
North Manchester General Hospital, Manchester, UK;
6
Tulane University, New Orleans, LA, USA. e-mail:
tom.wingfield@ifhad.org
Background: Social protection forms part of a key pillar

in the World Health Organizations post-2015 global TB
control strategy. However, evidence concerning TBspecific social protection is limited and it is unclear
whether social and/or economic support will have the
most impact on TB control. During the pilot of a
Community Randomized Evaluation of a Socioeconomic
Intervention: to Prevent TB (CRESIPT), we assessed
acceptability of a TB-specific social protection intervention and evaluated feedback on the differential importance of social vs. economic support in TB-affected
households.
Methods: 282 TB patients diagnosed by the Peruvian
National TB Program in 32 shantytown communities of
Lima, Peru, were recruited to the study. Patients were
randomly allocated to the comparison arm (standard of
care from the Peruvian NTP) or supported arm (standard
of care plus the social protection intervention). The social
protection intervention consisted of economic support
and social activities. Social activities included household
visits and participatory community meetings for infor-
S135
mation, mutual support, stigma reduction and empowerment. Economic support was provided through conditional cash transfers throughout treatment to reduce TB
vulnerability, incentivise and enable care through mitigation of TB-related costs. A mixed-methods approach
collected quantitative/qualitative feedback in focus
group discussions and exit questionnaires with all
participants.
Results: 135 patient households were randomised to the
support arm of whom 115 have had final follow up to
date. 8/135 (6%) recruited patient households did not
engage at all with the intervention and thus did not
receive any conditional cash transfers. 890/1157 (80%)
potential conditional cash transfers were made. Patient
households ranked the social activities of the social
protection intervention higher than the economic support (Figure). 88/92 (96%) patient households with data
reported that they would participate again in the
intervention if a household member fell ill.
Conclusion: In this setting, there was good uptake and
acceptance of the novel TB-specific social protection
intervention including conditional cash transfers. However, TB-affected household feedback suggested that the
social activities of the intervention were perceived to be
more important than the economic support. This
evidence should be considered in future planning of
social protection interventions incoporated into TB
control strategies.
S136
08. The molecular basis of PZA resistance:

from bench to bedside
OA-356-04 Mutations in gyrA and gyrB among
fluoroquinolone- and multidrug-resistant
Mycobacterium tuberculosis isolates
J-Y Chien,1,2 S-T Chien,3 C-J Yu,1 P-R Hsueh1 1National
Taiwan University Hospital, National Taiwan University
College of Medicine, Taipei, 2Graduate Institute of Clinical
Medicine, National Taiwan University College of Medicine,
Taipei, 3Chest Hospital, Ministry of Health and Welfare,
Tainan, Taiwan. e-mail: jychien@ntu.edu.tw
Background: To evaluate the correlation between mutations in the gyrA and gyrB genes and the level of
resistance to ofloxacin (OFX) and moxifloxacin (MFX)
in isolates of multidrug-resistant Mycobacteria tuberculosis (MDR-TB).
Design/Methods: Using minimum inhibitory concentrations (MICs) measured by the broth microdilution
method, we categorized the isolates into OFX-susceptible (MIC2 mg/L), low- (MIC 4-8 mg/L) and high-level
(MIC 316 mg/L) OFX-resistant isolates, and MFXsusceptible (MIC 0.5 mg/L), low- (MIC 1-2 mg/L) and
high-level (MIC 34 mg/L) MFX-resistant isolates.
Results: Of 111 isolates, mutations in the gyrA gene were
found in 30.2% of OFX-susceptible isolates, 72.5% of
low-level OFX-resistant isolates and in 72.2% of highlevel OFX-resistant isolates, and in 28.6% of MFXsusceptible isolates, 58.1% of low-level MFX-resistant
isolates and in 83.9% of high-level MFX-resistant
isolates. Compared with OFX-susceptible isolates, lowand high-level OFX-resistant isolates had a significantly
higher prevalence of mutations at gyrA codons 8894 (9/
53, 17.0% vs. 26/40, 65.0% and 13/18, 72.2%,
respectively; P , 0.001) and a higher prevalence of the
gyrB G512R mutation (0/53, 0.0% vs. 1/40, 2.5% and 3/
18, 16.7%, respectively; P 0.006). Similarly, compared
with MFX-susceptible isolates, low- and high-level
MFX-resistant isolates had a significantly higher prevalence of mutations at gyrA codons 8894 (7/49, 14.3%
vs. 16/31, 51.6% and 25/31, 80.6%, respectively; P ,
0.001) and a higher prevalence of the gyrB G512R
mutation (0/49, 0.0% vs. 0/31, 0.0% and 4/31, 12.9%,
respectively; P 0.011). D94G and D94N mutations in
gyrA and the G512R mutation in gyrB were correlated
with high-level MFX resistance while the D94A mutation was associated with low-level MFX resistance. We
also found that the prevalence of those mutations was
higher among fluoroquinolone-susceptible East Asian
(Beijing) and Indo-Oceanic strains than among fluoroquinolone-susceptible Euro-American strains.
Conclusion: Mutations at codons 88-94 of the gyrA gene
and the G512R mutation in the gyrB gene were
associated with resistance to OFX and MFX, although
the level of resistance was not equally affected by those
mutations. Our finding that the prevalence of gyrA and
gyrB gene mutations was higher among fluoroquinolonesusceptible East Asian (Beijing) and Indo-Oceanic strains
than among fluoroquinolone-susceptible Euro-American
strains imply that molecular techniques to detect FQ

resistance may be less sensitive in areas with a high
prevalence of those strains.
OA-357-04 Identification of pncA gene regions

assoicated with high-level pyrazinamideresistant TB in China
D Li,1,2,4 Y Hu,1,2,3 J Werngren,4 M Mansjo,4 S Hoffner,3,4
B Xu1,2 1School of Public Health, Fudan University, Shanghai,
2
School of Public Health, Fudan University, Shanghai, China;
3
Microbiology and Tumor Biology Center, Karolinska
Institutet, Stockholm, 4The Public Health Agency of Sweden,
Stockholm, Sweden. e-mail: yhu@fudan.edu.cn
Background: A mutation in the pncA gene is a major

mechanism of pyrazinamide (PZA) resistance. However,
the mutations are highly diverse and scatter over the full
length of the 561 bP of the pncA gene. This may
complicate the development of rapid molecular diagnosis.The present study aimed to investigate the genetic
characterization of pyrazinamide resistant M.TB isolated
from China as well as to analyze its relationship to the
minimum inhibitory concentration (MIC), in attempt to
search for the valuable hotspot regions associated with
high-level PZA resistance.
Design/Methods: A population-based multi-center study
was carried out in 4 Chinese counties. Pulmonary TB
cases registered in the local TB health facilities from 2011
to 2013 were enrolled in the study. Mutations in the
pncA gene were identified by whole gene sequencing.
MIC test for pncA mutated isolates were performed on
the mycobacterial growth indicator tube (MGIT) 960
platform according to the standard manufacturer protocol.
Results: Of 912 cases enrolled within the studied period,

878 were culture positive and 136 isolates had the
nonsynonymous pncA mutation including 133 with
Single Nucleotide Polymorphism (SNP) and 3 with the
large fragment insertions or deletions. Of the 147 PZA
resistant isolates, we identified 131 pncA nonsynonymous mutation, with a sensitivity of 89.1% (95%CI:
83.05%-93.19%) and a specificity of 99.3% (95%CI:
98.41%-99.71%) to detect the PZA resistance. Besides,
all 133 SNP distributed between position 11 and
position 180 of pncA gene and concentrated in five
regions with statistically different MIC level distributions
(P , 0.001): 73.2% (71/97) isolates with high-level MIC
(MIC.500lg/ml) and 95.1% (58/61) isolates with
S137
extreme high-level MIC (MIC.1024lg/ml) had the

mutation at three regions (position 51 to 76, position 130
to 142 and position 163 to 180); 72.2%(26/36) isolates
with MIC between 128 to 500lg/ml were observed to
have the mutation in the other two regions (position 94
to 97 and position 11 to 16).
Conclusion: The specific regions of pncA gene have
approved their value for the molecular rapid diagnosis of
PZA resistance in M.TB, especially for three regions to
define the high PZA resistance.
Conclusions: Among tuberculosis patients in Uganda,

resistance is mainly due to common Mycobacterium
tuberculosis resistance mutations, and mostly to first-line
treatment. PZA resistance was found in 52% of MDRs,
while injectable resistance is low, and the unusual gyrA
mutation unlikely confers fluoroquinolone resistance.
OA-358-04 Prevalence of common and rare

Mycobacterium tuberculosis resistanceconferring mutations in Uganda
M De Vos,1 J Rice,2 N Ismail,3 B Kreiswirth,4 T Dolby,5

P Van Helden,1 R M Warren,1 L Wangh2 1Stellenbosch
University, Cape Town, South Africa; 2Brandeis University,
Waltham, MA , USA; 3National Institute for Communicable
Diseases, Johannesburg, South Africa; 4University of
Medicine and Dentistry of New Jersey, New Jersey Medical
School, Newark, NJ, USA; 5National Health Laboratory, Cape
Town, South Africa. e-mail: margab@sun.ac.za
W Ssengooba,1,2 F Cobelens,1,3,4 D Lukoye,3,5

G W Kasule,5 M Joloba,2,5 B De Jong6,7 1Academic Medical
Center, Amsterdam Netherlands; 2Makerere University
College of Health Sciences, Kampala, Uganda; 3Amsterdam
Institute of Global Health and Development, Amsterdam,
Netherlands; 4KNCV Tuberculosis Foundation, The Hague,
Netherlands; 5Ministry of Health National TB program,
Kampala, Uganda; 6Institute of Tropical Medicine, Antwerp,
Belgium; 7New York University, New York, NY, USA. Fax:
(256) 414 533 033. e-mail: willyssengooba@gmail.com
Background: Understanding the circulating Mycobacterium tuberculosis (MTB) resistance mutations is vital for
better TB control strategies, especially in early stages of
second-line TB treatment programme introduction.
Design/Methods: We performed whole genome sequencing on 112 rifampicin and/or isoniazid-resistant isolates
from two drug resistance surveys done in Uganda. We
compared MTB strain lineages with drug resistance,
previous TB treatment and HIV-infection. We screened
for common and rare drug resistance mutations and their
relation with MTB-lineage and HIV-infection.
Results: 73 (65.2%) patients were male, median age was
35 years (IQR; 26-45), 33 (29.5%) were HIV-infected
and 77 (68.8%) were previously treated for tuberculosis.
Phenotypic mono-resistance to rifampicin (RIF) was 6
(5.0%), to isoniazid (INH) 67 (60.0%) and 39 (35.0%)
were MDR-TB. MTB lineages were L2 (East-Asian)
4.5%, L3 (East-African-Indian) 17.8% and L4 (EuroAmerican) 77.7%. Of the 70 sequencing results available
to date, drug resistance mutations were found in 42.9%
for RIF, 80% for INH, 28.6% for ethambutol (EMB),
19% for pyrazinamide (PZA), and 17.1% for streptomycin (STR). The most prevalent first-line mutations per
drug were RIF rpoB.S450L 47% (present in 92% of
MDR), INH katG.S315T 75% and inhA.pro-15T 20%,
EMB embB360 70%, STR rpsL.K43R 58% and pyrazinamide pncA.V180F 23%. Only two patients had RIF
compensatory mutations, rpoC.N698S and rpoC.V483A, both with rpoB.S450L mutation against a L4
background. One MDR-TB patient had kanamycin/
amikacin resistance rrsk.R468S. There was an unusual
QRDR gyrA.T80A mutation in 48% of the TB patients,
mostly in L4 (89%) and among previously treated
patients. None of the other mutations were associated
with HIV, previous treatment or MTB lineage.
OA-359-04 Construction and validation of a

three-color single-tube assay for the detection
of resistance to first- and second-line antituberculosis drugs
Background: The Genotypew MTBDRplus line probe

assay and the XpertwMTB/RIF assay, currently endorsed
by the WHO only provide evidence for resistance to
isoniazid and/or rifampicin. These data are insufficient to
ensure optimal treatment of MDR-TB when resistance
extends beyond those first line drugs. Moreover, treatment of MDR-TB in the absence of comprehensive drug
susceptibility testing promotes the emergence of XDRTB. Therefore there is a critical need for a convenient,
single-tube diagnostic technology that can test all known
mutations conferring TB drug resistance. The present
report describes such a reaction for first- and second-line
antibiotics.
Design/Methods: Using a combination of multiplexed
LATE-PCR and Lights-on/Lights-off probe technology
we have constructed a three color single tube assay to
screen and identify mutations in seven genetic targets
known to confer resistance to first-line (rifampicin and
isoniazid) and second-line (fluoroquinolone, aminoglycoside/polypeptide and ethionamide) antibiotics. The
assay was evaluated on DNA extracted from well
characterized clinical M. tuberculosis isolates harboring
all known mutation types in inhA and eis promoters,
rpoB, katG, gyrB, gyrA, and rrs genes. This comprises a
total of 44 possible allelic variants in a multitude of
possible combinations. One thousand blinded DNA
samples have been collected to evaluate the diagnostic
performance of the assay.
Results: Analysis of 81 resistant isolates generated a
library of signatures reflecting all of the known
mutations in inhA and eis promoters, rpoB, katG, gyrB,
gyrA, and rrs genes. Each unique mutation had its own,
highly reproducible fluorescent signature which was
distinct from that of pan-susceptible laboratory strain,
H37Rv, as well as all other isolates with different
mutations at the same or different codons.
Conclusion: This study demonstrates that a single tube
multiplexed assay can simultaneously distinguish different mutations that confer resistance to rifampicin,
isoniazid, ofloxacin, moxifloxacin, amikacin, kanamycin, capreomycin and ethionamide in less than three
S138
hours. This assay is currently being validated with 1000

DNA samples. Automation of this methodology will
enable clinicians to optimise treatment of drug resistant
TB thereby ensuring improved outcomes. This study is
funded by the NIH R01 A1099532.
OA-360-04 Association between genotypic and

phenotypic pyrazinamide resistance in INH and
RMP mono-resistant and MDR-TB isolates
M Whitfield,1 E Streicher,1 L Scott,2 W Stevens,2
S Sampson,1 P Van Helden,1 R M Warren,1 A Van Rie3,4
1
Stellenbosch University, Tygerberg, 2National Health
Laboratory Services, South Africa, Johannesburg, South
Africa; 3University of Antwerp, Antwerp, Belgium; 4University
of North Carolina at Chapel Hill, NC, USA. e-mail:
mwhit@sun.ac.za
Background: Pyrazinamide (PZA) plays an integral part

in anti-tuberculosis treatment. PZA will likely remain an
important component of treatment regimens for drugsusceptible and multidrug-resistant tuberculosis (MDRTB) because of its distinctive mode of action. However
little is understood about the prevalence of PZA
resistance with rifampicin (RIF) and isoniazid (INH)
mono-resistant cohorts.
Methods: A total of 87 isolates that were INH monoresistant, 391 isolates RIF mono-resistant and 370
MDR-TB isolates according to the MTBDRplus line
probe assay were identified from a culture bank housed
at Stellenbosch University, South Africa. The pncA gene
was sequenced for all isolates including up- and
downstream regions. Isolates were phenotypically tested
using the BACTEC MGIT 960 PZA drug susceptibility
test (DST) at a concentration of 100lg/ml. INH or RIF
mono-resistant were subjected to culture based phenotypic DST (BACTEC 960 SIRE kit) to confirm INH or
RIF mono-resistance, respectively.
Results: 79 (90.8% - 95%CI 99.1%) of the 87 INH
mono-resistant isolates were found to be wild type for
pncA; 318 (81.5% - 95%CI 93.2%) of the 391 RIF
mono-resistant isolates were found to be wild type for
pncA; and 184 (49.7% - 95%CI 96.4%) of the 370
MDR-TB isolates were found to be wild type for pncA.
All isolates with a wild type pncA genotype were
confirmed as PZA sensitive using the PZA DST. A total
of 62 different mutations were identified in the pncA
gene.
Conclusion: A clear increase in PZA resistance is
observed from INH mono-resistance (9.2%) to RIF
mono-resistance (18.5%) and finally to MDR-TB
(50.3%). This questions the utility of PZA in the
WHO recommended standardised MDR-TB treatment
regimen. The development of simplified routine diagnostic methods for PZA susceptibility will be essential if
new treatment regimens continue to rely on the inclusion
of PZA.
OA-361-04 Genotypic testing using nextgeneration sequencing is superior to

phenotypic testing for pyrazinamide
susceptibility testing
S Omar,1 N Ismail,1 A W Dreyer,1 N Ismail1 1National
Institute for Communicable Diseases, Johannesburg, South
Africa. e-mail: shaheedvo@nicd.ac.za
Background: Pyrazinamide (PZA) is an important drug

for the treatment of Mycobacterium tuberculosis infection, with efficient sterilization ability. This pro-drug is
converted to pyrazinoic acid by the enzyme pyrazinamidase (encoded by the pncA gene) produced by the
organism. Phenotypic testing is technically challenging
and often produces unreliable results due to dependence
on media-specific conditions. The genes of interest
(pncA) together with other genes (rpsA and panD) are
molecular markers for PZA resistance. Development of a
rapid, reliable molecular assay for determination of
resistance is difficult as the mutations associated with
resistance are spread across the entire length of the pncA
gene. We determined the value of whole genome
sequencing (WGS) using the next-generation method
for the determination of PZA resistance from isolates
collected in high-burdened TB South Africa.
Design/Methods: A total of 860 M. tuberculosis isolates
collected from a province in South Africa, a subset of a
larger national survey, were included in this study. All
isolates were phenotypically tested for PZA drug
susceptibility using the BD BACTEC 960 mycobacterial
detection system and the BD MGIT 960 PZA kit. WGS
was performed on the Illumina MiSeq instrument with
libraries prepared using Nextera XT chemistry. The
modified Waynes test for pyrazinamidase activity was
performed on all PZA resistant isolates by either method.
All test methods were performed at the National
Tuberculosis Reference Laboratory, South Africa.
Results: Of the 860 isolates tested, PZA resistance was
prevalent at 3.37% phenotypically and 2.44% genotypically. Resistance associated mutations were only found
in the pncA gene and none in the rpsA or panD genes.
Agreement between the phenotypic method and WGS
was 99% for susceptible isolates and 55% for resistant
isolates. Pyrazinamidase activity testing showed greater
agreement with WGS compared to MGIT-based phenotype; 91% for PZA resistant and 100% for PZA
susceptible (Mcnemar; P 0.5) compared to 52% and
20% (McNemar; P 0.03) respectively for the MGITbased phenotypic method. This implies an approximate
50% false resistant rate when using the phenotypic
method.
Conclusion: The phenotypic method lead to a significant
number of false PZA resistance. Molecular PZA susceptibility testing by WGS is a more accurate tool for this
important drug compared to the phenotypic method,
thereby improving benefit to TB patients.
OA-362-04 Large pncA gene deletions in

Mycobacterium tuberculosis: a novel
mechanism of pyrazinamide resistance
E Streicher,1 R Van Der Merwe,1 S Sampson,1
A Dippenaar,1 M Whitfield,1 N Da Camara,1
P Van Helden,1 P Arnab2 1Biomedical Sciences,
Stellenbosch University, Tygerberg, South Africa; 2King
Abdullah University of Science and Technology,
Thuwal-jeddah, Saudi Arabia. e-mail: lizma@sun.ac.za
Background: PZA has been used for the treatment of TB

for the past 50 years and therefore its continued use in
new and second line-treatment regimens requires accurate drug susceptibility testing to ensure optimal treatment. PZA susceptibility testing is currently not done
routinely due to the complexity of the culture based
methods. Molecular methods offer a rapid means to
identify mutations in the pncAgene which have been
previously associated with PZA resistance. This study
aimed to investigate the failure to PCR amplify the pncA
gene in selected clinical isolates using the recently
reported DNA sequencing method to detect pncA
mutations.
Design/Methods: A convenient sample set of 669 clinical
strains with various drug susceptibility profiles and
representing different strain lineages of Mycobacterium
tuberculosis were sequenced using the Illumina HiSeq
sequencing platform. An in-house pipeline was used to
align sequencing reads to the reference M. tuberculosis
H37Rv genome and identify SNPs, insertions and
deletions. Deletions were confirmed by Sanger sequencing. A phylogenetic tree was constructed to investigate
the relatedness of strains.
Results: Whole genome sequence analysis identified 23
clinical isolates in which the pncA gene was either
partially or completely deleted. Six different deletion
boundaries were identified, of which the deleted area
ranged from 298 to 17133 bp in length. Two isolates
showed mixed infections where the deletion was absent
in one strain and present in another. Phylogenetic
reconstructions confirmed clonality of strains with
identical deletions.
Conclusion: This study adds to the complexity of the
molecular basis of PZA resistance. The absence of part of
or the entire pncA gene in clinical isolates needs to be
considered when using molecular methods to determine
PZA resistance. This study also highlights that nonessential genes can be deleted without compromising the
ability of these strains to transmit. However, an in vitro
growth deficit was observed in isolates harbouring these
deletions.
S139
OA-363-04 Virtual sequencing of the entire

pncA gene target in a single tube using LatePCR and Lights-On/Lights-Off probes to predict
PZA susceptibility
B Kreiswirth,1 J Rice,2 L Wangh,2 M Whitfield,3
R M Warren,3 J Posey,4 P Bifani,5 S Marras1 1PHRI - Rutgers
University, Newark, NJ, 2Brandeis University, Waltham, MA,
USA; 3Stellenbosch University, Tygerberg, South Africa; 4U.S.
USA; 5Novartis Institute, Singapore, Malaysia. e-mail:
kreiswba@njms.rutgers.edu
Background: Pyrazinamide (PZA) is a key first line

antibiotic to treat tuberculosis and it has become a
critical drug in recent clinical trials with experimental
agents. Historically, phenotypic testing for PZA resistance has been problematic and in most settings the drug
is used empirically with no laboratory support. Genotypic analysis of resistance has revealed over 371
mutations in pncA, the vast majority iq non-synonymous
and correlate with PZA resistance. Importantly, sequencing data also revealed that a wild-type pncA gene
correlates with PZA susceptibility. We have developed
a LATE-PCR assay using Lights-On/Lights-Off probes to
distinguish the wild type drug susceptible pncA gene
sequence from all strains that have non-synonymous
genetic alterations.
Design/Methods: LATE-PCR is used to amplify a 685
base long amplicon encompassing the entire coding
region of the pncA gene, including the upstream flanking
sequence. Following amplification, the temperature of
the reaction is lowered to 25oC resulting in the coating of
the single-stranded DNA templates with 35 Lights-On/
Lights-Off fluorescent probes labeled in three colors. The
probes are then melted off of the target and the resulting
fluorescent signature reveals the presence/absence of
mutations, which are all distinguished from a highly
reproducible wild-type profile. Total time per assay 3.5
hrs.
Results: The assay, run in triplicte, has been tested
against a blinded panel of 206 DNA samples isolated
from 103 in vitro resistant PZA mutants in the CDC1551
background, and 103 PZA resistant and susceptible
clinical mutants. Among the 206 DNAs, 196 (95%)
samples were correctly genotyped as wild type (susceptible) or mutant (presumed non-susceptible). One susceptible isolate with a wild type pncA was incorrectly
called mutant and 9 isolates with mutations were
incorrectly called wild type.
Conclusion: We have successfully designed and tested a
single closed-tube assay that is able to identify and
distinguish the wild type pncA gene from strains with any
genetic alterations and provide an accurate and rapid
method to predict a PZA susceptible strain. Additional
refinements will improve identification of the rare
synonymous mutations in pncA and the few nonsynonymous that were not correctly scored as PZA
resistance. Supported by NIH Grant #R21AI106551.
S140
09. HIV and lung health

OA-364-04 CD4 count and risk of incident
tuberculosis in HIV-infected individuals: a doseresponse meta-analysis
W-C Lo,1 S-C Pan,1,2 H Fu,1 H-H Lin1 1Institute of
Epidemiology and Preventive Medicine College of Public
Health, National Taiwan University, Taipei, 2Department of
Internal Medicine, National Taiwan University Hospital, Taipei,
Taiwan. e-mail: nicholaslo0114@gmail.com
Background: The CD4 count level is an indicator for the

risk of tuberculosis among HIV-infected individuals, but
the dose-risk relation remains unclear. We systematically
reviewed and meta-analyzed literature to examine and
quantify the dose-response association between CD4
count level and risk of tuberculosis among HIV-infected
individuals.
Design/Methods: Prospective cohort studies were
searched for in PubMed until April 2014. Three
independent reviewers conducted abstract and full-text
review and extracted relevant information. Quality of
included studies was assessed using Newcastle-Ottawa
scale. Data were pooled using dose-response randomeffects models. When nonlinear effect was identified, the
generalized least squares cubic spline model was used for
dose-response analysis. Subgroup analysis was performed stratifying by the status of antiretroviral therapy
(ART) use.
Results: We identified 27 prospective cohort studies,
which included 141,894 HIV-infected individuals with
the follow-up period ranging from 6 to 72 months. In the
pooled analysis, the risk of tuberculosis increased with
decreasing CD4 count: the summary relative risk (RR)
was 1.41 [95% confidence interval (CI) 1.30-1.54, I2
77.5%] per 100 cell/lo decrease in CD4 count. Nonlinear association was noted with Pnon-linearity 0.005.
Dose-response analysis suggested that the risk of
tuberculosis increases rapidly with declining CD4 count
when the level of CD4 count was below 350 cell/lo
(Figure 1a). When stratifying by the status of ART use,
the rise of tuberculosis risk with declining CD4 count
was faster in those with partial (,57%) or no ART
(Figure 1b). Even among those with ART, the risk of
tuberculosis still increased with declining CD4 count,
suggesting that delaying the initiation of ART might
increase tuberculosis risk (Figure 1c).
Conclusion: This dose-response meta-analysis of prospective studies revealed a nonlinear inverse association
between CD4 count and active tuberculosis risk.

Information from this analysis could help determine the
potential benefits of early ART initiation on reducing
tuberculosis risk.
OA-365-04 Synergic effect on survival of

streptomycin and an intensified antituberculosis regimen in HIV-infected patients
with tuberculous meningitis
G Alvarez-Uria,1 M Midde,1 R Pakam,1 P S Yalla,1
P K Naik1 1Rural Development Trust Hospital, Bathalapalli,
India. e-mail: gerardouria@gmail.com
Background: The mortality of tuberculous meningitis

(TM) in HIV infected patients is extremely high, and
there has not been any major therapeutic breakthrough
in the last decades.
Design/Methods: This is an observational study from a
cohort study in Anantapur, India. We compared the
mortality during the first nine months after TM diagnosis
of 228 HIV-infected patients treated with a standard
anti-tuberculosis therapy (sATT) (79 patients), an
intensified ATT (iATT) (119 patients) and an iATT with
streptomycin (iATTSTM) (30 patients). The sATT
included rifampicin, isoniazid, ethambutol and pyrazinamide. The iATT included levofloxacin, ethionamide,
pyrazinamide and double dosing of rifampicin and
isoniazid, and was given only during the hospital
admission (median 7 days, interquartile range 6-9).
Univariate and multivariate analyses were performed
using Cox regression proportional hazard models.
Continuous variables that did not have a linear
relationship with the log-hazard were transformed using
restricted cubic splines.
Results: The mortality was similar in patients who
received the sATT and the iATT. However, patients in
the iATTSTM had significant lower mortality risk than
patients in the sATT group (hazard ratio [HR] 0.47, 95%
confidence interval [CI] 0.24 to 0.93), and the mortality
difference at nine months was 23.3% (95%CI 2.1 to
44.5). After adjusting for other covariates, the mortality
hazard of the iATTSTM vs. the sATT remained
statistically significant (adjusted HR [aHR] 0.2, 95%CI
0.09 to 0.46). Other factors associated with mortality
were being previously treated of tuberculosis (aHR 3.23,
95%CI 1.68 to 6.24), low albumin concentrations (aHR
0.74 per increase of 1 g/dl, 95%CI 0.58 to 0.95) and low
CD4 lymphocyte concentrations. The risk of death
increased exponentially only with CD4 lymphocyte
concentrations below 100 cells/ll. High C-reactive
protein concentrations were not associated with mortality in a subgroup of patients who received the iATT (with
or without STM).
Conclusions: In this observational study in a resourcepoor setting, the use of iATT with STM resulted in a
clinically important reduction in mortality compared
with the standard of care in HIV infected patients with
TM. The results of this study deserve further research.
OA-366-04 One third of in-patients diagnosed

with HIV-associated TB in Khayelitsha, South
Africa, have mycobacteraemia associated with
high mortality
S Janssen,1,2 C Schutz,2 A Ward,2,3 R Burton,3
R Wilkinson,2 M Grobusch,1,2 T Van Der Poll,1,2
G Meintjes2 1Academic Medical Center, University of
Amsterdam, Amsterdam, Netherlands; 2Clinical Infectious
Diseases Research Initiative, Institute of Infectious Disease
and Molecular Medicine, University of Cape Town, Cape
Town, 3Khayelitsha Hospital, Khaylitsha, South Africa. e-mail:
s.janssen@amc.nl
Background: Mortality in patients with HIV-associated

tuberculosis (TB) remains high despite availability of
effective therapy, particularly amongst patients ill
enough to require hospitalisation. We investigated
prevalence and prognostic significance of TB mycobactaeremia in association with sepsis biomarkers.
Design/Methods: We conducted a prospective cohort
study at Khayelitsha Hospital in Cape Town, South
Africa. Hospitalized HIV-infected patients with CD4
counts ,350 cells/uL that were diagnosed with, or had
high clinical suspicion of TB on admission were selected
randomly. At enrolment, sputum and blood cultures (BC)
for TB, and sputum and urine Xpert were performed
when specimens could be obtained. We restricted
analyses to patients with microbiologically proven TB,
who had a TB BC performed and did not have a BC
positive for other pathological bacteria. Clinical characteristics and biomarkers in mycobactaeremic patients
were compared to TB cases with negative TB BC. We
used Cox proportional hazard models to assess the
association of mycobactaeremia with time to death,
adjusted for pre-specified co-variates (sex, age, antiretroviral therapy (ART) status, CD4 count, HIV viral load
and haemoglobin).
Results: From January-September 2014 142 individuals
were recruited, 116 had a discharge diagnosis of TB.
Mycobacteraemia was found in 33% (n 39). 73
microbiologically confirmed cases with a TB BC result
were included in subsequent analyses. Of them, 53%
were male; median age was 35 years (interquartile range
(IQR) 28-43) and median CD4 count 59 cells/lL (IQR
17-102). Mycobacteraemia was found in 53%; 28% (n
S141
11) of mycobacteraemic patients died within 12 weeks,

vs. 17% (n 6) of patients with negative TB BC.
Mycobactaeremic patients had elevated plasma concentrations of procalcitonin (median 4.4 vs. 0.8 ug/L, P ,
0.005), CRP (median 176 vs. 105 mg/l, P 0.02) and Ddimer (1.3 vs. 1.1 mg/L, P 0.09), when compared with
TB BC negative patients. Haemoglobin and lactate
concentrations were not different between groups, but
mortality was associated with higher lactate (median 2.9
vs. 1.7 mmol/L, P 0.003). Mycobactaeremia independently predicted mortality. Older age, female sex and
defaulting ART were also independently associated with
mortality (Table).
Conclusion: In hospitalized patients with HIV-associated
TB, mycobactaeremia was present in one-third, and
independently predicted of mortality. Mycobactaeremia
was associated with elevations of biomarkers that also
characterize bacterial sepsis.
Table: Time to death in patients with microbiologically
confirmed TB
Variable
Univariate analysis
Mycobacteraemia
Female sex
Age (per 10 years increase)
CD4 count (per 100 cells/uL
increase)
HIV viral load (per 100 000
copies/mL increase)
Haemoglobin (per 1 g/dL
increase)
ART status
ART nave vs on ART
Default vs on ART
Multivariate analysis
Mycobacteraemia
Female sex
Age (per 10 years increase)
CD4 count (per 100 cells/uL
increase)
HIV viral load (per 100 000
copies/mL increase)
Haemoglobin (per 1 g/dL
increase)
ART status
ART nave vs on ART
Default vs on ART
Crude HR
95% CI
P value
1.77
2.23
1.93
0.66-4.79
0.82-6.03
1.31-2.83
0.26
0.11
0.001
0.57
0.26-1.24
0.15
1.00
0.95-1.04
0.89
1.05
0.85-1.30
0.66
0.81
2.12
0.25-2.65
0.64-6.95
0.73
0.22
4.69
3.50
3.03
1.15-19.19
0.03
1.12-10.96
0.03
1.76-5.23 ,0.001
0.92
0.40-2.09
0.84
0.97
0.92-1.03
0.37
0.93
0.73-1.17
0.52
1.11
4.73
0.29-4.31
1.15-19.52
0.88
0.03
S142
OA-367-04 Low implementation of XpertW

MTB/RIF among HIV-TB co-infected adults: a
survey of 19 low/middle-income countries
from the IeDEA Consortium
operational research must address implementation challenges of Xpert in order to adhere to the WHO
recommendations.
K Clouse,1 M L Lindegren,1 M Yotebieng,2 N Dung,3

A Omondi,4 D M Zannou,5 G Carriquiry,6 A Pettit1
1
Vanderbilt University School of Medicine, Nashville, TN, 2The
Ohio State University College of Public Health, Columbus,
Ohio, USA; 3National Hospital of Tropical Diseases, Hanoi,
Viet Nam; 4Academic Model Providing Access To Healthcare
(AMPATH), Eldoret, Kenya; 5Universite dAbomey-Calavi,
Cotonou, Benin; 6Instituto de Medicina Tropical Alexander
von Humboldt, Ingenieria, Peru. e-mail:
kate.clouse@vanderbilt.edu

recommends Xpert MTB/RIF (Xpert) as the initial TB
diagnostic for HIV-infected individuals but there are
limited data describing real-world implementation of
Xpert. We present Xpert implementation data from a
large, global network of HIV treatment sites.
Methods: We conducted an observational cohort study
among HIV-infected adults enrolled in HIV care/antiretroviral (ART) program sites participating in the International Epidemiologic Databases to Evaluate AIDS
(IeDEA) Consortium and clinically diagnosed with TB
from 2012-2014. Data were collected on HIV-TB coinfected adults (age 718) regardless of ART status using
a standardized electronic form. TB case validation was
performed at sites by file review. We provide proportions
for categorical variables and medians and interquartile
ranges (IQR) for continuous; we use log-binomial models
to estimate adjusted risk ratios (aRR) and 95%
confidence intervals (95%CI) for unfavorable TB treatment outcome (death, default, failure, or undocumented).
Results: Twenty-two participating sites in 19 low/middle
income countries provided data on 2780 patients; we
excluded 283 (10.2%) missing Xpert utilization data,
leaving 2497 participants. A majority was male (60.1%).
At TB diagnosis, median age was 35 years (IQR: 29-42)
and median CD4 count was 115 cells/ll (IQR:38-245).
Of 2497 patients, 123 (4.9%) had a documented Xpert
result; the remainder either did not have Xpert performed (70.2%) or had no recorded Xpert results
(24.9%). Among 123 with Xpert performed, 75
(61.0%) were positive, 48 (39.0%) were negative, and
none were invalid. Rifampicin resistance (RIF-R) was
identified in 15 (20.0%) of Xpert-positive TB cases.
Overall, 1626 (65.1%) TB cases were not bacteriologically confirmed by Xpert, AFB smear, culture or other
nucleic acid amplification test. Adjusted for age, sex and
CD4 at TB diagnosis, cases that had no documentation of
an Xpert test or result were more than twice as likely to
have an unfavorable treatment outcome than cases with
an Xpert test (aRR:2.12, 95%CI: 1.14-3.92).
Conclusions: Use of Xpert for TB diagnosis among HIV
co-infected patients in care in this global network is low.
A majority of TB cases in this analysis had no
documentation of Xpert testing on file, which was
associated with unfavorable treatment outcomes. Future
OA-368-04 Wide variations in compliance with

tuberculosis screening guidelines and
tuberculosis incidence between antiretroviral
therapy clinics, Ivory Coast
A Auld,1 M Blain,1 K Ekra,2 J Kouakou,2 R Shiraishi,1
`
3 T Ellerbrock1 1Centers for
M Tuho,3 V Ettiegne-Traor
e,
Disease Control and Prevention, Atlanta, GA, USA; 2Centers
for Disease Control and Prevention, Abidjan, 3Ministry of
Health, Abidjan, Ivory Coast. e-mail: aauld@cdc.gov
Background: In Ivory Coast, tuberculosis (TB) is a

common cause of death among antiretroviral therapy
(ART) enrollees. In accordance with World Health
Organization (WHO) guidelines, Ivorian guidelines
recommend screening for TB and initiation of treatment
for active TB before ART initiation. Compliance with
these guidelines can reduce TB-related morbidity and
mortality during ART.
Design/Methods: In a retrospective cohort study, 34
from 78 ART facilities with at least 50 ART enrollees
were randomly selected using probability-proportionalto-size sampling; at these facilities 3682 adults (715
years old) starting ART during 20042007 were randomly selected. Controlling for survey design, TB
screening compliance, prevalence of active TB at ART
initiation, and incidence of TB during ART were
evaluated.
Results: At ART initiation, median age was 36 years,
67% were female, and median CD4 count was 135 cells/
ll. Among all 3682 enrollees, 73 (2%) were on TB
treatment at the time of referral to the ART clinic.
Among the 3609 not on TB treatment, 1,263 (36%) were
documented to receive some TB screening before ART
initiation; 21% were screened for cough, 21% for weight
loss, 18% for fever, 18% for TB contacts, and 12% for
night sweats. All five TB screening questions were
administered to only 11% of patients. Among the 1,263

screened, 111 (11%) were diagnosed with TB and started
TB treatment before ART initiation. No associations
between patient characteristics and probability of being
screened were noted. However, TB screening compliance
varied widely by ART facility from 0100%. At ART
initiation, WHO stage IV was more common among
patients taking vs. not taking TB treatment (63% vs.
19%, P , 0.001). TB incidence during ART was 3.04 per
100 person-years but varied widely by ART facility from
0/100 person-year to 13.1/100 person-years.
Conclusion: Compliance with TB screening guidelines
was low. No evidence that clinicians were selectively
screening patients with more advanced disease was
found. Rather, facility-level factors such as training
programs, or patient-to-provider ratios might explain
facility variations in TB screening compliance. Targeting
under-performing sites with improvement activities is
needed. Facility variations in TB incidence warrant
further research; they could reflect facility differences
in TB screening or diagnostic practices, documentation,
or TB risk possibly related to infection control practices
or local community TB incidence.
OA-369-04 Routine implementation of 6-month

isoniazid preventive therapy among HIVinfected patients in seven pilot sites in
Zimbabwe
S143
compared to pre-ART patients. Having 71 documented

default visits from ART in the year prior to commencing
IPT [AOR 5.25, 95%CI (2.10-13.14)] was associated
with five-fold higher risk of non-completion in comparison to timely attendance of review visits in the past year.
Patients who received IPT drugs for 72 months upon
initiation of IPT were 48% less likely not to complete IPT
[AOR0.52, 95%CI (0.28-0.95)] when compared to
those who received one months supply of IPT drugs
upon initiation of IPT.
Conclusion: The completion rate of IPT in Zimbabwe
was high, and therefore expansion of IPT countrywide is
feasible. Patients with a previous history of defaulting
ART and those in pre-ART care have a higher risk of
non-completion and hence require close monitoring to
ensure they complete their treatment. Though dispensing
a months supply of isoniazid prophylaxis in the first
month of treatment is recommended for close patient
monitoring, this was linked with non-completion of
treatment and may require synchronisation with ART
drug pick-up times to achieve better IPT completion
rates.
OA-370-04 The effect of antiretroviral therapy

on chronic lung disease in vertically HIVinfected children and adolescents
K Takarinda,1,2 R Choto,2 A D Harries,1,3 T Apollo,2

C C Musanhu4 1International Union Against Tuberculosis and
Lung Disease, Paris, France; 2Ministry of Health and Child
Care, Harare, Zimbabwe; 3London School of Hygiene &
Tropical Medicine, London, UK; 4World Health Organization
Country Office, Harare, Zimbabwe. e-mail:
ktakarinda@theunion.org
G Mchugh,1 J Rylance,2,3 E Majonga,1,4 J Metcalfe,5

T Bandason,1 H Mujuru,6 K Kranzer,4 R Ferrand1,4
1
Biomedical Research and Training Institute, Harare,
Zimbabwe; 2University of Liverpool, Liverpool, 3Liverpool
School of Tropical Medicine, Liverpool, 4London School of
Hygiene & Tropical Medicine, London, UK; 5University of
California San Francisco, San Francisco, CA, USA; 6University
of Zimbabwe, Harare, Zimbabwe. e-mail:
graceandandre@gmail.com
Background: Isoniazid preventive therapy (IPT) is

recommended for HIV-infected individuals because of
their susceptibility to tuberculosis (TB). As from December 2012, Zimbabwe piloted the roll-out of IPT to HIVinfected patients enrolled at 10 public-sector antiretroviral therapy (ART) clinics. We therefore set out to
determine the extent of non-completion of IPT and the
associated factors in order to inform the scale-up of IPT
countrywide.
Design/Methods: This study was a retrospective cohort
study and data were abstracted from patient records for
all HIV-infected patients enrolled on IPT between 31
December 2012 and 01 March 2013 in 7 of the 10 IPT
pilot sites. Descriptive statistics were calculated using
univariate analysis and factors associated with noncompletion of IPT were established using multivariate
logistic regression.
Results: Of the 578 patients, 466 (81%) completed the 6
month IPT course. Of the 112 patients who failed to
complete IPT, 69 (60%) were lost to follow-up, 30 (27%)
stopped treatment but with undocumented reasons and 8
(7%) developed adverse reactions and /or TB. In
multivariate analysis, patients on ART were 93% less
likely [adjusted odds ratio (AOR) 0.09, 95% confidence interval (CI) (0.03-0.28)] not to complete IPT
Background: HIV infection is a recognised cause of

chronic lung disease. We conducted a cross-sectional
study to investigate whether antiretroviral therapy (ART)
reduces morbidity associated with chronic lung disease.
We report findings comparing older children and
adolescents with HIV receiving ART and a frequency
aged matched group of children who have been newly
diagnosed with HIV
Design/Methods: HIV- infected older children and
adolescents (age 6-16years) on ART for at least 6 months
and newly diagnosed ART- nave HIV-infected children
and adolescents underwent clinical examination including spirometry, shuttle walk testing and pulse oximetry.
Laboratory examinations included CD4 count and HIV
viral load testing.
Results: 200 participants on ART, median age 11 years
(IQR 9-12), 104 (52%) male, mean age of diagnosis was
5 years old were recruited along with 385 ART nave
participants, median age 11 years (IQR 8-13) of whom
186(48%) were male. The median CD4 count in the
treated group was 729 cells/llL (IQR 479-935) compared to 375cells/ ll (IQR 215-599) in the untreated
group (P 0.01).Of those who had results available,
26% of ART nave participants showed either obstruction or restriction on spirometry with poor reversibility.
S144
Similarly, 25% of ART experienced participants showed

either obstruction or restriction, again with poor
reversibility. ART nave n 385 ART experienced n
200, P value Age,Y, median (IQR) 11(8-13) 11(9-12)
0.67 Sex male (%) 48 52 0.37 CD4 Count, cells/ lL,
median (IQR) 375(215-999) 729(479-935) 0.01 History
of pulmonary tuberculosis (%) 5 38 ,0.01 Cough (%)
53 15 ,0.01 Breathless (%) 18 6 ,0.01 Smokers in
Household (%) 25 21 0.32 MRC Dyspnoea score 72
(%) 16 14 0.58 SpO2 ,88%( %) 14 1 ,0.01
Respiratory Rate .30/min (%) 5 3 0.28 75% desaturation post exercise (%) 27 12 ,0.01 FEV1 z score,
mean(sd) 0.73 (1.40) 0.73 (1.20) 0.99 FVC z score,
mean (sd) 0.63(1.35) 0.79(1.24) 0.22 FEF 25%-75%
Z score, mean (sd) 0.19 (1.13) 0.08 (1.20) 0.34
Conclusion: Although ART nave group reported a
higher rate of symptoms, there was no difference in lung
function between the two groups. While ART improves
immune status, it appears to have little impact on
improving lung function in children with longstanding
HIV infection. Interventions to treat and reduce the
significant burden of chronic lung disease are required.
pared to passive case finding and HIV-TB screening

respectively) were younger (median age: 32y vs. 34y vs.
35y; P 0.046) were as likely to be men (58% vs. 60%
vs. 59%; P 0.718), were less likely to be HIV positive
(59% vs. 73% vs. 86%; P , 0.001), but had similar rates
of taking antiretroviral therapy if HIV-positive (67% vs.
68% vs. 65%; P , 0.490). On age- and sex-adjusted
analysis, HIV prevalence remained significantly lower in
adults with TB identified through cwACF (RR: 0.52:
95%CI: 0.36-0.74) and highest in adults identified
through HIV-TB screening (RR: 2.33, 95%CI: 1.872.90) compared to self-presenting patients.
Conclusion: In this large citywide comparison, patient
characteristics differed substantially under three different
TB case-finding strategies. Passive case-finding remained
the dominant mode of detection, but cwACF identified
greater proportions of HIV-negative patients, and
showed additional major indirect benefits in addition
to the direct yield, likely from a lasting impact on healthseeking behaviour. Combined TB case-finding strategies
are likely to have complementary benefits.
OA-371-04 Patient characteristics and

contribution to overall caseload from three
different TB case-finding strategies in Blantyre,
Malawi
10. Using media for communication on

TB
OA-372-04 Mass media communication
campaign, a solution for reaching the
unreached: an experience from six states in
India
P Macpherson,1,2 E Webb,3 A Choko,4 M Nliwasa,3,4,5

A Mdolo,4 J Mpunga,6 L Chiume,4 L Corbett3,4 1University
of Liverpool, Liverpool, 2Liverpool School of Tropical
Medicine, Liverpool, 3London School of Hygiene & Tropical
Medicine, London, UK; 4Malawi-Liverpool-Wellcome Trust
Clinical Research Programme, Blantyre, 5University of
Malawi, Blantyre, 6Ministry of Health of Malawi, Lilongwe,
Malawi. e-mail: p.macpherson@liverpool.ac.uk
S Pandurangan,1 S Mohanty1 1International Union Against

Delhi, India. Fax: (91) 11 46 05 40 30. e-mail:
sripriya14@gmail.com
Background: In high HIV-TB burden settings, slow

progress towards targets indicates that more proactive
TB case-finding strategies are required. Community-wide
interventions may affect subsequent health-seeking
behavior, as well as providing diagnosis.
Design/Methods: Between January 2011 and August
2014, electronic records including mode of detection
(passive case finding, HIV-TB screening, or communitywide active case-finding (cwACF) were collected from all
registering TB patients in 11 facilities in Blantyre,
Malawi. cwACF used 6-monthly door-to-door enquiry
for TB symptoms and smear microscopy in 3 informal
settlements (total adult population 114 450). Multinomial logistic regression was used to compare characteristics of individuals registering for TB treatment by mode
of detection.
Results: Overall, 10 240 adults were registered with TB,
with 9308 (90.9%) identified through passive case
finding, 144 (1.4%) through cwACF and 785 (7.7%)
through HIV-TB screening. Overall smear-positive casenotification rates increased by 38% in the first year for
communities provided with cwACF (230/100 000 in
2010 to 316/100 000 in 2011) and by 11% for noncwACF communities (144/100 000 in 2010 to 160/100
000 in 2011). Adults identified through cwACF (com-
Background: TB remains one of the Indias greatest

public health challenges and accounts about 300 000
deaths per year. Lack of awareness about tuberculosis is
one of the main reasons for huge TB burden in the
country. The Union, through the project Axshya conducted a mass media campaign to improve the awareness
level among the general population in six states of India.
The campaign was conducted through a combination of
mid-media, static media and mass media campaign in the
program states. The first round of campaign was
conducted in March 2012 and the second was in the
month of December 2012. This paper presents the level
of exposure of the IEC campaign in increasing the
knowledge about early diagnosis of Tuberculosis and its
treatment services in different localities. Methodology:
Two rounds of cross sectional community based survey
was conducted adopting multistage systematic random
sampling in 30 districts of 6 states of India. The estimated
sample size was 2426 from 30 districts in the 6 states
considering 5% change in two years. A structured
pretested questionnaire was administered to capture
information on media exposure and awareness.
Results: Nearly 60 % of the general population in urban
areas reported that they had seen or heard any mass
media or mid-media or static media campaign on TB
during the last one year. In rural areas nearly one third
population reported the same. Under mass media, the
exposure of TV Commercial in comparison to radio
commercial and radio jingles is nearly five times higher in
both the areas i.e. rural and urban. Around 50%
respondents reported that they have seen TV commercial
on TB in the last one year whereas in rural areas only
26% reported so. Overall, nearly one third respondents
reported to have watched TV Commercial and only
6.6% reported to have heard radio commercial/jingle on
TB during the reference period.
Conclusion: This study highlights the reach of IEC
campaigns in urban and rural areas. The influence of
mass media campaign is seen more in urban areas than in
rural areas. This study emphasizes the need for designing
suitable campaigns to reach out to rural areas.
Table Exposure to campaign in program areas
Campaign exposure
Total
Urban Rural
Exposed to any campaign activity
42.2
59.9
34.0
Exposed to Mass Media (TVC & Radio)

Exposed to TVC
Exposed to Radio
34.9
33.8
6.6
51.1
50.3
11.4
27.5
26.1
4.4
14.0
9.3
6.7
22.3
16.1
9.9
10.4
6.1
5.3
Exposed to Mid Media

(Street theater/ Mobile van)
Exposed to street theater
Exposed to activation car activity
Exposed to Static Media
(Hoarding/Poster/wall painting)
n
18.1
2551
29.5
807
12.8
1744
OA-373-04 Are TB-related issues still not an

agenda for Indian media? Analysis of TBrelated news content in the last year
S Satapathy,1 S Chadha1 1International Union Against
Delhi, India. e-mail: Sachi.Satapathy@theunion.org
Background: The responsibility of TB control needs to

move from doctors to medical administrators and
politicians and most importantly solutions should be
found by involving all stakeholders from health care
professionals and policy makers to the people and
communities. This is where role of media paramount
importance. The studies have been made from 15April
2014 to 15 April 2015 to find out how national print
media in India have flashed the issue of TB to analyse
what is the present media interest on the issue and how
and what more can be done to address the gap if any.
Design/Methods: The newspapers of one year (Feb 2014Feb 2015) have been scanned to review the nature of
articles and accordingly a classification has been made on
the basis of news from government, civil society, private
practitioner, community, editorial piece, event, TB
patients, drugs and controversies and an attempt has
been made to find the knowledge, interest of media on
this issue.
Results: Government (govt.) related program and launch
of govt. led initiatives have got highest coverage in the
national media (39%) of all the news on tuberculosis in
India followed by corporate partnership issue (28%) and
S145
celebrity involvement/engagement (16%). The news on

new research, drugs, and initiatives has been covered by
3 media houses once in whole last year. The worrying
factor is that issue of tuberculosis-diabetes linkages issue,
which is going to hit the country like India in coming
years has so far not become a media interest in the
country. The local edition of national newspapers have
tried to focus issue of city specific TB cases, how elders in
a particular states have become great contributors to
tuberculosis, prisoners infected on TB, private providers
issues in some point of time. It shows that the main
agenda of the tuberculosis control program as being
strategized in Stop TB partnership has not so far become
an issue for media.
Conclusion: The research on media interest shows that
mainstream media are not so far open up their spaces for
tuberculosis issues as required to reach out to communities. This requires to strategize and make a plan to
engage media houses in an innovative campaign strategy
to achieve the desired objective.
OA-374-04 Building partnerships with

community radio for TB control: a field report
from India
A Srinivasan,1 N Krishnan,1 R Ananthakrishnan1 1REACH
Resource Group for Education and Advocacy for Community
Health, Chennai, India. e-mail:
anupamasrinivasan.reach@gmail.com
Background: Community-owned and operated media

are uniquely positioned to disseminate information
about TB, thereby improving awareness of the disease
among communities. Community radio stations (CRS) in
India have been broadcasting programmes on health
since their inception in the early 2000s. CRS have unique
relationships with their local communities who are active
listeners and often contribute to the process of creating
content.
Intervention: As part of countrywide efforts to involve
civil society in TB control activities through Project
Axshya, CRS have been trained to produce high-quality
radio programming on TB. Their primary mandate is to
address gaps in knowledge on TB and connect people to
local services. All CRS chosen as partners participate in
an intensive workshop and receive a customized toolkit
with talking points about different TB topics as well as
mentoring support through the project period. CRS are
linked with District TB Officers and NGO partners in
their respective locations. In addition, CRS are mandated
to organize regular community meetings that feature live
broadcasts and Q & A sessions with public health
providers.
Results: In five years, 40 stations have produced and
broadcast 550 original episodes on different TB-related
topics. Repeat broadcasts have attempted to enhance
listenership and ensured over 900 hours of programming
on TB on air. 120 community meetings have been held,
providing listeners with an opportunity to interact
directly with local service providers. Through their
interactions with listeners, CRS have focused on two
S146
key messages: that TB is curable and that high quality,

free TB services are available.
Conclusion: While community radio faces considerable
competition from private and commercial radio stations,
it remains one of the few sources of credible health
information in remote, hard-to-reach areas as is evident
from the example of Kerala-based stations serving
primarily tribal populations. However more research is
required to understand the exact nature of CRS
listenership and thereby customize programmes on TB
for local communities. Through this intervention, CRS
have built mutually rewarding relationships with district
and state TB centres, with potential for long-term
partnerships. This intervention has demonstrated the
potential of CRS as a medium for social mobilization and
sustained advocacy on TB.
OA-375-04 Keeping TB alive in the public

domain through sustained media engagement:
a report from India
A Srinivasan,1 R Ananthakrishnan,1 N Krishnan,1
S Prasad2 1REACH-Resource Group for Education and
Advocacy for Community Health, Chennai, 2Lilly MDR-TB
Partnership, New Delhi, India. e-mail:
anupamasrinivasan.reach@gmail.com
Background: Public health issues continue to find limited

space in newspapers in India. Given that newspapers
remain a trusted source of health information in India,
improving the frequency and quality of reporting on TB
is imperative. This can help meet the knowledge gap on
TB and give readers access to accurate information on
symptoms, diagnosis and treatment.
Intervention: The media engagement programme was
initiated in 2010 to improve the quality and frequency of
media reporting on TB, particularly among local
language media. Through annual fellowships programme, 62 mid-career journalists have received fellowships to investigate TB-related issues. All Fellows
participate in an intensive orientation, learning about
the science of TB and identifying locally relevant TB
stories. In addition, Fellows, current and former, receive
resources on TB and have access to key leaders in the TB
community. The best reporting on TB is recognized
through annual awards presented on World TB Day.
Most recently, a pilot initiative has involved former
Fellows as media trainers at journalism schools, thereby
giving students an introduction to reporting on TB.
Results: In all, 62 Fellows have written 257 stories on
different aspects of TB, all published in leading dailies
and magazines. 46% of Fellows have continued to report
on TB beyond the Fellowship period and 36% have
demonstrated interest in improving their capacity to
report on TB. In addition, over 10% of Fellows have
voluntarily expanded their mandate to advocacy, such as
lobbying for inclusion of TB in political manifestos. 23
journalists have received awards for effective reporting
on TB. 92 journalists have received regular resources and
information on TB. 15 resources have been published
including a comprehensive FAQ, an introduction to
ethical reporting on TB and an overview of World TB

Day themes. Overall, the initiative has engaged journalists based in 13 states in India and generated reporting on
TB in at least seven different languages.
Conclusion: This initiative has drawn attention to the
necessity of sustained engagement with the media if TB is
to remain alive in the public domain. The Fellowship
programme has demonstrated the long-term impact of
investing in building the capacity of local language
media. This process of engaging media in a sustained
manner can potentially be replicated in other highburden countries where there is limited reporting on TB.
OA-376-04 TB in the spotlight: media

fellowships as a model for raising public
investment in TB in India
A Buragohain,1 A Jacob,1 P Lal1 1International Union
Against Tuberculosis and Lung Disease, South-East Asia
Office, New Delhi, India. Fax: (91) 11 46 054 430. e-mail:
aburagohain@theunion.org
Background: Indias devastating TB burden over 1.5

million new patients annually is overwhelmingly borne
by the countrys poor and marginalised communities
giving rise to a popular conception of TB as a disease of
the poor. Despite being one of Indias greatest public
health challenges, TB lacks attention in the public
domain. Media reports on TB tend to be infrequent
and reflect news of organisations working in TB control.
Realities of the programme and patients on the ground
are almost completely absent from mainstream media
and consequently the public imagination.
Intervention: A media fellowship was designed to
promote reporting on TB, through which reporters and
researchers were invited to report on contemporary
practices around TB in India and the ways in which it
constituted a disease of the poor. The budget of USD 10
000 for the fellowship was drawn from the WHO DOTS
Expansion TB and Poverty sub-working group secretariat, which was housed at The Union South-East Asia
Office between 2008 and 2014.
Methodology: Through a national call for applications,
applicants were asked to submit a letter of editorial
support and propose at least two story ideas for
investigation. Field visits were stipulated as a necessary
component. After a screening process, nine regional and
national journalists from across India were selected.
They were linked to technical staff at The Union for
access to the research and resource persons to develop the
focus of their stories. Each fellowship included a stipend
of USD 800.
Results: A total of 24 in-depth feature articles/ episodes
were completed by 9 journalists over 9 months.
Chronicling issues of access and quality of TB care in
India, these were published/broadcast in leading media
outlets and gained significant traction. Topics ranged
from Non-clinical barriers to TB treatment (stigma,
psychosocial support, malnutrition); TB as an occupational health hazard (weavers, miners, bidi rollers); to TB
in special zones (refugee colonies, urban slums).
Conclusion: The reports thus produced were unique in

highlighting the broader issues of human development
the structural and socio-political conditions that
influence the condition for TB treatment in India. Several
of the fellows continue to report on TB today, attesting to
returns beyond the duration of the intervention. A focus
on the complex realities of TB care on the ground was
informative both for public health advocates and raising
public consciousness on TB.
OA-377-04 Promoting evidence-based

reporting on lung health diseases and
enhancing visibility in news media
R Dwivedi1 1Citizen News Service - CNS, Lucknow, India.
Fax: (91) 522 235 8230. e-mail: rahul@citizen-news.org
Background and challenges to implementation: As

Citizen News Service (CNS) specializes in evidence based
health reporting, it strides to identify structural barriers
by engaging with affected populations; and listen, learn
and document their voices along with national and
international health experts on a range of lung health
diseases; and getting news stories published through CNS
syndication in order to enhance more visibility on lung
health issues.
Intervention or response: CNS facilitates a Health
Writers Network that connects health writers/ reporters,
particularly from the developing nations, and hosts
online dialogue and debate on a range of cross-cutting
priority health issues for building the understanding and
sensitivity of CNS Health writers around lung health
issues. With support from the Union, CNS develops issue
brief on specific lung health issue each month to share
with health writers and help them in developing evidence
based news stories with inputs from issue brief, health
experts and affected people. The news stories get
published in newspapers, websites and widely disseminated through social media with specific hashtag
(#Lunghealth). Apart from this CNS also organizes
Media Dialogues on specific issues engaging journalists
of English and vernacular languages at local level with a
view to increase sensitive reporting on lung health
diseases.
Results and lessons learnt: In the last two year CNS has
circulated around 36 issue briefs on specific lung health
diseases, among health writers which have generated
more the 500 news stories from various Asian and
African countries mainly. CNS has also organised more
than 40 media dialogues on the health issues, which have
generated more than 400 news items in print and online
media in vernacular and English languages at local level.
Apart from these news stories CNS also documents Best
Practices in Programmatic Management of Drug Resistance Tuberculosis (PMDT) from 14 sites in India,
Voices of unheard Report on Childhood TB and Voices
from the Field of Childhood Pneumonia
Conclusions and key recommendations: CNS strongly
believes that it is critically important to ensure evidence
based health reporting and engagement with affected
S147
population to identify structural drivers and specific

linkages between health and development issues.
OA-378-04 Use of commercial marketing and

mass media tools for TB health promotion and
education
M Villapando,1 E B Generoso,1 C Cotingting1
Communication for Health Advancement through
Networking and Governance Enhancement (CHANGE)
Project, Makati, Philippines. e-mail:
crystle.cotingting.change@grey.com
1
Background: Among the barriers associated with poor

utilization of TB diagnosis and treatment services in the
Philippines include misconceptions and social stigma
surrounding TB, the unpopularity of visiting public
health facilities and the tendency of Filipinos to selfmedicate. Past TB communication strategies relied more
heavily on interpersonal communication at the local
health center, overlooking mass media.
Intervention: Under the USAID/Philippines-supported
CHANGE project, two television campaigns were
developed using a commercial marketing perspective.
The target audiences of the campaigns were mostly from
lower socioeconomic classes (Class D, E). The first
campaign, a brand ambassador campaign, employed a
likeable, credible entertainment personality to deliver the
message that TB is curable. The message was set to
upbeat music and comedic tones, in contrast to previous,
more somber campaigns. The second TV ad employed
the classic side-by-side / split-screen styles (i.e. Brand X
vs. Y campaigns) to contrast successful recovery with the
Directly Observed Treatment-Short Course against the
deteriorating condition of one resorting to ineffective
remedies. The ad sought to encourage seeking consultation when suspecting TB. Both ads underwent a series of
pre-testing for message comprehension and acceptability
before airing. The campaigns were aired in two phases,
averaging 20 spots per week over a 15-week period for
the first phase and for 14 weeks, in a flighting
scheduling pattern that created the semblance of
continuous airing, for phase two. An external research
firm was employed to conduct a nationwide evaluation
of the campaign with a survey of 2000 respondents.
Results: At the end of the campaign, a majority (54%) of
those who had experienced at least one TB symptom now
sought consultation, an increase from the baseline of
45%. A significant decrease from 42% to 34% in
agreement to the statement, TB is very contagious, so
we should keep away from people with TB indicated
softening of attitudes towards people with TB, aiding to
the objective of alleviating stigma. Ninety-six percent of
people reached were now aware that TB is curable vs.
91% prior to the campaign. Of those interviewed, 70%
were able to correctly identify the ads key messages.
Conclusions: The results indicate that both ads had
successfully communicated the key messages, demonstrating that mass media is a viable vehicle to encourage
better health seeking behavior and practices.
S148
Poster discussion sessions, Friday, 4 December
POSTER DISCUSSION SESSIONS

01. Extra-pulmonary tuberculosis:
detection and treatment
PC-700-04 Potts disease: epidemiological
aspects, diagnosis and prognosis in
Ouagadougou
G M E Badoum Ouedraogo,1 A Kabre,1 K Boncoungou,1
G Ouedraogo,1 A R Ouedraogo,1 S Zabsonre,1 M
Ouedraogo1 1Centre Hospitalier Universitaire Yalgado
Ouedraogo,
Burkina Faso. e-mail: gisebad@yahoo.fr
Background: Tuberculosis spondylitis or Potts disease

corresponds to the appearance of an infectious tubercular process on one or several disco-vertebral joints. The
purpose of this study was to specify the epidemiological,
diagnosis, therapeutic and prognosis aspects of Potts
disease at the University Teaching Hospital Yalgado
Ouedraogo of Ouagadougou.
Methodology: This study was retrospective in nature. It
involved 103 patients and was carried out over a period
of 6 years in the Department of Neurosurgery.
Results: The average age of participants was 46.2 years.
Men were more affected, the sex ratio was 1.6. Most
patients (73%) were from disadvantaged social and
professional groups. Spinal pain was the most common
reason for consultation in 93 cases (90.3%). An average
delayed patient consultation period of 10 months was
noted. The lumbar spine was most affected in 54.4% of
cases. The results of the tuberculin skin test were positive
in 91.35% of cases. Even though no culture was made, a
sputum examination for acid-fast bacilli was positive in
4.8% of cases. The treatment consisted in a 9-month
antituberculosis chemotherapy associated or not to an
orthopedic (12 cases) and/or surgical treatment (1 case).
The evolution was good in 61% of cases and sequelae
were observed in 37% of cases. In 1.9% of cases,
evolution resulted in the death of the patient.
Conclusion: Potts disease continues to be a public health
problem because it affects a young working population.
Antituberculosis chemotherapy is the primary method
and often a sufficient one in treating the disease.
However, surgery is still an option especially in a context
where complications are frequent at the diagnosis stage.
PC-701-04 Detection of multidrug-resistant

tuberculosis in extra-pulmonary specimens by
line probe assay: experience from a National
Reference Laboratory
R Singhal,1 J Anand,1 M P Bhalla,1 G Kumar,1 A K Verma,1
J Arora,1 V Myneedu,1 R Sarin1 1National Institute of
Tuberculosis and Respiratory Diseases, New Delhi, India.
e-mail: drritugo@gmail.com
Background: Extra-pulmonary tuberculosis (EPTB) accounts for 10-15% of all cases. GenoType MTBDRplus;
Line probe assay (LPA) is used as primary test for the
detection of multi-drug resistant tuberculosis (MDR-TB)
in pulmonary cases by National Programme in India. As
no experience is available on direct testing of EP samples

using MTBDRplus, present retrospective study was
conducted to determine its utility in these pauci-bacillary
samples.
Materials and methods: The present retrospective study
was conducted in Department of Microbiology, National
Reference Laboratory under Revised National TB
Control Programme. 644 EP samples from presumptive
MDR-EPTB patients were received from 1st January
2014 to 31st August 2014. The samples were subjected
to smear microscopy and processing as per RNTCP
guidelines. Processed smear positive specimens were
subjected to LPA and smear negatives were inoculated
in MGIT 960 culture. Positive tubes identified as M.
tuberculosis were subjected to LPA.
Results: Majority of patients (407/ 644; 63.2%) were in
15-35 years followed by 124/644; 19.3% in 36-60.
Pediatric and geriatric patients were 90/645; 14.0% and
23/645; 3.6% respectively. Male: female ratio was 1:1.
Commonest sample was fine needle aspiration (FNA) of
lymph nodes (349/ 644; 54.2%), followed by pleural
fluid (255/644; 39.6%), majority being exudative (157/
255; 61.6%). Other samples included biopsies (12/645;
1.9 %), ascitic fluid (10/644; 1.6%), CSF (8/644; 1.2%),
synovial tissue (4/644; 0.6%), urine (2/644; 0.3%), bone
tissue (2/644; 0.3%) and one each of abscess (1/664;
0.15%) and endometrial aspirate (1/664; 0.15%). Total
of 185; 28.7% samples were smear positive, of which
167; 90.3% gave valid results directly from sample in an
average turn-around time of 3.3 days. 35/185 (20%)
cultures were positive for mycobacteria of which 1 was
non-tuberculous mycobacteria and remaining were M.
tuberculosis. Total of 32/220 (15.8%), 10(5%) and
15(6.8%) strains were MDR, rifampicin mono-resistant
and INH mono-resistant respectively. Commonest mutation in rifampicin and isoniazid were missing wild type
with specific mutation S531L (33; 76.7%) in rpoB and
missing wild type with specific mutation in S315T1 in
katG 26; 57.8% respectively.
Conclusion: Genotype MTBDRplus is a useful rapid test
for detection of MDR-TB in smear positive EP-TB
samples. It is essential to perform structured studies to
generate validation data of test against reference
standard methodology.
PC-702-04 Predictors of success factors of extrapulmonary tuberculosis treatment using the

DOTS strategy in Dr. Cipto Mangunkusumo
Hospital, Jakarta
T Kamelia1 1Pulmonology Division, Dr. Cipto
Mangunkusumo Hospital, Jakarta, Indonesia. Fax: (62) 878
8159 7886. e-mail: tellybahar@gmail.com
Background: Extrapulmonary tuberculosis (EPTB) is a

common presentation found in Indonesia, besides
Tuberculosis (TB). We found that there were only few
researches on EPTB in Indonesia, especially regarding the
success of EPTB treatment using the DOTS strategy and
its predictor factors. We aim to determine predictors of
TB treatment success factors such as age, sex, diabetes
mellitus, HIV and anti-tuberculosis records. We also

want to determine the success rate of EPTB treatment
using DOTS strategy which is administered for a
minimum of 9 months in our institution.
Design/Methods: A retrospective cohort study was
conducted on medical records from January 1st, 2008
to December 31th, 2012. Medical records were taken
from Dr. Cipto Mangunkusumo Hospital, one of the top
national referral hospitals in Indonesia. A total of 542
patients of EPTB were identified, 193 patients were pure
EPTB while 279 had mix of pulmonary TB and EPTB.
Seventy records were not eligible due to incomplete data.
Results: The majority of patients were female (52.3%)
and age 18-60 years old, mean 31,34 11,64 years old,
(95.9%). The most common site of EPTB were lymph
(56.9%). There are 61, 66% of EPTB patients who
received treatment for ,9 months. In bivariate analysis,
successful treatment of EPTB among age of 18-60 years
rated 49.7% (OR 2.97, 95%CI 0.59 to 15.09, P
0.313). In patients treated for 79 months, success rate
was 94.6%. Successful treatment of EPTB among female
rated 55.4% (OR 1.77, 95%CI 1.00 to 3.13, P 0.05).
Successful treatment using DOTS strategy among patients with diabetes mellitus rated 33.3% (OR 1.96,
95%CI 0.48-8.06, P 0.546) and the rate in patients that
had past antituberculosis medication was 55,6% (OR
0.74, 95%CI 0.28-1.96, P 0.717). Successful treatment
in EPTB patients among HIV-seropositive rated 32,1%
(OR 2.59, 95%CI 1.33-5.04, P 0.007). In multivariate
analysis, EPTB with HIV co-infection factor had OR
2.59, CI 95% 1.33-5.04, P 0.005, for the success rate.
EPTB among HIV-seropositive patients had lower
therapy success rate using DOTS strategy and were
associated with unsuccessful therapy and poor prognosis.
Conclusion: HIV was a predictor factor that may
decrease therapy success rate of EPTB using DOTS
strategy. Success rate of extrapulmonary TB treatment
using DOTS strategy for minimum 9 months was good
(94.6%) in our institution.
S149
epidemiology of EPTB by means of national tuberculosis

surveillance data in golestan province. This review is
limited to reported public health surveillance system data
and is not a clinical case review. Our goal is to describe
EPTB epidemiology, compare the incidence and characteristics of EPTB with those of PTB, and suggest possible
factors that are contributing to the increased proportion
of EPTB-attributable tuberculosis cases.
Intervention or response: National tuberculosis cases
reported from 2005 to 2012 were classified as either
EPTB or PTB. EPTB encompassed lymphatic, pleural,
bone and/or joint, genitourinary, meningeal, peritoneal,
and unclassified EPTB cases. We excluded cases with
concurrent extrapulmonary-pulmonary tuberculosis and
cases of disseminated (miliary) tuberculosis. Demographic characteristics, drug susceptibility test results, and risk
factors, including human immunodeficiency virus (HIV)
status, were compared for EPTB and PTB cases. We
analyzed tuberculosis cases reported from January 1,
2005, through December 31, 2012. Cases were categorized by major disease site, reported as either PTB or
EPTB. The PTB group comprised cases with PTB listed as
the only disease site
Results and lessons learnt: Among 6879 cases, 79.55%
were PTB and 20.45% were EPTB, including pleural
(26.3%), lymphatic (24.8%), bone and/or joint (13%),
genitourinary (6.7%), meningeal (4.6%), peritoneal
(4.2%), and other types of EPTB (20.4%) cases.
Compared with PTB, EPTB was associated with female
sex (odds ratio [OR], 1.44; 95% confidence interval [CI],
1.2-1.6) and foreign birth (OR, 0.68; CI, 0.5-1), almost
equally associated with HIV status (OR, 1.1; CI, 1.11.1). Slower annual deincreases in EPTB case counts,
compared with annual deincreases in PTB case counts,
from 2005 through 2012 have caused EPTB to increase
from 24.03% of tuberculosis cases in 2005 to 19.78% in
2012.
Conclusions and key recommendations: EPTB epidemiology and risk factors differ from those of PTB, and the
proportion of EPTB has increased from 2005 through
2012. Further study is needed to identify causes of the
proportional increase in EPTB.
PC-704-04 Extra-pulmonary tuberculosis: case

notification trend and treatment outcomes in
Southern Nigeria, 20092013
PC-703-04 Epidemiology of extra-pulmonary
tuberculosis in Golestan Province, Iran, 20052012
S Rafiee,1,2 H R Kamalinia2 1Golestan University of Medical
Sciences, Gorgan, Iran. e-mail: soheil_rafiee@yahoo.com
Background and challenges to implementation: Almost

one-fifth of golestan tuberculosis cases ( 1:4.9 ) are
extrapulmonary; unexplained slower annual case count
increases have occurred in extrapulmonary tuberculosis
(EPTB), compared with annual case count increases in
pulmonary tuberculosis (PTB) cases. We describe the
C Ogbudebe,1 J Chukwu,1 C Nwafor,1 A Meka,1

N Ekeke,1 N Madichie,1 D Oshi,1 J Ikebudu1 1German
Leprosy and TB Relief Association, Enugu, Nigeria. e-mail:
dadasky22@yahoo.com
Background: According to the WHO, Nigeria currently

ranks third among the 22 TB burden countries globally
and first in Africa.1 This is sequel to a 2012 national TB
prevalence survey which returned incidence and prevalence rates of 338 and 322 cases per 100 000 respectively.
With a projected population of 170 million, this means
that the country can be saddled with over half a million
new cases of TB annually. The situation is further
complicated by a generalised HIV epidemic currently
S150
running at a median prevalence of 3.4%.2 It is estimated

that extra pulmonary (EP) TB accounts for 15%-36%of
all cases of TB.3,4 Studies in some sub-saharan African
countries have reported rates ranging from 25%-44%
among childhood EPTB cases.5 Objective: To determine
the proportion and trend of EPTB among notified TB
cases as well as the treatment outcomes in the national
TB programme (NTP).
Methods: A retrospective cohort study of EPTB patients
enrolled under the NTP from 1 January 2009 to 31
December 2013 was conducted in 14 states in Southern
Nigeria. Data were extracted from TB registers. National
census figures were used as denominators to calculate
rates. Outcomes were categorised as favourable (cure
and treatment completed) and unfavourable (died, loss to
follow-up, failure or transfer out).
Results: Of the 126 806 TB patients (all forms) registered
during the 5-year period, 10 395 (8.2%) had EPTB. This
translates to an average annual notification rate of 4.1
per 100 000 population (IQR 3.25.8). The median age
was 35 years (IQR 2754), female were 6,341 (61%) and
children (,15 years) were 1,247 (12%). HIV test was
documented for 8,960 (86.2%) patients, 2,331 (26%)
were HIV positive. CPT and ART uptake were 94.7%
and 51.9% respectively. Overall, there was significantly
more children ,15years (P 0.002), females (P 0.03)
and HIV positive patients (P 0.01) with EPTB
compared with PTB patients. Average treatment success
rate was 74%, while loss to follow-up rate was 16%
compared to treatment success rate of 84% and loss to
follow-up rate of 9% among PTB patients. HIV co
infection was associated with unfavourable treatment
outcomes (RR 1.4, CI 1.11.9, P , 0.02). Among them,
those who received ART had better outcomes (P 0.04).
Conclusion: Although overall proportion of EPTB cases
shows a rising trend, especially among children, there is
need for the NTP to pay greater attention to diagnosis
and case-holding of patients with EPTB.
PC-705-04 Epidemiology of urogenital

tuberculosis in Siberia
E Kulchavenya1 1Novosibirsk Research TB Institute, Medical
University, Novosibirsk, Russian Federation. e-mail:
ku_ekaterina@mail.ru
Background and challenges to implementation: Urogenital tuberculosis (UGTB) is the second most common
form of TB in countries with severe epidemic situation
and the third most common form in regions with low
incidence of TB.
Intervention or response: Estimates of incidence and
spectrum of extrapulmonary tuberculosis (EPTB) in
Siberia have been made on the basis of the data available
in the official reporting forms. Also 131 history cases of
UGTB patients, who were revealed in Novosibirsk (large
industrial center of Siberia) in 2009 2011 years, were
analyzed retrospectively.
Results and lessons learnt: In 1999 the incidence rate of
TB as whole in Siberia was 116 per 100 000 inhabitants;
in 2011 it increased up to 124. Every year in Siberia there
are about 1000 new revealed EPTB patients. Within the

last decade the spectrum of EPTB has changed significantly. TB of the central nervous system almost doubled
from 4.9% to 8.7%. Bone and joints TB increased from
20.3% to 34.5. The proportion of UGTB decreased from
42.9% to 31.7%, as well as peripheral lymph nodes TB
from 16.7% in 1999 to 11.2% in 2011. Among 131
patients with UGTB in 88 (67.2%) the isolated kidney
TB was diagnosed, in 33 (25.2%) male genital TB; 10
(7.6%) men had generalized UGTB. We found the
change in gender distribution from male : female of 1:2
in 1999 to 1:1.2 in 2011. Among all cohort of UGTB
asymptomatic course was in 12.2%, among kidney TB in 15.9%. Every third patient complained of flank pain
and dysuria (accordingly 35.2% and 39.8%), 17%
presented toxicity symptoms, 9.1% - renal colic, 7.9%
- gross-hematuria. MBT in urine was found in 31.8% in
isolated kidney TB as whole. The onset of TB orchiepidydydmitis was in 35.7% of patients, hemospermia
was in 7.1%, dysuria - in 35.7%. Most common
complaints for prostate TB were perineal pain (31.6%),
dysuria (also 31.6%), hemospermia (26.3%). MBT in
prostate secretion / ejaculate was revealed in this group in
10.5%.
Conclusions and key recommendations: UGTB remains
an important problem. All urogenital tract infections
should be suspected on UGTB in patients living in the
region with high incidence rate, who had contact with TB
infection, who has recurrent course of the disease,
resistant to standard therapy.
PC-706-04 Optimization of therapy for prostate

tuberculosis
E Kulchavenya,1 A Osadchiy1 1Novosibirsk Research TB
Institute, Medical University, Novosibirsk, Russian Federation.
e-mail: ku_ekaterina@mail.ru
Background: In its 2012 global report on tuberculosis,

WHO estimates that 3-7% (range 2.1-5.2%) of new
cases and 20% (range 1326%) of previously treated
cases have multidrug-resistant (MDR) tuberculosis (defined as tuberculosis caused by M. tuberculosis isolates
that are resistant to rifampicin and isoniazid). In many
countries in Eastern Europe and central Asia, 932% of
new patients and more than 50% of previously treated
patients have MDR-TB. One of the reasons for treatment
failure is non-optimal therapy. The therapy of prostate
tuberculosis (PTB) is very difficult as a few antibacterial
agents are able to distribute to the prostatic tissue and
achieve sufficient concentrations at the site of infection.
Design/Methods: Efficiency of optimized scheme for the
therapy of prostate tuberculosis (PTB) was estimated in
53 pts in the age of 24 - 69 (on average 32.167.2).
Infiltrative PTB was diagnosed in 24, cavernous - in 29.
M. tuberculosis was found in 21 pts (39.6%): in 6 by
culture and in 15 by PCR; histological verification was
in 8 (15.1%), in 29 patients (54.7%) diagnosis was
established by X-ray. First group (25 pts) was treated
with a standard scheme of chemotherapy: in intensive
phase 4 anti-TB drugs daily (isoniazid pyrazinamid
rifampicin per os and streptomycin i.m.). In phase

continuation the therapy includes only isoniazid and
rifampicin in intermittent regime (3 times a week).
Second group (28 pts) alongside with standard chemotherapy received ofloxacin during intensive phase, and
isoniazid was given as intravenous infusion. The phase
continuation was in both groups identical. The efficiency
was estimated in 2 and 8 months. The criteria of the
efficiency were: resolution of pain, dysuria, pyospermia,
negative results of bacteriology.
Results: Patients in the first group had faster positive
dynamic concerning pain, dysuria, pyospermia and
negative growths on M. tuberculosis as well as PCR in
ejaculate; in 8 months best results of optimized therapy
were as well. In 1st group 77.8% of pts had good results;
moderate results had 6 patients (22.2%), while in the
control group, who received standard therapy good
results showed 11 patients (44.0%) only, moderate
results 13 patients (52.0%), one patient (4.0%) had
poor result after completion full course of standard
therapy.
Conclusion: Optimization of the standard therapy of
prostate TB by additional administration of ofloxacin
improved treatment results in 33.8%: patients in the
group with optimized therapy.
PC-707-04 Comparison of outcomes of extrapulmonary and pulmonary tuberculosis among

HIV patients in an HIV clinic in Uganda
D Nalwanga,1 S Okware,1 P Kizito,1 A Von Braun,1,2
C Wiltshire,1 B Castelnuovo1 1Infectious Diseases Institute,
Kampala, Uganda; 2University of Zurich, Zurich, Switzerland.
e-mail: dnalwanga@idi.co.ug
Background: The prevalence of extra-pulmonary tuberculosis (EPTB) among patients with human immunodeficiency virus (HIV) is increasing while that of pulmonary
tuberculosis (PTB) is declining. Patients with EPTB are
often very sick and in resource limited settings diagnosis
is difficult, therefore treatment is presumptive. We
hypothesize that this may lead to poor treatment
outcomes in those with EPTB compared to those with
PTB. We set out to compare treatment outcomes of
patients treated for EPTB and PTB in an urban HIV
outpatient clinic.
Design/Methods: The study was carried out at the
Infectious Diseases Institute. All HIV positive patients
treated for tuberculosis (TB) between 1st January 2012
and 31st December 2014 who had a documented World
Health Organization (WHO) TB treatment outcome
were included in the analysis. Patients clinical and
demographic data were extracted from an electronic
database. Comparisons were made between baseline
characteristics and treatment outcomes (death, favorable
outcome which was defined as patients who were cured
or completed treatment as defined by WHO, and
unfavorable outcomes including transfer out, loss to
follow up and treatment failure) of patients with PTB
and EPTB . v2 test was used to measure the association
between TB type and each of the variables.
S151
Results: Of 417 patients with TB, 164(39%) had EPTB

with the majority having abdominal TB (60%) and TB
adenitis (20%).The patient gender, age and anti-retroviral treatment (ART) status were similar at TB diagnosis.
Patients with EPTB had significantly lower baseline
CD4, median (IQR) 130 (55-266) compared to PTB
193(58-343) P 0.03. There was no significant
difference in treatment outcomes between patients with
EPTB and PTB (77% vs. 70% for favorable outcomes,
11% vs. 12% for death, 12% vs. 18% for unfavorable; P
0.22). More patients with EPTB gained weight and had
a rise in CD4 count following TB treatment (P 0.007
and P 0.002 respectively). The median CD4 count gain
was higher in those with EPTB (99.9, IQR: 29-177)
compared to those with PTB (60, IQR: 11-135).
Conclusion: TB treatment outcomes in HIV patients with
EPTB and PTB are similar. Presumptive treatment of
EPTB is therefore justified in these patients to reduce
delays associated with confirming a diagnosis of EPTB.
02. Drug development: something old,

something new
PC-708-04 Laboratory and non-clinical
evaluation of pulmonary drug delivery in TB
S Pandya,1,2 A Gupta,1 R Ranjan,1,2 M Sachan,1
A Srivastava,1,2 A Misra1 1CSIR-Central Drug Research
Institute, Lucknow, 2Academy of Scientific & Innovative
Research (AcSIR), New Delhi, India. e-mail:
amit.cdri@gmail.com
Background: Particles capable of targeting lung and

airway macrophages when administered as a dry powder
inhalatin (DPI) have successfully completed pre-clinical
development. We demonstrate that micro-doses of antiTB drugs delivered in inhalable particles are safe,
efficacious and capable of recruiting innate, potentially
bactericidal macrophage responses in pre-clinical models.
Design/Methods: Studies were designed in accordance
with the Indian Drugs and Cosmetics Act. Particles were
prepared by spray-drying and subjected to Pharmaceutical Quality/CMC (Chemistry, Manufacturing and
Controls). Animal studies were conducted with permission from the Institutional Animal Ethics Committee.
Pharmacokinetics and biodistribution, especially to lung
tissue and lung/airway macrophages was established in
rodents and non-human primates at three logarithmic
dose levels. Acute, sub-acute and 90-day repeated dose
toxicity studies were carried out in rodents and
macaques. Efficacy was investigated in mice and guinea
pigs infected with M. tuberculosis H37Rv through a lowdose aerosol.
Results: Particles containing 50% by weight of equal
amounts of isoniazid and rifabutin were amenable to
deep lung delivery as a DPI. Inhaled particles were avidly
taken up by airway and lung macrophages and established high drug concentrations in the lungs. They also
evoked nitric oxide, oxidative free radicals, T-helper 1
S152
cytokines and pro-apoptotic gene expression. Infected

macrophages underwent apoptosis and displayed signs of
intensive Golgi activity suggestive of microautophagy
upon taking up the particles. Once-daily inhalations
elicited no observable toxicity in macaques at doses up to
100 mg/day. Mice/guinea pigs treated for four weeks
with once-a-day inhalations completely cleared bactreia
from their lungs, but not spleens. Combining inhalations
of 100 lg/day with oral administration of 5 mg/Kg of
isoniazid and rifabutin each for four weeks cleared the
lungs and spleens, and did not allow recrudescence of
infection up till four weeks after cessation of treatment.
Conclusion: Radical cure of pulmonary tuberculosis can
be achieved within four weeks if inhaled particles are
combined with half the standard oral dose of first-line
anti-TB drugs. Data for an Investigational New Drug
Application for clinical studies on safety and efficacy of a
formulation that has potential to significantly reduce
duration of drug-sensitive TB is now in place and a
clinical testing plan needs to be evolved.
the Fenton reaction, sterilizes cultures of drug-susceptible and drug-resistant Mycobacterium tuberculosis.
Design/Methods: Non Randomized controlled trial
study design was used. New Sputum Positive (NSP) TB
patients registered during June 2013 to December 2014
from two Designated Microscopy Centres supported by
Damien foundation India Trust were included. NSP TB
patients from Centre-A were allocated to study group
while those from Centre-B to control group. Study group
received two grams of Vitamin C daily along with ATT
and the controls had only ATT. Sputum microscopy for
Acid Fast Bacillus (AFB) was done at the time of
diagnosis, end of intensive phase, end of 2 months of
continuous phase and end of treatment
Results: Study group had 43 NSP TB patients and control
group had 56. Sputum AFB level of 2 and 3 was
observed in 53.4% of study group and 51.7% of control
group. Sputum conversion at 60 days was observed in
82.6% in study group with AFB level of 2 and 3 while
it was 66.6% in control group. The overall sputum
conversion at 60 days in study group was 88.4% and it
was 75.9% in control (P 0.059, Fisher exact 0.089).
Conclusion: There is indication that Vitamin C has
beneficial effect when administered with ATT. It seems to
augment sputum conversion and hence bacterial killing.
However a detailed study is required with adequate
sample size and confounding factors need to be
considered.
PC-710-04 Identification of optimal dose and

dosing regimen of clofazimine for the
treatment of MDR-TB based on
pharmacokinentic modeling
G Subramanian,1 G Sunkara,2 D Mcneeley,2 D Hughes3
1
Novartis, Hyderabad, India; 2Novartis, East Hanover, NJ,
USA; 3Novartis, Basel, Switzerland. e-mail:
subramanian.ganesan@novartis.com
PC-709-04 Can vitamin C augment sputum

conversion in TB patients while on antituberculosis treatment?
L Gujral,1 S Mugudalabetta,1
R K Ramalingapuram Muthusloahy1 1Damien Foundation
India Trust, New Delhi, India. e-mail: lgujral@gmail.com
Background: TB control programmes have been successfully implemented with short term treatment regimens.
The duration of treatment is still long. Treatment
adherence remains as a challenge. This study aims at
exploring the efficacy of Vitamin C in augmenting
sputum conversion in TB patients while on Anti
Tuberculosis Treatment (ATT). A common mechanism
of cell death by bactericidal antibiotics involves the
generation of highly reactive hydroxyl radicals via the
Fenton reaction. Vitamin C, a compound known to drive
Background: Currently, clofazimine is classified as a

WHO Group 5 (unproven efficacy), second-line drug
recommended for the treatment of MDR-TB. Clofazimine is absorbed relatively slowly and it is strongly
lipophilic and accumulates mainly in fatty tissue and in
macrophages of the reticuloendothelial system. Steadystate concentrations can take around 110 days to be
reached. Therefore, in order to reach early steady-state a
loading dose is essential to reach the target plasma
concentrations in humans during the initial months of TB
treatment. Clofazimine has not demonstrated early
bactericidal activity in a recent phase 2a study. The
objective of this research work was to determine the
optimal dose and dosing regimen of clofazimine based on
its pharmacokinetic properties to achieve the target
concentration expected to achieve anti-mycobacterial
activity.
Methods: A population pharmacokinetic model for
clofazimine was developed by pooling the data from
two clinical studies of clofazimine after single and
multiple dosing in healthy (200 mg dose once a day)
and leprosy patients (50 or 100 mg once a day),
respectively. Plasma concentration data from these two

studies were analyzed by non-linear mixed effect
modelling with first order conditional estimation method
with interaction. The population estimates in combination with in vitro MIC99 data (1.18 lg/ml) in drug
resistant clinical MDR-TB isolates and in vivo data from
mouse efficacy model, several simulations were performed to identify the dose and dose regimen that reach
target concentration of 1.18 lg/ml in earliest treatment
duration (~ 6 weeks) and sustain it for 18 weeks of the
intensive phase of TB treatment.
Results: The clofazimine plasma concentration time data
were best described by a two-compartment model with
first order absorption. Inter-individual variability component was included in all parameters. TSeveral clinical
scenarios with different dose and dosing durations were
simulated during the initial assessment. The time to reach
the target concentration of 1.18 lg/ml and the duration
for which the plasma concentrations are the selection of
dose and dosing regimen. The loading dose of 200 mg
once daily administration of clofazimine for 18 weeks
followed by once daily administration of 100 mg of
clofazimine till 24 weeks can provide target plasma
concentration in approximately 6 weeks. This dose is
planned in the clinical study planned in pivotal study.
PC-711-04 A comparison of the effectiveness of

a six- vs. eight-month anti-tuberculosis
regimen for pulmonary tuberculosis under
programme conditions
K Ukwaja,1 S Oshi,2 I Alobu,3 D Oshi2 1Department of
Internal Medicine, Federal Teaching Hospital, Abakaliki,
Ebonyi State, 2Centre for Development and Reproductive
Health, Enugu, Enugu State, 3National Tuberculosis and
Leprosy Control Programme, Ministry of Health, Abakaliki,
Ebonyi State, Nigeria. e-mail: ukwajakingsley@yahoo.co.uk
Background: The WHO recently changed the guidelines

of first-line treatment of tuberculosis (TB) from an 8month regimen containing rifampicin for two the first
two months only (2RHZE/6EH) to a 6 -month rifampicin containing regimen (2RHZE/4RH). However, the
impact of the shorter anti-tuberculosis regimen (2RHZE/
4RH) on treatment outcomes in Nigeria has not been
assessed. Local evidence conducted under routine conditions in high TB burden settings to support the current
WHO-recommended changes in anti-tuberculosis treatment policy is crucial but lacking. The aim of this study
was to compare the epidemiological characteristics and
treatment outcomes of TB patients treated with
(2RHZE/6EH) and those who received the 6-month
regimen (2RHZE/4RH) under programme conditions.
Design/Methods: This was a retrospective cohort study
of newly diagnosed, smear-positive TB patients treated in
two large hospitals in Ebonyi State, Southeastern Nigeria, between 1 January 2011 and 31 December 2012. The
diagnosis and treatment of TB in Nigeria uses standard
WHO/NTP methods. Standard treatment outcome categories according to the WHO/NTP guidelines (cured,
treatment completed, died, treatment failure, loss to
follow-up [default] and transferred out) were used
S153
Results: A total of 928 newly-diagnosed, smear-positive

TB patients were treated with daily ethambutol, isoniazid, rifampicin, and pyrazinamide for two months
followed by ethambutol and isoniazid for six months
(2RHZE/6EH), or the same intensive phase as the first
regimen followed by four months of daily rifampicin and
isoniazid (2RHZE/4RH). The proportion of successful
outcomes was 381/490 (77.8%) with 2RHZE/6EH and
373/438 (85.2%) with 2RHZE/4RH (P 0.004). The
proportion of loss-to-follow-up was significantly higher
in patients who received 2RHZE/6EH (14.3% vs. 5.5%;
P , 0.001). Treatment failure was not significantly
higher in patients who received 2RHZE/6EH (2.9% vs.
1.6%; P 0.15). After adjusting for confounders, older
age (adjusted odds ratio (aOR) 1.7), 2RHZE/6EH
treatment (aOR 1.6), and male gender (aOR 1.5)
independently predicted unsuccessful outcomes in HIVnegative TB patients.
Conclusion: Newly-diagnosed TB patients on 2RHZE/
4RH have a higher treatment success rate compared to
those treated with 2RHZE/6EH under programme
conditions in a low-resource, high-burden setting.
Current WHO recommendations should be maintained.
PC-712-04 Therapeutic drug monitoring of

isoniazid and rifampicin
L Moreno Exebio,1 L Lecca,2 G R Davies3 1National
Institutes of Health, Peru, Lima, 2Partners in Health, Peru,
Lima, Peru; 3University of Liverpool, Liverpool, UK. e-mail:
lemoreno70@hotmail.com
Background and challenges to implementation: Peru has

made considerable progress on prevention and control of
tuberculosis (TBC); the rate of mortality in 1990 was
198.6 per 100 000 inhabitants, for 2012 it was reported
105.2 cases per 100 000 inhabitants. Lima, capital of
Peru, concentrates 54% of total TB cases, 82% of MDRTB and 89% of extensively drug-resistant tuberculosis
(XDR-TB).Historical analysis show that best way countries struggle TBC is the implementation of national
programs of control, with national coverage using appropriate technologies and including during current activities
methods of monitoring and permanent evaluation.
Intervention or response: We monitored plasma levels of
rifampicin (RFP) and isoniazid (INH) and correlate
concentrations with clinical evolution of 204 nave
patients (older than 18 years old) from National
Tuberculosis Program from Ministry of Health, Peru.
Patients were from south and north area of Lima.
Dosages in plasma were done by liquid chromatography
method (HPLC), for rifampicin an own method was
developed; for isoniazid we used a validated HPLC
method from literature. Dosages were done at Day 5
(First Phase) and at Month 2 (Second Phase) of
treatment. Two samples of blood (5 ml) were collected
each time (2 and 4 hours after drug intake). During 6
months, patients were monitored for adverse drug
reactions, abandoned or failure of treatment. At the
end of 6 months of treatment, we reviewed charts to see if
the patients cured or not.
S154
Results and lessons learnt: Median concentration (2.3 lg/

ml) for INH were below recommended values for daily
administration (3-6 lg/l) and below (8.2lg/ml) from
recommended values for twice a week schedule (9-18 lg/
ml). Median concentration for RFP 34.4lg/ml (First
phase) and 41.4 lg/ml (Second phase) exceeded recommended ranges (8-24 lg/ml); however, concentrations
above 40.6 lg/mL were related with success outcomes at
6 months (P , 0.05).A delay of absorption of RFP was
detected, concentration of rifampicin were significantly
higher at 4 hours in comparison than 2 hours, either
phase of treatment. Moreover, concentrations of INH
and RFP at 2 hours were significantly higher during
second phase than first phase of treatment.
Conclusions and key recommendations: This study
demonstrated its utility to dose RFP and INH in plasma
and determine which cases constitute a risk for treatment
failure.
PC-713-04 Plasma concentration of first-line

anti-tuberculosis drugs in relation to minimal
inhibitory concentrations: a prospective
observational study
K Niward,1 L Davies-Forsman,2 J Bruchfeld,2 J Paues,1
E Eliasson,3 J Werngren,4 U Simonsson,5 T Schon
6
1
Department of Infectious Diseases, Linkoping

University
2
Hospital, Linkoping,
Unit of Infectious Diseases, Institute of
Clinical Medicine, Karolinska Institutet and Department of
Infectious Diseases, Karolinska University Hospital,
Stockholm, 3Department of Laboratory Medicine, Division of
Clinical Pharmacology, Karolinska University Hospital,
Stockholm, 4The Public Health Agency of Sweden,
Stockholm, 5Department of Pharmacology, Uppsala
University, Uppsala, 6Department of Clinical Microbiology
and Infectious Diseases, Kalmar, Sweden. e-mail:
tschon@hotmail.com
Background: Individualized treatment against tuberculosis (TB) based on therapeutic drug monitoring could
enhance the efficacy of current treatment. So far, few
studies have related plasma concentrations to the
minimal inhibitory concentrations (MICs) of M. tuberculosis. Our objective was to investigate the distribution
of plasma drug concentrations of isoniazid (INH),
rifampicin (RIF), pyrazinamide (PZA) and ethambutol
(EMB), in relation to the MICs.
Methods: Adult patients (n 33) receiving first line antiTB treatment were included. Plasma samples were
obtained after an overnight fast, at 0, 2, 4 and 6 hours
after drug intake (week 2) and at 0 and 2h during week 4
and 12. The drug concentrations were determined by mass
spectrometry and the MICs by serial dilutions in BACTEC
960 MGIT. Sputum samples were obtained at day 0 and 2,
as well as week 1, 2, 4 and 8. Sputum culture conversion as
well as time to culture positivity (TTP) were recorded.
Results: The median age was 34 years, 67 % (22/33)
were females and 58% were originating from subsaharan Africa. Most patients had pulmonary TB
(73%), of whom 94% had attained sputum culture
conversion by week 8. The MIC levels for INH and RIF
were low, ranging from 0.032-0.064 mg/l and 0.032-
0.125 mg/l. The median peak plasma concentrations for

INH, RIF, PZA and EMB evaluated as the highest
concentration at 2, 4 or 6h after drug intake at two weeks
of treatment were 5.3, 10.0, 39.3 and 2.5 mg/l,
respectively. Low plasma levels of RIF (,8mg/l) was
found in more than one third of the patients (13/33; 39
%), whereas 6% had low levels of INH (,2mg/l).
Importantly though, most of the patients with low RIF
concentrations (83%), had high ratios between peak
concentrations and individual MICs (.10). Low plasma
drug concentrations of RIF were observed for the only
two out of the 17 sputum culture positive patients who
exhibited an increasing bacterial load during the first two
weeks of treatment, as well as the only patient who did
not sputum culture convert by week 8.
Conclusion: In this study, where drugs were taken after
an overnight fast, low drug concentrations of RIF were
observed in 39% of the patients. Interestingly, the ratio
of drug concentrations of RIF and INH in relation to the
MICs, was high even in patients with low plasma drug
concentrations. The findings highlight the potential
importance of MIC determination in order to enable
individualized treatment of TB through therapeutic drug
monitoring.
PC-714-04 The role of wild-type distributions

and epidemiological cut-offs in determining
breakpoints for susceptibility testing against
M. tuberculosis
T Schon
1,2 1Department of Clinical Microbiology and
Infectious Diseases, Kalmar County Hospital, Kalmar,
2
Department of Medical Microbiology, Linkoping

University,
Linkoping
University, Linkoping,
Sweden. e-mail:
tschon@hotmail.com
Background: The current drug susceptibility breakpoints

(BPs) for Mycobacterium tuberculosis) are well established for some of the first line drugs but less so for the
second line drugs. Along with increasing drug resistance,
it is important that BPs are systematically reviewed and
revised. In this project, the aim is to define the susceptible
population of M. tuberculosis (i.e. the wild-type (WT)
distribution) for currently used drugs which is one of the
cornerstones for setting accurate BPs together with
pharmacokinetics/pharmacodynamics and clinical outcome data.
Methods: We have investigated the minimal inhibitory
concentration (MIC)-distributions of consecutive WT
and resistant isolates of M. tuberculosis (n 225) from a
mixed population of patients from all continents for
more than 20 first- and second line drugs, primarily by
using solid Middlebrook 7H10 medium (7H10) but also
BACTEC MGIT 960. Additionally, resistance genes have
been sequenced (inhA, katG, gyrA, rpoB and embB).
From the literature, MIC-data for the major first- and
second line drugs were extracted.
Results: For ethambutol, the current BP (5mg/L for
7H10) splits the upper end of the WT distribution which
leads to poor reproducibility within and between
laboratories. Additionally, up to 40% of isolates at
4mg/l have resistance mutations in embB. For important

second line drugs such as fluoroquinolones, current BPs
classify up to 26% of isolates with resistance mutations
in gyrA as susceptible which have implications for
treatment and the definition of extensively drug resistant
TB. The currently used breakpoint for rifabutin (RIB;
0.5mg/L for 7H10) is substantially higher than the
epidemiological cut-off (ECOFF), i.e. the highest MIC
within the wild-type distribution (0.064mg/l). Among 73
rifampicin resistant isolates with mutations in rpoB, 26%
of the isolates were susceptible to RIB at 0.5mg/L
although there are no clinical data to support that such
isolates could be treated by RIB. MIC distributions from
the literature search were recently published at the
EUCAST home page such as 1151 MIC-values from 12
laboratories for levofloxacin which could be used for
defining the ECOFF.
Conclusion: In this project, we defined the susceptible
population of M. tuberculosis for the majority of the
drugs currently used. Such data are important for
defining susceptibility breakpoints, for epidemiologic
surveillance and to resolve issues where there are
discrepancies between phenotypical and mutational
resistance analyses.
PC-715-04 Spectinamides: a new class of

semisynthetic protein synthesis inhibitors
against tuberculosis that overcome native drug
efflux
R E Lee,1 B Meibohm,2 A Lenaerts,3 E C Boettger,4
M M Butler,5 T L Bowlin5 1St. Jude Childrens Research
Hospital, Memphis, TN, 2University of Tennessee, Memphis,
TN, 3Colorado State University, Ft. Collins, CO, USA;
4
University of Zurich, Zurich, Switzerland; 5Microbiotix, Inc.,
Worcester, MA, USA. Fax: (1) 970 491 5125. e-mail:
lenaerts@colostate.edu
There is an urgent need for new tuberculosis (TB) drugs

with a novel mechanism of action which are effective
against multidrug and extensively drug resistant TB, that
are safe, and do not interact with HIV/AIDS medications,. The spectinamides, a new class of semisynthetic
analogs of the natural product spectinomycin, have been
modified to overcome efflux in Mycobacterium tuberculosis, thereby improving their activity dramatically. The
spectinamides are protein synthesis inhibitors which bind
to a distinct region of the 30S ribosome without showing
cross resistance to other aminoglycoside anti-tubercular
agents. Studies demonstrated no appreciable cytotoxicity
and no in vitro activity against eukaryotic cytosolic or
mitochondrial ribosomal translation. This reduces the
potential for side effects, such as ototoxicity and myeloid
suppression, which are liabilities commonly associated
with other protein synthesis inhibitors. Spectinamides
are highly soluble, with low protein binding, are not
subject to hepatic metabolism and are largely excreted
unchanged in the urine. The many compounds in the
series show robust in vivo activity in acute and chronic
mouse TB infection models when administered by
injection and inhalation. The lead compounds 1599
and 1810 demonstrate a promising therapeutic profile in
S155
mice as a single agent and combined with current and

new TB drugs. Combinations with rifampin and pyrazinamide showed additive to synergistic efficacy in
chronic Balb/c and C3HeB/FeJ mouse models suggesting
that the spectinamides may be good combination partner
agents. Current preclinical work ongoing includes: dose
fractionation studies in mice, GLP PK and rat toxicity,
neuromuscular blockade and nephrotoxicity safety
testing, manufacture optimization, and further combination efficacy trials in mice. The current development
plan is to develop a spectinamide combination regimen
with current and pipeline TB drugs suitable for treatment
of MDR tuberculosis without the noted ototoxicity
liability associated with aminoglycosides, which will be
replaced with the spectinamides in these regimens.
PC-716-04 Comparing the efficacy of drug

regimens for patients with drug-sensitive
pulmonary tuberculosis an aggregate data
meta-analysis of phase II studies
L Bonnett,1 G R Davies1 1University of Liverpool, Liverpool,
UK. e-mail: L.J.Bonnett@liv.ac.uk
Background: First-line therapy for tuberculosis (TB) has

remained unchanged for forty years. Most clinical drug
development programs begin with a short monotherapy
study using quantitative sputum bacteriology in a small
number of patients. Such studies of early bactericidal
activity (EBA) are then extended into Phase IIB studies
with longer treatment periods, using combinations of
drugs and larger numbers of patients. More commonly,
Phase IIB studies are based around culture conversion at
fixed time-points, usually two months. A complete
understanding of the performance of TB treatment
regimens in early phase clinical trials is crucial to
understanding their usefulness in predicting Phase III
trial results. We undertook a systematic review of Phase
II trials to determine the overall ranking of efficacy of
drugs and combinations.
Design/Methods: Phase II studies evaluating key drugs in
first-line treatment regimens for drug sensitive individuals with pulmonary TB were included. Outcomes of
interest were EBA over 2, 7 and 14 days, and the
proportion of patients with negative culture results at 8
weeks. Pooled estimates were obtained via the generalised inverse variance method.
Results: 131 relevant studies were identified presenting
data on 33,537 patients and 66 drug combinations. EBA
over 2 days was considered in 24 studies (32 drug
combinations). EBA over 7 and 14 days were less
frequently reported (6 and 8 studies respectively). The
proportion of patients culture negative by 8 weeks was
reported in 107 studies (46 combinations). Only 10
distinct regimens were considered across multiple studies
and phase II endpoints. All 10 were represented by EBA
0-2, 8 in EBA 0-7, and 7 in EBA 0-14. However, data for
2 month culture conversion was available for only 3
regimens. While rankings across EBA endpoints were
fairly consistent, it was not possible to compare these to 2
month culture conversion results.
S156
Conclusion: This is the first review to systematically

appraise the performance of single drugs and combination regimens in Phase II trials of treatment of pulmonary
TB. While the results partly reflect the history, development and differing goals of such trials, the narrowness of
this evidence base is concerning. Likewise, while rankings of the efficacy of treatment were reasonably
consistent on different Phase IIA endpoints, the existing
dataset is too small to provide convincing support to the
effectiveness of the current typical drug development
pathway.
03. Treatment outcomes and ethics:

bedaquiline
PC-717-04 The ethics of risk-benefit analysis of
new tuberculosis drugs
D Silva,1 R Upshur,2 A Dawson3 1Hannover Medical School,
Hannover, Germany; 2University of Toronto, Toronto, ON,
Canada; 3University of Birmingham, Birmingham, UK. e-mail:
diego.silva@utoronto.ca
Background: Since 2012, two new antitubercular drugs

have been approved for use to treat M/XDR-TB, namely
bedaquiline and delamanid; however, both have serious
side effects (e.g., unaccounted deaths in the treatment
arm of the bedaquiline study). Given the current
uncertainty surrounding TB, patients and communities
are required to balance the risks and benefits of new
treatments with incomplete or suboptimal information,
which requires normative judgments as to what is worth
risking and for how much. However, the very use of the
words risk and benefit should give pause: how do we
understand these terms, and from whose persepctive?
Design/Methods: A philosophical analysis using Jonathan Wolffs framework for evaluating risk and benefit
from an ethics viewpoint.
Results: What constitutes a risk or benefit related to
taking bedaquiline or delamanid may depend on, among
other things, whose risk and benefit is taken into account
and the context in which that risk evaluation is made.
Whether to take risks with new antitubercular drugs will
depend not only on whether one is the patient, the
healthcare worker, or member of the community, but will
also depend on the interpretation of the moral landscape
that led to contracting TB and the very need for
suboptimal pharmaceuticals in the first place.
Conclusion: Building on Jonathan Wolffs distinction
between an individuals perception of risk vs. the
perception of the imposition of risk, we argue that the
balancing of risks and benefits, both from the perspective
of individuals with TB and their communities, requires a
careful understanding of the social and political context
in which the risk of new TB drugs are introduced. How
supported or unsupported a TB patient feels may affect
not only psychologically, but also morally, how patients
and the public do, and ought to, balance the risks
associated with new TB drugs and the disease itself.
PC-718-04 Treatment interruption and directly

observed treatment of multidrug-resistant
tuberculosis patients in China
X Wei,1 J Yin,1 G Zou,2 Z Zhang,2 J D Walley,3 J Harwell,1
Q Sun,4 H T Li5 1Chinese University of Hong Kong, Hong
Kong, 2COMDIS-HSD China Research Consortium, Nuffield
Centre for International Health and Development, University
of Leeds, Shenzhen, China; 3Nuffield Centre for International
Health and Development, University of Leeds, Leeds, UK
4
Center for Health Management and Policy, Shandong
University, Jinan, 5Center for TB Control, Shandong Provincial
Chest Hospital, Jinan, China. Fax: (86) 755 2225 0390.
e-mail: zgy1021@hotmail.com
Background and challenges to implementation: China

has nearly one fifth of global multidrug-resistant
tuberculosis (MDR-TB) cases, and follows the 24-month
World Health Organization (WHO) standardised regimens. The paper aims to assess treatment interruption
among MDR-TB patients and its association with the
provision of directly observed treatment (DOT).
Intervention or response: We reviewed clinical charts and
conducted a questionnaire survey among all confirmed
MDR-TB patients who had been treated for at least 6
months from 1 January 2009 to 30 April 2012 in
Shandong Province. Treatment interruption was defined
as missing a dose for at least 1 day but for ,8 consecutive
weeks; the subset severe interruption was defined as
missing doses for 28 consecutive weeks.
Results and lessons learnt: Of 110 patients, 75 (68%)
interrupted treatment; 19 (17%) reported severe interruption, with a median duration of 30 days. Of the 110
patients, 26 (24%) received injections from family
members and 55 (50%) received DOT, 7 (13%) from
village doctors and 48 (87%) from family members.
Patients who underwent DOT with a family member had
less severe interruptions (OR 0.25, 95%CI 0.050.98)
than those who were given DOT by a village doctor or
who did not undergo DOT.
Conclusions and key recommendations: Family members
may act as treatment supporters for MDR-TB patients to
reduce treatment interruptions, but require orientation
on their role.
PC-719-04 Compassionate use of bedaquiline:

interim outcomes from the Armenian National
Tuberculosis Control Office
C Hewison,1 J Faqirzai,2 A Hayrapetyan,3
N Khachatryan,2 S Qayyum,1 F Varaine1 1Medecins

Sans
`
Frontieres
(MSF), Paris, France ; 2MSF, Yerevan, 3National
Tuberculosis Control Office, Yerevan, Armenia. e-mail:
cathy.hewison@paris.msf.org
Background: In 2013, confronted by the growing

challenge of drug resistant TB, the National Tuberculosis
Control Office of Armenia, supported by Medecins Sans
Frontières (MSF), introduced bedaquiline (Bdq) containing treatment through compassionate use for extensively
drug resistant (XDR-TB) or MDRTB patients with
additional resistance to fluoroquinolones (MDRFQ)
or injectables. We present results after two years.
Design/Methods: Eligible patients were enrolled to

receive 6 months of Bdq added to a background regimen
according to WHO recommendations. Two WHO
Group 5 drugs (imipenem and linezolid) previously not
used in Armenia were included in the regimen when
needed. Monitoring included monthly sputum smear,
culture, biochemical tests and ECG monitoring. Culture
conversion was defined as 2 consecutive negative culture
results one month apart for culture positive patients at
start of treatment.
Results: Between April 2013 and March 2015, 60
patients with a mean age of 42 years (range: 18-68)
initiated a Bdq containing regimen; 88% (53/60) were
male; 41.6% (25/60) had XDR-TB and 50% (30/60) had
MDRFQ. Three cases were HIV positive (5%). All 60
patients also received linezolid and 43/60 (71.6%)
received imipenem-cilastatin. Of the 38 patients enrolled
before July 2014, 30 were culture positive at start of
treatment. Culture conversion was achieved in 53.3%
(16/30) at 2 months and 80% (24/30) at 6 months.
Serious adverse events were reported in 4 patients, 1
attributed to Bdq QT effects without clinical consequences for the patient who completed Bdq treatment.
Fifty five percent (21/38) had an increase in QTcF from
baseline with a mean increase of 37.4 milliseconds (ms)
(range: 4-96 ms), 86% (18/21) of whom were taking
clofazimine concomitantly. Results at 14 months of
treatment were assessed for the 28 patients who started
treatment before February 2014: 20(71%) culture
negative, 1 (4%) died, 1 (4%) lost to follow up, 6
failures (21%). Amongst the failures, 3 never converted
and subsequently died and 3 reverted after initial
conversion and are still under treatment.
Conclusion: These preliminary outcomes of a Bdq
containing regimen show good tolerability and early
treatment response. However reversion rates highlight
the need for further studies on the optimum length of Bdq
treatment and the importance of considering prolongation of Bdq beyond 24 weeks on a case by case basis for
patients with otherwise weak regimens.
PC-720-04 Is bedaquiline as effective as

fluoroquinolones in the treatment of MDR
tuberculosis? A microbiological appraisal
through a case-control study
3,4 N Veziris,1,2,5 E Caumes,5
L Guglielmetti,1,2 D Le Du,
6
Y Yazdanpanah, J Robert,1,2,5 M Jachym,3 for the
MDR-TB Management Group of the French National
Reference Center for Mycobacteria and the Physicians
of the French MDR-TB Cohort5 1Institut National de la
Sante et Recherche Medicale,

U1135, Team E13
UPMC Univ Paris
(Bacteriology), Paris, 2Sorbonne Universites,
06, CR7, Team E13 (Bacteriology), Paris, 3Centre Hospitalier
de Bligny, Briis-sous-Forges, 4APHP, Centre Hospitalier
Garches, 5APHP, Hopitaux
Raymond Poincare,
Universitaires
ere-Charles
`
Pitie Salpetri
Foix, Paris, 6APHP, Hopital
Bichat
Claude Bernard, Paris, France. e-mail:
lorenzo.guglielmetti@gmail.com
Background: Bedaquiline (Bdq) is a new drug approved

to treat multidrug-resistant tuberculosis (MDR-TB,
S157
resistant to isoniazid and rifampicin). Because of the

paucity of drugs available for MDR-TB treatment,
interventional studies aimed at evaluating the activity
of Bdq are difficult. Therefore, we designed a retrospective study to compare the microbiological efficacy of
Bdq- and fluoroquinolones (Fq)-containing regimens, as
Fq are a mainstay of the treatment of MDR-TB.
Design/Methods: A nested case-control study was
performed within the cohort of MDR-TB cases hospitalized from 2006 to 2014 in a single referral TB center in
France. Case-patients were defined as those treated with
Bdq without Fq, or with Fq when isolates were
moxifloxacin (Mfx)-resistant. Control-patients were
those treated with Fq but without Bdq, and harboring
ofloxacin (Ofx)-susceptible isolates. All patients had
positive sputum cultures at treatment start, and received
at least one second-line injectable drug and linezolid.
Time to culture conversion was measured from treatment
start to the first of two consecutive negative examinations. Kaplan-Meier curves were estimated. Mantel-Cox
test was used to compare time to culture conversion
between cases and controls. A Cox proportional hazards
model was used to assess the association between time to
culture conversion and explanatory variables.
Results: A total of 25 cases and 42 controls were eligible
among a cohort of 172 MDR-TB patients. The culture
conversion rate at 3 months was higher in controls (74%)
than in (44%) cases, while no statistical difference was
found at 6 months (93% vs. 96%, respectively). The time
to culture conversion was significantly shorter for
controls (60 days, IQR 35-89) than for cases (98 days,
IQR 70-124; P , 0.01) (Figure). Characteristics associated with slower time to culture conversion were
presence of lung cavities (Hazard Ratio (HR) 0.15,
95% Confidence Interval (CI) 0.07-0.31), sputum-smear
positivity at treatment start (HR: 0.21, 0.05-0.99), and
male sex (HR: 0.31, 0.16-0.61).
Conclusion: Patients treated with Bdq and those treated
with Fq reached a similar 6-month culture conversion
rate, although Fq-treated patients had a faster time to
culture conversion. Cavitary, smear-positive pulmonary
TB, together with male sex, was associated with slower
time to culture conversion.
Kaplan-Meier analysis of time to culture conversion of

cases (solid line) and controls (dashed line). Dots indicate
censoring.
S158
PC-721-04 Reasons for not participating in a

tuberculosis clinical trial, Lima, Peru 2014
PC-722-04 Bedaquiline-containing treatment

regimens for multidrug-resistant tuberculosis
C C Contreras,1 C Moran,1 M Lindeborg,2 F Garcia,1

M Milstein,3 L Lecca,1 C Mitnick3 1Socios En Salud Sucursal
Lima, Peru; 2Harvard College, Boston, MA, 3Harvard
Peru,
Medical School, Boston, MA, USA. e-mail:
ccontreras_ses@pih.org
S Borisov,1 T Ivanushkina,1 D Ivanova,1 N Litvinova,1

A Filippov1 1Moscow Research and Clinical Center for TB
Control, Moscow, Russian Federation. Fax: (7) 499 785
2082. e-mail: sebarsik@gmail.com
Background: Clinical trials are essential to optimize TB

treatments and to minimize safety concerns from
medication use. Social and personal factors, such as
stigma, barriers to access, and lack of patient education,
can impede trial participation. This study seeks to better
understand the underlying reasons for non-participation
in a TB clinical trial in Lima, Peru.
Design/Methods: Data were collected from adults (718
years old) with newly diagnosed, drug-susceptible,
pulmonary TB in metropolitan Lima who were invited
to participate in a clinical trial between September 2013
and February 2015. Participants chose not to partake at
three main times: 1) refusing a screening invitation at the
health center, 2) not signing the informed consent form
(ICF), and 3) screen failure after enrollment.
Results: 542 potential candidates were invited to
screening, of which 392 accepted the invitation to go to
the study site and 150 (27.7%) did not because: 66
(44.0%) prefer health center treatment/have different
insurance; 33 (22.0%) lack of time; 26 (17.3%) family
influence, 7 (4.7%) lack of desire to be in a trial; 6
(4.0%) site decision; 5 (3.3%) drug susceptibility test not
available; and 7 (4.7%) other reasons. Of the 392 who
accepted the invitation, 345 signed the ICF and 47
(12.0%) did not because: 10 (21.3%) want immediate
treatment; have negative feelings about sputum (4,
[8.5%]) and blood collection (3, [6.4%]); 3 (6.4%)
dislike study-specific requirements; 3 (6.4%) family
influence; 2 (4.3%) lack of desire to be in a trial; 13
(27.7%) pre-screening criteria not met; and 9 (19.1%)
site decision. Finally, of the 345 who signed the ICF, 165
(47.8%) did not pass screening and get randomized to
study drug because: 146 (88.4%) met at least one
exclusion criterion, 4 (2.4%) family influence; 3 (1.8%)
prefer health center treatment/have different insurance; 1
(0.6%) dislike of home visits; 4 (2.4%) study suspension/
enrollment already closed; 1 (0.6%) drug susceptibility
test not available; 1 (0.6%) loss-to-follow-up; 5 (3.0%)
other.
Conclusion: It is important to distinguish between
personal reasons (lack of time, negative view of trials
or study-specific protocol) and social reasons (family
influence, proximity to health center vs. study site) that
TB patients choose not to participate in TB trials. To
improve study participation, researchers should aim to
strengthen patient/family education and address social
gaps impeding access to health services.
Background: The second-line drugs optionally combined

with the WHOs fifth group drugs (especially linezolid)
are effective in MDR-TB patients. In some cases, the
adequate chemotherapy (min 4 drugs) is unrealizable due
to the extended drug-resistance (XDR) and/or treatment
intolerance. The new drug bedaquiline (Bdq) is
essential, but it is not yet thoroughly tested in regimens
with traditional TB-drugs and new anti-mycobacterial
agents.
Design/Methods: The prospective unblinded non-randomized one-centered study include 54 pulmonary TB
pts, 18-73 y.o., 68.5% male; 20,4% (11 pts) new and
79,8% (43 pts) retreated. MDR was identified in 29.6%
(16 pts) and XDR in 70.4% (38 pts). Lung cavities
detected in 87.0% (47 pts), bilateral lesions in 40.7%
(22 pts). Bdq was included in regimens in all patients, the
next one was, if possible, linezolid (96.3% 52 pts),
cycloserine/terizidone (81.5% 44 pts), fluoroquinolones (70.4% 38 pts), capreomycin (46.8% 25 pts),
prothionamide (29.6% 16 pts), PAS (27.8% 15 pts),
pyrazinamide (18.7% 10 pts), ethambutol (13.0% 7
pts). If the regimen didnt include at least four drugs, we
add macrolides (38.9% 21 pts)), meronem (9.5% 5
pts), amox/clav (1 pt), isoniazid in high doses (1 pt). The
evaluation performed after 24 weeks of Bdq administration and all patients will be followed up to 24 months.
Results: The treatment with Bdq was completed in
87.0% (47 pts), in 4 pts (7.4%) treatment was
interrupted (but they werent lost to follow-up) and in
3 pts (5.%) Bdq was excluded due to the serious adverse
events. The evident involution of symptoms were
obtained in 85.1% (40/47 pts), the X-ray improvement
- in 80.5% (38/47 pts). The sputum smear conversion
total score (on liquid media): 87.2% (41/47 pts) with
Med4 wks (IQR 2.0-11.5). The time of conversion was
prolonged in pts with large cavities: 8 vs. 2 wks, P
0.001 (log-rank test) and bilateral involvement (4 vs. 2
wks, P 0.05) The SAE III-IV were obtained in 11 pts
(20,4%), include hepatitis in 3, hypereosinophilia (19%
and more) in 3, QTc prolongation up to 520 ms in 2
(without arrhytmia), obstinate vomiting in 2, anemia
(Hb,69 g/l) in 1 pt. SAE III-IV were attributed to Bdq
in 4 pts and only 3 stopped the treatment with Bdq.
Conclusion: The regimens, based on Bdq, second-line
TB-drugs and the fifth group drugs are well-tolerated
and high effective in XDR-TB. We need obtain data to
argue the prolongation of Bdq course.
PC-723-04 Bedaquiline: new capabilities to

treat MDR-TB patients
I Vasilyeva,1 A. Samoilova,1 T. Bagdasaryan,1 V Testov,1
A. Tikhonov,1 L Chernousova1 1Central TB Research
Institute, Moscow, Russian Federation. Fax: (7) 499 785
9108. e-mail: vasil39@list.ru
Background: Study treatment efficacy of MDR-TB

patients under chemotherapy regimen with bedaquiline.
Design/Methods: 54 MDR/XDR-TB culture positive
patients treated in CTRI during 2010 2014 participated
in the study. DR to 1st and 2nd line drugs was confirmed
for all patients with culture method on liquid media in
the certified high quality laboratory. Patients were
divided into two groups according to DR patterns: 1st
group 23 MDR-TB patients without additional DR to
fluoroquinolones, 2nd group 31 MDR-TB patients
with additional DR to fluoroquinolones (including 21
XDR-TB patients). Kanamycin or amikacin or capreomycin, pyrazinamide, protionamide or aminosalicylic
acid, cycloserine or terizidone, levofloxacin in the dose of
1.0 were included in the chemotherapy regimen of all
patients during first 6 months of treatment with further
transfer to moxifloxacin. Bedaquiline was appointed to
all patients during first 6 months of treatment according
to the following treatment plan: 400 mg daily during 2
weeks, then 400 mg 3 times per week. Intensive phase of
treatment continued till getting two-four negative
cultures but not less than eight months. Treatment at a
continuation phase lacked injection drug. Culture
negativation was an indicator of TB patients treatment
efficacy. Treatment efficacy of TB patients included into
the study also was assessed according to WHO definitions (2013) in 24 months after treatment initiation and
full treatment course finalization.
Results: 39.1% (9 of 23) of patients in the 1st group and
19.4% (6 of 31) of patients in the 2nd group had culture
negativation in 2 months of treatment, 73.9% (17 of 23)
and 61.3% (19 of 31) in four months, 87% (20 of 23) b
90.3 (28 of 31) in six months correspondingly, P .
0.05. Full recovery was observed among all patients
included into the study.
Conclusion: 88.9% of MDR-TB patients regardless
additional DR as a result of inclusion of bedaquiline in
a chemotherapy regimen had culture negativation by six
months of treatment. Observations made in 24 months of
treatment initiation and after full treatment course
finalization demonstrated 100% recovery of MDR-TB
patients including XDR-TB patients and gave hope to
patients, who used to be considered hopeless.
S159
PC-724-04 Geographical differences in baseline

drug resistance patterns and their impact on
treatment outcomes in a single-arm
bedaquiline phase II trial (C209)
A Pym,1 A Diacon,2 F Conradie,3 S Tang,4
M Haxaire-Theeuwes,5 N Lounis,5 B Van Baelen,5
B Dannemann6 1Medical Research Council and Kwazulu
Research Institute for TB and HIV (K-RITH), Durban, 2Division
of Medical Physiology and the Department of Science and
Technology/National Research Foundation Centre of
Excellence for Biomedical Tuberculosis Research and Medical
Research Council Centre for Tuberculosis Research, Cape
Town, 3Clinical HIV Research Unit, Department of Medicine,
Faculty of Health Sciences, University of the Witwatersrand,
Johannesburg, South Africa; 4Tuberculosis Department,
Beijing Chest Hospital, Capital Medical University, Beijing
Tuberculosis and Thoracic Tumor Research Institute, Beijing,
China; 5Janssen Infectious Diseases BVBA, Beerse, Belgium;
7
Janssen Research & Development LLC, Titusville, NJ, USA.
Fax: (27) 31 203 4702. e-mail: alexanderpym@hotmail.com
Background: Global variations in the distribution of

drug resistance patterns could potentially influence
regional treatment outcomes for multi-drug resistant
tuberculosis (MDR-TB). C209 (NCT00910871) was a
Phase II, open-label, single-arm trial of bedaquiline
(BDQ) as part of an individualized MDR-TB treatment
regimen in patients (pts) most of whom had received
prior MDR-TB treatment.
Design/Methods: 233 pts 718 yrs from 11 countries
with confirmed sputum-smear positive pulmonary
MDR-TB (either newly diagnosed or previously treated),
including pre-extensively (pre-XDR) or extensively
(XDR) drug-resistant TB received BDQ for 24 wks
(400mg qd for 2 wks, 200mg tiw for 22 wks) with an
individualized background regimen (BR), followed by
BR only for up to 120 weeks (final assessment). Baseline
resistance patterns by region (defined as China, other
Asian countries, Eastern Europe, South America, South
Africa) and their impact on treatment response (sustained culture conversion at 120 wks) were assessed in
the modified intent-to-treat population (n 205)
excluding 28 pts with drug-susceptible TB or negative
MGIT cultures at baseline.
Results: Baseline rates of second-line drug resistance
(Table) were highest in China: 73% fluoroquinolone
resistance, 32% aminoglycoside resistance and 27%
ethionamide resistance. Eastern Europe had the highest
rate of aminoglycoside resistance (44%). Baseline rates
of pre-XDR-TB and XDR-TB were higher in China
(48% and 28%, respectively) than in other Asian
countries (37% and 26%) and in Non-Asians (14%
and 17%). Surprisingly higher rates of baseline resistance
patterns in these regions did not lead to lower conversion
rates at 120 wks (MF): China (86%), other Asian
countries (77%), Eastern Europe (76%), South Africa
(60%) and South America (62%).
Conclusions: Addition of BDQ to an MDR-TB regimen
resulted in high conversion rates at 120 wks in a
population who had mostly received previous secondline drugs. There were wide regional differences in
baseline drug resistance, with the highest rates of
S160
fluoroquinolone resistance and XDR-TB in China, and

highest rate of aminoglycoside resistance in Eastern
Europe. Regional differences in the distribution of
baseline drug resistance patterns did not result in poorer
treatment outcome in regions with more extensive drug
resistance patterns.
Table. Second-line drug resistance at baseline; mITT, 7H11
agar proportion method*
Asia (other) China
n 26
n 41
Drug resistance
n (%)
n (%)
Eastern South
South
Europe America Africa
n 34 n 10 n 63
n (%)
n (%) n (%)
Ethionamide
Capreomycin
Kanamycin
Streptomycin
Ofloxacin
Pyrazinamide
8
8
15
34
13
30
4
2
7
19
14
18
(17)
(8)
(29)
(79)
(58)
(72)
11
7
13
37
30
30
(27)
(17)
(32)
(90)
(73)
(73)
(24)
(24)
(44)
(100)
(38)
(88)
0
3
3
8
1
9
(0)
(30)
(30)
(80)
(10)
(90)
6
5
8
36
7
48
(10)
(9)
(14)
(62)
(12)
(76)
Mortality at 9 months was 20% and 31% (P 0.2) and

treatment failure was 11% and 21%, in the ISR and
standard cohorts respectively (P 0.08). More patients
in the ISR cohort were culture negative and on treatment
at 9 months (64% vs. 40%, P 0.002). Among the 37
patients given an ISR (excluding 2 transferred out), 2
(5%) had died and 30 (81%) had negative cultures at 9
months (Figure 1). Median time to initiation of an ISR
following standard MDR treatment initiation was 1.9
months (IQR 1.1-3.5).
Conclusion: Interim outcomes among patients starting
ISRs containing BDQ and LZD are encouraging,
showing increased culture conversion and a trend
towards lower mortality and treatment failure. A
significant impact on mortality might be delayed due to
slow implementation or ISRs and short follow-up time.
As more patients are eligible and access ISRs, continued
analyses of outcomes are required.
*Drug susceptibility testing was performed at a central laboratory (Institute

of Tropical Medicine, Antwerp, Belgium).
Pyrazinamide drug susceptibility testing was performed using the MGIT960

PZA kit.
PC-725-04 Encouraging early outcomes for

individualized treatment of multidrugresistant tuberculosis using bedaquiline and
linezolid in Khayelitsha
E Mohr,1 H Cox,2 L Wilkinson,1 G Van Cutsem,1,3 V Cox,1
`
J Daniels,1 O Muller,1 J Hughes1 1Medecins
Sans Frontieres
(MSF), Khayelitsha, 2University of Cape Town, Cape Town,
3
MSF, Cape Town, South Africa. e-mail:
mohrek27@gmail.com
Background and challenges: Current treatment options

for multidrug resistant tuberculosis (MDR-TB) with
second-line (SL) resistance are limited and yield poor
outcomes. We describe early outcomes of patients treated
with individualized strengthened treatment regimens
(ISRs) containing bedaquiline (BDQ), linezolid (LZD),
clofazimine and/or rifabutin, following detection of SL
resistance.
Intervention or response: Patients with any SL resistance
(ofloxacin, amikacin) who started MDR-TB treatment
(January 2009-September 2011) were compared to those
starting treatment after ISR availability (October 2011July 2014). Patients with previous MDR-TB treatment
failure were excluded from the analysis. Prior to ISR
availability, standardized regimens were adapted with
addition of one or more of PAS, clofazimine and
capreomycin. LZD and BDQ became available in 2011
and 2013 respectively; implementation of ISRs was slow
due to cost and administrative constraints. Patients were
offered ISRs if SL resistance was confirmed, whether or
not current treatment was failing.
Results: Overall, 176 patients were assessed (72% HIV
positive); 80 in the standard cohort and 96 in the ISR
cohort. The proportion of previously treated MDR-TB
cases was 14% in the ISR and 1% in the standard cohort
(P 0.003). Thirty nine (41%) patients in the ISR cohort
received ISR; 20 (51%) received LZD and BDQ, 18
(46%) received LZD only and 1 (3%) received rifabutin.
04. New approaches to health systems,

care and management for drug-resistant
TB
PC-726-04 Occluding cavities with
endobronchial valves for the treatment of
multidrug-resistant tuberculosis: early clinical
experience
Z Wang,1 H An,1 L Liu,1 T Wang1 1309th Hospital of
Chinese Peoples Liberation Army, Beijing, China. e-mail:
wzy2004177@sina.com
Background: Multi-drug resistant tuberculosis (MDRTB) is a serious clinical problem in many parts of the
world i.e., 650 000 patients according to WHO (2010).
Because the development of new anti-TB drugs has not
kept pace with drug resistance, and also patient poor
treatment compliance, alternative approaches are needed.
Design/Methods: We investigated whether placement of
endobronchial valves (EBV, Zephyrw, Pulmonx, Redwood City, CA, USA) to induce selective atelectasis in
lobes that have TB cavities may be effective in improving
the treatment of MDR-TB. Zephyr EBV are one-way

valves that allow air and secretions to exit a lobe, but
prevent air from entering, inducing a lobar collapse or
atelectasis and an anaerobic environment. This minimally invasive form of treatment, carried out via bronchoscopy, is usually indicated for reducing hyperinflation in
COPD patients with severe emphysema.
Results: We present here our first 3 cases (Chinese, 1M,
2F) with MDR-TB that were treated with EBV valves.
They had previously received optimal TB treatments for
an average of 9.0 6 4.7 y (range: 3-16 y) but remained
multi-drug resistant. EBV implants to collapse TB
cavities was considered. Age at EBV treatment was
29.7 6 5.1 y. On average, 3.4 valves were implanted
(range: 1-5) per patient in 1 to 4 sequential sessions. In
total, 4 cavities were treated i.e., 2 in one patient and 1 in
the others. Full cavity collapses (100%) were achieved in
two instances and a partial collapse (i.e., 87% and 58%)
for the other cavities. TB status improved and all patients
became smear negative after EBV treatments. The
average follow-up period at the time of abstract
submission was 9.7 6 4.6 mo (range: 5-16 mo).
Conclusion: The early experiences reported here indicate
a promising potential for EBV implantation in the
management of multi-drug resistant lung tuberculosis.
Prior to the implantation of EBV in these patients, antituberculosis drugs failed to limit the number of lesions
and the development of cavities, but these conditions
were reversed post EBV implantation. There also was
conversion to smear negative sputum, thus potentially
limiting contamination risks. These results suggest that
lobar exclusion could be effective in the successful
control of tuberculosis and that sequential treatments
with several valves may be required. Longer follow-up
periods and additional cases will be available by the time
of the conference.
PC-727-04 Decentralized MDR-TB service model

increased access to case finding in the Amhara
and Oromia regions of Ethiopia
Y Molla,1,2 M Aseresa Melese,1,2 I Jemal Abdulahi,2,3
D Habte,1 Y Haile,4 D J Dare,1 Y Anteneh Kassie,1
P G Suarez2 1Management Sciences for Health, Help
Ethiopia Address the Low Tuberculosis Performance (HEAL
TB) Project, Addis Ababa, Ethiopia; 2Management Sciences
for Health, Center for Health Services, Arlington, VA, USA;
3
Oromia Regional Health Bureau, Addis Ababa, 4USA Agency
for International Development (USAID), Addis Ababa,
Ethiopia. e-mail: ymolla@msh.org

prevalence of MDR-TB in Ethiopia among re-treated
cases was 17.6% and that of new patients was 2.3%,
with over 3000 patients estimated to have MDR-TB in
2013. However, the number of treatment centers was
only two in 2011 for a country of 90 million population,
with huge geographic barriers to access. To address this
challenge, the country developed more treatment initiating centers, and for daily dose of directly observed
treatment of stabilized patients decentralized to health
centers.
S161
Intervention or response: In 2012, the FMoH began

working to decentralize MDR-TB services with support
from the USAID-funded Help Ethiopia Address the Low
TB Performance (HEAL TB) project. First, we renovated,
equipped three MDR TB centers, and also provided
technical support to sites already renovated by other
partners. The partners also trained health workers
throughout the regions to identify presumptive MDRTB cases and refer sputum samples for culture. The
sample transport was strengthened for presumptive
MDR-TB cases to Regional Reference Laboratories and
we assisted with placement of 40 GeneXpert machines in
the two regions. The project also supported targeted
clinical mentoring, organized special clinic days, provided continuing medical education system for the health
workers, and instituted a mortality audit.
Results and lessons learnt: Between January 2012 and
December 2014, these interventions helped expand
MDR-TB treatment service: treatment initiating centers
(TIC) increased from 2 to 17; testing for presumptive
MDR-TB from 662 to 12, 369; and patients ever enrolled
from 56 to 548 (Table). More than 300 health facilities
are providing monthly follow up to patients who began
treatment in the TICs, but were discharged after culture
conversion. Six-month culture negativity rate improved
from 41% to 69%. The cure and treatment success rate
for the first cohort of 36 patients were 61.1% and 74.9%
respectively. The proportion of death, failure of treatment and lost to follow up for the first cohort were
11.1%, 2.8% and 11.1% respectively.
Conclusions and key recommendations: Decentralizing
MDR-TB services to secondary and primary care levels
improved access to services, and improved treatment
outcomes. Designated clinic appointment days for follow
up, continuing medical education for health workers,
and death audits were critical in improving the quality of
care.
PC-728-04 Drug use review program in

Uzbekistan: pathway to improved rational use
of anti-tuberculosis medicines
M Tillyashaykhov,1 I V Liverko,1 V Belotserkovets,1
M Kavtaradze,2 A Salakaia2 1Republican Specialized
Scientific and Practical Medical Center of Phtisiology and
Pulmonology, Ministry of Health, Tashkent, Uzbekistan;
2
Systems for Improved Access to Pharmaceuticals and
Services (SIAPS) Program, Management Sciences for Health
(MSH), Arlington, VA, USA. Fax: (995) 322 232 490. e-mail:
kavtaradze.maya@gmail.com
Background: When a comprehensive indicator-based

assessment of Uzbekistans tuberculosis (TB) pharmaceutical system was conducted in 2014, findings revealed
gaps in the rational use of anti-TB medicines: only 60.4%
of prescriptions for intensive phase and 70.6% of
continuation phase for multidrug-resistant (MDR)-TB
patients were in accordance to the national standard
treatment guidelines. To respond to this problem, the
National TB Program (NTP) of Uzbekistan decided to
implement a drug use review (DUR) programa quality
S162
assurance intervention that identifies and remedies

problems related to drug use.
Intervention: In February 2015, with the technical
assistance of the US Agency for International Development (USAID)-funded Systems for Improved Access to
Pharmaceuticals and Services (SIAPS) Program, the NTP
piloted DUR at 3 TB facilities of Tashkent City. The
objectives of the review were to identify gaps in the
rational use of second line anti-TB medicines through
criteria-based review including prescribing of medicines,
monitoring of treatment and management of side effects.
The methodology included data collection according to
preset criteria through review of treatment cards and
patient interviews followed by analyses of collected data.
Results: Justification criteria for prescribing drug-resistant TB treatment regimen were met 100%. On average,
67% of baseline tests were conducted for patients before
start of treatment. The average percentage of patients
receiving the appropriate dose of all anti-TB medicines
was observed in 94% cases. During the intensive phase of
treatment, the treatment regimen was changed at least
once in 38% cases; of these cases, only 58% were
justified by a doctor in the treatment card. During the
continuation phase, the regimen was changed at least
once only in 17% of cases, of which only 40% was
justified in the treatment card. Average percentage of side
effects most reported by patients during the interviews
was 57% each for vestibular disorders and arthralgia; the
average percentage of records for these side effects in the
treatment cards was 22% and 11%, respectively.
Conclusion: Based on the DUR results, an improvement
plan was developed and disseminated to appropriate
treatment facilities staff. The DUR is scheduled to be
repeated in six months after the improvement interventions are in place. The DUR program will also be
expanded to other TB facilities, making the process
systematic.
PC-729-04 Systematic expansion of drugresistant tuberculosis services reduced

treatment initiation delay by five fold in
Jharkhand, India
V Ghule,1 A Sreenivas,1 P Parmar,1 R Dayal,2 R Pathak,1
S Saruk,3 A Mitra2 1World Health Organization Country
Office for India, New Delhi, 2Government of Jharkhand,
Ranchi, 3Government of Maharashtra, Washim, India. e-mail:
ghulev@rntcp.org
Background and challenges to implementation: Multi

drug resistant tuberculosis (MDR-TB) is a major
challenge for Indias Revised National TB Control
Program (RNTCP). New TB patients are treated free
under program with first line drugs. Non responders are
screened by culture and drug sensitivity testing (C&DST)
in quality assured laboratory. TB patients with known
resistance to first line drugs (isoniazid and rifampicin) are
known as MDR-TB patients. Programmatic Management of drug resistant tuberculosis (PMDT) - a systematic approach to address MDR-TB care was launched by
Government of India in 2007. Establishment of quality
assured laboratory and treatment initiation centres - drug

resistant TB (DR-TB) centres as per guidelines; early
diagnosis and treatment initiation within a week of
diagnosis were challenges in resource poor settings of
Jharkhand state. We sought to assess program performance and impact of systematic service delivery on
MDR-TB care in the state.
Intervention or response: Jharkhand is predominantly a
tribal province of eastern India with poor geopolitical
settings and difficult hilly terrain. With establishment of
one certified laboratory and one DR-TB centre, phase
wise implementation of PMDT was launched in 2010 to
cover 34.4 million population of 24 districts. (Table 1)
One more diagnostic technology and DR-TB centre was
established to improve access to MDR-TB care. PMDT
services were launched after technical assessment and
approval by Govt. of India. Systematic record review of
treatment register (417 patients) and performance
reports (2011 to 2013) was conducted to assess
performance trend of MDR-TB care.
Results and lessons learnt: Systematic roll out of PMDT
services resulted in 100% access to care (additional 34%
screening and 29% treatment initiation), reduction in
treatment initiation delay (by 18 weeks) and decreased
death proportions (by 31%) of MDR-TB patients in
Jharkhand India. (Figure 1)
Conclusions and key recommendations: Systematic roll
out of PMDT services improves access to MDR-TB care.
Decentralised care delivery saves additional lives.
PC-730-04 Need to improve knowledge about

management of MDR-TB among medical
personnel in India
P Mithra,1 P Banuru Muralidhara,2 B Unnikrishnan,1
R Thapar,1 N Kumar1 1Manipal University, Manipal,
2
South-East Asia Office, New Delhi, India. Fax: (91) 114 605
4430. e-mail: bmprasad@theunion.org
Background: Globally and in India, there is a rise in

number of Multiple Drug Resistant (MDR) Tuberculosis
patients. Many argue that this increase is also contributed due to irrational use of medicines; mainly antibiotics
by private medical practitioners. Following graduation
almost 90% of students from medical college adopt
private practice. Prior to the graduation, it is therefore
necessary to assess their knowledge about management
of MDR-TB.
Methods: A cross sectional study was conducted in a
medical college from south Indian city of Mangalore. A
total of 188 students included; Interns (n 98) and Post
Graduates [PG] (n 90) were interviewed through a semistructured questionnaire from January to March 2015.
They were selected using a convenience sampling
method. The collected data was analysed using Statistical
Package for Social Sciences [SPSS] version 16.0.
Results: Overall, 61.8% had adequate knowledge about
MDR-TB and knowledge was higher among PGs. In
total, .80% of them knew correct diagnostic methods.
However, only 21% of Interns and 34% of PGs were
aware of correct duration of treatment. Management
about MDR-TB cases was known only to about 24% of
Interns and 13% of PG students.
Conclusion: Knowledge and management of MDR-TB
among students Interns and PGs is limited. There is a
great need to have Continued Medical Education on
MDR-TB and include Standard Treatment Guidelines
into curriculum of medicine.
PC-731-04 Pathways of care for DR-TB in

Mumbai slums: case for use of rapid diagnostic
tests
E Lobo,1 A Patil,1 S Rangan,1 S Shah,1 Y Dholakia,1
N Mistry1 1Foundation for Medical Research, Mumbai, India.
Fax: (91) 222 493 2876. e-mail: fmr@fmrindia.org
Background: The rising threat of drug resistant TB (DRTB) in India necessitates early detection of drug
resistance and appropriate treatment initiation. In recent
times Mumbai has become the epicenter of various forms
of DR-TB with about 60% of Mumbai population living
in overcrowded slums with poor hygiene, sanitation and
ventilation. The objective of this study was to record the
duration of first care seeking, diagnosis and subsequent
initiation of DR-TB treatment of vulnerable patients
from Mumbai slums. Twenty-three DR-TB patients
identified by a household level survey of 15 high burden
TB wards in Mumbai were interviewed using a semistructured interview schedule. Median durations of first
care-seeking, diagnosis, initiation of DR-TB treatment
and total pathway were calculated.
S163
Results: The entire pathway from onset of symptom/s till

initiation of DR-TB comprised of 109 days ((IQR 52,
277). Delay was more pronounced for first care seeking
(20 days, IQR 12, 46), and DR-TB diagnosis (83 days,
IQR 14, 200). Treatment initiation however was rapid (0
days, IQR: 0, 1) with most patients initiated on treatment
on the same day as diagnosis. The total DR-TB pathway
time of patients with first episode of TB disease (179
days, IQR 122.5, 345.5) was found to be more than
twice as that of patients with previous episodes of TB (69
days, IQR 41.5, 126).
Conclusion: The unacceptable delay for diagnosis and
treatment of DR-TB in Mumbai advocates for stronger
implementation of early screening of patients for DR-TB,
especially in vulnerable setting through use of rapid genebased technologies.
PC-732-04 A novel therapeutic vaccine against

tuberculosis in the cynomolgus monkey model:
preclinical study and clinical trial
M Okada,1 T Nakajima,2 Y Kaneda,3 Y Inoue,1
K Tomono,3 K Tsuyuguchi,1 S Shoji,4 T Saito5 1National
Hospital Organization Kinki-chuo Chest Medical Center,
Osaka, 2Genomidea Co., Osaka, 3Osaka University, Osaka,
4
National Hospital Organization Tokyo Hospital, Tokyo,
5
National Hospital Organization Ibaraki-Higashi Hospital,
Ibaraki, Japan. Fax: (81) 722 512 153. e-mail:
okm@kch.hosp.go.jp
Background: Multi-drug resistant (MDR), especially

extremely drug resistant (XDR), Mycobacterium tuberculosis is a big problem in the world. We have developed
novel TB therapeutic vaccine (HVJ-E/HSP65 DNA IL12 DNA vaccine) to eliminate MDR-TB.
Design/Methods: DNA vaccine (so called HSP65 vaccine) expressing M. tuberculosis heat shock protein 65
and IL-12 was delivered by the hemagglutinating virus of
Japan (HVJ)-envelope. Human M. tuberculosis was
intratracheally instilled into cynomolgus monkeys and
then the monkeys were treated with the vaccine i.m.
Results: HSP65 vaccine provided remarkable protective
efficacy and strong therapeutic efficacy against MDR-TB
and XDR-TB in murine models (prolongation of survival
time and the decrease in the number of MDR-TB).
Furthermore, we extended our studies to a cynomolgus
monkey model, which is currently the best animal model
of human tuberculosis. HSP65 vaccine provided therapeutic efficacy of prolongation of survival time (100%
survival) compared to saline control group (60%
survival) of TB infected monkeys. This vaccine augmented the immune responses (IL-2 production and the
proliferation of PBL) in TB-infected monkeys. In
addition, the IL-2 production from PBL correlated with
the survival. Therefore, the preclinical tests were studied
in the good laboratory practice (GLP) level. The injection
of 100lg of the vaccine /mouse i.m. three times in two
weeks was optimal for the production of IFN-c and IL-2.
The GMP-level DNA vaccine was made by constructing
the master cell bank. By using this vaccine, safety
pharmacology and toxicology tests are studying using
monkeys. High dose 5mg DNA vaccine/monkey (kg)
S164
showed little toxicity (body weight, food consumption,

hematology, blood chemistry). Furthermore, we have
planned to do clinical phase I trial. Targets are human
patients with MDR-TB. 8 weeks after 5 times injection of
this vaccine, the number of MDR-TB and sputum-culture
conversions of MDR-TB are defined as an efficacy
(primary efficacy of end point) of the vaccine. Adverse
events will be also examined.
Conclusion and recommendations: These data indicate
that our novel vaccine might be useful against tuberculosis including XDR-TB and MDR-TB for human
therapeutic clinical applications. (Co-workers: A Kumanogoh, A Mikami, T Matsumoto, Y Kita, S Hashimoto,
Y Nishida, H Nakatani, S Nishimatsu, Y Kioka, Seiji H,
P Saunderson, E V. Tan and D McMurray.)
PC-733-04 Universal access to the MODS assay

reducing the incidence of MDR-TB cases in
Arequipa, Peru
A Mendoza-Ticona,1 C Figueroa,1 M Perea,2 V Vargas,1
D Zavala,1 A Ortega,2 V Alarcon1 1Peruvian National
Tuberculosis Program, Lima, 2Regional Tuberculosis Program,
Arequipa, Peru. e-mail: mendozalberto@hotmail.com
Background: Arequipa is the second-largest city of Peru,

with a TB incidence of 48 cases per 100 000 inhabitants
in 2014. In 2008, the Peruvian Government implemented
the universal access to the MODS assay for rapid
detection of isoniazid and rifampicin resistance from
sputum samples. Objective: to evaluate the impact of
MODS universal access on MDR-TB incidence in
Arequipa
Methods: This is an operational research project. We
reviewed NETLAB and RME systems from the National
TB Program in Peru, looking for MODS assay production and results from the Arequipa Regional Laboratory
and the incident MDR-TB cases reported by Arequipa
Region between 2008 and 2014. We reported the number
of cases and the proportion of TB patient that were
MDR-TB per year. The trends were evaluated by v2 for
trends.
Results: There was a significant reduction of notified

MDR-TB cases in Arequipa Region (P 0.0043) after the
implementation of universal access of the MODS assay,
between 2008 and 2014, with lowest number being in
2014, with 5 cases (Figure).
Conclusion: MODS assay implementation as a universal

access method was a successful public health intervention
supported by the Government of Peru. MODS is currently
available in six Regional Public Health Laboratories in
Peru and is being rolled out in another seven regions.
PC-734-04 Modeling scenarios to determine the

optimal MDR-TB treatment regimens in India
L Smith,1 P Yadav,1,2 N Arinaminpathy,3 R Anupindi,2
B Balogh1 1William Davidson Institute, Ann Arbor, MI,
2
University of Michigan, Ann Arbor, MI, USA; 3Imperial
College London, London, UK. e-mail: lisacsmi@umich.edu
Background and challenges to implementation: In many

countries there is considerable variation in the multidrug
resistant tuberculosis (MDR-TB) regimens used for
treatment. In a given setting, the optimal MDR-TB
regimen requires consideration for recommended first
line drugs, drug resistance, and regimen success. In
addition to accurate transmission modeling, a model to
help determine optimal MDR-TB regimens needs to
incorporate important facets of the market and production economics for different product markets (e.g., drug
price, price-volume relationships). However data scarcity makes this exercise challenging and requires thoughtful consideration for specific scenarios of drug resistance.
Intervention or response: The authors built a mathematical model that optimizes regimen choice as shaped by a
variety of factors, including: the transmission implications of non-adherence (for example as driven by adverse
events); the profile of resistance in a population; the way
in which procurement cost scales with the numbers of
different drugs being procured; and the programmatic
costs of treating drug resistance (whether incurred or
averted). Our model examines the optimal MDR-TB
treatment regimen in India under several different
epidemiological and market scenarios. The objective
for the model is to minimize the cost per QALY gained
over a ten-year period, under different choices for how a
given budget is apportioned between different treatment
regimens.
Results and lessons learnt: This model highlights the
trade-off that while extensive regimen variation may
further fragment what are considered small-demand drug
markets, a certain degree of tailoring treatment regimens
aids in better adherence and individual treatment
outcomes, thus potentially reducing opportunities for
onward transmission. This decision space has become
increasingly important to consider with the inclusion of
increasing numbers of drugs, and combinations thereof.
Conclusions and key recommendations: The model will
provide National Tuberculosis Treatment Programmes
with decision support when considering modifications to
existing MDR-TB drug regimens. This model may inform
regimen design particularly in light of new product
introduction and improved access and use of drug
susceptibility testing. Procurement groups at the country
level may also use the described modeling approach to
inform purchasing decisions through an understanding of
both the market and public health impact.
05. Collaborative services for TB and

diabetes
PC-735-04 Body mass index and risk of
tuberculosis: analysis from two prospective
cohorts
H Fu,1 H-H Lin1 1National Taiwan University, Taipei City,
Taiwan. e-mail: b97801047@ntu.edu.tw
Background: Previous studies suggested that people with

higher-than-normal level of body mass index (BMI) had a
lower risk of tuberculosis compared with those with
normal BMI. Few of these studies controlled for
important confounding factors including tobacco smoking and alcohol use.
Design/Methods: We conducted two cohort studies in
Taiwan to investigate the association between different
levels of BMI and risk of incident tuberculosis. The first
cohort used adult participants from two rounds of
National Health Interview Surveys (NHIS) (n 30
486). The second cohort consisted of general population
from community-based health screening service (n 116
893). Baseline information on BMI and other covariates
was collected through structured questionnaires. The
incidence of tuberculosis was ascertained from the
National Tuberculosis Registry (NHIS cohort) and the
national health insurance database (screening cohort).
Cox proportional regression was used to estimate the
adjusted hazard ratio (aHR) and 95% confidence
intervals for different levels of BMI.
Results: After a median follow-up of seven years, 85 and
388 cases of active tuberculosis developed in the NHIS
and screening cohorts respectively. Compared with
normal BMI, underweight was associated with increased
risk of tuberculosis, and overweight and obesity were
associated with decreased risk of tuberculosis (Table).
The analyses were adjusted for age, sex, smoking,
alcohol use, diabetes, and education level.
Conclusion: Our studies confirmed the findings from
previous studies that high-than-normal level of BMI was
associated with decreased risk of tuberculosis. The
implication of global epidemic of obesity on tuberculosis
epidemiology, especially in the lower and middle income
countries should be further investigated.
Table. Association between BMI and risk of incident
tuberculosis
NHIS cohort (n30,486)
2
BMI (kg/m )
aHR
,18.5
18.5 - 22.9
23.0 - 24.9
25.0 - 29.9
7 30
1.45
1.00
0.40
0.29
0.41
95%CI
Screening cohort (n116,893)

BMI (kg/m2)
0.70-3.01 ,18.5
ref
18.5 - 24.9
0.23-0.71
0.16-0.53
25.0 - 29.9
0.13-1.31 7 30
aHR
95%CI
2.85 1.96-4.16
1.00
ref
0.39 0.30-0.50
0.17 0.08-0.36
S165
PC-737-04 Screening of tuberculosis patients

for diabetes in public-private partnership
projects in India
S Mugudalabetta,1 L Gujral,1 S K Muthusamy,1 R Begum1
Damien Foundation India Trust, Chennai, India. Fax: (91) 44
2836 0496. e-mail: admin@damienfoundation.in
1
Background and challenges to implementation: Diabetes

is emerging as a major threat to progress made in
tuberculosis control. Screening of all TB patients for
diabetes and ensuring strict blood glucose control is
important to improve the treatment outcomes. Revised
National Tuberculosis Control Program has issued
guidelines for screening and management of diabetes
among TB patients. The aim of the study is to share the
experience in screening of TB patients for diabetes in
projects managed under public private partnership.
Intervention or response: Damien Foundation India
Trust organized training for all the staff involved in TB
control program in Salem (Tamil Nadu), Nellore
(Andhra Pradesh) and South West Delhi. In 2014, after
a brief training all the TB patients registered in third and
fourth quarter were screened for diabetes using glucometer. At the time of registration, all the TB patients
were enquired about history of diabetes. TB patients who
were not aware of their diabetic status underwent
random blood glucose test. Patients with random glucose
value more than 110 mg/dl were screened with fasting
blood glucose. If fasting blood glucose was more than
126 mg/dl, patients were considered as suffering from
diabetes.
Results and lessons learnt: During the third and fourth
quarter 2014, 1533 TB cases were registered for
treatment. Among them 139 were found to be already
suffering from diabetes. Out of 1394 patients who didnt
know their diabetic status, 1347 (97%) underwent
random blood glucose test. It was found that 380
patients had random blood glucose level .110 mg/dl,
among them 322 were screened with fasting blood
glucose. This screening initiative was helpful in diagnosing 37 new diabetic cases among TB patients. Overall
11.5% of TB patients were found to be suffering from
diabetes. Most of the patients who already knew their
diabetic status were found to be taking treatment from
private sector, therefore coordination with them remains
as a major challenge.
Conclusions and key recommendations: Screening of
tuberculosis patients for diabetes is feasible. Future
research should focus on coordination with private
sector to improve management of diabetes.
S166
PC-738-04 Study of TB-diabetes collaborative

activities under the Revised National TB
Control Programme, Maharashtra, India
PC-739-04 Performance of laboratory glycated

haemoglobin (HbA1c) as a screening test for
diabetes mellitus in pulmonary tuberculosis
B Pawar,1 S Bharaswadkar,2 O Bera,3 S Kamble1 1State TB

Office, Pune, 2World Health Organization South East Asia
Office, New Delhi, 3International Union Against Tuberculosis
and Lung Disease, South-East Asia Office, New Delhi, India.
e-mail: bhagawanpawar.gbvb@gmail.com
S Imaculata,1 N Nathalia,1,2 S Mcallister,3

R Koesoemadinata,1 R Ruslami,1 B Alisjahbana,1,2 P Hill,3
R Van Crevel4 1Universitas Padjadjaran, Bandung,
2
Universitas Padjadjaran/ Hasan Sadikin General Hospital,
Bandung, Indonesia; 3Otago University, Dunedin, New
Zealand; 4Radboud University Nijmegen Medical Center,
Nijmegen, Netherlands. e-mail: sofia.imaculata@gmail.com
Background: Revised National TB Control Programme

(RNTCP) is implementing National strategic plan (NSP)
for universal access of TB care. Management of
comorbidities is a part of NSP. RNTCP has developed
framework for TB- Diabetes collaborative activities in
collaboration with National Programme for Prevention
and control of cancer, Diabetes, cardiovascular diseases
and stroke (NPCDCS). This study was undertaken with
the following objectives: To assess completeness of TBDiabetes collaborative activities in Central India; and to
assess outcome of NSP patients with respect to diabetic
status to ascertain impact of diabetes and treatment of
diabetes on TB treatment outcome.
Intervention: Record Review of TB patients registered in
5 districts of Maharashtra in India in year 2013. Records
like Treatment card, TB register and reports like
quarterly report of TB- Diabetes collaborative activities
were reviewed for the year 2013. Data entry and analysis
were performed in Excel.
Results: 7145 TB patients were registered in 5 districts in
year 2013. 1% (98) patients were known diabetic
patients. 59% (4082) TB patients were screened for
diabetes. 19% (784) TB patients had random blood
sugar more than 110mg/dl. 64% of 784 TB patients
underwent fasting blood sugar.136 (27%) of the patients
were diagnosed as diabetic. Total 231 Diabetic patients
were diagnosed in TB patients .206 (89%) of the patients
reached diabetic care. Review of treatment outcome of
New smear positive TB patients who are diabetic found
that conversion rate and success rate of New smear
positive TB patients who are diabetic is similar to the
success rate of new smear positive TB patients who are
not diabetic i.e. 88% each.
Conclusions and key recommendations: TB Diabetes
collaborative activities have been taken up well under
programme. This initiative may be expanded in entire
country. Coverage may be strengthened by improving
supervision. Establishment of district coordination committee for monitoring of activities may be considered for
the same. Good comparable outcome in between diabetic
NSP patients and non-Diabetic NSP patients draws
attention importance of complete treatment of both TB
and diabetes. Moreover it underlines significance of early
and prompt diagnosis of diabetes in TB patients.
Background: Screening for diabetes mellitus (DM) in

pulmonary tuberculosis (PTB) patients is an increasing
concern due to the rise of DM and PTB co-existence.
Fasting plasma glucose (FPG) and two hours postprandial plasma glucose, while commonly used, are less
practical for screening. Glycated haemoglobin (HbA1c)
should be considered as an alternative but its performance can be influenced by abnormal haemoglobin (Hb)
level. This study examined the anaemic status and
performance of HbA1c in PTB patients.
Design/Methods: Screening for DM among PTB patients, as part of TANDEM program (www.tandem-fp7.
eu), was performed in Bandung from December 2013 to
April 2015. Routine haematological examination, random capillary glucose (RCG) using AccuChek, laboratory HbA1c using high-performance liquid chromatography analyzer (NGSP standardized), and point-of-care
(POC) HbA1c using HemoCue HbA1c analyzer were
conducted. DM was considered if RCG was 711.1
mmol/l and/or HbA1c 76.5%, which was confirmed by
a FPG and/or second HbA1c measurement. Sensitivity
and specificity of laboratory HbA1c were calculated
separately.
Results: From 349 PTB patients screened for DM, 319
cases with complete data were included in this analysis.
Mean age was 41.1 (SD 14.3) years and 54.2% were
male. HIV co-infection and kidney disease were found in
7 (2.2%) and 10 (3.1%) patients, respectively. Anaemia
was found in 181 (56.7%) patients. Between those with
and without anaemia, there was no significant difference
in age (mean 40.5 vs. 41.8 years, P 0.405), sex (male
50.3% vs. 59.4%, P 0.104), ethnicity (Sundanese
90.6% vs. 83.3%, P 0.097), and HIV co-infection
(3.3% vs. 0.7%, P 0.118). DM was confirmed in 26
(14.4%) patients with anaemia and 36 (26.1%) patients
without anaemia. Sensitivity of laboratory HbA1c in
both anaemia and non-anaemia group was 100.0%,
whereas specificity was higher in the non-anaemia group
(98.1% vs. 100%). Positive predictive value (PPV) and
negative predictive value (NPV) of laboratory HbA1c in
the anaemia group were 89.6% and 100%, respectively.
In the non-anaemia group, PPV and NPV of laboratory
HbA1c were both 100%.
Conclusion: Regardless of the anaemia status of PTB
patients, laboratory HbA1c could be used as screening
test for DM. HbA1c provides high sensitivity and is a
practical tool as it does not need to be measured in a
fasting condition and the blood sample does not need to
be processed within two hours.
S167
PC-740-04 Age, body mass index and family

history of DM as predictors of DM for TB
patients
PC-741-04 Avances de la estrategia nacional de

integrada de la comorbilidad
atencion
tuberculosis-diabetes mellitus en Mexico
L Prasiddha,1 T Pakasi,1 I S Widyahening1 1Faculty of

Medicine Universitas Indonesia, Dki Jakarta, Indonesia.
e-mail: lorens.prasiddha@gmail.com
Y Davila,1 M A Garcia,1 M Castellanos1 1Ministerio de
Salud, Mexico
City, Mexico. e-mail: gpejazz@gmail.com
Background: Tuberculosis (TB) and diabetes mellitus

(DM) are diseases commonly found as comorbidities. TB
patients are known to be prone to DM. DM may cause
further complications in TB and therefore, to prevent
further complications, early screening of DM is required
for TB patients. One of several methods of screening is
the utilization of predictors of DM. This study aimed to
develop age, body mass index, and family history of DM
as predictors of DM among TB patients.
Design/Methods: This is a cross-sectional study among
TB patients registered in in primary health facilities and
Diabetes Center of Ternate, one of DM endemic areas in
Indonesia. All subjects were assessed for DM based on
Indonesian National Consensus of Diabetes Mellitus. We
took note of every subjects age and family history of
DM. Height and weight of subjects were measured using
the proper examination tool. The height and weight data
were used to calculate the subjects body mass index.
Data of subjects age, family history of DM, and body
mass index were analyzed as individual and combination
predictors of DM among TB patients.
Results: Among 55 subjects of TB patients, 31 subjects
(56.4%) were diagnosed with DM whereas 24 subjects
(43.6%) were excluded from diagnosis of DM. TB
patients aged over 44 years higher risk (P , 0.001, RR
3.8) over TB patients aged below 44 years to develop
DM. TB patients with body mass index ,18.5 kg/m2
years had higher risk (P 0.004, RR 3.2) over TB
patients with body mass index 718.5 kg/m2 to develop
DM. TB patients with positive family history of DM had
higher risk (P 0.035, RR 1.9) over TB patients with
negative family history of DM to develop DM. TB
patients aged over 44 years having body mass index
718.5 kg/m2 had higher risk to develop DM (P 0.038,
RR 2.0). TB patients aged having body mass index
718.5 kg/m2 with positive family history of DM had
higher risk to develop DM (P 0.049, RR 1.8). Age
and family history of DM did not have significant
relationship as predictors (P 0.277).
Conclusion: This study concludes that age, body mass
index, and family history of DM were available as
predictors of DM for TB patients. Combination of age
and body mass index and also body mass index and
family history of DM were highly suggested as combination predictors of DM for TB patients. TB patients
aged over 44 years, body mass index ,18.5 kg/m2, or
positive family history of DM are more susceptible to
DM and therefore should always be screened for DM.
Background and challenges to implementation: La

Tuberculosis (TB) y la Diabetes Mellitus (DM) representan dos problemas de salud publica,
cada uno constituye
un desafo para su abordaje y control. En Mexico, 21%

de los casos de TB estan asociados a la DM. De 2000 al
2013 se registro incremento de 208%. Durante el 2014 se
identificaron 4,269 casos nuevos con TBTF y DM. Desde
el ano
2010 el Programa Nacional de Tuberculosis y el
Programa Nacional del Adulto y Anciano han colaborado para impulsar estrategias que coadyuven en el
control de la comorbilidad TB-DM. En 2011 la OMS y
la Union
Internacional contra la Tuberculosis y Enfermedades Respiratorias reconocieron la necesidad de
guas internacionales sobre el manejo conjunto de la
TB y la DM. Asimismo puntualizo que el establecimiento
de mecanismos de colaboracion
entre los programas
nacionales de TB y DM es fundamental para el
seguimiento y control de ambas patologas.
Intervention or response: Implementacion
de la propuesta mexicana para la Estrategia Nacional de atencion
integrada TB-DM. Desde 2014 se incluyo el indicador

de desempeno
nacional del programa de
en la evaluacion
TB. Deteccion
de DM en pacientes con TBTF de 20 anos
y mas de edad, aplicable para todas las instituciones del

Ministerio de Salud. Se emitieron recomendaciones de
atencion
integrada para garantizar el seguimiento y
control glicemico en las personas con TB-DM. Realizar
glicemia capilar de manera semanal, glicemia central de
manera mensual y hemoglobina glicosilada de manera
trimestral. Se elaboraron algoritmos de atencion
integrada para la deteccion,
monitoreo y control de las
personas con TB-DM.

Results and lessons learnt: - Mejora en la cobertura del
indicador de desempeno:
La deteccion
de DM en
personas con TB fue de 83.1% para el tercer trimestre

de 2014 en todo el pas, el ano
previo la actividad alcanzo
el 72.7%. -Aumento en el e xito terapeutico
de la cohorte
de tratamiento primario TB-DM de enero a junio 2014:

87.9%. El resultado de la corhorte anterior fue de:
86.2%. -Fortalecimiento de la atencion
integrada de la
comorbilidad TB-DM.
Conclusions and key recommendations: Mexico ha
avanzado en la recomendacion
internacional de integracion
de servicios para personas con TB y DM al ofertar el
tamizaje bidireccional e impulsar algoritmos de atencion
integrada de la comorbilidad. La atencion

integrada de la
comorbilidad TB-DM favorece la deteccion
oportuna de
ambos padecimientos, as como la adherencia terapeu
tica y la curacion.
S168
PC-742-04 Life after TB: a systematic review of

imaging defined post-TB lung disease
06. TB surveillance from Swaziland to

Japan
J Meghji,1 H Simpson,2 K Mortimer,3 S B Squire3

1
University of Liverpool, Liverpool, 2London School of
Hygiene & Tropical Medicine, London, 3Liverpool School of
Tropical Medicine, Liverpool, UK. e-mail:
jamilah.meghji@gmail.com
PC-743-04 District-level performance

monitoring improved the accuracy of TB
program reporting in Ethiopia
Background: Pulmonary and pleural TB cause changes in

lung structure and function that persist after treatment
completion. The pattern and prevalence of structural
damage, its relation to airway physiology, and the
associated morbidity are poorly defined. A systematic
review was conducted to identify studies describing the
prevalence of imaging defined (CT/CXR) post-TB lung
disease.
Design/Methods: A protocol driven literature search was
used to identify studies of pulmonary or pleural TB using
CXR or CT imaging. 2 reviewers selected English
language studies in which consecutive patients were
imaged after completion of TB treatment, and the
prevalence of abnormal imaging or severity of residual
disease were reported. Studies of children and those with
non-HIV immunosuppression were excluded. Additional
articles were found using reference and citation reviews.
Data were extracted and study quality was assessed.
Results: Of the 4495 articles identified, 35 were selected.
These were published between 1973-2014 with data for
4273 patients, and included 6 CT studies. Imaging was
performed at treatment completion in 19/26 studies of
pulmonary TB sequelae. The prevalence of cavitation
was higher in studies of multi-drug resistant (MDR)
disease (51.5-69.7% vs. 7.4-55.0%). The prevalence of
fibrosis (70.0-92.6%), bronchiectasis (35.0-85.2%), and
emphysema / bullae (15.4-45.0%) was high on CT
studies. Residual pleural thickening .10mm was seen in
19.6-46.0% of patients on treatment completion in 9
studies of pleural disease. Only 2 studies from subSaharan Africa (sSA) were found, both of which were
from South Africa. Disaggregated data were available for
96 HIV-infected people. Only 1 study included serial
imaging. 3 studies described the relationship between
imaging findings and spirometry, and 2 described the
relationship with morbidity: impairment of functional
capacity and symptom burden were greater in those with
abnormal imaging.
Conclusion: A high burden of structural lung damage
persists after TB disease, for which CT is more sensitive
than CXR. The high prevalence of damage after MDR
disease is a concern for areas with high rates of drug
resistance. There is a paucity of data from sSA and in
HIV-infected patients. Phenotyping of post-TB lung
disease along multiple respiratory parameters is needed
to optimise patient management. Further investigation of
the evolution of lung disease over time, its associated
morbidity, and its impact on patients lives is needed.
D Habte,1 N Demmelash,1 G Negussie,2

I Jemal Abdulahi,3 D J Dare,1 M Aseresa Melese,1
Y Haile,4 P G Suarez5 1Management Sciences for HealthHelp Ethiopia Address the Low TB Performance (HEAL TB)
project, Addis Ababa, 2Amhara Regional Health Bureau,
Bahar Dar, 3Oromia Regional Health Bureau, Addis Ababa,
4
United States Agency for International Development
(USAID), Addis Ababa, Ethiopia; 5Management Sciences for
Health, Center for Health Services, Arlington, VA, USA.
e-mail: dhabte@msh.org
Background and challenges to implementation: Health

programs are highly dependent on data for planning,
implementation, monitoring and evaluation purposes.
The USAID-funded Help Ethiopia Address the Low TB
Performance (HEAL TB) has been supporting the
Amhara and Oromia Regions of Ethiopia in TB program
including data quality.
Intervention or response: HEAL TB project implemented
a standard of care performance monitoring tool to track
the performance of health facilities in relation to TB
program implementation as well as the accuracy of
selected TB program indicators. District TB focal persons
visited health facilities in their jurisdiction on quarterly
basis to validate the accuracy of the TB program report
submitted by the health facilities against the re-count
made on the records kept at the health facilities. The
district TB focal persons provided feedback on data
quality issues based on the consistency check and decided
on improvement plan whenever there is a problem of
data quality. We present the trend of data quality on 3
selected indicators in health facilities in 10 zones of
Amhara and Oromia from 2011-2014.
Results and lessons learnt: In the 10 zones, 691 health

facilities have been supported with a comprehensive
package of TB program support. Over 85% of the health
facilities were mentored each quarter by the district TB

focal person where data quality check was also conducted. The accuracy of reports (reported and re-counted) in
the health facilities improved significantly over the three
years: 80.6% (95%CI: 77.5-83.4%) to 97.7% (95%CI:
96.3-98.6%) for all forms of TB; 86.6% (95%CI: 83.888.9%) to 92.6 % (90.4-94.3%) for TB cure rate; and
80.8% (95%CI: 77.7-83.5%) to 97.2% (95%CI: 95.898.2%) for HIV testing among TB patients (P , 0.05).
Conclusions and key recommendations: The use of
routine standard of care performance monitoring and
onsite feedback significantly improved the quality of
data and reports at health facility levels. The data quality
improvement is also attributed to a strong program.
PC-744-04 Surveillance system evaluation of

the integrated disease surveillance project
(IDSP) in Shimla district, Himachal Pradesh,
India
O K Bharti1 1State Health and Family Welfare Training Centre
Parimahal, Shimla, India. e-mail: bhartiomesh@yahoo.com
Background: The major objective of IDSP is to recognize

early warning signals of impeding outbreak and help
initiate an early response in a timely manner. A project to
study the disease surveillance project in reference to
tuberculosis was undertaken to know the causes of
increase in the incidence of tuberculosis
Design/Methods: A protocol for evaluation was prepare
and shared with all the stakeholders to get their
involvement. We mailed self administered questioners
to various health functionaries and also had informal
discussions with various stakeholders in the district. We
also assessed the level of integration as well as sensitivity
of IDSP with respect to TB as component of Revised
national tuberculosis programme, RNTCP. We entered
the data in Excel software and calculated various
indicators as proportions using Epi Info.
Results: The prevalence of TB in the district is 257/Lakh
population and is higher than the state prevalence of 210/
lakh population. 208 of 275 health workers in place
responded to the questioner with a response rate of 78%.
IDSP reported 249 cases of cough more than 3 weeks out
of that only 131 were sputum positive, while those
reported by RNTCP for similar period was 2333 chest
symptomatic screened and 263 cases were confirmed,
and the difference was significant. (c2 297; Df 1, P
0.00). This translates into sensitivity of 10% of suspect
cases and 50% of confirmed cases. Positive predictive
value is only 20% at MPW level and only 57% reports
are timely. 52 out of total 90 posts of lab technicians
(58%) are vacant in the district, even 7 DMCs are
without lab technicians. Mean population served by a
single lab is 69203/- 162458. Vacancy of MPW is
related to the cure rate of TB that is as the vacancy
increases cure rate decreases and is inversely correlated (r
0.99). This is due to the fact that MPWs are also DOTS
providers. Only 33% of the health workers are able to
visit their villages in a week. 59% of them said that
S169
people try to hide the disease even from them due to

prevalent stigma in the community.
Conclusion: There was mismatch between cases reported
by IDSP and those reported by official channels of the
Revised National Tuberculosis Programme (RNTCP). To
fill all the vacancies and to have more involvement of
private sector and medical colleges is the key to have a
sensitive IDSP and better RNTCP. Need to have less area
for monitoring and some mobile allowance for MPWs to
SMS data to higher authorities.
PC-745-04 Accelerating TB notification among

private providers using ICT based applications
in Delhi, India
D Kundu,1 K Chopra,2 A Khanna,3 T J Padmini,3 B Neeti3
World Health Organization Country Office for India, Delhi,
2
New Delhi Tuberculosis Centre, Delhi, 3Directorate of Health
Services, Delhi, India. e-mail: debashishkundu@yahoo.com
1
Background: In India almost half of all patients with

Tuberculosis (TB) seek care in the private sector as the
first point of care. The National Program is unable to
support TB patients and facilitate effective treatment as
there is no information on TB and M/XDR-TB diagnosis
and treatment in private sector. To improve this situation,
Government of India declared TB a notifiable disease for
establishing TB surveillance system, to extend supportive
mechanism for TB treatment adherence and standardised
practices in the private sector. But, TB notification from
the private sector is a challenge and still a lot needs to be
done to accelerate TB notification.
Intervention or response: Delhi State TB Control
Programme has taken special initiatives for accelerating
notification of TB cases from the private sector in 2014
by constituting a state level TB notification committee,
collaborating with the interface agencies like Delhi
Medical Association (DMA) for placing TB notification
advertisements periodically in their news bulletin and
Initiative for Promoting Affordable Quality TB Tests
(IPAQT) to facilitate notification from the IPAQT
partner laboratories. A most important initiative was
direct one to one sensitization of private providers (PP)
by state level TB notification teams for providing options
on notification modalities using ICT based applications Nikshay (Case-Based Web Online application) and
convenient web-login or mobile apps for the purpose of
direct notification of TB cases in Nikshay TB notification
portal [Table]. A total of 102 PPs were sensitised on TB
notification modalities, of which 40 PPs were also given
live on-line demonstration using web-login. Eight private
laboratories were sensitised in Nikshay and direct weblogin process.
Results and lessons learnt: 83% (85/102) of PPs post
sensitisation started to notify TB cases in Nikshay and
total of 2837 TB cases were notified by the private sector.
38% (15/40) of PPs directly notified 110 TB cases using
web-login access and 100% (70/70) PPs notified 2727
TB cases in Nikshay using active support from the field
staff, including of IPAQT.
S170
Conclusions and key recommendations: Direct one to

one sensitisation of the PPs can accelerate TB notification, especially, with the help of active support from the
field staff in entering the notified TB cases using ICT
based applications. State level initiatives and collaborations with interface agencies can improve awareness on
TB notification.
Table Notification by private providers (PPs) based on
notification modalities using ICT based applications Nikshay, Direct Web-Login or Mobile Apps
No. of
No.
PPs
Sensistarted
No. tised to noSensi- One
tify
TB
tised - to
No. using
NotifiOne One
of active
cation
to using
PPs supfrom
One webstart- port
active
using login
ed to from
supNik- and
di- the
TB port
shay Nikrect- field To- Notifi- from
and shay
ly staff tal cation the
Guid- Guid- To- noti- (TBH- PPs using proance ance tal fy
Vs star- Web- gramTool Tool PPs using and ted Login me us- Total
TB
on TB on TB Sen- web- Dis- to directing
Type of Notifi- Notifi- si- log trict No- ly by web- NotifiHealth cation cation tised in DEO) tify PP login cation
[h]
Facilities [a]
[i]
[b] [c] [d] [e] [f] [g]
PPs Clinic
(Single) 24
14
PPs Hospital
(Multi) 38
18
PPs Laboratory
8
8
Total
70
40
Proportion of PPs
started to notify
post-sensitisation
38 4
24
28
12
532
544
56 11
38
49
98
897
995
8 0
102 15
8
70
8
85
110
1298 1298
2727 2837
38% 100% 83%
PC-746-04 Recommendations for TB intensified

case finding based on national TB prevalence
survey results in Indonesia
B Dwihardiani,1 D B Lolong,2 l Pangaribuan,2
O S Simamarta,2 M Farid,3,4 F Ahmadi,2 T N Dinihari,e
C Widaningrume 1World Health Organization, Indonesia
Country Office, Jakarta Pusat, 2Ministry of Health Republic
Indonesia, Jakarta Pusat, 3National Bureau of Statistics,
Jakarta Pusat, 4Tuberculosis Operational Research Group,
Jakarta Pusat, Indonesia. e-mail:
bintari_dwihardiani@yahoo.com
Background: As a country with TB burden among the

highest in the world, Indonesia needs to innovate its TB
control strategies. The national TB prevalence survey of
2013-2014 found that TB prevalence was higher than
previously estimated. Intensified case finding (ICF) in key
population of TB is one of the strategies proposed to
reduce significantly the TB incidence according to the
country target. National TB prevalence survey serves as a
model of intensified case finding.
Design/Methods: National TB prevalence survey was

conducted based on WHO guidelines. In this nationwide
survey, X-ray and chronic cough were used as screening
criteria for sputum submission. Smear microscopy and LJ
culture were conducted to all positively screened
participants and Xpert MTB/RIF was used to confirm
positive smear microscopy and inconclusive culture.
Analysis on the survey results were conducted to
understand key populations and interventions related
to intensified case findings.
Results: Among 76 576 eligible individuals, 67 946
(88.7%) participated in the survey. Percentage of TB
cases among participants with positive screening symptoms and positive screening X-ray results were 2.74 and
3.58% respectively. Among those with negative screening symptoms, 2.78% of positive screening X-ray
participants were confirmed TB cases. Positive X-ray
screening without positive screening symptoms contributed to 44% of bacteriologically confirmed cases. Higher
TB cases found among participants with previous TB
history, had close contacts with TB patients, and current
smoker. TB prevalence was higher among men and older
age, however the burden is still very high among
productive age population. Prevalence per notification
ratio is high nationally, and among men and elderly
population. Among positive smear specimens from 292
participants, only those from 164 (58%) participants
were TB confirmed by Xpert or culture LJ. About 52.4%
participants who had non TB smear positive reported
chronic cough or hemoptysis.
Conclusion: ICF which targets specific key populations
with difficult access and incorporating X-ray screening
may contribute to increased case detection. In the setting
of Indonesia, using more sensitive and specific diagnostic
test during ICF may reduce false positive of smear
microscopy.
PC-747-04 Age progression among notified

tuberculosis patients and its implications in
Japan
A Ohkado,1,2 K Uchimura,1 S Yoshimatsu,1 K Izumi,1,2
L Kawatsu,1,2 K Ito1,2 1Research Institute of Tuberculosis /
Japan Anti-Tuberculosis Association, Kiyose, 2Nagasaki
University Graduate School of Miomedical Sciences,
Nagasaki, Japan. Fax: (81) 424 93 5529. e-mail:
ohkadoa@jata.or.jp
Background: The notification rate of tuberculosis (TB) in

Japan has been declining, to as low as 16.1 cases per 100
000 population in 2013. While the rates among patients
aged 765 years (the elderly) have been consistently high,
the older age group, e.g. 785 years, shows the higher
rates. The present study aimed to describe the current age
progression among TB patients in Japan and its
implications to TB control.
Design/Methods: This study is a descriptive summary of
publicly available TB surveillance data in 2013 that
focused on age structure change. The annual trend of TB
notifications and proportions by age group and patient
characteristics upon registration such as bacteriological
test results, chest radiography (CXR) findings, symp-
toms, diagnosis delay, and time of death during the

treatment were summarised. The v2 test for categorical
data was applied to compare proportions.
Results: In 2013, the proportion of pulmonary TB (PTB)
patients with bacteriologically positive results was higher
(89.9% vs. 77.5%, P 0.00), but that of patients with
cavitary lesions on CXR was lower (29.1% vs. 35.0%, P
0.00) among those aged 765 years than among those
aged 1564 years. Among all the symptomatic PTB
patients, the proportion of those with non-respiratory
symptoms only was higher (26.5% vs. 18.0%, P 0.00)
in those aged 765 years, increasing with age from
21.6% in those aged 6575 years to 30.3% in those aged
785 years. The proportion of TB patients with patient
delay (duration between onset of symptoms and the first
consultation visit) of 72 months was lower (14.2% vs.
25.6%, P 0.00), whereas that with doctor delay
(duration between the first consultation visit and TB
diagnosis) of 31 month was slightly higher (23.0% vs..
20.3%, P 0.00) among the patients aged 765 years
than among those aged 1564 years. Among the TB
patients aged 765 years who were notified in 2012,
32.2% died within 1 year after the initiation of TB
treatment.
Conclusion: The elderly tended to have a lower
frequency of respiratory symptoms, to have a late TB
diagnosis after the first consultation visit, and to die at
the early stage of TB treatment compared with the
patients in the other age groups. Health-care workers
and those in other sectors, including the welfare service,
need to be aware of the characteristics of the elderly at
the time of TB diagnosis and throughout the treatment
course, not only in Japan but also in countries that
experience age progression among TB patients.
PC-748-04 Use of the national electronic

database to map TB patients with transfer out
outcomes in Kenya
J Kiarie,1 D Mibei1 1Ministry of Health, Nairobi, Kenya.
e-mail: jckiarie@gmail.com
Background: Tuberculosis being a disease of public

health concern has further been complicated by the
emergence of MDRTB. There has been lots of effort
globally to fight the disease and Kenya has made much
progress in this. TB patients in Kenya are managed
according to WHO guidelines. While the TB program
recommends follow up, its hard to control movement of
patients within the country and across borders. Patients
who move to another district permanently to continue
with treatment are allocated a transfer out outcome.
Therefore, proper referral for these patients by the
facility of origin is vital to control default rate, treatment
interruption, disease transmission and risk to MDRTB.
Transfer out outcome is inevitable since patients move
due to various reasons during their treatment period
which impacts negatively on TB treatment success rate.
Among adverse TB treatment outcomes, Transfer outs
are low hanging fruits since a follow up can be done at
the facility of origin to improve treatment success rate
S171
hence the need to assess whether transfer out TB patients

arrive at their destination facilities
Design/Methods: This study was done to determine the
proportion of smear positive patients with a transfer out
outcome who could be matched in the national database,
determine the treatment outcomes among patients who
recorded a transfer out outcome and establish gaps in
documentation of data for patients who were transferred
out. A cross sectional descriptive study was done using
secondary data from the national electronic database to
match patients with a transfer out outcome using Names,
Age, Type of TB, Smear results at month 0, HIV status
and date of treatment initiation variables
Results: From 366 smear positive patients registered in
quarter 1 2013, most cases with transfer out outcomes
were from the prison sector, 10% under nutrition cases,
10% HIV- and 13% those not initiated on ARVs. 27% of
the total of cases was found with 76% of them having a
cured treatment outcome. Key gaps identified included
documentation of names and patient type classification.
After incorporating the mapping results the National
treatment success rate increased by approximately 2%.
Conclusion: There is need to strengthen the patient
referral system, capacity building to health care workers
and standardize documentation procedures to improve
TB treatment outcomes in Kenya
PC-749-04 Swaziland active surveillance and

patient safety monitoring for patients on ARVs
and TB drugs
K Kunene,1 N Shongwe2 1SIAPS, Management Sciences for
Health, Mbabane, 2Ministry of Health, Mbabane, Swaziland.
Fax: (268) 2404 2787. e-mail: kkunene@msh.org
Background: Swaziland remains one of the countries

with the highest dual human immunodeficiency virus
(HIV) and tuberculosis (TB) burden globally. The HIV
incidence rate is 26% and there are around 1382 per 100
000 incident TB cases annually. Exacerbating this
challenge is the TB-HIV co-infection rate that remains
high at around 80% and 66% of the co-infected patients
are receiving treatment for both. The continuous
introduction of new essential medicines necessitates the
concurrent monitoring of adverse effects and safety of
the patients on antiretrovirals (ARVs) and TB medicines.
As such, the Ministry of Health (MOH), supported by
the USAID-funded Systems for Improved Access to
Pharmaceuticals and Services program (SIAPS) implemented an active pharmacovigilance and patient safety
monitoring system for patients on ARVs and TB
medicines.
Design/Methods: SIAPS supported the MOH to develop
the protocol and tools to operate the Sentinel Site-based
Active Surveillance System for Antiretroviral and AntiTB (SSASSA) electronic system. The cohort event
monitoring was introduced at 5 pilot hospitals in June
2013. Clinicians enrol (and follow up) treatment nave
HIV patients and TB patients starting a new regimen.
Relevant patient information is captured onto the system
by data clerks. Data is collected monthly and analysed
S172
centrally using the electronic data collection and analysis

tool. Supportive supervisory visits are conducted every
month. Data is disseminated quarterly through a
newsletter.
Results: A total of 2001 patients have been enrolled from
June 2013 to March 2015 (67% females and 33% males)
and there have been 936 adverse events (ADRs) reported.
Peripheral neuropathy (15%) and rash (10%) are the
most common, with kidney failure, ototoxicity and death
being the most severe. ADRs accounted for 2% of 956
regimen switches. The causality assessment showed that
90% of the reported ADRs were probably or possibly
caused by the medicines using the WHO and Naranjo
causality assessment methods.
Conclusion: The active surveillance system is systematically documenting and quantifying the incidence rate of
adverse events associated with ARVs, 1st and 2nd line
anti-TB medicines. In addition, the active surveillance
system is generating local data to provide better estimates
of risk-benefit profiles and help prevent and minimized
such risks at sentinel sites. The MOH is in the process of
expanding the active surveillance system to four more
regional hospitals.
PC-750-04 Evaluation of the tuberculosis

surveillance system in Shamva District,
Zimbabwe, 2014
used in any way at their health facility. Only 19.5% had

access to TB screening flow chart. 85% of the healthcare
workers were willing to complete TB notification forms.
The median time for sample to reach the lab was 1 day; a
result to reach health facility was 3 days and monthly
report to reach district was 8 days. There was integration
of HIV-TB activities in the district. Majority of health
facilities did not have computers, internet access, faxes
and landlines. Shortage of various TB surveillance tools
was reported.
Conclusion: Knowledge levels were poor on the objectives of the surveillance system. The TB surveillance
system was not simple, stable and timely but was useful,
flexible, and acceptable. Mentoring, training and capacity building of healthcare workers in TB surveillance and
provision of surveillance tools will improve the system
performance.
07. Tuberculosis: from the bench to the

grave
PC-751-04 Recurrence rate of tuberculosis and
mortality among smear-positive patients in
Viet Nam
R Makurumidze,1 N Gombe,1 D Bangure,1

M Tshimanga,1 T Magure2 1University of Zimbabwe,
Harare, 2National AIDS Council of Zimbabwe, Harare,
Zimbabwe. e-mail: ricmanie2000@yahoo.com
G Fox,1 N Nguyen,2 T A Nguyen,1 T L Phan,1 B H Nguyen,2

N A Le,1 W Britton,3 G Marks1 1Woolcock Institute of
Medical Research, Sydney, NSW, Australia; 2National Lung
Hospital, Hanoi, Viet Nam; 3Centenary Institute of Cancer
Medicine and Cell Biology, Sydney, NSW, Australia. Fax: (61)
2 9114 0010. e-mail: foxsimile@gmail.com
Background: A review of tuberculosis (TB) surveillance

data for 2013 in Shamva District showed shortcomings.
An average of 10 health facilities out 15 was reporting
TB data on time. There was incomplete and incorrect
filling of TB surveillance tools. We therefore evaluated
the tuberculosis surveillance system in the district.
Design/Methods: A descriptive cross sectional study
was carried out using Centre for Disease Control
Guidelines on Evaluation of a Public Health Surveillance
System. Study participants were healthcare workers and
key informants. Data was collected using questionnaires
from healthcare workers and interview guide from key
informants. Quantitative data was analyzed using EpiInfo. Permission to carry out the study was sought from
the relevant authorities. All ethical considerations were
duly observed.
Results: 41 healthcare workers were interviewed. Females constituted (70.3%) and males 29.3%. Median
age was 39 years (Q1:33; Q3:45) and median length of
service was 6 years (Q1:4; Q3:9). The majority of the
healthcare workers were primary care nurses (43.9%).
Only 22 % were trained in TB disease surveillance.
Knowledge levels on TB symptoms and signs were above
90% and for objectives of the surveillance system was
between 60 - 90%. Lack of transport; lack of stationery;
delay in compilation; negative attitude and motivation
were mentioned as reasons for reporting delays. Thirty
three (80.5%) mentioned that the TB data collected is
Background: The recurrence rate of tuberculosis (TB) is

an important indicator of the effectiveness of a TB
control program. Throughout Vietnam, 7.1% of all
notified TB patients treated by the National TB Program
(NTP) are reported to develop recurrent TB. This may
underestimate the true recurrence rate, as it excludes
patients lost to follow-up. We aimed to determine the TB
recurrence rate and mortality among patients with smear
positive TB treated by the Viet Nam National Tuberculosis Program (NTP) in the first 24 months or more of
follow-up.
Design/Methods: This prospective cohort study, with
nested case-control study, was conducted in 70 districts
selected at random from eight provinces in Viet Nam,
with the probability of district selection in proportion to
the total population. Patients with smear positive
pulmonary tuberculosis, enrolled in a cluster-randomized
controlled trial of household contact investigation, were
re-traced for between two and four years after treatment
initiation and recurrent episodes of TB and mortality
were documented. In addition, all patients with incident
TB, and a random sample of healthy contacts, completed
a more detailed questionnaire.
Results: 11 200 patients with smear positive TB were
enrolled between 1st September 2010 and July 31st
2013. Of 9331 index patients enrolled, 8480 (90.9%)
were re-traced and 851 (9.1%) were lost to follow-up.
Median follow-up for patients was 34 months (inter-
quartile range 28 40 months). Among 7327 patients

who completed treatment, 510 (7.0%, 95%CI 6.4
7.5%) developed recurrent TB during follow-up. Of
these, 66 individuals died as a result of TB. Among retraced patients, 895 (10.6%, 95%CI 9.9-11.2%) died
during the follow-up period, including patients died with
recurrent TB. The age and gender Standardized Mortality Ratio for all patients compared to their household
contacts was 5.
Conclusion: The recurrence rate of TB among smear
positive patients was similar to official estimates. Further
research is required to identify factors associated with
recurrence and mortality, in order to strengthen TB
control in Viet Nam.
S173
and the lowest in Ponta Negra (1.18 deaths/100 000

inhabitants per year) (Figure 1).
Conclusion: The study revealed that deaths occurred
mostly in males, with working ages and pulmonary TB.
In addition, there was a heterogeneous distribution of
deaths in space, appearing more susceptible or vulnerable
areas to the event than others, which suggests those
places prioritization in terms of disease control measures.
PC-752-04 Deaths due to tuberculosis in Natal,

Rio Grande do Norte, Brazil, 2008-2014
A Queiroz,1 M C C Garcia,1 L H Arroyo,1 A Belchior,1
J Crispim,1 M Touso,1 M Popolin,1 R Arcencio1 1Ribeirao
Preto School of Nursing, University of Sao Paulo, Ribeirao
Preto, SP, Brazil. e-mail: aninha_arego@hotmail.com
Background: Despite the advances in prevention and

treatment people are still dying due to tuberculosis (TB)
around the world, mainly in 22 High-TB Burden
countries. The World Health Organization has just
approved the ambitious goal of reducing 95% of TB
deaths by 2015. It is necessary to know why and where
these people die due to TB, in Brazil. Thus, this study
aimed to characterize the clinical and epidemiological
profile of patients who died due to TB and to identify the
areas where these events occurred more frequently.
Design/Methods: Descriptive and ecological research
developed in Natal - Rio Grande do Norte Brazil, which
has a population of 817 590 inhabitants (2010), and it is
considered a priority city for TB control with an
incidence of 46.6 per 100 000 inhabitants. It was
possible to include all cases of TB death occurred during
the period 2008-2014 obtained through the Mortality
Information System (MIS), whose primary and associated causes were pulmonary and/or extrapulmonary TB
with or without bacteriological confirmation. Descriptive statistics was applied with calculating measures of
central tendency for continuous and categorical variables, absolute and relative frequency counts. TB
mortality rates were estimated by neighborhood. Corochromatic maps were developed by using the free
software Quantum GIS (QGIS) 2.6.1 in order to identify
sites with the highest occurrence of deaths.
Results: 235 cases of TB death were identified. 169
(71.9%) were male, with a median age of 51 years; 125
(53.1%) were race/brown color. Concerning the clinical
form, 196 (83.4%) were pulmonary ones, and it was
possible to note that 203 cases (86.3%) died at the
hospital. HIV/AIDS appeared as the principal cause of
death in 48 cases (58.5%). Regarding the mortality rate,
16 neighborhoods showed higher coefficients than the
municipality (4.17 deaths/100 000 inhabitants per year).
The highest coefficient was in the neighborhood of
Santos Reis (17.42 deaths/100 000 inhabitants per year),
PC-753-04 Excess mortality of tuberculosis

patients in Shanghai, China: a 10-year followup study
W Wang,1 Q Zhao,1 B Xu1 1Fudan University, Shanghai,
China. Fax: (86) 215 423 7811. e-mail: wwb@fudan.edu.cn
Background: China has the second highest tuberculosis

(TB) burden worldwide, and the premature mortality of
TB patients makes this a critical issue.
Methods: The study cohort was drawn from residents in
4 of 19 districts in Shanghai City. All subjects in the
cohort were registered with TB between January 1, 2004
and December 31, 2008. Baseline data were collected at
the most recent TB diagnosis, and patient mortality was
assessed between March and May of 2014.
Results: Overall, the cohort had a standardized mortality
ratio (SMR) for all-cause deaths of 5.2 (95%CI: 4.8-5.6).
Males had a higher SMR than the females (6.1 vs 3.0).
Patients with hemo-disseminated pulmonary disease had
an SMR of 18.5, and those with secondary pulmonary
disease had an SMR of 5.3. Patients living with other
comorbidities also had high SMRs, especially those with
chronic obstructive pulmonary disease (22.6) and cancer
(30.1). The 1-, 5- and 10-year cumulative case fatality
rates were 7.48%, 17.20%, and 21.23%, respectively.
With adjustment for age and sex, all-cause hazard ratios
(HRs) were significantly greater previously treated
patients (HR 1.26, P 0.005) and smear-test positive
patients (HR 1.55, P , 0.001); the risk to TB-caused
deaths were also significantly greater for previously
treated patients (HR 1.88, P 0.003) and smear
positive patients (HR 3.16, P , 0.001).
Conclusion: Numerous groups of TB patients from
Shanghai have increased risk of mortality. Identification
S174
and more vigilant post-treatment follow-up of these

high-risk patients may help to reduce mortality.
PC-754-04 Analysis of factors associated with

overdiagnosis of tuberculosis
Z Laushkina1 1Novosibirsk Research TB Institute,
Novosibirsk, Russian Federation. Fax: (7) 383 203 8675.
e-mail: zlaosh@list.ru
Background: Differential diagnosis of pulmonary TB is

still difficult. We found the tendency to overdiagnosis of
pulmonary TB in TB hospital.
Design/Methods: Noncomparative retrospective study,
subject of interest were medical records of 230 most
difficult cases, in which previously established diagnosis
of pulmonary TB was rejected. The chances for
establishment of true diagnosis and influencing factors
were estimated. Odds ratios (OR) and nominal 95%
confidence intervals were presented.
Results: Males 81 %, mean age 48.3616.5 yrs. Diseases,
most often initially misdiagnosed as tuberculosis, were
pneumonia (46 %), lung cancer (24 %), and sarcoidosis
(15 %). Clinico-radiological signs in these cases were
more characteristic for TB, than for other pulmonary
diseases. 49 % of patients were previously treated of an
assumed pneumonia before hospitalization in TB hospital, 36 % of patients up to making a hospital diagnosis
received TB treatment. All patients have been hospitalized with wrong diagnosis pulmonary TB, a principal
cause misinterpretation of chest radiogram. In 5 % of
all cases a few acid-fast bacilli were found in sputum by
luminescent microscopy. Factors found to be associated
with false-positive diagnosis of TB are: old age (OR 0.57,
P 0.01), detection of AFB in sputum (OR0.5, P
0.037), pulmonary dissemination (OR 0.57, P 0.045),
inconspicuous disease onset (OR 0,54, P 0.009),
nonspecific inflammatory findings detected by flexible
bronchoscopy (OR 0.46, P 0.000).
Conclusion: Several reasons for long delays were found,
but the main reason were the error of radiologist and
false-positive detection of AFB. Due to high incidence of
tuberculosis in our country there is a tendency to
overdiagnosis tuberculosis. Application of contemporary
methodologies of diagnostic tests interpretation, use of
rapid TB diagnostic is highly needed.
PC-755-04 Feasibility of a streamlined SInglesaMPLE (SIMPLE) TB diagnosis and treatment

initiation strategy in Uganda
P Shete,1,2 E Ochom,3 P Howlett,4 P Haguma,3
L Chaisson,1 A Katamba,5 D Moore,4 A Cattamanchi1,2
1
San Francisco General Hospital, San Francisco, CA,
2
University of California, San Francisco, San Francisco, CA,
USA; 3Infectious Diseases Research Collaboration, Kampala,
Uganda; 4London School of Tropical Medicine & Hygiene,
London, UK; 5Makerere University College of Health
Sciences, Kampala, Uganda. Fax: (1) 415 206 1551. e-mail:
pbshete@gmail.com
Background: In high burden countries, many patients

with tuberculosis (TB) who present to community health
centers are lost to follow-up before TB can be diagnosed

or treated, leading to ongoing transmission. A primary
reason is that the standard approach of collecting sputum
specimens over multiple days for microscopic examination is not only insensitive but also inconvenient and
costly for patients. We report on the feasibility of a
patient-centered, SIngle-saMPLE (SIMPLE) TB diagnosis
strategy.
Design/Methods: The SIMPLE TB strategy includes: 1)
Single-sample LED fluorescence microscopy (analysis
and reporting of smear results from the initial specimen
within two hours); 2) Daily transport of smear-negative
sputum samples to GeneXpertw MTB/RIF (Xpert)
testing sites; 3) SMS-based reporting of Xpert test results
to patients and health centers; and 4) Routine feedback
of TB evaluation metrics to health center staff. In a
single-arm interventional pilot study, we evaluated the
feasibility of the first two components at 4 community
health centers in Uganda. Using data from TB laboratory
and treatment registers, we evaluated process measures
that reflect implementation of the two intervention
components and TB case detection.
Results: Of 145 consecutive patients referred for TB
testing, 87 (60%) were female and their median age was
39 years (IQR 28-49). Two smears from the initial spot
specimen were analyzed and reported on the same-day
for 125/145 (86%) of patients. Overall, 11 (9%) were
smear-positive and all results from smear-positive patients were obtained on the same day as presentation.
Sputum was transported to an Xpert testing site for 65
patients (3 smear-positive, 55 smear-negative and 7
smear not done). All samples were transported on the
same day as the initial health center visit. Xpert results
were positive in 7 (11%) patients including 4/55 (7%)
smear negative patients. Follow up of patient treatment
initiation is currently underway.
Conclusion: The SIMPLE TB strategy enables same-day
diagnosis of TB cases by facilitating access to and
utilization of novel diagnostics such as Xpert, even when
not available at the point-of-care. By strengthening
health systems and reducing the cost and burden of TB
diagnostic evaluation for patients, the SIMPLE TB
strategy has strong potential to ensure that more people
with TB presenting to community health centers are
diagnosed and treated.
PC-756-04 Tuberculosis mortality in Puerto Rico,

20092013
D Thomas,1,2 T Chorba,2 M Bermudez,1 S Mase,2
O Joglar,1,2 A Khan,2 B Rivera-Garcia1 1Puerto Rico
Department of Health, San Juan, Puerto Rico, 2U.S. Centers
for Disease Control and Prevention, Atlanta, GA, USA.
e-mail: wii6@cdc.gov
Background and challenges to implementation: Puerto

Rico (PR) is a U.S. territory in the Caribbean with a
history of significant tuberculosis (TB) burden prior to
the development and effective use of chemotherapeutic
agents in the 1950s. Greater than 90% of the foreignborn population in PR originates from a high burden TB
country, Dominican Republic. PR ranks among the top
ten states and territories for persons living with HIV

(PLHIV) and the territory has an HIV associated
mortality rate four times the national average. However,
the reported rate of active TB is only two-thirds the U.S.
rate, 2 cases per 100 000 suggesting that there is underreporting.
Intervention or response: To assess the extent of underreporting, we evaluated PR specific TB mortality data
compared to those of the USA, during 20102013 as
mortality is associated with delays in diagnosis. TB
mortality was defined as cases reported as dead at
diagnosis or during TB treatment through the national
Report of Verified Case of Tuberculosis.
Results and lessons learnt: TB mortality occurred in 61 of
294 TB cases (21%) in PR compared with 4259 of 52
654 (8%) in the USA. The majority of deaths had
pulmonary TB (84% in both groups) and occurred in
men (80% in PR, 67% in U.S.). In PR, 92% of TB
mortality was among U.S. born (including PR), compared to 53% in the USA. In PR, 3% of TB mortality was
among persons born in Dominican Republic. Of persons
with a known HIV status, 37% of TB mortality in PR
was among PLHIV compared with 14% in the USA.
Conclusions and key recommendations: TB demographics in PR are significantly different as compared with the
USA. TB cases in PR are more than twice as likely to die.
TB mortality is 1.7 times more likely to occur among
U.S.- born in PR compared to the USA. Mortality among
foreign born cases is highest for Dominican Republic
nationals, however, this represents only 3% of all known
TB mortality in PR. Due to PRs high HIV incidence and
mortality rates and the high likelihood of co-infection
with TB, and we recommend annual TB symptom
screening and radiographic evaluation if symptomatic
for all PLHIV. Based upon mortality data, PR likely has
significantly more tuberculosis disease than previously
identified. Outreach to clinicians in Puerto Rico will be
important to prevent, detect, and cure TB.
PC-757-04 Mortality amongst TB cases in the

UK: an analysis of surveillance data
D Pedrazzoli,1 R Houben,1 R White,1 M Lalor,2 L Thomas2
1
London School of Hygiene & Tropical Medicine, London,
2
Public Health England, London, UK. e-mail:
debora_pedrazzoli@hotmail.com

estimates that globally the TB mortality rate fell by 45%
between 1990 and 2013; therefore progress needs to
accelerate to reach the Stop TB Strategy target of a 50%
reduction by 2015. Although this target is within reach in
half of the WHO Regions, it is unlikely to be met in the
European Region where the majority of countries
(including the UK (UK)) are classified as having a low
TB incidence. Risk factors for death in the UK have not
been fully investigated.
Design/Methods: Cases notified to the Public Health
England (PHE) Enhanced Tuberculosis Surveillance
System with death as reported outcome between 2003
and 2012 were included in the analysis. Vital registration
S175
data were matched to national surveillance data to

estimate the underreporting of mortality associated with
TB. Univariate and multivariate analyses were performed
to describe the characteristics of those who died from TB
in the UK, to identify risk factors for having the outcome
of death compared to another outcome, and to compare
the characteristics of those reported to vital registration
systems compared to those only reported to national
surveillance.
Results: UK-born patients aged 65 years or over, those
with social risk factors, patients with sputum smear
positive pulmonary disease and patients with MDR-TB
remained at increased risk of death. We found no
evidence of any difference across quintiles of deprivation.
We also found that 43% of deaths were only reported to
vital registration systems but not to national surveillance.
Analyses to explore the characteristics of cases notified to
PHE vs. cases reported to vital registration systems is
currently underway.
Conclusion: Despite an overall decline over time in TB
mortality, the number of patients who still die from this
disease remains significant, and efforts to control TB and
improve diagnosis and treatment outcome in the UK
should be sustained. If we wish to achieve a reduction of
95% in TB deaths from 2015 to 2035 more will need to
be done. Missing deaths in notification remain worrying;
this is a limitation that national TB surveillance aims to
address by linking TB and vital registration data on a
routine basis in order to ensure notification and reporting
of all TB deaths in the UK in future.
PC-758-04 Occupational exposure to indoor

congregate settings and tuberculosis mortality
J Wolny,1 C Jackson,1 H Stagg,1 I Abubakar,1,2 T Yates1,3
1
University College London, London, 2MRC Clinical Trials Unit
(at University College London), London, UK; 3Africa Centre
for Health and Population Studies, Nr Somkhele, South
Africa. e-mail: t.yates@ucl.ac.uk
Background: In high burden settings, molecular epidemiology suggests that most Mycobacterium tuberculosis
transmission occurs outside the household. Modelling
suggests much transmission occurs in indoor congregate
settings with workplaces being important sites of
transmission between adults. Data on age-specific TB
mortality by occupation for men in England and Wales
were recorded in Registrar Generals decennial supplements for 1890-2, 1900-2 and 1910-2 (when TB rates
may have been even higher than in Southern Africa
today). We examined these data to see if there was an
association between TB mortality and working in a
crowded indoor environment.
Design/Methods: Data on follow up time and deaths
from phthisis (TB) were extracted. Occupations were
coded as exposed (involving exposure to crowded indoor
environments) or unexposed (outdoor occupations or
those involving little exposure to indoor congregate
settings). We described crude TB mortality rates by
exposure status. In regression analysis we assessed the
association between occupational exposure to indoor
S176
congregate settings and mortality from TB adjusting for

decade, age group and socioeconomic position.
Results: There were 24 366 TB deaths over 15.2 million
person-years among men in exposed occupations (160
deaths per 100 000 person-years). There were 28 775 TB
deaths over 20.1 million person-years follow up time
among men in unexposed occupations (143 deaths per
100 000 person-years). However, most men worked in
occupations that we were unable to classify and these
men had the highest mortality. These trends persisted in
adjusted analysis but were attenuated if occupations with
additional risk factors for TB, such as silica exposure,
were excluded.
Conclusion: These data are consistent with crowded
indoor workplaces being important sites of M. tuberculosis transmission. Non-differential misclassification of
cause of death and the fact that men may have worked in
both exposed and unexposed occupations may have
caused us to underestimate the strength of the association. We were unable to adjust for urban or rural
residence nor address the possibility that men might have
moved into more or less exposed occupations after
developing TB disease. The substantial differences
between England and Wales a century ago and high
burden settings today should caution against generalisation of these findings. Future research should assess
whether crowded indoor workplaces remain important
sites of M. tuberculosis transmission.
08. Trends, transmission and tuberculosis

case finding
PC-759-04 Offer incentives and directly
observed treatment of tuberculosis in primary
health care in a Brazilian city
A Beraldo,1 N H Orfao,1 A Wysocki,1 M E Brunello,1
R Andrade,1 L M Scatena,2 T C Scatena Villa1 1College of
Nursing of Ribeirao Preto University Sao Paulo, Ribeirao
Preto, SP, 2Federal University of Triangulo Mineiro, Uberaba,
Minas Gerais, Brazil. e-mail: li_aab@yahoo.com.br
Background: The Directly Observed Treatment is a

strategy used in the organization of care for tuberculosis
patients, and should be seen not only as the supervision
of the dose, but the involvement of health professionals,
and community service. Allied granting incentives offer
the opportunity to identify other patients life needs,
helping to control the disease.
Aim: To characterize the offer incentives and the directly
observed treatment (DOT) for tuberculosis (TB) in
primary health care (PHC) services by patients and
professionals nursing team perspective in a Brazilian city.
Methods: Descriptive epidemiological study, conducted
in Campinas-SP (Brazil) city, priority for TB control,
between August 2012 and May 2013. The study
population was composed of professionals nursing team
of the PHC and TB patients. A questionnaire with 4
questions with dichotomous response (yes or no), related
to the offer incentives and DOT for TB patients in the
PHC services was used. These questions were designed

based on the National Manual Recommendations for
the TB control. For data analysis was used descriptive
technique, frequency distribution, measures of central
tendency and variability.
Results: 183 professional nursing team and 165 TB
patients were interviewed. The perceptions of TB
patients showed that 94 (57.0%) underwent the DOT,
155 (93.9%) reported that the intake was observed by a
health professional, averaging receiving TDO 2.2 times a
week (SD 1.6), 136 (82.4%) were given a breakfast kit
and 89 (54.0%) were treated by the same health
professional (Table). The perceptions of health professionals showed that 178 (97.3%) professionals offered
DOT for patients in treatment; with an average of 3.3
observations of intake of DOT in the week; 158 (86.3%)
of professionals offered the Breakfast Kit (Table).
Conclusions: There was disagreement between the
perceptions of patients and professionals nursing team
for the realization of DOT, and the number of intake of
observations in the week. For the effective management
of the disease, it is essential the partnership between TB
patients, health professionals and health service, regarding the implementation of these actions, taking into
account the opinion of the patient as to the place and
time of observation of dose, aimed at strengthening
treatment, decreased disease transmission and consequent reduction in morbidity and mortality.
Table Frequency distribution of the offer incentives and directly
observed treatment of the tuberculosis in the primary health
care services by patients and professional nursing team
perspectives, Campinas-SP, 2013.
Indicators
Providing DOT *
to tuberculosis
patients.
Observation the
Ingestion of
Medicines**
Categories of
responses
Yes
No
Health
Professional
Family
Nobody
People in the
Community
Offering
Breakfast Kit
Incentives**
Basic Food
Basket
Transportation
vouchers
No
Care by the same Yes
health
No
professional
Not applicable
during the
DOT *
Patients
n
94 57,0
71 43,0
178
5
97,3
2,7
155 93,9
172
94,0
26 15,8
8 4,9
-
96
7
2
52,5
3,8
1,1
136 82,4
22 13,3
158
37
86,3
20,2
13
7,1
0,6
-
Professionals
Nursing
Team
1,2
29 17,6
89 54,0
3 1,8
73 44,2
S177
PC-760-04 Temporal trends of drug-resistant

tuberculosis in eastern rural China, 20052012
PC-761-04 Is tuberculosis declining in West

Papua? Challenges in disease control
Y Hu,1 L Wu,1 Q Zhao,1 W Jiang,1 B Xu1 1School of Public

Health, Fudan University, Shanghai, China. e-mail:
bxu@shmu.edu.cn
T Pakasi,1,2 HA Papendang1 1Universitas Papua, Sorong,

2
Faculty of Medicine Universitas Indonesia, Jakarta,
Indonesia. Fax: (62) 021 315 3550. e-mail:
tpakasi@yahoo.com
Background: To analyze the longitudinal trend of drug

resistant tuberculosis (TB) in duration of 8 years in
eastern rural China.
Design/Methods: The study was carried out in two rural
counties of eastern China during 2005-2012. All
registered TB patients received bacteriologic-based TB
diagnosis. Drug susceptibility test to first-line anti-TB
drugs, including isoniazid, rifampin, ethambutol and
streptomycin were applied to all mycobacteria tuberculosis isolates. The temporal trend of drug resistant TB
during the 8-year study period were simulated and
plotted by non-linear regression model using exponential
equation. Growth ratio and doubling time of drug
resistant TB were calculated using Graphpad Prism 5.
Results: The prevalence of overall drug resistance was
45.3%, and the prevalence of multidrug-resistant TB
(MDR-TB) was 13.4% during the study period (10.4%
in new patients and 26.0% in previously treated
patients), which was higher than national average level.
The prevalence of overall drug resistance and MDR-TB
among bacterial culture positive TB patients increased
from 33.6% and 3.6% in 2005 to 57.3% and 27.3% in
2012. The growth ratio of overall drug resistance was
1.09 (95%CI: 1.058-1.126), and the doubling time was
7.54 years. The growth ratio of MDR-TB was 1.25
(95%CI: 1.185-1.323), and the doubling time was 2.730
years. The prevalence of resistant to isoniazid, rifampin
and ethambutol exhibited an upward trend, while it was
downward in streptomycin. In newly diagnosed TB
patients, the growth ratio of overall drug resistance and
MDR were 1.20 (95%CI: 1.006-1.172) and 1.23
(95%CI: 1.055-1.187) respectively, whereas they were
1.10 (95%CI: 1.055-1.187) and 1.27 (95%CI: 1.1551.386) respectively in previously treated patients. The
growth ratio of overall drug resistant TB was significantly higher in male than female (F 23.640 P ,
0.0001). Compared to previously treated patients, the
newly diagnosed TB patients had a significantly higher
growth ratio of overall drug resistance (F 15.750, P
0.002), but a significantly lower growth ratio of MDRTB (F 26.380, P , 0.001).
Conclusion: The temporal trend of drug resistant TB in
eastern rural China suggests that the burden of drug
resistant TB would be increasing rapidly, and the
prevalence of MDR-TB might be doubled in less than
three years. An MDR-TB control program targeting high
risk TB patients should be scaled up in the basic TB
control unit of Chinas TB control system.
Background: West Papua Province is one among the

youngest provinces in Indonesia, situated in Papua
Island, small population living in wide area. Recently,
the ministry of health published in 2013, a populationbased study about health situation (Riskesdas 2013). The
report mentioned that the tuberculosis (TB) prevalence in
West Papua was similar as the national rate, declined
from the same study in 2007 (0.9% to 0.4%). It was
quite premature to conclude the declining was true and
therefore we studied further the report.
Methods: We reviewed several reports regarding prevalence, case finding, services of TB besides the methodology being implemented in the forming of the data. We
limited the search from 2000-2014.
Results and lesson learnt: The Riskesdas 2013 reported
0.4% respondents were TB in West Papua and 3.5% had
the main symptom in which 2.7% had bloody coughing.
It showed clearly that 3.1% of the respondents might
have TB but had never been diagnosed. The Directorate
General of Disease Control found the decrease of case
finding in West Papua from 621/100 000 population
(2012) down to less than 500/100 000 population
(2013), compared to the Papua province, the achievement was only half. The estimated population of West
Papua was 846 711 people with density of 8.5 people/
Km2. With a wide area and scattered population, the
issue of accessibility and availability of the TB service is
very important to overcome. For more information, the
positivity rate in West Papua achieved 14.8%, almost
5% higher than the national rate (10.4%). The estimated
TB patients might be very high. It could be problems in
the laboratory or access to health care, thus only the very
needed patients went to seek medical care and increase
the chance to become positive TB. The Riskesdas 2013
was a paper-based study and the category of TB was selfreported by the respondents, whether respondents had
ever been diagnosed as TB. Sputum test was done
whenever problems occurred but did not use in the
calculation to correct the prevalence. It was different
from another population based-study in 2004, the
Household and Health Survey, which used sputum test
and culture for correction of the diagnosis.
Conclusion: Decrease prevalence of TB reported in 2013,
revealed an accessibility issue to health care rather than
the successful of the Program in West Papua. Scattered
population and fewer facilities were the main reasons as
reviewed in many reports.
S178
PC-762-04 Increased case finding among highrisk groups by mobilizing frontline health
workforces in Kathmandu Valley
PC-763-04 Diagnostic value of clinical features

of presumptive TB patients in Bagamoyo,
Tanzania: a prospective cohort study
R S Gopali,1 G Shrestha,1 B Lamichhane,2 N Ishikawa,3

L Ditiue4 1Japan-Nepal Health & TB Research Association,
Kathmandu, 2National Tuberculosis Centre, Kathmandu,
Nepal; 3Research Institute of Tuberculosis/Japan
Anti-Tuberculosis Association, Tokyo, Japan; 4UNOPS/Stop TB
Partnership, Geneva, Switzerland. e-mail:
gwala73@hotmail.com
F Mhimbira,1,2,3 J Hella,1,2,3 M Sasamalo,2 K Ibrahim,1,2,3

E Mfinanga,2 F E Haraka,2 L Jugheli,1,2,3 K Reither1,2,3
1
University of Basel, Basel, Switzerland; 2Ifakara Health
Institute, Bagamoyo, Tanzania; 3Swiss Tropical and Public
Health Institute, Basel, Switzerland. e-mail:
fmhimbira@gmail.com
Aim: To detect additional TB case finding in Kathmandu

Valley.
Methodology: This cross sectional study was done in
Kathmandu valley focusing Garbage collectors, suspects
visit in the Private pharmacies and street childrens. The
trained team of health care providers and volunteers
working in the Kathmandu valley conducted intensive
community mobilization activities, performing door-todoor visit, refer the suspects found in community to near
by microscopy centres or camps for the sputum
examination. The intervention target to screen approximately 80 000 from the above mentioned groups and
detect the additional cases in the valley in one year. Two
sputum samples including one early morning and 1 spot
specimen were collected in same day. For chest X-ray and
Xpert test suspects were advised to visit hospitals and
near by health facilities where the facilities are available.
Results: Among total (53 717) persons screened in 8
month of period, 9585 (18%) were identify with TB sign
and symptoms. 83% (8003) suspects identified in the
community and private pharmacy actually visited for the
sputum examination. Among total examine 120 people
were diagnosed as New Smear Positive (NSP) patients.
Nearly two third of diagnosed cases were male and their
mean age was 36. The prevalence of active pulmonary
tuberculosis in those targeted areas is 2 times higher
comparde to other general groups. The intensive ACSM
activities and door-to-door visits improve the health
seeking behavior of the target groups. Among total
diagnosed cases 116 (97%) TB patients enrolled their
treatment from the near by DOTS centres. This
intervention contributes 10% of the additional TB case
finding in the Kathmandu valley.
Conclusion: The Public-Private Mix (PPM) activity in the
community support to maintain the motivation level of
key stakeholders. Regular mobilization of key stakeholders can contribute to increase additional case finding
rate from the hard to reach area and it can be sustain by
endorsing similar types of the programme in the regular
activity of National Tuberculosis Programme.
Background: The hallmark of tuberculosis (TB) control

is early diagnosis and treatment of infectious pulmonary
TB patients. Symptom-based screening recommended by
World Health Organization, is simple and cost-effective
way of identifying presumptive TB patients to be tested
for TB. Objective: To determine the diagnostic value of
clinical features in a microbiologically and clinically well
characterized cohort of presumptive TB patients.
Methods: A prospective study adult presumptive TB
patients was done between September 2010 and September 2013. Presumptive TB was defined as having any
cough and one or more of any other TB symptom.
Participants were followed up for 18 months. At five
months of follow-up, participants were classified into
seven categories based on clinical assessment, radiological findings, sputum smear microscopy, liquid and solid
culture results. Sensitivity and specificity were estimated,
using culture-positive TB patients and controls (no
symptoms at five months without anti-TB treatment) as
reference standard.
Results: A total of 499 adult presumptive TB patients had
a median age of 39 years (interquartile range; 30-50
years); 259 (51.9%) were females. The estimated HIV
prevalence was 45.7 [95% Confidence Interval: 41.2%
to 50.2%]). Of 499 enrolled, 114/499 (22.8%) were
culture-positive TB patients and 118/499 (23.6%) were
non-TB controls at five months (total, n 232). Among
the culture-positive TB patients, 84/114 (73.6%) were
smear-positive and culture-positive and 30/114 (26.4%)
were smear-negative culture-positive. Chest pain was the
commonest reported symptom (75.9%) followed by
fever in 162/232 (69.8%) and weight loss in 152/232
(65.5%) patients. Weight loss and fever have high
sensitivity of 80.7% and 78.9%, respectively. Specificity
was high for haemoptysis 85.6% and fair for fatigue
56.8% (Table). In a multiple logistic regression; male,
excessive night sweats, weight loss and fatigue and age
were significantly associated with culture-positive TB
patients with P , 0.05.
Conclusion: Symptom screening performs with a fairly
good sensitivity but predominately poor specificity to
identify TB patients in resource-poor setting with high
HIV prevalence. Symptom screening can contribute to
ruling out TB in many cases, but better TB point-of-care
diagnostics tools are still urgently needed.
Table Patient classification, clinical features at recruitment and

their diagnostic value in presumptive TB patients against liquid
and solid culture combined as gold standard
Symptom
Culturepositive
PTB
n 114
n (%)
Non-TB
controls
n 118 Sensitivity Specificity
n (%) % (95%CI) % (95%CI)
Productive cough 114 (100) 117 (99.2)

Haemoptysis
12 (10.5) 17 (14.4)
Night sweats
83 (72.8) 60 (50.9)
Fever
90 (79.0) 72 (61.0)
Weight loss
92 (80.7) 60 (50.9)
Chest pain
85 (74.6) 91 (77.1)
Fatigue
76 (66.7) 51 (43.2)
100
(96.8100)
10.5
(5.617.7)
72.8
(63.780.7)
78.9
(70.386.0)
80.7
(72.387.5)
74.6
(65.682.3)
66.7
(57.275.2)
0.8
(0.04.6)
85.6
(77.991.4)
49.2
(39.858.5)
39.0
(30.148.4)
49.2
(39.858.5)
22.9
(15.731.5)
56.8
(47.365.9).
PC-764-04 Seasonal drivers of tuberculosis:

evidence from 100 years of notifications in
Cape Town
J Andrews,1 S Hermans,2,3 R Wood3 1Stanford University,
Stanford, CA, USA; 2Amsterdam Institute for Global Health
and Development, Amsterdam, Netherlands; 3University of
Cape Town, Cape Town, South Africa. e-mail:
jasonandr@gmail.com
Background: Tuberculosis has been observed to follow a

seasonal pattern in many settings; however, the impact of
HIV on seasonality and the contribution of seasonal
factors influencing reactivation risk vs. variation in
transmission risk have not been elucidated. We examined
trends in seasonality of tuberculosis before and during
the HIV epidemic, and according to age and HIV status.
Design/Methods: We reviewed data on monthly tuberculosis notifications in Cape Town from 1903 through
2013, exclusive of 1995-2001, which were unavailable.
Data from 2003 onward were stratified by HIV status
and age. We decomposed time series into seasonal,
random, and trend components, and performed timedependent spectral analysis using wavelets to assess
periodicity over time.
Results: In the 20th century, there were intermittent
periods in which tuberculosis followed an annual
seasonal pattern; however, from 2003 onward, consistent seasonal periodicity was observed, with the highest
number of cases occurring in August through November
(peak: October), following the coldest months (Figure).
Seasonal effects were similar between men and women.
Among children under 5 years of age, a second seasonal
peak was observed in March and was higher than the
October peak; this second peak was not seen in other age
groups. Evidence for seasonality was stronger among
HIV co-infected tuberculosis cases than in HIV uninfected cases. The amplitude of seasonality was comparable
between HIV positive and negative cases, with peak case
months recording 47% and 42% times more cases than
trough months, respectively.
S179
Conclusion: There is substantial seasonal variation in the

burden of tuberculosis in Cape Town in the HIV-era.
Stronger seasonal effects were seen in HIV-infected
individuals and those in younger age groups, who
progress more rapidly to tuberculosis following infection. These findings may support increased transmission
in the winter as an important driver of tuberculosis in
Cape Town.
PC-765-04 The importance of household vs.

community tuberculosis transmission in
Peruvian shantytowns
L Martinez,1,2 S Isaacson,2 R Gilman,3 M Saito,4
L Cabrera,5 R Oberhelman,2 C Evans6,7 1University of
Georgia School of Public Health, Athens, GA, 2Tulane
University School of Public Health and Tropical Medicine,
New Orleans, LA, 3Johns Hopkins Bloomberg School of Public
Health, Baltimore, MD, USA; 4Tohoku University Graduate
Benefica
School of Medicine, Sendai, Japan; 5Asociacion

Proyectos en Informatica, Salud, Medicina y Agricultura (AB
PRISMA), Lima, 7Universidad Peruana Cayetano Heredia,
Lima, Peru; 8Imperial College London, London, UK. e-mail:
leomarti@uga.edu
Background; Active case finding (ACF) for tuberculosis

(TB) has been widely proposed for strengthening TB
control. However, it is unclear whether it would be more
effective to target ACF in households or community
areas. The comparative effectiveness of ACF in community vs. households is likely to depend on relative TB
transmission levels in these settings.
Objectives: To compare the prevalence of childhood TB
infection in households with and without a known past
TB case, calculate the odds of household TB transmission, and assess changes in these data comparing years
1990 vs. 2000.
Methods: Two community-based tuberculin surveys
were performed in 1990 and 2000 in Pampas de San
Juan de Miraflores, a low-income shantytown in Lima.
Tuberculin skin tests were performed in children (0-14
years old) to assess latent TB infection (LTBI) as evidence
of transmission. Prevalence of latent TB infection from
both surveys were calculated and children with household exposure to a known case of TB disease were
identified. We stratified children in both studies by
household TB contact and calculated odds ratios of new
infection in children with vs. without household TB
S180
exposure. Two-tailed P values were calculated using twosample t-tests.

Results: The study sample sizes were 238 in 1990 and
624 in 2000. In children 0-4 years old, overall prevalence
of LTBI was significantly higher (P , 0.0002) in 1990
(24/149 infected; 16% prevalence [95%CI, 10 22])
compared to 2000 (4/188 infected; 2.1% prevalence
[95%CI, 0.1 4.1]). Similarly, prevalence was significantly higher (P , 0.0002) in children 5-14 years old in
1990 (44/238; 19% infected [95%CI, 14 23]) vs. 2000
(44/624; 7.1% infected [95%CI, 5.1 9.1]). The
increased odds of infection in children exposed to a
household TB case was 2.4 (95%CI, 1.7 3.5) in 1990
and 5.5 (95%CI, 2.7 11) in 2000.
Conclusion: In these two community-based studies from
the same Peruvian shantytown, we found that the odds of
household TB transmission doubled over 10 years while
the prevalence of new TB infection decreased significantly. This suggests that household contact investigations may have higher yield in preventing transmission in
low TB incidence areas compared to those with highburden.
21 406]). There was an increase in the number of longterm migrants from high incidence countries (151-250
per 100 000) until 2009 (from 143 987 to 200 944),
followed by a drop in the number of migrants from these
countries to 102 837 in 2013.
Conclusions: The changes in TB incidence in the UK over
the past decade are due to changes in the numbers of nonUK born TB cases only. Over the same time period, the
numbers of TB cases in the UK born population has
remained stable, but has not seen the reduction observed
in most other comparable countries. Changes in the
pattern of migration to the UK over the past decade,
particularly an initial increase, and more recent decrease
in the number of migrants from high TB incidence
countries, is likely to have had an important impact on
TB in the UK. However, additional factors, including
transmission within the UK and changes to TB control
activities are may also have played a role.
PC-766-04 The impact of changes in patterns of

migration on TB incidence in the UK
M Muzyamba,1 D Pedrazzoli,1,2 D Zenner,1 M Lalor,1
J Davidson,1 I Abubakar,1 L Thomas1 1Public Health
England, London, 2London School of Hygeine & Tropical
Medicine, London, UK. Fax: (44) 20 8327 6112. e-mail:
morris.muzyamba@phe.gov.uk
Background: The incidence of TB in the UK increased by

50% from the mid-1980s to 2011, followed by a small
reduction in 2012 and 2013. In contrast, most comparable countries saw a steady decline in incidence over the
same time period. The majority of TB cases in the UK are
non-UK born, and migration is likely to have a major
impact on the changing epidemiology. This study aims to
estimate the impact of changes in migration patterns
from countries with different levels of TB burden on TB
incidence in the UK.
Methods: Data on tuberculosis notifications from the
Public Health England Enhanced Tuberculosis Surveillance system from 2004-2013 were analysed. Annual TB
rates and TB rates by country of birth were calculated
using population estimates provided by the Office of
National Statistics and the Labour Force Survey respectively. Data on the annual number of long term migrants
to the UK were obtained from Home Office and stratified
by the 2012 WHO TB incidence estimates in their
country of birth.
Results: From 2004 to 2011 there was an increase in the
number of non-UK born TB cases from 4789 to 6,335;
from 2011 to 2013 there was a decrease in the number of
non-UK born TB cases, to 5529 in 2013. In contrast, the
number and rate of TB cases in the UK-born population
has remained stable over this period, with just over 2000
cases a year, a rate of 4 per 100 000 (Figure). Over the
past 10 years, there has been a steady decrease in the
number of long-term migrants to the UK from very high
(.250 per 100 000) incidence countries [from 67 917 to
09. Public-private strategies to End TB:

the case of India
PC-767-04 AYUSH providers involvement in TB
prevention and care under PPP in India
B Entoor Ramachandran,1 S Chadha,1 S Waikar2
1
South-East Asia Office, New Delhi, 2Population Services
International, Delhi, India. Fax: (91) 1 14605 4430. e-mail:
ebabu@theunion.org
Background: AYUSH is an acronym that is used to refer

to the non-allopathic medical systems in India. It includes
the Indian medical system of Ayurveda, Yoga, Unani,
Siddha, and also Homeopathy. (AYUSH mean life in
Sanskrit). India, Every year, around 30,000 persons are
graduated from 494 AYUSH medical colleges as per
(2012-2013) Central Council of Indian Medicine,
Department of AYUSH, Ministry of Health and Family
Welfare. India. In community, AYUSH practitioners are
first point of care for any health problems. Provides the
bulk of health care, particularly for the poor and their
presence in Vulnerable and Marginalised areas or
locations and this hold true for TB symptomatics as well.
Intervention: Axshya (TB Free) is a Global Funded

round 9 TB project launched in India in 2010, it is
covering 300 districts across 21 states in the country. The
objective of Project Axshya is to expand the reach and
visibility of Revised National Tuberculosis Control
Programme (RNTCP) of India. Considering the role of
AYUSH in TB prevention and care, project Axshya
trained and engaged AYUSH practitioners into TB care in
60 Districts in 10 states in India. Axshya intervention
focused on following components. Recognition by the
Government authorities (National TB Programme, and
District Tuberculosis Office [DTO]), Capacity building,
Regular coordination with District TB staff, Providing
Incentives and Motivation.
Results: Totally 3165 AYUSH practitioners have been
sensitized during the period April 2013 to March 2015.
AYUSH practitioners were expected to offer following
three TB services with support of financial incentives: 1)
referral of persons with symptoms of TB, 2) sputum
collection and transportation, and 3) provision of DOT
to TB patients. Out of trained, 40% of AYUSH
practitioners were engaged by providing at least one of
the TB service. During the reporting period 10848
persons with symptoms of TB have been referred for
diagnosis. Out of the referred, 6508 (60%) persons have
undergone diagnosis tests. 1065 (16%) persons were
diagnosed as having TB and 95% (1007) of them have
been initiated on treatment.
Conclusion: Indias National Tuberculosis control programme is aiming towards Universal Access to TB Care.
In this context, every provider from different discipline
pays a role in identification and serving persons with TB.
Ongoing Axshya intervention is a small effort like a drop
in bucket towards Mainstreaming AYUSH with RNTCP
to support END strategy.
S181
PC-768-04 Why are private practitioners not

willing to be part of the tuberculosis control
programme in India?
P Banuru Muralidhara,1 S Pattanshetty,2 P Mithra2
International Union Against Tuberculosis and Lung Diseases,
South-East Asia Office, New Delhi, 2Kasturba Medical
College,Manipal University, Manipal, India. Fax: (91) 114
605 4430. e-mail: bmprasad@theunion.org
1
Background: Engaging private practitioners is the key

strategy in reaching the missing 3 million tuberculosis
patients and much of this is coming from India.
Although, tuberculosis control programme has developed and implemented strategies to engage private
sector; results are not encouraging. A study was
conducted to answer why private practitioners were
not willing to be part of programme?
Design: A cross sectional qualitative survey was conducted among 116 private practitioners (General (GP)
and Specialists (SP)) in Udupi district of Karnataka, India
from September 2008 to October 2009. The investigator
visited each practitioner in their respective clinics and
administered pretested questionnaire tool after taking
the consent. The data was analysed under patient related,
practice related and regimen related indicators as
expressed by practitioners.
Results: The sample constituted 41 GPs and 75 SPs and
33% of them did not refer any presumptive TB persons.
Only 2% of practitioners cited patient related issues
(Patients dont believe in government drugs) for not
referring to programme. Nearly 56% of them informed
practice related (We lose patients, why to refer to
DOTS? We can also think better ways to treat) and 36%
to regimen related (Patients are comfortable with the
treatment regimen which I am giving). 90% of the
referrals made by practitioners were in view of ability-topay by patients (We refer only poor patients as they get
free drugs). A total of 53 private practitioners responded as not referring presumptive TB persons for programme or DOTS; of which 36% were related to
regimen (Side effects are more with DOTS regimen as
it is not adjusted per Kg body weight) and practice
(Every physician in private sector has his own method of
treating tuberculosis, no clear guidelines. CME programme would be better). Other responses of which
19% were related to patients (Confidentiality issues are
not maintained well as health worker visits patients
home) and 9% programme related (DOTS programme
is for masses not for individual patients, we private
practitioners deal with individual patients).
Conclusion: The results draw our attention to practice
related issues as key concern for private practitioners not
willing to be part of the programme. Practice related
objectives derive the choice of regimen and ability-to-pay
by patient. A regulation policy on use of antibiotics by
private practitioners and policy on CME is the need of
hour.
S182
PC-769-04 Have I notified a TB patient today?

Study of TB notification by private sectors after
the introduction of Nikshya in India
S Satapathy,1 B Thapa,1 S Chadha1 1International Union
Against Tuberculosis and Lung Disease South-East Asia
Office, New Delhi, India. e-mail:
Sachi.Satapathy@theunion.org
Background: Tuberculosis (TB) is a notifiable disease in

India since 2012. The private sectors cater to almost 50%
of TB patients however the notification is minimal. To
strengthen the TB case notification from private sectors,
National TB Programme has initiated a web-based
notification system-Nikshya in 2013 where the private
sectors are registered. The objective of the study was to
examine the registration and notification of TB cases
from the private sector before (2012) and after the
Nikshya (2013).
Design/Methods: This is retrospective study carried to
analyse the secondary data on private sector registration
and Tb case notification from NTP reports published in
2014. The variables studied were, registration and TB
case notified from private practitioner, hospital and lab.
The outcome measures were number and proportionate
increase in registration and Tb case notification.
Results: Altogether, private sectors registered in the web
based system were 55 513 and 1839 in 2013 and 2012,
respectively with an absolute increase of 53 674. Of
those registered in 2012, 74% (1360/1839) were private
practitioners, 22 % were private hospitals and 4% were
private labs while in 2013, 64% were private practitioners, 27% followed by 27% private hospitals and 9%
private labs. Similarly, a total of TB case notifications
reported were 38 596 and 3106 in 2013 and 2012
respectively with an absolute increase of 35490. Of those
notified in 2012, except private practitioners, whose
contribution is 91%, private laboratories and private
hospitals contribution is 4% each, while in 2013, share
of private hospitals is more (50%), followed by private
practitioner (45%) and private Lab (9%). Within the
private sector participation the proportionate figure of
registered vs. notification, private hospital has shown an
increase (0.32 in 2012 & 1.29 in 2013), however, the
other private laboratories (1.8 in 2012-0.29 in 2013) and
private practitioner (2.0 in 2012 - 0.49 in 2013) have
shown a negative trend. Within the states in India, out of
36 states and union territories (UT), notification rate of
private sector per 100 000 population 16 of them have
register the contribution of , 1.0.
Conclusion: A revised strategy to engage private sector
needs to be taken up by the NTB program with a focus on
private practitioners and private lab engagement with
more priority to 16 states, where the contribution of
notification is less than one per cent for lakh population.
PC-770-04 How to better equip private health

service providers to fight tuberculosis in India:
a study of the PHSP perspective on the
challenges
S Satapathy,1 B Thapa,1 S Chadha1 1International Union
Office, New Delhi, India. e-mail:
Sachi.Satapathy@theunion.org
Background: With more than one million TB cases being

missed in India, the increasing role of Private Health
Service Providers (PHSP) is considered as a potential
strategy to reach out to these missing cases. In spite of
governments consistent efforts to enhance participation
of PHSP in the countrys TB control program, the
engagement of PHSP does not show much improvement
as expected (current national average of PHSP in 201314 data is 3.3 per cent per one lakh population). So, there
are assumptions that PHSPs perspective on the challenges of working with communities influence casefinding and case-management. However, there is hardly
any information about the PHSPs perspective on the
challenges they face in working with communities.
Design/Methods: A cross sectional community-based
survey in 30 districts of India for Tuberculosis Knowledge, attitude and practices was conducted in 2013. The
PHCPs were identified during a door-to-door survey and
interviewed using a semi-structured questionnaire.
Results: Of 195 PHSP interviewed in this study, 36%
PHSPs have managed TB patients in their practice. Of
these, 11% have diagnosed and initiated the treatment
on their own. Among these 57% used allopathic
medicine for TB treatment. 35% PHSPs interviewed
knew about microscopy centres (MC) and knew diagnosis is for free and of these 23% did not used the
diagnostic services from the microscopy centres. Moreover, 14% of the PHSPs, who knew of MC, revealed that
they face problem in getting sputum smear examination
done. The key challenges faced by the PHSPs were; 1)
people hesitate to visit their clinic for treatment 45% 2)
non-availability of anti TB drugs 24% with PHSPs 3) No
pathology or X ray lab available in their vicinity 20%
and 4) people do not reveal their history of TB easily
25% .
Conclusion: If resources are provided and challenges
faced are addressed, they can contribute to find missing
cases.
PC-771-04 Public health marketing: engaging

private providers for public health the pharma
way
S Vijayan,1 R Gandhi,1 V Oswal,2 G Ghatawat,2
R Chopra,1 M Datta,1 J Thakker,1 S Pandey1 1PATH,
Mumbai, 2Mumbai Mission for TB Control, Mumbai, India.
Fax: (91) 114 605 4430. e-mail: shibuvij@gmail.com
Background and challenges to implementation: Effective

engagement of private providers (PP) to appropriately
diagnose, treat and notify TB cases has been a challenge
in India. Conventional methods like sensitization and
training workshops have seen limited success. While such
workshops may guarantee participation, there is limited

scope for continued engagement with providers.
Through Private Provider Interface (PPIA), the engagement of private providers is designed to strengthen the
service delivery for TB through increasing the outreach of
the services and achieving higher case notifications of TB
Intervention or response: PPIA engaged with PP including less-than-fully qualified (LTFQ), fully qualified
providers, chemists, and laboratories for providing
subsidized and quality TB diagnostic and treatment
services. PPIA field team consists of experienced officers
as medical representatives (detailers) with pharmaceutical marketing experience. The officers visit providers in
their clinics and explain about the PPIA services like
subsidized diagnostic and free treatment. Doctors are
also sensitized on Standards of TB Care in India by
sharing an information bookletand detailing aids.
Doctors contribution to the program viz. number of
patients referred, number of diagnostic tests conducted
etc. is shared. Each field officer is allocated a particular
geographical area and given per day target visits.
Typically, a field officer visits a particular provider twice
a month. Each field officer is responsible for engaging
with 100-120 providers monthly with daily visits to 8-10
providers. The health provider is ensured accessibility to
subsidized services for patients, awareness about the
related services and thus acknowledgement of treating
the afflicted community at a subsidized cost.
Results and lessons learnt: PPIA has reached out to 2500
private providers through in-clinic visits in six months of
operations.This covers 42% of relevant provider universe in Mumbai. We have been able to engage 1145
providers in the PPIA network which is 19% of the
relevant provider universe.
Conclusions: Private provider engagement ensures regular interaction with the private providers. Repeat visits to
provider clinics enables to capture provider feedback and
improve service delivery.The success and cost effectiveness of this model of marketing public health in similar
lines with Pharma industry need to be evaluated before
expanding
PC-772-04 TB notification in south India: what

do private practitioners anticipate?
A Trivedi,1 S Burugina Nagaraja,2 S Chadha,1 S Prasad,3
Sury Kanth,4 K Sagili1 1International Union Against
Tuberculosis and Lung Disease South-East Asia Office, New
Delhi, 2Employees State Insurance Corporation Medical
College and Post Graduate Institue of Medical Sciences &
Research, Bangalore, 3Eli Lilly and Co., New Delhi, 4State TB
Cell, Bangalore, India. e-mail: atrivedi@theunion.org

Government of India has made TB notification by
private healthcare providers mandatory from May
2012 onwards. The National TB Programme has
developed a case based web based online reporting
mechanism Nikshay to facilitate notification. Objective is to determine the awareness, practice and
anticipated enablers related to TB notification by private
practitioners.
S183
Intervention or response: A cross-sectional study was

conducted among private practitioners of Mysore city in
south India. All the private practitioners were mapped
and the practitioners were interviewed using semistructured questionnaire. Systematic stratified sampling
was adapted to attain the desired sample size and the
practitioners were line-listed for interview.
Results and lessons learnt: There were 1249 private
practitioners at Mysore city of which 155 practitioners
were interviewed. Majority 142 (92%) of the private
practitioners were aware that TB is a notifiable disease;
however, 82% of them were not aware of Nikshay.
Only one in six practitioners were registered in Nikshay, while one in three practitioners notified the
disease. The anticipated enablers from the programme
were free drugs, training and timely feedback for their
referred cases. Nearly 47% of them opined that it should
also be backed by legal punitive measures
Conclusions and key recommendations: The programme
should develop innovative strategies that provide enablers, address concerns of practitioners while having
simple mechanisms for TB notification. The programme
should strengthen its inherent capacity to monitor TB
notification.
PC-773-04 Public private partnership in HIV and

TB: a case study of India
R Deshmukh,1 A Amar Shah,1 A Elizabeth,2 A Sreenivas,1
K S Sachdeva3 1RNTCP, World Health Organization Country
Office for India, New Delhi, 2National AIDS Control
Organisation, New Delhi, 3Central TB Division, New Delhi,
India. e-mail: deshmukhr@rntcp.org
Background and challenges to implementation: India has

second highest burden of TB and HIV in the world. India
has one of the largest private sectors and this is often the
first point of contact for a significant number of TB and
HIV patients. There is felt need for decentralised service
delivery with involvement of Private, NGO and corporate sector for stregthening the Services of HIV and TB.
Initiative were taken by National AIDS Control Program
(NACP) and Revised National TB control Program
(RNTCP) in India to involve Private practioners, private
medical colleges and public sector institutions as Public
Private Partnership to provide decentralised service
delivery for TB & HIV and as per the END TB strategy
engage all care providers for integrated patient centered
care and prevention.
Intervention: To address the challenge of providing
decentralise services to in India private sector hospitals
were engaged to start Integrated counselling and testing
centres and designated microscopic services . Memorandum of understanding was signed with 11 ministries
under the government of India and public sector
institutions were sensitised for service delivery of
RNTCP and NACP. Revised flexible schemes have been
initiated to engage the private sector.Standards of TB
care in India have been developed and dissemeniated to
private sectors.Sensitisation of professional bodies like
Indian Medical Association is ongoing and private
medical institutions are involved in program.
S184
Results and lessons learnt: 5086 PPs involved in RNTCP

schemes and 1033 PPP ICTCs in Private Sector were
initiated for HIV and counselling and testing services.RNTCP DMCs provide quality assured microscopic
services as per the scheme .Quality services are provided
by specialist trained in TB and HIV. Partnership has
shown increase case detection rates, reduce diagnostic
delays and cost to the patients.
Conclusions and key recommendations: Private sector
involvement leads to quality healthcare delivery as seen
in the case of the India .PPP approaches are necessary for
early diagnosis, decentralised TB and HIV treatment
services, to reduce out of pocket expenses to patients.
Training and sensitisation meetings between the public
and the private sectors are essential prerequisites for the
success of the collaboration.Promoting Standards of TB
care and adress HIV and TB in integrated manner is
essential for patient centred care and prevention as per
the END TB Strategy.
PC-774-04 Urban free TB service delivery via

engagement with private providers in
Mumbai, India
S Vijayan,1 T Shah,2 P Keskar,2 A Bamne,2 M Khetarpal,2
J Salve,2 R Chopra,1 J Thakker1 1PATH, Mumbai, 2Municipal
Corporation of Greater Mumbai, Mumbai, India. Fax: (91)
114 605 4430. e-mail: shibuvij@gmail.com
Background: Private providers (PP) account for 1 million

of the 3 million missing TB case notifications globally.
An estimated 60% of urban population seeks care from
PP. Inaccurate and delayed diagnosis and lack of
treatment adherence are key challenges in the private
sector. We implemented an urban PP engagement model
as Private Provider Interface Agency (PPIA), with
subsidized TB diagnostic and free anti-TB drug services
through vouchers and user-friendly e-health systems for
early TB diagnosis and quality care.
Intervention: Objective was to increase: TB case notification from private sector, microbiological detection,
number of patients accessing Drug sensitivity testing,
treatment adherence, and reduce delay in diagnosis. Thus
private provider engagement and provision of patient
subsidies as incentives to providers for detecting and
treating TB cases as per the Standards of TB Care in India
(STCI) was sought. PPs, laboratories and chemists in the
intervention area were mapped, and prioritized for
engagement. Field officers engaged and provided PP
services through a voucher system, provisioning free
digital Chest X-rays, subsidized CB-NAAT testing, and
free standardized treatment through e-vouchers at
chemists. An efficient sputum sample collection and
transportation system was established to reduce the delay
in diagnosis.Treatment adherence was facilitated by
SMSs and adherence calls by contact center coupled
with health worker personal contact through 15 days
follow-up system established by the NGO stafffor
patients on TB treatment, and patient counselling.
Results: In 7 months, PPIA reached out to 2500 PPs
comprising 42% of relevant provider universe in
Mumbai, and engaged 1426 providers for TB service

delivery. So far, PPs have redeemed ~3500 X-ray
vouchers (9500 USD), 1750 CB-NAAT vouchers (36
000 USD) and ~9000 treatment vouchers (55 000
USD).Out of 3328 TB positive cases notified by PPIA,
40% (1326) cases have been confirmed microbiologically. TB notification increased from 1070 to 6849 in 7
months.
Conclusions: This unique TB service delivery platform
demonstrated an effective model of PP engagement for
quality diagnosis and achieving higher case notification
as per STCI in an urban slum. With appropriate model
tweaking, this could be replicated in other urban setting.
Improving PP quality and achieving scale in PP engagement and patient coverage will further enhance the
program.
10. Whats app doc? Use of m-health

PC-775-04 Effects of electronic tuberculosis
information system on quality of TB data in
Afghanistan
G Q Qader,1 S D Mahmoodi,2 H Manochehr,2 D Azimi,2
M Rashidi,1 M Seddiq,2 P G Suarez3 1Challenge TB,
Management Sciences for Health, Kabul, 2Ministry of Public
Health of Afghanistan, Kabul, Afghanistan; 3Challenge TB,
Management Sciences for Health, Arlington, VA, USA.
e-mail: gqader@gmail.com
Background: The TB data collection and reporting from

health facility till national level were vertical; also, there
were higher discrepancy between data reported by
national health management information system (HMIS)
and TB surveillance data. Also, the timeliness of TB data
was only 65%. To address these challenges, NTP with
technical and financial assistance from USAID funded
TB CARE I project designed and integrated TBIS with
national HMIS. The objective of this assessment was to
assess the outcomes of the integration of TBIS with
national HMIS on the quality of TB data.
Methodology: A team from Challenge TB, NTP and
HMIS unit of MOPH reviewed the data reported
electronically with that send to NTP. We used the
standard TB surveillance reporting and recording forms
and TB electronic reporting system as data collection
tool. The ever first electronic reporting system made in
2010 and piloted in few provinces, revised based on
assessment results. Later, there was shift in implementation strategies from provincial TB coordinators to NGOs
HMIS officers and all trained on it.
Results: Just after the training, 540 (80%) health
facilities send their electronic data within the deadlines
(timeliness) that is 15% improvements compared to the
data accuracy assessment of 2013. Also, the comparison
of electronic data with that of hard copies of the reports
reached to NTP showed a completeness of 95% and
accuracy of 97%. Three (0.4%) health facilities were not
reported electronically.
Conclusion: The integrated electronic reporting system

with national HMIS improved the discrepancy in TB
data between HMIS and NTP surveillance, significantly.
Also, it progressed into improved data accuracy, completeness and timeliness and advanced evidence based
decision making at NGO, PHO and national levels.
Thus, we strongly recommend the maintaining of this
system, its scale up and upgrading with new definitions
for TB reporting.
S185
year in 4Q13 and 1Q14 respectively. After intervention,

annualized NSN case notification rate increased with 15
cases per lakh per year and 14 cases per lakh per year in
2Q14 and 3Q14 respectively.
Conclusions: Due to efforts of using WHATSAPP
application (ICT) in X-ray reading, NSN case notification rate is improved under RNTCP. This study
highlights the need of ICT and innovation to improve
case findings in RNTCP.
PC-776-04 Improving smear-negative case

detection using social media (WhatsApp)
under the Revised National TB Control Program
in Ahmedabad district, Gujarat, India
D Kapadiya,1 P Dave,2 B Vadera,2 P Prakash,1 R Solanki,3
S Yadav,4 B Dave4 1State TB Training and Demonstration
Center, Ahmedabad, 2State TB Cell, Gandhinagar, 3BJ
Medical College, Ahmedabad, 4District Panchayat (Health
Branch), Ahmedabad, India. e-mail: gutbhiv@rntcp.org
Background: Smear negative case detection is low in

Gujarat with annualized new smear negative case (NSN)
notification rate of 11 per lakh population. It is even low
in Ahmedabad district with annualized NSN case
notification rate of 6 per lac population. Lack of
confidence in diagnosing TB based on chest X-ray by
medical officers (MO) posted at peripheral health
institute (PHI) was perceived to be one of the reasons
of low detection of smear negative TB cases. WHATSAPP is a social media mobile application used to share
Images, Video, Messages etc. and majority of mobile
users are familiar with using this application on routine
basis. Use of this social media app was explored to
improve smear negative case detection by transferring
images to referral health centre in Ahmedabad district of
Gujarat, India.
Intervention: The study was conducted between April to
October 2014 at all community health centers where Xray facility and trained X-ray technicians were available.
One health staff was designated as nodal person at
Government X-ray facility who captured and sent images
of chest X-ray in coordination with MO-PHI and X-ray
technician to District TB Officer through WHATSAPP
mobile application. DTO transferred X-ray image from
mobile phone to computer through data cable or
Bluetooth. Then, Chest X-ray reading was done by TB
and Chest specialist available at B.J. Medical College and
State TB training and demonstration center (STDC).
Results of X-ray chest was informed to respective PHI by
message though WHASAPP or telephone call. Quarterly
data reported routinely under programme on smear
negative TB case notification were reviewed and
notification rates before and after intervention was
reviewed.
Results: Total 260 chest X-ray images were received
during study period and all images were reviewed by TB
& Chest specialist and results of chest X-rays were
communicated in time to concern health facilities. Before
intervention, annualized NSN case notification rate was
8 per lakh population per year and 6 cases per lakh per
PC-777-04 The usefulness and feasibility of

mobile phone voice based system for
notification of tuberculosis by private medical
practitioners: a pilot project
B Velayutham,1 B Thomas,1 D Nair,1 K Thiruvengadam,1
K Thiruvengadam,1 S Prashant,2 S Kittusami,2
A Jhunjhunwala,2 S Swaminathan1 1National Institute for
Research in Tuberculosis, Chennai, 2Indian Institute of
Technology Madras, Chennai, India. Fax: (91) 442 836
2525. e-mail: dinanair73@yahoo.co.in
Background: Tuberculosis (TB) is a notifiable disease and

health care providers are required to notify every TB case
to local authorities. TB patients in India are diagnosed
and managed both in the public and private sector.
Various modalities for TB notification is currently being
implemented of which Interactive Voice Response
System (IVRS) has not been explored. We conducted a
pilot study (Mobile interface in TB notification MITUN) to determine the usefulness and feasibility of
mobile phone voice based system for notification of TB
cases by private medical practitioners. This technology
has a unique Multi-lingual Speech recognition Program
for data capture and intelligent analytics tools at the
backend for speech to text conversion enabling remote
voice based data collection and viewing of real time data
on a web portal.
Design/Methods: The study was conducted during
September 2013 to October 2014 in three zones of
Chennai, an urban setting in South India. Private clinics
wherein services are provided by single private medical
practitioners were approached. The steps involved in
MITUN included: Registration of the practitioners and
notification of TB cases by them through mobile phone
based voice interactions. Pre and post-intervention
S186
questionnaires were administered to collect information

on TB notification practices and feasibility of MITUN
after an implementation period of 6 months.
Results: A total of 266 private medical practitioners were
approached for the study. Of them, 184 (69%) participated in the study; of whom 11 (6%) practitioners used
MITUN for TB notification. There were 138 (73%) of
184 practitioners who were aware of the Govt. Order on
mandatory TB notification and 154 (81%) who considered mobile phone as a convenient modality for TB
notification. Reasons for not using MITUN include lack
of time, referral of patients to government facility, issues
related to patient confidentiality and technical problems.
Suggestions from private practitioners for making mobile
phone based TB notification process user-friendly included reducing call duration, including only crucial
questions and using missed call or SMS options.
Conclusion: Mobile phone voice based TB notification
system has to be simple and user friendly for effective use.
Lack of knowledge, perceived problems, logistical and
practical issues preclude scale up of notification of TB
by private practitioners.
is encouraged to use the SMS reporting and is helped by

DP. The Senior Treatment Supervisor (STS) of the
programme monitors a population of 2.5 lakhs with
android phones and electronic treatment cards of all drug
resistant TB patients in his area. When the patient sends
SMS, the treatment cards get automatically updated on
real time basis. Daily log auto generated from the server
enables comprehensive monitoring of patients missing
pills, having side effects and requiring help. STS can then
mobilize resources effectively to address these problems
promptly
Results: Some encouraging results are seen from the
intervention. Since January 2015, 60 MDR-TB patients
reported daily DOT through SMS; 5 reported side effects
and 2 requested medical help which were promptly
attended. Counselling messages were appreciated by all
Conclusions: The SMS/IVRS gateway is a simple,
promising tool to effectively monitor DOT and promote
treatment adherence. Further large scale deployment is
needed to document evidence
PC-778-04 Short message services and

interactive voice response system: gateways to
quality care and TB treatment adherence in
Andhra Pradesh, India
S Achanta,1 J Jaju,1 M Parmar,1 A Sreenivas1 1World
Health Organization Country Office for India, New Delhi,
India. e-mail: achantas@rntcp.org
Background: One important challenge in Programmatic

Management of Drug Resistant TB (PMDT) in India as it
expands is monitoring treatment adherence. There is no
formal counselling system and need of behavioural
change in TB patients or promoting patient participation
in treatment adherence is not systematically addressed.
Programme depends on skills of DOT providers (DP),
largely community health volunteers with limited ability
and informal training to recognize side effects and
counsel. Comprehensive and real time information from
the point of event where DOT occurs could improve
quality of care to patients. TB control measures would be
ideal if situated at sub-district or district level, programme supervisors or managers could know if a patient
in a far-off village has taken his daily medicine today and
if he requires any assistance. Patient must get direct and
easy access to health system to register his experience and
seek help
Intervention: Andhra Pradesh uses a web based programme e-SMARTS for monitoring PMDT .The technical platform of e-SMARTS was used to create 1) An
SMS gateway for drug resistant TB patients to report
daily pill consumption, side effects or need for help
through simple SMS templates- outgoing feature2) An
interactive voice response system (IVRS) service to
inform patients about their test results, send follow up
reminders, auto alerts on missed/delayed doses and
periodic counselling messages incoming feature. The
DP and patient are trained in sending daily SMS. Patient
PC-779-04 Mobile vs. non mobile user rural

health care providers diagnosing TB patients:
comparative analysis from a tribal district in
India
A Trivedi,1 S Chadha,1 S Satapathy1 1International Union
Office, New Delhi, India. e-mail: atrivedi@theunion.org

project funded through Lily Foundation has trained
Rural Health Care Providers (RHCPs), first point of
contact for curative services in many villages especially
in tribal and remote geographic areas in India. A paper
based referral mechanism is used to capture the data on
referrals made, results of sputum examination and the
management of those diagnosed with TB. However
RHCPs and Lab Technicians (LTs), often lack information about referred chest symptomatics. The objective is
to pilot innovative use of mobile technology to improve
referrals and scale up on the basis of results achieved.
Intervention or response: Mobile application developed

is being piloted in three blocks of Khunti, a tribal district
in Jharkhand, India. This mobile platform is easily
customizable- tracks the referred symptomatics, supports
RHCPs and creates a central database on real-time basis.
Two interoperable applications have been developed.
One application is being used by 25 RHCPs and 1 NGO
supervisor and other by 3 LTs.64 trained RHCPs are
referring through non-mobile.
Results and lessons learnt: In 21 months of implementation, 762 (74%) symptomatics have been referred by
25 RHCPs and 264 (26%) by 64 RHCPs. On an
average 32 referrals have been made by each RHCP
through mobile and 4 through non-mobile. Smear
positive 77 (15% sputum positivity) and 27 (11%)
through non-mobile. 86 (71%) TB patients have been
diagnosed through mobile and 35 (29%) through nonmobile. 32 patients are cured and 14 completed
treatment through mobile.14 cured and 18 completed
treatment through non-mobile. Proportion of referrals
reaching DMCs is around 96%-98% from April 2014
onwards. Despite few challenges in implementation like
training of RHCPs on use of android phones and
maintenance of hand sets, this mobile based management
system developed is assisting in follow up of referred
chest symptomatics by RHCPs for diagnosis and
treatment of Tuberculosis efficiently.
Conclusions and key recommendations: Add on benefits
are there in using mobile application. It improves
detection, treatment adherence and better outcomes.
This platform facilitates compilation and analysis of
data. The application is being scaled in other sites. The
pilot project has explored possibility of sustainability.
Implementing NGO Sankalp Jyoti has received adherence scheme which will help to sustain the intervention
after the project ends.
S187
PC-780-04 WhatsApp: could real-time data

sharing enable a complementary surveillance
system? Future of health information
management
S Vijayan,1 A Karad,2 R Gandhi,1 P Kandasamay,1
J Thakker,1 R Chopra,1 M Datta,1 V Jondhale1 1PATH,
Mumbai, 2World Health Organization Country Office for
India, Delhi, India. Fax: (91) 114 605 4430. e-mail:
shibuvij@gmail.com
Background: Real time flow of information is crucial for

healthcare programs heavily based on field workers.With
increasing penetration of internet and smartphones, the
use of Whatsappwhich allows seamless flow of messages, photos and geo location makes it a very handy
tool. Private Provider Interface Agency (PPIA) has set up
virtual communities for sharing information and tracking programme activities via WhatsApp
Comparative analysis: mobile vs non-mobile users at tribal

district in India: Khunti ( May 2013Jan 2015)
Mobile Users
Number of trained
RHCPs referring
symptomatics
Number of referrals
by engaged
RHCPs
Referrals per
engaged RHCP
Referrals reaching
DMC
Smear positive
Sputum positivity
Diagnosed TB
Cured
Completed
treatment
25
762 (74%)
32
534 (70%)* 96-98%
from April 2014 onwards
77
15%
86 (71%)
32
14
Non-mobile
users
64
264 (26%)
4
242 (91%)
27
11%
35 (29%)
14
18
Intervention: PPIA is usingWhatsApp for tracking

sputum collection and transportation, and for real time
flow of information regarding registration of new
patients (paper based registration). Sputum agents collect
the samples from all PPIA networked health facilities and
deliver to nearest laboratory. The PPIA WhatsApp
group of sputum agents and supervisors share key
information for sample testing process such as referral
hospitals, referral physicians name, testsrequested,
laboratorys name, patients name, case ID, test voucher
details with the NGO central office. Real time information enables NGO staff to follow up with the designated
laboratory for test results on the same day. The
information shared on this group is accessible to key
project staff and data entry operators. This also enables
the supervisors to manage and track field activities on a
regular basis. The sputum agent can also share a photo of
a doubtful sample and ascertain the quality of the sample
over a call with the supervisor. WhatsApp group is also
used to track the number of new patients registered daily
S188
and the vouchers issued at each level of diagnosis and

treatment for suspect group. It provides a platform to
share information viz. suspects registered at the health
facility, vouchers issued for- Chest X-ray, Gene Xpert test
(CBNAAT),first line of treatment drugs, TB positive
status and drug resistance status of the patient.
Results: Over 60 program staff are linked viaWhatsApp
data covering 1000 TB suspects was reported and
monitored over 7 months of operation. The real time
data are then collated, and analysed via MS Excel.
Technical difficulties and flaws during field operations
are solved immediately through the experts engaged in
the project
Conclusion: This shows that instant messaging applications establish efficient platforms for bidirectional
communication and data acquisition from an ongoing
program. Advancing this model as a complementary
surveillance approach may increase outcome efficiency
of a program due to instantaneous data and strong
follow up system.
PC-781-04 Optimizing an electronic TB register

to increase tuberculosis case notification and
treatment success in Uganda and South Sudan
M Mubeezi,1 N Sewankambo,2 E Obuku,2 C Nalugwa,1
J Mogga,3 S Macharia,4 H Lasu,3 F Mugabe5 1Wild
Innovations, Kampala, 2Makerere University College of
Health Sciences, Kampala, Uganda; 3Ministry of Health, Juba,
4
Management Sciences for Health, Juba, South Sudan;
5
Ministry of Health, Kampala, Uganda. e-mail:
micah.mubeezi@tbregister.org
Background: Tuberculosis remains a major public health

problem particularly in the 22 high burden countries, of
which Uganda and South Sudan are part. Despite
significant progress in the Directly Observed Treatment
Short Course strategy, with a downward trend of the
Tuberculosis global burden, recording and reporting has
been neglected. Current tools for recording and reporting
Tuberculosis notification and treatment outcomes are
archaic and paper based particularly in the low and
middle income countries. This has resulted in serious
delays in reporting, and when reports are submitted in
time the quality of such reports is below par with poor
estimation of tuberculosis diagnosis and treatment
outcomes. Further, the existing reporting framework
does not allow for disaggregation of individual patient
data to permit more robust analysis for decision making,
and better planning for appropriate interventions, at all
levels of the National TB Control Programs in Uganda
and South Sudan. The advantages of applying an
electronic tuberculosis register have been scarcely documented and not rigorously evaluated. As such, decision
makers in National Tuberculosis Control Programs in
low and middle income countries lack the research
evidence to facilitate the uptake of electronic recording
and reporting of tuberculosis diagnosis and treatment
outcomes.
Intervention: We conducted a study to evaluate the
effectiveness of an electronic tuberculosis register over a
24 months period in 28 health units in Uganda and
South Sudan. We used a mixed-methods operations

research study adopting a quasi-experimental design,
with two intervention and two control TB management
units (between unit comparisons) plus a pre- and postintervention study period (within unit comparison).
Preliminary results: Over the study period, the TB
management units in the intervention area consistently
submitted their quarterly reports within 2 days after the
end of each quarter. The cure rates for the bacteriologically confirmed TB cases improved by at least 70% in
each of the study sites. The treatment success rates for the
newly detected TB cases improved by over 100%. There
was no significant increase in the case notification rates.
Conclusion: Electronic TB recording and reporting
significantly improves TB treatment outcomes, accuracy
and timeliness in reporting.
PC-782-04 Characteristics associated with

mobile phone access among tuberculosis
patients in Pune, India
Y Ogale,1 J Elf,1 R Lokhande,2 S Roy,2 V Mave,2 A Gupta,1
J Golub,1 J Mathad3 1Johns Hopkins Bloomberg School of
Public Health, Baltimore, MD, USA; 2Byramjee Jeejeebhoy
Medical College, Pune, India; 3Weill Cornell Medical College,
New York, NY, USA. e-mail: yasmin.ogale@gmail.com
Background: Mobile health (mHealth) technology is

being considered as a strategy for tuberculosis (TB)
treatment adherence. We aimed to describe mobile phone
ownership and usage among TB patients in Pune, India to
assess the feasibility of a short message service (SMS)
mHealth program to improve TB patient care.
Methods: Consenting adults presenting with active TB
were administered a brief questionnaire capturing sociodemiographics, mobile phone ownership and comfort
using SMS messaging. Those with access to a mobile
phone were contacted via phone call 4 times over 6
months to understand long-term access to the provided
number. Recruitment and enrollment is ongoing; interim
analyses are presented.
Results: A total of 73 adults have been enrolled thus far;
53 (73%) were male and the median age was 32 (IQR:
24, 45) years. Fifty (71%) participants had 5 or more
years of education, and 45 (64%) were currently
employed. Nearly all participants (96%) reported access
to a mobile phone, with the majority (78%) personally
owning a mobile device and the remaining (22%)
borrowing from a family member. Among all participants, 44 (63%) felt uncomfortable sending an SMS
message; 32 (46%) felt uncomfortable receiving an SMS
message. The majority of participants could be reached
by voice call 1 week (86%), 1 month (90%) and 3
months (83%) after initial contact. In multivariate
logistic regression, female participants tended to have
lower odds of mobile phone ownership (OR 0.4; 95%CI:
0.07, 2.0) while those with more than 5 years of
education (OR 6.4; 95%CI: 1.3, 32.6) and those
employed (OR 2.3; 95%CI: 0.5, 10.4) tended to have
higher odds of ownership. Older age groups (31-50 yrs
OR 0.4; 95%CI: 0.08, 1.9 and 51 yrs OR 0.2; 95%CI:
0.02, 2.3) tended to have lower odds of comfort using

SMS messages, while those with more than 5 years of
education (OR 39.9; 95%CI: 4.3, 370.8), those who
were employed (OR 4.5; 95%CI: 0.8, 25.8), and those
who owned a mobile phone (OR 2.5; 95%CI: 0.4, 14.5)
had greater odds of comfort.
Conclusion: TB patients in this population had high
access to, and ownership of, mobile devices. Heterogeneity in ownership and comfort using SMS highlight
potentially vulnerable populations, including women,
older individuals, and those of lower sociodemographic
status. These trends are expected to reach statistical
significance with increasing enrollment. Highlighted
groups may require additional interventions or training
prior to using SMS programs.
11. Basic science: bugs and drugs

PC-783-04 In vitro anti-mycobacterial activity of
selected medicinal plants against
Mycobacterium tuberculosis and
Mycobacterium bovis strains
A G Mekala,1 G Ameni,2 M Giday,2 A Worku2 1Haramaya
Unversity, College of Health & Medical Sciencs, Harar, 2Addis
Ababa University, Aklilu Lemma Institute of Pathobiology,
Addis Ababa, Ethiopia. Fax: (251) 256 668 081. e-mail:
abdgemechu@gmail.com
Background: Tuberculosis (TB) is a global burden with

one third of the world s population infected with the
pathogen Mycobacterium tuberculosis complex and
annually 1.4 million deaths occur due to the disease.
This high incidence of infection and the increased rate of
multi-drug resistant and extensively-drug resistant
strains of the organism further complicated the problem
of TB control and have called for an urgent need to
develop new anti-TB drugs from plants. In this study, the
in vitro activity of root of Calpurnia aurea, seeds of
Ocimum basilicum, leaves of Artemisia abyssinica,
Croton macrostachyus, and Eucalyptus camaldulensis
were evaluated against M.tuberculosis and M. bovis
strains.
Methods: Five Ethiopian medicinal plants, root of
Calpurnia aurea, seeds of Ocimum basilicum, leaves of
Artemisia abyssinica, Croton macrostachyus, and Eucalyptus camaldulensis used locally for the management of
TB. They were investigated for in vitro antimycobacterial
activity against M. tuberculosis and M. bovis strains.
80% methanolic extracts of the plant materials were
obtained by maceration. The antimycobacterial activity
was determined using 96 wells of microplate with the
help of visual Resazurin Microtiter Assay.
Results: The crude 80% methanolic extracts of the root
of C. aurea, seeds of O. basilicum, and leaves of A.
abyssinica, C. macrostachyus, and E. camaldulensis had
anti-mycobacterial activity with minimum inhibitory
concentration (MIC) ranging from 6.25 100 l g/ml.
The MIC of 80% methanol extracts in the order
mentioned above ranged 25-100 l g/ml and 12.5-75 l
S189
g/ml, 25 100 l g/ml and 25 50 l g/ml, 6.25-50 l g/ml

and 12.5-50 l g/ml, 12.5-100 l g/ml and 18.25-50 l g/ml
and 6.25-50 l g/ml and 12.5-50 l g/ml, respectively for
M. tuberculosis and M. bovis strains.
Conclusion: The results support the local use of these
plants in the treatment of TB and it is suggested that these
plants may have therapeutic value in the treatment of TB.
However, further investigations are needed on isolating
chemical constituents responsible for eliciting the observed activity in these plants.
PC-784-04 New horizons for quinolone-class

agents
K Drlica,1 A Mustaev,1 R Kerns,2 X Zhao,1 J Berger3 1Public
Health Research Institute, Rutgers University, Newark, NJ,
2
College of Pharmacy, University of Iowa, Iowa City, IA,
3
Johns Hopkins School of Medicine, Baltimore, MD, USA.
e-mail: nicolas.veziris@upmc.fr
Fluoroquinolones are important second-line agents for

treatment of tuberculosis. However, resistance is a
growing problem that is unlikely to be controlled. As a
key part of the quinolone action mechanism, which is
understood largely from work with non-tuberculosis
bacteria, the compounds traP gyrase on DNA as
reversible complexes. X-ray crystallography of these
complexes has defined the primary binding mode for the
drugs. In that mode, the carboxyl end of the drug forms a
water-magnesium ion bridge with amino acids in the
GyrA subunit of gyrase. Mutation of these amino acids
eliminates the bridge, lowers drug-binding strength, and
causes target-based resistance. Fluoroquinolone-like
compounds called quinazolinediones lack the carboxyl
and serve as a way to bypass existing resistance. Efforts
are being made to increase the binding strength of
quinazolinediones. We recently found a second mode of
quinolone binding through quinolone-gyrase crosslinking: the C7 end of quinolones rather than the C3
carboxyl interacts with GyrA. One speculation is that
the second mode involves a 180 degree inversion of the
drug. Surprisingly, adding a phenyl ring to the C7 end
removes much of the protective effects of GyrA resistance
substitutions, thereby providing a second way to bypass
existing target-based resistance. Diones and C7-phenyl
derivatives are being modified to improve rapid lethality,
limit the mutagenic SOS response, and thereby restrict
the emergence of new resistance. Pathways leading to cell
death are also being studied to identify targets of small
molecule enhancers of lethality (one enhancer targeting
DNA recombinational repair increases the lethal action
of quinolones with mycobacteria). Two lethal pathways
exist, one that involves the accumulation of toxic reactive
oxygen species and one that does not. How drug
structure and binding to gyrase influences lethal action,
which is mechanistically distinct from blocking growth,
is not known. Work on action mechanism is leading to
new ways to design quinolone-class compounds to
bypass existing resistance. Attention on lethal action,
which is crucial for slowing the emergence of resistance,
is the subject of ongoing work. Involvement of reactive
oxygen in drug lethality suggests follow-up work to
S190
determine whether consumption of anti-oxidants interferes with drug lethality.
PC-786-04 Phosphate ABC transporter systems

regulate the level of rifampicin resistance in
Mycobacterium tuberculosis
PC-785-04 Study of herbal bioenhancers extract

on various characteristics of isoniazid and
rifampicin microspheres
M Grobbelaar,1 S Sampson,1 E Louw,2 T Victor,1

P Van Helden,1 R M Warren,1 M De Vos1 1Stellenbosch
University, Cape Town, South Africa; 2National Institutes of
Health, Bethesda, MD, USA. e-mail: melgrob@sun.ac.za
P Pingale,1 R.P. Ravindra2 1GES Sir Dr M S Gosavi College of

Pharmaceutical Education & Research, Nashik, 2Dharamitra,
Wardha, India. Fax: (91) 253 223 2799. e-mail:
prashant.pingale@gmail.com
Background: Bioenhancers are agents when combined

with an active drug lead to potentiate the pharmacological effect of the drug but themselves are not having any
therapeutic effect. Recent development in bioavailability
enhancement of drugs by herbal compounds has produced a new shift in the way of therapeutics. Bioenhancers improve permeability and influence bioavailability of the drug for therapeutic efficacy which may
result into decrease in dose ensuring therapeutic availability. The major problems associated with the anti-TB
drug therapy include loss of efficacy through bacterial
resistance, side effects, and low patient compliance. In
the cited research work, we have attempted to solve the
above mentioned problems by simultaneous use of both
the concepts of microsphere drug delivery (higher
bioavailability) and herbal bioenhancers (improved
bioavailability).
Methods: Complex coacervation method and modified
emulsion method. Materials: Isoniazid and rifampicin as
a model drug used in TB therapy. Hydro-alcoholic
extracts of Carum carvi and Ocimum sanctum were
used as bioenhancer.
Results: The in-vitro drug release of isoniazid and
rifampicin from formulations where bioenhancer extract
is used was found to be about 85-90% in 12 hrs. While
the same was about 45-50% in case of formulations
where bioenhancers were not used. Other parameters
like percentage bioadhesion, permeability study using
intestinal sac method etc. were also studied.
Conclusion: For microspheres prepared by both complex
coacervation and modified emulsion methos, the particle
size was uniform and found to be less than 130 micron in
size; the particle size may alter on addition of bioenhancer extract to the formulations. The drug encapsulation efficiency was found to be in the range of 43-67%
(increased on addition of bioenhancer in the formulations). The percentage bioadhesion of the microsphere
where bioenhancers were used found to increase by 20%
as compared to basal value (upto 67% from 42%). The
in- vitro release study by USP paddle apparatus and the
most important results from the in-vitro release study
relates to the very significant enhancement in drug
release (45 to 90% for microspheres prepared by
modified emulsion method and 47 to 90% complex
coacervation method), due to presence of bioenhancers
alone and in combination.
Background: Evidence demonstrating the differential

dynamics in the population structure of Mycobacterium
tuberculosis within a clinical specimen has been reported
recently. It is thus suggested that the phenotype of an
entire population reflects the average of the phenotypic
characteristic of individual clones within a population.
We hypothesise that the resistance phenotype, as
measured by the minimum inhibitory concentration
(MIC), reflects an average of the MICs of individual
clones present in the clinical specimen.
Design/Methods: Single cell colonies from M. tuberculosis rifampicin mono-resistant clinical isolates with
genetic backgrounds representing the X and Beijing
genotypes were purified. Subsequent selection criteria
included the presence of the rpoB S531L mutation and
the differential low and high rifampicin MICs for each
genotype. Genomic DNA was extracted from each single
cell culture and sequenced using the Ion Torrent
platform. Single nucleotide polymorphisms and InDels
were called using the bioinformatic suite, Ion Reporter.
All variants were confirmed by PCR and Sanger
sequencing.
Results: The level of rifampicin resistance for both the X
and Beijing genotype were as follows: the low MIC
40lg/ml and high MIC .150lg/ml. In the X genotype, 2
unique variants were identified in the high MIC single
cell, with 3 variants in the low MIC. Additionally, in the
Beijing genotype, 5 unique variants were identified in the
high MIC single cells, while 8 variants were identified in
the low MIC. The low MIC colony with an X genotype
showed a 7bP insertion in phosphate ABC transporter
phoT, while the low MIC Beijing genotype colony
showed a 6251bP deletion spanning the phosphate
transporters pstC2, pstS2, pstA1, pstB, and pstS1.
Conclusion: The insertion in or deletion of ABC
transporters may be linked to the low levels of rifampicin
resistances within the single cell colonies isolated from
two different clinical isolates. We postulate that these
natural knockouts demonstrate low rifampicin MICs as
they are unable to extrude rifampicin efficiently. Phosphate ABC transporters may be important targets for the
development of adjuvant drugs to improve the efficacy of
TB treatment.
PC-787-04 Clofazimine: mechanism of

resistance within M. tuberculosis
H Visser,1 M De Vos,2 R Van Der Merwe,3 T Victor,4
P Van Helden,5 R M Warren,6 L Paul7 1University of
Stellenbosch, Cape Town, South Africa. e-mail:
hvisser@sun.ac.za
Background: The observed increase in the global

incidence of drug resistant tuberculosis (TB) has renewed
the search for new anti-TB drugs and the repurposing of

drugs that are currently used to treat other pathogenic
infections including other species of mycobacteria. One
such repurposed drug, is clofazimine (CFZ), which is
currently used for the treatment of leprosy, caused by
Mycobacterium leprae. The mechanism of action of CFZ
is not clear, but it is hypothesized that CFZ is reduced by
a mycobacterial type II NADH oxidoreductase. The
reduction of CFZ drives the production of reactive
oxygen species which are toxic to the pathogen. The aim
of this study was to elucidate the mechanism of CFZ
resistance.
Methods: Spontaneous in vitro CFZ resistant mutants
were selected. Whole genome sequencing was done to
identify SNPs which may cause CFZ resistance.
Results and Discussion: Mutations were identified in the
transcriptional regulator Rv0678 and a fatty-acid-AMP
ligase (FadD28, Rv2941). Mutations in Rv0678 have
previously been shown to play a role in both CFZ
resistance and bedaquiline (BDQ) cross-resistance. No
link has been found between CFZ resistance and
mutations in fadD28. The novel SNP identified in
Rv0678 resulted in the replacement of an alanine residue
with threonine at codon 84, located in the DNA binding
domain. Virtual modeling of the mutated Rv0678
protein showed that the A84T mutation caused a change
in hydrophobicity, which may influence binding to DNA.
Previous studies showed that mutations in Rv0678
resulted in the upregulation of mmpL5, a putative efflux
pump. However, the mechanism whereby CFZ resistance
occurs via increased abundance of this efflux pumP in the
cell wall is not clear and needs further investigation.
Conclusion: Cross-resistance between CFZ and BDQ,
caused by mutations in Rv0678, is of concern and may
influence the planning of anti-TB drug regimens for the
future. The role of the other mutation which causes CFZ
resistance is not clear and requires further investigation.
Finally, the findings of this study support the role of
Rv0678 in CFZ resistance thereby suggesting that this
gene could be useful as a diagnostic marker to test for
CFZ resistance in clinical isolates.
PC-788-04 Whole genome analysis of IS6110

insertion sites in two closely related
Mycobacterium tuberculosis Beijing strains
with distinct pathogenic phenotypes
K Siame,1 R M Warren,1 P Arnab,2 A Abdallah,2
A Christoffels,3 N Gey Van Pittius,1 S Sampson1
1
Biomedical Sciences, DST-NRF Centre of Excellence for
Biomedical Tuberculosis Research/SAMRC Centre for
Human Genetics, Faculty of Medicine and Health Sciences,
Stellenbosch University, Tygerberg, Cape Town, 2Pathogen
Genomics Laboratory, BESE Division, King Abdullah
University of Science and Technology, Thuwal, 3South African
National Bioinformatics Institute, University of the Western
Cape, Cape Town, South Africa. e-mail: kksiame@sun.ac.za
Background: A recent molecular epidemiological study

showed that Mycobacterium tuberculosis Beijing genotype strains, isolated from patients in Cape Town, South
S191
Africa, were either clustered or unique according to

IS6110 RFLP. Two isolates reflecting either ongoing
transmission (clustered) or the absence of transmission
(unique) were shown to have a hyper-virulent or hypovirulent phenotypes in a murine infection model,
respectively. Additionally, the hyper-virulent strain elicited an anti-inflammatory TH2 immune response in the
murine model whilst the hypo-virulent strain had a proinflammatory TH1 immune response. The genetic
mechanism(s) underlying these contrasting phenotypes
remain to be elucidated. Objective: This study aimed to
use whole genome sequencing to determine whether
IS6110 transposition could account for the different
virulence phenotypes seen in these Beijing genotype
clinical isolates.
Design/Methods: Whole Genome Sequencingof the
hypovirulent and hypervirulent strains was performed
on an Illumina platform generating 105 bP paired-end
reads and on a PacBio platform generating long single
end raw reads. De novo assembly was performed using
MIRA software to generate a hybrid assembly using
Illumina and PacBio raw reads. Gap closure among
contigs from the MIRA assembly was done using
Platanus genome assembly software; ordering of the
assembled genome was done with ABACAS software.
The first and last 140 bases of IS6110 in the super-contig
were then searched for to reveal the sequence adjacent to
the IS6110 location. NCBI BLAST was used to determine
the location of the IS6110 element with respect to M.
tuberculosis H37Rv or CDC1551 reference genomes
Results: The hypo-virulent and hyper-virulent strains
shared 22 IS6110 insertion positions and differed only by
an IS6110 insertion at position 2,165,961 in the
hypervirulent strain. This insertion disrupted the gene
ppe34.
Conclusion: Our results showed that ppe34 was uniquely
disrupted by an IS6110 element in the hyper-virulent
strain. Previous studies have suggested that a domain
encoded by ppe34 elicits a TH2 immune response, which
is also the case for the hypervirulent strain albeit having
the aforementioned domain disrupted. Further investigation will be required to determine whether the IS6110mediated disruption of ppe34 plays a role in modulating
the TH1/TH2 balance, or whether other mechanisms
independent of ppe34 involvement are driving this
differential host response.
PC-789-04 Global high-throughput analysis of

DNA-binding proteins in Mycobacterium
smegmatis
N Steyn,1 T Heunis,1 P Van Helden,1 R M Warren,1
M Williams,1 S Sampson1 1Biomedical Sciences,
Stellenbosch University, Cape Town, South Africa. e-mail:
nastassja@sun.ac.za
Background: Elucidating changes in the Mycobacterium

tuberculosis transcriptome in response to stresses encountered within the host environment, including drug
treatment, is critical for understanding pathogen adaptation to these adverse conditions. These studies are often
S192
limited to micro-array or ChIP-seq analysis where the

transcription factor initiating a transcriptional response
is known. A large number of predicted genes in the M.
tuberculosis genome have unknown functions, and it is
therefore plausible to hypothesize that a subset of these
proteins are involved in transcriptional regulation. In this
study we aim to develop a method that will identify novel
DNA binding proteins in M. tuberculosis. For the
purposes of establishing this method we used the closely
related, non-pathogenic and fast growing organism
Mycobacterium smegmatis.
Design/Methods: M. smegmatis was cultured under
standard conditions to early log phase and then treated
with formaldehyde to crosslink protein-DNA complexes.
The complexes were captured onto a solid matrix using
an anti-RNA polymerase antibody and an on-matrix
tryptic digestion performed. Eluted peptides were analyzed by high resolution LC-MS/MS and DNA binding
proteins were identified using the Andromeda search
algorithm as part of MaxQuant 1.2.2.5.
Results: Preliminary data obtained using M. smegmatis
confirmed that our method selectively enriched for
proteins known to be bound to DNA. These included
the RNA polymerase complex, ribosomal proteins and
other RNA associated proteins. In addition, we identified
proteins required for energy metabolism. The reproducibility of this method is currently being evaluated.
Conclusion: The identification of proteins that are
associated with DNA in M. tuberculosis will provide
insights into the essential cellular processes of transcription and potentially identify novel proteins involved in
the regulation of these processes.
PC-790-04 Expression and purification of

Mycobacterium tuberculosis Rv1460, a possible
SUF system regulator
(600 nm) at 308C. Protein expression was induced by the

addition of 0.4mM IPTG and incubation at 13oC for 24
hrs. Cell extract from 2.5L of culture was loaded onto a
5ml HiTraP nickel IMAC column in the presence of
20mM imidazole, washed and then eluted with a 20 to
500mM linear imidazole gradient. Fractions containing
the fusion protein were pooled, concentrated and applied
to a Sephadex S200 gel filtration column. Protein purity
was assessed by SDS-PAGE.
Results: SDS-PAGE and western blot analysis revealed
that the majority of the Rv1460-fusion protein was
present in the insoluble fraction, although sufficient
soluble protein was present for purification. Interestingly, the protein was visible as a monomer (~35 000 Da)
and a homodimer (~70 000 Da) on a reducing SDSPAGE gel. The elution profile from the IMAC column
showed two major protein peaks eluting within the
imidazole gradient, and the second peak was confirmed
to contain the Rv1460-fusion protein by SDS-PAGE and
western blot analysis. SDS-PAGE analysis revealed that
peaks eluting from the S200 gel filtration column
corresponding to the size of the monomer and dimer
both contained the fusion protein. This is consistent with
the finding that cyanobacterial SufR functions as a dimer.
The purity of the protein following gel filtration was
assessed to be approximately 90%.
Conclusion: Optimised purification of Rv1460 will
enable us to characterise its DNA-binding properties
and provide insight into its role as a mycobacterial
transcriptional regulator.
PC-791-04 Persistence of bacilli during

treatment of TB is associated with changes in
colony lag phase
D Willemse,1 B Weber,2 R M Warren,1 M Williams1

1
Biomedical Sciences, Stellenbosch University, Cape Town,
2
University of Cape Town, Cape Town, South Africa. Fax:
(27) 219 389 863. e-mail: dwillemse@sun.ac.za
D Barr,1,2,3 M Kamdolozi,4 Y Nishihara,4 V Ndhlovu,4

J Meghji,1 M Khonga,4 G R Davies,2 D Sloan3 1Wellcome
Trust Liverpool Glasgow Centre for Global Health Research,
Liverpool, 2University of Liverpool, Liverpool, 3Liverpool
School Tropical Medicine, Liverpool, UK; 4College of
Medicine, University of Malawi, Blantyre, Malawi. e-mail:
david.barr@liv.ac.uk
Background: Iron-sulphur clusters (Fe-S) are protein

cofactors involved in multiple cellular processes. Clusters
are sensitive to oxidation and are therefore synthesised
by Fe-S cluster biogenesis systems. M. tuberculosis
contains only the sulphur mobilization (suf) Fe-S cluster
biogenesis system, encoded in an operon (Rv1460Rv1461-Rv1462-Rv1463-csd-Rv1465-Rv1466) (suf operon) consisting of seven co-transcribed genes. Rv1460
encodes a probable transcriptional regulator that shares
homology with the cyanobacterial sufR gene. The
Rv1460 encoded protein of M. tuberculosis has predicted
DNA- and iron-binding domains, but its role in
regulating the suf operon and in physiology as a whole
is unclear.
Design/Methods: The Rv1460 gene was amplified and
cloned into the pET28a expression vector to generate an
N-terminal His-tagged fusion protein. The Escherichia
coli ArcticExpress (DE3) strain transformed with the
construct was cultured in LB broth to an OD of ~0.5
Background: To meet global TB control targets, shorter

TB treatment regimens that can still reliably deliver low
relapse rates are needed. Sterilising cure without relapse
is hypothesised to depend on eradication of drug tolerant
sub-populations of M. tuberculosis (MTB), made up of
slower growing persister bacilli. No well validated
biomarker for sterilising activity of anti-tuberculosis
drugs exists. Indices derived from serial sputum colony
counts and time to detection of growth in liquid culture
(TTD) show potential as surrogates, but neither directly
measures elimination of sub-populations. Measuring
time to colony appearance (lag time) on solid agar might
identify slower growing persister sub-populations because they will appear later on the plate.
Design/Methods: Sputum samples from smear or
pulmonary TB patients in Blantyre, Malawi were
collected from patients at 0, 2 and 4 weeks into standard
short course therapy and inoculated without decontamination at serial dilution onto selective 7H11 plates.
Resulting colony growth was digitally imaged. Colony

measures were made with OpenCFU software. Lag time
and radial growth rate of individual colonies were
extracted using custom written R programming scripts
to attribute unique identities to individual colonies in
serial images.
Results: Measurable colony growth was recorded for 20
patients at baseline, 10 patients at two weeks, & 6
patients at four weeks. 1400 colonies were observed.
Mean lag time of observed colonies increased between
baseline and two weeks then plateaued. Growth rate was
higher for colonies from week 4 samples. A mixed-effects
model, accounting for random variation between patients, suggested that colony lag time increased by 6.6
days between baseline and two week samples (fixed
effect, P 0.01). While total colony counts per ml
sputum showed typical biphasic elimination, counts of
colonies with lag time .20 days declined more slowly
and linearly (Figure).
Conclusion: Study of individual TB colony growth may
identify drug tolerant MTB sub-populations, as shown
by differing bacillary elimination profiles when colonies
are disaggregated by lag time. This is the first quantitative description of TB colony dynamics, and our findings
suggest that colony phenotyping could add value to
surrogate markers of efficacy in phase II TB treatment
trials, which would be an important step towards shorter
regimens.
Results: Our results showed that 34.13% (71/208) of

recruited cases had a positive result for Mycobacterium
tuberculosis culture from pus, while 76.44% (159/208)
of recruited cases showed positive results for Xpert.
Xpert is not only an alternative diagnostic approach for
BJTB but also can be further used for the detection of
drug resistance, the results of which were consistent with
the classic drug susceptibility test. However, it further
provided the molecular profile of the mutations can be
conducted much faster than the classic drug susceptibility
test can. In the studied 208 pus samples from patients
with BJTB, we found a rate of 14.08% for rifampicin
(RIF) resistance, and14.08% for multidrug-resistant
tuberculosis, which is similar to the reported resistance
in pulmonary tuberculosis. The molecular profile indicated that probes associated with the observed RIFresistance were as follows: E (79.17%, 19/22), D
(8.33%, 2/22), B (4.17%, 1/22) and no RIF-resistance
was associated with probe A and C.
Conclusion: Our data suggest that Xpert is capable of
detecting drug resistance for pus samples from BJTB. We
recommend Xpert as routine assays for patients with
suspected BJTB.
Table Demographic profile and clinical features of patients
with BJTB who tested positive by culture and/or Xpert
Characteristics
No. (%)
of culturepositive
patients
(n71)
No. (%)
of Xpert positive
patients
(n88 )
P values
Median age, years (rang)
44(14-84)
29(3-80)
0.029
Male sex
35(49.3)
46(52.27)
0.709
Concomitant TB
36(50.7)
57(64.77)
0.073
History of TB contact
12. TB diagnostic challenges: molecular

and other techniques
PC-792-04 Molecular profile of drug resistance
in bone and joint tuberculosis in China
G Wang,1 H Huang1 1Beijing Chest Hospital, Capital
Medical University, Beijing Tuberculosis and Thoracic Tumor
Institute, Beijing, China. e-mail: 23332978@qq.com
Background: Bone and joint tuberculosis (BJTB) constitutes about 5-10% of the extrapulmonary tuberculosis
(EPTB). Xpert MTB/RIF (Xpert) has been endorsed by
WHO for diagnosis EPTB. Using Xpert, we investigated
the drug resistance in BJTB patients.
Design/Methods: 208 pus specimens were obtained from
orthopedics patients. Culture and Xpert were performed
for each specimen, while MGIT960 based drug susceptibility testing (DST) was conducted for all the recovered
isolates. The efficiency of Xpert was evaluated based on
the bacteriological examination outcomes.
S193
2(2.82)
4(4.55)
0.693
Smoking
Alcohol
13(18.31)
8(11.27)
9(10.23)
5(5.68)
0.142
0.201
Average treatment time at

admission (day)
68
Clinical features at admission

Fever
Night sweat
Weight loss
Lethargy
Localised pain
Swelling
Restriction of movement
Neurological deficit
MRI and CT abnormalities
Bony destruction
Soft-tissue involvement
Histopathological features
consistent with TB
107
0.12
18(25.35)
20(28.17)
25(35.27)
19(26.76)
67(94.37)
19(26.76)
57(80.28)
27(38.03)
40(45.45)
29(32.95)
30(34.09)
12(13.64)
79(89.77)
22(25)
57(64.77)
29(32.95)
0.009
0.516
0.883
0.038
0.293
0.801
0.031
0.443
64(90.14)
51(71.83)
70(79.55)
67(76.14)
0.068
0.537
62(87.32)
70(79.55)
0.194
S194
PC-793-04 Comparison between fine needle

aspiration cytology and GeneXpertW in the
diagnosis of TB lymphadenitis in Ethiopia
Table Summary of readings of FNA Cytology and Gene Xpert

test in the diagnosis of TB lymphadenitis
T Gemechu,1 M Aseresa Melese,2 B Tessema,3

N Demmelash,2 D Habte,2 D J Dare,2 P G Suarez4
1
Department of Pathology, College of Health Sciences, Addis
Ababa University, Addis Ababa, 2Management Sciences for
Health, Help Ethiopia Address the Low TB Performance (HEAL
TB) Project, Addis Ababa, 3Department of Medical
Microbiology, School of Biomedical and Laboratory Sciences,
College of Medicine and Health Sciences, University of
Gondar, Gondar, Ethiopia; 4Management Sciences for
e-mail: keeti.gemechu@gmail.com
Cytology TB
Cytology non-TB
Total
Background: Lymphadenitis is the most common form of

extra-pulmonary tuberculosis in Ethiopia. Fine-Needle
Aspiration (FNA) offers a feasible and safe option for
specimen collection for cytological examination, culture
and GeneXpert tests. Cytology microscopic reading is an
old diagnostic tool, which is simple, rapid, and performance-friendly method. However, FNA cytology is
performed only in few hospitals in Ethiopia due to a
limited number of pathologists and the introduction of
Xpert might be an alternative test.
Design/Methods: A comparative cross-sectional study
was conducted in a tertiary hospital between September
2014 and February 2015 to compare the cytological
results of the senior pathologists and Xpert results. A
senior pathologist took a split sample of FNA from
patients clinically suspected of TB lymphadenitis. The
cytology smears were stained with Wrights stain and
read by the same senior pathologist. A senior microbiologist ran the Xpert test following the standard
procedure. Both the pathologist and the microbiologist
were blinded for the results of each other. We calculated
the concordance rate of the two test results using Kappa
statistics.
Results: A total of 118 TB lymphadenitis suspected
patients were included in this study with male to female
ratio of 0.72. From the total of 118 TB lymphadenitis
suspects, 89 (75.4%) and 64 (54.2%) were diagnosed as
TB by cytology reading and Xpert, respectively. Four
(6.3%) of the TB patients diagnosed by Xpert were found
to have Rifampicin Resistant Tuberculosis. Out of 89
patients diagnosed as TB lymphadenitis by cytology, 63
(70.7%) were diagnosed as TB by Xpert. The cytology
labeled 29 of the suspects as no TB of which 28 (96.6%)
were also negative for TB by Xpert. The agreement level
between the two tests was moderate (Kappa 0.52, P ,
0.001). The TB positivity rate is significantly less in
caseous aspirates samples as compared to the pus
(Cytology 65.3% vs. 88.3% and Xpert 38.8% vs.
70.5%, respectively)
Conclusion: Gene Xpert was also able to detect a larger
proportion of the TB cases diagnosed by Cytology. In
settings where there is no pathologist to provide cytology
service, the new diagnostic tool could fill the gap. Getting
enough samples for caseous aspirates was a challenge
and the low yield among caseous samples could be due to
insufficient sample.
Xpert TB
Xpert non-TB
Total
63
01
64
26
28
54
89
29
118
PC-794-04 Need to adapt better diagnostic

tools to improve the diagnostic accuracy of
sputum-negative TB cases in India
S Dapkekar,1 R Yeole,1 S Kamble,2 P Parmar,1
A Sreenivas1 1World Health Organization, Country Office for
India, New Delhi, 2State Tuberculosis Office, Pune, India.
e-mail: drshankardapkekar@gmail.com
Background: National Tuberculosis Programme of India

(RNTCP) has vision to accurately diagnose and effectively treat all TB patients including drug-resistant (DR)
cases. From inception of RNTCP, diagnosis of TB
depends on sputum microscopy. Presumptive pulmonary
TB case found positive on sputum microscopy are typed
as sputum positive TB and those tested negative on
sputum microscopy but X-ray finding suggestive of TB
are been typed as sputum negative TB as per current
diagnostic algorithm under RNTCP. Previously treated
pulmonary TB cases are been offered drug susceptibility
testing through molecular diagnostic tool through
programme assuming they are presumptive MDR-TB
cases. Concerns have been raised about the high
proportion of Mycobacterium tuberculosis not detected
results on Xpert MTB/RIF testing for those already
diagnosed as pulmonary TB cases in the field. We felt this
study as unique opportunity to compare results of Xpert
MTB/RIF test with basis of diagnosis of TB.
Methods: In this cross sectional study, we collected data
on presumptive pulmonary MDR-TB cases with history
of previous anti-TB treatment having MTB not detected
results from January to December 2014 at 3 Xpert MTB/
RIF lab sites and compared it with basis of diagnosis of
TB in them.
Results: Total of 1553 pulmonary TB cases with past
history of anti-TB treatment were tested for DST at 3
Xpert MTB/RIF lab sites in 2014. Among them 891
(57%) were sputum positive TB cases and rest were
sputum negative. 603 (39%) of total tested for DST had
M. tuberculosis not detected results. M. tuberculosis not
detected results proportions were 19% and 66% among
sputum smear positive and sputum smear negative
previously treated pulmonary TB cases respectively.
Conclusion: World Health Organization has recommended use of Xpert MTB/RIF test for diagnosis of TB
depending on its accuracy. It is concerning to know very
high M. tuberculosis not detected results on Xpert MTB/
RIF test among those diagnosed as TB on the basis of Xray findings. Many studies have indicated that dependency on X-ray for diagnosis of TB has risk of false
positive diagnosis to the extent of 40%.
Results of this study reiterate their findings. Though
resource intensive, there is urgent need to adopt
molecular tests like Xpert MTB/RIF instead of depending
S195
on X-rays in RNTCP diagnostic algorithm for improved

accuracy. Further research is needed to understand why
19% of cases in the study were sputum positive in the
field but M. tuberculosis negative on Xpert MTB/RIF
test.
TB could be taken early during the disease. Our results

concur with the newly formulated national policy to
provide Xpert MTB/RIF upfront for diagnosis of EPTB.
To resolve test failures, procedures of retesting & time
lag till retesting need further standardisation.
PC-795-04 Diagnosis of drug-resistant

tuberculosis by XpertW MTB/RIF (CBNAAT) in
extra-pulmonary samples: a retrospective
analysis from Central India
PC-796-04 Utility of GeneXpertW for diagnosis

of rifampicin-resistant extra-pulmonary
tuberculosis: a retrospective review of 70 cases,
SGH, Pune, India
P Desikan,1 N Panwalkar,1 S Dhawan,2,3 K Rade,3

S Mannan,3 L Mehandru,3 N Kulshreshta,4 M Pandey1
1
National Reference Library, Bhopal Memorial Hospital and
Research Centre, Bhopal, 2International Union Against
Delhi, 3World Health Organization, Country Office for India,
New Delhi, 4Central TB Division, New Delhi, India. e-mail:
prabhadesikan@yahoo.com
S Girish,1 A Kagal,1 S Dharmshale,1 R Bharadwaj,1

A Chandanwale,1 R Lokhande1 1Byramjee Jeejeebhoy
Government Medical College, Pune, India. Fax: (91) 20
2612 6868. e-mail: sunitagirish@rediffmail.com
Background: 10 to 15% of TB cases in India are

estimated to have extrapulmonary TB (EPTB). Its
diagnosis is challenging due to the variety of sample
types and protocols for sample processing. In March
2014, Indias National TB Control Program adopted the
Standard Operating Procedure for use of Xpert MTB/RIF
in extrapulmonary (EP) samples. We aim to understand
the performance of Xpert MTB/RIF on EP samples in the
Central state of Madhya Pradesh, India.
Design/Methods: We retrospectively reviewed the results
obtained on EP samples by Xpert MTB/RIF at a National
Reference Laboratory for TB, Bhopal Memorial Hospital
and Research Centre (BMHRC), Bhopal from April 2014
to March 2015. In addition to valid results, there were
errors, invalids and no results. Samples with such results
were retested from the samples kept in the refrigerator.
We tried to understand the utility of retesting samples
kept at 4oC till the time of retesting.
Results: 317 EP samples (86 tissue biopsies, 115 fluids,
46 aspirates & 70 pus) were tested by Xpert MTB/RIF.
Of these, 18 (5.6%) were smear positive & 299 (94.3%)
smear negative. From 317 samples, TB was detected in
88 (27.7%) samples-Rif resistant (19.3%, 17/88), Rif
sensitive (80.6%, 71/88). TB was not detected in 175
(55.2%) samples, while there were 54(17%) samples
which needed retesting (8 indeterminates, 40 errors, 5 no
results & 1 invalid result).Of the 17 Rif resistant
samples, 2 were pleural fluids, 1 empyema pus, 2 CSF,
6 lymph nodes, 2 granulation tissue from spine, 3
osteomyelitic bones & 1 synovial fluid.Of the 54 samples
needing retesting, 15 samples were not available for
retesting. On retesting the remaining 39 samples kept at
48C, 85 % samples (33 out of 39) could be resolved
giving an additional gain of 5/ 54(9%) TB cases.Unresolved results out of the 317 tests were 6 (2%) (2 pus, 1
CSF, 1 Ascitic Fluid, 1 Granulation Tissue, 1 FNAC).
Conclusion: Our results underline the utility of Xpert
MTB/RIF in diagnosis of TB including Rif resistance in
EP samples. Xpert MTB/RIF positivity over microscopy
was 22% (70/317). By Xpert MTB/RIF, 19.3% Rif
resistance was detected from EP samples within 2 hours
and decision for initiation of treatment of drug resistant
Background: India lacks evidence-based guidelines for

the diagnosis of extra-pulmonary TB (EPTB), resulting in
empiric, standard TB treatment of patients, based on
clinical symptoms and for prolonged duration of 12-24
months. Culture isolation rates are usually poor as is
smear positivity. With the current MDR tuberculosis
crisis, it has become even more important to not only
isolate the organisms but also do a sensitivity testing. To
address these issues we evaluated the sensitivity and
specificity of GeneXpert diagnosis of EPTB and the
prevalence of Rifampicin resistant tuberculosis in patients with extra pulmonary tuberculosis in a large,
urban Indian hospital setting.
Design/Methods: During the period January 2013 to
August 2013,Extra pulmonary samples (70) which
included pus,pleural fluid, Urine, pericardial fluid,CSF,lymphmode aspirate, endometrial tissue and ascitic
fluid were obtained from patients, clinically diagnosed as
EPTB. All specimens were tested by GeneXpert, ZiehlNeelsen stained and cultured using standard protocol
methods.
Results: A total of 70 patients of EPTB were identified.
The mean age of the patients was 34 6 16.4 years, and
male to female ratio was 1:1. Lymph node aspirates (23),
pleural fluid (20) and cerebrospinal fluid (13) were the
most common samples. All 70 samples were subjected to
processing by GeneXpert and 62 were cultured on LJ
medium. GeneXpert could detect Mycobacterium tuberculosis in 29 samples (41.4 %) and 9 (31%) of these were
Rifampicin resistant. Out of 62 samples, cultured growth
was obtained in 16 (25.8%) samples. The sensitivity and
specificity of GeneXpert as compared to culture was
81.3% and 72.7%.respectively and negative predictive
value was 91.43 % (*)indicating probability that the
disease is not present when the test is negative.
Conclusion: The current study showed the prevalence of
31% of Rifampicin resistant TB, and picked up an
additional 13 cases which were negative by culture.
Hence GeneXpert is a rapid and sensitive method for
detecting M. tuberculosis in extra pulmonary samples
which will allow early and appropriate treatment of TB.
The National TB Program and chest and infectious
disease societies should effectively implement GeneXpert
for the early diagnosis and initiation of treatment of
EPTB.
S196
Table RIFA resistance and sensitivity in total EPTB cases

GeneXpert Report
not detected
MTB detected Rif sensitive
MTB detected Rif resistance
MTB detected Rif indeterminate
Total
No. of cases
41
20
9
4
70
58.6
27.1
8.6
5.7
100
PC-797-04 Clinical features of urogenital

tuberculosis
E Kulchavenya,1 D Kholtobin1 1Novosibirsk research TB
Institute, Medical University, Novosibirsk, Russian Federation.
e-mail: ku_ekaterina@mail.ru
Background and challenges to implementation: Urogenital tuberculosis (UGTB) is one of the most common
forms of tuberculosis (TB) after pulmonary TB.
Intervention or response: With purpose to estimate
clinical features of UGTB we analyzed history cases of
131 patients who were under supervision in Novosibirsk
anti-TB dispensary in 2008-2011 years.
Results and lessons learnt: Among 131 pts with UGTB 88
(67.2%) had isolated kidney TB (KTB): 10 pts (10.2%)
TB of parenchyma, 35 pts (39.8%) papillitis, 22 pts
(22.4%) - cavernous KTB, 21 pts (21.4%) - polycavernous KTB; in 10 pts alongside with polycavernous KTB
male genital TB (MGTB) was diagnosed. In 33 pts (25.2)
MGTB only was revealed: in 14 orchiepidydidimitis,
and in 19 prostate TB. Main clinical features were pain
(flank or perineal), dysuria, hematuria, hemospermia,
toxicity, but their frequency varied from 0 till 60.0% in
different groups. Among all cohort of UGTB asymptomatic course was in 12.2%, among kidney TB - in 15.9%.
Every third patient complained of flank pain and dysuria
(accordingly 35.2% and 39.8%), 17% presented toxicity
symptoms, 9.1% - renal colic, 7.9% - gross-hematuria.
MBT was found in 31.8% in isolated kidney TB as
whole. Sterile pyuria was in 25%. The onset of TB
orchiepidydydmitis was in 35.7%, hemospermia - in
7.1%, dysuria - in 35.7%. Most common complaints for
prostate TB were perineal pain (31.6%), dysuria (also
31.6%), hemospermia (26.3%). MBT in prostate secretion / ejaculate was revealed in this group in 10.5%.
Conclusions and key recommendations: UGTB is multivariant disease, and standard unified approach is
impossible. Join term UGTB has insufficient information in order to estimate therapy, surgery and prognosis
as well as to evaluate the epidemiology. Using clinical
classification will improve the efficiency of the therapy of
UGTB.
tuberculosis and typical patomorphology are nonspecific

and vague. The diagnosis becomes evident only with
destruction, massive fibrosis and organs failure. Cases of
isolated prostate TB are sporadic and accidental histological findings. This is why studying the early diagnosis
of urogenital TB is important.
Design/Methods: A total of 83 males age from 36 to 72
years with prostate tuberculosis (n 45; prostate TB
group) and chronic inflammatory prostatitis (n 38;
chronic prostatitis group) were included in the study. All
patients attended the Novosibirsk TB Research Institute
between 2006 and 2014. Laboratory tests, ultrasound
and radiological examinations, standard 6-12 cores
prostate biopsies were done for all of them.
Results: Among the patients with prostate TB 66.7% had
active TB in another localizations (lung, urinary system,
scrotum). Additionally 8.9% had high density calcifications in a thoracic organ. Among the patients with
chronic prostatitis the same signs were found in 7.9%
and 15.7% of cases respectively. 35.5% of prostate TB
cases has been confirmed by isolating of M. tuberculosis
from ejaculate and prostatic fluid. Thus, sperm culture
and MBT identification by PCR in ejaculate were more
sensitive than the same study of prostatic fluid (sensitivity at 0.31 and 0.16, respectively). Specific symptoms
which were identified only in prostate TB group patients
were high density lesions and cavities in prostate
(detected by X-ray examination) and hypoechoic lesions
with high density wall (detected by ultrasonography).
Large calcifications (28.8% vs. 7.9%; P , 0.02),
hyperechogenic fibrosis areas (71.1 vs. 47.4%; P ,
0.05) and seminal vesicles lesion (17.7% vs. 2.6%; P ,
0.05) were detected more frequently in prostate TB
patient. Leukospermia was detected with the same
frequency in both groups (39.5% vs. 41.4%), but
hematospermia was detected significantly more frequent
in patients with prostate TB (in 22.2% vs. 7.8% of cases
respectively; P , 0.05). Prostate biopsy with subsequent
patomorfology, culture and PCR study of samples
confirmed the diagnosis only in 20% of patients.
Conclusion: The diagnostic of prostate TB is difficult and
based on M. tuberculosis detection, patomorfology,
radiological examination and patient history (TB in
another localization in present or past). But in the
absence of specific symptoms indirect sign, such as large
calcifications, hyperechogenic fibrosis areas in prostate,
seminal vesicles lesion and hematospermia must be taken
into account during the diagnostic.
PC-798-04 Differential diagnosis of prostate

tuberculosis
PC-799-04 A randomised controlled trial of the

impact of XpertW on tracheal aspirates within a
burden of disease study in South African
intensive care units
E Brizhatyuk,1 E Kulchavenya,1 A Baranchukova2

Novosibirsk TB Research Institute, Novosibirsk, 2Novosibirsk
State Medical Universaty, Novosibirsk, Russian Federation.
e-mail: elena.brizhatyuk@yandex.ru
G Calligaro,1 G Theron,1 H Khalfey,1 J Peter,1 M Miller,1

L Michell,1 I Joubert,1 K Dheda1 1University of Cape Town
Lung Institute and Groote Schuur Hospital, Cape Town,
South Africa. e-mail: greg.calligaro@uct.ac.za
Background: Criteria to diagnose urogenital tuberculosis

(TB) on its early stages except finding of Mycobacterium
Background: There is limited prospective data about

incidence and mortality associated with pulmonary
tuberculosis (PTB) in ICUs, and none on the accuracy

and impact of Xpert-MTB/RIF in this setting.
Design/Methods: Ventilated patients with suspected PTB
(n 341), admitted between August 2010 and July 2013,
to 4 hospitals in Cape Town, were, irrespective of the
reason of admission, prospectively investigated using
microscopy, Xpert-MTB/RIF and culture A. In a substudy, 242 patients were randomised to microscopy or
Xpert-MTB/RIF. The proportion of culture-positive
patients on treatment at 48 hours and mortality (28
and 90 days) were endpoints.
Results: 46/317 (14.51%) of TB suspects had culturepositive TB. Culture-positive patients who failed to
initiate treatment (aHR 4.49, 95%CI 1.45-13.89) or
received inotropes (AHR 4.33, 95%CI 1.62-12.54) were
more likely to die, however, TB was not itself associated
with death. In the RCT, Xpert-MTB/RIF sensitivity was
higher than microscopy (100% vs. 43%; P , 0.002).
Median time-to-diagnosis was shorter (0.2 days vs. 12.1
days, P , 0.001) in the Xpert-MTB/RIF arm. Xpert
MTB/RIF resulted in more culture-positive patients on
treatment only at 48 hours (92% vs. 53%, P 0.043).
There was no difference in stay duration, or overall 90day mortality (32% vs. 42%, Xpert MTB/RIF vs. smear,
P 0.1478).
Conclusion: PTB amongst patients admitted to ICUs in
South Africa is common, but mortality is unaffected by
PTB. Although Xpert-MTB/RIF on TA is more sensitive,
results in rapid treatment initiation and reduces incorrect
treatment, it does not increase the number of patients on
treatment due to high rates of empirical treatment.
S197
13. Diverse diagnostics

PC-800-04 Can line probe assay be used directly
on smear-positive extra-pulmonary samples?
P Dave,1,2 P Patel,1 B Vadera,3 M Ghedia,4 H Thanki,3
A Amar Shah,4 K Pujara2 1Intermediate Reference
Laboratory, Ahmedabad, 2State TB Cell, Gandhinagar,
3
World Health Organization Country Office, Gandhinagar,
4
World Health Organization Country Office, New Delhi,
India. e-mail: drpranavpatel09@gmail.com
Background: Line probe assay (LPA) has been endorsed

for the use of first line drug susceptibility testing (DST) of
smear positive sputum specimens of pulmonary TB
patients. For extra pulmonary and smear negative TB
patients, the samples are inoculated in culture and
isolates of MTB are tested on LPA. This procedure
results in considerable delay and loses the benefit of rapid
diagnostics. Hence, in the present study smear positive
extra pulmonary specimens were tested directly on LPA
and its results was compared with LPA conducted on
culture isolates of same specimens.
Design/Methods: In India, LPA has been used under
Revised National TB Control Programme since 2009 for
diagnosis of multi drug resistant TB patients. The study
was conducted at Intermediate Reference Laboratory in
the state of Gujarat which was one of the sites of
validation study of LPA in the country and has
experience of using technology since 2009. All extra
pulmonary specimens received in 2014 were subjected to
ZN smear microscopy and were inoculated on liquid
culture. Smear positive samples were directly tested with
LPA. Simultaneously, culture isolates of MTB were also
subjected to LPA. Results of LPA conducted on both
direct specimen and culture isolates were compared.
Results: A total 391 extra pulmonary specimens were
tested on ZN smear microscopy and culture. Out of
these, 177 were smear positive and direct specimen could
be tested on LPA and among them 88 were culture
positive and were tested on LPA. Among those specimen
tested directly on LPA, 127 had valid results. This
contributed to 32% of all extra pulmonary samples with
median time to diagnose rifampicin resistance using
direct sample was 5 days (IQR 2-7). In comparison, 88
(23%) extra pulmonary specimens with culture isolates
tested on LPA with valid results and with longer
turnaround time (18-40 days). Of these 88 specimens,
37 were smear negative and contributed to 14% valid
results in addition to smear positive direct specimen. Of
51 specimens tested directly or on isolates with valid LPA
results, 50 (98%) had concordance each for rifampicin
and isoniazid resistance.
Conclusion: The present observation suggests the advantage of testing extra pulmonary smear positive samples
directly on LPA and sequential testing of isolates on
specimens which did not have results on direct specimen
due to smear negative or invalid results.
S198
PC-801-04 Evaluation of colorimetric detection

based liquid culture system for rifampicin
testing for Mycobacterium tuberculosis
PC-802-04 Evaluation of the bacteriological

diagnosis of adult tuberculous meningitis in
Indonesia
R Singhal,1 M Bhalla,1 N Singh,1 R Sarin,1 D Behera,2

P Sharma3 1National Institute of Tuberculosis & Respiratory
Diseases, New Delhi, 2Post graduate Institute of Medical
Education and Research, Chandigarh, 3Statistics, National
Institute of Tuberculosis & Respiratory Diseases, New Delhi,
India. e-mail: drritugo@gmail.com
L Chaidir,1,2 J Annisa,1 S Dian,3 C Ruesen,2,6

B Alisjahbana,1 I Parwati,3 A R Ganiem,3 R Van Crevel2
1
Universitas Padjadjaran, Bandung, Indonesia; 2Radboud
University Medical Center, Nijmegen, Netherlands;
3
Universitas Padjadjaran/Hasan Sadikin Hospital, Bandung,
Indonesia. e-mail: lidya.chaidir@gmail.com
Background: Colorimetric tests employ oxidation-reduction indicator methodology for detection of growth of an
organism. World Health Organization has recommended
use of these tests for detection of drug resistant
tuberculosis as an interim solution in resource constraint
settings. In our country, as there is limited experience of
colorimetric tests, present study was undertaken to
evaluate the accuracy of detection of rifampicin (RIF)
resistance in Mycobacterium tuberculosis (MTB) in
comparison with the results of 1% proportion method.
Methods: Total of 263 MTB positive cultures obtained
over three months in the National Reference Laboratory
were processed on solid Lowenstein Media. Positive
cultures were considered for drug sensitivity by 1%
proportion method and tetrazolium assay. Tetrazolium
microplate assay was performed using Tetrazolium
bromide [3-(4.5-dimethylthiazol-2-yl) 2.5 diphenyl
tetrazolium bromide] prepared at the concentration of
RIF at 1 mg/ml in absolute ethanol. Final RIF concentrations in the wells were set from 4 lg/ ml to 0.06 lg/ ml.
Growth of MTB was detected by change of colour to
purple and was confirmed by demonstration of cords in
the break-point wells.
Results: The sensitivity, specificity, positive predictive
value and negative predictive value were found to be
94.1%, 88.2%, 93.6% and 89.1% with 1% proportion
method as the gold standard. Ten and 11 strains were
found to be false sensitive and false resistant respectively.
Cut off MIC for determining resistance was calculated at
. 0.05 ll using MedcalC software. Area under receiver
operating characteristic (ROC) curve analysis was
93.1% (95% Confidence Interval 89.9% - 96.4%). The
Diagnostic odds ratio (DOR) was 119 suggesting that
test has high discriminatory ability. Average turn-around
time from culture to drug susceptibility result was found
to be 7.2 days whereas average turn-around time for 1%
proportion method was found to be 42 days. The cost per
patient for drug susceptibility by tetrazolium assay from
culture was found to be $ 2.1
Conclusions: The tetrazolium assay was found to be
promising indirect susceptibility assay with turn-around
time comparable to commercial liquid culture systems.
The test is simple, inexpensive and sensitive for detection
of resistance. However, it is desirable to confirm the RIF
resistance by other methods. Stringent quality controls
and standardization is required for interpretation of the
result.
Background: Tuberculous (TB) meningitis is the most

severe form of TB and causes substantial morbidity and
mortality in adults and children. Early recognition and
treatment of the disease is believed to be able to reduce
the burden of this disease, but this is hampered by the fact
that it is often difficult to find bacteriological proof for
TB meningitis. Relatively little is known about the
performance of various methods for diagnosing TB
meningitis in Indonesia.
Methods: We retrospectively evaluated diagnosis among
all adult TB meningitis suspects at a referral hospital in
West Java, Indonesia, between 2006 and 2013 (n 916).
Signs and symptoms were recorded, and cerebrospinal
fluid (CSF) was obtained for routine examination and
microbiological testing. Different methods were applied
over time, including Ogawa solid culture (n 465), MB/
BacT (n 449) and MODS (n 482) liquid culture, and
in-house IS6110 TB PCR (n 207).
Results: Head to head comparison showed a higher yield
of liquid (40.6%) compared with solid culture (23.2%; P
, 0.01), and a shorter time to positivity for MODS
compared to Ogawa (14 vs. 33 days). Microscopy (ZN)
was positive in 4.9 % of clinical suspects, and in 22.9%
of Ogawa-positive and 13.8% of liquid culture-positive
patients. In-house IS6110- PCR (n 207) was positive in
92% of culture-confirmed TB meningitis; and 68% of
clinically diagnosed cases (vs. 44% for liquid culture; P
0.262); and 100% specificity in non-meningitis controls
patients. HIV-infected patients (19.6%) had a lower yield
of TB culture (21.2% vs. 34.1%, P , 0.01) but not
microscopy (6.1% vs. 4.8% ) compared to HIV-negative
patients, with significant numbers of clinically suspected
TB meningitis patients diagnosed with cryptococcal
meningitis or PCR-positive cerebral toxoplasmosis
(2.4%). Culture yield was higher with larger CSF
volumes and lower when TB treatment had been started
prior to lumbar puncture. CSF glucose more reliably
predicted culture results than protein and cell count.
Conclusion: Microscopy was much less sensitive compared to other methods. Liquid culture was reasonably
sensitive but not helpful in clinical decision making. Inhouse PCR had a good sensitivity for TB meningitis, but
too complex to implement in routine practice. There is
still an urgent need for faster and more sensitive
diagnostic tests for TB meningitis.
PC-803-04 Genotyping of Mycobacterium

tuberculosis isolated in a national reference
laboratory in Baghdad
R Ali,1 A Al Sudani,2 L Salihi,2 M Abdulrazaq2 1Veterinary
Directorate/ Ministry of Agriculture, Baghdad, 2Specialized
Center for Chest and Respiratory Disease, Baghdad, Iraq.
e-mail: ntp.mohammad@gmail.com
Background: Tuberculosis remains a major global health

problem. Worldwide it ranks as the second leading cause
of death from an infectious disease. Molecular characterization of Mycobacterium tuberculosis isolates has
helped the study of dynamics transmission of such strains
within defined geographical setting. Aim of the study:
find molecular epidemiology and phenotypic and genetic
characteristics of M. tuberculosis complex isolates
Methods: During January-May 2013, a study conducted
in Baghdad, 2 genotyping methods spoligo-typing and
Mycobacterial interspersed repetitive units variable
number tandem repeats (MIRU-VNTR), alongside with
patients demographic data have been utilized to study
samples from suspected TB patients.
Results: Spoligo-typing was performed for 106 M.
tuberculosis strains, the results showed (2.7%) samples
given negative and excluded from further investigation,
the result of MTC isolates showed the most predominant
genotype among tuberculosis strains were Delhi / CAS 37
(37.4%), TUR genotype was observed in (30.3 %) of
spoligo-type, while (15.2%) classified as Ghana. (5.1%)
isolates belong to Haarelem lineage. Uganda II was
found in (3%) isolates, S lineage were found in (2%),
NEW-1, LAM, lineage were observed in (1%) to each
spoligo-type. Remaining (2%) spoligo-type patterns that
did not match with any of the updated in the MIRUVNTR plus database were classified as Unknown and
multiple matches of genotypes respectively. MIRUVNTR was used to assess the diversity of M. tuberculosis; there are no clustering rate of this method was
observed from these strains. Five loci (QUB-11b, QUB26, Mtub-21, Mtub-29and Mtub-34) showed poor
discriminatory power (HGDI,0.3). Six loci (MIRU-26,
MIRU-39, MIRU-40, ETR-C, QUB-4156 and Mtub-04)
were dispersed to discriminate the isolated moderately
(0.3 6 HGDI 6 0.6). Lastly, thirteen loci (MIRU-2,
MIRU-4, MIRU-10, MIRU-16, MIRU-20, MIRU-23,
MIRU-24, MIRU-27, MIRU-31, ETR-A, ETR-B, Mtub30 and Mtub-39) were highly discriminate (HGDI .
0.6).
Conclusions: Genotyping of M. tuberculosis is important
in the identification of related strains and characteristic
disease phenotypes. It can augment classic epidemiology
and clinical practice to provide a holistic approach in the
investigations, treatment and control of tuberculosis.
S199
PC-804-04 Effectiveness of the detection of LTBI

infection using tuberculin skin test and CFP-10ESAT-6 allergen in children and adults in
Moscow, 2013-2014
L Slogotskaya,1 E Bogorodskaya,1 O Senchihina,1
G Nikitina,1 V Litvinov,1 P Seltsovs.ky,1 D Kudlay,2
N Nikolenko2 1Scientific and Clinical Antituberculosis Center
of Moscow Government Health Department, Moscow,
2
RESEARCH, OJSC Generium, Moscow, Russian Federation.
e-mail: chingy@bk.ru
Background: The Russian company has developed a

preparation known as Diaskintest (DST), which represents Mycobacterium tuberculosis-specific recombinant
proteins CFP10-ESAT6, which are absent in all M. bovis
BCG substrains and in most of non-tuberculous mycobacteria. It is for skin testing (0.2 mcg/0.1 ml in the same
way as the Mantoux test). Aim. To determine the size of
the group at highest risk of TB using two different skin
tests as a screening method in the children population.
Design/Methods: During 2013 in Moscow the Mantoux
test with 2 TU PPD-L was performed in 1 420 084
children and adolescents (97.4% of general children
population), positive reactions were observed in 1 019
814 individuals (71.8%), among them DST was performed in 131 361 or 12.9%, predominantly screened
people with suspected TB infection, excluding postvaccination allergy as much as possible. In Moscow more
than 90% of children are BCG-vaccinated. During 2014
Mantoux test was performed in 1 429 395 children and
adolescents (97.3% of children population), positive
reactions were observed in 1 058 191 individuals
(74.0%). Among them DST was performed in 88 000
people or 8.4%.
Results: In 2013 among 131 361 Mantoux-positive
individuals DST-positive reactions were observed in
3304 cases (2.5%). Among DST-positive persons new
cases of TB were identified in 163 (4.9%). Among 131
361 Mantoux-positive new cases of TB were identified in
0.12%. Detection rate was 40 times lower compared to
positive DST, P 0.000. In 2014 among 88 527
Mantoux-positive persons DST-positive reactions were
in 3573 (4.0%) cases. Among DST-positive persons new
cases of TB were identified in 138 (3.9%). Among 88 527
Mantoux-positive children new cases of TB were
identified in 0.16%. Detection rate was 24 times lower
compared to positive DST, P 0.000. In 2013 among
131361 children examined by 2 skin tests DST-positive
reactions were observed in 0.23% of the Moscow child
population and in 2014 among 88 527 examined by 2
tests DST-positive reactions were observed in 0.25% (P
. 0.05). In both years DST- positive reactions were
observed in 0.31- 0.35% of the Mantoux-positive
children.
Conclusion: Children with DST-positive reactions are the
group with the highest risk of TB and this proportion in
2013-2014 was approximately 0.31-0.38% of the
Mantoux-positive persons or 0.23-0.25 of the child
population.
S200
PC-805-04 A sensitive urinary

lipoarabinomannan test for tuberculosis in HIVnegative patients
1
A Westman,1 B Hamasur,1 G Froberg,
J Bellbrant,1
1,2,3
1
M Correia-Neves,
J Bruchfeld, G Kallenius1
1
Karolinska Institutet, Stockholm, Sweden; 2University of
Minho, Braga/Guimaraes, 3PT Government Associate
Laboratory, Braga/Guimaraes, Portugal. e-mail:
gunilla.kallenius@ki.se
Background: We have previously developed a diagnostic

test for tuberculosis (TB) based on detection of mycobacterial lipoarabinomannan (LAM) in urine.1,2 Commercial tests based on our findings have been developed,
but a range of studies have shown a very low sensitivity
in HIV-negative TB patients. We have now developed an
easy to perform test based on a magnetic immunoassay
platform, utilizing high avidity monoclonal antibodies
for the detection of LAM in urine.3 With this method the
analytical sensitivity of the assay was increased 50-100fold compared to previous tests.3. The method includes a
few pipetting and washing steps. A technician can handle
24-48 samples in about 2 h. An ELISA reader is used for a
quantitative measurement of the LAM concentration,
but can be replaced by either: 1) a battery operated
portable ELISA reader or 2) by reading the colour change
by the naked eye.
Design/Methods: The aim if the study was to evaluate

the efficacy of the urinary LAM assay (Uri-TBdiadirect)
in HIV-negative TB patients.The test was evaluated in
adult HIV negative patients with either microbiologically
verified TB or clinical TB at Karolinska University
Hospital, Stockholm. Healthy individuals with no
history of TB were included as controls. By reading the
absorbance at 450 nm a cut-off value for the sensitivity
and specificity of the test was set at an optical density
(OD) of 0.35.
Results: In HIV-negative patients with microbiologically
verified TB (n 32) the median OD value was 0.80 and in
healthy controls (n 44) the median OD was 0.24. The
sensitivity of the test was 91% and the specificity 98%.
The LAM concentration decreased significantly over
time in patients who responded to TB treatment.
Conclusion: Here we demonstrate the high sensitivity

and specificity of our new urinary LAM detection test.
This is in stark contrast to previous studies using
available commercial tests with polyclonal anti-LAM
Abs where the sensitivity in HIV-negative TB patients
was very low.
References:
1 Hamasur B, et al. Rapid diagnosis of tuberculosis by detection of
mycobacterial arabinomannan in urine. J Microbiol Meth 2001; 45:
4152.
2 Tessema T A, et al. Diagnostic evaluation of urinary lipoarabinomannan at an Ethiopian tuberculosis centre. Scand J Inf Dis 2001; 33:
279.
3 Hamasur B, et al. A sensitive urinary lipoarabinomannan test for
tuberculosis. PLOS ONE 2015; 10: e0123457.
PC-806-04 Bedaquiline minimal inhibitory

concentration quality control ranges for drug
susceptibility testing in a multi-laboratory
study
K Kaniga,1 D M Cirillo,2 S Hoffner,3 N Ismail,4
B Metchock,5 G Pfyffer,6 A Sloutsky,7 A Venter8 1Janssen
Research & Development, LLC, Titusville, NJ, USA; 2TB
Supranational Reference Laboratory, San Raffaele Scientific
Institute, Milan, Italy; 3Department of Microbiology, The
Public Health Agency of Sweden, Solna, Sweden; 4National
TB Reference Laboratory, Center for Tuberculosis, National
Institute of Communicable Diseases, Johannesburg, South
Africa; 5Reference Laboratory, Division of TB Elimination, U.S.
USA; 6Department of Medical Microbiology, Center of
Laboratory Medicine, Luzerner Kantonsspital LUKS, Lucerne,
Switzerland; 7Department of Public Health Massachusetts
State Laboratory Institute, Jamaica Plain, MA, USA; 8TASK
Applied Science, SA MRC Centre for TB Research, DST/NRF
Center of Excellence for Biomedical TB Research, Division of
Molecular Biology and Human Genetics, Tygerberg, South
Africa. e-mail: kkaniga@its.jnj.com
Background: To successfully address the global challenge

of multidrug-resistant tuberculosis (MDR-TB), reliable
and reproducible DST results are critical, both for patient
care and for surveillance. The advent of new anti-TB
drugs including bedaquiline (BDQ) has highlighted this
need. Two multi-laboratory studies were conducted to
determine BDQ MIC QC ranges using standardized
methods on solid and liquid medium.
Methods: Two multi-laboratory, reproducibility, tier-2,
QC studies of BDQ DST were conducted in 8 geographically diverse laboratories following the CLSI M23-A3
and M24-A2 guidelines and using the standard QC
reference strain Mycobacterium tuberculosis (MTB)
H37Rv. Three separate lots of 7H10 and 7H11 agar
from different manufacturers (agar dilution study) and
96-well frozen microtiter plates (Thermo Fisher; 7H9
broth microdilution study) were evaluated for the
performance on BDQ DST. In each study, BDQ
concentrations tested were 0.0082 lg/ml and 0.0084
lg/ml, respectively, and polystyrene ware was used. MIC
was defined as the lowest BDQ concentration that
completely inhibited growth by visual inspection. BDQ
MIC frequency, mode, and geometric mean were
calculated. Three- or 4-dilution QC ranges targeted

795% of the observed BDQ MICs centering on the
mode or geometric mean. If one laboratory produced
outlying data, their entire dataset (for all 3 media lots)
was excluded, and the analyses repeated with the
remaining laboratories.
Results: In the agar dilution study, a 4-dilution QC range
(0.015 to 0.12 lg/ml) centered on the geometric mean
included 95.8% of the BDQ MICs on 7H10 agar (204/
213 observations with one dataset excluded) and 95.9%
on 7H11 agar (232/242) (Table). In the 7H9 broth
microdilution study, a 3-dilution QC range (0.015 to
0.06 lg/ml) centered on the mode included 95.1% (235/
247) of BDQ MICs (Table). BDQ MICs were not affected
by the sources of 7H10 or 7H11 agar lots or the frozen
microtiter plate lot.
Conclusion: Based on the totality of the data from these
two multi-laboratory, multi-country reproducibility
studies, the BDQ MIC QC ranges against the MTB
reference strain H37Rv were defined (Table). The 7H9
broth microdilution, 7H10 and 7H11 agar dilution
methods and reference BDQ MIC QC ranges will
provide a standard for BDQ DST in various settings.
S201
surveyed and asked to rank TPP test attributes in order

of importance.
Design/Methods: An online survey was administered to
43 purposively selected government and non-government
participants representing 14 of the 22 TB high-burden
countries. Respondents included laboratory personnel,
national TB program managers, technical experts, patient
representatives and researchers. Participants were asked
to rank 21 TPP test attributes in order of importance
using a maximum differential (max-diff) method.
Results: The max-diff approach ranked the 21 TPP
attributes in order of relative importance, from most to
least important (Figure), Respondents identified sensitivity as the most important TB test attribute, followed
by maintenance/calibration, reagent kit storage/stability,
sample prep steps, time to results, and specificity (in that
order).
Conclusion: Developers can utilise the survey results to
prioritize their investments and guide decision-making
when trade-offs are required in test development.
Table: BDQ MIC QC ranges against MTB H37Rv

Quality
control
strain
Mycobacterium
tuberculosis
H37Rv
MIC (lg/mL)
7H10 agar
7H11 agar
7H9 broth
0.015 0.12
0.015 0.12
0.015 0.06
PC-807-04 Which attributes within target

product profiles for TB diagnostics are the most
important to focus on?
T Adepoyibi,1 L Lillis,1 H Greb,1 D Boyle1 1PATH,
Washington, DC, USA. e-mail: topeadepoyibi@yahoo.com
Background: In recent times, an unprecedented range of

tuberculosis (TB) nucleic acid amplification technologies
have entered the market, or are in late-stage development. To ensure that these tests meet the needs of the TB
community, sufficient information should be made
available to test developers in response to their key
questions, and potential barriers to market entry should
be addressed. In this regard, a 2013 publication outlined
the top ten questions of 25 surveyed TB test developers
(Pai, 2013). Among the questions posed included; What
are the unmet diagnostic needs and target product
profiles (TPPs) of greatest relevance? There were also
two sub-questions; Which attributes within the TPP are
the most important to focus on? and; What are the top
four to five features that are needed in a TB diagnostic
test for developing countries? (Pai, 2013). In 2014 the
WHO released a critical report following a consensus
meeting, which identified the highest priority TPPs for
new TB diagnostics. Using this information to assist in
further addressing the two sub-questions, a range of
stakeholders from high-TB burden countries were
PC-808-04 Use of automated MGIT culture

tubes with a conventional incubator and
manual reader for TB diagnosis in resourcelimited settings
A Kruuner,1 B Kosloff2,3 1University of Alabama at
Birmingham, Birmingham, AL, USA; 2ZAMBART, UNZA
School of Medicine, Lusaka, Zambia; 3London School of
Hygiene & Tropical Medicine, London, UK. Fax: (26) 211
254 710. e-mail: bkosloff@yahoo.com
Background: TB programs in resource-limited settings

are often unable to equip all of their diagnostic
laboratories with automated MGIT 960 instruments.
Central laboratories might use a MGIT 960 instrument
and automated tubes while peripheral laboratories use
the manual MGIT system or solid media with a
conventional incubator and manual reading. It would
be simpler and more efficient to use the same tubes and
reagents in all laboratories. Therefore, we compared use
of 7mL automated MGIT tubes incubated in a conventional incubator and read manually to automated MGIT
tubes incubated in a MGIT 960 instrument.
Design/Methods: 299 consecutive specimens (238 sputum, 59 urine, 1 bronchial lavage and 1 body fluid) were
S202
processed for TB culture. Two automated MGIT tubes

and one Lowenstein-Jensen
(L-J) solid medium slope
were inoculated with 0.5mL of sediment. One MGIT

tube was incubated in a MGIT 960 instrument (automated method). The second tube was incubated in a
conventional incubator and read manually on weekdays
using a Micro-MGIT Fluorescence Reader (manual
method). Negative results were reported after 42 days
of incubation. Tubes with bacterial growth from any of
the three culture systems were submitted for further
workup to confirm the presence of mycobacteria.
Identification of all AFB-positive cultures was performed
using the Hain CM line probe assay.
Results: The recovery rates of M. tuberculosis using the
automated method and the manual method were similar:
3% (9/299) for the automated method vs. 2.7% (8/299)
for the manual method. Both methods detected the same
number of environmental mycobacteria (MOTT). There
was no difference in time to detection (TTD) of M.
tuberculosis complex (16 vs. 16.4 days) or MOTT (16.5
vs. 16.7 days) between the two methods. Contamination
rates of 21% for the automated method and 32% for the
manual method were much higher than for solid medium
(11%). The contamination rate after 35 days was much
higher using the manual method compared to the
automated method.
Conclusion: This study demonstrates that BD automated
MGIT culture tubes incubated in a conventional
incubator and read manually yield similar results to the
same tubes incubated in a MGIT 960 instrument. The
AFB-positivity rate and TTD of all AFB-positive cultures
were not significantly different. Additional studies are
needed to determine the reason for higher contamination
rates after 35 days for culture tubes incubated in a
conventional incubator.
mation on the molecular pattern of drug resistance of

MTBC strains in West Africa.
Design/Methods: Isolates from the West African Nodes of
Excellence for Tuberculosis Aids and Malaria (WANETAM) had previously undergone antimicrobial susceptibility testing. Whole genome sequencing was performed
for 51 MDR and 40 non-MDR M. tuberculosis complex
non-randomly selected from Nigerian. The maximum
likelihood phylogeny of the isolates was reconstructed
from single nucleotide polymorphisms (SNP) in the nonrepetitive core genome. Resistance conferring mutations
were detected by mapping sequencing reads to the H37Rv
reference genome using the software KVARQ. Phylogenetic lineage was inferred using a typing method.
Sequencing was unsuccessful for four isolates.
Results: The Euro-American lineage (lineage 4) was the
most prevalent in the dataset (n 78). Within lineage 4,
the Cameroon sub lineage (n 40) was the most
prevalent. Others including Turkish, Latin American &
Mediterranean, Ghana and Harlem were found. Eight
isolates belonged to M. Africanum West Africa 1 lineage
and one was from M. bovis. Out of 51 MDR isolates, 45
and 6 were found to belong to lineage 4 and Maf1
respectively. Rifampicin resistance was linked to mutations in rpoB while isoniazid resistance was linked to
mutations in katG and the inhA promoter. In a third of
isoniazid resistant isolates no resistance mutation was
detected. Similarly a known resistance mutation was
detected in only a third of streptomycin resistant isolates.
In 20 instances mutations in embB did not correlate with
resistance to ethambutol. Interestingly, in two patients,
phenotypic resistance to either ethambutol or streptomycin was developed over the course of treatment.
Despite the development of drug resistance there was no
change in the genome sequence.
14. Gene mutation: the engine behind

drug resistance and TB strain diversity
PC-810-04 Genomic characterisation of
multidrug-resistant (MDR) and non-MDR
Mycobacterium tuberculosis complex from two
urban referral centres in South Nigeria
M Senghore,1 J Otu,1 A Witney,2 A Kehinde,3 O Idigbe,4
B De Jong,5 M Pallen,6 M Antonio1 1Medical Research
Council The Gambia Unit, Bakau, Gambia; 2St Georges
University London, London, UK; 3University College Hospital,
Ibadan; 4National Institute of Medical Research, Lagos,
Nigeria; 5Institute of Tropical Medicine, Antwerp, Belgium;
6
University of Warwick, Coventry, UK. e-mail:
msenghore@mrc.gm
Background: The global emergence of multidrug resistant TB (MDR-TB) and extensively Drug-resistant TB
(XDR-TB) in the context of HIV threatens to undermine
the success that has been achieved in treating drug
sensitive tuberculosis. Acquired drug resistance in
Mycobacterium tuberculosis complex occurs due to
chromosomal mutations. However, there is little infor-
Conclusion: This preliminary data suggests that the

molecular mechanisms of resistance in Nigerian M.
tuberculosis complex may span beyond repository of
known resistance-conferring mutations. Enhancing our
knowledge of the molecular basis of resistance will aid in
development of kits for rapid detection of drug resistance
in clinical tuberculosis.
PC-811-04 Characterization of drug-resistant

Mycobacterium tuberculosis complex species:
higher rates of RMP resistance among M.
tuberculosis patients in Northern Ghana
1
I D Otchere, A Asante-Poku, S Osei-Wusu,

S Aboagye,1 A Forson,2 F Bonsu,3 S Gagneux,4
D Yeboah-Manu1 1Noguchi Memorial Institute for Medical
Research, University of Ghana, Legon, Accra, 2Korle-Bu
Teaching Hospital, Accra, 3National Tuberculosis Control
Program, GHS, Accra, Ghana; 4Swiss Tropical and Public
Health Institute, Basel, Switzerland. Fax: (233) 3021 502
182. e-mail: dyeboah-manu@noguchi.ug.edu.gh
Background: The control of TB is threatened by the

emergence of strains resistant to anti-TB drugs. This
study analysed for drug resistant conferring mutations
among Mycobacterium tuberculosis sensu stricto (MTB)
and Mycobacterium africanum (MAF) strains obtained
from two regions (Greater Accra; GAR and Northern;
NR) in Ghana.
Design/Methods: Isolates were tested for sensitivity to
isoniazid (INH) and rifampicin (RIF) by the micro-plate
alamar blue assay. Specific genetic elements of respective
drug resistant isolates (inhApro, inhA, oxyR-ahpC, katG
and ndH for INH and rpoA, rpoB and rpoC for RIF)
were sequenced for mutation analyses. The proportions
of drug resistant isolates were compared between species
and the two regions
Results: A total of 111 (7.4%), 10 (0.7%) and 39 (2.6%)
of the 1490 isolates analyzed were mono-resistant to
INH, RIF and MDR. The likelihood of an isolate to be
RIF mono-resistant was higher in the NR (1.4%; 2/140))
than GAR (0.6%; 8/1350) (P 0.0000) and a similar
trend was seen with MDR even though not statistically
significant (P 0.0636). However, all the RIF monoresistant strains were MTB. The propensity for an isolate
to harbor katG S315T mutation was significantly higher
in MTB (68.1%; 62/91) than MAF (63.6%) (P 0.0002)
whereas the opposite was the case for inhApro mutations
36.4%; 4/11 (MAF) and 23.1%; 21/91 (P 0.0082). We
detected two rpoC mutations (G332R and V483G),
which occurred independently with rpoB S450L respectively. Among the 160 cases infected with strains
resistant to at least one drug, 24 were lost to follow up
and among those monitored, 36 were cured of which 35
were INH mono-resistant. In addition, treatment for 42
cases failed among which 38 were MDR cases.
Conclusion: The level of drug resistant TB is still fairly
low in Ghana however, the rate of RIF mono-resistance
(1.4; 2/140) and MDR (5.0%; 7/140) in the Northern
region are above the national average. This calls for
strengthening of control activities especially DOTs in the
north by the NTP.
S203
PC-812-04 Molecular epidemiology of

extensively drug-resistant strains of M.
tuberculosis in Southern Kazakhstan
A Alenova,1 T Abildaev,1 P Tarlykov,2 D Raimbek,2
E Zholdybayeva,2 E Ramanculov2 1National Center for
Tuberculosis Problems, Almaty, 2National Center for
Biotechnology, Astana, Kazakhstan. e-mail:
arike.alenova@mail.ru
Background: The emergence of extensively drug-resistant tuberculosis (XDR-TB) inKazakhstan compromises

proper control of tuberculosis (TB) in the country.
However, molecular characterization of XDR-TB isolates can be used for identification of mutations in the
genes responsible for emerging patterns of antibiotic
resistance.
Design/Methods: Genomic DNA was extracted from 58
XDR-TB isolates using thermolysis method. PCR amplification of nine genetic loci was performed, namely,
rpoB, katG, embB, tlyA, eis, rrs, gyrA, gyrB and the
fabG-inhA promoter region. PCR products were sequenced using Big Dye Terminator Cycle Sequencing Kit
v3.1 (Life Technologies) with the same primers and run
on 3130xl Genetic Analyzer (Life Technologies). Mutations were detected by comparison with the M.
tuberculosis wild-type reference strain H37Rv using
SeqScape software v2.6. In addition, VNTR-typing of
24 loci was performed by PCR (MIRU02, MIRU04,
MIRU10, MIRU16, MIRU20, MIRU23, MIRU24,
MIRU26, MIRU27, MIRU31, MIRU39, MIRU40,
ETR-A, ETR-B, ETR-C, Mtub04, Mtub21, Mtub29,
Mtub30, Mtub34, Mtub39, Qub-11b, Qub-26 and
VNTR53).
Results: A combination of nine genetic loci has confirmed XDR-TB phenotype in 39 isolates (67.2%). The
most frequent mutations among the XDR-TB isolates
were 315 inkatG; 531, 526 and 516 in rpoB; 94, 90 and
91 in gyrA and 1401 in rrs. Novel mutation was detected
in 3 (5.2%) isolates causing change of aspartic acid to
cysteine at codon 94 of gyrA. The mutation may
represent additional molecular marker for fluoroquinolone resistance in M. tuberculosis. Also, among these
XDR-TB isolates, 55 (94.8%) were classified asBeijing
genotype, 2 (3.5%) LAM and 1 (1.7%) S genotype.
Conclusion: Sequencing has revealed the presence of
major mutations in the genes responsible for resistance to
the anti-TB drugs. This work may serve as a basis for new
technique suitable for diagnosis and epidemiological
control of XDR-TB inKazakhstan.
S204
PC-813-04 Detection of mutations among

rifampicin- and isoniazid-resistant
Mycobacterium tuberculosis isolates in
Kazakhstan
Y Tursynbay,1,2 A Akhmetova,1 L Chingisova,3
V Bismilda,3 A Akilzhanova,1 U Kozhamkulov1 1Center
for Life Sciences, Nazarbayev University, Astana, 2School of
Science and Technology, Nazarbayev University, Astana,
3
National Center for Tuberculosis Problems, Almaty,
Kazakhstan. e-mail: kzulan@mail.ru
Background: According to WHO data, Kazakhstan is

positioned as 27th among the highest multidrug-resistant
(MDR) TB burden countries worldwide. The aim of our
study was to characterize mutations associated with drug
resistance to the rifampicin and isoniazid among
Mycobacterium tuberculosis isolates from Kazakhstan.
Design/Methods: M. tuberculosis strains were isolated in
2013 from new TB patients in three different regions of
Kazakhstan (Almaty, Kyzylorda and Kostanay). Drug
susceptibility of M. tuberculosis to isoniazid and
rifampicin was determined using the absolute concentration method according to WHO recommendation.
Genetic determination of M. tuberculosis drug resistance
was identified through amplification and sequencing of
the rpoB and katG gene regions and promoter regions of
mabA (fabG)-inhA, oxyR-ahpC operons by automated
ABI 3730 (Applied Biosystems) Genetic Analyzer,
according to the manufacturers instructions.
Results: For genetic research, 119 clinical isolates of M.
tuberculosis was studied, including 92 MDR isolates, 23
isoniazid-resistant isolates and 4 rifampicin-resistant
isolates. Among 96 rifampicin-resistant clinical isolates
mutations were detected in 93.7% isolates. Six different
variants of mutations in 4 codons (Ser531, Asp516,
His526 and Leu533) were found. The majority of
nucleotide substitutions have occurred at codon 531 on
rpoB gene (82.3%), where 2 variants of single-nucleotide
substitutions were discovered with a predominance of
Ser531Leu (81.2%) replacement. In other cases, mutations were observed in less than 10% of cases and
comprised: 9.3% at codon His526; 1.04% at codon
Leu533; 1.04% at codon Asp516. In 6.2% cases
mutations in rpoB gene were not found. Mutations that
determine resistance to isoniazid were detected in 99.1%
cases among 115 isoniazid-resistant isolates. Analysis of
the katG gene revealed 112 isolates (97.4%) with
mutation in codon 315, which leads to amino acid
replacement of serine to threonine, including 6 cases
(5.2%) with double mutation, when replacement simultaneously takes place in katG gene and in 8 and 15
positions of promoter mabA (fabG)-inhA operon. Two
clinical isolates had a single mutation in the 15 and - 8
positions in the promoter mabA (fabG)-inhA.
Conclusion: Mutations in Ser531Leu (81.25%) of rpoB
gene and Ser315Thr of katG gene (97.39%) were the
most prevalent among -rifampicin and -isoniazid resistant clinical isolates in Kazakhstan.
PC-814-04 Trends in pattern of resistance to

first-line anti-tuberculosis drugs, Malawi
K Mbendera,1 L C Chisuwo,1 F Nyakwawa,1
L Chaponda,1 F Kassa,1 B G Belaineh,1,2 I Dambe,1
J Mpunga1 1National Tuberculosis Program, Malawi,
Lilongwe, 2International Training and Education Center for
Health (ITECH)-Malawi, Lilongwe, Malawi. e-mail:
belaineh@gmail.com
Background: Malawi is low MDR-TB burden country in

southern Africa. The prevalence of MDR(TB was 0.48%
among new and 4.8 % retreatment cases according to
drug resistance survey conducted in 2010/11. Although
the resistance level is low, there are different factors that
put Malawi with an increased risk of drug resistance TB.
Malawi has managed to shift to RH from EH just after
the DRS survey, and there is a continuous interaction
with countries in SADEC region that have high MDR(TB
burden. Hence the purpose of this analysis is to review
the pattern of resistance and inform decision.
Design/Methods: The national Tuberculosis program has
established a surveillance system to monitor drug
resistance among presumptive MDR TB cases in central
reference laboratory. Drug susceptibility tests (DST)
were performed on sputum culture positive isolates on
four first-line anti-TB drugs: rifampicin, isoniazid,
ethambutol and streptomycin. The pattern of resistance
to first line anti TB drugs was reviewed. This was also
compared to drug resistance survey result of 2010/11.
Culture Registers were the source of information.
Samples with any resistance for the aforementioned four
drugs were included in the analysis.
Results: The total number of samples with any resistance
for the four drugs was 68, 91 and 86 in 2012, 2013 and
2014 respectively. The number doesnt reflect the actual
number of samples with any resistance as 2014 are still
being processed. Resistance to rifampicin (67.44%) was
higher among the drugs. Followed by INH (47.44%) and
Streptomycin (25.58%) in 2014. This is contrary to DRS
done in 2010/11 where resistance to INH was 58.5 %
followed by streptomycin (42.3%) and Rifampicin
(34.2%) among samples with any resistance collected
from previously treated patients.
Conclusion: The analysis showed proportion of resistance to Rif is growing among presumptive MDR-TB
cases in the last 3 years. Proportion exceeded resistance
to INH. The change could be attributed to change in
regimen shift from Ethambutol/INH to Rifampicin and
INH. Proportion of samples with resistance to streptomycin has also declined. This may be related to reduce
use of streptomycin and increased resistance the other
drugs. Although, comparability of DRS with routine data
may be affected by different biases including selection
bias, the comparison revealed changes that signal
alterations in drug resistance pattern and magnitude.
Drug resistance survey will provide accurate information.
PC-815-04 The correlation between genetic

mutations for isoniazid resistance by line probe
assay and phenotypic resistance by drug
Z Puyen,1 H Ramirez,1 E Pacheco,1 J R Acosta Barriga1
1
Instituto Nacional de Salud, Lima, Peru. e-mail:
zpuyeng@gmail.com
Background: Genotype MTBDRplus is a comercial line

probe assay (LPA) utilized in Peru since 2011. This assay
allow the identification of the M. tuberculosis and its
resistance to Rifampicin and/or Isoniazid. This probe is
based on reverse hybridization of amplicons to immobilized membrane based probes covering wild type and
mutation sequences. Isoniazid resistance is most frequently associated with mutations in katG gene and inhA
gene. The detection of mutation in the katG gene is
directly related with high level isoniazid resistance and
the mutation in the promoter region
of the inhA gene is
related to the low level isoniazid resistance. The aims of
this study is to determine whether detected genetic
mutations for isoniazid (H) resistance by LPA GenoType
MTBDRplus correlate with high or low level phenotypic
resistance to isoniazid using Drug Susceptibility Test
(DST).
Design/Methods: We compare the results of 8942 smear
positive sputum specimens and 7227 culture sample
tested with GenoType MTBDRplus, in the National
Reference Laboratory of Mycobacteria in Peru from
2011 to 2013, against DST. We selected all the samples
with valid result in DST and LPA, 2712 samples
(excluding non-growing and contaminated samples).
Data of these samples were exported to an EXCEL file
and after analyzed with OpenEpi version 3. Comparisons
between gene mutations for H resistance using LPA and
the low and high phenotypic concentrations of H using
DST will be done using Cohens kappa coefficient,
sensitivity and specificity, and positive and negative
predictive values using DST as the referent. 95%
confidence intervals were used throughout
Table Sensitivity and specificity of LPA (as compared to DST)
for detection of levels of isoniazid resistance
CDST
INH 0.2 ug/ml
(Low)
Kat G mutation
Sensitivity
Specificity
VPP
VPN
Kappa Cohens
64.2%
96.5%
98.4%
45.1%
0.43
Inh A mutation
Sensitivity
Specificity
VPP
VPN
Kappa Cohens
35.0%
96.53%
97.06%
31.2%
0.18
INH 1.0 ug/ml

(High)
82.2%
95.6%
96.3%
79.2%
0.75
27.6%
72.3%
58.5%
41.4%
,-0.01
Results: Mutation in KatG was present in 64.2% (1334)

of the samples with low level isoniazid resistance, and
82.2% (1306) of the samples with high level isoniazid
resistance. Mutation in InhA was present in 34.9% (727)
S205
of the samples with low level isoniazid resistance, and

27.6% (438) of the samples with high level isoniazid
resistance.
Conclusion: This study has been done under programmatic conditions. The VPP values of KatG mutations for
predict resistance to H, are above 95% for low H
resistance as for high H resistance. On the other hand,
the VPP values of InhA mutations for predict resistance
to H, are above 95% only for low H resistance, however
the sensibility is only 35.0%. Finally, it is important to
perform more detail analysis of the mutation implicated
in these results.
PC-816-04 Potential of rifabutin in the

treatment of rifampicin-resistant tuberculosis
A Van Rie,1,2 M Whitfield,3 L Scott,4 F Sirgel,3
Y Voss De Lima,4 W Stevens,4 R M Warren3 1University of
Antwerp, Antwerp, Belgium; 2University of North Carolina,
Chapel Hill, NC, USA; 3SAMRC Centre for Tuberculosis
Research / DST/NRF Centre of Excellence for Biomedical
Human Genetics, Stellenbosch University, Stellenbosch,
4
Africa. Fax: (1) 919 966 2089. e-mail:
vanrie@email.unc.edu
Background: Resistance to rifampicin (RIF) greatly

jeopardizes the success of tuberculosis (TB) treatment.
WHO guidelines therefore recommend that patients
diagnosed with RIF resistant TB initiate treatment with
a 5-drug MDR-TB regimen, which is more expensive and
has higher level of toxicity compared to first line
treatment. Several rpoB mutations have been associated
with isolates that are RIF sensitive or RIF resistant but
rifabutin (RBT) sensitive. RIF or RBT could thus be used
in selected patients diagnosed with RIF resistance on
Hain MTBDRplus or Xpert MTB/RIF assay, two
genotypic assays that are commonly to screen for RIF
resistance. The proportion of such patient at population
level is not known.
Design/Methods: Individuals diagnosed with RIF resistant TB by Hain MTBDRplus or Xpert MTB/RIF assay
in 3 provinces in South Africa (Gauteng, Free State and
Eastern Cape) were enrolled in a prospective cohort
study. Samples of 314 patients were available for
targeted sequencing of the rpoB gene and minimal
inhibitory concentration (MIC) for RIF at concentrations
0.125-200 microgran/ml. MIC for RBT (at concentrations 0.125 2 microgram/ml) was performed on
samples representing the different rpoB mutations and
deletions and different levels of RIF MIC observed in the
study population.
Results: Among the 314 enrolled patients, MIC for RIF
and rpoB sequencing results were available for 291
(93%). Mutations in rpoB 531 (45%), 516 (30%) and
526 (13%) were the most frequent. RBT MIC was
performed in 43 isolates. Isolates carying rpoB mutations
L511P, D516A, D516C, H526N, and 3bP deletion in
518 were sensitive to both RIF and RBT. Isolates carrying
rpoB mutations D516V, D516Y, D516F, D516F, H526C,
dual mutations 511&516 or 512&516 or a 3bP deletion
S206
in 517 were resistant to RIF but susceptible to RBT.

Isolates carrying mutation H526L were sensitive to RIF
but mixed resistance results were observed for RBT. On
population level, 8% of patients diagnosed with RIF
resistance on Hain MTBDRplus or Xpert MTB/RIF can
be treated with rifampicin and an additional 32% can be
treated with Rifabutin.
Conclusion: Rifabutin holds great potential for treatment
of patients diagnosed with RIF resistant TB by Xpert
MTB/RIF or Hain but this would require knowledge of
the exact rpoB mutation/deletion.
modules presented signs of dysfunction during the first

three years after implementation. We showed that failing
modules present early signs of dysfunction which can be
identified as early as the first month after implementation
by using adequate automatic reporting systems. We
conclude that the one-year free-of-charge warranty from
the manufacturer will cover the vast majority of costs
related to sub-optimal equipment delivered. Nevertheless, as failing modules can be difficult to identify, it is
recommended to implement the adequate surveillance
solutions in order to maximize the benefit that can be
made from this warranty.
15. GeneXpert: the nuts and bolts of TB

diagnosis
PC-817-04 Failing XpertW MTB/RIF modules:
When ? What warranty is needed from the
customers perspective? Experience in the DR
Congo
E Renard,1 E Andre2,3 1Universite Catholique de Louvain,
Louvain-la-Neuve, 2Universite Catholique de Louvain,
Brussels, 3Cliniques Universitaires Saint-Luc, Brussels,
Belgium. e-mail: emmanuel.andre@uclouvain.be
Background and challenges to implementation: Between

2012 and 2014, the Democratic Republic of Congo has
gradually implemented 22 GeneXpert machines. The
overall proportion of non-valid results (i.e. hh error ii, hh
invalid ii and hh no result ii) was 8%. Among the 68
analytical modules composing the 22 GeneXpert, 32%
presented a significantly higher rate of non-valid results
during the first 3 years of implementation, leading to
unnecessary work and consumption of reagents as well
as loss of diagnostic opportunities. The objective of this
study was to determine whether it was possible to early
identify modules presenting sub-optimal performance.
Intervention or response: Data was collected using the
GenXchange (UC Louvain, Belgium; NTP, DRC) and
XpertSMS (IRD, Pakistan) technology. We separated the
modules in two groups: 22 modules identified with a high
proportion of non-valid results, and 46 modules presenting normal figures. For each group, we computed the
proportion of non-valid results based on time after
implementation of the GeneXpert. We used a moving
average on 30 days.
Results and lessons learnt: The figures below represent
the percentage of non-valid results between the two
groups of modules and the progressive decrease in
activity of the modules with the time. In the group with
a higher rate of non-valid results, all showed signs of
dysfunction (proportion of not-valid results higher than
10%) from the first month after implementation and
increased significantly with a peak of over 60% around
month 7 (210 days). After 7 months of activity, virtually
none of these modules was still in activity, compared to
the non-failing group of modules, which continued
reporting tests.
Conclusions and key recommendations: It had been
previously reported that 32% of GeneXpert MTB/RIF
PC-818-04 Stability of XpertW MTB/RIF reagents

in sub-optimal storage conditions:evaluation
based on three years experience in the
Democratic Republic of Congo
C Vande Kerckhove,1 E Andre2,3 1Universite Catholique de
Louvain, Louvain-la-Neuve, 2Universite Catholique de
Louvain, Brussels, 3Cliniques Universitaires Saint-Luc,
Brussels, Belgium. e-mail: emmanuel.andre@uclouvain.be
Background and challenges to implementation: Since

2012, the Democratic Republic of Congo (DRC) has
performed 23.227 GeneXpert MTB/RIF tests. Cartridges are the main consumables for this technology and
must be stored between 2 and 288C. In DRC, these
instructions cannot be fully controlled due to local
climatic and infrastructure constraints. Based on international quality standards, medical laboratories must
perform additional verifications when storage instructions cannot be strictly observed. The objective of this
study was to determine the impact of the sub-optimal
storage on the quality of reagents.
Intervention or response: Out of 22 lots of cartridges, 3
were considered as defective based on hypothesis testing
and were excluded from analysis. Data was collected
using the GenXchange (UCLouvain, Belgium; NTP,
DRC) and XpertSMS (IRD, Pakistan) technology. We
focused on two quality indicators. First, the evolution in
the proportion of non-valid results (i.e. Error, Invalid, No result): we computed for each test the
remaining days left before the expiration date of the

cartridge and whether the result was valid or non-valid.
The second indicator was the distribution of the internal
quality control (Cycle threshold or Ct of the SPC) among
all tests.
Results and lessons learnt: Figure A represents the
evolution of the proportion of non-valid results over
time. The expiration date is the graph origin for all lots of
cartridges. There was no significant increase of non-valid
results observed when the expiration date was closer. The
Histogram (Figure B) represents the Ct distribution for
the SPC internal quality control. The empirical probability density function was obtained by fitting a
distribution to the data histogram. The Ct values of
SPC ranged between 23 and 37, with a mean value at
27.3. We observed a normal distribution of these values.
Conclusions and key recommendations: In this study, we
evaluated the quality of Xpert MTB/RIF reagents during
their life-cycle post-shipment by the manufacturer. We
conclude that sub-optimal storing conditions did not
seem to impact the quality of reagents over time, as there
was no significant increase of non-valid results in the
time and the distribution of SPC Ct was considered as
normal, although we cannot definitively exclude from
this analysis that high SPC values were not linked to
cartridges exposed to particularly sub-optimal storage
conditions or linked to cartridges approaching expiry
date.
S207
Design/Methods: 67 souches de Mycobacterium tuberculosis provenant de patients tuberculeux recus pour

diagnostic, suivi ou en e chec de traitement ont e te testes
sur le système MGIT 960 (Streptomycine, Isoniazide,
Rifampicine, Ethambutol) qui est un test qualitatif sur
milieu liquide. Les resultats obtenus ont e te ensuite
compares a` ceux de lautomate GeneXpert qui est un test
de diagnostic in vitro par PCR semi quantitatif en temps
reel pour la detection de lADN du complexe tuberculosis
et la mutation du gène rpoB associe a` une resistance a` la
Rifampicine.
Results: Sur la totalite des souches testees par le système
MGIT 960 , 41 e taient sensibles a` tous les antibiotiques
testes et 26 presentaient une monoresistance a` au moins
un antibiotique. 16 souches avaient donne un profil de
resistance a` la Rifampicine sur le système MGIT 960
alors que le système Xpert en avait donne que 15 parmis
ces derniers. Deux souches resistantes a` lIsoniazide et
sensible a` la Rifampicine sur MGIT 960 e taient
resistantes a` la Rifampicine sur le systeme Xpert.
Conclusion: La sensibilite de la dectection des souches
resistantes a` la Rifampicine est plus importante avec le
systeme GeneXpert que le syteme automatise sur milieux
liquides MGIT 960 et devrait etre mis en place dans les
laboratoires pour une meilleure prise en charge des
patients tuberculeux sous traitement.
PC-820-04 Online XpertW recording and

reporting system in Kenya
J Ogoro,1 E Masini,1 M Kamene,1 T Miura,2 S Chebore,3
M Mburu4 1National TB Program Kenya, Nairobi, 2Japan
International Cooporation Kenya, Nairobi, 3USAID, Centre
for Health Solutions, Nairobi, 4Centre for Disease Control,
Nairobi, Kenya. Fax: (254) 271 3198. e-mail:
jeergo@yahoo.com
PC-819-04 Etude comparative de la sensibilite

de Mycobacterium tuberculosis a` la rifampicine
`
sur MGIT 960 versus systeme
XpertW
M Y Fall Niang,1,2 A Ba Diallo,1,2 A Ndiaye Diawara,1,2
G W Ossoga,1,2 A Thiam,1,2 A Gaye-Diallo1,2 1Bacteriology
and Virology Laboratory, Centre Hospitalier Universitaire
Aristide Le Dantec, Dakar, Senegal. e-mail:
ossogagedeon@gmail.com
Background: Le but de cet e tude est de faire une

comparaison des resultats des tests de sensibilite sur
milieux liquides MGIT 960 vs. ceux obtenus sur le
syst e me Xpert a` partir de souches provenant
dechantillons recus au laboratoire de BacteriologieVirologie de lhopital
Aristide Le Dantec.
Background: Kenya ranks 13th in 22 high TB burden

countries and currently has 70 GeneXpert testing sites.
For the last four years the recording and reporting has
been paper based in diagnosis and surveillance of
tuberculosis. Patient/clinician data collection is currently
captured online and relay of results through email and
text messages to the clinicians immediate they are
available through automation. All the GeneXpert equipment are bundled to support the system at minimal costs
Design/Methods: All the generated laboratory request
forms have patient demographics, clinicians mobile
numbers and their emails and tests required. Laboratory
technologists transcript the request form details onto the
same soft copy in the system before they commence
running the test. After the testing is completed, the
automated system relays results electronically to the
emails and mobile numbers of the requesters. Consumption data is captured in the same system through site
monthly updates.
Results: Out of the 70 sites trained and supported to use
the system, 63 are actively utilizing the system. The
online System has prevented missing records and reduced
clerical errors. Management of Commodities has been
made easier since sites update their monthly stock status
before testing since this is an automatic prompt by the
S208
system hence making Forecasting and Quantification

process is accurate. Data on actual testing and commodity availability onsite is available too
Conclusion: The adoption of online recording and
reporting tools into the testing system has markedly
reduced TAT for TB diagnosis and DR-TB patients to be
put on treatment faster. Additional benefits include
reduction of missing laboratory results.
accompanied by wider efforts at strengthening the

broader laboratory system itself.
PC-821-04 Understanding the introduction of

new laboratory technology in a poorly
resourced setting: the case of GeneXpertW in
Nepal
Background: In sub-saharan Africa tuberculosis is a

significant occupational problem with the laboratory
considered a work area with one of the highest risks
regardless of immune status. Staff shortages, and drug
resistance confound the problem. There is no safe level of
exposure; often it is difficult to directly associate work
exposure with disease. Resources limit optimal interventions but simple practical measures can mitigate risk.
Aims and Objectives: The aim was to investigate a
potential tuberculosis outbreak in our laboratory with
the objective of breaking transmission. A laboratory that
performs 2500 M. tuberculosis cultures per month.
Culture manipulation for drug susceptibility testing is
done. 56 staff, including trainees, work here. One nontechnical staff member had confirmed tuberculosis in
2015, 2 others had disease in the preceding 5 years.
Methods: Laboratory confirmation was by the Xpert
MTB/RIF assay and culture. Staff education sessions on
tuberculosis and biosafety were held and symptomatic
screening administered. Xpert MTB/RIF and chest
radiography was performed on screen positives. Review
of previous safety audit reports, risk analysis and
environmental testing was carried out. Respirator fit
testing was performed. Procedure specific interventions
were enhanced and some implemented.
Results: Work areas had no mechanical ventilation.
Regularly serviced Class II Biosafety cabinets were in use.
Only 5/31 (16%) workers passed fit testing using the TSI
PORTACOUNT PRO. None had pre-employment chest
X-rays. 100% had completed quarterly health questionnaire,7 (12.5%) workers reported risk factors for TB. 4
suspects tested negative on Xpert MTB/RIF. Poor access
control, inconsistent use of PPE and BSCs were among
risky work practices identified through audit.
Discussion: An epidemiological link was not established
but 12.5 % had high risk co-morbidity. While regular
health surveillance of workers was done, processes for
pre-employment and pre-placement assessments have
had to be developed. Testing for latent TB infection was
not established. Available respirators were unsuitable
resulting in more expensive alternatives being sort.
Ethnic variations in facial dimensions need to be
considered when selecting this PPE. Education of staff
and behavioural change was identified as important in
ensuring a safety culture. Inadequate infrastructure and
poor equipment should not be over emphasised to the
neglect of these.Centralised testing is an option.
R Khatri,1 K Dahal,1 S Baral,1 I Harper2 1Health Research

and Social Development Forum, Kathmandu, Nepal;
2
University of Edinburgh, Edinburgh, UK. e-mail:
rrekha.khatri@gmail.com
Background and challenges to implementation: Nepal, a

country with a high TB burden, good TB treatment and
cure rates, aimed to increase its case finding rate to over
80% by 2015.
Intervention or response: GeneXpert, endorsed by the
WHO, was introduced in Nepal from 2011, in an
attempt to target the unreached populations and increase
the case finding rate through early detection of tuberculosis. A range of REACH funded projects introduced
within the National TB programme through its various
partners operated under various modalities. While some
of the partners have static centers for GeneXpert, one of
them is also running mobile services through labs in vans.
Results and lessons learnt: Extensive participant observation of the GeneXpert installation process in TB labs
across the country was undertaken. In addition, interviews were conducted with a range of participants in the
process, including lab staff, TB-leprosy officers and
managers of the organizations implementing the TB
REACH programmes. Although the technology does
improve the speed of diagnosis and acts as a proxy for
MDR diagnosis, broader issues of general laboratory
capacity, sustainability and health systems strengthening
remained unaddressed. Staff demanded incentives for the
additional workload. The maintenance of the machine is
a key challenge in poorly resourced settings where basic
lab infrastructure is poor, was expensive, time consuming
and undermined performance. Nepal currently lacks
local capacity for maintaining the machine as the
cartridges and modules have to be sent to France when
there are problems. The cartridges are also expensive and
the question of sustainability as the REACH grants end
has yet to be addressed.
Conclusions and key recommendations: Examining the
implementation of GeneXpert technology in a poor
resource setting like Nepal offers valuable lessons for
broader attempts at laboratory and health systems
strengthening. Maintaining a technology like GeneXpert
(or any other new technological intervention), requires
certain preparedness on behalf of government to sustain
its effective use. While new technology does support the
detection and thus treatment of a disease, it should be
PC-822-04 Mycobacterium tuberculosis in the

laboratory: when the hunter becomes the
hunted
B Magazi1 1University of Pretoria, Pretoria, South Africa.
e-mail: beki106@gmail.com
PC-823-04 Operational implementation of a

quality assurance monitoring protocol for
GeneXpertW in a hospital setting in Kampala,
Uganda
S209
should include parallel implementation of a standardized

QA protocol.
L Asege,1 D Armstrong,2 A Andama,1,3 M Joloba,3

F Semitala,1,3 J Lemaire,4 A Cattamanchi,5 C Yoon5
1
Infectious Diseases Research Collaboration, Kampala,
Uganda; 2Johns Hopkins University, Baltimore, MD, USA;
3
Makerere University College of Health Sciences, Kampala,
Uganda; 4 Foundation for Innovative New Diagnostics,
Geneva, Switzerland; 5University of California, San Francisco,
San Francisco, CA, USA. e-mail: lucyasege@gmail.com
Background: GeneXpert MTB/RIF (Xpert), a semiautomated nucleic acid amplification assay for rapid
tuberculosis (TB) diagnosis, has undergone widespread
scale-up in high TB burden countries since 2010. We
assessed the proportion of indeterminate results at the
national referral hospital of Uganda, before and after
implementation of a standardized quality assurance
(QA) protocol.
Methods: In July 2013, two Xpert platforms were
installed in Mulago Hospital. Routine monitoring of
test results and device maintenance was not performed
until November 2014, when a standardized quality
assurance (QA) protocol was implemented. QA activities
included implementation of manufacturer-recommended
daily, weekly and monthly device maintenance and the
development of a monitoring toolkit to capture key
quality indicators including the number of tests performed and the number of indeterminate results (errors
and invalids), per month. To evaluate the impact of these
QA activities, we extracted test results from archived
data stored on both platforms to calculate the monthly
frequency of indeterminate results. We then evaluated
trends in the type and proportion of indeterminate results
before and after implementation of the QA monitoring
protocol.
Results: From August 2013 to March 2015, 4362 tests
were performed (2448 on machine A, 1914 on Machine
B) of which 727 (17%) were positive for M. tuberculosis,
560 (13%) showed test errors, and 70 (2%) were
invalids. From August 2013 to March 2014, the median
error rate was 6% (IQR 4-8%) and exceeded the 5%
upper limit established by the manufacturer for 5/8
months (range 2-11%; Figure). Despite calibration and
replacement of failed modules in September 2014, error
rates continued to increase, peaking at 69% in October
2014. The most common error codes were #2127
(module communication loss, n 349) and #1004
(internal instrument temperature out of range, n 83).
Following implementation of our standardized QA
protocol, the median Xpert error rate decreased to 4%
(IQR 3-4%; range 2-4%).
Conclusion: In the absence of an Xpert QA protocol,
error rates rapidly increased and routinely exceeded 5%,
resulting in wasted cartridges/lab resources and possible
delays in TB diagnosis. For Xpert to achieve maximal
impact, a systematic process for monitoring test results
and routine device maintenance is required; annual
calibration alone is insufficient. Future scale-up of Xpert
PC-824-04 Evaluation of the impact of XpertW

testing on the diagnosis and cascade of care of
drug-susceptible tuberculosis in Epworth,
Zimbabwe
R Harrison,1 C Ssonko,2 J Greig,2 J Arhem,1 T Maparo,1
`
E C Casas,3 D Legrand,1 D Fusco3 1Medecins
Sans Frontieres
(MSF), Operational Centre Amsterdam, Harare, Zimbabwe;
2
MSF, London, UK; 3MSF, Amsterdam, Netherlands. e-mail:
rebecca8harrison@gmail.com
Background: This retrospective study aimed to evaluate

how the introduction of GeneXpert in 2012 influenced
the diagnosis, treatment and outcomes of patients with
drug sensitive Tuberculosis in Epworth, Zimbabwe. In
2013, Zimbabwe reported an HIV prevalence of 14%
and a TB incidence rate of 409 per 100 000. 69% of TB
patients were HIV co-infected. TB case detection rate
was estimated at 42%.
Design/Methods: All smear microscopy and GeneXpert
results recorded for the 4683 patients registered for any
type of TB treatment in Epworth and Overspill polyclinics from 15/3/2010 until 31/12/2013 were recorded.
GeneXpert was introduced in 2012 to the onsite
laboratory in one of the clinics, whereas previously only
fluorescent Smear Microscopy was utilised for bacterial
confirmation. Summary descriptive statistics were calculated comparing the pre-GeneXpert and GeneXpert era.
Results: After the introduction of GeneXpert the number
of sputum confirmed patients put on TB treatment per
month rose from 39 to 49, but the total number of
treatments started went from 117 to 88 per month. The
median time from sputum test to treatment initiation for
smear positive patients was 3 days before and after
GeneXpert implementation. Before the implementation,
196 people were tested for TB per month vs. 189 after.
84% (95%CI: 82-86%) of the first sputum submitted
during the pre-GeneXpert era were of good quality
compared to 67% (95%CI: 64-69%) during the GeneXpert Era. Of all the patients starting treatment at
Epworth Polyclinic, 72% (95%CI: 70-73%) successfully
S210
completed treatment before implementation compared

to 66% (95%CI: 64-69%) after. Treatment success for
pulmonary confirmed MTB went from 71% (95%CI:
68-74%) to 69% (95%CI: 65-72%) and for smear
negative cases from 70% (95%CI: 67-73%) to 65%
(59%-70%). Death rates of MTB confirmed went from
5% (95%CI: 4%-7%) to 8% (95%CI: 6%-10%), LTFU
from 12%(95%CI: 10%-15%) to 9%(95%CI: 7%11%) and patients with outcomes not evaluated went
from 8%(95%CI: 6%-10%) to 12%(95%CI: 10%14%).
Conclusion: In this context, the introduction of GeneXpert was associated with an increase in number of
sputum confirmed TB patients but not with an increase in
total TB treatments started. Median time between test
and start of treatment remained unchanged while
treatment success rate for all cases decreased slightly.
Further investigations will be carried out into changes in
TB prevalence, false positives amongst smear negative
cases and changes in quality of care provision.
16. Retaining our patients in care

PC-825-04 Factors contributing to good TB
treatment adherence among patients in North
West Region of Cameroon
A Z Achidi,1 W Mbacham,2 I Ndong,1 M Sander,3 L Ayuk,4
P Aseh1 1Catholic University of Cameroon, Bamenda,
2
University of Yaounde 1, Yaounde, 3Bamenda Regional
Hospital, Bamenda, 4Regional Delegation of Public Health,
Bamenda, Cameroon. e-mail: achidiasanga@gmail.com
Aim: To investigate aspects of patient and healthcare

worker knowledge and behaviour that is contributing to
high adherence to TB treatment in the Northwest region
of Cameroon, where only 3% of patients are lost to
follow-up as compared to the national average of 8%
(NTP data).
Methods: A convenient sample of 38 adult TB patients
and 18 TB treatment nurses were recruited in the study.
Indepth semi-structured interviews were conducted for
all study participants. In addition, four focus group
discussions (FGDs) were held for three groups of 10
patients each and one group of 10 TB nurses.
Results: From this qualitative assessment, several factors
were identified that may potentially contribute to the
relatively high TB treatment adherence in this region.
From the in-depth interviews administered to TB
patients, a majority of patients reported taking their
medicine on their own, without a designated treatment
supporter, by using an alarm system to remind them of
the time to take their treatment (21 of 38 respondents). In
addition, the TB nurses surveyed indicated a high level of
knowledge about the correct administration of TB
treatment including the importance of TB treatment
adherence. The patients surveyed also had good knowledge of most aspects of tuberculosis treatment reported
little or no feeling of stigmatization by their families or
communities due to the disease and reported a good

patient-health service staff relationship during the FGDs.
Conclusion: These results suggest that in similar settings
where TB treatment administration is not directly
supervised by a healthcare worker, high rates of
treatment adherence and completion may be achieved
by other methods, including patient use of mobile phone
alarms as treatment reminders and strong engagement by
health care workers in patient education at treatment
initiation.
PC-826-04 Use of counselling to retrieve and

retain lost to follow-up multidrug-resistant
patients in India
S Waikar,1 A Kapoor,1 B Entoor Ramachandran,2
S Chadha2 1Population Services International, New Delhi,
2
International Union Against Tuberculosis and Lung Disease
8779. e-mail: swaikar@psi.org.in
Background and Challenges: India accounts for 64 000

MDR-TB cases accounting for 22% of global burden.
One of the major challenges in the management of MDRTB is lost to follow-up (LTFU) patients. Evidence from
HIV/AIDS programming indicates that counselling can
support adherence to long-term treatment by addressing
psychosocial issues. Yet the RNTCP program in India
does not offer counselling services for TB and MDR-TB
cases.
Intervention: Since April 2014, Population Services
International (PSI), under the Global Fund supported
Project Axshya is implementing MDR-TB counselling in
28 districts across 12 States. PSI recruited and trained
professional counsellors with masters degrees in social
work and 2-3 years of experience in counselling. The key
roles of MDR-TB Counsellors include providing facility
and home based counselling to all MDR TB patients
currently on treatment with an objective to ensure
adherence and treatment completion, linking MDR TB
patients to appropriate social welfare schemes and
retrieving lost to follow up patients before and after the
initiation of treatment.
Results and Lessons Learnt: A total of 9878 MDR-TB
cases are registered in 28 intervention districts as of
December 2014. 275 patients were recorded as LTFU in
intervention districts before April 2014. From May14 to
Dec14, MDR Counsellors tracked and counselled 55%
(153) of patients LTFU. Non availability of proper
addresses and migration are two major factors for not
being able to track the remaining 45% of LTFU cases.
49% (74) of tracked LTFU cases were retrieved by MDR
counsellors and brought back on treatment. Empathetic
counselling, regular follow-up, catalysing family and
social support and MDR patient support group meetings
helped to bring these patients back on treatment. MDR
counsellors continue to make efforts to retrieve the
remaining LTFU cases in the districts, however loss of
faith in treatment, intolerance to side effects, substance
abuse and regular supply of MDR medicines are some of
the key barriers that limit success.
S211
Conclusion: Empathetic counselling, regular follow up,

facilitating family and social support and MDR patient
support group meetings helps to retrieve LTFU cases.
Counselling should be an integral part of RNTCP
program. Based on the initial success in MDR-TB
counselling, the counselling component is going to be
scaled up at 30 additional districts under Project Axshya.
PC-827-04 Poverty induced migration causes TB

treatment default in tribal district, Lohardaga,
Jharkhand, India: calls to enforce TB and
migrant strategy
1
food incentive to check migration along with adequate

counselling facility during treatment.
Conclusion: Default in tribal area is deep rooted with
issues of migration/poverty, knowledge gap, poor health
seeking behaviour & systemic gaps in counselling,
follow-up & barrier managements. Tribal TB control
needs an integrated style with policy/strategy for
migrants, extra HR, fund, admin committment &
linkage mechanism.
Table Cases registered and default status in Lohardaga Tribal
district of Jharkhand, India
2012
S Nayak, R Dayal, S Chadha, V Ghule, O Bera,

R Pathak,3 A Das,1 P Mishra,2 K Rout4 1International Union
Office, New Delhi, 2State TB Cell, Ranchi, 3World Health
Organization Country Office India, Ranchi, 4Catholic Health
Association of India, Lohardaga, Jharkhand, India. Fax: (91)
114 605 4430. e-mail: snayak@theunion.org
Background: Jharkhand (India) a tribal dominated state

with 26% tribal population (of 35 million) live in hilly &
hard to reach terrains, registered 72 227 TB cases
including 44 874 (62%) smear positive (S ve) in 2012
& 13. Lohardaga with 0.5 million population, is a tribal
district & comes under 5th schedule of the Indian
Constitution with 56% tribal. The paradox is while
default rates of new (5%) & re-treatment (RT) (10%) S
ve cases of Jharkhand remains same of 2012 & 13, that
of Lohardaga was 10% (new) & 21% (RT) in 2012;
16% and 22% respectively in 2013. The objective of this
study is to find out the barriers to high default in tribal
areas & strategy to overcome those.
Method: It includes desk review of 811 cases (527 S ve)
with 610 (75%) tribals registered during 2012 & 13 in
the context of tribal vis-a-vis
non-tribal whose outcomes
`
declared. 25 default patients selected out of 75 S ve
defaulted in a proportion 1:3 and interviewed through a
semi-structured questionnaire; 3 Focused Group Discussions (FGDs) were conducted among 7 DOTs Providers,
6 NGOs & 8 programme implementers to address the
objective.
Results: Among 811 (550M, 261F) cases, 117 defaulted
(15%) that include 94M (17%) & 23F (9%). The case
ratio of tribal to non-tribal is 3:1 and in case of default, it
is 3.7:1. Among male, default is 17% & 9% in female.
Male to female ratio is 2:1 in total case & 4:1 in
defaulters, the same in case of tribals are 2:1 & 9:2
respectively. Out of 25 tribal default patients interviewed, 10 (40%) defaulted because of migration to
other states for felt need of survival. All (100%) are of
education of below 10th standard with less than $100
monthly income. Further, 16% found to be chronic
alcoholic & 12% defaulted due to drug side effect and
shifted to private. Most have inadequate knowledge on
TB & think of cure after 3/4 months. 7 cases (28%)
found dead after defaulting, 3(12%) complete the
treatment after re-registration and 1 switched to MDR.
The FGD findings stressed upon addressing issues of
migration & alcohol practice. It suggests for monetary/
Particulars
2013
Total
FeFeFeMale male Total Male male Total Male male Total
Total
Number of
TB cases
Registered 297 135 432 253 126 379 550 261 811
Tribal
226 96 322 191 97 288 417 193 610
NonTribal
71 39 110 62 29 91 133 68 201
Smear
Positive TB
Cases
198
Tribal
156
NonTribal
42
Total Default
Tribal
NonTribal
Smear
Positive
Default
Tribal
NonTribal
67 265 185
52 208 140
77 262 383 144 527

63 203 296 115 411
15
57
45
14
59
87
29 116
47
34
11
6
58
40
47
41
12
11
59
52
94
75
23 117
17 92
13
18
19
25
28
20
8
4
2
32
22
35
30
8
7
43
37
63
50
12
9
75
59
10
13
16
PC-828-04 Axshya Project (TS&AP): experience

of retrieval of potential TB drug interrupters
A Kumar,1 P Varma,1 PSundaresh,1 V Panibatla,1
S Mukhopadhyay,2 S Cornelius,2 B Samuel2 1TB Alert
India, Hyderabad, 2World Vision, New Delhi, India. e-mail:
emailtoarunk@gmail.com
Background and challenges to implementation: Sub

Recipient (SR) for Global Fund for Andhra Pradesh
State, TB Alert India is implementing Axshya India
Project in 7 districts of Telangana & Andhra Pradesh
states. . The phase 2 of project is being implemented since
April 2013. Despite decades of intense drive two deaths
occur every 3 min from tuberculosis (TB). The drug
resistant cases threaten to undermine the program at
various levels. At one level there is the difficulty in
treating the cases and on another level are the misconceptions in the community. In the present scenario of
universal access to TB care and TB declared a notifiable
disease, a look into the challenges faced by the ones
struggling to complete the treatment.
Intervention or response: One of project activity objectives is to increase TB treatment adherence among
potential Drug Interrupters (missing one week dosage
to More than 2 months) retrace them and bring them
S212
back to DOTS treatment. Retrieval of Drug Interrupters,

bringing them back to DOTS and ensuring that they have
taken the medicines at least for one week time (minimum
3 doses).The project collected the information about the
reasons of default during Drug Interrupter retrieval
activity. As a part of activity, project staff advise the TB
patient on importance of continuation of long TB
treatment regimen and counsel addressal of side-effects
of drugs , nutritional education, family education, anger
management, usage of motivation card for self-motivation, confidence building and restoring self-esteem.
Results and lessons learnt: Between Apr13 to December
2014, TB Alert has retrieved 619 Potential TB Drug
Interrupters (age ,18 Yrs- 4%, 19-45 Yrs- 46% , 46-60
Yrs 30% and .60 Yrs 20% , Average 42 ). Of them
472 (74%) number male, and 14 7(24%) were females.
The major reasons mentioned for interrupting the drugs
is due to side effects 65% (Male- 405 and Female-96) and
21% (Male-35 Female 97) interrupted thinking that
they got cured after initiation of treatment
Conclusions and key recommendations: The experiences
of Drug Interrupter Retrieval analysis stressed the need
for prompt assistance and greater addressal of side effects
and counselling of patients on importance of TB
treatment completion.
PC-829-04 Risk factors for loss to follow-up

among DR-TB patients on second-line
treatment in Kyrgyzstan
A Toktogonova,1 A Estebesova,2,3 E Kotysheva,4
S Fedorova,5 E Maliukova,1 N Kurmanova,6 A Kadyrov,1
S Verver7,8 1National Phthisiatry Center under the Ministry of
Health, Bishkek, 2KNCV Tuberculosis Foundation, Country
Office, Bishkek, Kyrgyz Republic; 3Open Society Foundations,
New York, NY, USA; 4Kyrgyz State Medical Academy,
Bishkek, 5International University of Kyrgyzstan, Bishkek,
6
Kyrgyz State Medical Institute on Postgraduate Education
and Re-training, Bishkek, Kyrgyz Republic; 7KNCV
Tuberculosis Foundation, Hague, 8Academic Medical Centre,
AIGHD, Amsterdam, Netherlands. e-mail:
aida.estebesova@gmail.com
Background: Kyrgyzstan has one of the highest rates of

MDR-TB in the world, with 68% MDR-TB among
previously treated patients and 26% MDR-TB among
new cases. About 20% of MDR-TB patients are lost to
follow-up of treatment (LTFU). Reasons and risk factors
are often unknown. The objective of this study was to
assess risk factors associated with loss to follow-up from
TB treatment among drug resistant patients on second
line treatment in Kyrgyzstan.
Design/Methods: Four out of nine regions (oblasts) in
Kyrgyzstan were selected for feasibility reasons. Within
these regions all new and previously treated drug
resistant (DR) TB patients who were over 18 years of
age and started second line drugs (SLD) in 2011 and first
half 2012 were included from the main national level
registry. We studied risk factors by comparing patients
successfully treated with patients LTFU.
Results: Out of the expected 625 patients fulfilling the
selection criteria, 347 patients records were found
(56%). Among the 347 records, 288 patients were

included for risk factor analysis including 73 LTFU
(21%) and 215 successfully treated (62%). 59 patients
had other treatment outcomes. The following factors
were associated with LTFU in multivariate analysis:
being treated in the intensive phase in ambulatory care
had a lower risk of LTFU vs. those hospitalized (OR for
0.34, 95%CI 0.13-0.91), having sputum conversion after
6 months or never had a higher risk of LTFU vs. those
with sputum conversion within 6 months (OR 1.72,
95%CI 0.93-3.19).
Conclusion: In Kyrgyzstan, drug-resistant TB patients
getting ambulatory treatment in intensive phase and
patients with early sputum conversion have a lower risk
on LTFU than other DR patients. It should be more often
considered to treat DR patients on ambulatory care.
Patients who do not have sputum conversion after 6
months should get extra attention to prevent LTFU.
PC-830-04 Common difficulties in tracing

tuberculosis treatment interrupters
M Mohamad,1 N A Noor,1 N Abdullah,1
W Al-Kubaisy-Abd-Al-Qahar1 1Universiti Teknologi MARA,
Sungai Buloh, Malaysia. e-mail: yammo33@yahoo.com
Background: It is stated in the International Standards

for TB Care, that healthcare providers (HCP) have the
responsibility to address factors of treatment interruption (TI) or discontinuation which failure in doing so is
considered as not meeting the international standard.
Managing TI is vital in preventing relapse, drug
resistance and death due to TB. It is an important
activity and that it should be done in organized manner
in order to meet the standard care. Our intention in this
study is to get our HCP perceptions on the difficulties in
tracing TB cases with TI. We hope from the information
gathered in this study, we could aid our local health
system in addressing any difficulties in tracing TI cases.
Design/Methods: A cross sectional study was carried out
in one of the states in Malaysia involving primary HCP
who were involved in managing TB patients. Each of the
HCP was given a self administered questionnaire which
contained demographic profile and their rankings on the
reasons of difficulties in tracing TI. The respondents were
required to rank 1-10 where 1 being the most important
and 10 the least important of 10 reasons of difficulties in
tracing TI.
Results: From the 236 healthcare personnel responded in
this study, 162 (68.4%) admittedly had some involvement in managing TI. Majority of them are Malays
(84.0%), more than half (55.6%) are females and 89.5%
of them have attended TB management course. Their
median age was 30 years old (IQR 9) and median
duration of service was 5 years (IQR 7). The median
duration of involvement with TB care was 24 months
(IQR 36months). The median number of days for the
HCP to act after noticing a TI was 3 days (IQR 6). Only
86 (53.1%) of the HCP, responded to the questions on
ranking the reasons of difficulties in tracing TI. The
reasons of difficulties according to the most top to the
lowest ranks (mode) were as follows: 1. Incomplete

address, 2. Homeless, 3. Non-existent address, 4. Wrong
address, 5. Patient from other district, 6. Legal Immigrants, 7. Drug abusers, 8. Illegal immigrants, 9. Patient
busy working, and lastly, 10. Patients in detained
institutions.
Conclusion: The common difficulties in tracing TI in our
setting seem to be around the address issues and being
homeless. The incomplete, wrong and non-existent
address can be dealth with during the first encounter
with TB patients. We recommend that our HCP to
develop a system to verify the addresses given by the
patients and also obtained other addreses.
PC-831-04 Social issues are an important

contributor to lost to follow-up during TB
treatment: experience from the North and
Northeast of Namibia
L Petersen,1 S Dalebout,2 S Neri,1 D Moongo,1
R Indongo,3 F Mavhunga,4 T N Shilongo4 1Project HOPE,
Windhoek, Namibia; 2Project HOPE, Millwood, VA, USA;
3
United States Agency for International Development,
Windhoek, 4National TB and Leprosy Programme, Windhoek,
Namibia. e-mail: sneri@projecthope.org
Background and challenges to implementation: Project

Hope Namibia (PHN) supports the MoHSS in the
provision of community-based TB care in 6 districts
across 4 Regions. Our work focuses on advocacy,
treatment adherence, early identification and referral
for TB treatment as well as continued adherence support
through DOT and follow up. The program is successful
with LTFU of under 5%, but this still represents a
significant number that warrants review.
Intervention or response: Community health workers
implement the DOT strategy and trained DOT supporters reach the household providing adherence support
during treatment. When interruption occurs, tracing is
conducted. Once the patient is traced, interventions
include restart or continuation of treatment, referral for
adherence counselling, and revisiting the role of the
treatment supporter.
Results and lessons learnt: Patient records over the past
three years indicated 4760 patients registered for TB
treatment, of which a total of 95 (2%) were lost to
follow-up (LTFU) with more than half lost for less than 4
weeks. Of the 95, 91(96%) were successfully traced.
55% of them gave socio-economic problems as a reason
for treatment interruption, 21% gave no reason for
interruption, and 14% relocated. LTFU occurred after
the initial intensive phase of treatment, the first two
months. Side effects that may be is experienced during
this initial phase did not seem to contribute, however
patients may feel clinically better and thus may think
treatment is no longer required. The socio-economic
problems indicated by 55% of patients were either not
identified or attended to prior to initiation or during
treatment. In more than 80% of cases, referral for
problems, such as lack of finances to travel, alcohol
abuse, lack of treatment support, relocation for better
S213
economic opportunities, and others, did not have the

desired outcome.
Conclusions and key recommendations: Review the
current DOT strategy. Programming should be strengthened and tailored to address social issues at the onset of
treatment with structured counselling, referral and
follow up and on-going adherence education.
PC-832-04 Improving continuum of care for

mine workers on TB treatment across borders
during holidays
N Sigwebela,1 A Hani2 1University Research, Pretoria,
2
Aquity Innovations, Pretoria, South Africa. e-mail:
ntombim@urc-sa.com
Background and challenges to implementation: Managing TB across borders presents a challenge to National
TB programs (NTP) in Southern Africa. The circular
migration to and from labor-exporting countries exacerbates TB transmission across communities. For mine
workers with TB, the disruption in the continuum of care
results in treatment interruptions caused by various
challenges relating to unequal access to care and
perceived poor quality of health care services, specifically
treatment medicines.
Intervention or response: Aquity Innovations/URSA with
funding from the DGF project of the World Bank,
implemented a large scale intervention targeting miner
workers diagnosed with TB to ensure continued provision of comprehensive TB treatment services across the
borders. The primary objective was to promote continued patient support and ensure uninterrupted care for
mine workers on TB treatment during holidays. The
intervention was implemented in partnership with two
mining companies in South Africa as well as the NTPs in
each of the 4 countries. The 3 interventions implemented
were 1) provision of a survival kit to each mine worker
which contained educational materials on TB adherence,
TB medicines available in their country and a list of
health facilities 2) provision of an incentivized mobile
phone to receive sms reminders, confirm compliance
with treatment and call for help if needed 3) Follow up by
nurses to remind the mine worker to take their
medications. The nurse was incentivized with $3 worth
of airtime for each follow up made. In addition local
media was used to broadcast messages on TB treatment
and importance of adherence.
Results and lessons learnt: 399 mine workers with TB
participated, .900 sms reminders sent over 10 days, 540
mobile phone responses received from mine workers
confirming treatment was taken
Conclusions and key recommendations: The existing
systems are adequate for ensuring continuum of care for
mine workers with TB. However, for optimal care and
support to be achieved, a multi stakeholder approach
involving employers, workers and the ministries of health
is critical. Incentives for both service providers and
patients can also be an effective way of ensuring
uninterrupted care.
S214
PC-833-04 Involvement of sub-county health

workers in improving cure rates
A Batwaula,1 A Mugume,1 R Kimuli,1 H Ndagire1 1John
Snow Inc JSI Research and Training Institute, JSI/STAR-EC,
Jinja, Uganda. e-mail: abatwaula@starecuganda.org

national cure rate stands at 40% 2012/2013 National
TB and Leprosy Programme (NTLP) reports. Whereas At
inception of STAR-EC program in 2009, the overall cure
rate stood at 30% for the nine districts of East- central
Uganda. The low cure rate was attributed to failure for
patients to turn up to facilities for follow up, inadequate
information given to the patient at initiation and
absenteeism of laboratory staff at time of patient visit.
Intervention or response: JSI with funding from USAID is
implementing the STAR-EC program that supports the
National TB and Leprosy program (NTLP) to strengthen
TB control activities in the nine districts of East Central
Uganda. STAR-EC supported a training of Sub-county
health workers (SCHW) on smear preparation and TB
infection control. This was followed by attachment to a
diagnostic facility with in their respective subcounties
for several days to perfect smear preparation skills. The
SCHWs were then supported to follow up patients in the
communities on a monthly basis to collect sputum and
prepare slides from clients that were due for sputum
follow up then transported to facilities to be examined by
the lab staff
Results and lessons learnt: Over a period of 5 years, there
has been progressive improvement in the cure rate cure
rate from 30% to 70%%. Active involvement of lay
providers trained on slide preparation in the communities in TB control activities improves cure rates as
opposed to waiting for patients to visit facilities
Conclusions and key recommendations: In resource
constrained settings where human resource for health is
a challenge, utilization of community volunteers to
support community TB sputum follow is a cost effective
strategy.
PC-834-04 An ethnographic account that

assesses drug-resistant tuberculosis patients
ability to adhere to treatment in Cape Town,
South Africa
L Winterton1,2 1University of Cape Town, Cape Town, South
Africa; 2University of Edinburgh, Edinburgh, UK. e-mail:
winterton.laura@gmail.com
Background and challenges to implementation: Approximately 13 000 news cases of Drug-Resistant Tuberculosis (DR-TB) are reported in South Africa every year.
Treating DR-TB requires an intensive drug regimen for
the duration of the 18-24 months. The pharmacological
benefits of the treatment cannot be understated. However, adherence to treatment remains a profound
challenge for both patients and their health care
providers. I explored how DR-TB patients not only
understand and cope with the extreme side effects, but
how they develop processes and strategies that alleviate
some of the physical and psychological symptoms that

surface during treatment.
Intervention or response: This was a four-month,
qualitative ethnographic study conducted in Khayelitsha,
Cape Town in 2012. I observed weekly DR-TB support
group sessions. I conducted 18 illness history interviews
with patients drawn from the support group sessions. I
also conducted 5 in-depth case studies where I witnessed
patients in their day-to-day lives moving from the clinic,
their homes, shopping centres and church. At the end of
my research I performed 2 female focus group discussions (FGD) and two male FGDs to elaborate on themes
of gender and identity that surfaced during interviews
and observations. All interviews and FGD were voice
recorded and detailed field notes were kept throughout
the research.
Results and lessons learnt: Participants in this study were
not labeled as Loss to Follow UP (LTFU), but all
participants expressed moments of pausing or experimenting with their treatment as a coping strategy. The
patients viewed DR-TB treatment as a conflicting process
whereby the side effects felt more damaging than the
infection. The profound psychosocial effects from toxic
drug regimens destabilise a patients sense of self and it
restructures their most intimate valued relationships.
Conclusions and key recommendations: This research
reveals the need for more and continued qualitative
research that can capture the social and psychological
complexities of treatment seeking behaviour. Side effects
are the most significant reason for DR-TB patients LTFU.
Patients expressed the need for more psychological and
social support that recognises the profound and complex
journey from diagnosis to cure. TB treatment strategies
require more robust psychological support that can
speak to range and complex side effects that surface
during treatment and after treatment is complete.
17. Including those affected in care

PC-835-04 Improved urban DOTS strategy:
strengthening referral linkages between
tertiary hospitals and treatment (DOTS) centres
S Islam,1 T Hirayama,2 M Akramul Islam,1 S Reza,1
N Ishikawa,2 M Haque,3 M Siddiqui1 1BRAC, Dhaka,
Bangladesh; 2Research Institute of Tuberculosis, Tokyo,
Japan; 3National Tuberculosis Control Programme, Dhaka,
Bangladesh. Fax: (88) 2988 8026. e-mail:
shayla.dhaka@gmail.com
Background and challenges to implementation: Urban

TB control is a major challenge for a developing country
like Bangladesh. The National Tuberculosis Control
Programme (NTP) of Bangladesh first adopted the
directly observed treatment short course (DOTS) strategy in 1993 as World Health Organization declared TB
as a global emergency. The programme rapidly scaled up
in the following years to almost all areas of the country
and DOTS was expanded to Dhaka metropolitan city in
2002 through partnership with different NGO partners
including BRAC. Initiatives taken to establish referral

linkage between tertiary hospital and peripheral DOTS
centres to initiate early treatment.
Intervention or response: TB symptomatic from different
corners of the city comes to these 15 tertiary hospitals for
diagnosis which is run by BRAC. After diagnosis TB
patients residing closely to the hospital are registered and
given DOT from the hospital DOTS corners. Medicine is
also given to the admitted TB patients. Discharging
patients are referred to the peripheral DOTS center
according to patients convenience. Counseling to the TB
patient is done to report the referral center in time and to
initiate treatment promptly. A referral form (TB-07) is
given to the patient to send its part back after reporting
into the peripheral DOTS centre.
Results and lessons learnt: In 2014, a total of 7589 TB
patients were diagnosed in 18 tertiary level hospitals of
Dhaka city. Among them, only 122 (1.6%) were
registered in the hospital and most of the TB cases
(6789; 89%) were referred to different DOTS centres
according to the patients convenience. Among the
referred patients 4667 (69%) were reported to the
different DOTS centers following NTP established
referral system, 280 (3.6%) left from hospital after
discharge without referral siip and others 398 (5%) are
died/medicine stopped/self medicated. In 2010, reported
case in different DOTS centres was 28% which has been
improved in 2014 (69%).
Conclusions and key recommendations: As a large
proportion of TB patients are diagnosed from these
tertiary hospitals, linkage with the DOTS centers is
crucial following NTP established referral mechanism
which reduces the number of missing cases after
diagnosis. Proper counseling of TB patient is necessary
during referral to peripheral DOTS centres.
PC-836-04 Can contact investigation play a role

in missing one million in India?
B Entoor Ramachandran,1 S Chadha,1 S Waikar2
South-East Asia Office, New Delhi, 2Population Services
International, New Delhi, India. Fax: (91) 114 605 4430.
e-mail: ebabu@theunion.org
Background: The Global TB Report 2013 estimates 2.9

million cases are missed every year (out of the estimated 9
million incident cases) and 31% of these missing cases (~
1 million) are in India. A recent study in India found
8.7% of household contacts were diagnosed with TB. As
per policy, contacts of TB cases should be screened.
However, the practice is not systematic across the
country.
Intervention: Project Axshya focuses on engaging all
sectors to strengthen TB care and control in 300 districts
across 21 states of India. This project is implemented by
The Union with support from the Global Fund. In order
to facilitate early diagnosis, customized Contact Investigation (CI) intervention was designed and implemented
in 120 Tuberculosis Units (TU) in 60 project districts
with the involvement of community based volunteers
S215
Results: In project locations, 19 269 TB patients (all

types) were registered during the period Jul Dec 2013.
The project aimed to reach 12 126 (63%) smear positive
TB cases (30 % female) with contact investigation
services but only 8425 (69%) TB patients have been
reached. There were 5866 children household contacts
and 26% (1489) of them were on Isoniazid Prophylaxis
Treatment (IPT). 1857 (4%) contacts out of 51 411
(including 12% children below 6 years) were having
symptoms during the visit. 997 (54%) adult persons have
undergone the diagnosis and 74 (7.42%) of them were
diagnosed as having TB. The prevalence of TB among
symptomatic children was 7.32%. It was found that 8.02
% of persons were having TB out symptomatics tested
under special intervention. Services and cost of CI
activity implemented by Project Axshya in India Services
offered Rates (covering travel and honorarium) Visit to
TB patient household Creating awareness on TB
Preparing symptomatics for TB diagnosis 1 US$ (0 - 20
kilometer) per household 2 US$ (beyond 20 km) per
household Sputum collection and transportation symptomatic from persons to designated microscopy center
(DMC) Sharing of lab reports with symptomatics Refer
Not TB patients to health facilities Linkage of identified
TB patients with treatment centers 2 US$ (0- 10 km) per
symptomatic 3 US$ (beyond 10 km) per symptomatic.
Conclusion: In India 1 410 880 total TB cases were
notified during the year 2014. As per this study,
systematic implementation of CI strategy could have
identified additional 17 337 TB patients during the year
2014 and this will also increase in propionate to index
case identification.
S216
PC-837-04 A study on awareness and stigma

related to tuberculosis among patients
attending a district TB centre, Kollam, India
P Teena,1,3 P Rakesh,2 K Viswanathan,4 C Joltin,3
V Sameer,3 P Bhat,3 T Thomas3 1School of Medical
Education, Kottayam, 2District TB Control Office, Kollam,
3
Catholic Health Association of India, Secunderabad, 4Indian
Medical Association, Delhi, India. e-mail:
sameer-valsangkar@chai-india.org
Background: TB stigma leads to TB diagnostic delay and

treatment non-completion and has a more significant
impact on women and poor or less-educated community
members, which is especially concerning given that these
groups are often at higher risk for health disparities. TB
stigma may worsen pre-existing gender- and class-based
health disparities. The study was conducted to evaluate
stigma related to tuberculosis among patients in District
TB Centre, Kollam.
Design/Methods: The study was conducted over a period
of three months among patient attending the District TB
Centre, Kollam. Convenience sampling method was used
to select 50 of the 120 TB patients attending the center
during the study period. Data was collected through
personal interview with the help of a structured schedule.
Results: The sample consisted of predominantly (20%)
elderly population between the age of 60 to 65 years,
among whom 72% were males and 28% were females.
Awareness was measured by asking simple questions
regarding TB disease and only 30% of people were well
aware of TB. 14% of the people agreed and 18% of
people strongly agreed that they felt ashamed for having
TB disease.
Conclusion: Stigma and lack of awareness are still a
factor in tuberculosis which impact the life of these
patients and can hinder seeking care. Awareness generation and stigma redressal components must be a part of
initiatives against tuberculosis in India.
PC-838-04 Civil society TB forum gears up to

protect the rights of TB patients and help
ensure social security of TB patients
S Nayak,1 S Chadha,1 S Mohanty,1 A Das,1 S Diwakar,2
P Banuru Muralidhara,1 K Deepak,2 J Thithio3
1
South-East Asia Office, New Delhi, 2Catholic Health
Association of India, Ranchi, 3Emanuel Health Association,
Jharkhand, India. Fax: (91) 114 605 4430. e-mail:
snayak@theunion.org
Background: The Union implements Project Axshya

(Global Fund supported) through its eight sub recipients
in 21 states of India covering 300 districts. The project is
in operation in 15 districts of Jharkhand. Catholic
Health Association of India (CHAI), Emanuel Health
Association (EHA) and Population Service International
(PSI) are the Sub Recipients (SRs) to The Union in
Axshya Project in Jharkhand. As part of the project
strategy, district TB Forums have been formed in each
district with representatives from various cross-sections
of the society, such as social repute opinion leaders,
Doctors, Media persons, retired bureaucrats, business

men, philanthropists and more importantly cured TB
patients.
Intervention: Axshya Project through its SRs has formed
district TB forums in all 15 district of the state. The TB
forums mission is to work for ensuring rights and
entitlements of TB patients and strengthening the TB
control measures as civil society representatives. The
forum advocates with TB patient, TB department, other
departments for convergence and social protection of TB
patients and civil society groups including media. This
paper is an outcome of the qualitative analysis of visit
and discussion with members of five District TB Forums
in Jharkhand.
Results and lessons Learnt: The forums are successful in
taking the cause of TB emergency to the higher level and
dissemination at the public by advocating for social
security schemes for TB patient, sensitising patients on
patient charter, meeting and local political representatives for involvement in TB, providing nutritional
support to TB/MDR TB cases (from red cross, their
own fund, corporates and advocate with Social Welfare,
PRI & Women Child departments) and raising the issues
in the media. In Hazaribagh, Garwa & West Singhbhum
districts, the forum has raised nutritional support for
MDR patients. In Palamu and West Singhbhum districts,
the forum has given memorandum to the administration
for financial support from the district welfare fund to
below poverty line TB patients. Across the state ~30
patients benefitted during 2014. The forum in Hazribagh
has taken up issues of notification.
Conclusion: The District TB Forum represents a solid
body representing civil society for the benefits of the TB
patients and equally supports to the TB control measures
as well. More representation of TB patients in the forum
will further strengthen in taking up gaps and barriers
starting from cough to cure.
PC-839-04 Systematic screening of contacts of

DR-TB patients in Andhra Pradesh: results of a
rapid survey
C Chatla,1 J Jaju,1 S Achanta,1 C Purad,1
R Chepuri Nagaraj,1 R Samyuktha,2 A Sreenivas,1 P
Parmar1 1World Health Organization Country Office Inida,
New Delhi, 2State TB Cell, Hyderabad, India. e-mail:
chatlac@rntcp.org
Background: India has the worlds highest estimated

burden of multi-drug-resistant tuberculosis (MDR-TB).
While prevalence of MDR-TB is known to be 2 3%
among new TB cases and 12-17% in previously treated
cases, programmatic information on the prevalence of
MDR-TB/ TB among contacts of known MDR-TB cases
is not available. A state wide, house to house systematic
screening of household contacts was conducted by
RNTCP (Revised National Tuberculosis Control Program), Andhra Pradesh to understand the prevalence of
TB/MDR-TB among contacts of known MDR-TB
patients under treatment.
Intervention: All MDR-TB patients in the State registered

between July 2011 and Sep 2013 and currently on
treatment under the programme were listed. House to
house visit was made as per pre-defined survey protocol
by the programme supervisory staff. All household
contacts were screened for symptoms and if found
positive, 2 sputum samples were collected (one spot
and one morning) and transported for testing on
GeneXpert.
Results: A total of 2750 MDR-TB patients were
registered between July 2011 and Sep 2013. Of these,
2167 MDR-TB patients were included as per survey
protocol and 7324 contacts of these patients were
screened for past history of diagnosis or treatment of
TB. After excluding those with past history of diagnosis
or treatment for TB, a total of 6823 contacts were
screened for current signs and symptoms of TB. Of these,
785 had at least one of the symptoms suggestive of TB
and their sputum samples were transported and processed on GeneXpert. A total of 34 contacts among 785
tested (4.3%) were diagnosed with TB. Of these 17 were
Rif. sensitive (2.2%), 15 were Rif. resistant (1.9%) and 2
were indeterminate results. Of the diagnosed TB cases,
50% (n 17) were found to be Rif sensitive TB cases, and
44.1% (n-15) Rif resistant TB while the remaining 5.9%
Rif indeterminate TB cases.
Conclusions: Based on the estimation of 203 TB cases/
lakh population as per Annual Risk of Tuberculosis
infection in India and estimation of 2% MDR-TB among
new cases (4 MDR-TB cases/lakh population); it can be
concluded that being a contact of MDR-TB patient
increases the chance of being diagnosed with TB by 2.4
times and MDR-TB by 55 times. Rapid diagnostics must
be utilized for early diagnosis in this vulnerable group. It
reinforces the need of regular contact screening for early
detection and prompt treatment besides educating the
patients and family on infection control measures.
S217
PC-840-04 Factors associated with performance

of tuberculosis support volunteers in Lusaka,
Zambia, 20142015
Y Toyama,1 M Ota,2,3 Y Onoe,1,3 I Njovu,3 Y Takemura,1,3
A Ito,3,4 G Samungole,5 S Hirao,2,3 H Ogata2,3 1Japan
Anti-Tuberculosis Association, Tokyo, 2Research Institute of
Tuberculosis, Tokyo, Japan; 3RIT/JATA-Zambia, Lusaka,
Zambia; 4Japan International Cooperation Agency, Tokyo,
Japan; 5Lusaka District Community Health Office, Lusaka,
Zambia. e-mail: ytoyama@jatahq.org
Background: Mobilising tuberculosis support volunteers

(TSs) is a widely employed approach for tuberculosis
(TB) control especially in communities that traditionally
suffer shortages of health care workers. Yet, individual
performance of the TSs and its associated factors have
not been fully identified. Since 2012, Japan AntiTuberculosis Association conducted a project aiming at
early detection of presumptive TB cases as well as
providing treatment support through mobilising TSs for
urban marginalized sectors at the selected health centres:
Bauleni, Chelston, and Chilenje in Lusaka, Zambia
through Japan International Cooperation Agency Partnership Programme. Objective of this study is to identify
associated factors influencing attendance of TSs to their
activities after 3 years of project implementation.
Methods: A cross-sectional study was conducted among
TSs at the three clinics. Data were collected through
structured interviews. TSs individual performance was
measured by latest 6 month attendance rate to weekly
patients support activities, such as health education and
drug adherence support. Descriptive and multiple logistic
regression analyses were done to identify factors
influencing individual attendance of TSs.
Results: Of the 74 (97% of the total) respondents, the
average monthly attendance rate between September
2014 and February 2015 was 83.2% (95% Confidence
Interval (CI) 79.2-87.2). The attendance over 80% was
likely to be achieved by the TSs who were in a temporary
or no work (adjusted odds ratio [aOR]: 8.9, 95%CI 6.8
45.5) and have experienced the executive members
among the groups (aOR: 7.0, 95%CI 5.437.3).
However, the TSs who had secondary education or
above were less likely to attend over 80% than those who
did not completed the primary education (aOR: 0.10,
95%CI 0.083-0.77). No associations were found between the level of attendance and gender, distance to the
designated health facilities, marital status, TB knowledge
and satisfaction of project activities.
Conclusion: The study suggested in recruiting TSs, the
priority should be those who have a temporary or no
work, rather than a fulltime work. Selecting an
individual as an executive member of the TSs may
facilitate the person to more commit the activities. Those
who have secondary education might be more likely to
commit other activities, such as income generation
activities, than the TS activities and should be selected
as TSs with caution.
S218
PC-841-04 Adherencia al tratamiento antituberculoso en pacientes de un ensayo clnico
que recibieron acompanamiento

comunitario
vs. pacientes que reciben tratamiento en
condiciones programaticas
J Jimenez,1 R Yataco,1 B Martel,1 D Garcia,1 L Panduro,1
E Osso,2 C C Contreras,1 C Mitnick2 1Socios En Salud
Lima, Peru; 2Harvard Medical School, Boston,
Sucursal Peru,
MA, USA. e-mail: jjimenez_ses@pih.org
Antecedentes: La adherencia al tratamiento y resultados

oportunos para TB, disminuyen los riesgos de morbilidad, mortalidad y adquisicion
de drogorresistencia. El
acompanamiento
comunitario a traves de promotores de
salud busca mejorar la adherencia al tratamiento,

brindar soporte emocional y educacion
al paciente
durante y despues de la administracion
del tratamiento.
El presente estudio describe los resultados de tratamiento
de pacientes que se incluyen a un ensayo clnico con
acompanamiento
comunitario vs. pacientes que reciben
tratamiento en condiciones programaticas.

Metodologia: Se incluyeron pacientes nunca tratados que
iniciaron tratamiento anti-TB, esquema 1, en el periodo
setiembre 2013 - agosto 2015, en Lima - Peru.
En el mes
de marzo del 2015 se revisaron las condiciones de egreso
del tratamiento de 160 pacientes, identificandose:
pacientes que se incluyeron a un ensayo clnico con
acompanamiento
comunitario y pacientes que recibieron
tratamiento en los EESS de acuerdo a los estandares

programaticos. Se colectaron datos de los registros de
tratamiento, registrados por personal de investigacion
o
por el personal de salud de los EESS.
Resultados: 127 pacientes fueron incluidos a un ensayo
clnico y 33 pacientes recibieron su tratamiento en los
EESS del MINSA. De los 127 pacientes que participaron
en el ensayo clnico, todos accedieron a un resultado de
Prueba de Sensibilidad (PS) Rapida antes de iniciar el
tratamiento; y las condiciones de egreso fueron: 94
(74%) curados y 33 (26%) se transfirieron a su EESS por
terminacion
temprana (9 criterios de exclusion
del
estudio, 17 presencia de eventos adversos, 3 retiro
voluntario, 1 recluido en penal, y 3 abandonaron el
tratamiento debido al consumo de drogas). De los 33
pacientes que recibieron tratamiento en los EESS, solo
17
(52%) tuvo un resultado de PS, siendo las condiciones de
egreso: 18 (55%) curados, 9 (27%) continuan
en
tratamiento (3 irregulares y 6 iniciaron tratamiento

tardamente, 2 (6%) cambiaron de esquema de tratamiento por resistencia a INH, y 4 (12%) no tienen datos.
Conclusion: El acompanamiento
comunitario y el acceso
a la prueba de sensibilidad rapida podran mejorar la

adherencia al tratamiento obtenie ndose un mayor
porcentaje de curacion
y termino de tratamiento en los
tiempos establecidos.
PC-842-04 Improving TB control in urban

settings through engagement of community
linkage facilitators: the case of Kampala City
C Nanziri,1 D Lukoye,1 A Etwom,1 S Ntudhu,2 D Okello,3
B Woldemariam,1 P G Suarez,1 E Birabwa4 1Management
Sciences for Health, Kampala 2AIDS Information Centre,
Kampala, 3Kampala City Council Authority, Kampala, 4USAID
UGANDA, Kampala, Uganda. Fax: (256) 414 235 035.
e-mail: cnanziri@msh.org
Background and challenges to implementation: Over

20% of patients that were registered for TB treatment in
Kampala were lost to follow up in 2012, DOT was less
than 27% and cure rate at 24%, barely half of the
national target of 65%.With a daily 1:20 health worker
patient ratio (TRACK-TB quarter 2015 report), patient
retention and provision of sufficient patient counseling
and adherence support through the course of TB
treatment was very challenging.
Intervention or response: Kampala Capital City Authority with support from USAID funded TRACK TB
project, received 46 community linkage facilitators
(CLFs) who actively followed up bacteriologically
confirmed TB patients at 20 selected high TB burden
health facilities. A monthly average of 12 patients were
assigned to each CLF to monitor directly observed
treatment (DOT) and sputum smear follow ups at
months 2, 5 and 6/8. DOT was considered if the patient
treatment supporter was counseled, received health
education and ascertained daily patient observation of
swallowing TB medicines with in the first 6 weeks of
treatment. Counseling and health education was done by
the CLFs at first contact, at a home visit during the
intensive phase of treatment and the biweekly phone calls
during the course of continuation phase treatment.
Follow up fortnight telephone calls were made to ensure
continuity of DOT, track patients who missed appointments and remind patients to keep sputum monitoring
appointments.
Results and lessons learnt: By December 2014, the
number of new bacteriologically confirmed TB patients
receiving directly observed treatment had increased to
95% and the cure rate to 73%. The improvement was
consistent at health facilities that were allocated CLFs
compared to those that had no CLFs. This was attributed
to follow up efforts for DOT and sputum monitoring
through home visits and telephone calls. Follow up
activities were affected by patient mobility in Kampala;
5% were lost to follow up, 1% were transferred out, 5%
died and 16% had a treatment completed outcome.
Patients who maintained residence and contacts were
easily followed up at CLF supported sites.
Conclusions and key recommendations: Community
linkage facilitators helped to improve DOT and cure
rates for TB patient in Kampala. This may be a useful
intervention that needs to be expanded to other facilities
and urban settings with similar challenges.
PC-843-04 Self-reported health of TB patients in

Republic of Macedonia: is less more, or is it
even more complicated?
D Gudeva Nikovska,1 F Tozija1,2 1Faculty of Medicine,
University Ss Cyril and Methodius, Skopje, 2National Institute
of Public Health, Skopje, Macedonia, Yugoslav Republic. Fax:
(389) 2317 7983. e-mail: dgnikovs.ka@gmail.com
Background: Studying self-reported health is considered

an indicator for morbidity and mortality that may be
used in primary health care to detect poor health in
certain population groups that predicts health care
utilization. Prognostic value of self-rated health is
particularly important among older adults and literature
acknowledge self-rated health as lying at the cross-roads
of culture and biology, therefore it needs a collaborative
effort between different disciplines to improve understanding of this key measure of health status. The goal of
the survey is to assess the socio-economic self-rated
health gradient and to describe contribution of behavioral risk factors to this gradient among TB patients in
Republic of Macedonia.
Design/Methods: Data is collected through a nested
case-control study, conducted in the period March
December, 2013. Cases are households with TB
patient(s) registered in the period July, 2012 June,
2013 and controls are households randomly chosen in
cases immediate vicinity. The total study population is
605 households with total of 2720 respondents. Selfreported health is assessed through the question How
will you rate your health status today? with a five-point
scale ranging from 1 very satisfied to 5 very
dissatisfied). Data was analyzed with SPSS 19.0, utilizing
logistic regression to measure predictive value of most
important factors that determine self-reported health.
Results: Self-rated health was reported as excellent or
good by only half of the respondents, with slightly less
positive answers among TB cases compared to controls
and evident differences in responses for poor or extreme
difficulties in everyday life provided by TB patients.
Analyzed by group of respondents, 50.8% cases and
49.1% controls assess their health status as good, but
differences are significant in categories bad with dominant responses by TB cases with 9.3% vs. 4.9% controls
and very bad with 10.3% positive answers among TB
cases vs. no such response in controls with statistically
significant difference in answers among groups
(X226.410, df4, P , 0.001). Positive association
was found between poor rated health and longstanding
diseases, such as asthma and arthritis, and education was
associated with poor self-rated health. Adding questions
on mobility, self-care, pain, cognition, interpersonal
activities and affect has only reaffirmed poorer health
of TB patients, with statistically significant differences
among study groups along all six dimensions assessing
mobility.
Conclusions: The study is the first step towards
establishment of evidence that self-rated health is an
important indicator to be followed by both clinicians and
health planners in Republic of Macedonia. Further
research is needed to determine if self-rated health
S219
assessment in routine clinical setting can be used to

identify groups at risk for increased mortality and other
important health outcomes.
18. Dont forget us: migrants and

PC-844-04 Do migrants from developing
countries transmit M. tuberculosis to locals in
the host country? Results of a molecular
epidemiology study in Israel
Z Mor,1,2,3 E Rorman,4 D Chemtob,1 P Freidlin,4
N Cedar,1,4 H Kaidar-Shwartz1,4 Z Dveyrin,4 D Goldblatt4
1
Ministry of Health, Jerusalem, 2Ministry of Health, Ramla,
3
Tel Aviv University, Tel Aviv, 4Ministry of Health, Tel Aviv,
Israel. Fax: (972) 3683 6632. e-mail:
turkiz1@netvision.net.il
Background: Health authorities in developing countries

raise the concern of a possible transmission of M.
tuberculosis from migrants who originate in high
incidence countries to the local population. This study
aimed to assess the penetration of imported isolates of
migrants to the local Israeli population and describe the
predominant strains in Israel.
Methods: Molecular characterization was employed by
43-spacer spoligotyping and 16 loci MIRU-VNTR typing
to all new culture-positive strains reported between 2008
and 2010. Patients were classified by those who were
members of a cluster, and to those who were not. In case
Israeli-born patients were members of a cluster of M.
tuberculosis isolates with non-Israeli born patients, they
were questioned if they could recall a possible contact
with a non-Israeli born individual.
Results: Of 741 culture confirmed cases reported during
the study period, 684 (92.3%) were genotyped and 83
different clusters were identified. The major spoligotype
families in Israel included Central Asian (CAS) (n 140,
20%), Beijing (n 101, 15%) and T (n 160, 23%).
CAS strains were mainly (74%) from Ethiopia, Eritrea
and Sudan, while most (66%) Beijing strains were
isolated from patients originating from former Soviet
Union. Membership in a cluster was more common
among younger patients, males, Ethiopian-born, migrants who had longer stay in Israel, smear positive and
MDR cases. Of all 83 clusters, 30 (36%) were mixed,
and included 40 Israel-born and 291 non-Israeli born
patients. Of these Israeli-born patients, 24 (55.8%)
shared the cluster only with non-Israeli born citizens,
12 (27.9%) shared the cluster with both non-Israeli born
citizens and non-Israeli born non-citizens and 4 (9.3%)
shared the cluster with non-Israeli born non-citizens. Of
all the Israeli-born, 15 (37.5%) could recall a close
contact with an individual who was non-Israeli born
citizen, but not with non-Israeli born non-citizen.
Conclusions: Cluster analysis indicated that membership
in a cluster was more common in migrants, suggesting a
probable transmission among communities of foreignborn. Despite the concern of a possible penetration of M.
S220
tuberculosis from migrants to Israeli-born in the host

country, only limited transmissions were observed from
non-Israeli born citizens, yet none from non-Israeli born
non-citizen.
PC-845-04 Spatiotemporal analysis of

tuberculosis in South American immigrants in
Sao Paulo municipality, 2006-2013
P Pinto,1 M Ribeiro,1 M J Penon Rujula,1 C Silveira1
Faculdade de Ciencias
Medicas
da Santa Casa de Sao Paulo,
Sao Paulo, SP, Brazil. e-mail: priferscaff@hotmail.com
Background: Sao Paulo is the biggest municipality in

Brazil and the principal economic center in the country.
Since 1980, Sao Paulo has been an important destination
to immigrants from other countries of South America,
especially Bolivia. These immigrants are attracted to
work in sewing workshops, where they are subjected to a
working system of semi-slavery. This precarious situation
exposes them to many health risks and therefore to the
tuberculosis (TB) infection. The aim of this study is to
describe the spatiotemporal distribution of tuberculosis
in South American immigrants in Sao Paulo from 2006 to
2013.
Methods: Secondary data were obtained from an
information system of Sao Paulo State called TB-WEB.
It was considered to the survey all new TB cases in South
American immigrants, residents in Sao Paulo municipality. Socioeconomic variables associated to the monitoring
of TB were used to describe the sample. Data were
analyzed on EpiInfo. The geocoding of the cases was
performed through the QGIS software.
Results: During the study period, 1904 cases of
tuberculosis in South American immigrants were reported, representing 4.2% of total cases in Sao Paulo. In
2006, there were 155 cases, in 2011, there were 307
cases. 63% were male, the dominant age group was 2039 years (76.5%), the most reported skin color was
indigenous (26.7%). In 30.8% of the cases, the education
level was 8-11 years of study. 87% were employees and
46% were couturiers. The majority of cases (84.3%)
were classified as pulmonary tuberculosis, the subjects
discovered they were ill after finding a primary health
care service (57.2%) and a emergency care service
(25.1%). 9.5% did not realize the sputum smear
examination, 62.2% had manifestations of TB in results
of chest X-ray, 63.4% did the TB supervised treatment,
the abandon rate was 15.2% and the cure rate was 74%.
On the beginning of the survey, the geocoding showed
that the South American immigrants with TB were
concentrated in the downtown region, but over the years
they are moving to the suburbs especially in the Northern
part of the city.
Conclusions: The number of South American immigrants
with TB is increasing in Sao Paulo municipality, and the
municipal health system must establish a strong policy to
control the disease in these vulnerable groups and
improve cure rates. It must also direct health interventions to the most affected areas of the city.
PC-846-04 Enhanced tuberculosis management

among Syrian refugees in Jordan
L Kutkut,1 A Galev1 1International Organization for
Migration, Amman, Jordan. e-mail: lkutkut@iom.int
Background and challenges to implementation: Four

years of civil war has depleted the health infrastructure in
Syria; the exposure to tuberculosis (TB) emerges as a
grave public health concern for the Syrian refugees &
host populations throughout Jordan (including Zaatari
Camp {83 501 refugees} & Azraq Camp{17 738} of the
628 427 registered Syrian refugees in Jordan). Due to the
dynamic movement of Syrian refugees, some TB patients
migrate to other countries without prior notice, making
them difficult to trace. The International Organization
for Migration (IOM) in collaboration with Jordanian
National Tuberculosis Program (NTP) & with funding
support of partners (United Nations & Global Fund),
work towards strengthening NTPs capacity to address
TB-related challenges that exceed the existing capacities
of Jordanian health authorities.
Intervention or response: Targeted TB activities include: TB Prevention: Starts through IOMs First Health
Screening as refugees arrive at the Jordanian Transit
Center. Awareness is raised about signs & symptoms of
TB & some effective practices to avoid the spread of TB.
IOM conducts TB training to primary health care
providers to ensure effective referrals of presumptive
TB cases.TB Screening: IOM medical mobile teams
conduct CX-rays with a portable CX-ray unit coordinated by Jordanian Ministry of Health. TB Diagnosis &
treatment: IOM facilitates diagnostic investigations: CXray, 3 direct sputum smears, cultures & GeneXpert. All
confirmed cases are then referred to NTP to confirm the
diagnosis & avail drugs. IOM medical teams monitor
Directly Observed Treatment (DOT), side effects &
treatment progress. Challenges arise when TB patients do
not adhere to treatment thus needing close follow up. 50
% of the 62 extra-pulmonary TB patients suffered from
comorbidities & complications like vertebral deformities
& paralysis, requiring costly medical interventions
ranging from 70-14,000 $. Transportation for follow
up visits & nutritional support with a high protein diet

are provided.
Results and lessons learnt: In spite of all the challenges,
the treatment success rate for 2013 cohort was 96% with
1 death & 3 % default rate. The figure summarizes data
from March 2012- March 2015.
Conclusions and key recommendations: Joint efforts by
the NTP & International community are necessary to
address challenges of TB & reduce susceptible TB
transmission, morbidity & mortality among Syrian
refugees residing in Jordan.
PC-847-04 Overview of the TB situation among

migrants in Kazakhstan
S Ussembayeva,1 E Berikova,1 S Pak,2 T Markabayeva2
National Tuberculosis Center, Almaty, 2Representative Office
of the Royal Netherlands Tuberculosis Association (KNCV) in
Central Asia, Almaty, Kazakhstan. Fax: (7) 727 291 8658.
e-mail: saule_usem@mail.ru
S221
Conclusion: There is a need in Kazakhstan of development and implementation of complex measures on

prevention of TB spread among migrants service, PHC
network, international and non-government organizations. The activities on quality of TB care for migrants
are provided in the Complex Plan to combat Tuberculosis in Kazakhstan for 2014-2020 years.
PC-848-04 Challenges in managing MDR-TB in

Dadaab refugee camp on the Kenya-Somali
border
D Miriti,1 H Mohammed,1 Y Mohamed,1 A Abdullahi,1
D Nyachieo,1 B Opare,1 N Marwan1 1International
Organization for Migration, Nairobi, Kenya. Fax: (254) 272
2818. e-mail: dmiriti@iom.int
Background: The successes on stabilization and improvement of TB epidemiological situation were

achieved since implementation of WHO strategy in
Kazakhstan. For last 10 years the rate of TB population
morbidity was reduced for 54.9% and mortality for
77.4%. At the same time for the last years due to
economic growth Kazakhstan shifted to category of the
countries with active inflow of migrants who are related
to vulnerable population with high risk of TB. According
to various expert evaluations Kazakhstan annually
receives from 300-500 thousand up to 1 million working
migrants, the flow of which is weakly regulated by the
state.
Design/Methods: In the frame of implementation of 8
Round of the Global Fund project the operational
research on the spread of TB, MDR-TB and the TBHIV among migrants in Astana and Almaty. One of the
goals of the research is investigation of the sociodemographic and health characteristics in external
migrants diagnosed with tuberculosis for the period
2009-2012.
Results: Results of the research allowed to define the
portraits of migrants arrived in Kazakhstan and taken for
treatment to TB facilities. Usually they are young people
at able-bodied and reproductive age (81%), at the age of
21-50 with secondary education (63.6%). The mostly are
males (63.6%) arrived from Uzbekistan, Kyrgyzstan and
Russian Federation (86.3%). 42,7% of migrants do not
have families and majority (80%) of them are living in
Kazakhstan less than 1 year. Third of them are working
migrants and more than half (60%) have legal status of
staying in the country. More than 90% arrived migrants
are renting dwelling with accommodation of more than 3
people. 87% of migrants have duration from the moment
of identifying the first symptoms to medical treatment
about more than 1 month. TB diagnosis often takes 1
week. Treatment success registered in 72.2% cases but
violation of regime is observed in 7,3% cases. Results of
cases outcomes of defaulters demonstrated that all of
them were lost and went out of the country to unknown
direction.
Background: Dadaab refugee camp is located near the

relatively porous Kenya-Somalia border with 351 446
registered refugees as of March 2015. Various agencies
and partners under the overall coordination of UNHCR
provide various services to refugees in the camp including
health. The MDR-TB treatment centre is located within
the camp. It was opened in 2009 with 20 beds capacity
and later expanded to now cater for an average 100
patients. In a partnership with Kenya NTP, CDC,
UNHCR, IOM took over the provision of clinical
services at this MDR-TB ward in January 2013. The
NTP provides the TB medicines and some laboratory
reagents, while UNHCR supports the infrastructure
maintenance and feeding patients through its implementing health partner in the camp.
Intervention: A dedicated clinical team of 1 clinician, 2
nurses and 6 incentive workers carry out the daily tasks
of patient care and management. External consultation
services requested as needed. Security provisions with
armed escort accompany staff to the ward daily. Directly
Observed therapy of TB treatment, continued education,
clinical review with adverse effects monitoring and
scheduled investigations form part of daily routine tasks.
Diagnosis is based on clinical evaluation, sputum testing
with smear microscopy, molecular gene Xpert testing as
well as culture and DST for both first line and second line
TB medicines. The ward is supported by IOM TB
laboratories with capacity for all above TB investigations. Baseline blood tests are done in partner laboratories in the city - Nairobi and results sent back to health
care worker by email. Moving to a larger facility 2014
made it possible to separate infectious/intensive phase
patients from those long converted. The ongoing patient
care and management seems to act as a pull factor
continually attracting ever increasing patient flow.
Results: 2009 to date, the facility has enrolled a total of
184 patients for MDR-TB treatment. 47 (26%) of these
were enrolled in the first four years 2009-2012, 137
(74%) were enrolled in 2013-2015. All the 184 patients
tested HIV negative. Female to male ratio is 1:3 (50 and
134 respectively). 5 out of 184 patients were ,15 yrs old.
144 patients (78%) are NOT registered refugees but are
migrants who crossed into Kenya solely in search of
MDR-TB treatment. 94% of all patients were treatment-
S222
after-failure patients. GeneExpert tests supported by

DST confirmed MDR-TB for all the patients. No XDR
patient identified so far. Approx 80% were malnourished
at initiation of treatment. Currently 101 patients have
come to a definitive outcome (84% cured, 5% defaulted,
12% died). 83 are still ongoing with treatment. The
influx of patients from Somalia continues.
Conclusions: There is need for all partners in Kenya,
Somalia and beyond to source for funds to strengthen
management of TB both pan susceptible and drug
resistant & cut down transmission.
PC-849-04 Tuberculosis among the key health

hazards of labor migrants from Central Asia in
Moscow
D Kashnitsky,1 E Demintseva1 1Higher School of
Economics, Moscow, Russian Federation. e-mail:
kashnitsky@gmail.com
Background: Previous studies of migrants from Central

Asia living and working in Russia suggest that their life is
characterized by limited resources and social exclusion.
Lack of health insurance, scarce information about
medical infrastructure and discrimination practices turn
out as barriers on migrants way to timely and quality
medical care. Heavy physical work, cramped living
conditions and poor nutrition are the factors that largely
contribute to the new cases of tuberculosis. Central Asian
countries have much higher prevalence of TB than Russia
(Kyrgyzstan 190 per 100 000, Russia 121 per 100
000).1 [
Design/Methods: The study is based on the analysis of
qualitative interviews with 30 caregivers working in
Moscow-based medical facilities including TB clinics.
Results: We have identified several strategies of seeking
medical care in Moscow. Migrants tend to disregard an
occurring disease as they often do not have time, money
and information to seek treatment. Only emergency care
is fully accessible to migrants in Moscow free of charge.
When diagnosed with an open form of TB migrants still
had to be accorded permission from the Moscow city
health department before they could be placed in the TB
clinic. This often took time so migrants had to remain
without treatment while waiting. During this period they
continued to work being exposed to their colleagues and
roommates and posing a threat to their health. Migrants
diagnosed with TB most often preferred to leave Russia
to pursue treatment in their home country. Young female
migrants are most vulnerable. Besides the pressure of
limited resources and discrimination they also experience
double stigma both from the host society and from their
community.
Conclusion: Migrants with TB lack prompt access to
treatment while working in Russia. Their health condition is greatly endangered as well as they pose a serious
threat to their roommates, co-workers and all who are
around them. The study provides recommendations for
city health authorities with regards to addressing the
needs of migrants living and working in Moscow
including the health rights of migrant TB patients.
1) World Health Organization Country Data 2013. www.who.int/tb/

data. Accessed 22 January 2015.
PC-850-04 Utilizing existing health care

workers to find TB cases in high-risk migrant
populations in Swaziland
V Masuku,1 P A Dlamini2 1University Research South Africa,
Mbabane, 2National TB Control Program, Manzini,
Swaziland. Fax: (268) 2505 3248. e-mail:
victoriam@urc-sa.com
Background and challenges to implementation: Swaziland has highest burden of TB with an incidence of 1349/
100 000. The National Tuberculosis Strategic Plan 20152019 lacks community approaches that can increase
active case finding. The existing treatment supporters
cadre focuses only on TB DOTS and the National TB
program (NTCP) sought to investigate the potential use
of this cadre in improving active TB case finding (ACF),
including amongst high risk groups such as miners and ex
miners.
Intervention or response: In June 2014, University
Research South Africa (URSA) under the regional
Development Grant Fund worked with ex miners
associations in Swaziland to identify members to be
trained as community TB treatment supporters with an
ACF role. The training was done in collaboration with
the NTCP and covered basic TB facts, treatment
adherence and the TB screening tool. Treatment supporters were deployed to 14 communities (with hot spots
identified through ex-miners registration data) and
community entry was led by the ex-miners associations.
Health facility introductions were done by URSA at 14
health facilities identified to receive referrals for all TB
presumptive ex miners identified at household level.
Door to door activities were done focusing on TB
education, TB screening and referral for all TB presumptive for follow up diagnosis. Care givers were given a
target of 10 households each month.
Results and lessons learnt: From July to December 2014,
across 2500 households, more than 6000 ex miners were
reached with TB education, 5013 screened for TB, 307
found to be TB presumptive and 56 bacteriologically
confirmed and initiated on TB treatment, and 148 were
already on TB treatment. As a result, the NTCP is now
introducing community TB screening as another function
for the MoH supported treatment supporters.
Conclusions and key recommendations: The pilot among
ex-miners demonstrated that community cadres can be
utilized to improve national TB ACF efforts.
PC-851-04 Improving access to TB care services

with community extension workers in tribal
populations in Central India
A Vyas,1 V Jain,1 C Andrew,2 C Jacob,2 S Sahu2 1Asha Kalp,
Gwalior, India; 2Stop TB Partnership, Geneva, Switzerland.
e-mail: andrewc@stoptb.org
Background: Indias tribal and indigenous people often

have poor access to TB care services because they live in
isolated and insular communities; the Saharia tribes of

Madhya Pradesh province are no exception. Many of
their villages are only accessible by foot and are more
than 40km away from the closest government health
facility. As a result of these barriers, TB prevalence in
Saharia communities is thought to be many fold higher
than Indias national prevalence estimate and few TB
patients in these communities receive the care they need.
Design/Methods: A case finding initiative was set up
which employed community health workers (CHWs) to
act as an extension of the RNTCP in Saharia communities. CHWs verbally screened the tribes people for
symptoms of TB and symptomatic individuals were
either referred to the closest microscopy centre or a
sputum specimen was collected in the community and
transported by CHWs to the laboratory. Smear-positive
patients were started on treatment in the community.
CHWs were initially given a fixed salary, but salaries
were later revised downward and an incentive of INR
225 (USD 3.5) was provided for each patient started on
treatment.
Results: 25 679 individuals were verbally screened over 6
months, resulting in the detection of 4744 suspected TB
patients. 913 (19.2%) individuals were simply referred to
the microscopy centre for further evaluation. This
strategy was quickly stopped as it became clear zero
referred individuals were tested. 3831 (80.8%) people
were asked to provide a sputum sample in the
community. 2574 sputum specimens were transported
and tested at the nearest laboratory, resulting in the
detection of 454 (17.6%) smear-positive TB patients.
447 (98.4%) of these patients were then started on
treatment in the community by CHWs. These activities
resulted in a 89.4% increase in the number of people
treated compared with the year prior, indicating that the
patients detected by this initiative would likely not have
been treated without CHWs.
Conclusion: This initiative shows that large numbers of
previously missed TB patients can be detected and
treated using CHWs as an extension of existing health
facilities to target tribal populations. We must move past
the clinic-based model of care and strengthen community
engagement to achieve our ambitious post-2015 targets.
Community interventions which show impact, such as
this one, should be prioritized for scale-up and replication in other tribal populations.
S223
19. TB in children: epidemiology - I

PC-852-04 Tuberculosis in children in Brazil,
2014
D Gomes Dellorti,1 A Torrens,1 F. Dockhorn,1 M S
Evangelista,1 R Ruy De Souza,1 N M Saita,1 F M Reis,1 D
Barreira1 1National Tuberculosis Programme Brazil, Brasilia,
Distrito Federal, Brazil. e-mail: danielle.dellorti@saude.gov.br
Background: The magnitude of childhood tuberculosis

(TB) has not been well established yet. It is estimated
that, each year, more than half a million TB cases affect
children around the world. Consequently, the disease is
an important cause of childhood morbidity and mortality
in endemic countries, which makes it a public health
problem. The objective of this study is to describe
tuberculosis cases in children in Brazil during 2014
Design/Methods: A cross-sectional study was developed
of new TB cases in children of up to 14 years of age in
Brazil. The data were obtained from the National
Notification System, based on the tuberculosis cases
notified to the National Tuberculosis Control Program
(NTP). The clinical and demographic characteristics of
tuberculosis were analyzed, as well as the case outcomes.
To compile the data, the software Microsoft Excelw was
used
Results: In 2014, 67 885 new cases of tuberculosis were
registered in Brazil, 3.3% of which were children (2263
cases). The highest incidence occurred in the age group
from 10 to 14 years (39.9%) and the lowest in children
under one year of age (17.8%). Most children were male
(52.8%) and mulatto (46.8%). Concerning the clinical
form, 76.0% of the cases were pulmonary and 24.0%
extrapulmonary. When considering the total number of
pulmonary and extrapulmonary cases per age group,
most TB cases affected the age group from 10 to 14 years,
with 39.8% and 40.6%, respectively. In addition, an
increase was observed in the number of pulmonary TB
cases between the ages of less than one year and the
group from 10 till 14 years. Most notified cases had a
favorable outcome, with 84.7% of cure. Treatment
abandonment corresponded to 6.1% and death due to
tuberculosis to 1.9%
Conclusion: Different strategies need to be considered to
strengthen the disease control and surveillance actions in
this vulnerable group, mainly in children under one year
of age and in the age group between 10 and 14 years,
which showed significant differences in the case incidence rates. Although the outcome of the cure indicator
was favorable, it should be highlighted that the
abandonment percentage is a source of concern for TB
control in children.
S224
PC-853-04 Lessons learned in a childrens

tuberculosis clinic
A Rossoni,1,2 C Oliveira Rodrigues,1 T Tannus Tahan,1
B Gabardo,3 M Rossoni,2 A Ruffino-Netto4 1Federal
University of Parana, Curitiba, Parana, 2State University of
Ponta Grossa, Ponta Grossa, Parana, 3State Department of
Health of Parana, Curitiba, Parana, 4University of Sao Paulo,
Ribeirao Preto, SP, Brazil. Fax: (55) 41 3339 1002. e-mail:
dearossoni@gmail.com
Background: The diagnosis of tuberculosis (TB) in

children is difficult to perform and all efforts to improve
it should be encouraged.
Design/Methods: This is a descriptive study, which
analyzed data of children treated at an outpatient
references for children with suspected of tuberculosis in
south of Brazil. We followed up children from 0 to 14
years, from January 2005 to July 2010. The Research
Ethics Committee of the hospital approved this study.
Results: In this period in the study 186 children were
included: 30 cases (16.1%) were classified as TB and 127
(68.3%) with latent infection and 29 (15.6%) with other
diagnoses; 50.5% were male, 81.1% white and median
age was 5.8 years. Regarding the demographic characteristic there was no difference between the children sick
and not sick. In the TB group 85% showed history of
contact with tuberculosis case (especially parents), with a
higher frequency of multiple contacts (P 0.02); 85.7%
had signs or symptoms and, when present, all of them
were statistically more frequent in this group. But when
observing the reason for referral the child, only 19%
were due to clinical symptoms. Among all children
directed by clinical symptoms, after detailed analysis of
the case and in some cases performed therapeutic
measures nontuberculous, only 30% remained as symptomatic (found lower in the TB group P , 0.01). Among
the tests performed 41.5% of children who had
abnormal X-ray, with repetition of the exam only 1
remained abnormal; 23.2% of TST need to be repeated
for being initially non-reactors, of which 14% turned
positive.
Conclusion: The main reason patients referral was the
history of contact with tuberculosis, thus concluded the
importance of active search of cases. As well, the
importance of judicious analysis of these children, by
trained professionals, because a lot of tests or clinical
symptoms may be erroneously interpreted as suggestive
of TB, when in reality there arent; being necessary,
sometimes, the repetition of the exams and previous
therapies before the specific drugs for TB.
PC-854-04 Risk factors for unfavorable

treatment outcomes in DOTS-treated pediatric
tuberculosis patients
S Shelke,1 R Bollinger,2 A Chandanwale,1 A Deluca,3
N Gupte,2 P Adhav,1 D Kadam,1 A Gupta2 1B.J.Medical
College, Pune, India; 2Johns Hopkins University School of
Medicine, Baltomore, MD, 3Johns Hopkins Bloomberg School
of Public Health, Baltimore, MD, USA. e-mail:
drshelke@rediffmail.com
Introduction: Indian children with tuberculosis (TB)

treated with DOTS under programmatic conditions have
1020% unfavorable treatment outcomes. Ascertaining
risk factors for poor outcomes will help authorities take
corrective actions to improve the program.
Methods: We conducted a retrospective analysis of
children with TB diagnosed and treated in the year
2013, in 6 TB Units of Pune city, India. Unfavorable
treatment outcome was defined as failure to respond
clinically, death or loss to follow-up. Categorical
variables were compared by Fishers exact test. Treatment completion rates, adjusted odds ratio (AOR) with
confidence intervals (CI), univariate and multivariate
logistic regression analysis was done. We included age,
sex, clinical category, type of TB treatment, HIV
infection and type of TB in the model.
Results: A total of 3645 TB patients were registered in
2013, under the Revised National TB Control Programme (RNTCP) in Pune city. Of these, 212 (5.82%)
were children and 125 (59%) of these were female.
Pulmonary TB was diagnosed in 110 (50.5%) and 102
(46.8%) were diagnosed as extra pulmonary TB (EPTB).
Among the EPTB 71 (69.6%) had lymphadenitis. Twenty
(9.4%) of the children were HIV-infected. Twenty
(9.4%) of children had an unfavorable treatment
outcome .Of these 8 were loss to follow up, 9 died and
2 failed to respond clinically. WHO Category II TB
patients were at greater risk for unfavorable treatment
outcome, compared to category I newly diagnosed
patients (AOR8.4, CI: 2.1-12.4, P , 0.001). Compared
to uninfected children, HIV-co-infected children had a
greater risk of unfavorable outcome (AOR 3.1, CI:1.2
3.3, P 0.045).
Conclusions: The study illustrates HIV children and
Category II patients had increased odds of unfavorable
outcomes and these patients may benefit from more
stringent frequent monitoring and counselling. This
indirectly accentuates that patients in Category I should
also be thoroughly encouraged and supervised so that we
have very few Category II patients. New guidelines by
RNTCP recommends daily regimen to HIV pts. As an
extension of this policy, we recommend daily observation
of category II patients, as they may benefit from the
same. Further studies need to be done to understand
whether daily treatment for Cat II and HIV children
reduces unfavorable outcomes.
PC-855-04 The epidemiology of childhood

tuberculosis in Japan
S Yoshimatsu,1 A Ohkado,1,2 K Uchimura,1 K Izumi,1,2
L Kawatsu,1 K Ito1 1Research Institute of Tuberculosis (RIT)/
Japan Anti-Tuberculosis Association (JATA), Kiyose, Tokyo,
2
Biomedical Sciences, Nagasaki University Graduate School,
Nagasaki, Japan. e-mail: yoshoji@jata.or.jp
Background: Annual numbers of newly notified tuberculosis (TB) cases and its rate have been decreasing in
Japan, but the notification remains high (16.1/100 000
population in 2013) because of the high notification
among the elderly. Although TB is prevalent in the
society, childhood TB cases have become very rare in
Japan. We report the current epidemiological trend of
childhood TB in Japan to identify its uniqueness and the
factors relating to childhood TB control.
Design/Methods: The childhood TB data from 1987 to
2013, derived from the data posted on a website of the
Tuberculosis Surveillance Centre, the Research Institute
of Tuberculosis and data from a statistical yearbook on
TB in Japan and other publicly available, were analyzed.
Results: The annual number of notified childhood TB

patients has been less than 100 since 2006 and
notification rate has been less than 1.00 per 100 000
child population since 2005 (Figure).In 2013, there were
66 newly notified childhood tuberculosis patients aged
014 years, corresponding to a notification rate of 0.40
per 100 000 child population, which were 957 and 2.9 in
1987, respectively. The notification rate of childhood TB
in Japan is one of the lowest in the world in spite of high
notification rate among the elderly. Only one or no cases
of TB meningitis had been found in children since 2006,
though two cases were reported in 2013. In 2009, 4
S225
miliary TB were reported; only one or two miliary cases

have been found in children since 2006. Regarding the
modes of detection, almost half of the childhood cases
were found by household contact investigation and one
third were identified at medical institutions in the recent
years. BCG coverage among TB patients aged 0 to 5
months was extremely low (0 50 %), because current
vaccination policy in Japan recommends not to give BCG
for infants less than 3 months old. This is to avoid BCG
vaccination for immunodeficient infants before their
diagnosis and its fatal side effects. BCG coverage among
TB patients aged more than 6 months was above 90 %.
Conclusion: Japan is a still middle burden country of TB
as a nation, but achieved very low notification rate
among children. This may partially be attributable to
BCG vaccination, along with contact investigation based
on routine monitoring and evaluation.
PC-856-04 Treatment outcomes among children

with tuberculosis in Malawi
B G Belaineh,1,2 I Dambe,1 K Mbendera,1 L Mlauzi,1
J Mpunga1 1National Tuberculosis Programme, Malawi,
Lilongwe, 2International Training and Education Center for
Health ( ITECH) -Malawi, Lilongwe, Malawi. e-mail:
belaineh@gmail.com
Background: Children account for more than 10 % of

notified TB cases in Malawi. Childhood Tuberculosis is
also one of the thematic areas addressed in the Malawian
national strategic plan 2015 to 2020. As part of the
national strategy, NTP focus on case finding mainly using
contact investigation, IPT provision and provision of
firrst line Anti TB treatment. The treatment outcome of
children with TB has not been monitored and reported
regularly. Hence this analysis was done to inform
performance and indetify gaps to improve performance.
Methods: NTP has revised the national TB-HIV supervision tool to allow collection compilation and review of
treatment monitoring and follow up among children. As
part of the first cycle of supervision, 109 health facilities
were covered. The treatment outcome module covered
treatment outcome by category, diseseas classification
and age catetory . Treatment outcome was assessed for
patients enrolled during Oct- Dec 2013. The source of
data was TB unit register. Comparison was made among
children aged 0-4 yrs , 5-14 years and all age groups.
Results: During the period under review, a total of 347
children were evaluated for their treatment outcome. Of
these 133 were children (0-4 years old). The treatment
success rate was higher among children (5-14 yrs).
Children with , 5 yrs tend have higher risk of death
than chlidren with 5-14 years old. The unfavorable
treatment outcome among 0-4 years were death
(10.52%) and lost to follow up (6.79%) and these were
higher than those children(5-14 years).
Conclusion: More than 80% of children with tuberculosis have successfully completed treatment. Younger
children have increased risk of having unfavorable
outcome. The risk of death was 3 times higher in 0-4
years than older children. They also had increased risk of
being lost to follow up than older children. Specific cause
S226
of death among young children needs to be explored.

Prevention of new tuberculosis infection and early
diagnosis and treatment initiation among these populations need to be strengthened through strengthening of
TB contact investigation and IPT provision.
Item
Number in the cohorts, n
Treatment success rate, %
Death rate, %
Lost to follow-up, %
Not evaluated, %
04 years
514 years
All age
groups
133
81.9
10.5
6.79
6.76
214
86
3.72
3.25
6.5
3748
79.88
9.52
6.96
2.88
TB among children may serve as a proxy indicator of

Tuberculosis transmission in the community. NTP is
planning to improve capacity of helath workers in
dignsosis and management of children. The contact
investigation and expansion of IPT provision will also be
strengthened.
PC-857-04 Epidemiology of childhood

tuberculosis in Malawi: what more can we do?
J Mpunga,1 I Dambe,1 L Vito,1 K Mbendera,1 F Kassa,1
B G Belaineh1,2 1National Tuberculosis Control Programme,
Malawi, Lilongwe, 2International Training and Education
Center for Health (ITECH) -Malawi, Lilongwe, Malawi. e-mail:
belaineh@gmail.com
Background: Malawi is one of the high TB and HIV

burden countries. The latest national TB prevalence
survey revealed that Tuberculosis is 2 times higher than
the current estimate. However, the national prevalence
survey doesnt provide estimate of the magnitude of
Tuberculosis among children. Addressing childhood TB
is a high priority agenda for the Malawian tuberculosis
control program, Hence, analysis is the notification data
is vital to understand TB epidemic and come up with
appropriate measures.
Methods: A descriptive analysis was done using routine
TB surviellance data for period of January 2014 to
December 2014. All TB cases are diagnosed according to
national TB diagnositic algorithms. AFB microscopy,
GeneXpert, clinical assessment, Chest X-ray and FNA
are the main diagnostic tools that assist TB diagnosis.
The data was summarized by age group, disease category
and district. Percentage of children among notified cases
was determined for each district.
Results: Of 17 353 TB cases notified in one-year period,
childhood tuberculsosis constituted 10.65% of all cases.
There was a wide variation in the proportion of
childhood Tuberculosis cases across districts. The highest
proportion of childhood TB was noted in Lilongwe
district (19%) followed by Zomba (17%) and the least
was noted in Chiradzulu (4.5%). Children 0-4%
contributed the least among all age group (27/100 000
vs. all age group (104/100 000). Among children (0-4
yrs), smear negative pulmonary TB cases dominate
(69.8%). And the highest proportion EPTB (extrapulmonary TB case) (37.9%) cases were reported among
children age 5-14 yrs.
Conclusion: Children contribute to 10.69% of the
notified TB cases. Notification doesnt seem to correlate
with overall notifification rate. The variation among
districts seems to be related to the capacity to diagnosis
and compliance to national algorithm. With a strong
surviellance and access to service the notificaton rate of
PC-858-04 3-year analysis (20122014) of

childhood TB notification in Akwa Ibom State,
Nigeria: implications for childhood TB scale-up
in a resource-limited area
A F Omoniyi,1 A Awe,1 V Obot,2 G Akang,3 J Kuye,3
R Eneogu,3 M Gidado,4 E Oyama1 1World Health
Organization Country Office, Abuja, 2Akwa Ibom State TBL
Control Programme, Uyo, 3NTBLCP, Abuja, 4KNCV, Abuja,
Nigeria. e-mail: omoniyifadare@yahoo.com
Background: Akwa Ibom state has a population of about

5.9 million, a higher percentage of whom belongs to the
relatively young age group (age 0-9 years). TB services
are provided in about 150 health facilities across all the
31 LGAs in the state. The state is considered as one of the
6 high TB-HIV burden states in Nigeria. There has been
an increased in the proportion of childhood (0-14years)
TB cases notified in the state from 5% in 2012 to 7% in
2014, with some LGAs reporting as high as 18%. The
aim of this study is therefore to review the childhood TB
notifications in each of the 31 LGAs in the state with the
aim of providing evidence based information to enhance
childhood TB programme in resource limited settings.
Design/Methods: A retrospective review of childhood TB
notifications reports from all the DOTS centers in the 31
LGAs in the state from 2012-2014.
Results: There was an increased in the proportion of
childhood (0-14 years) TB cases notified in the state from
5% in 2012 to 7% in 2014, the proportion of childhood
TB cases notified range from 0% to 15% in 2012, from
0% to 12.1% in 2013 and from 0% to 18.2% in 2014.
10% (3) of the LGAs did not report any cases of
childhood TB in 2012, 32% (10) did not report in 2013
and 23% (7) did not report in 2014. The LGA with the
tertiary institution recorded the highest number of
childhood TB cases in the state while the rural LGAs

recorded the lowest number of cases. LGAs with 100%
smear positivity rate among notified cases in 2014
reported no case of childhood TB, this is because the
diagnosis of TB still largely depends on medical officers.
The implication of this is that most childhood TB are still
being missed in the state.
Conclusion: Notification of childhood TB cases is higher
in urban LGAs and LGA with tertiary facility where a
high number of medical officers are located. The TB
programme must therefore urgently task shift diagnosis
of TB among children to Nurses and other cadre of
health workers for the programme to be able to rapidly
increase childhood TB notification, the use of rapid
diagnostic test tool for TB diagnosis among children
must be scale up in resource limited settings to address
the low childhood TB case finding.
PC-859-04 Assessment of management of

tuberculosis among HIV-infected vs. noninfected children at Kalembe-Lembe Pediatric
Hospital, Kinshasa, R D Congo
P Lelo,1 C Akele,1 A Shoma,2 D Muteteke,2 G Bakaswa2
1
Kalembelembe Pediatric Hospital, Kinshasa, 2Programme
National de Lutte contre la Tuberculose, Kinshasa,
Democratic Republic of the Congo. e-mail:
fkitetele@hotmail.com
Background and challenges to implementation: To assess

the manadgement (clinical profile, evolution and issue)
of tuberculosis (TB) treatment among HIV infected vs.
uninfected children.
Intervention or response: Retrospective chart review of
children treated for TB at Kalembe Lembe Pediatric
Hospital in Kinshasa, Democratic Republic of Congo,
between January 2008 and December 2012. Data was
collected on use of status HIV, TB presentation and
outcome of treatment.
Results and lessons learnt: During the study period, 1009
children were treated for TB. Median age was 4.33 years
(range 0.08 to 15); 46.6% were female. Screening HIV
test was realized to 809 (80.2%) children after parents
consent. 441 (54.5%) of them had a positive serological
result for HIV. In clinical presentation, among HIV
infected children treated for TB, 252 (57.1%) were aged
less than 5 years, 106 (24.0%) were between 6 - 10 years
and 83 (18.9%) were between 11 and 15 years. Among
them, 52 (11.6%) children had extra-pulmonary TB, 5
(1.1%) had a positive sputum, 28 (6.3%) died during the
TB treatment whom 17 (60.7%) aged between 0.08 and
5 years and 18 (4%) were loss to follow up. Compared to
HIV uninfected children treated for TB, the profile of age
was nearly similarly (76.4%, 16.3% and 7.3%), no
sputum positive and the rate of loss to follow up was
similarly (5.2% vs. 4.0%).However the cases of extrapulmonary TB was higher (19. 9% vs. 11.6%), No
deaths in those non-infected by HIV.
Conclusions and key recommendations: In our population, HIV infection is one of considerable causes of
morbidity and mortality for children coinfected by TB.
S227
To promote HIV testing of children treated for TB and

screening of TB of HIV children infected should be an
integral component of the care of children in TB endemic
areas.
20. TB in children: diagnosis

PC-860-04 The diagnosis of tuberculosis in
children in the province of South Kivu,
Democratic Republic of Congo
Y Lufungulo Bahati,1,2 F Zech,1 E Musafiri,4
D Kalumuna,3 G Bakaswa,3 A Murhula,2
D Van Der Linden,1,4 E Andre1,4 1Cliniques Universitaires
Saint-Luc, Brussels, Belgium; 2Universite Catholique de
Bukavu, Bukavu, 3Programme National de Lutte contre la
Tuberculose, Kinshasa, Democratic Republic of the Congo;
4
Universite Catholique de Louvain, Brussels, Belgium. e-mail:
emmanuel.andre@uclouvain.be
Background and challenges to implementation: Tuberculosis (TB) in children is particularly difficult to

diagnose. In Africa, while the paediatric population
often represents the majority of the population and has
the same degree of exposure as adults, the infection is
often under-diagnosed. The aim of our work is to analyse
the epidemiology of paediatric TB and the determinants
of diagnosis in the province of South Kivu of DRC.
Intervention or response: We carried out our analysis on
TB detection data from the National TB Program
covering a period of 12 calendar quarters from 2010 to
2012. We also evaluated the impact of interventions
specifically targeting detection of difficult forms of TB.
The changes over the period covered by the study were
examined using quasi-least squares (generalised linear
regression). We chose the hypothesis of an autoregressive
between-year correlation matrix. The cases in each area
had a gamma-Poisson distribution with a risk correction
depending on the population size of each health area.
The percentages followed a beta-binomial distribution.
Results and lessons learnt: During this period, the
incidence of TB per 100 000 inhabitants in the general
population showed no significant change. A significant
fall was, however, seen during the period in the number
of notified cases of TB in the population under 15 years
of age, with the incidence falling from 14.7 to 10.4
during the period being studied. The proportion of
paediatric TB varies significantly between urban and
rural health areas, with predominantly urban health
areas (Bukavu, Uvira and Kamituga) showing the highest
proportion of paediatric cases. In the rural areas, the
health areas supported by a non-governmental organisation and those in which Xpert MTB/RIF is available
had the largest proportion paediatric TB, while the other
rural health areas had lower detection rates.
makes it possible to highlight the difficulties encountered
in the diagnosis of paediatric TB in rural areas in the
Democratic Republic of Congo. The presence of a higher
density of qualified healthcare workers and access to
S228
appropriate diagnostic tools are factors that positively

impact the detection rate of TB in children. We have also
shown that the performance of health areas in terms of
detection of paediatric TB was linked to their performance in terms of detection of other forms of TB that are
considered difficult to diagnose.
symptomatic TB-exposed children, with suspected intrathoracic TB disease. Model discrimination was assessed
by area under the receiver operator characteristic curve
(AUC).
Results: Overall, 150 symptomatic TB-exposed children
with suspected intrathoracic TB disease were included in
this analysis with a median age of 6 years (IQR 3 9
years), an approximately equal gender distribution and
none were HIV infected. Thirty-five (23%) were
diagnosed with active TB disease and started on TB
treatment, while 115 (77%) had other respiratory tract
infections (OD). In the final, most parsimonious clinical
prediction model, the combination of age ,5years (AOR
4.8; 95%CI 2.0 11.5) and lymphadenopathy on clinical
examination (AOR 4.9; 95%CI 1.8 13.0) discriminated
active TB from OD with an AUC of 0.70 (95%CI 0.61
0.80). Internal validation using bootstrapping technique
with 1000 repetitions gave a normal-based 95%
confidence interval of the AUC (AUC 95%CI 0.61
0.79; bias 0.014; bootstrap SE 0.048) that was similar to
those of the non-bootstrapped clinical prediction model.
Conclusion: We developed a simple and internally
validated clinical algorithm that includes age ,5 years
and lymphadenopathy on clinical examination which
reliably classified 70% of active TB from OD among TBexposed children and could be an efficient screening tool
in resource limited settings.
PC-862-04 Pediatric population: expediting the

diagnosis of Intra-thoracic, multidrug-resistant
tuberculosis
U Singh,1 S Kabra,1 R Lodha,1 H Gautam,1 Y Verma,1
R Jain,1 G Mehta,1 P Pandey1 1All India Institute of Medical
Sciences, AIIMS, New Delhi, India. e-mail:
drurvashi@gmail.com
PC-861-04 Utility of a clinical model for

predicting the risk of tuberculosis in child
contacts of tuberculosis cases in The Gambia
T Togun,1 U Egere,1 M P Gomez,1 A Sillah,1 M Daramy,1
J Sutherland,1 P Hill,2 B Kampmann1,3 1Medical Research
Council (MRC) Unit - The Gambia, Banjul, Gambia;
2
University of Otago School of Medicine, Dunedin, New
Zealand; 3Imperial College London, London UK. Fax: (220)
449 5919. e-mail: ttogun@mrc.gm
Background: It has been suggested that clinical features

may offer good prognostic value if they are well defined
and appropriate risk stratification is applied. A documented history of TB exposure in a child changes the pretest probability of disease and the positive predictive
value of subsequent investigations, therefore it provides
an appropriate parameter for risk stratification. We
established a multivariable clinical prediction model for
active TB disease in symptomatic TB-exposed children
identified from an active case-finding study in The
Gambia.
Methods: Backward stepwise logistic regression and
bootstrapping techniques were used for development
and internal validation of a clinical prediction model for
active TB in a prospective cohort of actively-traced
Background: Rapid and accurate diagnosis of Intrathoracic Tuberculosis (TB) in children is challenging due
to difficulties in obtaining good quality sputum specimen, pauci-bacillary nature of disease, low sensitivity of
smear microscopy and poor access to culture. We
compared the diagnostic accuracy of Xpert MTB/RIF
assay with microscopy and culture for diagnosis of PTB
and multi-drug resistant TB (MDR-TB) in children.
Design/Methods: Two hundred eighty-eight children
(mean age, 7.2 yrs.) with suspicion of TB were grouped
as Confirmed, Probable and No-TB based on microbiological and clinical findings. Patients were classified as
progressive pulmonary disease (PPD, parenchymal lesion/pleural effusion/ pneumothorax) and primary pulmonary complex (PPC, hilar lymphadenopathy alone) on
Chest X-ray. Gastric aspirates and induced-sputum
samples were taken on two consecutive days and
subjected to Ziehl Neelsen stain, Xpert MTB/RIF-assay
and MGIT-960 culture (reference standard) with Drug
sensitivity testing.
Results: Thirty-seven children (12.8%) were confirmed
TB, 224 (77.7%) were probable TB and 27 (9.3%) were
No-TB. Eight children (2.7%) were positive by smear, 70
(24.3%) tested positive by Xpert assay and 37 (12.8%)
by culture. Xpert- assay detected 30 of 37 culture

confirmed cases (sensitivity 81% and NPV 96%). Xpert
assay detected 62 cases more than smear microscopy and
was more sensitive (81% vs. 18%). Xpert assay detected
rifampicin resistance in additional 4 cases which were
negative on culture. Among Xpert-MTB/RIF Positive
cases, 26/30 (86.6%) had parenchymal involvement in
Confirmed TB group, and 36/40 (90%) in Probable TB
group.
Conclusion: In our study, up-front access to Xpert-MTB/
RIF testing in clinically suspected pediatric Intra-thoracic
TB cases demonstrated approximately two-fold increase
in microbiologically-confirmed PTB, and increased
detection of MDR-TB cases. Xpert proved to be more
useful in patients with parenchymal involvement. Patients with only hilar lymphadenopathy but no parenchymal lesion had fewer microbiological confirmations.
Xpert assay significantly improved case detection among
probable TB cases hence improving early detection and
treatment. Nevertheless a negative Xpert outcome does
not always rule out TB, hence culture continues to be the
gold standard. In addition, good clinical correlation is
needed for treatment initiation. Research should continue for better diagnostics.
PC-864-04 Radiological findings in young

children investigated for tuberculosis in
Mozambique
A L Garcia-Basteiro,1,2 E Lopez Varela,1,2 O Joaquim
2
P Alonso,1,2
Augusto,1 K Gondo,1 J Sacarlal,1 J Munoz,
3 1
J L Ribo Centro de Investigacao
em Saude de Manhica,
Manhica,
Mozambique; 2Barcelona Institute for Global
Health (ISGlobal), Barcelona, 3Hospital Sant Joan de Deu,
Hospitalet De Llobregat, Barcelona, Spain. e-mail:
alberto.garcia-basteiro@manhica.net
Background: Chest radiography remains a critical tool

for diagnosing intrathoracic tuberculosis (TB) in young
children who are unable to expectorate. We describe the
radiological findings in children under 3 years of age
investigated for TB in the district of Manhica, southern
Mozambique, an area with a high prevalence of TB and
HIV.
Design/Methods: Digital antero-posterior and lateral
projections were performed and reviewed by two
independent readers, using a standardized template.
Readers included a local pediatrician and a pediatric
radiologist blinded to all clinical information. International consensus case definitions for intra-thoracic TB in
children were applied.
Results: A total of 766 children were evaluated of whom
43 (5.6%) had TB. The most frequent lesion found in TB
cases was air space consolidation (65.1%), followed by
suggestive hilar lymphadenopathy (17.1%) and pleural
effusion (7.0%). Air space consolidation was significantly more common in TB cases than in non-TB cases (odds
ratio 8.9; 95%CI: 1.6-50.5), as were hilar lymphadenopathy (OR 17.2; 95%CI: 5.7-52.1). The only case
with miliary infiltrates and 3 with pleural effusions
occurred in HIV-infected children.
S229
Conclusion: Frequent air space consolidation complicates radiological distinction between TB and bacterial
pneumonia in young children, underscoring the need for
epidemiological contextualization and consideration of
all relevant signs and symptoms.
PC-866-04 Approaching a diagnostic point-ofcare test for pediatric tuberculosis through

evaluation of immune biomarkers across the
clinical disease spectrum
S Jenum,1 S Dhanasakeran,1 A Mukherjee,2 S Singh,2
R Lodha,3 S K Kabra,2 C Ritz,3 H Grewal1 1University of
Bergen, Bergen, Norway; 2All India Institute of Medical
Sciences, New Delhi, India; 3University of Copenhagen,
Copenhagen, Denmark. e-mail:
harleen.grewal@gades.uib.no
Background: The World Health Organization calls for an

accurate, rapid and simple point-of-care (POC) test for
the diagnosis of childhood TB in order to make progress
Towards Zero Deaths. A POC test based on detecting
the presence of MTB is likely to have poor sensitivity in
childhood TB as 70-80% of children having culturenegative disease. Host biomarkers reflecting the on-going
pathological processes from Mycobacterium tuberculosis
(MTB) infection throughout the spectrum towards TB
disease might hold greater promise.
Design/Methods: We analyzed transcriptional immune
biomarkers direct ex-vivo applying dual-colour Reverse
Transcriptase-Multiple Ligation-dependent Probe Amplification (dcRT-MLPA) in 88 Indian children with
probable intra-thoracic TB (40 culture confirmed, 48
culture negative) aged 6 months to 15 years and 39
healthy siblings, having a positive (n 15) or negative (n
24) for the Tuberculosis Skin Test (TST).
Results: We identified 12 biomarkers consistently associated with clinical groups upstream towards culturepositive TB on the TB disease spectrum or downstream
S230
towards a decreased likelihood of TB disease (Figure).

Differences in biomarker profiles between the groups
were assessed by hierarchical clustering and Lasso
regression analysis. Notably, the latter method identified
a biomarker signature with a very satisfactory capacity to
discriminate children with intra-thoracic TB from
healthy siblings regardless of culture result (area under
the receiver-operator characteristic curve of 88%).
Conclusion: Consistent correlation with the identified
biomarkers with MTB-related pathology, indicates a
relevance to future POC-test for pediatric TB. Furthermore, the specific biomarker signature reported could
fulfill the requirements for a POC-test, which, if
validated, would contribute significantly towards a
reduction of the worldwide TB attributable deaths in
children.
PC-867-04 Effectiveness of screening questions

used in enhancing pediatric TB detection in a
rural setting
S Saleem,1 M Jaswal,1 A Malik,1,2 F Amanullah,1
H Hussain1 1The Indus Hospital, Karachi, 2Interactive
Research & Development, Karachi, Pakistan. e-mail:
saniya.saleem@industbcontrol.org
Background: Pediatric tuberculosis comprised 10% of all

new case notification in Pakistan in 2013, which is an
underestimate. We implemented a mass screening program to enhance child TB detection in rural Sindh,
Pakistan. The objective of this study was to assess the
effectiveness of screening questions in identifying presumptive TB patients and likelihood of TB confirmation
in this cohort.
Design/Methods: A chart review was conducted of all
children ,15 years screened and identified as presumptive TB patients by trained screeners and referred to the
physician for further evaluation, at three participating
treatment centers in Jamshoro, Sindh from October 2014
to March 2015. The screening questions included
presence of cough, fever, night sweats and weight loss.
Results: We found that out of the 859 children screened
and identified as presumptive TB patients, 25.7% (221/
859) were later diagnosed with TB and started on TB
treatment. The most commonly reported symptom by the
children identified as likely TB patients was fever in 66%
(567/859) and 20% (113/567) of these children were
diagnosed with TB by the physician. Female children
were more likely to report fever .2 weeks as a presenting
symptom, 61% (61/100) compared to 43% (52/121)
male children (P , 0.01). In the 0-4 year age group 29%
female children presenting with fever were diagnosed
with TB compared to 16.7% male children (P , 0.05).
Cough .2 weeks was a more specific TB screening
question reported by 23.7% of children who were later
diagnosed with TB compared to 12% children who
reported cough ,2 weeks and were found to have TB (P
, 0.05). Also of interest was the finding that 26.8% (29/
108) children who presented with cough .2 weeks and
no fever were diagnosed with TB making it an important
stand-alone symptom.
Conclusion: In mass screening efforts to find previously

undiagnosed children with TB, symptom screens are an
invaluable tool. We found that fever and/or cough of
greater than two weeks duration were presenting
symptoms in 98.2% of children diagnosed with TB.
Fever and/or cough of two weeks or more are essential
screening questions and the presence of either symptom
in a child can be an entry point for TB evaluation in high
burden settings.
PC-868-04 Advanced TB diagnostics in a

resource-limited setting: findings from the first
two years of the pediatric TB Centre of
Excellence in Mbeya, Tanzania
J Bacha,1 J Benjamin,1 L Mwita,1 P Clowes,2 C Mangu,2
A Dinardo,3 K Ngo,3 A Mandalakas3 1Baylor College of
Medicine Childrens Foundation - Tanzania, Mbeya, 2National
Institute of Medical Research - Mbeya Medical Research
Centre (NIMR-MMRC), Mbeya, Tanzania; 3Baylor College of
Medicine, Houston, TX, USA. e-mail: bacha@bcm.edu
Background: In 2013, the Baylor Tanzania Center of

Excellence (COE) in Mbeya, NIMR-MMRC and the
Mbeya Referral Hospital created a centre providing
advanced TB diagnostics and treatment for children with
presumptive TB. The program offers sputum analysis,
tuberculin skin testing (TST), fine needle aspiration
(FNA), and chest X-ray (CXR) interpretation as well as
treatment. This abstract provides a description of robust
patient assessments, diagnostic findings and treatment
during the first 22 months of the program.
Design/Methods: Retrospective chart review at the COE
between March 2013 and December 2014. Baseline and
outcome data were captured using a standardized data
collection tool and unique data base. Diagnostic test
performance used clinical diagnosis of TB disease as a
reference standard, which incorporates both CXR and
TST results.
Results: Baseline: 504 patients with presumptive TB were
referred. 48% female (241/504); age 0.3-19.7 years
(median 4.8yr). 57% (288/504) HIV-infected, 4% (20/
504) HIV-exposed, HIV-uninfected, and 39% (196/504)
HIV-negative. Of HIV-infected, 63% (181/288) were on
ART when referred. 20% (103/504) reported a known
TB contact, and 3% (16/504) had history of prior TB
treatment. Of all evaluated, 70% (355/504) provided
sputum (42% induced, 58% spontaneous), 93% (470/
504) completed TST, 5% (26/504) FNAs, and only 56%
(281/504) had CXR results due to mechanical and
logistical challenges. Outcomes: 176 of the 504 (35%)
were diagnosed with TB disease and 98% (172/176)
initiated TB treatment. 19% (33/176) were bacteriologically-confirmed TB, 62% (109/176) probable TB, and
19% (34/176) possible TB. 14 patients were LTFU or
died (1 of which had bacteriologically-confirmed TB
results). Of those diagnosed, 86% (152/176) had
pulmonary TB, 5% (8/176) lymph node TB, and 9%
(16/176) extrapulmonary TB (EPTB). HIV-infected
accounted for 60% (91/152) of PTB, 38% (3/8) of
LNTB and 44% (7/16) EPTB. The Table lists the results
and performance of the TB diagnostics. Median time
from referral for TB diagnostics to initiation of TB

therapy was 3 days (range 0-142 days).
Conclusion: Effective pediatric TB diagnosis is possible
in our resource-constrained setting and improves TB case
finding and prompt initiation of TB treatment. However,
despite access to state of the art diagnostics, performance
of GeneXpert and culture are still sub-optimal, regardless of HIV and referral status, stressing the importance
of clinical TB diagnosis in children.
Table TB diagnostic test results and test performance for
pediatric patients with presumptive TB (with TB Disease
diagnosis used as reference standard)
Diagnostic Test
(# Performed)
All patients (n504)
Sputum GeneXpert (n355)
Sputum smear (n355)
Sputum culture (n308)
TST (n470)
FNA (n 26)
Chest x-ray (n281)
% Tests
with
positive
result
(n) PPV NPV
3% 1.00
(11/355)
4% 1.00
(13/355)
7% 0.95
(21/308)
14% 0.78
(65/470)
35% 0.89
(9/26)
31% 0.94
(87/281)
0.63
0.07
1.00
0.66
0.10
1.00
0.68
0.18
0.99
0.71
0.31
0.95
0.64
0.57
0.92
0.80
0.68
0.97
0.67
0.08
1.00
0.67
0.07
1.00
0.69
0.20
0.99
0.69
0.25
0.97
0.70
0.63
0.88
0.82
0.72
0.97
4% 1.00
(5/128)
3% 1.00
(4/129)
2% 1.00
(2/108)
9% 0.63
(19/204)
38% 1.00
(3/8)
28% 0.94
(31/111)
0,67
0.11
1.00
0.66
0.09
1.00
0.68
0.06
1.00
0.68
0.17
0.95
0.80
0.75
1.00
0.76
0.60
0.97
4% 1.00
(8/223)
3% 1.00
(7/226)
10% 0.95
(19/200)
19% 0.85
(46/245)
33% 0.83
(6/18)
33% 0.95
(56/170)
0.66
0.10
1.00
0.65
0.08
1.00
0.67
0.23
0.99
0.74
0.43
0.95
0.58
0.50
0.88
0.83
0.74
0.97
HIV-infected patients only (n288)

Sputum GeneXpert (n216) 3% 1.00
(6/216)
Sputum smear (n220)
2% 1.00
(5/220)
7% 0.93
(15/194)
TST (n250)
11% 0.85
(27/250)
FNA (n16)
38% 0.83
(6/16)
Chest x-ray (n156)
34% 0.94
(53/166)
Inpatient Referrals (n227)
Sputum smear (n129)
TST (n204)
FNA (n 8)
Chest x-ray (n111)
Outpatient Referrals (n277)
Sputum smear (n226)
TST (n245)
FNA (n18)
Chest x-ray (n170)
Sensi- Specitivity ficity
TST-Tuberculin Skin Test; FNA- Fine Needle Aspirate; PPV- Positive

Predictive Value; NPV- Negative Predictive Value
S231
PC-869-04 Barriers to childhood tuberculosis

diagnosis: a qualitative study
S Chiang,1 S Roche,2 C C Contreras,3
A Valentina Antonieta,4 H Del Castillo Barrientos,5
M Becerra,3,6,7 L Lecca3 1Baylor College of Medicine,
Houston, TX, 2Boston University School of Public Health,
Boston, MA, USA; 3Socios En Salud (Partners In Health), Lima,
4
Ministerio de Salud, Lima, 5Instituto Nacional de Salud del
Lima, Peru; 6Harvard Medical School, Boston, MA,
Nino,
7
Brigham and Womens Hospital, Boston, MA, USA. e-mail:
schiang@alumni.stanford.edu
Background: According to Perus National Tuberculosis

Program (NTP), children aged 0-14 account for 7.9% of
the countrys tuberculosis (TB) incidence. This figure
likely reflects suboptimal diagnosis of childhood TB. The
challenges of childhood TB diagnosis have been well
described by clinical studies and expert opinion. However, the perspectives of front-line healthcare providers
and affected families have been infrequently reported. An
understanding of these groups experiences is crucial to
increasing case detection. In this qualitative study, we
sought to identify obstacles to childhood TB diagnosis
faced by primary care providers and patients families in
Lima, Peru.
Design/Methods: El Agustino and La Victoria, two of
Limas 43 districts, were chosen as the setting because of
their high TB incidences: 224 per 100 000 person-years
in El Agustino and 195 per 100 000 person-years in La
Victoria. We purposefully sampled NTP providers,
community health workers, and parents of children with
TB at the eight (of 16 total) health centers in El Agustino
and La Victoria with the highest pediatric TB caseloads.
Using semi-structured guides, moderators conducted ten
focus group discussions with 53 primary care providers,
community health workers, and parents. In addition,
moderaters conducted nine in-depth interviews with
NTP administrators and pulmonologists to add a
broader perspective to the issues explored in the focus
groups. Two authors independently performed inductive
thematic analysis and identified emerging themes.
Results: Participants identified six barriers to the
diagnosis of childhood TB. These barriers were (1)
ignorance and stigma among the community; (2)
insufficient contact investigation; (3) limited access to
diagnostic tests; (4) insufficient health center staff; (5)
inefficient specialist referrals; and (6) provider shortages.
While participants mentioned the challenge of sputum
collection and the low sensitivities of smear microscopy
and culture, these issues received less attention than the
other six barriers.
Conclusion: Recent efforts to increase childhood TB
detection have centered on the development of new
technologies. However, our findings demonstrate that
many diagnostic barriers are rooted in socioeconomic
and health system problems, which must be addressed
alongside efforts to improve technological tools.
Strengthening providers ability to diagnose childhood
TB at the primary care level may be the best starting
point for intervention.
S232
Results: COPD developed in 46 adult smokers, which

corresponds to a 2-year cumulative incidence of 7.1%,
giving an average annual incidence of 3.55%. Mean
FEV1 and FVC among incident COPD cases were 1.80l
(SD 0.54) and 2.34L (SD 0.51) respectively. Both
FEV1 and FVC were significantly lower (P 0.000 and P
0.008 respectively) among the incident COPD cases.
Incident COPD is significantly associated with age,
occupation, income, respiratory symptoms and BMI on
bivariate analysis. In the adjusted model, older age (760
y: OR 4.2, 95%CI 1.58-12.5) and presence of respiratory
symptoms in last 12 months (OR 3.86, 95%CI 1.748.58) are independent risk factors for incident COPD
among the adult ever smokers.
Conclusion: COPD incidence is very high among
smokers in Bangladesh. Ever smokers at older age group
and/or presence of respiratory symptoms should be
priority individuals to reduce COPD incidence in
Bangladesh.
PC-871-04 DNA damage and cellular

abnormalities detected by micronucleus assay
in patients with COPD, tuberculosis and lung
cancer
21. COPD, lung function and pulmonary

embolism
PC-870-04 Incidence and risk factors of chronic
obstructive pulmonary disease among smokers
in Bangladesh
A Siddiquee,1 M A H Chowdhury,1 S Ahmed,1 S Pervin,1
K Hasan,1 A Naheed,1 D S Alam2 1International Center for
Diarrheal Disease Research, Bangladesh Dhaka, Bangladesh;
2
Centre for Global Health Research, Li Ka Shing Knowledge
Institute, St. Michael Hospital, Toronto, ON, Canada. e-mail:
tanweer@icddrb.org
Background: Chronic obstructive pulmonary disease

(COPD) is a major cause of morbidity and mortality
worldwide; the brunt of which is in the low- and middleincome countries. The prevalence of COPD among
Bangladeshi adults is 13.5%. Smoking is highly prevalent
in Bangladesh, especially among males. There is scarcity
of population based data on COPD incidence among
smokers. The goal of this study is to estimate incidence of
COPD and its risk factors among adult (40 years or
above) ever smokers.
Design/Methods: We conducted a population based
cohort study in urban Dhaka and rural Matlab where
652 adult smokers underwent spirometry, who had
normal lung-function detected 2 years back in an ongoing COPD prevalence study. COPD was defined
according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria. Socio-economic and
lifestyle variables included educational status, household
income, occupation, exposure to biomass fuel burning,
and smoking history. Risk factors of incident COPD
cases were identified using multiple logistic regression
model.
A Goncalves Da Silva,1 M J Bresciani,1 T E Karnopp,1

A Ferreira Weber,1 J H Ellwanger,1 J A P Henriques,2
A R Moura Valim,1 L Possuelo1 1University of Santa Cruz do
Sul - UNISC, Santa Cruz Do Sul, Rio Grande do Sul, 2Federal
University of Rio Grande do Sul - UFRGS, Porto Alegre, Rio
Grande do Sul, Brazil. Fax: (55) 145 137 171 855. e-mail:
andreag@unisc.br
Background: The inflammation plays a central role in

many lung diseases, including COPD, tuberculosis and
lung cancer. Currently, studies have related inflammation
with the increasing cancer incidence. The study of DNA
damage at the chromosome level is very important
because chromosomal mutation is an event of carcinogenesis.
Design/Methods: We investigated the DNA damage
(peripheral blood lymphocytes) in patients with chronic
obstructive pulmonary disease (COPD, n 28), lung
cancer (LC, n 18) and tuberculosis (TB, n 22) and
control individuals (n 17) by comet assay. The
cytogenetic DNA damage was evaluated by buccal
micronucleus cytome assay (BMCyt).
Results: A higher percentage of apoptotic cells was found
in patients with respiratory disease compared to control
subjects. Some significant differences were also observed
among the case groups (COPD, TB and CP) and the
control group. COPD patients had a higher percentage of
apoptotic cells and necrosis compared to the LC patients
and a higher percentage of apoptotic cells compared to
the TB group. The TB group had a higher percentage of
DNA damage compared to the COPD and LC groups
and defects in cytokinesis and apoptotic cells compared
to the COPD group. The TB group also had a higher
percentage of apoptotic cells and necrosis compared to
the LC group.
Conclusion: Our results indicated that patients with
COPD, TB and LC had a low frequency of permanent
DNA damage. Patients with COPD presented cellular

abnormalities that corresponded to death by apoptosis
and necrosis, while patients with TB presented defects in
cytokinesis and dysfunctions in DNA repair that resulted
in the formation of micronuclei.
PC-873-04 Prevalence of COPD among subjects

with symptoms or exposure to risk factors in
the city of Telavi, Georgia
M Maglakelidze,1,2,3 N Maglakelidze,2,3
T Maglakelidze,2,4 I Chkhaidze2,5 1The Black Sea Health
Alliance, Tbilisi, 2Georgian Respiratory Association, Tbilisi,
3
National Center for Disease Control and Public Health,
Tbilisi, 4LTD Diagnostic Center, Tbilisi, 5Tbilisi State Medical
University, Tbilisi, Georgia. e-mail: maka_mag@hotmail.com
Background: Chronic Obstructive Pulmonary Disease

(COPD), which is defined as chronic airflow limitation is
one of the leading causes of mortality and morbidity. The
major risk factors of COPD are smoking, outdoor and
indoor pollution. The prevalence of COPD in Georgia is
suspected to be high due to high smoking rates and air
pollution. There is no exact definition for the diagnosis of
COPD in the Georgian State Medical Standards list; the
present definition is Other Chronic Obstructive Diseases
of Lung, official prevalence rate of which is low 0.097%
Design/Methods: A quantitative cross-sectional study
was carried out in Telavi family medicine center, for
determining the prevalence of COPD and relation of its
risk factors in Telavi. The subjects over 35 years of age
who had symptoms of prolonged cough, sputum
production, or dyspnea, and/or a history of exposure to
disease risk factors were included. COPD diagnosis was
made through a questionnaire and spirometric examination using the criteria of Global Initiative for Chronic
Obstructive Lung Disease (GOLD).
Results: 101 patients were consecutively included in the
study, in the period of February 15-April 15, 2010. The
sex ratio M/F was 1.24. The mean age of study
participants was 55 years. COPD was confirmed in
35.6%. The risk of COPD in men was 4.6. times higher
than in women. Seven subjects (age range 35-39 years)
were diagnosed with COPD. They did not report to be
current or past smokers. There were no patients detected
with very severe form of COPD. The results demonstrated neither, statistically significant association between
the duration of cough and the severity of COPD, nor
statistically significant association of smoking and the
severity of COPD. A Chi-square test demonstrated that
COPD was dependent on outdoor pollution and the
dependency was statistically significant P , 0.05.
Statistically significant dependency between COPD and
indoor pollution was not found.
Conclusion: The prevalence of COPD among investigated population was 35.6%. The estimate prevalence rates
for COPD in Telavi appeared to be 1.7 times higher than
the official rates for Other Chronic Obstructive Diseases
of Lung. The results of the study can serve as baseline
data for broader studies on COPD, which should focus
on the impact of various risk factors. Specific public
S233
policy issues need to be re-examined from the angle of

environmental contribution to the burden of COPD and
definition of COPD has to be included in Georgian State
Medical Standards list.
PC-874-04 Chronic obstructive pulmonary

disease screening among elderly in an urban
community of South India using the IPAG
questionnaire
R C Chauhan,1 A J Purty1 1Pondicherry Institute of Medical
Sciences, Puducherry, India. e-mail:
rcchauhan21@gmail.com
Background and challenges to implementation: Underdiagnosis of chronic obstructive pulmonary disease

(COPD) is common as this disease is often not recognised
and diagnosed until it is moderately advanced. One of
the barriers to the diagnosis of COPD is the difficulty
with which spirometry is performed in the general
practice/community setting. International Primary Care
Airways Guidelines (IPAG) questionnaire is a feasible
validated screening tool for COPD diagnosis in community setting.
Intervention or response: This community based observational study was conducted among 1037 elderly people
in urban areas of Puducherry in India. Participants were
recruited using multistage random sampling technique.
IPAG questionnaire [Cut off Score 17 (Positive
predictive value: 92%)] was administered among all
study subjects. We compared quantitative parameters
between COPD and non-COPD subjects using the
student t-test and v2 test was applied for qualitative
variables.
Results and lessons learnt: Data from 1037 subjects (435
males) were analysed. The proportion of smokers among
male was 16.5%. A positive IPAG questionnaire for
possible COPD (717 points) was obtained in 342
(33.1%) subjects. Based on the answers to the questionnaire given by the subgroup of possible COPD subjects,
69 (22.2%) had no cough, 215 (62.9%) reported sputum
production in the absence of a cold, 186 (54.4%) cough
up phlegm (sputum) from their chest first thing in the
morning, 162 (47.4%) had frequent or occasional
wheezing and 113 (33 %) reported to have or had any
allergies. Age and body weight found to be associated
with positive IPAG questionnaire. Possible COPD was
significantly higher among smokers (52.9%) as compared to non-smokers (31.7%) (P , 0.01).
Conclusions and key recommendations: Proportion of
subjects with possible COPD was high in community.
IPAG questionnaire was an easy to administered tool for
screening of COPD and little expertise is sufficient for its
administration. IPAG questionnaire could be used for
screening and planning of health interventions in
communities in developing countries.
S234
PC-875-04 Silico-tuberculosis and associated

risk factors in Iran
A Farazi,1 M Jabbariasl2 1Arak University of Medical
Sciences, Arak, Iran. Fax: (98) 861 313 4002. e-mail:
farazialiasghar@yahoo.com
Background and challenges to implementation: Coexistence of silicosis and tuberculosis is known as

silico-tuberculosis. This article review the burden of
silicosis and tuberculosis in workers who exposed to
silica and evaluate influencing factors that may increase
the risk of silico-tuberculosis.
Intervention or response: An analytical cross-sectional
study was performed in silica exposed workers in
Markazi province of Iran during 2011-2012. Sampling
method was un-randomized and considering all workers
who at least 6 months exposed to silica. The study was
done via questionnaire, clinical examination, spirometry,
chest X-ray and tuberculosis investigations.
Results and lessons learnt: A total of 3121 workers were
included in the study, the mean age of participants was
43.1612.4 years, and mean employment duration
14.966.8 years. Prevalence of TB in silica-exposed
workers without silicosis was 172 cases per 100 000
people and prevalence in silicosis cases was 917 cases per
100 000 people. Incidence of TB in silica-exposed
workers without silicosis was 69 cases per 100 000
people and incidence in silicosis cases was 459 cases per
100 000 people. In the current study, frequency of LTBI/
TB was significantly higher in subjects with silicosis
workers compared to the non-silicosis group (P 0.005).
Prevalence of TB in silica-exposed workers without
silicosis was 5.2 fold and prevalence in silicosis workers
was 27.8 fold higher than general population. Incidence
of TB in silica-exposed workers without silicosis was 3.3
fold and incidence in silicosis cases was 21.8 fold higher
than general population. The frequency of LTBI/TB was
higher in age over thirty years old (P 0.02), in workers
with employment duration over 10 years (P 0.004), in
workers with exposure duration over 5 years (P 0.03)
and smokers with over 5 pack-years (P 0.01).
Conclusions and key recommendations: Exposure to
silica causes a renewed multiplication of bacilli in the
healing TB lesions. Prevalence of pulmonary tuberculosis
in Silicosis is more common when compared to
prevalence in general population, hence all should use
prophylactic measures intensification of work place.
PC-876-04 Validation of the American Thoracic

Society respiratory questionnaire for lung
function assessment among an occupational
group of textile workers
T Jamali,1 D A Nafees1 1Aga Khan University, Karachi,
Pakistan. e-mail: tanzil.jamali@aku.edu
Background: The objective of this study was to

determine the correlation of spirometric lung pattern
with respiratory symptoms and to validate the American
Thoracic Society (ATS-DLD-78A) respiratory questionnaire for lung function assessment among an occupational group of textile workers.
Design/Methods: This was a cross-sectional survey

conducted in 2009-2010 which included 372 workers
from 15 textile mills in Karachi. Data was collected
through the ATS-DLD-78A respiratory questionnaire
and lung function was assessed by using a portable
spirometer.
Results: A general trend of decrease in lung function
(predicted FVC, FEV1 and FEV1/FVC ratio) was
observed among participants who had chronic respiratory symptoms. Predicted FEV1 was significantly reduced for chronic cough (aOR: 3.09, CI: 1.26, 7.56),
chronic wheeze (aOR 1.98, CI: 1.05, 3.71) and shortness
of breath grade 2 (aOR: 2.07, CI: 1.05, 4.07). The risk of
decrements in lung function significantly increases with
the addition of chronic symptoms on predicted FEV1;
cough wheeze (aOR: 2.08 CI: 1.05, 4.10), cough
shortness of breath grade 2 (aOR: 2.47, CI: 1.18, 5.18),
phlegm shortness of breath grade 2 (aOR: 2.59, CI:
1.23, 5.43) cough wheeze shortness of breath grade 2
(aOR: 4.64, CI: 1.97, 10.93) (aOR: 4.18, CI: 1.68,
10.37). We found low sensitivity and high specificity and
negative predicted values for the respiratory symptoms.
Conclusion: A combination of chronic respiratory
symptoms was best associated with decrements in lung
function. These findings have relevance for developing
evidence-based approaches for the respiratory health
surveillance among textile workers.
PC-877-04 Prevalence of chronic obstructive

pulmonary disease among patients previously
treated for pulmonary tuberculosis in Manila: a
multi-center study
C Gulay,1 H Delfin,1 J Benedicto,1 L Fernandez,1
R Pagcatipunan, Jr1 1Philippine General Hospital, Manila,
Philippines. e-mail: coleengulay@gmail.com
Background: Chronic obstructive pulmonary disease

(COPD) has been a major cause of death worldwide,
ranking as the fourth leading cause of death worldwide.
In several epidemiologic studies, pulmonary tuberculosis
(PTB) is currently an emerging potential risk factor in the
pathogenesis and severity of COPD. This study aims to
determine the prevalence of COPD among patients who
were previously treated for PTB in various Directly
Observed Treatment Short-Course (DOTS) Centers in

Manila.
Design/Methods: We performed a prospective crosssectional study using purposive randomized sampling
among patients managed in different DOTS centers in
Manila. Participants were aged . 40 years and completed a questionnaire to assess their demographic profile,
exposure to possible risk factors for COPD, presence of
other confounding factors which may present with
obstructive changes, and TB-treatment related variables.
A diagnosis of COPD was based on the post bronchodilator FEV1/FVC , 70 as recommended in the Global
Initiative for Chronic Obstructive Lung Disease
(GOLD). Pulmonary function tests among matched
population without history of PTB were also gathered
to compare prevalence of COPD among these patients
against those with history of PTB.
Results: Among 224 participants who were previously
treated for pulmonary tuberculosis, 20.47% met the
criteria for COPD with an odds ratio of 1.10 (95%CI
0.68, 1.76). This is compared with 19.38% who met the
criteria for COPD among 227 participants without
history of PTB. This may imply that the risk for COPD
is around 10% higher for patients previously treated
with PTB. In this study, majority of the patients
diagnosed with COPD belong to GOLD Stage III.
Conclusion: The prevalence of COPD in patients
previously treated for PTB in this study is not statistically
significant. Nevertheless, our data suggest that previous
history of PTB posed a risk for development of
obstructive airways disease that warrants screening to
provide appropriate intervention and maximized therapy.
PC-878-04 Activity of blood phagocytes in

sarcoidosis and tuberculosis: phagocytic
activity of neutrophils as simple test for
differentiation both diseases?
A Dubaniewicz1 1Medical University of Gdansk, Gdansk,
Poland. Fax: (48) 583 491 513. e-mail:
aduban@gumed.edu.pl
Background: Clinical and histopathological similarities

between sarcoidosis (SA) and tuberculosis (TB) suggest a
participation of Mycobacterium tuberculosis in the
etiopathogenesis of SA. Because these resemblances
may be cause of a dangerous misdiagnosis both diseases,
there are searches for new biomarkers useful for the
discrimination of sarcoidosis from tuberculosis. We
recently revealed that genetically different (HLA and
non-HLA, e.g. NRAMP1 and FCGR) patients with SA
and TB induce dissimilar immune responses to the same
mycobacterial heat shock proteins, which are implicated
in forming of immune complexes (CIs). The complexemia in both diseases may be a result of an alteration in
the abilities of monocytes and/or neutrophils to phagocytosis/clearance antigen(s)/CIs with following persistent
antigenemia and granuloma formation.
Design/Methods: Therefore we evaluated phagocytic
activity of blood monocytes and neutrophils in 25 SA
S235
patients, 25 TB patients, and 20 healthy volunteers using

the PHAGOTESTw kit by flow cytometry.
Results: We revealed that the total number of white
blood cells was significantly higher in TB than in SA and
healthy controls (P 0.003, P 0.02, respectively),
whereas there was no differences between SA and
controls. The total number of blood monocytes was
higher in SA than in TB and controls (P 0.002, P
0.002, respectively) and also in TB compared to controls
P 0.04). We revealed increased of phagocytic activity of
monocytes in SA than healthy individuals (P 0.03) and
insignificantly than in TB, whereas there was no
differences between TB and healthy controls. We also
revealed higher total number of neutrophils in TB than in
SA and controls (P 0.0004, P 0.01, respectively),
whereas there was no differences between SA and healthy
individuals. Phagocytic activity of neutrophils in TB was
higher than in SA and controls (P 0.004, P 0.01,
respectively), whereas there was no differences between
SA and healthy controls.
Conclusion: Our results revealed increased phagocytic
activity of monocytes in both disease, but in contrast to
SA, increased phagocytic activity of neutrophils was
detected in TB. Current study suggest different role of
neutrophils in the etiopathogenesis of both diseases.
Analysis of phagocytic activity of blood neutrophils may
be useful for a differentiation of SA from TB. Funding:
grant 5160/B/P01/2010/39.
PC-879-04 Determinants of pulmonary

embolism in Kinshasa hospitals
A Bakebe Mbaku,1 M Kashongwe Murhula,1
M Tshiasuma,1 C Mulenga,1 S Bisuta Fueza,1
B Kabengele,1 Z Kashongwe Munogolo,1
J M Kayembe Ntumba1 1Kinshasa university hospital,
Kinshasa, Democratic Republic of the Congo. e-mail:
alainbakebe@yahoo.fr
Background: Pulmonary embolism (PE) is no longer a

curiosity in our environment; this reality could be
explained by the increasingly higher frequency of its
predisposing factors.
Design/Methods: Multicentric cross-sectional study undertaken between January 2011 and April 2014. Patients
aged 18 years or more admitted for acute dyspnea and /
or acute chest pain suspected of respiratory origin were
eligible. The diagnosis of PE was based on computedtomographic pulmonary angiography or on a set of
arguments involving clinical, echocardiographic, lower
limb compression venous ultrasonographic features. The
v2 test of Pearson and Wald with 95%CI were used in the
statistical analysis. Significance was accepted at P ,
0.05.
Results: 158 patients were included in our study. Male
gender was predominant with a sex ratio of 1.2. The
mean age of patients was 54.8 6 14.8 years. The main
risk factors including risk surgery (P 0.018), immobilization (P , 0.001), obesity (P , 0.001), diabetes
mellitus (P , 0.001) and heart disease (P 0.014).
Immobilization (OR: 8.33; P , 0.001) and obesity (OR:
S236
8.06; P , 0.001) were found to be major determinants of

the EP in multivariate logistic regression.
Conclusion: The EP is common in our environment. The
emergence of certain factors related to the epidemiologic
transition in black Africa might partly explain the
increased prevalence of venous thromboembolic disease.
22. Smoking cessation

PC-880-04 A European Respiratory Society
Consensus Document to help smokers with
pulmonary disorders to quit
C Vardavas,1 C Gratziou,2 C Jimenez-Ruiz,3
J I De Granda 4 1University of Crete, Rethymnon, 2Athens
de
University, Athens, Greece; 3Sociedad Madrilena
Neumologa y Ciruga Toracica, Madrid, 4Hopital 12 de
Octubre, Madrid, Spain. e-mail: maeve.barry@ersnet.org
Background and challenges to implementation: It is

known that tobacco smoking is the major etiological
factor for the development of COPD and about 15-20%
of all smokers will develop COPD. On the other hand,
smoking cessation is the only therapeutic intervention
that can avoid chronic progression of COPD. Its also
known that smoking cessation reduces the annual
decrease of FEV1, improves responses to bronchodilator
drugs and inhaled corticosteroids, reduces the incidence
of acute exacerbations and reduces bronchial infections.
The ERS Consensus Document to help smokers with
Pulmonary Disorders to quit is a qualitative review
regarding smoking cessation in patients with COPD and
other pulmonary disorders. It describes the epidemiological links between smoking and pulmonary disorders, the
evidence for benefits of stopping smoking, how best to
assess tobacco dependence and what interventions
currently work best to help pulmonary patients quit.
Finally, it describes characteristics and management of
smokers who finds it difficult to quit.
Intervention or response: Lung cancer, asthma and
tuberculosis are disorders that are related with tobacco
consumption. More than 85% of lung cancer is due to
smoking and around 15% of lung cancer patients still
smoke 6 months after diagnosis. Many benefits appear
after smoking cessation in patients with lung cancer. The
most relevant are as follows: reduction in surgical
complications, improvement of responses to chemotherapy and radiotherapy and increase in survival time.
Results and lessons learnt: Smoking rate among asthma
patients is similar to the general population and smoking
contributes to deter the prognosis of this disorder and
avoid the benefits of bronchodilators drugs and inhaled
corticosteroids.
Conclusions and key recommendations: Assessments of
smoking characteristics in patients with pulmonary
disorders is a crucial intervention. Assessments should
include: number of cigarettes smoked daily, number of
years of smoking, CO index, meassurement of blood
cotinine levels, assessment of motivation and self-efficacy
to quit and assessments of depression, tobacco depen-
dence and previous attempt to quit. Smoking cessation

strategies for patients with pulmonary disorders should
include: a combination of counseling and pharmacological treatment (NRT, bupropion and varenicline).
PC-881-04 Impact of smoking cessation

intervention package on treatment outcome of
tuberculosis patients in North India: a cluster
randomized control trial
S Goel,1 J Kathiresan,1 A Garg,2,3 S Raj,4 R J Singh2 1Post
Graduate Institute of Medical Education and Research,
Chandigarh, 2International Union Against Tuberculosis and
Lung Disease, South-East Asia Office, New Delhi, 3Health and
Family Welfare, Chandigarh, 4Punjab University, Chandigarh,
India. Fax: (91) 172 2744 993. e-mail:
sonugoel007@yahoo.co.in
Background: Smoking is an independent risk factor for

the development of tuberculosis and also decreases the
effectiveness of the tuberculosis treatment. However,
cessation may yield substantial positive effects on TB
treatment outcomes, relapse and future lung disease. The
Union recently called for inclusion of brief smoking
cessation ABC package in standard TB case management. The objective of the present study was to ascertain
the effect of a comprehensive smoking cessation intervention package on treatment outcomes of pulmonary
tuberculosis patients.
Methods: This cluster randomized control trial was
conducted among all pulmonary tuberculosis patients
over 15 years of age, enrolled for DOTS in two quarters
(January till June 2013), in all 17 Designated Microscopy
Centres (DMCs) of Chandigarh, located in north India.
The clusters were defined at the Designated Microscopy
Centre (DMC) level with 9 DMCs as intervention arm
and 8 as control arm. Both the arms received the
established standards of care for Pulmonary TB patients
along with DOTS treatment. The intervention arm
received the standardized smoking cessation intervention
package in addition to control arm. ABC for TB is a
package that delivers the activities of Ask, give Brief
advice and provide Cessation support. It can be delivered
by any health care worker within as little as 2-5 minutes
within existing programme. These services are delivered
at the time of registration and during sputum reexamination visits. The two arms were followed up until
the treatment completion.
Results: Total 686 tuberculosis patients were enrolled in
the study. 12.6% (n 87) were daily smoker and 46.3%
(n 318) were smokeless tobacco users. Cure rate
(68.1% vs. 62.6%), transfer out rate (4.7% vs. 3.6%)
and defaulter rates (2.6% vs. 1%) was higher in
intervention group than in control group but was not
statistically significant. Treatment failure rate was
statistically significantly higher in intervention group
(4.9%) as compared to control group (1.7%). Patients
who were willing to quit were significantly more in
intervention group (84.8%) than control group (40.5%).
Conclusions: Participants who received comprehensive
smoking cessation intervention package had a better
outcome than who received the usual T.B care. The
comprehensive smoking cessation package should be

integrated within existing Revised National Tuberculosis
Control Programme (RNTCP) of country.
PC-882-04 In-school cessation in low-income

schools in Mumbai
A Pilankar,1 D Chadha1 1Salaam Bombay Foundation,
Mumbai, India. e-mail: ajay@salaambombay.org
Background and challenges to implementation: 14.6%

of youth (age 13-15 years) use tobacco in India. Despite
the high rate of tobacco use among children and youth,
there are no cessation programmes available for youth
and the costs of nicotine replacement therapies are
prohibitively high. To curb the tobacco epidemic among
youth, cessations programmes are urgently needed.
Intervention or response: Salaam Bombay Foundation
(SBF) began an in-school cessation programme in 2012
in public schools in Mumbai for students in grades 5 to
10. The objectives of the cessation programme are to
increase quit attempts, motivate users to quit, and
prevent the transition from betel nut use to smokeless
tobacco use. The SBF cessation programme provides an
orientation for all students, and a 6- month counseling
programme for students who register. Registered students participate in 6 group sessions designed to sensitize
students to tobaccos harmful health effects, motivate
them to quit, and coach them to build refusal skills and
resist peer pressure.
Results and lessons learnt: Approximately 4000 students
have been oriented to the cessation programme. 312
(7.8%) students have registered. Most of the registered
students report betel nut consumption (n 241, 77.2%),
tobacco consumption (n 15, 4.8%) or both (n 56,
17.9%). Following six counseling sessions (6 months
after registration), students have ceased betel nut use (n
210, 87.1%), tobacco use (n 1 1, 73.3%), or both (n
41, 73.2%).
Conclusions and key recommendations: Cessation programmes for children and youth are needed in countries
where youth tobacco use is high. While MPOWER
measures advocate for cessation efforts in primary
healthcare services and through free telephone quit lines,
in- school cessation programmes should be adopted as
well. School-based cessation programmes can be effective and should be designed specifically for child and
youth users with an emphasis on: Role-playing to build
refusal skills and address peer pressure Activity-based
sessions to keep students engaged Orientation for school
administrators and teachers.
PC-883-04 Smoking cessation outcomes among

in-patients in a tertiary care hospital: learning
from LifeFirst Programme in India
P Todankar,1 V Thawal,1 H Gupte,1 L Choudhuri1
1
RESEARCH, Narotam Sekhsaria Foundation, Mumbai, India.
e-mail: priyamvada.todankar@gmail.com
Background: It is well documented that in India tobacco

usage kills about 1 million people each year which is
S237
more than TB, HIV/AIDS and malaria combined. It is

estimated that if the trends continue then tobacco will
account for 13% of all deaths in India by 2020. Taking
into consideration the gravity of the issue, Narotam
Sekhsaria Foundation (NSF), a leading grant making
organisation in India is playing a pivotal role in devising
and testing tobacco dependence treatment models in
various healthcare and non healthcare settings under its
flagship programme LifeFirst.
Design/Methods: Inpatient phase is seen as a vital
opportunity to encourage and prepare patients to enroll
in tobacco cessation programme. One of the models of
LifeFirst is implemented in a tertiary care hospital in
Mumbai for the inpatients. Training of resident doctors,
nurses and other para-medical staff was conducted by
NSF experts trained at Mayo clinic (USA). It emphasized
on screening the admitted patients for tobacco use,
providing brief advice and referral to specialist services.
The other part of treatment protocol includes understanding patient inclination towards cessation; enrolling
him/her in LifeFirst programme and proving detail
follow up counseling over a span of 6 months through
6 follow up sessions.
Results: The discussion here includes smoking cessation
among the inpatients reached till date through this
model. Total numbers of inpatients who reported any
form of tobacco use was 1443 cases. Of these, 397 (28%)
were classified as smokers. When posed with the
question on intention to join LifeFirst, 175 (45%) have
stated that they are keen however only 109 (27%) joined.
For these 109 cases, efforts were made to reach them
with 6 input follow up sessions as per the treatment
protocol. After 6 months completion, it is observed that
40% of them have received all 6 sessions and 60% have
received less than that due to patient related issues. A
comparative analysis of those who received all 6 sessions
as against those who received less than 6 sessions
indicates that the 7 day PPA at 6 months among those
input completion cases is 47% whereas among the
comparative group it is 28%. In both the groups, no
remarkable difference was observed in terms of gender,
marital status, age of initiation of tobacco use. Most of
them were admitted in the cardiology department. These
observed analyses of service data will guide to study the
variation in the outcome in a more scientific manner.
PC-884-04 National survey of smoking

cessation services in Irans primary health care
system
Z Hesami,1 M Aryanpur,1,2 M R Masjedi2 1National
Research Institute Tuberculosis and Lung Diseasese,Tobacco
Prevention and Control Research Center,Shahid Beheshti
University or Medical Sciences, Tehran, 2Iranian Anti Tobacco
Association, Tehran, Iran. Fax: (98) 2120 10484. e-mail:
zahra_hessami@yahoo.com
Background: Delivery of smoking cessation supports via

primary health care settings could be an effective way to
increase people access to cessation services. This study
was aimed at evaluating structural characteristics of
S238
smoking cessation services established within the Iranian

Primary Health care system.
Design/Methods: In order to obtain structural information about smoking cessation services, firstly a phone call
was made with coordinating authorities of tobacco
control programs in each university which are under
supervision of Ministry of Health. Secondly, after
describing the objectives of project they were asked to
fill the related questionnaire. The questionnaire was
available at MOH website and follow-up for its
completion was done via telephone call.
Results: Smoking cessation centers started their activities
in 2007 and their number increased between 2008 and
2011. In all primary health centers, smoking cessation
services are provided free of charge .In sixty percents of
centers individual therapy was used, a combination
therapy (including pharmacotherapy) is highly preferred.
Nicotine patch was the most common drug which is used
(62 %) in smoking cessation clinics. General physicians
are the main providers of smoking cessation programs in
health care centers (87%).The number of smoking
cessation centers in primary health care system decreased
during years of 2011 (89 centers) and 2012 (79 centers)
compare with 158 centers in 2010. There isnt national
quit line in Iran.
Conclusion: This study shows smoking cessation programs provided in primary health care system in Iran.
The present study gathered useful and updated information on structural of smoking cessation services in Irans
primary health care system.
PC-886-04 Cross-sectional survey on quitting

attempts among adolescent smokers in
Dharan, Eastern Nepal
P M S Pradhan1 1Patan Academy of Health Sciences,
Lalitpur, Nepal. e-mail: pranil.pradhan@gmail.com
Background: Tobacco dependence is known to be a

chronic condition and only few of those who initiate
quitting become successful. Adolescent smoking is in
rising trend in Nepal. They often want to stop smoking
and frequently attempt smoking cessation but are unable
to maintain long term abstinence mainly because they are
addicted to tobacco and experience withdrawal symptoms just like adults. This study aimed to explore the
quitting attempts among adolescent smokers in Dharan
Municipality of Eastern Nepal
Design/Methods: A cross sectional study was conducted
using pre tested self administered questionnaire adapted
from Global Youth Tobacco Survey to assess current
smokers and quitting attempts among the representative
sample of 1312 adolescent students in middle (1415years) and late adolescence (16-19 years) selected by
stratified random sampling from July 2011 to June 2012.
Among them 182 were current smokers. Descriptive
statistics were used to explore the reasons behind
quitting, duration and willingness to quit in the future.
Results: The prevalence of current smoking was 13.7%.
Among the current smokers, 66.5% had attempted to
quit in the past because they believed smoking was
harmful to health (35.5%), their parents instructed them

to quit (21.5%) and they did not want to continue
smoking (10.7%). The median duration of quitting was
150 days (Inter-quartile range 60 to 315 days). Nearly
8% of the current smokers were unwilling to quit
smoking in the future because they thought it is already
a habit (60%) and they could not succeed to quit (20%).
Conclusion: Even after multiple quitting attempts at
smoking during adolescence, relapse often occurs.
Tobacco focused interventions to support abstinence
are of potential value at this stage where the adolescents
are most vulnerable to habituate smoking.
PC-887-04 Effective method for quitting

tobacco
A Tenna1 1ADIC Alcohol and Drug Information Centre,
Colombo 05, Sri Lanka. Fax: (94) 112 508 484. e-mail:
tennaamaranath@gmail.com
Background and challenges to implementation: In Sri

Lanka, Tobacco prevalence is in decline trend. Cigarette
consumption has reduced by 7% in 2014 compared to
2013. However, the usage is high in certain groups
compared to the general public (ADIC spot survey,
2014). Through this survey, it was identified that officers
in armed forces have a high prevalence rate of tobacco
use (63%). Therefore, an intervention was implemented
in Air Force and Navy camps in order to help smokers to
quit smoking. Objective: To reduce tobacco consumption
among soldiers in selected Air Force and Navy camps, Sri
Lanka
Intervention or response: The intervention was conducted with 712 officers in 2014. Majority of the tobacco
users were daily users (n 236, 53%). As the first step, a
discussion was conducted with Public Health Officers
(PHI) attached to the medical unit of the camps to
introduce the programme. Respective PHIs organized
awareness programme in their camps. ADIC conducted a
3-hour awareness programme to develop enthusiasm
among soldiers to quit smoking. The strategy used was to
calculate expenditure on tobacco and explain how they
have become victims of tobacco industry. While increasing the enthusiasm to quit smoking, ADIC had introduced 6 steps method for tobacco quitting. Followup
calls were given within 4 weeks.
Results and lessons learnt: 150 tobacco users were
selected to evaluate the project impact. The evaluation
was carried out in two stages. Firstly, the post test
conducted at the end of the awareness programme
revealed that 72(48%) took a decision to stop smoking
and 68 (45.33%) decided to reduce smoking, 10 (6.66%)
didnt respond. Second stage evaluation was conducted
over the telephone conversations within a month after
the awareness programme. 138 responded in total and 59
out of 72 smokers who decided to stop actually stopped
smoking while out of the 68 who decided to reduce, 51
reduced, 12 decided quit and 17 have given up their
attempt to reduce.
Conclusions and key recommendations: Most of the
smokers want to be free from smoking. What is required
is an intervention that doesnt portray quitting as a

difficult task rather an intervention that raises their
enthusiasm to quit.
PC-888-04 Developing and evaluating a

behavioural support intervention for
smokeless tobacco cessation in Pakistan and
the UK
O Dogar,1 C Jackson,1 R Bibi,1 H Thomson,2 I Kellar,3
K Siddiqi1 1University of York, York, 2Leeds City Council,
Leeds, 3University of Leeds, Leeds, UK. e-mail:
ofd500@york.ac.uk
Background: Around 300 million people worldwide

consume smokeless tobacco (SLT). Considered to be a
part of the culture, its use is particularly common in
people of South Asian-origin. A recent Cochrane review
suggests that behavioural support interventions (BSIs)
are likely to be effective in SLT cessation. However,
studies in this review were mainly conducted in the US
and European populations who, compared to South
Asians, use less addictive and less hazardous products
(e.g. moist snuff). Moreover, its consumption among
South Asians has strong socio-cultural dimensions,
which need to be considered in BSIs. We aim to evaluate
the acceptability, feasibility and fidelity of a new BSI for
South Asians trying to quit using SLT.
Design/Methods: This is a two-country, multi-methods
pilot study, using qualitative and quantitative methods to
develop a solid theoretical framework for development
of a SLT cessation intervention that is culturally
appropriate for use with the South-Asian population.
The BSI was developed, tested and refined in three
iterative phases: Development phase (6 months): Two
expert panel workshops were conducted to, first, select
the behavioural determinants of SLT use and related
behaviour change techniques (BCTs) based on their
importance and relevance and second, to inform the
structure and key messages for the intervention. Feasibility assessment phase (8 months): The BSI was tested at
five sites in the UK and Pakistan, delivered by trained
cessation advisors to 32 SLT users. All BSI sessions (prequit, quit and post-quit) will be audiotaped and analysed
for fidelity, 16 participants and four advisors will be
interviewed in depth to assess the feasibility and
acceptability. All study participants will be tested for
SLT use status via urine/salivary cotinine testing at 6
months. Refinement phase (4 months): The BSI will be
refined in light of the results from feasibility testing.
Results: Findings from the qualitative interviews, the
audio-taped sessions and participant SLT use status at 6
months will inform any further modifications needed in
the content, structure or format of the BSI.
Conclusion: The refined BSI will be the first behavioural
intervention of its kind developed specifically for SLT
users of the South Asian-origin, with a comprehensive
methodology involving the use of BCTs proven effective
for tobacco cessation, ready for effectiveness testing in a
randomised controlled trial.
S239
23. Social media campaigns

PC-889-04 Using a social media tool to enhance
formal media in exposing tobacco industries
interference in Bangladesh
M Mehedi,1 I Rasul1 1PROGGA Knowledge for Progress,
Dhaka, Bangladesh. e-mail: mehedi198598@gmail.com
Background and challenges to implementation: 23

million Bangladeshi people are actively using social
media in a regular basis. This has become a popular
mode of media linking people in and out of the country.
The formal media is not much aware how to use this
technology as a support for exposing tobacco industry
interference. Recently PROGGA, which runs a network
consists of 300 journalists all around Bangladesh
launched a social media tool (www.facebook.com/
tobaccoindustrywatch.bd) to expose tobacco industry
interference at present and share the stories for all
stakeholders including the journalist who are engaged in
tobacco control. As this is a new technology the
traditional journalists are not much used to using this
tool. This is a challenge to implement this type of tool at
this moment.
Intervention or response: The journalists were given
training on social media tools usages in their day-to-day
reporting in respective media houses. They were also
trained to keep them updated using different types of
social media tools.
Results and lessons learnt: News, stories, pictures, video
and audio clippings, which reflects tobacco industrys
interference were collected from different sources and
were exposed and disseminated in the above mentioned
social media tool and other electronic outlets. The
journalists of the network contacted PROGGA to know
more about this news stories and did further investigation for in-depth reporting on this issue. They published
these stories in formal media and repeatedly exposed the
companys ill tactics. Thus the tobacco industry had to
stop many of the campaign/promoting activities they
were running unlawfully. Thus how PROGGA exposed
the tobacco industrys ugly faces to the policy makers and
to the public showing concern for public health.
Conclusions and key recommendations: Exposing industry interference by social media hindered their business,
which brings success to public health by preventing more
people from tobacco usage. Usage of social media to
combat industry interference should be highlighted by
training more journalist and try to use this technique in
their day to reporting activities.
PC-891-04 Education, communication and

public awareness about tobacco control:
findings from the ITC Brazil survey
C Perez,1 F Mendes,2 T Cavalcante2 1Brazil Cancer
Foundation, Rio De Janeiro, RJ, 2National Cancer Institute of
Brazil, Rio De Janeiro, RJ, Brazil. e-mail: cperez@inca.gov.br
Background: Article 12 of the FCTC requires Parties to

promote and strengthen public awareness of tobacco
S240
control issues by providing broad access to public

awareness programs on the health risks of tobacco
consumption and exposure to tobacco smoke, and about
the benefits of cessation. The International Tobacco
Control Policy Evaluation Survey Brazil (ITC) measured
changes in smokers knowledge of specific health risks of
smoking and perceptions of harm.
Design/Methods: The ITC Brazil Survey is a longitudinal
cohort survey and was conducted in Brazil with 1,200
adult smokers and 600 adult non-smokers living in three
cities. Telephone-administered surveys were conducted
using an area stratified random sampling strategy,
yielding a representative sample of the four largest cities
in Brazil (Rio de Janeiro, Porto Alegre, and Sao Paulo).
The Waves included questions a variety of tobacco
control measures, including education, communication,
and public awareness related to smoking.
Results: The ITC Brazil Survey asked respondents
whether they had noticed any advertising or information
about the dangers of smoking, or encouraged quitting. At
Wave 1, 37% of smokers and 34% of non-smokers in the
combined sample often or very often noticed
information about the dangers of smoking. These
percentages decreased to 25% of smokers and 21% of
nonsmokers at Wave 2. At Waves 1 and 2, smokers were
asked to report on various reasons that led them to think
about quitting smoking in the last 6 months. There was a
decrease in all reasons to quit between Wave 1 and Wave
2.Significantly, fewer smokers at Wave 2 stated that
advertisements or information about the health risks of
smoking somewhat or very much led them to think
about quitting.
Conclusions: The findings suggest that noticing advertisements related to the dangers of smoking, or encouraging quitting, decreased between Waves 1 and 2. Only a
quarter of smokers surveyed frequently noticed antismoking information at Wave 2. Similarly, there were
decreases at Wave 2 in the percentage of smokers who
thought about the harm smoking might be doing to them
or to others. These decreases are of concern and suggest
that anti-smoking campaigns and continued strengthening of educational activities are needed to increase the
visibility of anti-smoking information, which is aimed
especially to prevent smoking initiation and to encourage
quitting.
PC-892-04 Media intervention as a tool to

oppose tobacco industry interference
B Sharma1 1Freelance, Indore, India.e-mail:
silkdrops@rediffmail.com
Background and challenges to implementation: In India

one of the member of parliamentary committee to review
tobacco control in India said that there are no Indian
studies to prove link between tobacco and cancer.Later
on it was found out that the statement was given by
committee member who had bidi manufacturing unit. As
India is signatory to Framework Convention on Tobacco
Control (FCTC) and article 5.3 of FCTC prohibits
tobacco industry interference, the comment by a
committee member was taken seriously by Indian media

and lot of media intervention was done in the matter.
Intervention or response: Media response regarding the
comment of parliamentary committee member was
observed in Dainik Bhaskar newspaper which is one of
the leading newspaper in India. News was analysed for a
period of 15 days.
Results and lessons learnt: Front page news was
published with catching headlines like smoking 60
cigarettes a day with no cancer, no death. The headline
was very effective in generating public interest and
putting pressure on government. Newspaper also called
all the 15 members of the committee and asked them
about their views on the matter, their views published on
front page. Statement of opposition leaders and ex
Health Minister in favour of tobacco control also
published. A special viewpoint of newspaper editorial
in support of tobacco control was also published.
Newspaper highlighted the link between parliamentary
committee members and tobacco industry and direct
conflict of interest was exposed. The newspaper also
exposed the fact that as government has to implement
pictorial warning on 85 % of tobacco package and in
order to prevent that the committee members were trying
to postpone the implementation of pictorial warning.
Emotional stories of tobacco victims were published.
Government of India had to intervene in the matter and
Prime Minister Said that there should be no conflict of
interest involved in taking decisions related to tobacco
control.
Conclusions and key recommendations: Media stories
when drafted in a good manner and published regularly
with good space can play an important role in exposing
tobacco industry interference.
PC-893-04 Media advocacy for the

implementation of regulating smoking in films
B Mathew1 1Voluntary Health Association of India, Delhi,
India. e-mail: binoymathew84@gmail.com
Background and challenges to implementation: India,

the worlds largest producer of movies produces more
than 1000 movies a year in several languages. Bollywood
represents the Indian Hindi movie industry and the
worldwide viewership for their movies is estimated to be
about 3 million. Bollywood movie stars in India are
public figures, have large fan followings and exercise
tremendous influence on the behavioural attitudes of
adolescents. One of the major influences on the uptake of
teen tobacco use is the glamorization of tobacco use in
movies and on television. Movies are seen as very
influential for kids and teens.
Intervention or response: Using earned media to create
pressure on policymakers and to put indirect pressure on
the Government to ensure implementation of the rules
notified by the Government of India which firmly
intended to regulate the depiction of tobacco use in
Films and Television with effect from the prescribed date
of implementation. The strategy was to ensure that news
items or stories come out to attract the attention of the
government and the public. I increased consumer

awareness about the issue of smoking scenes in films
through a sustained strategy of media engagement. We
increased our interactions with the media, both on a oneto-one basis.
Results and lessons learnt: In order to sustain the medias
engagement with the issue, over the period of time, about
8 to 10 press releases were shared with the journalists
and this strategy of media advocacy resulted in nearly 40
45 stories in National and Vernacular Media during
May to December 2012.
Conclusions and key recommendations: Extensive media
coverage pan India (spanning national, regional and local
media), Ensured that news items or stories on prohibition
of smoking in films are published/aired regularly to
attract the attention of the government and the public.
From November 2012 onwards, the new films screened
at movie halls started showing the disclaimers of 30
seconds each on the ill-effects of tobacco use with strong
graphic pictures of cancer affected mouth, pictures on
how smoking causes heart attacks, effects of second hand
smoke etc both in English & Hindi that appeared in the
beginning and during the interval of the movie.
PC-894-04 #AnswerSunita: a social media

advocacy campaign to protect new pack
warning notification in India
P Puri,1 S Hamill,1 D Svenson,1 P Chaturvedi,2
N Murukutla,1 S Mullin,1 V Mallik,1 T Johnston1 1World
Lung foundation, New York, NY, USA; 2TATA Memorial
Hospital, Mumbai, India. e-mail:
ppuri@worldlungfoundation.org
Background: In August 2014, then Health Minister of

India launched a national tobacco control testimonial
campaign Sunita who after being diagnosed with
tobacco-related oral cancer at a young age of 27 years
underwent major surgery and then agreed to tell her story
in a PSA to raise awareness about the harms of tobacco.
Two months subsequent to campaign launch, new larger
and stronger pack warnings were notified to be
implemented from 1st April15. However, the move
was delayed after tobacco industry interference. Ironically, Sunita died on the same day of implementation,
leaving a letter with dying request of larger graphic Pack
Warnings to PM Modi. WLF and partners launched
#AnswerSunita, a social media advocacy campaign
aimed at higlighting the cost
Intervention or response: The key message of the
campaign was PM Modi: #AnswerSunitas dying
request and pick childrens health over tobacco industry
profits by passing larger graphic pack warnings. The
ongoing effort was intended to be intense but short lived,
working in synergy with advocacy efforts from other
partners. The campaign comprised of four main pieces
an online #AnswerSunita microsite with death clock
and profit clock, social media tools like ThunderclaP
(an online flash mob application), Facebook and
Twitter, traditional media PR and daily social media
messages.
S241
Results: The campaign generated more than 200 news

stories in three days and helped build pressure on the
government. The campaign reached over 275 000 people
through social media, with many signing up to a
ThunderClap campaign to simultaneously tweet Prime
Minister Narendra Modi calling for larger graphic
warnings. Monetization of the media coverage is
underway.
Conclusions: This paper demonstrates how social media
tools can be used to target messages to policy makers in
India. The #AnswerSunita campaign served as a framing tool, which allowed advocates to portray the human
face of the issue and the health costs. The campaign
underscores how social media can hold significant
promise for tobacco control advocacy.
PC-895-04 Media coverage of tobacco control

activities in Mizoram, India, 2011-2014
J R Ralte,1 T Lalbiaksanga,1 L Rokhum,1 H Renthlei,2
L Lalnuntluangi,3 Z Chhakchhuak,3 R J Singh4 1Mizoram
State Tobacco Control Society, Aizawl, 2Mizoram University,
Department of Management, Aizawl, 3Health & Family
Welfare, Govt. of Mizoram, 4International Union Against
Delhi, India. e-mail: tluangtlei@gmail.com
Background: Mizoram, a tiny Northeast Indian state has

the highest prevalence of adult tobacco use in India 67% (GATS 2009). Mizoram State Tobacco Control
Society (MSTCS) has undertaken the task of implementing tobacco control activities in the state. MSTCS has
worked in close partnership with the Mizoram Journalists Association (MJA) and as a result has garnered
strong support from the media. Since unpaid media
coverage is a good indicator of public education and
awareness regarding a particular issue, improving media
coverage of tobacco control activities in both print and
digital media is a positive development towards reducing
tobacco use prevalence. This study attempts to assess the
level of media coverage received by tobacco control
activities in the state from 2011 to 2014.
Design/Methods: A detailed analysis of tobacco related
news articles in print media (8 major newspapers) and
digital media (television & radio) was conducted for the
period 2011 - 2014.
Results: A steady increase in coverage of tobacco related
news and anti-tobacco squad (ATS) has been observed.
Over the study period, 31% (n 339) of all media
coverage is news report of tobacco control activities in
print media; 34% (n 375) is ATS related. 9% (n 102)
are slogans and 8% (n 92) are articles/essays. 6% (n
61) are related to trainings and workshops. Other print
media items are Public notices - 3% (n 33), Editorials 3.7% (n 41), advertisements (anti-tobacco messages) 0.7% (n 8). Digital media coverage is relatively lesser 1.1% (n 13) for All India Radio and 2.2% (n 25) for
Doordarshan (national television). The Figure depicts the
increasing trend in media coverage.
Conclusion: Due to close partnership and collaboration
with media personnel and MJA, MSTCS has made great
strides in making tobacco control visible and a part of
S242
daily life for Mizos through unpaid media coverage. This

will be instrumental in conditioning the social environment in relation to smoking and aid in reducing the
overall prevalence of tobacco use in the future. Future
efforts should focus on increasing digital media coverage
and more articles on harmful effects of tobacco in print
media.
Tobacco Control Networks are among the results of this

Social Media Workshop for Tobacco Control.
Results and lessons learnt: Six Communication Plans
were developed to address the problems identified in the
SWOT Analysis. More hands-on time should have been
given to participants to enable them to experience the
actual procedures in posting contents using Facebook,
Twitter and other social media platforms.Having a
contest within the workshop pushed the participants to
engage and ask support within their own network of
friends in their social media circles.With the pool of new
social media warriors, online campaigns can easily be
launched and get national and local support.
Conclusions and key recommendations: More and more
tobacco control advocates all over the world should be
given this kind of training to be effective and stay
relevant in maximizing the use of social media in tobacco
control.
24. Engagement of CSOs, communities

and patients in TB control and
management
PC-897-04 Social media workshop for tobacco
control
C Garcia1,2 1FCTC Alliance, Philippines, Quezon City,
Philippines; 2World Lung Foundation, New York, NY, USA.
Fax: (632) 441 4383. e-mail: kaloi.garcia@gmail.com
Background and challenges to implementation: In this

day and age, the potential of social media cannot be
taken for granted. With the Youth as the main users of
Social Media, tobacco control advocates should know
how to maximize the necessary tools to reach out to the
Youth, which are also the main target of the tobacco
industry. This training workshop aims to arm selected
Tobacco Control advocates with appropriate knowledge
on the use of Social Media by increasing their capacity in
public engagement to promote tobacco control issues
using digital social media tools. Health Promotion
Officers from different Regional Department of Health
offices, Tobacco Control Focal Points and Local Legislators coming from 16 Regions of the Philippines
attended the workshop.
Intervention or response: The 8-step Communciation
Planning template of the World Lung Foundation was
used in this workshop. To wit: Step 1: Define the problem
Step 2: Identify target audience Step 3: Develop
campaign goals Step 4: Create key messages Step 5:
Partnerships & resources Step 6: Campaign materials
Step 7: Implementing the campaign Step 8: Evaluate the
campaign In addition, a SWOT (Strength-WeaknessOpportunities-Threats) Analysis using World Cafe methodology was incorporated in the workshop to probe
deeper in defining the tobacco problem in the Philippines. Establishment of a pool of social media campaigners in expanding the tobacco control movements sphere
of influence, create demand for stronger smoke-free
policies, and help develop a digital communication plan
amongst advocates of the Regional and Provincial
PC-898-04 Outcomes of private practitioners

engagement in tuberculosis case notification in
Afghanistan: a document review
H Akhgar,1 H Manochehr,1 M Rasoli,1 G Qader2 1Ministry
of Public Health, Kabul, 2Management Sciences for Health,
Kabul, Afghanistan. e-mail: ntp.mhakhgar@gmail.com

private health sector progressed significantly in Afghanistan. According to an assessment private sector is the first
contact for 65% of Afghans. Also, large number of
persons presumptive to have TB attending these facilities
that usually go unprofessional identification, diagnosis
and treatment that usually resulted in delayed diagnosis
and treatment initiation of TB. Addressing these challenges, NTP conducted national situation analysis (NSA)
for public private (PPM) DOTS strategy, policy, and
practical guidelines drafted, to establish a framework for
TB services by private health provider, indicating the
actions required for implementation and successful
achievement of the strategy. We assessed the outcomes
of implementation of PPM-DOTS strategy on TB case
notification and treatment in three provinces of Afghanistan.
Intervention or response: The team used standard TB
programs recording and reporting forms as data
collection tools. Between January to December 2014, a
total of 140 private practitioners oriented on TB DOTS,
conducted regular quarterly review workshops aiming at
monitoring DOTS implementation through private
practitioners.
Results and lessons learnt: The data for Jan-Dec 2014
shows that in 4806 presumptive TB cases referred by
those engaged in PPM, of them 2034 cases diagnosed as
TB all forms that made 6% of total diagnosed TB cases in
the country and 712 cases as bacteriological confirmed
TB. Also, 819 cases diagnosed as sputum smear negative

TB cases and 503 extra pulmonary TB cases.
Conclusions and key recommendations: The above
findings shows that engagement of private practitioners
contributed to significant improvement in TB case
notification in intervention provinces, improved referrals
and strengthened coordination mechanism between
public and private health facilities, also, a sustainable
partnership for TB control established with the private
practitioners. Thus, we strongly recommend scale up of
PPM DOTS to other provinces and similar settings
elsewhere.
S243
The cure rate for smear positive migrants increased from

11.36% in 2003 to 84.49% 2013, with no significant
difference with local residents.
Conclusions and key recommendations: The communitybased multi-department collaborative tuberculosis control model has a positive impact on TB control and
prevention.
PC-899-04 A community-based multidepartment linked collaborative tuberculosis

control model in China
H Su,1 l Tang1 1Minhang District Center for Disease Control
and Prevention, Shanghai, China. Fax: (86) 215 488 5229.
e-mail: suhualin@sina.com
Background and challenges to implementation: China

has the worlds second largest tuberculosis epidemic
(after India). After 2004, the government increased its
commitment and leadership to tackle public health
problems and, among other efforts, increased public
health funding, revised laws that concerned the control
of infectious diseases. Before 2000, specialized TB
hospitals in Shanghai were in charge of the diagnosis
and treatment of TB. With the development of society of
China, the system faced with many challenges. One
group of special concern is work migrants, most often
poor men, who leave the countryside to join the wage
economy in towns and cities all over China. The Chinese
experience has shown that investment in communitybased intervention control programs of TB was needed,
and indeed essential, to improve the adherence to anti-TB
treatment, especially for migrants.
Intervention or response: A community-based multidepartment linked collaborative tuberculosis control was
established in the Minhang District. Under the system,
district CDC are in charge of planning, surveillance,
management, monitoring, assessment, health education
and faculty training of TB control and prevention; the TB
designated hospitals are responsible for the reporting,
diagnosis, treatment of TB patients; the community
health centers are the venues for delivery of medications,
observed treatment, contact tracing and health education
of TB patients. The core characteristics of the system are
multi-department linkage and decentralization of the
facilities related to the TB control. All TB suspects are
compulsorily referred to designated hospitals for diagnosis and treatment. Once a patient is diagnosed as TB,
community health centers will perform the management
and follow-up for the patients. The CDC will train the
staff, monitor DOTs implementation and conduct
assessment for TB program regularly.
Results and lessons learnt: Since the program was
initiated in 2003, referral arrive rate increased from
64.91% to 95% in 2014. The coverage of monitoring for
the migrants increased from 54.98% to 98% in 2014.
PC-900-04 Using community health volunteers

to reach TB patients living in remote areas: the
islands of South Nias Regency, Indonesia
E Siahaan,1 J Ake,1 S Ruschak1 1Yayasan Menara Agung
Pengharapan Internasional, Medan, Indonesia. Fax: (62)
787 7636. e-mail: esiahaan@map.org
Background and challenges to implementation: South

Nias Regency is one of the most under-developed areas in
Sumatera Province, Indonesia. It contains over 100
islands, many of which are inhabited by indigenous
people. TB prevalence in the Regency is higher than the
revised national estimate, but due to the lack of human
resources and the high transportation costs between
these remote islands, many TB patients do not access
care.
Intervention or response: Indigenous community volunteers were recruited and trained to screen for symptoms
of TB. Individuals suspected of having TB were then
referred to a health facility for smear microscopy.
Community volunteers facilitated transportation to
ensure testing occurred. When a patient was detected,
the volunteers helped to initiate and follow up treatment
on the remote islands. In addition to these communitybased activities, laboratory technicians were trained and
monitored to improve the quality of diagnostic services
offered through government health facilities.
Results and lessons learnt: In a span of just 9 months,
1536 community volunteers screened over 50 000 people
S244
for symptoms of TB across 112 villages. These activities

resulted in the detection of 579 suspected TB patients
who were referred for testing. Over 225 smear-positive
TB patients were diagnosed and 44 of these newly
detected patients were among children aged less than 14
years. These active case finding activities resulted in
86% increase in new smear-positive case notifications
compared with the same quarters in the prior year,
indicating that the TB patients detected among people
living on remote islands would likely not have been
treated in the absence of the project.
Conclusions and key recommendations: This project
demonstrates that community health volunteers can be a
highly effective solution for expanding TB care services
into remote and underserved populations in conjunction
with strengthening existing government laboratory
services. In light of the significant upward revision of
Indonesias prevalence estimate and subsequent decline
in its case detection rate, initiatives which are able to
expand services and demonstrate increases in anti-TB
treatment uptake should be prioritized for scale-up.
PC-901-04 Does timing of household visits for

active tuberculosis case finding affect
outcomes in a community-driven TB control
program in India?
N Singh,1 B Thapa,1 D R Mishra,1 A Das,1 S Chadha,1
S Mohanty,1 J Tonsing1 1International 4Union Against
Delhi, India. e-mail: bthapa@theunion.org
Background: A civil societies are engaged in community

driven Tuberculosis (TB) care and control programme in
northern Himalayan state Uttarakhand since April 2103.
Every month, 13 000 households are visited by community volunteers to disseminate TB related messages, find
presumptive TB patients (PTBP) and link to the diagnosis
and treatment services. Regular transportation is an issue
and activity is conducted during the afternoon hence
reaching all individuals in a household is a challenge. The
study aims to determine whether the intervention
conducted in the morning will lead to higher reach of
household individuals, dissemination of TB related
messages, identification of PTBP and TB diagnosis.
Design/Methods: The intervention was carried in 2
marginalized villages 15 kilometres from nearest
microscopy centres-in Haridwar Districts of Uttarakhand. The number of households were 900 (Bhogpur)
and 1100 (Shapura). The household visits were made by
community volunteers between 7 11 am (morning) in
Bhogpur and between 11 am-4 pm (afternoon) in
Shapura. The intervention included; dissemination of
TB related messages (symptoms, free diagnosis and
treatment services), distribution of information, education and communication (IEC) materials, identification
of presumptive TB patients (PTBP) through screening of
household individuals for cough or 2 or more than 2
weeks and linking to diagnosis and treatment services.
Results: Community volunteers visited 887 (98%) and
880 (80%) households in the morning and afternoon,
respectively. Of those visited, 60% (3316/5566) and

55% (2569/4701) individuals were directly reached
during the intervention with 5% increase (P , 0.001)
and an additional 747 individuals were reached in the
morning. Interestingly the morning and afternoon
interventions reached more children (70% and 67%),
and females (56% and 62%), respectively. Proportion of
PTBP identified in the morning was 1.4% (45/3316) in
comparison to 2.5% (65/2569) identified in the afternoon (P . 0.05). Of those, 2 (7.4%) and 3 (4.5%) TB
patients were diagnosed and initiated on treatment,
respectively (P . 0.05).
Conclusion: 747 individuals were additionally reached
through household visits conducted in the morning.
Household visits in the morning could reach more
individuals in a community driven TB control programme in geographically constrained villages of Uttarakhand.
PC-902-04 Local participation for TB control by

using a community-based organization model
in migrant populations
R Chantichai,1 C Thamsuwan,1 T Samart2 1World Vision
Thailand, Bangkok, 2World Vision Thailand, Ranong,
Thailand. Fax: (662) 229 211. e-mail:
ruangtong.ann@gmail.com
Background: Since 2007, the World Vision Foundation

of Thailand (WVFT) has received the trust and budget
support from the Global Fund - Round 6 as the Principal
Recipient for the project entitled TB Reduction among
Non-Thai Migrants (TB-RAM). WVFT received additional funding as a Sub-Recipient to DDC, MOPH as
part of the project entitled Universal access to quality
TB control & care and empowering communities in
Thailand. Both projects ended as of Sep14, but there is
continuing implementation to develop the cadre of
migrant workers (MW) to function as migrant health
volunteers (MHV) for referral community members who
have symptoms of TB for diagnosis. These MHV then
administer DOT for those persons diagnosed with TB.
Intervention: Project implementation in 5 border provinces are Phuket, Phang Nga, Chumporn, Ranong, and
Tak. Start from recruitment of MHV who then formed
groups of MW to establish CBO, with WVFT as the
coordination focal point and mentor in defining the nine
sections of the program as follows: (1) Community
Participation; (2) Conduct the activities in community
such as revolving funds to assist TB and HIV cases,
occupational development,etc.; (3) Form task forces for
ad hoc activities; (4) Recruit general membership;(5)
Conduct finance &accounting & auditing; and (6)
Perform miscellaneous tasks. Members of the program
share information on activities with stakeholders on a
regular basis at various local forums. Implementers use
these program meetings to provide educational information on TB and other prevalent diseases, and information
on rights of MW related to health promotion. There is
capacity building of the network of MHV, and an
emphasis on TB case-finding at the earliest stage of illness

for immediate referral for proper therapy.
Results: The program conducted training for over 5000
MHVs. A total of 56 CBOs have been established with
nearly 2000 members. There is a set of SOP for
coordination of case referral for those suspected of
having TB based on case-finding activities of the MHV.
Conclusions: The creation of this CBO is continuing
beyond the end of the external Project funding.The MHV
continue to provide TB education, screening and referral
for their communities.Thus, this is an example of a
locally-sustainable CBO for TB control which can be
expanded to address other infectious and non-communicable diseases that affect these expanding communities
of MW in Thailand.
PC-903-04 Integrating HIV and malaria into TB

outreach: lessons learnt from TB elimination
campaigns conducted in 2014 in Imo State,
Nigeria
A F Omoniyi,1 G Akang,2 A Awe,1 C Orji,3 G Madu,3
M Gidado,4 J Kuye,2 E Oyama1 1World Health Organization
Country Office, Abuja, 2 National Tuberculosis and Leprosy
Control Programme, Abuja, 3Imo State TBL Control
Programme, Owerri, 4KNCV Tuberculosis Foundation, Abuja,
Nigeria. e-mail: omoniyifadare@yahoo.com
Background and challenges to implementation: Imo state

with population of 3.9million ranked high among the 6
high TB-HIV burden states in Nigeria. The state notified
8% of the estimated TB cases in 2013. The programme in
collaboration with the HIV and malaria programme in
2014 conducted an outreach in one of the LGAs with the
lowest case finding in the state. The aim of this study is to
assess what worked, challenges and lessons learnt in the
integration of HIV and Malaria into TB outreach.
Intervention or response: Desk review was done to
identify LGA and community with high TB burden,
which led to the selection of the Amakohia/Akauma
community in Owerri North LGA. PHC Health workers
and community workers were trained; microscopes, antiTB drugs, HIV and malaria rapid test kits, posters in local
language and other tools were provided. Community
workers went around the community for one week
inviting community members with TB signs and symptoms to the PHC. TB diagnosis performed with sputum
AFB using front-loading system, presumptive TB cases
were routinely screened for HIV and malaria. Diagnosed
TB cases were started on anti-TB drugs, those with
malaria positive results were giving anti-malaria drugs.
HIV positive clients were linked to the nearest HIV
treatment center. GeneXpert test was done for chronically ill patients with AFB negative results.
Results and lessons learnt: 308 TB suspects (125 males &
183 females) were seen during the one week TB
elimination campaign, 60% whom were females. 5%
(15) of the TB suspects were diagnosed with TB.
14(93%) of the diagnosed TB cases were bacteriologically positive. 82% (253) of the TB suspects were tested
for HIV, 2% (5) were HIV positive. 14.4% (2) of the TB
patients was found to be HIV positive. All the TB-HIV
S245
co-infected patients were started on CPT. 70% (217) of

the presumptive TB cases were tested for Malaria; 17%
(36) of whom were positive and treated for malaria.
Conclusions and key recommendations: TB outreaches
provide opportunity for HIV and malaria programmes to
reach members of the community with their services at
the same time in a cost effective manner using the same
structure. The integration also ensure access of TB cases
to HIV and Malaria services. Most of the presumptive TB
cases with history of fever could also be having malaria
as shown by 17% malaria positivity rate among
presumptive TB cases. Collaboration with malaria
programme should also be strengthened in addition to
that with HIV programme.
PC-904-04 Increasing access to TB diagnostic

services through remote smearing stations in
Zamboanga City, Philippines
R Agbulos,1 J Bue,1 T Yu,2 S Dela Cruz,2 S Masulit,2
A Lagos3 1Systems for Improving Zamboanga City Health
Office, Zamboanga, 2Philippine Business for Social Progress,
Manila, 3Management Sciences for Health - SIAPS, Quezon,
Philippines. e-mail: virgilbelen@gmail.com
Background and challenges to implementation: Microscopy services are available in all 16 health districts, 10
hospitals, and 31 private clinics in Zamboanga City
(population: 907 725) in southern Philippines. However,
these are mainly in the urban center, out of reach of
people living in island and remote barangays (villages).
TB diagnostic services are inaccessible due to distance,
limited availability and high cost of transportation, poor
road infrastructure, and high cost of staying in the city.
This contributed to missed opportunities for TB diagnosis and subsequent treatment.
Intervention or response: In 2008, USAID/Philippines
supported the establishment of 12 remote smearing
stations (RSS) in access-poor villages. The RSS made
use of the barangay health station or any appropriate
pre-existing structure which was designated as smear
preparation area, and engaged barangay health workers
(BHWs) as informal laboratory workers (ILWs) after
completing a rigorous three-day training. Quality assurance and infection control practices ensured quality
smears and safe handling of specimens. The RSS
initiative applied a systems approach consisting of (i) a
situational analysis to determine the need for, feasibility,
and acceptability of RSS in an area; (ii) advocacy to
secure the barangay governments commitment to
provide logistical and community support; (iii) training
of ILWs in sputum smear preparation, infection control,
and recording and reporting using a modified training
curriculum designed specifically for ILWs; and (iv)
strengthening the system for RSS operations, including
a system for the referral and transport of slides with
smeared sputum, local financial support, supply management, recording and reporting, quality assurance,
supervision, and monitoring and evaluation.
Results and lessons learnt: The RSS brought diagnostic
services considerably closer and made access more
S246
convenient for patients and health workers in remote

areas. This led to increased detection of new smearpositive cases among residents living in remote areas. In
2008, the RSS contributed 3% (33/1095) to the citys
total smear-positive cases, and nearly half or 46% (417/
901) in 2014.
Conclusions and key recommendations: The RSS brings
quality-assured diagnostic services to underserved TB
clients in remote and inaccessible areas where microscopy services are unavailable, thus addressing equity issues.
PC-905-04 Piloting Photovoice as a

participatory research methodology in a
cookstove adoption study in Malawi
J Ardrey,1 N Desmond,1 R Tolhurst,1 K Mortimer1
1
Liverpool School of Tropical Medicine, Liverpool, UK. e-mail:
jane.ardrey@lstmed.ac.uk
Background: Many rural Malawians cook on open fires

and are exposed to high levels of air pollution from
burning biomass fuels (crop residues, wood, charcoal).
The WHO Global Health Observatory Report 2014
estimates that household air pollution (mainly from
cooking) resulted in 4.3 million deaths worldwide in
2012. The Wellcome Trust has a specific challenge of
connecting environment, health and nutrition and
supports research that bridges these themes. The
Cooking and Pneumonia Study (CAPS) is a Wellcome
Trust, MRC and UKAid funded village-level (n 150)
randomised controlled trial underway in Malawi (www.
capstudy.org). The CAPS intervention is an efficient
burning advanced cookstove that may reduce exposure
to household air pollution.
Intervention: The primary aim of CAPS is to study the
impact of the intervention on pneumonia in the under 5s.
It also provides a unique opportunity to gain understanding about the social and cultural factors that may
facilitate sustained use of improved cookstoves. In
January 2015 the use of Photovoice as a participatory
research methodology was piloted at the CAPS Chikhwawa site. Photovoice is a photographic technique that
allows communities and particularly women to share
knowledge about their perspectives and priorities in
areas such as food, cooking and household labour. For
this Wellcome Trust-funded pilot study 4 households (6
individuals) were trained to use a basic digital camera
and asked to collect images related to food and drink.
After 24-48 hours the cameras were collected and the
images processed. Each household was then interviewed
about why they took the images and this process was
captured on film. These interviews were conducted by
CAPS fieldworkers who were given guidance about the
methodology and provided with a topic guide.
Results: The pilot Photovoice work generated over 400
images and a 1 hour film that are revelatory of
community concerns about food and cooking and can
be thematically analysed. The participants were able to
use the technology after training (only 1 participant had
used a camera before), to produce clear and relevant
images and to explain their selection. The collection of
interview data through film was useful for capturing

discussion and was acceptable to participants and field
staff.
Conclusion: Photovoice is a feasible participatory
research methodology that can play a valuable role in
qualitative studies of improved cookstove adoption in
challenging developing country settings.
25. HIV and lung health: something for

everyone
PC-906-04 Spirometric restrictive pattern in
HIV-infected subjects: a cross-sectional
comparative study
P-Y Walter,1 E W Pefura-Yone,1,2 V Poka,1
A D Balkissou,1,2 A P Kengne3 1Faculty of Medicine and
Biomedical Sciences, University of Yaounde 1, Yaounde,
2
Yaounde Jamot Hospital, Yaounde, Cameroon; 3Medical
Research Council, Cape Town, South Africa. e-mail:
pefura2002@yahoo.fr
Background: Restrictive spirometric pattern is a risk

factor for all-cause and cause-specific mortality. We are
not aware of studies evaluating restrictive syndrome in
HIV-infected individuals. Accordingly, we conducted a
cross-sectional study to determine the prevalence of
spirometric restrictive pattern according to HIV infection
status, and investigated the determinants of restrictive
pattern in HIV-positive individuals in Yaounde.
Methods: A group of HIV-positive subjects, and age and
sex matched HIV-negative controls group were enrolled.
HIV-positive subjects were recruited in the approved
center for HIV treatment of Yaounde Jamot Hospital
between November 2012 and April 2013, and controls
were recruited through multilevel stratified random
sampling in general population in Yaounde, Cameroon
between December 2013 and April 2014. Spirometric
restrictive pattern was defined by a forced expiratory
volume in 1s/forced vital capacity (FEV1/FVC) ratio 7
lower limit of normal (LLN) and FVC , LLN (LLNrestrictive pattern) or by FEV1/FVC ratio 7 LLN and
FVC , 80% (fixed cut-off -restrictive pattern). Logistic
regression models were used to investigate the determinants of restrictive pattern.
Results: Final sample comprises 430 HIV-positive
subjects (68.4% women) and 444 HIV-negative subjects
(68.9% women). The mean % predicted FEV1 was lower
in the HIV-positive group compared to HIV-negative
group (90.12% vs. 95.01%, P , 0.001). The mean %
predicted FVC was also lower in HIV-positive subjects
(88.51% vs. 95.44%, P , 0.001). The prevalence of
LLN-restrictive pattern was 37.7% among HIV-positive
subjects and 23.2% in HIV-negative patients (P ,
0.001). Similarly, the prevalence of restrictive pattern
based on the fixed cut-off was higher in HIV-positive
subjects (28.8%) compared to HIV-negative subjects
(18.5%), P , 0.001. In multivariable analysis including
age, sex, education level, smoking, exposure to biomass
and body mass index, HIV infection remained associated
S247
to restrictive pattern regardless of definition used. In

further multivariable analysis in the HIV-positive group,
independent determinants of LLN-restrictive pattern
were low formal education [odds ratio (95%CI):1.66
(1.08 2.57), P 0.022] and smoking [1.97(1.08-3.59),
P 0.027]. Independent determinants of fixed cut-offrestrictive pattern in HIV patients were the low formal
education [2.25(1.40-3.62), P 0.001)] and chronic
respiratory symptoms [(2.00(1.26-3.15), P 0.003].
Conclusion: HIV infection is associated to high prevalence of spirometric pulmonary restriction. The determinants of restrictive pattern should allow early identification of HIV-positive subjects necessitating specific
interventions.
thrush (P , 0.05), pallor (P , 0.03), and failure to thrive

(P , 0.002) were significantly more common in children
with TB-HIV coinfection. Median (IQR) of weight in Kg
was significantly lower (P , 0.023) in TB with HIV. The
frequency of Ghon focus, adult-type infiltrate, cavitation, lymphadenopathy, pleural effusion, or pericardial
effusion on chest radiographic examination were not
different by HIV status (P . 0.05).
Conclusion: Overall, the clinical manifestations appeared similar between the groups. However, the
presence of skin rashes, failure to thrive, pallor or oral
thrush should prompt evaluation for HIV coinfection to
enable early initiation of cotrimoxazole and concurrent
antiretroviral therapy.
PC-907-04 Impact of HIV-coinfection on the

clinical presentation of childhood tuberculosis
in Ghanaian children
PC-908-04 Baseline anaemia among HIVinfected and non-infected patients initiating

treatment for rifampicin-resistant tuberculosis
in Khayelitsha, South Africa
A Enimil,1,2 F Gallani,3 A Sarfo,2 D Bosumtwe,2 T Opoku,2

A Badu-Peprah,1,2 S Antwi,1,2 A Kwara3,4 1Kwame
Nkrumah University of Science and Technology, Kumasi,
2
Komfo Anokye Teaching Hospital, Kumasi, Ghana; 3Miriam
Hospital, Providence, RI, 4Warren Alpert Medical School of
Brown University, Providence, RI, USA. e-mail:
tonash@gmail.com
Background: Childhood Tuberculosis (TB) remains a

major cause of morbidity and mortality globally. HIVinfected children are at a higher risk of developing active
TB and TB-related death. To understand the effect of
HIV disease on the clinical manifestations of TB in
children, we compared presenting symptoms and signs of
TB in the children with and without HIV coinfection.
Design/Methods: Children aged 3 months to 14 years old
with clinical diagnosis of TB with or without HIV were
enrolled into a prospective pharmacokinetic study of
first-line anti-TB drugs at Komfo Anokye Teaching
Hospital, Kumasi, Ghana. Baseline medical history and
examination were documented by the attending clinician, entered into Microsoft access and analyzed in R
statistical software. Descriptive statistics were used to
summarize the patient characteristics. v2 test was used to
compare dichotomized clinical variables. Continuous
variables between the two groups were compared using
Students t-test or Mann-Whitney U test (if data
distribution was skewed).
Results: Of the 144 children, 52 (36.1%) had TB alone,
82 (56.9%) were females and 117 (81.2%) had
pulmonary TB. There was no difference in median age,
age distribution or sex distribution between the two
groups. Pulmonary TB was diagnosed in 88 (95.7%) of
92 children with TB and HIV coinfection and in 29
(55.8%) of 52 of children with TB alone (P , 0.00). The
TB diagnosis in children with TB alone vs. TB-HIV
coinfection was clinical in 71.7% vs. 67.4%, radiographic in 23.1% vs. 28.3% or smear-positive in 5.8%
vs. 4.3%, respectively. There was no significant difference in the proportion of children presenting with cough,
weight loss, diarrhea, headaches, lymph nodes enlargement, and shortness of breath, vomiting and fever among
the groups (P . 0.05). Skin rashes (P , 0.023), oral
J Hughes,1 E Mohr,1 V Cox,1 L Wilkinson,1
`
G Van Cutsem,2 H Cox3 1Medecins
Sans Frontieres
(MSF),
Khayelitsha, 2MSF, Cape Town, 3University of Cape Town,
Cape Town, South Africa. e-mail: mohrek27@gmail.com
Background: Anaemia is commonly reported among

patients with rifampicin-resistant tuberculosis (RR-TB)
and human immunodeficiency virus (HIV) as a result of
the disease or drugs used for treatment (zidovudine,
isoniazid, linezolid). Since tenofovir is contraindicated
with aminoglycosides, baseline anaemia may limit
potential drug choices for either condition and may
necessitate additional interventions or regimen modifications. We investigated the frequency of baseline
anaemia in a setting of high RR-TB-HIV co-infection in
Khayelitsha.
Methods: We conducted a retrospective cohort study of
patients initiating RR-TB treatment in 2013. Frequencies
of HIV infected and uninfected RR-TB patients with
baseline anaemia were calculated.
Results: Among 197 patients starting standardised RRTB treatment, 139 (71%) were HIV-infected and 83
(60%) of these were on antiretroviral therapy (ART) at
the time of RR-TB treatment start. Although ART
regimen data was unavailable, .90% of HIV-infected
patients in Khayelitsha are initiated on tenofovir-based
regimens. Baseline haemoglobin (Hb) measurements
were available for 48 (83%) HIV-uninfected patients,
63 (76%) HIV-infected patients on ART and 47 (84%)
not on ART. Anaemia was graded according to
documents for the division of AIDS. Overall, 31/158
(20%) patients had at least moderate baseline anaemia;
8/48 (17%) vs. 23/110 (21%) HIV uninfected and
infected, P 0.5 (Table).
Conclusion: One fifth of patients initiating RR-TB
treatment had at least moderate baseline anaemia; there
was no significant association with HIV infection.
Concurrent use of haemotoxic drugs in co-infected
patients could contribute to worsening anaemia. Due to
limited treatment options and poor outcomes associated
with RR-TB, linezolid and isoniazid may still need to be
S248
used in anaemic patients. In these cases zidovudine

should be avoided and routine monitoring of Hb is
recommended throughout treatment. Countries with
high burdens of RR-TB and HIV co-infection, using
zidovudine in first line ART regimens, should consider
substitution with tenofovir. This will aid in reducing the
risk of baseline anaemia and drug interactions during
RR-TB treatment.
Table
HIV
infected
on ART
n (%)
No anaemia
Mild anaemia
Moderate
anaemia
Severe anaemia
Potentially lifethreatening
anaemia
Total
34 (54%)
16 (25%)
5 (8%)
6 (10%)
2 (3%)
63
HIV
infected
not on
ART
n (%)
HIVuninfected
n (%)
P value
27 (57%)
10 (21%)
33 (69%)
7 (15%)
0.2
0.5
4 (9%)
4 (9%)
3 (6%)
5 (10%)
0.9
0.9
2 (4%)
47
0
48
0.4
PC-909-04 Including traditional health

practitioners in community-based public
private partnership to provide HIV and TB
services in KwaZulu-Natal
N Gqaleni,1 W Mbokazi,1 B Buthelezi,2 S Radebe,3
M Shabangu,4 N Buthelezi,5 V Ngobese,5 G Nene6
1
Durban University of Technology, Durban, 2KwaZulu-Natal
Department of Health, Ulundi, 3KwaZulu-Natal Department
of Health, Newcastle, 4KwaZulu-Natal Traditional Health
Practitioners, Paulietersburg, 5KwaZulu-Natal Traditional
Health Practitioners, Newcastle, 6KwaZulu-Natal Traditional
Health Practitioners, Vryheid, South Africa. Fax: (27) 8657
70748. e-mail: nceba5850@gmail.com
Background and challenges to implementation: Traditional Health Practitioners (THPs) constitute an extensive network potentially capable of expanding and
simplifying access to comprehensive TB-HIV prevention,
care, treatment and support through various entry
points. Through the USAID TB Program funding, the
Durban University of Technology has conducted a
project on the engagement of THPs in improving access
to HIV and TB services.
Intervention or response: A pictorial HIV-TB Screening
Tool and a Referral Form were developed for use by the
THPs when referring their clients to health facilities.
Operating in two district of Amajuba and Zululand, 800
THPs were trained in their wards, linked with their warrooms, clinics, and community health care workers.
Training among others included referral, TB screening
and infection control, and HIV Counselling and Testing
(HCT).
Results and lessons learnt: A recent National Workshop
on THPs has endorsed the project including the referral
systems developed. The KwaZulu-Natal Department of
Health has incorporated the projects referral system into
its Referral Policy. During October 2013 to March 2015,
800 trained THPs consulted with 22 879 clients,

screened 1271 for TB-HIV and referred 818 suspects
for TB-HIV related symptoms. Using radio, community
dialogues and health talks we were able to reach 700 788
people. Twenty percent (160) THPs voluntarily undertook HCT. As part of promoting infection control 7630
N95 masks, were handed to THPs
Conclusions and key recommendations: We have demonstrated the potential role that can be played by THPs
in enhancing community based TB prevention, care, and
support programs which may now be scaled up.
PC-910-04 Longitudinal assessment of healthrelated quality of life of HIV-infected patients

treated for TB and HIV in South Africa
T Mthiyane,1 C Connolly,2 Y Balakrishna,2 A Pym,3
T Reddy,2 R Rustomjee1 1Medical Research Council, Cape
Town, 2Medical Research Council, Durban, 3KwaZulu Natal
Research Institute for Tuberculosis and HIV, Durban, South
Africa. e-mail: thulimthiyane24@gmail.com
Background: Measuring Health Related Quality of Life

(HRQOL) of human immunodeficiency virus (HIV)
infected patients diagnosed with tuberculosis (TB) and
receiving combined treatment has been rarely conducted
in public health facilities in South Africa.
Design/Methods: This clinical trial enrolled TB and HIV
coinfected and only HIV infected participants before
treatment commencement. Patients with CD4 count
.200 were randomized to ARV therapy or no ARV
therapy. HRQOL was measured at baseline, 3, 6 and 12
months. Participants were divided into four arms; Arm 1
TB-HIV on TB and ARV therapy, CD4 .200; Arm 2 TB-HIV on TB therapy only, CD4 .200; Arm 3 - TBHIV on TB and ARV therapy, CD4 ,200; Arm 4 nonTB-HIV on ARV therapy, CD4 ,200. Using the
Functional Assessment of HIV Infection (FAHI) instrument we calculated composite physical, functional,
emotional and mental health scores for participants that
completed at least three months of treatment, compared
scores between arms at month 0, 3, 6 and 12. The paired
t-test was used to test for change in total score between
each time point and baseline. GEE models which take
into account correlated observations were used to model
the total score as a function of time.
Results: Of 78 participants 76 (97.4%) participants
completed all 5 domains of HRQOL (physical, emotional, functional, social and cognitive). There was a
significant change in total score between month 0, 3, 6
and 12 (all P , 0.0001) and a significant increase in the
total score over time (P , 0.001). There was no
significant difference between the total score in arm 1
and arm 3 (P 0.719) or between arm 2 and arm 4 (P
0.860). Adjusting for baseline total score, baseline CD4
count had a significant effect on the total score (P
0.002) and the rate of change in total score over time
(interaction effect - P , 0.001).
Conclusion: Low CD4 count at enrolment was associated with poorer HRQOL at baseline and a higher rate of
change in HRQOL over time. There was a general
S249
improvement in quality of life of all patients, irrespective

of CD4 count, in all domains except cognitive domain.
This study should be conducted in a larger cohort using
closer time points to verify these results.
Table Immunologic and Nutritional Characteristics of ART

Nave HIV-infected Children with TB Disease
PC-911-04 Characteristics and outcomes of HIVinfected, ART-nave children diagnosed with TB

disease in Mbeya, Tanzania
WHO Immunosuppression
Severe
Advanced
Mild
Not significant
WHO Staging
Stage 4
Stage 3
Stage 2
Stage 1
J Bacha,1 J Benjamin,1 L Campbell,1 P Clowes,2

C Mangu,2 A Dinardo,3 K Ngo,3 A Mandalakas3 1Baylor
College of Medicine Childrens Foundation - Tanzania,
Mbeya, 2National Institute of Medical Research - Mbeya
Medical Research Centre, Mbeya, Tanzania; 3Baylor College
of Medicine, Houston, TX, USA. USA, e-mail:
bacha@bcm.edu
Background: HIV-infected children have increased risks

for TB exposure, infection and disease, as well as poor
response to treatment compared to their HIV-uninfected
peers. Treatment is often complicated by delays in TB
diagnosis, initiation of anti-TB treatment (ATT), and
initiation of antitretroviral therapy (ART). In ART-nave
HIV-infected children with TB, guidelines favour earlier
vs. delayed ART initiation following initiation of ATT.
However, the optimal timing of ART initiation is
unknown and limited data exist describing outcomes of
this paediatric population.
Design/Methods: Retrospective chart review between
March 2013 and December 2014 of patients at the
Baylor Tanzania Center of Excellence (COE) in Mbeya,
Tanzania. Inclusion criteria: HIV-infected children diagnosed with TB disease with no prior exposure to ART
(ART nave). Baseline, 6- and 12-month outcome data
were captured using a standardized data collection tool
and unique data base.
Results: Baseline data: 39 ART nave patients were
diagnosed with TB. 44% female (17/39); median age 6.1
years (0.7-17.8 yr). 97% (38/39) initiated ATT. Type of
TB: 85% PTB (33/39), 10% EPTB (4/39), 5% LNTB (2/
39). Diagnostic certainty: 23% confirmed (9/39), 56%
probable (22/39), possible 21% (8/39) TB. 46% (18/39)
were inpatients at time of diagnosis. All patients were
treated with RHZE and none reported prior ATT. The
Table compares the immunologic and nutritional characteristics of the cohort. Outcome data: 44% (17/39)
completed ATT, 5% (2/39) cured, 21% (8/39) with
ongoing ATT, 8% (3/39) transferred out, 8% (3/39) lost
to follow up, and 16% (6/39) died. Median time to death
after ATT initiation was 14.5 days (3-26d). A total of
87% (34/39) initiated ART, with median time from ATT
to ART of 15 days (6-55days).
Conclusion: Our cohort of ART-nave, HIV-infected
children with TB disease exhibited TB disease associated
with severe immunosuppression, malnutrition, and
inpatient admission, as well as high rates of early
mortality and LTFU, highlighting the importance of
early disease recognition and treatment. However, early
initiation of ATT and ART were associated with good
immunologic and nutritional outcomes at 6 and 12
months, supporting the recommendation of early ART
initiation following ATT in ART-nave children.
6 Month
Outcomes
(n20)
12 Month
Outcomes
(n12)
n32 with
CD4 data
31% (10/32)
22% (7/32)
22% (7/32)
25% (8/32)
n 17
w/CD4
18% (3/17)
12% (2/17)
35% (6/17)
35% (6/17)
n9 w/CD4
0% (0/9)
11% (1/9)
44% (4/9)
55% (5/9)
72% (28/39)
28% (11/39)
0% (0/39)
0% (0/39)
0 % (0/20)
0 % (0/20)
0 % (0/20)
100% (20/20)
0 % (0/12)
0 % (0/12)
0 % (0/12)
100% (12/12)
N/A
187 (7%)
60 (4%)
Baseline
(n39)
Median CD4 rise

absolute (CD4 %)
Nutrition Status
Severe malnutrition 64% (25/39) 0 % (0/20)
0 % (0/12)
Moderate
10% (4/39) 0 % (0/20)
0 % (0/12)
malnutrition
Normal nutrition
26% (10/39) 100% (20/20) 100% (12/12)
Anthropometrics
(median, range)
WHZ score
BMI
Weight-for-age
Z-score
1.61
(5.55
to 1.36)
15.0
(10.7
to 23.7)
3.73
(6.60
to 0.36)
Change:
Change:
1.29
1.40
1.75
2.12
1.03
1.31
PC-912-04 Smoking and tuberculosis among

HIV-infected men in Soweto, South Africa: a
case-control study
L Bronner,1,2 N Martinson,2,3 R Moloney,1 R Msandiwa,3
M Mashabela,3 J Golub1,2 1Johns Hopkins Bloomberg
School of Public Health, Baltimore, MD, 2Johns Hopkins
School of Medicine, Baltimore, MD, USA; 3Perinatal HIV
Research Unit, University of the Witwatersrand,
Johannesburg, South Africa. e-mail: lizabronner@gmail.com
Background: Although there is ample evidence that

smoking increases the risk of tuberculosis (TB), the
magnitude of impact on TB risk among HIV-infected
persons is poorly described. Given that a high proportion
of patients with TB are co-infected with HIV in South
Africa, this study investigates the association between
smoking and pulmonary TB (PTB) among HIV-infected
men in South Africa.
Design: We conducted an unmatched case-control study
of HIV-infected men in Soweto, South Africa. Eligible
subjects were antiretroviral therapy nave with confirmed
HIV. Cases had laboratory-confirmed PTB and controls
had no evidence of active TB. Participants were interviewed to collect data on socio-demographics, TB disease,
detailed smoking history, and alcohol usage. We defined
current smoking as self-reported smoking within 12
months prior to PTB diagnosis. We defined heavy alcohol
consumption as consuming 712 drinks per week. v2 and
non-parametric Wilcoxon tests were used in bivariate
analyses followed by multivariable logistic regression.
S250
Results: We enrolled 284 HIV-infected men; 146 cases

with PTB and 138 controls. Median age, CD4 count and
BMI of the cases and controls were both 38 years, 60 and
84 cells/mm3, 23 and 24 kg/m2, respectively. Forty-three
percent were current smokers; 48% among cases and
39% among controls (P 0.15). Heavy alcohol
consumption did not differ significantly between cases
and controls (75% vs. 69%), but among participants
who consumed the equivalent of 712 alcoholic drinks/
week, those who were also current smokers were more
likely to be cases with PTB (69% vs. 42%, P 0.04). In
adjusted analyses, smokers had twice the odds of PTB
(aOR 1.9; 95%CI: 1.1-3.4), while smokers who consumed 712 alcoholic drinks/week were 4.3 times more
likely to develop PTB (95%CI: 1.2-15.3) (Figure).
Conclusions: HIV-infected male smokers are at greater

odds of developing PTB than non-smokers, and heavy
alcohol consumption significantly increases this likelihood. Our data suggests the deleterious effects of
smoking in HIV-infected men with PTB are exacerbated
by heavy alcohol consumption. This relationship, we
posit, is likely due to a complex interaction of
environmental and social conditions, and immunological
responses. Our findings support calls for smoking and
alcohol cessation and prevention efforts to be scaled up
among HIV-infected patients in epidemic settings as part
of efforts to control PTB.
PC-913-04 High prevalence of depression and

hazardous/harmful alcohol use among TB-HIV
patients initiating ART in Lesotho
E Hayes-Larson,1,2 Y Hirsch-Moverman,1 S Saito,1,2
B Pitt,1 L Maama-Maime,3 K Frederix,1 A Howard1,2
1
Columbia University, ICAP, Mailman School of Public Health,
New York, NY, 2Columbia University, New York, New York,
USA; 3National Tuberculosis Control Programme, Lesotho
Ministry of Health, Maseru, Lesotho. e-mail:
elh2167@cumc.columbia.edu
Background: Mental health issues have not been well

characterized in TB-HIV co-infected patients in resourcelimited settings. We sought to describe the prevalence
and correlates of depression and hazardous/harmful (H/
H) alcohol use among TB-HIV patients initiating ART
within 8 weeks of TB treatment initiation who enrolled
in the Start TB patients on ART and Retain on Treatment
(START) Study.
Design/Methods: START is an ongoing cluster-randomized trial in 12 health facilities in Lesotho, evaluating the
effectiveness, cost-effectiveness and acceptability of a
combination intervention package vs. standard of care to
improve early ART initiation, retention and TB treatment success in TB-HIV patients. Consenting adults with
TB-HIV initiating ART within 8 weeks of TB treatment
initiation were enrolled in a measurement cohort (MC).
Data from MC interviewer-administered baseline questionnaires (collected 4/2013-3/2015) were analyzed to
describe prevalence of depression (PHQ-9, with cutoffs
for mild, moderate and severe of 5, 10, and 15,
respectively) and H/H alcohol use (AUDIT score 78).
Correlates of moderate/severe depression and H/H
alcohol use were assessed with unadjusted generalized
linear mixed models with a random effect for study site.
Results: Among 363 participants with available data, the
median age was 35 y, 56% were male, and 54% were
married/living with a partner. Overall, 42% reported
symptoms consistent with mild depression, 29% reported moderate/severe depression, and 29% reported H/H
alcohol use. Moderate/severe depression and H/H
alcohol use were not associated; 7% of participants
reported both conditions. Statistically significant correlates of moderate/severe depression included no education, more frequently needing or wanting help or being
stressed by family/friends, and low health literacy
(Table). Participants who had disclosed their HIV status
or planned to disclose their TB diagnosis had lower odds
of reporting moderate/severe depression. Significant
correlates of H/H alcohol use included male sex and
greater perceived TB stigma. Socially desirable response
tendencies were associated with reduced odds of
reporting both conditions.
Conclusion: Prevalence of moderate/severe depression
and H/H alcohol use were high in this sample, and were
correlated with different demographic, psychosocial, and
TB-HIV-related factors. Interventions to prevent and
treat H/H alcohol use and depression are urgently needed
for TB-HIV patients in these settings.
Abstract presentations, Saturday, 5 December
SATURDAY
5 DECEMBER 2015
e-POSTER SESSIONS
06. Preventing infection and
transmission of TB in health facilities and
the community
EP-144-05 Transmission of multidrug-resistant
tuberculosis among household contacts
previously cured of tuberculosis in Delhi, India
1
3,
R Singla, N Singla, J Caminero Luna, D Behera,

A Kumar,1 J Arora,1 M Varma Basil,4, R Sarin1,5 1National
India; 2University General Hospital of Gran Canaria Dr.
Negrin, Las Palmas de Gran Canaria, Spain; 3Post Graduate
Institute of Medical Education and Research, Chandigarh,
4
Vallabh Bhai Patel Chest Institute, New Delhi, 5National
India. Fax: (91) 1 126 517 834. e-mail:
drrupaksingla@yahoo.com
Background: There are some reports of transmission of

multi-drug resistant tuberculosis (MDR-TB) among
household contacts (HHCs) of MDR-TB patients.
However, transmission of MDR-TB among previously
tuberculosis (TB) successfully treated HHCs of MDR-TB
patients has not been described earlier.
Methods: HHCs of MDR-TB patients, enrolled between
September 2011 and August 2012 under national
programme, in a tertiary referral institute in Delhi, India,
were prospectively followed up for 2 years. Previously
TB successfully treated HHCs, who developed TB now,
were evaluated. The transmission dynamics was studied
by spoligotyping and mycobacterial interspersed repetitive units (MIRU) typing of cultures of index patients and
their corresponding HHCs, whenever possible.
Results: Among 839 HHCs of 299 MDR-TB patients
followed up in the study, nine HHCs, who had previously
been TB successfully treated with first line drugs, were
found to have developed TB now. Age of nine contacts
ranged from 14- 65 years, median age being 23 years
with Inter Quartile Range (IQR) 18-31 years. All were
found to have MDR-TB. All nine HHCs were HIV
negative. Two of nine HHCs had Beijing strain and two
had spoligotyping results of unknown strains. In these
four HHCs, including those with unknown strains, the
spoligotyping and MIRU typing of the HHCs matched
with their corresponding index confirming the possibility
of transmission of MDR-TB from the index to the
respective HHCs in the house hold in these cases. In rest
five HHCs the, index and HHC pair could not be
matched as the culture of index cases were not viable at
the time when the contact developed TB. Time gap
S251
between diagnosis of their respective index cases and

development of MDR-TB in all HHCs ranged from 2-16
months, median duration being 10.5 months with IQR 311 months. Among two HHCs with Beijing strains time
gap was just 3 months each.
Conclusions: There is high probability of transmission of
MDR-TB from the index cases to the HHCs in the house.
This occurs even if the HHCs have previously been cured
of TB. The Beijing strain, in particular, is highly
infectious leading to development of MDR-TB among
HHCs within few months of their exposure to the index
case. All HHCs, including the previously TB treated
HHCs of MDR-TB cases, should be actively investigated
for multi-drug resistance TB as a public health priority
under national programmes.
EP-145-05 HCW screening

L Anande,1 R Nanavare,1 V Chavan1 1Group of TB
Hospitals, Mumbai, India. e-mail: drlalitanande@gmail.com
Background and challenges to implementation: Tuberculosis is an infectious disease caused by bacteria called
Mycobacterium tuberculosis. Today tuberculosis has
been recognized as an important international concern
due to increase in incidence of acutely rapidly progressing forms, late detection of disease and wide spread of
strains resistant and multiresistant to the treatment. In
view of HCW, the challenge lies in combating the
increase in number of multidrug resistant tuberculosis,
prevent progression of latent infection to active tuberculosis, provision of insurance schemes and TB leave for
individuals under treatment, active participation of
NGOs, media, web domains for training modules,
inclusion of treatment for extra pulmonary tuberculosis
in government run programs and unstinted support from
government to develop thoracic surgery departments in
hospitals are vital.
Intervention or response: Health care workers are
exposed to TB bacilli most often. Group of TB hospitals
in Mumbai, India has initiated Infection control trainings
conducted by in-house Infection control committee of the
hospital. The committee is inclusive of internal and
external speakers providing training to health care
workers and patients to educate respiratory hygiene
and coughing etiquettes. These include wearing N95
mask for HCW, other masks for patients and visitors,
provision of high protein diet and regular health checkup
with records maintained in diaries are implemented in
hospital. The screenings of HCW was initiated in 2011
and since then have continued the trend quarterly. HCW
is inclusive of ward attendants, sweepers, barber, time
keeper, gardener, radiographer, doctor, nurse, clerk,
peon, operation theatre assistants, contract basis labor
and pharmacists to name a few. Until March 2015, total
number of HCW acquired TB infection are 63. Further,
the total number of HCW acquired MDR-TB is 36 where
as Non-MDR are 27. Total number of HCW cured of
infection is 20 and death due to infection is 11.
Results and lessons learnt: Consequently, the death rate
of TB cases has decreased and cure rate has increased in
S252
hospital. HCW are expected to abide respiratory hygiene,

update knowledge with education programs and participate in maximum numbers for screening activities to
maintain their good health record.
Conclusions and key recommendations: It is vital to
recognise good health of HCW imparting knowledge,
keep the trend of health checkup & maintain records
thus make India TB Free.
Conclusion: We found a 7-fold higher risk of TB among

HCWs compared to the community risk and identified a
~2-fold higher risk of TB among medical trainees
compared to the general HCW group. Our study is the
first to report several INH-resistant TB and MDR-TB
cases among HCWs in India and highlights the immediate need to implement appropriate administrative,
personal and environmental infection control measures.
EP-146-05 Tuberculosis outbreak among health

care workers and trainees in a tertiary care
hospital in Pune, India
EP-147-05 Tuberculosis and latent tuberculous

infection among health care workers in
Kisumu, Kenya
A Basavaraj,1 D Kadam,1 A Patil,1 A Chandanwale,1

V Mave,2,3 D Jain,2 A Deluca,4 R Bollinger3 1Byramjee
Jeejeebhoy Government Medical College, Pune, 2Byramjee
Jeejeebhoy Government Medical College Clinical Trials Unit,
Pune, India; 3Johns Hopkins University School of Medicine,
Baltimore, MD, 4Johns Hopkins University Bloomberg School
of Public Health, Baltimore, MD, USA. e-mail:
anita.basavaraj@rediffmail.com
J Agaya,1 C Nnadi,2 C Obonyo,1 V Obiero,1 V Lipke,2

E O Okeyo,1 K Cain,3 J Oeltmann2 1Kenya Medical
Research Instutute, Kisumu, Kenya; 2US Centers for Disease
Control and Prevention, Atlanta, GA, USA; 3U.S. Centers for
Disease Control and Prevention, Kisumu, Kenya. e-mail:
kcain@cdc.gov
Background: India has the worlds largest burden of

tuberculosis (TB) with an estimated 26% of the global
burden of TB and 2 million new cases every year. TB risk
among health care workers (HCWs) is known to be high,
and documenting the risk for HCWs in India may drive
improvements in TB guideline implementation. We
therefore set out to investigate the disease prevalence in
a group of at-risk HCWs. We also aimed to determine the
frequency of treatment failure, mortality, and drugresistance among HCWs at Sassoon General Hospital.
Design/Methods: A retrospective study of HCWs was
conducted between June 2011 and December 2013 in a
public teaching hospital. The primary outcome-TB
prevalence and the incidence estimate- were measured
by those self-reporting TB. Study assessments included
HCW interviews and clinical exams, as well as medical
record data abstraction. The primary outcome was to
find TB prevalence estimate, calculated as the number of
TB cases divided by estimated total HCWs. Descriptive
statistics were used to measure secondary outcomes for
frequency of treatment failure, adverse drug reactions,
mortality and drug-resistance.
Results: Of 1886 HCWs evaluable, 47 cases of TB were
identified. Fifteen (32%) were residents, 12 (26%) were
interns, 13% were nursing students and the rest were
staff nurses, medical students and other faculty. The
estimated incidence of TB was 1611 per 100 000 HCWyears for all HCWs and 2719 per 100 000 HCW-years
for medical trainees. The majority of TB occurred among
general medical residents (15 [32%]) and in general
medicine wards (23[49%]); 20 (43%) had pulmonary
TB, 4 (11%) had monoresistance to isoniazid and 3
(6%) had multi drug resistant (MDR)-TB. Two HCWs
died from co-morbid conditions of cirrhosis of liver and
ascites in alcoholic liver disease, 3 HCWs had persistent
positive cultures, and 31 (66%) developed adverse drug
reactions. One HCW developed severe hepatotoxicity in
the form of acute fulminant hepatic failure and underwent hepatic transplant.
Background: Health care workers (HCWs) in settings

with high tuberculosis (TB) prevalence are at an elevated
risk for TB compared to the general population. Data on
baseline community-acquired TB infection in some of
these settings, and on health facility types that may
confer the greatest nosocomial risk for TB, are limited.
Methods: We conducted a cross-sectional survey among
HCWs in Western Kenya during 2013. HCWs were
recruited from dispensaries, health centers, and hospitals
that offer both TB and HIV services. School workers
(SWs) from the health facilities catchment communities
were randomly selected to serve as the community
comparison group. Latent TB infection (LTBI) was
diagnosed by tuberculin skin testing (TST) and subsequent examinations to exclude TB disease. Human
immunodeficiency virus (HIV) status of participants
was assessed. Using an adjusted logistic regression
model, we determined the adjusted odds of LTBI among
HCWs compared to SWs; and among HCWs only, we
assessed work-related risk factors for LTBI.
Results: We enrolled 1005 HCWs and 411 SWs.
Approximately 60% of both groups were female. A
total of 211 of 958 HCW (22%) and 48 of 392 SW
(12%) were positive on HIV evaluation. Prevalence of
LTBI was 60% among HCWs and 48% among SWs.
Adjusted odds of LTBI was 1.5 times higher among
HCWs than SWs (95% confidence interval 1.22.0).
HCWs at all three facility types had similar prevalence of
LTBI (P 0.72), but increasing years of employment was
associated with increased odds of LTBI (P , 0.01).
Conclusion: HCWs at health care facilities in Western
Kenya which offer TB and HIV services are at increased
risk of LTBI, and the risk is similar across facility types.
The World Health Organization-recommended TB infection control measures are urgently needed in health
facilities to protect HCWs.
EP-148-05 Occupational TB as measure of

impact of TB infection control program
G Volchenkov,1 P Jensen2 1Vladimir TB Control Center,
Vladimir, Russian Federation; 2US Centers for Disease
Control, Atlanta, GA, USA. Fax: (7) 492 232 3265. e-mail:
vlchnkv@yahoo.com
Background and challenges to implementation: Tuberculosis (TB) infection transmission in health care settings
plays significant role in ongoing drug resistant TB
epidemic of high burden countries. Health care workers
may be one of the occupational groups most affected by
high TB transmission risk. A combination of TB infection
prevention and control (TB IPC) interventions are
recommended to reduce or eliminate TB transmission
in various settings.
Intervention or response: Intensive and complex TB IPC
program supported by US CDC and WHO was
implemented in the Regional TB Dispensary and other
TB facilities of Vladimir region, Russia in 20032014.
The program included FAST interventions and other
administrative controls, selected environmental controls
in high TB transmission risk areas and personal
respiratory protection for high risk staff and surgical
masks for contagious patients. We performed retrospective analysis of active TB notification data among health
care workers of Vladimir region in 1994 2014 to assess
the impact of these interventions on occupational TB rate
as a measure of nosocomial TB transmission level.
Results and lessons learnt: Implementation of TB IPC
program in TB facilities of the region resulted in sharp
reduction of active TB disease notification among health
care workers of both Regional TB Dispensary and all
other TB facilities of the region. Average TB notification
rate for Regional TB Dispensary staff in 19942003 was
1080 per 100K, and after TB IPC program implementation zero TB cases has been registered in 20082014 in
this facility. No TB cases have been registered among
health care workers of other TB facilities of the region
since 2009. At the same time average annual relative risk
of TB for physicians and nurses of the emergency
hospital, where no airborne precautions has been used,
was over 6,5 comparing to residence population of
Vladimir city.
Conclusions and key recommendations: Occupational
TB notification rate can be used as sensitive indicator of
both TB transmission risk in the facility, unit or
profession and as measure of effectiveness of TB IPC
interventions. A comprehensive and effective TB IPC
program may reduce nosocomial TB transmission risk in
few years.
EP-149-05 Strengthening and accelerating the

implementation of an integrated HIV-TB
workplace response for health sector workers
D Elray,1 F Fayers,1 M T Mbatha,2 S Mabhele3 1Hospersa,
Pretoria, 2Urc, Pretoria, 3Ilo, Pretoria, South Africa. e-mail:
hsc@hospersa.co.za

Africa currently ranks third in the world, in terms of the
S253
tuberculosis (TB) burden. It is estimated that 80% of its

52 million people are infected with the TB bacillus.Furthermore 10% of these infected with the TB bacillus
(latent TB) will develop TB in their lifetime. This
percentage increases dramatically if the person is human
immunodeficiency virus (HIV) positive. Health Care
workers (HCWs) in South Africa are at a risk of up to 4
times higher than that of the broader public to
contracting TB. It therefore becomes imperative that
we ensure that HCW remain occupationally TB-HIV
free. The incidence of Occupational TB in South Africa is
increasing but remains grossly underreported .This
underreporting is due to a range of contributing factors;
fear of stigma, lack of knowledge on TB-HIV risk, lack of
knowledge on occupational hazards and a lack of routine
medical surveillance for HCW.
Intervention or response: Scaling up the engagement of
health sector workers in fighting TB in both public and
private sector health facilities
Methods: There were 18 workplace wellness events
hosted by HOSPERSA, ILO and the USAID TB CARE II
Project, and these targeted both public and private health
facilities throughout the country, particularly aiming for
increased uptake for TB & HIV Screening and testing
amongst health sector workers. Training was held to
empower 270 Nurses from various clinics/hospitals
throughout the country on the clinical management of
TB-HIV, MDR-TB and improved infection control
measures. Three focus groups were held with young
health care workers to understand their perceptions of
TB-HIV risk at work. A baseline survey by way of a selfadministered questionnaire was conducted and 250
HCW responded, the survey aimed to find out about
structures in place to curb Occupational TB-HIV
exposure in public health facilities.
Results and lessons learnt: Health care workers are still
practicing inadequate implementation of infection control measures in healthcare facilities, which poses a risk
for contracting TB. Routine reporting of occupational
TB-HIV cases, to increase in the health facilities that this
project reached.
Conclusions and key recommendations: The need for a
national policy that promotes the protection of HCWs in
terms of occupational exposure to TB and HIV is evident.
Routine TB testing of HCWs does not exist. HCWs know
that they are at Risk of Contracting TB-/HIV but are not
equipped to take preventative steps.
EP-150-05 Systematic review of the

epidemiology and programmatic response to
TB in South African health care workers
L Nicol,1 S Mehtar,1 K Dheda,2 S Adams,2
M Van Der Walt,3 M Osman4 1Stellenbosch University,
Cape Town, 2University of Cape Town, Cape Town, 3South
African Medical Research Council, Pretoria, 4City Health
Directorate, Cape Town, South Africa. e-mail:
liesl.nicol@gmail.com
Background: Tuberculosis in health care workers is the

third most commonly reported occupational disease in
SA. Transmission of TB in health care facilities to both
S254
patients and health care workers has been reported from

every country of the world, regardless of local TB
incidence. The risk for transmission of TB in health care
facilities varies by setting, occupational group, local
prevalence of TB, patient population and effectiveness of
TB infection control measures. With the aim of
contributing to evidence-informed TB policy development in South Africa this systematic review collated all
relevant research on the epidemiology of and programmatic response to TB in healthcare workers in South
Africa.
Design/Methods: A comprehensive search for both
published and unpublished research was conducted in
April 2015. Based on specific inclusion criteria, research
investigating TB incidence, prevalence, risk factors,
prevention, treatment and care in health care workers
in South Africa and research reporting on TB-specific
infection prevention and control issues in South African
health care facilities were systematically reviewed.
Results: Twenty-eight studies were included in the
systematic review. Table 1 outlines the number of studies
addressing each of the topics, the study dates (range) and
setting. Most of the studies conducted in KwaZulu-Natal
showed a higher incidence of TB disease in health care
workers compared to the local population. A national
survey of 133 primary health care facilities, conducted
between 2006 and 2008, reported a 23 times higher
incidence rate of TB disease in health care workers
compared to the general population. Studies investigating TB infection, prevention and control (IPC) in South
African healthcare facilities are relatively dated, with the
most recent study conducted in 2011.
Conclusion: The optimal approach for TB screening in
South African health care workers, the use of isoniazid
preventative therapy (IPT) in South African health care
workers and the implications of providing IPT to health
care workers require investigating. More effective ways
of promoting, teaching and implementing TB IPC
processes in South African healthcare facilities need to
be investigated.
EP-151-05 Tuberculosis infection control:

knowledge and attitudes among health
workers in Uganda
E Buregyeya,1 E Mitchell2 1Makerere University School of
Public Health, Kampala, Uganda; 2KNCV Tuberculosis
Foundation, The Hague, Netherlands. Fax: (256) 414 53
1807. e-mail: eburegyeya@musph.ac.ug
Background: The WHO recommends TB infection

controls (TBIC) in health care facilities and these include;
managerial, administrative, environmental and personal
protective. In 2008, the Ministry of Health Uganda
(MOH) and the Tuberculosis Assistance Programme
initiated efforts to implement TBIC by training of
HCWs. Good knowledge of a health problem, accompanied with the right attitude, can result in desired
practices. Research among HCWs has found them to
often lack knowledge about TB and infection control,
which contributes the risk of TB transmission. This study
was carried out to assess HCWs knowledge and attitudes
towards TBIC.
Design/Methods: We conducted a cross-sectional study
among HCWs in health facilities in the districts of
Mukono and Wakiso in Uganda, from October 2010 to
February 2011. We assessed HCWs knowledge in basic
TB (standards of TB diagnosis and treatment) and TB
infection control (knowledge about TBIC measures such
as use of masks/respirators, triaging and ventilation) and
attitudes towards TBIC (perceptions about TBIC measures).
Results: Twenty four percent of the participants answered correctly all the questions about basic TB
knowledge. Overall, 62% of the HCWs were judged to
have adequate basic TB knowledge. At multivariable
analysis, non-clinical cadres, were more likely to have
poor basic TB knowledge [adjusted OR0.43(95%
CI0.27-0.68)]. Only 7% of the respondents answered
all the questions on TBIC correctly. Almost all the
respondents knew that TB was transmitted through
droplet nuclei, while only a third (34%; 174/532) knew
that masks do not protect the wearer from getting TB.
Overall, 69% (355/512) of the HCWs were judged to
have adequate TBIC knowledge. At multivariable
analysis, non-clinical cadres were more likely to have
poor TBIC knowledge [adjusted OR0.57(95%
CI0.360.91)]. Twenty-one percent (112/537) of the
HCWs mentioned that their risk of contracting TB was
the same whether the consultation window was open or
closed. Just over half of the respondents (264/513) were
rated as having positive attitudes towards TBIC. At
multivariable analysis,non-clinical cadre AOR
0.48(95% CI0.300.77) was more likely to have poor
attitudes towards TBIC.
Conclusion: Our study findings highlight the inadequacy
of HCWs knowledge in TBIC and poor attitudes towards
TBIC. Poor knowledge about TB and TBIC, and poor
attitudes towards TBIC were associated with being nonclinical. MoH should emphasis TBIC training among
non-clinical staff.
EP-152-05 Transmission of tuberculosis in

households in England
M Lalor,1 L Anderson,1 E Hamblion,1 A Burkitt,2
J Davidson,1 I Abubakar,1,3 L Thomas1 1Public Health
England, London, 2Public Health England,
Newcastle-upon-Tyne, 3University College London, London,
UK. e-mail: maeve.lalor@phe.gov.uk
Background: A reliable estimate of the proportion of

tuberculosis (TB) cases from recent transmission in
England would inform public health measures to control
TB. Estimates of transmission using strain typing data
alone have several limitations. The aim of this analysis
was to estimate the proportion of TB due to household
transmission, by combining strain typing data with
epidemiological data on links between cases.
Design/Methods: Household linked TB cases between
2010 and 2012 were identified from addresses reported
at notification or during cluster investigation. Household
linked TB cases were classified by 24 MIRU-VNTR
strain typing data into four transmission categories;
confirmed, probable, unknown and refuted.
Results: 7.7% (1849/24 051) of cases notified between
2010 and 2012 lived in a household with at least one
other TB case. For 67% of these cases, strain typing data
was unavailable for one or both cases; therefore
transmission could not be confirmed or refuted. For
those with strain typing data; 64% had confirmed, 11%
probable, and 25% refuted transmission. 718 households where transmission was not refuted were identified, 20% contained more than two cases. The median
time from symptom onset to treatment start was 65 days
for the first case in households, decreasing with
subsequent cases. The proportion of TB cases that were
in a household varied by country of birth; Bangladesh
(2.7%), Pakistan (5.0%), India (5.4%), Nepal (8.7%)
Somalia (10.7%), UK (11.1%).
Conclusion: It is widely accepted that the majority of TB
in England occurs due to reactivation of latent TB
acquired aboard. Given that almost 8% of notified TB in
England occurs in cases with household links, and that
the majority of those with strain typing were confirmed
to be due to transmission, demonstrates the important
contribution of household transmission to TB in England. The recently launched Tuberculosis Strategy for
England (20152020) outlines the need to improve TB
control in England to prevent transmission including
enhanced contact tracing and interventions to reduce
diagnostic delay.
S255
07. The rainbow session: treatment

vaccination, surveillance or tuberculosis
and everything in between
EP-153-05 Clinical profile and outcome of extrapulmonary tuberculosis: a 3-year study from
south Kerala
J Cherian,1 J Subramoniyapillai,2 K Periasamy,1
P Dhayakar,1 P Sadasivanpillai3 1Government Medical
College Thiruvananthapuram, Trivandrum, 2State
Tuberculosis Cell, Trivandrum, 3T.D. Government Medical
College, Alappuzha, India. e-mail: dr.jjcherian@gmail.com
Background: Although the overall tuberculosis (TB)

notification has decreased recently, extrapulmonary
tuberculosis (EPTB) is on the rise. Various factors
influencing the site of involvement and treatment
outcome are age, sex, HIV status and geographical
location. The objective of this study was to analyse the
treatment outcome of various sites of EPTB and the
factors influencing it. Methodology The study included
patients diagnosed with EPTB from the southern seven
districts of Kerala, between January 2010 and December
2012. Patients who registered for treatment under
Revised National Tuberculosis Control Programme
(RNTCP) and administered Directly Observed Treatment Short course (DOTS) were chosen. Data regarding
their demographic profile, type of patient, Category of
treatment, HIV status, site of involvement and treatment
outcome was collected from the treatment registers. The
variables unavailable in the registers were recorded as
data unavailable.
Results: Of the 39 267 patients registered for treatment
for TB, EPTB constituted 8610 (22%) cases. 53.1% were
males. 51.6% of the cases belonged to the age group 30
60 years. 90.7% of cases were newly detected and most
of the cases (84.9%) received Category 1 regimen. Only
53.6% of the 8610 cases had their HIV status tested and
among them 5.9% (n 271) were positive. The most
common site of involvement was lymph node (31.5%),
followed by pleural effusion (18.6%), bone & joint
(8.7%), intestine (5.1%), central nervous system (CNS)
(4%) and genitourinary (GU) (1.8%). Other less
common sites constituted 5.8%. Data regarding site
was unavailable for 24.5% cases. 87.4% completed
treatment, 4.3% defaulted, 3.7% died and 2.3%
transferred out. Eleven patients failed treatment, three
patients were shifted to MDR TB regimen and 190
patients did not have their outcome available. Factors
that were associated with poor treatment completion
rates included male gender (P , 0.001), age over 60 years
(P , 0.001) and a positive HIV status (P , 0.001). The
best treatment completion rates were seen in GU TB
(92.3%) followed by lymph node TB (90.4%). CNS TB
had the highest mortality (14.5%) and default (5.5%)
rates.
Conclusion: The treatment completion rates among
EPTB patients in this study were remarkable (87.4%).
However, the partial incompleteness of data was a
disadvantage while drawing conclusions on different
S256
sites of involvement. The rising proportion of EPTB

demands special guidelines from RNTCP to cater to this
growing subset.
urgently addressed by the National Tuberculosis Programme.
Table Treatment outcomes of various sites of EPTB
EP-155-05 Predictors of mortality and

immediate causes of death in hospitalised TB
patients undergoing an intensive phase of
treatment in Kigali University Teaching
Hospital
Treatment Completed
Yes
Site
Lymph node
Pleural Effusion
Bone and Joint
Genitourinary
CNS
Intestine
Others
Data unavailable
Total
No
Total
2450
1379
640
143
251
372
428
1847
7510
90.4
86.2
85.1
92.3
73.0
84.9
85.6
87.5
87.2
260
220
112
12
93
66
72
265
1100
9.6
13.8
14.9
7.7
27.0
15.1
14.4
12.5
12.8
2710
1599
752
155
344
438
500
2112
8610
EP-154-05 Risk factors of loss-to-follow-up and

death among tuberculosis patients receiving
treatment in a resource-limited setting
I Alobu,1 S Oshi,2 D Oshi,2 K Ukwaja3 1National
Tuberculosis and Leprosy Control Programme, Ministry of
Health, Abakaliki, Ebonyi State, 2Centre for Development
and Reproductive Health, Enugu, 3Department of Internal
Medicine, Federal Teaching Hospital, Abakaliki, Ebonyi State,
Nigeria. e-mail: ukwajakingsley@yahoo.co.uk
Background: Treatment interruptions and death during

care are major challenges facing tuberculosis control.
The aim of this study was to describe the epidemiological
characteristics, timing and determinants of being lost-tofollow-up or death among adult tuberculosis patients in
Nigeria.
Design/Methods: Using routine surveillance data, we
conducted a retrospective cohort study of adult tuberculosis patients treated during 2011 and 2012 in two large
health facilities in Ebonyi State, Nigeria. Multivariable
logistic regression analyses were used to identify independent predictors for loss-to-follow-up and death
during treatment.
Results: Of 1668 treated patients, the rate of loss-tofollow-up was 157 (9.4%), whilst 165 (9.9%) died. Also,
35.7% (56) of the patients with loss-to-follow-up and
151 (91.5%) of deaths occurred during the intensive
phase of treatment. Risk of loss-to-follow-up increased
with increasing age (adjusted odds ratio (aOR) 1.2; 95%
confidence interval (CI) 1.11.9), smear-negative TB case
(aOR 2.3; CI 1.53.6), extrapulmonary TB case (aOR
2.7; CI 1.35.2), and patients who received the longer
treatment regimen (aOR 1.6; 1.12.2). Risk of death was
highest in extrapulmonary TB (aOR 3.0; CI 1.46.1) and
smear-negative TB cases (aOR 2.4; CI 1.73.5), rural
residents (aOR 1.7; CI 1.22.6), HIV co-infected (aOR
2.5; CI 1.73.6), not receiving antiretroviral therapy
(aOR 1.6; CI 1.12.9), and not receiving cotrimoxazole
prophylaxis (aOR 1.7; CI 1.22.6).
Conclusion: Targeted interventions to improve treatment
adherence for patients with the highest risk of loss-tofollow-up or death are urgently needed. This needs to be
O Manzi,1 L Bitunguhari,1 J Clerinx2 1University of

Rwanda, Kigali, Rwanda; 2Institute of Tropical Medicine,
Antwerp, Belgium. e-mail: oliviermanzi@yahoo.fr
Background: Tuberculosis (TB) and HIV coinfection are

major causes of mortality in Africa, and inter-related.
Predictors of mortality are not well characterised,
especially during early inpatient treatment. We aimed
to examine predictors of mortality in these patients in
Rwanda.
Design/Methods: We conducted a prospective observational cohort study for consenting adult hospitalised TB
patients diagnosed from January 2013 to December
2013; we assessed predictors of mortality among
demographic, clinical, laboratory and medical imaging
data. Immediate causes of death were constructed from
analysis of cardinal symptoms in the week prior to death.
Results: A total of 152 hospitalized TB patients (of 208
recruited) had full data available for analysis. HIV
coinfection was present in 95 (62%), and 49 died during
followup (overall mortality rate 32%). Pure pulmonary
TB affected 31 (20%) patients, whereas extrapulmonary
TB was seen in 84 (55%) and mixed TB in 54 (36%).
Auramine stains were positive in 47/108 (44%) patients.
In univariable analysis, dyspnea, abdominal tenderness,
low platelet count, low oxygen saturation, TB pericarditis, disseminated TB and miliary TB were associated
with death. In a multivariable analysis, only thrombocytopenia , 150000/lL (HR4.08, P , 0.001; 95%CI
1.827.31) and dyspnea (HR 3.65, P 0.009; 95%CI
1.319.24) were independent predictors of mortality.
HIV status, low CD4 count and high abdominal
ultrasound lesion score were not associated with excess
mortality. Septic shock was observed in 61% of patients
who died. Half of TB associated mortality occurred
within 15 days from TB diagnosis.
Conclusion: In hospitalised TB patients in Rwanda,
extrapulmonary TB is the main clinical presentation,
disseminated TB disease is frequent. Mortality is high,
occurs early after diagnosis, and is not associated with
HIV status or a low CD4 count. Thrombocytopenia and
dyspnoea are independent predictors of mortality. Septic
shock is the main clinical event preceding death.
EP-156-05 Prevalence of non-tuberculosis

mycobacteria in patients with suspicion of
tuberculosis in a high tuberculosis burden
country: a prospective cohort study
M Bonnet,1,2 Y Pho,1 J-P Dousset,3 S Heng,4 S Godreuil,5
WW Yew,6 S Kim Chamroeun,3 C Hewison7 1Epicentre,
Paris, 2Institut de Recherche pour le Developpement,
`
Montpellier, France; 3Medecins
Sans Frontieres
(MSF), Phnom
Penh, 4Institut Pasteur Cambodge, Phnom Penh, Cambodia ;
5
INSERM U1058, Monptellier, France; 6Hong Kong
Tuberculosis, Chest and Heart Diseases Association, Hong
Kong, China; MSF, Paris, France. e-mail:
Background: Prevalence and case management of nontuberculosis mycobacteria (NTM) are poorly documented in high tuberculosis (TB) burden countries. We present
the prevalence, risk factors and clinical implications of
NTM in the district of Kampong-Cham, Cambodia.
Methods: Consecutive adult patients with clinical
suspicion of TB were enrolled in a prospective cohort
with baseline epidemiological, clinical, radiological and
sputum mycobacterial investigations, and 12 months
follow-up of patients with NTM isolates. Culture used
Lowenstein-Jensen
and MGIT and identification used the
GenoTypew Mycobacterium CM/AS assay. An advisory

committee classified patients as NTM lung disease or
colonisation using the 2007 American Thoracic Society
guidelines and advised on NTM treatment.
Results: Of 1188 patients (51.4% female, 54 years
median age), 218 (18.3%) were diagnosed with Mycobacterium tuberculosis and 124 had NTM isolates
(10.4%). Among them, 11 were classified as NTM lung
disease (8.6%) and 113 as colonisation (87.8%).
Amongst smear positive patients, 10/196 (5.1%) grew
only NTM. Most common NTM were M. fortuitum
(23.4%), M. intracelluare (16.1%), M. abscessus
(11.3%), M. scrofulaceum (10.5%) and M. gordonae
(10.5%). HIV-infection (aOR 3.4, 95%CI [1.29.8]) and
past TB history (aOR 1.6, 95%CI [1.12.7] were
associated with isolation of NTM. Patients with NTM
lung disease were more likely to have past TB history
(63.4% vs. 6.0%), history of hospitalisation for respiratory disease (54.5% vs. 6.4%), HIV infection (27.3% vs.
1.1%), low body mass index (median 14.8 vs. 18.1 Kg/
m2) and bronchiectasis (60.0% vs. 13.4%) or pleuropulmonary sequelae (60.0% vs. 31.9%) on chest X-ray
compared to TB patients. Eight patients were started on
NTM treatment (5 cured, 1 death and 2 still on
treatment) and 8/116 (6.9%) non-treated patients died
during follow-up.
Conclusion NTM represent more than 1/3 of sputum
mycobacterial growth from TB suspects. The 5% of
smear-positive patients due to NTM are at risk of being
over treated for TB. Post tuberculosis lung sequelae and
HIV infection are important NTM predisposing factors.
Decision to start NTM treatment should be rapid and
systematic in advanced HIV-infected patients. However
in HIV-negative patients, the decision for starting
treatment also depends on the NTM species and host
co-morbidity, especially bronchiectasis. Addressing ben-
S257
efit versus risk of therapy, notably with age consideration

and medication tolerability is warranted.
EP-157-05 Abrupt decline in tuberculosis

among foreign-born persons in the United
States
B Baker,1 C Winston,1 Y Liu,2 A France,1 K Cain1 1U.S.
Centers for Disease Control and Prevention (CDC), Division of
Tuberculosis Elimination, Atlanta, GA, 2CDC, Division of
Global Migration and Quarantine, Atlanta, GA, USA. e-mail:
izj4@cdc.gov
Background: While tuberculosis (TB) cases in the USA

have declined over the past two decades, TB morbidity
among the foreign-born has remained persistently
elevated. A recent abrupt decline in TB cases among
foreign-born persons provided an opportunity to examine contributing factors and inform future TB control
strategies. We investigated the relative influence of four
factors upon the decline: 1) changes in the foreign-born
population through immigration and emigration, 2)
changes in distribution of country of origin among the
foreign-born population, 3) changes in the TB case rate
among foreign-born subgroups, and 4) changes in the
extent of TB transmission.
Design/Methods: Using data from the U.S. National
Tuberculosis Surveillance System and the American
Community Survey during 2000 to 2013, we examined
TB case counts, TB case rates, and population estimates,
stratified by years since U.S. entry and country of origin.
Joinpoint regression modeling was used to assess
significant changes in trend. We estimated the proportion
of the case count decrease due to changes in TB case rate
versus population shifts. We estimated the number of
cases due to recent transmission in the United States
using TB genotyping data.
Results: A 39.5% decrease in cases among recent
entrants (,3 years since U.S. entry), beginning in 2007
and ending in 2011 (P , 0.05), accounted for 69.6% of
the overall decline (Figure). Most (80.7%) of the decline
among recent entrants from Mexico was due to a
decrease in population, but the decline in cases among
recent entrants from the other top five countries of origin
was almost exclusively (95.5100%) due to decreases in
TB case rate. Among recent entrants, 6.3% of cases were
likely due to recent transmission. Among non-recent
entrants, an 8.9% decrease in cases was entirely (100%)
due to a decrease in TB case rate and accounted for
30.4% of the overall decline.
Conclusion: Both recent and non-recent entrants contributed to the decrease in TB cases. The factors
contributing to the decline among recent entrants varied
by country of origin, but the decline among non-recent
entrants was entirely due to a decrease in TB case rate.
Strategies that address recent and non-recent entrants
(e.g., investment in overseas TB control, testing and
treatment of latent TB infection for high-risk subgroups
among non-recent entrants) will be necessary to accelerate the decline in TB among foreign-born persons in the
United States.
S258
EP-158-05 Trends of tuberculosis treatment

outcome following community involvement in
an under-resourced setting: the case of Bauleni
compound, Lusaka, Zambia
Treatment outcomes were categorised according to the

national TB control program (NTP) guidelines. Microsoft Excel software was used to manage the data
Results: The annual average treatment success rate
increased from 47.0% in 2007 to 82.3% in 2013, whilst
the rate of persons lost to follow up declined from 45.3%
in 2007 to 0% in 2013. The death rate reduced too from
4.5% in 2007 to 1.8% in 2013. Compared to the values
of the preceding years, treatment success showed a
consistent upward trend from 2007 (47.0%) to 2013
(82.25%)
Conclusions and key recommendations: After the implementation of the project, the treatment success rate of
new smear-positive TB patients improved significantly.
Monitoring, supervision and home visits and community
education by TS may have contributed towards the
achievement of 85% treatment success target and
declining default and death rates.
V Mfungwe,1 M Ota,1 Y Onoe,1 G Samungole,2,3

M Masaninga2 1Research Institute for Tuberculosis (RIT)/
Japan Anti Tuberculosis Association (JATA) Zambia, Lusaka,
2
Ministry of Community Development, Mother and Child
Health, Lusaka, 3Ministry of Community Development,
Mother and Child Health, Lusaka, Zambia. e-mail:
mfungwev@gmail.com
Background and challenges to implementation: Supervised tuberculosis (TB) treatment, which may have to
include direct observation of therapy (DOT), helps
patients take their drugs regularly and complete treatment, thus achieving cure and preventing the development of drug resistance TB. Supervision may be
undertaken at a health facility, in the workplace, in the
community. It should be provided by a treatment partner
or supporter who is acceptable to the patient and is
trained and supervised by health services. In Zambia,
standardised tuberculosis prevention and control programme, incorporating Directly Observed Treatment,
Short Course (DOTS) started in 1993. In 2006 and 2007,
the treatment success rate of TB patients was unsatisfactorily low (47.0%) in Bauleni, one of the urban slums
located in Lusaka urban district. A high proportion of
this was persons lost to follow up (45.3%), which is a
serious public health concern that needs to be addressed
urgently.
Intervention or response: We embarked on a 6-year
project to improve access of the community members to
TB and TB-HIV co-infection case finding and patient
care and support in Bauleni by training health workers
and involving the community through the training of TBHIV treatment supporters (TS) in community sensitization, case finding, patient care and support. This study
investigated the outcomes of TB treatment in Bauleni to
measure the impact of the implementation of the project.
We analyzed the records of all new smear-positive TB
patients registered in Bauleni from January 2007 to
December 2013 based on the health centre TB register.
EP-160-05 The duration of protection of infant

BCG vaccination in England
P Mangtani,1 R Keogh,1 P Nguipdop-Djomo,1 L Trinder,1
L Rodrigues1 1London School of Hygiene & Tropical
Medicine, London, UK. e-mail:
punam.mangtani@lshtm.ac.uk
Background: Studies in Brazil and American Indian

populations in the USA suggest duration of protection
from BCG lasts several decades but evidence is scarce and
relevance to other settings uncertain. In the UK, the BCG
vaccine was given to school children, with selected
vaccination to high-risk, usually ethnic minority, infants
since the 1980s, covering half the health districts by
1992. School vaccination was replaced in 2005 by a
national programme of BCG vaccination in infancy for
high-risk populations.
Design/Methods: We carried out an observational study
in England of cases of tuberculosis and frequency
matched population based controls in minority ethnic
groups aged 019 years. Trained interviewers conducted
face to face interviews and requested permission to
examine upper arms for a BCG scar. Records of past
BCG vaccination are known to be incomplete and BCG
scar is a traditional way to establish past BCG
vaccination with high validity as an indicator of BCG
vaccination in many populations. Information on potential confounders was collected including demographic
and social variables. A complete case analysis is currently
being carried out before multiple imputation for missing
scar information. Results of vaccine effectiveness by

small periods of time since vaccination will be examined
and presented, starting with the first 5 years and
compared with the well established protective effect in
the literature to assess robustness of the study.
Results: From 2012 to 2014 we recruited over 720 cases
and 720 controls both with a response rate of over 60%.
Interviewers noted a scar in 56% of cases and 65% of
controls but with missing information in 16% and 10%
respectively. Patterns of missingness indicated teens and
young adults females of Bangladeshi and Pakistani
women were less willing than other groups to be
examined for a BCG scar. Initial complete case analyses
suggested a protective effect in the first 5 years since
vaccination.
Conclusion: We will present new evidence on how long
BCG protects against disease which is important to
reduce uncertainty. It will support appropriate decisions
on any modification of future TB vaccine programmes eg
universal infant BCG vaccination may be more economic
than assumed. New improved vaccines may also need to
show they offer protection against pulmonary disease
greater than that offered by BCG alone.
08. High-definition insights into drug

resistant tuberculosis
EP-161-05 Specific gyrA gene mutations predict
poor treatment outcome in drug-resistant
tuberculosis
L Rigouts,1,2 N Coeck,1,2 G Gumusboga,1 A Maug,3
MA Hossen,3 H L Rieder,4 B De Jong,1,5,6 A Van Deun1,7
1
Institute of Tropical Medicine, Antwerp, 2Biomedical
Sciences, Antwerp University, Antwerp, Belgium; 3Damien
Foundation, Dhaka, Bangladesh; 4University of Zurich,
Zurich, Switzerland; 5Medical Research Council, Banjul,
Gambia; 6New York University, New York, NY, USA;
7
Paris, France. Fax: (32) 476 333. e-mail: lrigouts@itg.be
Background: Increasing evidence points to the role of

specific mutations in the gyrase genes of M. tuberculosis
in phenotypic resistance to fluoroquinolones. However,
the predictive value of these markers for clinical
outcomes in multi-drug resistant patients is to date
unknown. In this study, we aimed at determining
molecular markers and breakpoints predicting secondline treatment outcomes in M. tuberculosis patients
treated with fluoroquinolones.
Design/Methods: We analyzed treatment outcome data
in relation to the gyrA and gyrB sequences, and minimal
inhibitory concentrations for ofloxacin, gatifloxacin and
moxifloxacin in pre-treatment Mycobacterium tuberculosis isolates from 181 multi-drug resistant tuberculosis
patients in Bangladesh whose isolates were sensitive to
injectable drugs.
Results: Resistance levels for ofloxacin were systematically higher relative to moxifloxacin and gatifloxacin.
The gyrA 90Val, 94Gly and 94Ala mutations were most
S259
frequent, with the highest resistance levels for 94Gly

mutants. Increased pre-treatment resistance levels correlated with specific mutants, were associated with lower
cumulative cure percentages, with no more cure among
cases found resistant at 4 lg/ml. Any gyrA 94 mutation,
except 94Ala, showed a significantly lower proportion of
cure compared to all other gyrA mutations taken
together (all non-94 mutants 94Ala) (OR 4.2).
Conclusion: Our study suggests a breakpoint of 2 lg/ml
on Lowenstein-Jensen
and mutations at position 94 other
than Ala to indicate high level resistance to gatifloxacin

or moxifloxacin, predictive of poor treatment outcome in
multi-drug resistant tuberculosis patients in whom the
injectable agent is still effective.
EP-162-05 Contribution of efflux pumps to

rifampicin resistance in clinical isolates of M.
tuberculosis
A Narang,1 S Porwal,1 K Garima,1 A Bhandekar,1
K Shrivastava,1 R Prasad,1 M Bose,1 M Varma Basil1
1
Vallabhbhai Patel Chest Institute, Delhi, India. Fax: (911)
2766 7420. e-mail: mandirav@rediffmail.com
Background: Mutations in a specific 81-base pair region

of the rpoB gene, the Rifampicin (RIF) resistance
determining region (RRDR), are responsible for 95%
of rifampicin-resistant strains. Approximately 5% of
clinical RIF-resistant M. tuberculosis isolates do not
harbor mutations in the RRDR region, suggesting the
existence of other mechanisms that may play a role in
development of drug resistance. The role of efflux pumps
in facilitating the emergence of resistance is now being
recognised as a cause of drug resistance in M. tuberculosis. The present study investigated the expression of
putative efflux related genes under RIF stress to
investigate their role in resistance to RIF in M.
tuberculosis isolates obtained from patients of pulmonary tuberculosis.
Design/Methods: We investigated the role of four
putative efflux pump genes Rv0676, Rv0507, Rv1250
and Rv0194 in conferring RIF resistance to clinical
isolates of M. tuberculosis. The study included 10 RIF
susceptible and 19 RIF resistant isolates. The RRDR of
the isolates was sequenced. The RIF resistant isolates
included seven isolates with no mutations in the RRDR.
The Minimum Inhibitory Concentration (MIC) of the
isolates was analysed by Microplate Alamar Blue Assay
and the isolates were exposed to subinhibitory concentrations of RIF for 24 hours. The expression of efflux
genes was studied by Real Time Analysis using sigA as a
normalizer and the results correlated.
Results: Amongst the 19 RIF resistant isolates, Rv0676c,
Rv0507, Rv1250 and Rv0194 were upregulated in 10, 5,
7 and 10 isolates respectively. Five of the seven RIF
resistant isolates, selected due to their wild type
sequence, overexpressed one or more efflux pump genes.
The MICs of all these five isolates were .1lg/ll. The
efflux gene RV0194 was overexpressed in all these five
wild type RIF resistant strains. Of the eleven RIF
resistant M. tuberculosis isolates with mutations at the
S260
531 codon, 7 overexpressed Rv0676c and four of these

had high level resistance (MIC.64 lg/ll). Rv0194 was
upregulated in 4 isolates with 531 mutations. An isolate
with mutation at 511 codon overexpressed Rv0676c,
Rv1250 and Rv0194. The efflux genes were upregulated
in only one of the RIF susceptible isolates, though the
difference between the resistant and susceptible isolates
was not statistically significant.
Conclusion: Efflux pumps may be upregulated as an
initial response to drug stress in an individual and may
thus play a role in RIF resistance.
EP-163-05 Use of whole genome sequencing to

identify mutations related to phenotypic
resistance to delamanid and bedaquiline in M.
tuberculosis
A Cabibbe,1 S Battaglia,1 E Schena,1 E Borroni,1
P Miotto,1 D M Cirillo1 1San Raffaele Scientific Institute,
Milan, Italy. e-mail: battaglia.simone@hsr.it
The rise of multi- and extensively- drug resistant

tuberculosis (M/XDR-TB) is reducing the spectrum of
available drugs for treatment, thus highlighting the need
of new effective compounds. Among new anti TB drugs,
delamanid (DLM) and bedaquiline (BDQ) are conditionally approved. Mutations responsible for BDQ-R are
found in atpE gene. However, other spontaneous BDQ-R
mutants harbour mutations in the non-target Rv0678
gene, a negative regulator of the MmpS5-MmpL5 efflux
pump. These mutations may be linked to clofazimine
resistance. DLM is a pro-drug requiring intracellular
activation by Rv3547 protein. Mutations in Rv3547
gene or in fbiA, fbiB, fbiC and fgd1 genes (F420
biochemical pathway) are involved in DLM-R phenotype. Taking advantage of Whole Genome Sequencing
(WGS) approach, we analysed 9 genes putatively
involved in BDQ-R and DLM-R on 236 M. tuberculosis
isolates (96 susceptible to rifampicin (RIF) and isoniazid
(INH), 11 RIF-R/INH-S, 96 MDR, 11 MDRAG-R, 12
MDRFQ-R and 10 XDR) representing 14 different
lineages. We focused on atpE, Rv0678, mmpL5, mmpS5
genes for BDQ-R; ddn, fgd1, fbiA, fbiB and fbiC genes
for DLM-R. We identified several new non-synonymous
mutations in genes putatively involved in BDQ-R. In
particular, although no mutations were observed in atpE,
the other analysed genes carried substitutions: 7 strains
have mutations in Rv0678 gene [V3I, N4T (2), M17V,
L32S, G87R, R134G]; 18 in mmpL5 [P70L, D128N (3),
S325G, A349T, E552G, L634Q, F696L (8), Q702K,
Q707K]; 2 in mmpS5 (G109S). We observed mutations
possibly involved in DLM-R in: Rv3547 (3 isolates:
R72W, E83D, W88STOP); fgd1 (1: G314E); fbiA [8:
Q120R (4), T302M (4)]; fbiB [64: F220L (56), D315A,
L447R, K448R (6)]; fbiC [6: V41M, T273A (4), I693V].
From this data subset, we didnt observe any correlation
among mutations in candidate genes for BDQ-R or
DLM-R and the resistance pattern or genotype of strains
included in the study. Previously reported lineage-specific
substitution in fgd1 (K270M: Haarlem; K296E: West
Africanum 2) and mmpL5 (D767N: Beijing) were also
detected. In addition to the phenotypic testing performed

on a subset of 17 strains for DLM, drug susceptibility
testing is under investigation on the other strains to better
understand the role of these mutations in the emergence
of BDQ-R and DLM-R. The WGS approach showed its
usefulness in pre-treatment simultaneously analysing of a
wide range of DR-related genes, reducing time for
analysis of mutations linked to TB drug-resistance.
EP-164-05 Characterisation of resistance to

rifampicin and isoniazid among Beijing
Mycobacterium tuberculosis isolates in
Kazakstan
U Kozhamkulov,1 A Akhmetova,1 V Bismilda,2
L Chingisova,2 A Akilzhanova1 1Center for Life Sciences,
Nazarbayev University, Astana, 2National Center for
Tuberculosis Problems, Almaty, Kazakhstan. e-mail:
kzulan@mail.ru
Background: The incidence of tuberculosis in 2013 was

73.4 cases per 100 000 people, and the mortality rate was
5.6 per 100 000 people in Kazakhstan. Beijing/W family
strains of Mycobacterium tuberculosis associated with a
high risk of drug resistance are widely distributed in
many parts of the world, including areas where it was not
found previously. The aim of this study is characterization drug resistance to the rifampicin and isoniazid
among Beijing/W family M. tuberculosis clinical isolates
collected in Kazakhstan.
Design/Methods: The M. tuberculosis strains collected
for this study were isolated from new TB cases patients in
different regions of Kazakhstan. Spoligotyping was
performed on DNA by using the standard method
(Kamerbeek et al., 1997) using a reverse dot-blot
spoligotyping commercially available kit (Ocimum Biosolutions Inc) with positive controls (M. tuberculosis
strain H37Ra and M. bovis BCG ) and negative control
without DNA. Drug susceptibility of M. tuberculosis to
isoniazid and rifampicin was determined using the
absolute concentration method according to WHO
recommendation. Genetic determination of M. tuberculosis drug resistance was identified through amplification
and sequencing of the rpoB and katG gene regions and
promoter regions of mabA (fabG)-inhA, oxyR-ahpC
operons by automated ABI 3730 (Applied Biosystems)
Genetic Analyzer, according to the manufacturers
instructions.
Results: The structure of the DR-region was determined
by spoligotyping and identified 117 W-Beijing M.
tuberculosis isolates including 68 (58.12%) MDR
isolates, 30 (25.64%) susceptible isolates and 19
(16.24%) resistant other than MDR isolates. 69 M.tuberculosis clinical isolates were investigated to DNA
sequencing analysis of the hypervariable (hot-spot) rpoB
region and mutations in 4 codons were identified. Most
of them were nucleotide substitutions in 531 codon
(95.65%), especially Ser531Leu mutation (92.7%).
Mutation in katG associated with a resistance to
isoniazid among 78 isoniazid-resistance strains showed
prevalence mutation in 315 codon (Ser315Thr substitu-
tion). This type of mutation noted in 98.71% cases. In

one case detected mutation in 8 (T-C) position of
promoter region mabA (fabG)-inhA.
Conclusion: Prevalence of MDR and other drug resistant
clinical isolates (58.12% and 16.24%) among Beijing
family strains of M. tuberculosis from different regions of
Kazakhstan was shown.
EP-165-05 Sequencing-based detection of rpoB

gene mutations in Mycobacterium tuberculosis
strains from Malawi
T Chikaonda,1,2 I Kets,2 N Nguluwe,1 I Thengolose,1
F Nyakwawa,3 W Stevens,2 L Scott,2 M Hosseinipour1
1
UNC Project, Lilongwe, Malawi; 2University of the
Witwatersrand, Johannesburg, South Africa; 3National
Tuberculosis Control Programme, Lilongwe, Malawi. Fax:
(265)1755954. e-mail: tarchikaonda@gmail.com
Background: Detection of resistance to anti-tuberculosis

drugs remains a significant challenge in Malawi due to
limited diagnostics. New diagnostics, like the Xpert can
both detect Mycobacterium tuberculosis and rifampin
(RIF) resistance in a single, rapid assay. RIF resistant TB
(RR-TB) has not been well studied in Malawi. In order to
understand mechanisms of RIF resistance, we characterized 43 isolates from adult patients (718 years) for
mutations in the 81bp region (codons 507 533) of the
rpoB gene by DNA sequencing.
Methods: Sputum pellets were randomly selected from
specimens sent for Drug Susceptibility Testing (DST) at
National TB Reference Laboratory (NTRL) for processing at UNC Project laboratory. Pellets were re-suspended
in 1.5ml phosphate buffer and an aliquot of 0.5ml was
processed on Xpert. Another 0.5ml was cultured on
MGIT 460 system. DNA was extracted from corresponding positive cultures of TB samples identified as
RIF resistant by Xpert for sequencing in an automated
DNA sequencer ABI Prism 3700 using BigDye terminator kit V3.1
Results: Mycobacterial strains (n 43) defined as RR-TB
by Xpert were eligible for sequencing. To confirm RIF
resistance, a 450bp fragment covering the core region of
the rpoB gene was sequenced and yielded seven different
mutations in 37/43 (86%) in codons 511, 512, 513, 516,
522, 526 and 531. The most common mutations were in
codons 526 (38%), 531 (29.7%) and 516 (16.2%). No
insertion, deletion or double mutations were observed.
However, mutations were not found in 6/43 (14%) of the
isolates, but were detected by Xpert probe B (4/6) and by
probe E (2/6). Complete drop out of probe B and E was
observed in one sample on either probe (1/4 and 1/2
respectively). Cycle threshold (Ct) values between 17.0
and 32.7 were observed in the remaining samples (4/6).
Conclusions: This is the first comprehensive molecular
analysis of rpoB gene in RR-TB isolates from patients
known to be at increased risk of drug resistance in
Malawi. All isolates had one of the previously described
rpoB mutations, containing nucleotide changes. No new
mutations were identified in the current study. Interestingly, mutations were not detected in six strains despite
being RIF resistant on Xpert. Resistance mechanism
S261
remains unknown in isolates lacking rpoB gene mutations observed in this study. Further molecular cluster
analysis such as spoligotyping and DNA finger printing is
required to determine transmission dynamics of the
mutated strain.
EP-166-05 Genetic markers of drug resistance

and their association with clinical outcomes in
patients with MDR-TB or XDR-TB
J Limberis,1,2,3 E Pietersen,1,2,J Jayakumar 1,2,3
G Theron,1,2,3 K Dheda,1,2,3 L Smith,1,2,3,4 R M Warren,4
T Clark5 1University of Cape Town, Cape Town, 2University of
Cape Town, Cape Town, 3University of Cape Town, Cape
Town, 4University of Stellenbosch, Cape Town, South Africa.
5
e-mail: lmbjas002@myuct.ac.za
Background: Mycobacterial genetic associates of resistance to second line drugs and poor clinical outcomes
[~75% of extensively drug resistant (XDR-) TB patients
die within 5 years] are poorly understood.
Methods: We conducted a 5-year prospective study on
the outcome of XDR-TB patients in the Western Cape,
South Africa. Diagnostic culture isolates underwent
phenotypic susceptibility testing (DST), IS6110-RFLP
strain typing, and Illumina whole genome sequencing
(WGS). A bioinformatics pipeline was constructed and
the presence of high confidence resistance conferring
SNPs (listed in TBDream) described. Treatment outcomes were classified as favorable or unfavorable.
Results: Of the 145 isolates with WGS data, 112, 28, 3,
and 2 isolates had XDR-TB, pre-XDR-TB, MDR-TB and
mixed second line resistance, respectively. Compared to
phenotypic DST, sequencing for resistance had a
sensitivity of 99%, and 94% for resistance to the first
line drugs and the second line drugs, respectively.
Missense mutations were rarely identified in genes
related to new or repurposed drugs (9 for both delamanid
and pretomanid [none were in the high confidence
delamanid resistance gene, ddn], 4 for SQ109, 1 for
both bedaqualine and clofazamine, and none for linezolid; Figure). 49 isolates had a ponA1 fitness mutation
and none had rpoB compensatory mutations. 51/128
strains with a floroquinolone resistance-conferring mutation in gyrA had an rpoB mutation known to lower in
vitro fitness cost. Inter-isolate comparison showed 1% of
isolates with identical IS6110-RFLP patterns to have a
different SNP barcode, while 26% of isolates with
identical SNP barcodes had different IS6110-RFLPs.
22/129 patients had a favorable outcome. 93 SNPs were
more frequent (P 6 0.05) in patients with a favorable
outcome, 20 of which were in virulence-associated genes.
Conclusion: There was excellent concordance between
genotypic DST and phenotypic DST suggesting that
sequencing is likely to be useful clinically. Moreover,
SNP-based strain typing had high agreement with a
traditional fingerprinting method, thereby enabling
tracking of strains. The absence of mutations in genes
previously described to be associated with resistance to
newly developed drugs holds promise for the treatment
of drug-resistant TB. Known compensatory mutations
S262
associated with drug resistance were rare. We also

identified a preliminary SNP signature encoding genes
involved in fundamental cellular functions associated
with favorable outcome.
EP-167-05 Whole genome sequencing reveals

accumulation of single nucleotide
polymorphisms (SNPs) associated with
treatment failure
L Malinga,1 J Brand,1 T Abeel,2 A Earl,2 M Van Der Walt1
1
Medical Research Council, Pretoria, South Africa; 2Broad
Institute, Cambridge, MA, USA. Fax: (27) 12 324 7922.
e-mail: lesibana.malinga@mrc.ac.za
Background: Multidrug drug resistant tuberculosis

(MDR-TB) is a serious challenge, especially in high
burden countries such as South Africa. Diagnosis and
treatment is often delayed and this can lead to poor
treatment outcomes. Due to ineffective treatment,
Mycobacterium tuberculosis genome mutates and leads
to bacterial fitness under antibiotic stress. Accumulation
of short nucleotide polymorphisms (SNPs) could be
responsible for treatment failure. We used whole genome
sequencing (WGS) to analyse accumulation of drug
target and compensatory SNPs between serial isolates of
longitudinal cohorts failing treatment.
Design/Methods: MDR-TB strains were collected from
46 patients on second-line drug treatment in KwaZuluNatal (KZN) and Eastern Cape (EC) between 2005 and
2009. All patients had phenotypically determined MDRTB at baseline and had treatment outcomes documented.
Each strain had baseline and at least one strain collected
on follow up. From each strain, M. tuberculosis DNA
was extracted from colonies grown on Lowenstein
Jensen slants, and fragment and jumping paired-end
Illumina DNA libraries were constructed and sequenced

on the Illumina HiSeq 2000 (Broad Institute in Cambridge, MA). Sequences were aligned to H37Rv genome
and Pilon was run to generate a list of SNPs. Digital
spoligotyping was performed to a database 43 unique
spacer sequences.
Results: 92 sequences from 46 serial isolates of MDR-TB
patients from KZN (29) and EC (17) were analysed.
Among 46 patients, 11 (28%) had treatment success (i.e
cured or completed), 29 (63%) failures (i.e., died or
defaulted) and four (13%) were lost to follow-up.
Phylogenetic reconstruction revealed that primary genotype differed by province. The Beijing genotype was
predominant in EC while EuroAmerican lineage (S, T,
LAM, X) found in KZN. There was no difference in
treatment outcomes between Beijing and EuroAmerican
strains (55.2% vs. 44.8%, P 0.3). 76% of cases failing
treatment already had 75 SNPs at baseline. Among
treatment failure cases, 42% patients accumulated 1-2
SNPs on follow-up isolates. We further identified positive
epitasis between drug target and compensatory regions.
Conclusion: Continued evolution through SNPs accumulation and transmission of strains contributes to the
MDR-TB genetic population fitness in these regions. The
clinical significance of WGS for early detection of SNPs
predictive of treatment failure needs further investigation.
EP-168-05 Use of DNA from AFB smears to

perform high-throughput sequencing based
surveillance for drug-resistant Mycobacterium
tuberculosis in Afghanistan
J Mancuso,1 M Rowneki,2 N Hamraz,3 P Du,2 G Davis,4
R Blakemore,5 D Alland,2 N Aronson1 1Uniformed Services
University, Bethesda, MD, 2Rutgers New Jersey Medical
School, Newark, NJ, USA; 3National Military Hospital, Kabul,
4
Combined Security Transition Command-Afghanistan,
Kabul, Afghanistan; 5Sackler School of Graduate Biomedical
Sciences, Tufts University, Boston, MA, USA. e-mail:
james.mancuso@usuhs.edu
Background: Although Afghanistan is one of the 22

countries with the highest burden of tuberculosis (TB),
little is known about the prevalence of the drug-resistant
forms of TB. Since understanding the prevalence of
resistance to drugs such as pyrazinamide and the
fluoroquinolones is critical to devising new treatment
regimens, surveillance of resistance trends must expand
to cover more drugs. The objective of this study was to
evaluate the prevalence of drug-resistant tuberculosis in
Afghanistan using molecular techniques that can be
adopted to detect resistance to a wider range of drugs.
Design/Methods: We obtained slides identified as positive for acid-fast bacilli during routine clinical practice
from two provinces in Afghanistan. Slides were re-scored
by study personnel according to the WHO/IUATLD
standard semi-quantitative grading system. We extracted
DNA from all slides graded 1 or higher. Sample DNA
was amplified and uniquely barcoded via multiplex PCR.
Barcoded samples were pooled and high-throughput
targeted re-sequencing of 16 drug-resistance associated
loci was performed on the Illumina MiSeq platform.

Identified mutations were compared with a library of
known drug-resistance associated mutations.
Results: We collected and examined 200 slides, of which
160 (80%) were 1 or above. Sequence depth and quality
varied between loci and samples. We obtained interpretable sequence data for between 5% and 75% of our
samples, depending on the loci. One of 73 samples
(1.4%) with interpretable sequence data had a mutation
to rpoB, 1 of 8 samples (12.5%) had a mutation to gyrA,
and 4 of 51 samples (7.8%) had a mutation in pncA.
These mutations are associated with resistance to
rifampicin, fluoroquinolones, and pyrazinamide respectively. We did not observe any mutations associated with
resistance to isoniazid, ethambutol, aminoglycosides, or
other TB drugs.
Conclusion: Few mutations associated with drug resistance to TB were seen in samples from these provinces of
Afghanistan, but technical challenges in extracting and
amplifying DNA from sputum microscopy slides remain.
Changes to the PCR amplification strategy currently
being tested should further improve sequencing coverage. Our estimates of drug resistance were lower than
previous estimates reported by WHO in 2013, which
were based on only 45 samples. High-throughput
sequencing has the potential to provide information that
can be used to guide both clinical management and
public health policy.
EP-169-05 Association between specific

mutations conferring resistance to rifampicin
and isoniazid and baseline resistance to other
anti-tuberculosis drugs
E Click,1 E Kurbatova,1 H Alexander,1 T Dalton,1
J Ershova,1 P Cegielski,1 J Posey1 1U.S. Centers for Disease
Control and Prevention, Atlanta, GA, USA. Fax: (1) 404 639
1566. e-mail: eoc9@cdc.gov
Background: The Hain GenoType M. tuberculosis

DRplus assay (Hain test) uses line probe based technology to detect the specific mutations in the Mycobacterium tuberculosis rpoB gene that confer resistance to
rifampin, and in the katG and inhA genes that confer
resistance to isoniazid. The Hain test not only provides
important diagnostic information but may potentially
aid in understanding the functional and clinical significance of specific mutations conferring resistance at the
molecular genetic level. We investigated the association
between specific genetic mutations in rpoB, katG, and
inhA and baseline characteristics of cases of multidrugresistant (MDR) tuberculosis (TB).
Design/Methods: In a prospective cohort study in 7
countries, patients with MDR-TB diagnosed using
phenotypic methods and with concordant results for
isoniazid and rifampin drug resistance detected by the
Hain test, where available, (n 447) were included in the
study. We used bivariate analysis to estimate associations.
Results: Participants were predominately HIV negative
(58%, 260/477); 71% (317/447) had previously received
S263
first line anti-tuberculosis drugs and 13% (60/447) had

previously received second-line drugs. Relative to the
most common rpoB mutation, S531L (observed in 369/
447 83%), rpoB D516V was associated with baseline
resistance to three injectable drugs kanamycin, amikacin
and capreomycin (RR 15.6, 95%CI 7.26-33.4) and to
ethambutol (RR 1.4, 95%CI 1.2-1.6) and was protective
for resistance to rifabutin at a concentration of 2.0 lg/ml
(RR 0.36, 95%CI 0.26-0.49). Relative to cases with
isolates that have both inhA and katG mutations (n 86/
447, 27%), cases with isolates that only have inhA
mutations only were more likely to have resistance to
ethionamide at baseline (RR 1.79, 95%CI 1.2-2.67).
Conclusion: Specific mutations conferring resistance to
rifampin and isoniazid may be differentially associated
with baseline drug resistance patterns in MDR TB
patients and merit further investigation into the biological basis of these associations.
09. Behind bars: TB in correctional

settings - II
EP-171-05 Alarming rates of tuberculosis
transmission in Brazilian prisons
D Paiao,1, E Lemos,1 M Pompilio,1 S Simionatto,1
A Motta-Castro,1,2 J Andrews,3 A Ko,4 J Croda1,2 1Federal
University of Grande Dourados, Dourados, Mato Grosso do
Sul, 2Oswaldo Cruz Foundation, Campo Grande, Mato
Grosso do Sul, Brazil, 3Stanford University School of
Medicine, Stanford, CA, 4Yale School of Public Health, New
Haven, CT, USA. e-mail: juliocroda@gmail.com
Background: In Brazil, the incidence of active TB among

prisoners is .1000 cases per 100 000 population, which
is more than 25 times higher than in the general
population. Although prisoners represent only 0.3% of
the population, they accounted for 8% of notified cases
of TB in Brazil in 2013. Despite the high burden of
tuberculosis in prisons in Brazil and other countries,
there have been few empirical data on infection rates
within prisons, which are critical in evaluating the
relative contribution of recent transmission versus
reactivation in driving tuberculosis incidence.
Design/Methods: We performed a prospective cohort
study of prisoners from 12 prisons in the 5 largest cities in
the state of Mato Grosso do Sul, Brazil (Campo Grande,
Corumba, Dourados, Ponta Pora and Tres Lagoas) in
order to estimate the annual incidence of TST conversion
and active TB and evaluate risk factors for these
outcomes. A stratified proportional sample of 3380
subjects was selected from a population of 7221 inmates.
We administered a baseline questionnaire and tuberculin
skin test, followed subjects for 1 year in cohort study, and
performed a repeat TST. TST positivity was defined as
710 mm (75 mm for HIV-infected individuals). TST
conversion was defined as 710 mm induration and at
least 6 mm increase in size of induration in a subject
during follow-up who had a baseline TST ,10 mm. We
S264
used multilevel, multivariable logistic regression to assess

independent predictors of TST conversion and active TB.
Results: Among the 3,380 participants recruited for this
cohort, 1,666 were unavailable at one-year follow-up.
The overall TST conversion rate was 26.0% (95%CI
23.2%-28.0), with higher rates in male prisons (28.7%)
than in female prisons (9.7%) (P , 0.001). Mixed race
(AOR 1.51; 95%CI 1.22-2.04) and black (AOR 1.94;
95%CI 1.16-2.97) were positively associated with TST
conversion, while diabetes (AOR 0.29; 95%CI 0.090.96) was negatively associated. The incidence of active
tuberculosis was 2.8 (95%CI 2.0%-3.8%) in the cohort
and ranged from 1.5% to 5.9% among the 12 prisons (P
, 0.001). Baseline TST positivity (AOR 1.72; 95%CI
1.07-2.75), less than 4 year of schooling (AOR 2.33;
95%CI 1.37-3.85), knowing someone with TB (AOR
3.07; 95%CI 1.74-5.44) and drug use over the last year
(AOR 2.09; 95%CI 1.29-3.38) were associated with
active TB in the multilevel multivariate analysis. Prisoners accounted for 8.2%-42.2% of tuberculosis cases in
their respective cities.
Conclusion: Extremely high annual risk of infection and
incident active tuberculosis were observed in this twelveprison cohort. Recent transmission, rather than reactivation, is likely driving these alarming disease rates.
Urgent measures focused at disrupting the chain of
transmission are needed to address the crisis of tuberculosis in Brazilian prisons.
innovative strategies of Advocacy, Communication and

Social Mobilization (ACSM). We estimate that in the end
of the project, approximately 35 thousand inmates will
be examined.
Intervention or response: Initially, the project was
implemented in the Central Prison of Porto Alegre. From
October to December 2014, the TB case detection was
held in one of the nine galleries of the prison, by
spontaneous demand, at the time the inmate enters the
prison. All individuals did the X-ray exam and were
questioned about the presence of cough. For those with
radiological alteration, sputum sample was taken (when
existent) for GeneXpert, culture and antimicrobial
susceptibility testing. Samples without quality for the
Xpert were referred for sputum smear microscopy (BK).
Results and lessons learnt: 1849 people were examined
with 15 from spontaneous demand, 190 that were
already incarcerated and 1,644 at entry into the prison
system. Of 1,849 prisoners, 272 (14.7%) had altered
results in X-rays. 237 patient samples were collected for
holding BK, culture and GeneXpert. Of these, 27 were
positive for BK, culture or Xpert and 18 were positive
only by Xpert, representing a total of 45 laboratory
confirmed cases with 16.5% of positivity.
Conclusions and key recommendations: The results
suggested high TB positivity in Central Prison of Porto
Alegre. The screening for TB at the entry into the prison
system, galleries and spontaneous demand, proved to be
an important strategy for early TB detection and, in the
long run, to stop the transmission chain of the disease. In
addition, ACSM activities with inmates may have
contributed to their collaboration in the screening of
tuberculosis among the population already incarcerated.
EP-173-05 Tuberculosis screening in prisons in

Rwanda: the potential impact of combining
digital X-ray screening and XpertW MTB/RIF in
TB diagnosis
EP-172-05 Tuberculosis in prisons
C L P Oliveira,1 J L Da Silva,1 V Dos Santos,1
E S De Oliveira,1 R S De Jesus,1 D Pelissari,2,
2
D Chaves Kuhleis,
D Barreira2 1Porto Alegre Health
Secretary, Porto Alegre, Rio Grande do Sul, 2Ministry of

Health, Brasilia, Distrito Federal, Brazil. e-mail:
daniele.chaves@saude.gov.br
Background and challenges to implementation: Tuberculosis (TB) is a serious public health issue in Brazil,
especially in prisons. In 2013, the population of inmates
in the country totaled 600 000 people, which represents
0.3% of the Brazilian population. Of the 70 000 new TB
cases diagnosed per year nationwide, 7.6% comes from
prisoners. Recognizing the high incidence of TB in this
population and the difficulties to implement health
actions within the prison system, the National Tuberculosis Programme submitted a project for TB Reach in
2013, a financial program connected to the Stop TB
Partnership. The project aims to increase TB detection in
prisons from districts of Rio de Janeiro, Porto Alegre and
Charqueadas, by GeneXpert w implementation allied to
E Ruseesa,1 G Mutembayire,1 J Mugabekazi,2 M Gasana1

Ministry of Health, Kigali, 2World Health Organization,
Kigali, Rwanda. e-mail: ruseesa@yahoo.co.uk
Background and challenges to implementation: From

2005, the Rwanda National TB Programme (NTP)
introduced the policy of tuberculosis (TB) screening
using clinical signs for every new prisoner entering the
prison. Since December 2013, we introduced X-ray
screening for identified TB high risk groups.
Intervention or response: The implementation of X-ray
TB screening started in prisons. The screening was done
using digital X-ray technique where all inmates in all 14
prisons in Rwanda were targeted to be screened. All
inmates with abnormal chest x-ray images gave sputum
samples for both LED microscopy and Xpertw MTB/RIF
examination.
Results and lessons learnt: From December 2013 to April
2015, the activity has been implemented in all 14 prisons
in Rwanda. A total of 52 007 out of 54 560 (95%)
prisoners were X-ray screened for TB. Among them,
6362 (12%) had an abnormal chest x-ray image
S265
suggestive of TB disease, and were therefore classified as

presumptive TB. 6181 (97%) and 6052 (95%) presumptive TB gave sputum sample for microscopy and
GeneXpert examination respectively. A total of 336
new TB cases were diagnosed. 264 (79.5%) cases of them
were diagnosed only by Xpert examination. 10 cases
(3%) among the new TB cases were Rifampicin
resistance positive. Compared to the routine TB detection program where an average of 120 new bacteriologically confirmed pulmonary TB cases were detected
annually (2010-2012) in prisons, while in 2013-2014 we
detected 240.
Conclusions and key recommendations: TB screening
using digital X-ray technique may lead to detection of
many TB cases not routinely detected by the clinical
screening especially in TB high risk groups. The role of
GeneXpert in TB diagnosis is highly significant.
RIF test; 3 contacts (1%) were diagnosed with were

diagnosed with TB as shown in the cascade in Figure 1.
Conclusions and key recommendations: Contact investigation in Lira prison led to identification of inmates
with undiagnosed TB which increased the risk of
transmission within the congregate setting. We recommend TB contact screening for both household and close
contacts of inmates diagnosed with TB in addition to the
periodic and routine TB screening at entry into the prison
setting.
EP-175-05 Contact investigations in congregate

settings as a key strategy for TB case finding
and control: experiences from Lira prisons in
Uganda
Background: With an annual incidence of 1675 cases of

TB per 100 000 (2013), the National Penitentiary holds
the highest burden among 17 detention sites in Haiti. The
absence of intensive TB program and lack of funds
allowing the use of sophisticated diagnostics have
permitted to detect TB only among very ill and
symptomatic patients, although many cases were not
detected, especially negative cases. The introduction of
new technology from the National TB Program (PNLT),
and USAID-funded NGO, Health through Walls, at the
request of the Haitian Prison Authority, and use of new
strategy such as screening at entry, contact tracing have
increased detection of TB and improved outcomes. The
introduction of new technology in the largest penitentiary of Haiti has had a significant impact on the health
and living conditions in the prison of Port-au-Prince.
Results: 4718 prisoners entering in the penitentiary
system were screened in 2013 and 5703 in 2014. Of
the 10 421 prisoners screened, 141 of them have been
diagnosed with TB in 2013 and 318 in 2014. The tool
used in 2013 was symptom screening and digital X-ray,
followed by sputum smear and culture for positive
screenings. In 2014, GeneXpert was introduced to help
diagnose more Tb cases and detect MDR-TB according
to the algorithm of PNLT. We have noticed that the tool
introduced in 2014 mixed to the aggressive follow-up of
non symptomatic abnormal X-ray, the molecular test has
permitted to increase the detection rate and double the
number of cases. MDR-TB has been detected among two
through GeneXpert. All have been put under treatment
using DOTS. Mortality rate from TB has decreased by
38%.
Conclusion 141 TB cases have been diagnosed in 2013
and 318 in 2014, demonstrating an increase in TB
detection at Haitis National Penitentiary. While Haiti is
classified as a low-income country, the application of
new technology has been effectively implemented despite
few resources. The lesson learned is that new technology
and a good collaboration with the countrys national TB
program, NGOs and prison authority can help control
TB in overcrowded, low-income prisons. The model will
be expanded to others detention sites of the country.
E Kizito,1 P Donggo,2 M Kenneth,1 S Stavia,3

M Ruhweza,1 B Woldemariam,1 P G Suarez1 1Mangement
sciences for Health, Kampala, 2Lira Regional Referral
Hospital, Lira, 3Ministry of Health, Kampala, Uganda. e-mail:
kizitoen@gmail.com
Background and challenges to implementation: One of

the components of the WHO Stop TB strategy is to
address the needs of TB contacts, and of poor and
vulnerable populations. Inadequate ventilation, overcrowding and inadequate TB infection control practices
in congregate settings increase the risk of transmission of
all forms of Tuberculosis. During 2014, Lira regional
referral hospital in Northern Uganda diagnosed 12
patients with drug sensitive TB and 2 patients with
multi-drug resistant TB all of whom were serving jail
sentences in nearby prisons. Over the years, such patients
would be transferred to the hospital for treatment but
there was no documented contact screening and investigation for all inmates.
Intervention or response: Lira Regional Referral Hospital with support from the USAID funded TRACK TB
Project conducted education sessions about TB for the
prisons authorities and inmates with a focus on TB
transmission and infection control including the need to
promptly identify and evaluate presumptive TB cases.
Multi-disciplinary teams comprising of Nurses, clinicians
and laboratory technologists visited each of the prisons
to conduct the sensitization session and screening of all
inmates using the Intensified case finding tool (ICF).Sputum samples were collected from all inmates that
were found to have TB symptoms (presumptive TB cases)
and subjected to an MTB/RIF test.
Results and lessons learnt: A total of 300 inmates were
screened for TB, all were males just like the index cases
and their average age was 32 years. All these contacts
were HIV negative. Fifty-three symptomatic contacts
were identified and had samples subjected to the MTB/
EP-176-05 Bringing technology to prisons of

low-income countries to end TB: the
experience of Haiti
J May,1 K Duverger,1 M Bury,1 W Morose2 1Health through
Walls, Miami, FL, USA; 2National TB Program, Port-au-Prince,
Haiti. Fax: (1) 305 893 5009. e-mail: drjpmay@aol.com
S266
EP-177-05 Prisons and tuberculosis: developing

a prison-based tuberculosis support model in
Zambian prisons
EP-178-05 A first insight into high prevalence of

undiagnosed tuberculosis cases in Mbuji-Mayi
Central Prison in DR Congo
N Sanjobo,1 O Simooya,1 L Lochting2 1IN BUT FREE Prisons

Project, Kitwe, Zambia; 3LHL International, Oslo, Norway.
e-mail: nsanjobo@cbu.ac.zm
M Kaswa Kayomo,1 G Bakaswa,1 D Muteteke,1

M Boelaert,2 B De Jong2 1National Tuberculosis
Programme, Kinshasa, Democratic Republic of the Congo;
2
Institute of Tropical Medicine, Antwerp, Belgium. e-mail:
meckkay2002@yahoo.fr
Background and challenges to implementation: Prison

communities constitute high risk populations for tuberculosis (TB). Most prisons in Zambia are overcrowded
and coupled with poor ventilation and long hours of cell
confinement, such conditions facilitate the spread of TB.
Despite being exposed to such unpleasant living conditions, having prisoners in one place could be an ideal
environment for TB control and an opportunity to
facilitate effective TB control measures.
Intervention or response: Since 2012, IN BUT FREE
Prisons Project with support from LHL International has
been executing a TB programme in Zambia. The project
covers over 5000 prisoners and 430 prison officers. The
objective of the intervention is to build capacity of
prisoners and officers in TB. The programme uses
volunteer prisoners and officers trained as Peer Supporters. Each prison elects their executive committees
comprising prisoners and officers. These committees
are responsible for coordinating work at each prison and
report to the project team monthly. Peer supporters meet
weekly to review progress. Thursdays are dedicated as
cell teaching days and all trained peers are afforded
opportunity to lead discussions in their cells. Peer
supporters are also accorded opportunities to share
information on TB, HIV and AIDS with fellow prisoners
and officers at parade square every week. Sensitization
campaigns are carried out through one on one discussion,
drama performances, cell teachings, edu-sport, poems,
distribution of Information, Education and Communication materials. Peer supporters assist in identifying
possible suspected TB patients and refer them for
investigations and treatment. Peer supporters also
monitor and support their colleagues while on treatment.
Results and lessons learnt: 430 Prisoners and 105
Officers have been trained as Peer Supporters. A
computer has been procured for the largest prison.
Posters on TB and HIV, Brochures and TB Booklets have
been printed and distributed. A video on TB in prison has
been produced. Mentoring and support visits are
conducted quarterly. Drama Groups have been established. Health education and TB case detection is
available to prisoners. Patient referral system has been
developed and follow up is carried out.
Conclusions and key recommendations: The involvement of prisoners and prison officers is fundamental to
achieving an effective response to TB control in prisons
in the new agenda for lung health.
Background: In low income countries, such the DRC,

due to the financial and logistical issues, the current
Xpert w MTB/RIF machines available are located in
national or provincial reference laboratories and more
dedicated to RMP detection among MDR suspects.
Xpert MTB/RIF utilization in intensified case finding and
its impact on TB burden in prisons in DRC where a
generalized HIV epidemic does exist has not yet been
examined.
Design/Methods: We report the outcomes of an active
TB case finding performed in the Mbuji-Mayi Central
Prison in the Kasai Oriental Province, DRC in 2015. In
addition, because RMP is the pillar of first-line TB
treatment and the potential risk of spread of drugresistant M. tuberculosis within the prison and exportation to the general population exist, we also studied the
drug susceptibility (RMP) patterns of M. tuberculosis
isolated from inmates in this prison and the associated
risk factors.
Results: Of the 918 inmates found out in the prison, we
did collect 350 sputum specimens. Among the 336
specimens tested, we retrieved 147 TB cases and 14 TB
rifampicin resistant (TB-RR) cases. All TB and TB-RR
have been put in treatment. However, these numbers are
expected to change on the rise.
Conclusion: With overcrowding conditions, poor ventilation, and malnutrition, the risk of ongoing spreading is
still very high.
10. The neglected cousins: trends and

treatment of non-tuberculous
mycobacteria
EP-179-05 Identification and characterization
of non-tuberculous mycobacterial infection
among people living with HIV presenting with
TB symptoms in Botswana
J Basotli,1 T Agizew,1 K Radisowa,2 Z Rey,3 T Tamuhla,1
H Alexander,3 A Auld,3 A Finlay1,3 1Centers for Disease
Control and Prevention, Gaborone, 2National Tuberculosis
Reference Laboratory, Gaborone, Botswana, 3U.S. Centres
for Disease Control and Preention, Atlanta, GA, USA. Fax:
(267) 390 2713. e-mail: jbasotli@cdc.gov
Background: Botswana has the second-highest HIV

prevalence globally, with one in four adults HIV-infected,
and tuberculosis (TB) a common cause of morbidity and
mortality. Recently we reported a higher-than expected
prevalence of non-tuberculous mycobacteria (NTM)
isolation from symptomatic people living with HIV
(PLHIV) attending 22 HIV care and treatment centers.

We characterized the NTM species in this population
given NTM species differ in clinical relevance. Isolation
of Mycobacterium kansasii and M. malmoense from
pulmonary specimens often are associated with disease,
whereas M. gordonae and M. simiae are typically
contaminants. M. avium complex (M. intracellulare
and M. avium), and M. abscessus form an intermediate
category.
Methods: PLHIV attending 22 HIV care and treatment
centers were systematically screened for TB symptoms
and sputum collected for liquid culture and identification. Culture growth was identified as NTM using the
SD-Bioline TB Ag MPT64 Kit and Ziehl-Neelsen
microscopy. NTM isolates underwent species identification by the Hain GenoType CM and AS line probe assays
(LPAs).
Results: From 08/2012 to 10/2014, 1960 PLHIV
presented with TB symptoms submitted at least one
sputum specimen for culture testing. Among these, 429
(21.9%) had 71 positive culture result, 247 (57.6%) of
which were identified as NTM. NTM species were
identified for 205/247 (83.0%) patients, including five
mixed species. M. intracellulare was the most common
species isolated, 107/205 (52.2%). Other NTMs commonly associated with pulmonary disease included M.
malmoense 8 (3.9%), M. avium 4 (2.0%), M. abscessus 2
(1.0%), and M. kansasii 1 (0.5%) (Table). Typical
contaminants M. gordonae and M. simiae were identified
in 15 (7.3%) and 9 (4.4%) isolates, respectively. Also, 32
(15.6%) NTMs could not be identified by the LPAs. At
least 135 (65.9%) of NTM isolates were not obvious
contaminants and require clinical follow-up.
Conclusions: Our preliminary finding is the first report
on isolation and identification of NTM among a cohort
of HIV-infected persons with TB symptoms in Botswana.
Out of 13 species identified, M. intracellulare was
predominant as has been reported in similar settings.
Environmental sampling is underway to look for
potential sources of contaminants. The higher-than
expected prevalence of possibly disease-causing NTM
species warrants further investigation and consideration
during preparation of national TB guidelines.
Table NTM isolated among PLHIV presenting with TB
Symptoms in Botswana
NTM species
M. intracellulare
M. gordonae
M. simiae
M. malmoense
M. scrofulaceum
M. fortuitum
M. asiaticum
M. avium
M. abscessus
M. genavense
M. lentiflavum
M. kansasii
M. phlei
Mixed NTM*
Others
Total
Number
Frequency
107
15
9
8
6
6
5
4
2
2
2
1
1
5
32
205
52.2%
7.3%
4.4%
3.9%
2.9%
2.9%
2.4%
2.0%
1.0%
1.0%
1.0%
0.5%
0.5%
2.4%
15.6%
100.0%
S267
*Mixed species: more than one species identified per patient
NTMs that require additional tests for species identification
EP-180-05 Profile of non-tuberculous

mycobacteria patients in Brazil, 20122014
N Saita,1 D Pelissari,1 F. Dockhorn,1 D Gomes Dellorti,1
F M Reis,1 R Ruy De Souza,1 P Bartholomay Oliveira,1
D Barreira1 1National Tuberculosis Programme, Braslia,
Distrito Federal, Brazil. e-mail: nanci.saita@saude.gov.br
Background: Information System of Special Treatments

for Tuberculosis (SITE-TB) is an online system used by
secondary and tertiary reference units in order to notify,
monitor and evaluate the diagnosed cases requiring
special treatment, including cases of nontuberculous
mycobacteria (NTM). The NTM cases are included in
SITE-TB to facilitate the logistics of drugs, since in Brazil
the NTM is not a notifiable disease.
Objectives: To describe the profile of patients with NTM
recorded in SITE-TB in the period 2012 to 2014,
according to clinical and demographic information.
Design/Methods: This is a descriptive study of new NTM
cases in Brazil, from 2012 to 2014. We used data from
SITE-TB and the following variables were analyzed:
gender, age, race, educational level, form of the disease,
HIV status, type of NTM, comorbidities and outcomes.
Results: In the period from 2012 to 2014, 604 new NTM
cases were recorded in SITE-TB. 52% were male; 21%
had between 55 to 64 years; 50% were white; 24%
studied 4 to 7 years. As the clinical form, 78% were
pulmonary. The more common NTM species associated
were: Mycobacterium kansasii (16%), Mycobacterium
avium (12%) and Mycobacterium abscessus (10%). For
the comorbities: 15% were smokers; 7% had diabetes
mellitus; 7% were alcoholics; and 19% had HIV
positive. As to the outcomes, 48% are still in treatment
and 29% cured.
Conclusion: The incidence of tuberculosis by Mycobacterium tuberculosis is admittedly more common in men,
however, it was observed that the NTM affects similarly
men and women. The distribution according to type of
NTM in Brazil was similar to that reported in other
countries. Of note is the high percentage of smokers and
the TB-HIV co-infection. The National Tuberculosis
Program in Brazil encourages the notification of NTM
cases in SITE-TB, since 2013. Therefore, data reported
are limited and do not represent the totality of national
cases for the analyzed period.
EP-181-05 Rapid SPs identification of NTM by

molecular genetic assay from culture
C Iravatham,1 P Chelluri,2 M Vijaykumar,3
D Mallikarjuna4 1Dr.Iravathams Clinical Laboratory,
Hyderabad, 2Shadan Medical college, Hyderabad, 3AP Chest
Hospital, Hyderabad, 4Bhimavaram Hospital, Hyderabad,
India. Fax: (91) 40 6662 3533.
Background: Differential diagnosis between non tuberculous mycobacteria ( NTM and M. tuberculosis
complex has always been a catch-22 situation for
S268
clinicians and microbiologists.Suspected NTM isolates

had to be sent to refferal lab for sps identification.either
by HPLCor conventional biochemicalmethodwhich are
both time consuming and laborious. Advent of molecular
dignostic assays has contributed a great deal towards
early diagnosis and initiating treatment Design: To
evaluate DNA strip technology for sps identification of
NTM from culture slope
Methods: Patients not responding to Anti TB treatment
were reffered for further lab investigation (n 13).
Samples included pus from non healing sub cutaneous
abscess (sequelae to post laparascopic procedure)(n 9),
sputum (n 1), BAL (n 1), breast abscess (n 1), lymph
node tissue (n 1). HIV was status negative in all
patients. Male: female ratio was 5:8. Mean age male
48.5; female 38.7 Samples screened for AFB by ZN and
Fluorescent stain, Line probe assay (Lipa) done directly
on clinical samples to rule out M. tuberculosis complex
and also check for HR resistance. AFB culture done by
conventional LJ and MGIT system. Antibiogram done by
MIC method for ATT first line drugs (SHER), quinolones
and aminoglycosides and by microdilution for Clarithromycin,Co-trimaxazole and Linezolid. sps identification from positive culture slope done by DNA strip assay
using Genotype mycobacterium CM kit ( Hain), simultaneously compared with conventional biochemical or
HPLC method.
Results: All samples screened were positive for AFB.
Culture positive within two weeks. Sps identified by
DNA strip assay were: M. fortutitum/M. mageritense (n
9) from post laparoscopic procedure, M. abscessus/M.
immunogenum (n 2) breast abscess and sputum; M.
chelonae/M. immunogenum (n 1) BAL; M. scrofulaceum (n 1) from cx tissue. Turn around time for the
assay , 8 hours. Conventional biochemical and HPLC
tests confirmed the rapid growers and HPLC M.
scrofulaceum. Antibiogram profile: All isolates resistant
to SHER. Sensitivity pattern varied for quinolonoes and
aminoglycosides in rapid growers suggesting possibility
of mutation.
Conclusion: Early diagnosis upto sps level and intervention will have positive impact on treatment outcome.Limitation Assay can be performed from culture slope and
not from clinical sample.
the center of expertise for mycobacterial infections

(University lung center Dekkerswald, Nijmegen, The
Netherlands).
Methods: We conducted a retrospective study of all
patients seen between 2008 and 2013 with at least one
NTM isolate and diagnosed using the American Thoracic
Society (ATS) diagnostic criteria. Patients with non
classical radiographic features that were not described
in the ATS-criteria were included as well.
Results: 334 patients were reviewed. Sixty-three patients
with NTM-PD were included. Thirty-two of them were
female and mean age was 60.8612.9 years. M. avium
complex (MAC) was the predominant species (n 37,
58.7%); M. malmoense and M. kansasii were seen in 7
and 6 patients. M. abcessus was seen in 4 patients, M.
chimaera in 1 patient, M. simiae in 3 patients, M. xenopi
in 2 patients and 3 patients had combined infections. The
most predominant radiologic features were consolidations or infiltrates (61.9%) and nodules (60.3%). Most
patients had underlying COPD (73.0%), bronchiectasis
was present in 54.0%, 39.7% smoked at time of
diagnoses and 34.9% had a history of smoking. 53.1%
(n 17) of patients treated for MAC-PD were given a
rifamycin-ethambutol-macrolide regimen (REM). Five
patients received REM in combination with clofazimin
and 7 patients received REM in combination with
amikacin and clofazimin. Five patients with MAC were
not treated. All six patients with pulmonary M. kansasii
disease were treated with rifampicin and ethambutol;
four also received isoniazid. The overall cure rate was
74.1%; 71.9% for MAC, 83.3% for M. kansasii and
66.7% for M. malmoense. In our study population allcause mortality was 25.9%; during treatment; 10
patients experienced treatment failure and 4 patients;
7.4% relapsed.
Conclusions: In our setting, MAC was the predominant
NTM and treatment largely followed ATS guidelines.
Treatment success rates are comparable with data from
the literature. The poor outcomes and toxicity underline
the need for clinical trials addressing NTM-PD treatment.
EP-182-05 Characteristics and treatment

outcomes of nontuberculous mycobacterial
pulmonary disease in a mycobacterial disease
reference clinic
S Zweijpfenning,1 C Magis-Escurra,1 J Van Ingen,1
M Boeree,1 W Hoefsloot1 1Radboud University Medical
Center, Nijmegen, Netherlands. e-mail:
sannezweijpfenning@gmail.com
Background: The incidence of nontuberculous mycobacterial pulmonary disease (NTM-PD) is increasing in the
last few years in the Netherlands. Evidence concerning
treatment is poor. In this study we give an overview of
clinical aspects, treatment and outcome of all patients
with suspicion of clinical relevant NTM-PD referred to
EP-184-05 Non-tuberculous mycobacterial

disease: the reality in Lisbon
F Teixeira Lopes,1 S Alfarroba,1 A Dinis,1 N Ribeiro,1
M Gomes,1 A Tavares1 1Centro Diagnostico

Pneumologico
Dr. Ribeiro Sanches, Departamento de Saude

Publica
da
Administraca
o Regional de Saude
de Lisboa e Vale do Tejo,
Lisboa, Portugal. e-mail: franciscateixeiralopes@gmail.com
Background: The incidence of non-tuberculous mycobacteria disease (NTM), although not fully known, is
currently estimated as 1 to 1.8/100 000 inhabitants. As

NTM are ubiquitous in the environment, their presence
is not always synonymous of disease. There are clinical,
imaging and microbiological criteria that determine the
need for treatment and therapeutic regimens. The aim of
this paper is to characterize the NTM disease reported in
Lisbon in the last 3 years.
Methods: Retrospective study on NTM notified by the
CDP Dr. Ribeiro Sanches (Lisbon) from 2012 to 2014 to
characterize the population, clinical presentations, diagnostic criteria, treatment regimens and therapeutic
results.
Results: In the last three years 37 cases of NTM were
reported, corresponding to 0.7/100 000 cases/year. Most
were female (62%) with mean age of 56.7 yrs. Risk
factors were identified in 33 patients including bronchiectasis (17 patients), HIV infection (10), history of
neoplasm (7) and COPD (5) among others. More
frequent registered mycobacteria were Mycobacterium
avium complex (17), M. abcessus (4) and M. chelonae
(3). Isolation was obtained from sputum (16 cases),
bronchial lavage (16), blood cultures (2 M. avium),
pleural fluid (1 M. gordonae) and skin biopsy (1 m.
intracellular). Most cases were first infections (36). The
vast majority (89%) had only pulmonary manifestations,
with one case of pulmonary and cutaneous co-infection
and another with disseminated disease. Most patients
had radiological findings (34), the most common being
bronchiectasis with inflammatory changes (7) and
cavitating nodules (6). Therapy had a median duration
of 9.1 months. There were four therapy withdrawals and
two interruptions, one due to optic neuritis and another
for gastrointestinal intolerance. All therapeutic schemes
included a macrolide and the vast majority used
ethambutol l(33) and rifampin/rifabutin (28), according
to published guidelines. Recurrence was identified in
only 1 patient.
Conclusions: The diagnosis and treatment of NTM
disease remains a clinical challenge. Epidemiological
data are scarce and therapeutic standards require
updates. We admit an underdiagnosis of NTM in Lisbon
according to literature. In our center, most of the patients
were treated according to the recommendations although
the average length of treatment was shorter than
recommended. One of the possible reasons is the
difficulty to sensitize patients to therapeutic adherence
after symptomatic improvement.
S269

11. Improving TB treatment outcomes
OA-379-05 Effectiveness of a conditional cash
transfer programme in tuberculosis cure rate: a
retrospective cohort study in Brazil
A Torrens,1 M Sanchez,2 J Nery,3 D Barreira,1 D Rasella4
1
Brazil Ministry of Health, Braslia, Distrito Federal, 2University
of Braslia, Braslia, Distrito Federal, 3Federal University of
Braslia, Braslia, Distrito Federal, 4Independent, Rio De
Janeiro, Distrito Federal, Brazil. Fax: (556)183007775.
e-mail: ana.torrens@saude.gov.br
Background: Tuberculosis (TB) is a poverty related

disease. The Bolsa Famlia Program (BFP) is currently
the largest conditional cash transfer program in the
world that targets low income families. By improving the
material conditions and access to health services have the
potential to improve access to quality TB care. This study
aims to evaluate the effectiveness of the BFP in
tuberculosis treatment success.
Design/Methods: This study aims to evaluate the
effectiveness of the BFP in tuberculosis treatment success.
It is a retrospective cohort study. The study population is
newly diagnosed TB cases in Brazil in all months of 2010
exposed and not exposed to BFP income transfer during
TB treatment. The association measures relative risk and
confidence intervals between exposure to cash transfer
and treatment successes (proven cure plus treatment
complete) was estimated by Poisson regression with
robust variance model.
Results: The cure rate among those exposed to BFP
income transfer during TB treatment was 81.7%, 5.4%
higher than among those not exposed during follow up
which was 76.3%. Multivariate analysis estimated a
relative risk of 1.07 (95%CI) of the income benefit of
BFP in tuberculosis treatment successes when adjusted
for the covariates age, race, TB type, diabetes mellitus,
HIV status, directly observed treatment, illiteracy, zone
of residence, number of rooms and floor material of
household and family per capita income.
Conclusion: The results of the study show that conditional cash transfer programs may increase TB patients
cure rate of. This study, conducted with all TB patients in
Brazil eligible to PFB benefits, one of the largest cash
conditional transfer programs for low income families in
the world, is relevant because treatment success continues a great challenge for TB control in the country.
S270
Oral abstract sessions, Saturday, 5 December
Results: The characteristics of the 1544 (59.5% men)

patients included in the derivation sample were similar to
those of 1317 (60.1% men) patients included in the
validation sample. There were 110 deaths (cumulative
mortality rate: 7.1%) in derivation sample and 103
deaths (7.8%) in validation sample. The SPS was
calculated as follows: age (years)-53adjusted body mass
index (kg/m2) (clinical form of TB: smear negative
pulmonary TB, 30 points; extra-pulmonary TB, 15
points; smear positive pulmonary TB, 0 point) (HIV
infection: yes, 45 points; no, 0 point). The discrimination of this score was excellent in derivation sample (cstatistic [95%CI]: 0.803, 0.762-0.844, P , 0.001) and
validation sample [c-statistic: 0.817 (0.777-0.856), P ,
0.001]. Five mortality risk categories were identified:
class I (low risk, SPS , 60), Class II (moderate risk, 60
6 SPS , 30), Class III (high risk, 30 6 SPS , 0), class
IV (very high risk, 0 6 SPS , 30) and class V (critical
risk, SPS 7 30). The probability (standard error of
estimate) of death during TB treatment was 1.3% (0.2),
5.7% (1.0), 15.9% (2.6), 24.2% (3.7) and 29.5% (4.5)
respectively in the classes I, II, III, IV and V in the
validation population.
Conclusion: We have developed and validated a simple
prognostic score for TB using routinely data collected in
the TB programs. This score can allow rapid risk
stratification of TB patients for close monitoring and
appropriate management of these patients.
OA-380-05 Development and validation of a

simplified prognostic score for tuberculosis
under the tuberculosis programmme in a subSaharan African setting
OA-381-05 The impact of referring patients

with tuberculosis from a mega-tuberculosis
centre to a peripheral health care centre on loss
to follow-up
E W Pefura-Yone,1,2 A D Balkissou,1,2 V Poka,1

H K Fatime Abaicho,3 P T Enono-Edende,3 C Kuaban,1,4
A P Kengne5 1Faculty of Medicine and Biomedical Sciences,
University of Yaounde 1, Yaounde, 2Yaounde Jamot Hospital,
de Technologie Medicale,
Yaounde, 3Institut Superieur
Yaounde, 4Faculty of Health Sciences, University of Bamenda,
Bamenda, Cameroon; 5Medical Research Council, Cape
Town, South Africa. e-mail: pefura2002@yahoo.fr
E W Pefura-Yone,1,2 H K Fatime Abaicho,3

A D Balkissou,1,2 A P Kengne4 1Yaounde Jamot Hospital,
Yaounde, 2Faculty of Medecine and Biomedical Sciences,
de
University of Yaounde 1, Yaounde, 3Institut Superieur
Technologie Medicale,
Yaounde, Cameroon; 4Medical
Research Council, Cape Town, South Africa. e-mail:
pefura2002@yahoo.fr
Background: Death during treatment is a major challenge for tuberculosis (TB) control programs in developing countries. Currently, there is no simple tool for use in
routine TB programs in high TB burden countries to
stratify patients for mortality risk. The objective of this
study is to generate and validate a simple score for
predicting the risk of death during TB treatment.
Design/Methods: This was a cohort study based on data
from clinical charts of patients aged 715 years, who
undertaken diagnosis and treatment for tuberculosis at
the referral center of the Yaounde Jamot Hospital
between January 2012 and December 2012 (derivation
sample), and January 2013 to December 2013 (validation sample). Predictors of mortality during TB treatment
were obtained from logistic regression models. Regression coefficients (b) were used to generate the simple
prognostic score (SPS) for predicting death during TB
treatment. The c-statistic was used to assess the
discriminatory ability of the model.
Background: Several historical tuberculosis (TB) treatment centers in developing countries are still diagnosing
and treating a disproportionate number of patients with
TB. The Yaounde Jamot Hospital (YJH) is a referral and
historical hospital center for TB diagnosis and treatment
(TDC) in Cameroon. Over the last 5 years, 1600 to 1800
patients with TB were detected each year at the YJH,
with most of them subsequently receiving treatment in
this institution. Since three years, a decentralization
program has been initiated, aiming at referring patients
from YJH to other TDC in Yaounde. The objective of this
study was to assess the impact of referring patients with
TB from the mega-TDC of the YJH to other TDC on the
rate of lost to follow-up.
Methods: In this retrospective cohort study, the records
of patients diagnosed with TB at YJH from January 2012
to December 2013 were reviewed. Patients were divided
into two groups: those referred from HJY for another
TDC (referred group) and those treated at YJH (YJH
group). The Kaplan-Meier estimator and log-rank test
were used to compare the rate of lost to follow-up in both

groups. The proportional Cox regression model was used
to estimate the hazard ratio of lost-to follow-up in the
referred group in comparison to the YJH group.
Results: Of the 3794 patients enrolled, 59.7% were male
and their mean age standard deviation (SD) was 35.8
(12.9) years. During the study period, 553 (14.6%)
patients were referred to another TDC and 3241
(85.4%) were treated in YJH. Patients in the referred
group had a mean age (SD) of 34.2 (12.7) years and those
in YJH group had a mean age of 36.1 (12.9) years (P ,
0.001). The clinical forms of TB were distributed
similarly in both groups with 70.9% smear positive
pulmonary TB cases in the referred group and 67% in the
YJH group. The rates of treatment success, death and lost
to-follow-up were respectively 81.4%, 2.4% and 11.2%
in the referred group. The equivalent figures for the YJH
group were respectively 68%, 9.3% and 15% (P ,
0.001). The Kaplan-Meier estimator showed that the
probability of being lost to follow-up was lower in the
referred group (log-rank test, P 0.007). In univariable
analysis, the hazard ratio (95%CI) for lost to follow-up
in the referred group compared to YJH group was 0.70
(0.54 to 0.91), P 0.008. In multivariable analysis
including the baseline characteristics significantly associated with lost to follow-up in univariable analysis,
patients reference from YJH to other TDC remained a
protective factor of the lost to follow-up with adjusted
hazard ratio (95%CI) of 0.55 (0.41 to 0.74), P , 0.001.
Conclusion: The reference of patients from the mega-TB
center of YJH to other treatment centers significantly
reduces the risk of being lost to follow-up. An effort
should be made to ease the pressure on large TB
treatment centers by referring well-characterized TB
patients to small satellite centers for treatment and
follow-up.
S271
OA-382-05 High MDR-TB rates among

successfully treated new smear-positive TB
patients using the intermittent regimen under
the DOTS strategy
D Garg,1 N Garg2 1Chest Center Jhandewalan, New Delhi,
2
Health Center, Punjab Bhawan, New Delhi, India. e-mail:
dkgarg56@gmail.com
Background: Under Revised National Tuberculosis

Control Programme (RNTCP), India follows WHO
recommended DOTS strategy using intermittent regimen
to treat drug sensitive tuberculosis without any policy of
long term follow up of patients. The emergence of
resistance to drugs used to treat tuberculosis has become
a significant public health problem and an obstacle to
effective TB control. An MDR-TB case is defined as a
presumed MDR-TB patient who is culture positive & has
M. Tuberculosis resistant to isoniazid & rifampicin.
Sputum sample from an MDR-suspect of any criteria A,
B or C was tested by Culture & DST at an RNTCP
accredited laboratory. Criteria B which includes all smear
positive previously treated pulmonary TB cases at
diagnosis is being followed since January 2012.
Design/Methods: This study included total 1458 MDRTB suspects ( including 322 retreatment relapse cases )
during the period Jan. 1, 2012 to December 31, 2014 in
Jhandewalan district, New Delhi, India covering a
population of 5.5 lacs. Two sputum samples (one early
morning & other supervised spot) were sent in falcon
tubes to the IRL for diagnosis of MDR-TB by drug
sensitivity test LPA (Line Probe Assay) or CBNAAT
(Cartridge Based Nucleic Acid Amplification Test). The
patients were interviewed to evaluate the past history of
treatment and records were checked for data validation.
Results: Total 144 (9.9%) MDR-TB cases were diagnosed. Out of 322 retreatment relapse cases there were
47 (14.6 %) MDR-TB cases. 26 MDR cases (17 M, 9 F,
all above 14 yrs of age) were from new smear-positive
(NSP) cases successfully treated with Cat I. Thus MDR
cases from NSP form 18% (26/144) of the total and 55%
(26/47) of the MDR cases from retreatment relapse
patients. Moreover 34.6 % (9/26) of these cases were
cured less than 1 year back.
Conclusion: In view of the findings mentioned above, we
conclude that there is high MDR-TB among successfully
treated new smear-positive cases with intermittent
regimen under RNTCP. The treatment completion
declared based on sputum microscopy creates doubt.
There may be reactivation & genetic mutation of the
remaining bacilli. Recommendations: 1) the programme
may reassess its treatment regimen of Cat I regarding the
frequency of doses & duration of treatment, 2) treatment
completion may be declared based on liquid culture
technique, and 3) provision of timely follow up of
successfully treated cases to find out the relapse at the
earliest.
S272
OA-383-05 Low rifampicin levels in patients

treated with intermittent anti-tuberculosis
regimens in India
OA-384-05 Cost-effectiveness of different

models of treatment for DR-TB patients in
KwaZulu-Natal, South Africa
G Ramachandran,1 H K Agibothu Kupparam,1

C Vedachalam,1 J Lavanya,2 S Swaminathan1 1National
Institute for Research in Tuberculosis, Chennai, 2Chennai
Corporation, Chennai, India. e-mail:
geetha202@rediffmail.com
M Loveday,1 K Wallengren,2 J Brust,3 A Voce,4

N Padayatchi,5 J Ngozo,6 Z Radebe,6 E Daviaud1 1South
African Medical Research Council, Cape Town, 2ThinkSA,
Durban, South Africa; 3Montefiore Medical Center & Albert
Einstein College of Medicine, New York, NY, USA; 4University
of KwaZulu-Natal, Durban, 5Caprisa, Durban,
6
KwaZulu-Natal Department of Health, Pietermaritzburg,
marian.loveday@mrc.ac.za
Background: In the Revised National Tuberculosis (TB)

Control Programme (RNTCP) in India, patients are
treated for TB with thrice-weekly regimens. We determined plasma concentrations of rifampicin (RMP),
isoniazid (INH) and pyrazinamide (PZA) and examined
factors that influenced drug levels and treatment
outcomes.
Design/Methods: Adult TB patients with pulmonary/
extrapulmonary TB, receiving thrice-weekly anti-tuberculosis treatment (RMP 450mg/600mg for weight ,/.
60kg), INH 600 mg, PZA 1500mg) in the RNTCP in
Chennai Corporation were studied. At steady state, twohour post-dose concentrations of RMP, INH and PZA
were determined by high performance liquid chromatography, after directly observed drug administration.
Cured/treatment completed was considered a favourable
outcome and death/failure an unfavourable outcome.
Multivariate and logistic regression analyses by stepwise
method respectively were used to determine factors that
influenced drug levels and treatment outcomes.
Results: The study population consisted of 1266 patients
(mean age 39.6 years; males 67%; mean body mass index
(BMI) 19.4kg/m2). 86%, 11% and 19% of patients had
sub-therapeutic RMP, INH and PZA concentrations
(sub-therapeutic cut-off: RMP , 7lg/ml; INH , 2lg/
ml; PZA , 20lg/ml). Females had higher drug levels
than males (P 0.012 and , 0.001 for RMP and PZA
respectively), those with BMI . 18.5 kg/m2 had lower
drug levels (P 0.063, , 0.001, , 0.001 for RMP, INH
and PZA respectively) and those with blood glucose .
200mg/dl had lower INH and PZA levels (P , 0.001 for
INH and PZA). RMP levels were significantly influenced
by sex and BMI, INH by age, BMI and blood glucose and
PZA by sex, BMI, blood glucose, type of TB and
smoking. Outcomes were available in 212 smear positive
patients, of whom 17 had an unfavourable response.
RMP concentrations were significantly lower in unfavourable than favourable responders (0.67 vs. 3.5 lg/ml;
P , 0.001). Factors that influenced treatment outcome
were low 2-hour RMP concentration (P 0.015; 95%CI
1.1 1.8; AOR 1.4) and low BMI (P 0.033; 95%CI
1.1 23.7; AOR 5.2).
Conclusion: The majority of patients on thrice-weekly
anti-tuberculosis therapy had sub-therapeutic RMP. Low
RMP levels and undernutrition were associated with
unfavourable treatment outcomes. There is a need to
consider increasing drug doses in those with higher BMI
and diabetes. This study has important clinical implications as suboptimal blood levels could lead to failure and
acquired drug resistance.
Background: Given the size of the DR-TB burden in

South Africa and the cost of treating each DR-TB patient,
policy makers and TB programme managers need to
know the cost of each model of care, and which model of
care is most cost effective. In this study we report actual
costs together with the cost-effectiveness of a spectrum of
different models of care for DR-TB patients.
Methods: In our operational prospective cohort study
DR-TB patients were treated with 4 different models of
care. At the centralised hospital and decentralised sites
(district hospitals) patients were initially hospitalised. In
the two community-based models of care patients were
not hospitalised and the daily injectable was administered at either a local clinic or a mobile injection team.
Using a provider perspective, we collected costs to the
health service only, excluding household costs. We
performed a cost analysis to enable us to compare the
cost of each model of care together with the costeffectiveness. Cost-effectiveness was defined as the cost
per MDR-TB patient successfully treated, that is, cured
or completed treatment.
Results: Successful treatment rates varied across the
different models of care, from a low of 54% at the
centralised hospital to a high of 72% at Decentralised 1.
For each treatment model the cost per patient treated
varied considerably (Centralised $30 067, Decentralised
1 $24 516, Decentralised 2 $33 554, Clinic $9519,
Mobile $10 616) (Table). The cost per successful
treatment was highest in models of care in which patients
were hospitalised (Centralised $55 680 and Decentralised 2 $64 776), where it was more than 3 times higher
than the community-based models of care (Clinic $15
865 and Mobile $18 495). Hospitalisation was the major
cost driver accounting for 67-82% of the total costs in
the 3 models of care in which patients were hospitalised.
Length of hospitalisation and the subsequent costs varied
(Table). In the community-based models drugs and tests
were the major cost drive accounting for 40-52% of the
treatment costs.
Conclusion: Community-based care for patients with
MDR-TB is more cost-effective than either care in a
decentralised or centralised setting as evidenced by the
lower cost per patient successfully treated. This exploratory study suggests that community-based care is both
effective and cost-effective. However, our study sample
in the community-based sites was small and a larger
study is needed to confirm these findings.
S273
good outcome than patients with RMP AUCs in the

lowest quartile in the control and test arms respectively.
Conclusion: In TB patients on a standard regimen,
measures of RMP exposure better predict culture
conversion than non-PK factors including cavity and
HIV status. The role of RMP AUC and CMAX in
bacterial kill is consistent with HFS experiments. Thus,
for shortening therapy, the background regimen needs to
be dose optimized.
OA-385-05 Predictors of 2 month culture

conversion in pulmonary tuberculosis
W Smythe,1, J Pasipanodya,2 P Denti,1, C Merle,3
P Olliaro,4 T Gumbo,2 H Mcilleron1 1University of Cape
Town, Cape Town, South Africa; 2Baylor Research Institute,
Dallas, TX, USA; 3London School of Hygiene & Tropical
Medicine, London, UK; 4UNICEF/UNDP/World Bank/WHO
Special Programme on Research & Training in Tropical
Diseases (TDR), World Health Organization, Geneva,
Switzerland. Fax: (27) 214 481 989. e-mail:
smywyn001@myuct.ac.za
Background: Between patient pharmacokinetic (PK)

variability could lead to poor treatment outcomes. We
identified predictors of 2 month culture (defined as
outcome) in adults with newly diagnosed pulmonary
tuberculosis (TB) in the OFLOTUB phase 3 randomized
controlled trial at clinics in 3 African countries.
Design/Methods: Classification and regression tree
(CART) analyses was used to identify factors predictive
of outcome and their interactions. Potential predictors of
outcome included peak concentration (CMAX) and area
under the steady state concentrationtime curve (AUC)
for rifampicin (RMP), isoniazid (INH), pyrazinamide
(PZA), ethambutol (EMB) & gatifloxacin (GFX), age,
sex, body mass index, HIV status, baseline cavitation and
number of lung zones affected, baseline sputum smear,
and regimen standard including RMP, INH, EMB &
PZA vs. test with GFX replacing EMB. Ten-fold cross
validation, the receiver operating characteristics and
parsimony were used in model selection.
Results: CART identified steady state RMP AUC .33.45
mg*h/L and age ,30 years as best predictors of outcome
in the standard and test arms, respectively. Patients in the
test arm were 6.1 times more likely to have good
outcome if they were under 30 years of age (95%CI: 1.919.4). The odds of a good outcome in the control
regimen was 3.6 (95%CI: 1.6-8.2) when the steady state
RMP AUC was above 33.45 mg*h/L which corresponds
to a CMAX of 7.8 mg/L. Notably, CART analysis
demonstrated GFX exposure to have at best a relative
variable importance of 30% compared to the 74% of
RMP. Correspondingly, neither steady state GFX AUCnor CMAX were associated with outcome. Antagonism
between moxifloxacin and RMP demonstrated in hollow-fiber system model (HFS) experiments could explain
failure of RMP exposure to predict outcome in the test
regimen. Supporting this, the association between RMP
exposure and outcome was significantly less in the test
regimen, which included GFX. Patients with RMP steady
state AUCs in the top quartile were 6 times (95%CI: 222) vs. 1.1 times (95%CI: 0.4-3.5) more likely to achieve
OA-386-05 TB treatment outcomes among

participants in the XTEND trial
K Fielding,1 K Mccarthy,2,3 S Ginindza,2 V Chihota,2,3
S Charalambous,2,3 G Churchyard,1,2,3 A Grant1,3 1London
School of Hygiene & Tropical Medicine, London, UK, 2The
Aurum Institute, Johannesburg, 3University of the
Witwatersrand, Johannesburg, South Africa. Fax: (44) 20
7636 8739. e-mail: katherine.fielding@lshtm.ac.uk
Background: South Africa has replaced sputum microscopy with Xpert MTB/RIF as the first-line diagnostic test
for TB. The XTEND study, a pragmatic cluster
randomized trial, evaluated the effect of Xpert on patient
outcomes in persons with suspected TB. We report on TB
treatment outcomes of XTEND study participants.
Design/Methods: 20 laboratories in 4 provinces were
randomized to intervention (immediate Xpert implementation) or control (microscopy, with Xpert implementation deferred) arms. At 2 primary health clinics served by
each laboratory, a systematic sample of adults with
suspected TB, identified by clinic staff, was invited to
participate. Participants were followed up for 6 months.
TB treatment initiations and outcomes were identified
using paper and electronic TB registers and participant
self-report, excluding those with drug-resistant TB. Poor
treatment outcome was defined as death, lost to followup or failure (versus cure/completed treatment). The
intervention effect was estimated using methods for
cluster randomized trials, adjusting for baseline differences.
Results: Between June-November 2012, 4972 persons
with suspected TB were screened for the study. 4656
persons enrolled and followed for 6 months over which
time 522 (282 in microscopy and 240 in Xpert arm) had
started on TB treatment (excluding 19 with drugresistant TB). Of these, 43% were female, median age
36 yrs, 19% had body mass index (BMI) ,18kg/m2,
17% had previous TB and 69% were HIV-positive at
enrolment, of whom 26% had ever taken antiretroviral
therapy (ART). Among n 522, there was some
imbalance by arm for BMI, ART use and previous TB.
Treatment outcomes were known for 89% (466/522) of
participants; 87% (406) were cured/completed treatment
(78%; / 522), 8% (37) had died, 4% (17) were lost to
follow-up, 4 had transferred out and 2 had documented
treatment failure. By arm, 12.5% (31/249) and 11.7%
(25/213) had poor treatment outcome, in the microscopy
and Xpert arms; adjusted risk ratio (Xpert vs. microscopy) was 1.05 (95%CI: 0.70-1.59, P 0.8). Poor
treatment outcome was more common among those
HIV-positive, irrespective of ART status. Among those
S274
HIV-positive 33/43 (77%) of poor outcomes were

deaths.
Conclusion: Replacing smear microscopy with Xpert did
not alter outcomes of drug-sensitive TB. Mortality was
the leading cause of poor treatment outcomes; additional
interventions are needed to address this, particularly
among HIV-positive people.
12. MDR-TB: impacts of detection and

early treatment
OA-387-05 The yield of household contact
investigation of MDR-TB index cases in two
regions of Ethiopia
N Demmelash,1 D J Dare,1 M R Nigussie,2 I Jemal
Abdulahi,3 T Kebede,1,8 M Aseresa Melese,1 D Habte,8
Y Haile,4,5 P G Suarez1 1Management Sciences for Health,
Help Ethiopia Address the Low Tuberculosis Performance
(HEAL TB) Project, Addis Ababa, 2Amhara Regional Health
Bureau, Bahir Dar, 2 3Oromia Regional Health Bureau, Addis
Ababa, 4United States Agency for International Development
, Addis Ababa, Ethiopia; 5Management Sciences for Health,
Center for Health Services, Arlington, VA, USA. e-mail:
nhiruy@msh.org
Background and challenges to implementation: Systematic screening of high risk groups including contacts of
MDR-TB patients is a key strategy to detect MDR-TB
cases. However, there is limited experience with the
routine implementation of household contact investigation for MDR-TB index cases in low-income settings. In
this report, we present our experience in implementing
contact screening for MDR-TB index cases from two
regions of Ethiopia.
Intervention or response: Building on the experience of
contact screening for drug sensitive TB in Amhara and
Oromia regions of Ethiopia, the USAID funded MSH/
HEAL TB project supported health facilities to conduct
systematic TB contact screening for MDR TB patients
using a standardized tool. HEAL TB supported with
preparation, printing and distributing of standard
operating procedures (SoPs), oriented health workers
on their use, provided regular mentoring, and assisted
with data collection and reporting. According to the SoP,
treating clinicians requested all index cases to bring their
household contacts for screening at health facilities
where those with symptoms were further evaluated for
presence of TB. The site staff recorded patient data on a
logbook and did follow up screening on a quarterly basis.
Results and lessons learnt: During 2013-2014, we
screened 414 contacts of 130 index MDR-TB patients,
which gave an average of 3.2 contacts per index case. We
identified 30 active TB cases, either wouldnt have been
detected or detected late (23 MDR and 7 drug-sensitive).
The overall yield was 7.25% (95% Confidence interval
(CI); 5.12%-10.16%), with MDR-TB accounting for the
76% of the cases identified. The yield of MDR-TB was
thus 5.56%; CI, 3.73%-8.19% (5,555 MDR-TB cases
per 100 000 MDR-TB contact population) while that of
drug sensitive TB was 1.69% (CI, 0.82%-3.45%) which

is equivalent to 1691 per 100 000 MDR-TB contact
population. Of those diagnosed with MDR-TB, 4 (17%)
were children below 15 years of age. In comparison, the
yield of MDR-TB among non-contact presumptive
MDR-TB cases identified through the routine referral
system during the same time period was 2.94% (CI,
2.60%-3.34%).
Conclusions and key recommendations: The yield of
MDR-TB contact investigation was about twice higher
than the yield obtained from the screening among noncontact presumptive MDR-TB cases. We recommend
that MDR-TB contact investigation be a routine practice
of MDR-TB treating centers.
OA-388-05 Impact of a public-private mix

intervention on MDR-TB detection in Karachi,
Pakistan: experiences from the TB XpertW
Initiative
W M Awan,1 A Zaidi,1 A Malik,1,2 S Khowaja,1 U Khan,1
J Creswell,3 H Hussain,1,2 A Khan1 1Interactive Research &
Development, Karachi, 2The Indus Hospital, Karachi,
Pakistan; 3Stop TB Partnership, Geneva, Switzerland. e-mail:
wardah.awan@irdresearch.org
Background: Pakistan has the fourth highest multi-drug

resistant TB (MDR-TB) burden in the world. Casedetection for MDR-TB is low & treatment facilities
remain under-utilized due to operational limitations in
case-finding, MDR-TB suspect referrals and restrictive
diagnostic algorithms for Xpert South Africa, MTB/RIF.
Design/Methods: An active-case finding intervention
was carried out from July 2013 to March 2015 in
Karachi. Verbal screening for TB was carried out through
community screeners at a network of private providers in
low-income towns of Karachi. TB suspects were referred
to private diagnostic centers for X-ray and Xpertw MTB/
RIF testing. GeneXpert machines and technicians were
also provided to six public sector tertiary care hospitals,
one private laboratory and one PMDT site. Screening for
TB was carried out at various locations in the hospitals
and linkages were developed with neighboring public
and private hospitals, laboratories and pharmacies for
referrals of TB suspects. All TB suspects were tested
through GeneXpert irrespective of previous history of TB
treatment. Field staff supported sputum expectoration,
treatment initiation and case-registration activities.
Cases with rifampicin resistance were referred to one of
three PMDT sties in the city for further evaluation.
MDR-TB surveillance data was obtained from the NTP
to obtain case notifications from Karachi. A trend line
was fitted to historical case-notifications prior to the Q3
2013, and used to forecast notifications during the
intervention period to determine additionality of MDRTB cases.
Results: A total of 41 104 TB suspects were tested using
Xpert MTB/RIF assay of which 7260 were bacteriologically positive. Rifampicin resistance was found in 481
cases of which 425 were initiated on second-line drugs. A
total of 322 cases were confirmed for MDR-TB on drug-
susceptibility testing. Overall, 548 MDR-TB cases were

reported from the Karachi district during the intervention period , resulting in an trend adjusted additionality
of 216 cases, an increase of 67% over the baseline.
Amongst all MDR-TB patients detected, 68 cases
(12.4%) were primary MDR-TB.
Conclusion: A multi-faceted PPM intervention consisting
of private-sector engagement, screening at public sector
sites and testing through Xpert MTB/RIF significantly
increased MDR-TB case detection Karachi. Baseline
GeneXpert testing on treatment-naive cases yielded a
significant number of MDR-TB cases and should be
considered in high-MDR burden settings.
OA-389-05 Universalisation of rapid MDR-TB

testing is associated with more rapid initiation
of optimal patient therapy
M Tovar,1,2 E Ramos,1 T Valencia,1 A Valencia,3 Y Cortez,3
J Sandoval,3 J Agreda,3 C Evans4 1Innovation for Health
and Development, Lima, 2Innovacion Por la Salud Y el
Benefica
Desarollo, Asociacion
Prisma, Lima, 3DIRESA,
Regional National Tuberculosis Program, Callao, Peru;
4
Infectious Diseases & Immunity, Imperial College London,
and Wellcome Trust Imperial College Centre for Global
Health Research, London, UK. e-mail: marco.tovar@upch.pe
Background: Rapid multidrug resistant tuberculosis

(MDR-TB) testing, same-day molecular testing and
universalisation of MDR-TB testing are recommended
but are expensive and have uncertain effects on patient
care under operational conditions. In Peru, patients
usually commence empiric first-line therapy as soon as
TB is diagnosed, whilst MDR-TB testing is performed.
This first-line therapy is unlikely to cure MDR-TB so if
MDR-TB is diagnosed then the first-line therapy is
changed to specific second-line therapy for MDR-TB,
which is more likely to be curative if it commences earlier
in the patients illness. Objective: To study whether
universalization of MDR-TB and same-day MDR-TB
testing were associated with earlier initiation of appropriate second-line drug therapy for MDR-TB.
Design/Methods: From 2002-2014, all patients commencing TB therapy at 16 health post in shantytowns in
north Lima, Peru, were invited to give informed written
consent to participate in this study. We recorded the date
of starting empiric first-line therapy and the date of
starting second-line MDR-TB drugs, if they were
administered. According to NTP guidelines, in 2002
S275
only TB patients with known MDR-TB risk factors had

MDR-TB testing and this was done in the Peruvian
National Institute of Health (NIH) with the indirect agar
plate proportions assay that sometimes took months to
complete. From 2003-2008, a research project in the
Research and Development Laboratory of Universidad
Peruana Cayetano Heredia (LID) additionally offered all
patients with smear-positive TB Microscopic-Observation Drug-Susceptibility (MODS) testing, that takes 7-21
days. From 2008 to now, according to NTP guidelines,
all patients had MODS tests in the NIH laboratory. Since
early 2014, a research project in the LID additionally
offered all patients Xpertw MTB/RIF, which provides
same-day rifampicin-susceptibility testing, a surrogate
for MDR-TB testing.
Results: We recruited 5251 pulmonary TB patients and
368 of them initially received empiric first-line therapy
that was subsequently changed to second-line therapy for
MDR-TB. For these 368 patients, the delay from
initiation of inappropriate empiric first-line therapy for
initially unrecognised MDR-TB until change to appropriate second-line drugs decreased 4-fold from median
239 days in 2002 to 50 days in 2013 (P , 0.0001). The
introduction of Xpert same-day MDR-TB testing was
associated with a further halving of the delay until
appropriate second-line therapy for MDR-TB after
median 25 days in 2014 (P 0.006, Figure).
Conclusion: Both the universalisation of rapid MDR-TB
testing and the implementation of same-day molecular
rifampicin susceptibility testing, in association with
strengthening of the Peruvian National TB program,
were associated with statistically significant improvements in time to appropriate patient therapy.
S276
OA-390-05 The drug-resistant TB treatment gap

and treatment initiation delays in South Africa:
impact of XpertW implementation
implementation on RR TB will only be realised if care

linkage and scale up is improved.
H Cox,1 L Dickson-Hall,1 N Ndjeka,2 A Grant,3 A Vant

Hoog,4,5 W Stevens,6,7 J Workman,1 M Nicol1,8 1University
of Cape Town, Cape Town, 2National Department of Health,
Pretoria, South Africa, 3London School of Hygiene & Tropical
Medicine, London, UK; 4Amsterdam University, Amsterdam,
5
Amsterdam, Netherlands; 6University of the Witwatersrand,
Johannesburg, 7National Health Laboratory Services,
Johannesburg, 8National Health Laboratory Services, Cape
Town, South Africa. e-mail: lindydickson4@gmail.com
OA-391-05 Emergence of additional resistance

during standardised MDR-TB treatment
Background: Globally, the majority of drug-resistant

(DR) TB cases remain undiagnosed, and among those
that are, access to appropriate second-line treatment is
limited. South Africa has improved rifampicin-resistant
(RR) TB case detection from 7386 in 2010 to 26 023
cases in 2013 through universal access to RR-TB
diagnosis with the Xpertw MTB/RIF test, but only
41% were reported to access second-line treatment by
the National Department of Health in 2013.
Methods: To assess the extent of the gap between
diagnosis and treatment initiation and the time to
receiving the first dose of a standardised second line
regimen for RR-TB, we retrospectively identified two
nationally representative cohorts of newly diagnosed
RR-TB cases in 2011 (pre-Xpert) and 2013 (post-Xpert)
from laboratory records. Cases were followed through
registers, health facility records and interviews to
determine treatment initiation, timing and pre-treatment
loss to follow-up (including death).
Results: Among 2698 cases in 2011 and 2696 cases in
2013, 575 and 1154 initiated treatment within 28 days
respectively, following specimen registration at the
laboratory. This time to treatment is considered to be a
reasonably desirable period in developing countries. By 6
months 1,653 (61%) in 2011 and 1,863 (69%) (P ,
0.001) in 2013 had initiated an MDR treatment regimen,
although this proportion varied significantly across the 9
South African provinces (range 52-82%), and was
highest among provinces with greater access to decentralized treatment. In the 2013 cohort, 53% of cases who
initiated treatment were recorded on the national DR-TB
treatment register. The median time to treatment
initiation was 59 days in 2011 and 26 days in 2013 (P
, 0.001). Time to treatment in 2013 also varied
significantly across provinces (range of medians by
province was 19-51 days in 2013). Among the 1631
(60%) cases in 2013 in whom RR was diagnosed by
Xpert, the median time to treatment was 20 days,
compared to 43 days for those diagnosed with other
methods.
Conclusion While the treatment gap is lower than
routine reporting suggests, the proportion of diagnosed
cases not starting treatment remains high. However,
given the dramatic increase in case detection with Xpert
implementation, many more patients are receiving
treatment overall, and delay in treatment initiation has
been substantially reduced. The full benefit of Xpert
M Mphahlele,1 R M Warren,2 E Streicher,2

M Van Der Walt,3 P Van Helden,2 K Jacobson4 1Jhpiego
South Africa, Pretoria, 2University of Stellenbosch, Cape
Town, 3South African Medical Research Council, Pretoria,
South Africa; 4Boston University School of Medicine, Boston,
MA, USA. e-mail: mphahlelematsie@gmail.com
Background: To address the increasing incidence of

MDR-TB, the South African National TB Program
implemented a standardised treatment regimen in 2000
consisting of a 4-month intensive phase with the 5 drugs,
pyrazinamide, ethambutol, ethionamide, ofloxacin, and
either amikacin (AMK) or kanamycin (KAN), followed
by a 12- to 18-month continuation phase removing
AMK/KAN. This policy also advocated for MDR-TB
treatment to occur in hospitals until culture conversion.
The objective of the study was to assess the individual
patient impact of implementing a standardized MDR-TB
treatment regimen in two South African provinces, the
Eastern Cape (EC) and North West (NW), by retrospectively quantifying the acquisition of drug resistance to
second-line drugs in serial cultures.
Design/Methods: We enrolled 556 MDR-TB patients
from the Eastern Cape and North West Provinces. We
investigated whether baseline DST pattern, HIV status,
and treatment outcomes were associated with the
acquisition of drug resistance. To differentiate reinfection
from persistence, spoligotyping was performed on the
MDR-TB isolates obtained at the time of treatment
initiation and compared with last available isolate.
Results: Sixty-four percent (n 356) of patients had poor
treatment outcomes (died, defaulted, failed treatment).
Of 556 MDR-TB patients, 106 (64 females and 42 males)
had DST and spoligotyping results. Of these, 48 had an
initial and last available isolate. In 12 (25%) patients the
spoligotype patterns changed suggesting reinfection. In 9
(18%) of these patients the reinfecting strain had a
resistance profile that showed additional resistance
relative to the initial isolate. Of the remaining 36
patients, 13 (27%) acquired additional resistance while
on treatment as their isolates remained identical during
the course of disease.
Conclusion: Resistance amplification was observed in
46% of MDR-TB patients with available isolates.
Implementation of standardised MDR-TB treatment in
the absence of knowledge of baseline resistance in
circulating MDR-TB strains appears to increase the risk
of resistance amplification during treatment. Furthermore, hospitalisation of MDR-TB patients for treatment
appears to increases the risk of nosocomial reinfection
with strains that are more resistant than the initial
infecting strains. This work is supported by the South
African Medical Research Council.
S277
OA-392-05 An assessment of site readiness of

hospitals in South Africa to decentralize MDRTB services
OA-393-05 Variations in and risk factors for

drug-resistant tuberculosis in sub-Saharan
Africa: a systematic review and meta-analysis
M Mphahlele,1 A Peters,1 J Pienaar,1 N Ndjeka,2 J Farley3

1
Jhpiego South Africa, Pretoria, 2Department of Health,
Pretoria, South Africa; 3Johns Hopkins University, Baltimore,
MD, USA. e-mail: jacqueline.pienaar@jhpiego.co.za
G Yiga1 1Save a Life Foundation, Kalangala, Uganda. e-mail:

igerfrey@yahoo.com
Background and challenges to implementation: Prior to

2011, national policy in South Africa (SA) mandated
hospitalization of all drug-resistant TB (DR-TB) patients.
This resulted in long delays, inadequate bed capacity and
infection control and patients inconvenience.To address
these challenges the DR-TB programme plans to
decentralize DR-TB management to the community.
Intervention or response: The purpose of this descriptive
study was to evaluate clinical sites for DR-TB readiness
to ensure quality services prior to decentralization.Fiftytwo facilities (8 Community Health-Care centers, 3
Primary Health-Care clinics, 9 MDR-TB hospitals plus
33 district-based hospitals) were evaluted from July 2014
to March 2015. Facilities were prioritized based on: 1)
MDR-TB burden 2) MDR-TB quality care for sites
providing care and 3) absence of any decentralized unit
in the district. Nineteen of 53 (36%) facilities were
providing DR-TB services. A standardized decentralization tool was used to collect information on health
services and infrastructure status.Key informants were
interviewed and healthcare worker practices were
observed. Following the initial assessments, quality
improvement plans for each facility were developed.
Results and lessons learnt: Facility readiness to decentralize DR-TB services varied between the 52 facilities.
Overall, 83% had access to laboratory services for
diagnosis of DR-TB. Most sites had access to X-ray
services (79%) but very few (25%) had audio equipment
to assess ototoxicity. Thirty (58%) had DR paper-based
data reporting tools, of which five facilities (17%) had a
laptop for electronic data capturing, however none of
these sites were reporting data electronically.A limited
number of facilities had dedicated doctors (52%) and
nursse trained in initiating (10%) DR-TB. There was a
slow roll out of decentralization at PHCs level, among
the three primary health-care clinics only one had a
roving doctor who initiated treatment once a week.
Conclusions and key recommendations: Limitations
remain that may slow decentralization of DR-TB to the
community.While the majority of facilities meet minimum staff and infrastructure requirements, training on
DR-TB was inadequate and should be fast-tracked.Equipment for monitoring ototoxicity should be prioritized for implementation.Task-sharing with more nurse
involvement will be key to success, but is slow to
occur.Quality monitoring including electronic drugresistant registers should be implemented.
Background: Prevalence of multidrug resistant tuberculosis (MDR-TB), defined as in vitro resistance to both
rifampicin and isoniazid with or without resistance to
other TB drugs, in sub-Saharan Africa (SSA) is reportedly
low compared to other regions. These estimates are
based on data reported to the World Health Organization (WHO) on drug resistance surveys, which may
suffer from a reporting bias. We set out to evaluate the
variation in prevalence of drug resistant tuberculosis
(DR-TB) and its determinants across SSA countries
among new and previously treated TB patients.
Design/Methods: The aim was to perform a systematic
review and meta-analysis of DR-TB prevalence and
associated risk factors in SSA. PubMed, EMBASE,
Cochrane and bibliographies of DR-TB studies were
searched. Surveys at national or sub-national level, with
reported DR-TB prevalence (or sufficient data to
calculate a prevalence) to isoniazid (INH), rifampicin
(RMP), ethambutol (EMB), and streptomycin (SM)
conducted in SSA excluding the Republic of South
Africa, published between 2003 and 2013 with no
language restriction were considered. Two authors
searched and reviewed the studies for eligibility and
extracted the data in pre-defined forms. Forest plots of all
prevalence estimates by resistance outcome were performed. Summary estimates were calculated using
random effects models, when appropriate. Associations
between any DR-TB and MDR-TB with potential risk
factors were examined through subgroup analyses
stratified by new and previously treated patients.
Results: A total of 726 studies were identified, of which
27 articles fulfilled the inclusion criteria. Studies reported
drug susceptibility testing (DST) results for a total of 13
465 new and 1776 previously treated TB patients. Pooled
estimate of any DR-TB prevalence among the new cases
was 12.6% (95%CI 10.6-15.0) while for MDR-TB this
was 1.5% (95%CI 1.0-2.3). Among previously treated
patients, these were 27.2% (95%CI 21.4-33.8) and
10.3% (95%CI 5.8-17.4%), respectively. DR-TB (any
and MDR-TB) did not vary significantly with respect to
study characteristics. Variations between any DR-TB and
MDR-TB with potential risk factors were examined
through subgroup analyses stratified by new and
previously treated patients.
Conclusion: The reported prevalency of DR-TB in SSA is
low compared to WHO estimates. MDR-TB in this
region doesnot seem to be driven by the high HIV
prevalence rates.
S278
OA-394-05 A rifampicin-resistant diagnosis on

XpertW MTB/RIF frequently leads to initiation
of a weakened standardized MDR-TB regimen
K Jacobson,1,2 M Barnard,1 M Buckley,2 E Streicher,1
T Dolby,3 L Scott,4 A Van Rie,5,6,7 R M Warren1
1
Stellenbosch University, Tygerberg, South Africa; 2Boston
University School of Medicine / Boston Medical Center,
Boston, MA, USA; 3National Health Laboratory Service, Cape
Town, 4National Health Laboratory Service, Johannesburg,
South Africa; 5University of Antwerp, Antwerp, Belgium;
6
Gillings School of Global Public Health, University of North
Carolina, Chapel Hill, NC, USA. e-mail: marywb@bu.edu
Background: Xpert MTB/RIF detects Mycobacterium

tuberculosis complex (M. tuberculosis c) and rifampicin
(RIF) resistance, enabling physicians to rapidly initiate
multidrug resistant tuberculosis (MDR-TB) treatment.
South African guidelines instruct that RIF-resistant (RR)
patients start standardized MDR-TB therapy with
pyrazinamide (PZA), ethionamide (ETH), kanamycin
(KAN), moxifloxacin and terizidone before knowledge
of resistance to these drugs. We aimed to quantify the
time delay to reporting of second-line (SL) drug
susceptibility test (DST) results and the proportion of
patients placed on suboptimal therapy.
Methods: We retrospectively identified all patients
diagnosed with RR TB by Xpert at the National Health
Laboratory Service, Cape Town, South Africa, between
November 2011 and June 2013. We also prospectively
collected a sample of RR TB patients diagnosed in
Eastern Cape, Free State, and Gauteng Provinces. We
calculated the time from specimen collection to reporting
of SL DSTs and susceptibility patterns. We performed
targeted DNA sequencing when phenotypic susceptibility patterns were not available.
Results: During this time period, 904 samples were RR
by Xpert in the Western Cape (WC), of which 466
(55.8%) had SL DST results. For the 496 RR samples
collected in the other 3 provinces, SL DST results were
available for 141 (28.4%) patients. Median time from
sputum taken to second-line DST results was 54 days (IQ
range 27-259) in the WC and 60 days (IQR 42-77) in
other provinces. Time to DST results was shorter for
smear positive samples (P , 0.0001). Of the 171 WC
patients with DST results for PZA, ETH, KAN and
ofloxacin, 47% would be considered to have started an
ineffective regimen (, 4 effective drugs) with only 1
likely effective drug in 8%, 2 in 9%, and 3 in 30%. In the
other 3 provinces, 26% of 66 patients would be
considered to have started an ineffective regimen with
only 1 effective drug in 6%, 2 in 6% and 3 in 9%. This
difference is driven by rates of ETH resistance (6.6% vs.
60.5%).
Conclusion: Under current programmatic conditions,
many patients diagnosed with RR TB by Xpert are
started on a weakened 5-drug MDR regimen, with the
information to adjust therapy only available after 8
weeks or never at all. SL DSTs were less frequently
available in the 3 provinces where only 1 specimen is
requested at diagnosis rather than 2. Such time delays
and lack of SL results risk worsened outcomes and may

fuel resistance amplification in the population.
13. LTBI in high- and low-burden settings

OA-395-05 Early biomarkers associated with
progression of latent tuberculosis infection to
clinically active disease: a longitudinal cohort
study
N Rakotosamimanana,1 V Richard,1,2 V Raharimanga,1
T M Doherty,3 B Gicquel,4 A Zumla,5
V Rasolofo Razanamparany1 1Pasteur Institute of
Madagascar, Antananarivo, Madagascar; 2Pasteur Institute
of Dakar, Dakar, Senegal; 3GSK, Copenhagen, Denmark;
4
Pasteur Institute, Paris, France; 5University College London,
London, UK. e-mail: niaina@pasteur.mg
Background: An estimated one third of the worlds

population is latently infected with Mycobacterium
tuberculosis, of whom 10% will likely progress to active
tuberculosis (TB) disease at some time in their lives. We
performed a longitudinal cohort study on blood and
infection markers to address the need of sensitive and
clinically practical prognostic markers of at-risk individuals.
Design/Methods: A cohort of household contacts of
newly diagnosed pulmonary TB patients were recruited,
blood analyzed and investigated for TB symptoms at
different periods for up to two years in Antananarivo,
Madagascar. Controls were matched healthy individuals
from the same community. Immune status was assessed
by TST, and full blood counts (FBC) were determined.
Results: 595 HIV-negative individuals: 104 newly
diagnosed pulmonary TB patients, 305 household
contacts and 186 controls were enrolled. Age negatively
correlated with risk of progression to active disease
(OR3.98, 95CI 1.05-14.99; P , 0.05, for ,16 years).
Age paired analysis showed that in contacts, an elevated
monocyte count in peripheral blood was associated with
risk of developing active TB (P , 0.001). Further, of the
12 contacts that later developed TB, 10 (83%) had a TST
induration .14mm at the entry of the study (OR6.1,
95CI 1.31-28.32; P 0.02). A monocyte percentage
.7.4% together with a TST.14 mm was associated
with a greatly elevated risk of progression to TB disease
(OR14.13 [3.58-55.75], P , 0.001).
Conclusion: The routinely-used clinical tests, TST and
FBC may have potential as simple biomarkers for risk of
developing active TB in recently-exposed individuals.
Further evaluations of these biomarkers for identification
and treatment of these individuals are required.
OA-396-05 Understandings of tuberculous

infection, TB disease and isoniazid preventive
therapy in KwaZulu-Natal, South Africa
J Boffa,1,2 M Mayan,3 S Ndlovu,4 R Cowie,1,3 R Sauve,1
D Wilson,5 D Fisher1,3 1University of Calgary, Calgary, AB,
Canada; 2University of KwaZulu-Natal, Durban, South Africa;
3
University of Alberta, Edmonton, AB, Canada; 4Izimbali
Zesizwe, Pietermaritzburg, 5Edendale Hospital,
Pietermaritzburg, South Africa. e-mail:
jody.boffa@ucalgary.ca
Background: In 2011 the South African Department of

Health implemented 6-month isoniazid preventive therapy (INH) to prevent tuberculosis (TB). All people living
with HIV (PLWH) were eligible if active TB symptoms
were absent. Prior to this, there was no local language
equivalent for latent TB infection, as pulmonary TB was
only recognised once it had become infectious. We aimed
to describe community understandings of TB infection
and disease in light of preventive strategies and determine
perceptions of INH following implementation.
Methods: Guided by community-based research principles and the ethnographic method, we worked with three
communities of the Edendale Hospital catchment area,
where predominantly underemployed black South Africans reside in poor housing; HIV prevalence is 15% and
TB incidence is 1%, with 7% multi-drug resistance. We
conducted eight focus group discussions (FGDs) (57
participants total) with three population pools: PLWH to
learn about potential users perspectives, people without
HIV to glean peer perceptions, and taxi drivers to capture
male working class perspectives of TB and INH. FGDs
took place in isiZulu and were later translated into
English for qualitative content analysis in Nvivo 10.
Community Advisory Teams were consulted to assist
with and validate the interpretation of themes.
Results: Health was generally considered a lack of
disease or pain. Communities were thought to be
healthier in the past, when HIV, TB, and diabetes were
uncommon. Dirty environments or particular lifestyle
behaviours were thought to contribute to illness.
Participants recognised cough and weight-loss as TB
symptoms and knew the link and distinction between
HIV and TB. Only participants in the PLWH group had
heard of INH and grasped that some kind of pre-disease
TB state must exist, though none knew of the terms latent
TB infection or TB elele (sleeping TB). A minority of
PLWH had been offered or initiated on INH, although
members of all three populations inquired about
eligibility for enrolment. Barriers to clinic service
included wait times, poor communication with clients,
and feelings of inferiority preventing questions or
clarification.
Conclusion: Overall, common TB symptoms and the link
to HIV were well recognized. Some PLWH were aware of
INH, but knowledge of latent TB infection was
uncommon. A focus on patient-centred care and more
opportunities to access information about TB prevention
may improve uptake of future TB prevention regimens.
S279
OA-397-05 Optimal testing choices for latent

tuberculosis infection in high-risk populations
in a low incidence country: a latent class
analysis strategy
S Ghosh,1 M Johnson,2 D Katz,1 G Fanning-Dowdell,3
R Punnoose,3 C Ho,1 J Stout2 1U.S. Centers for Disease
Control and Prevention, Atlanta, GA, 2Duke University
Medical Center, Durham, NC, 3Northrup Grumman, Atlanta,
GA, USA. e-mail: smita.ghosh@cdc.hhs.gov
Background: The absence of a gold standard to diagnose

latent tuberculosis infection (LTBI) complicates efforts to
estimate sensitivity and specificity of commercially
available tests. Understanding the tests performance
characteristics is especially important in low incidence
countries where targeted testing and treatment of highrisk populations are key for tuberculosis control and
elimination. Researchers from the TB Epidemiologic
Studies Consortium (TBESC) conducted a head-to-head
performance comparison of the three widely available
tests among high-risk populations in the United States.
Design/Methods: Persons at high risk for LTBI (e.g.,
foreign-born, close contacts) or for progression to active
TB disease (e.g., HIV-infected) were enrolled from July
2012September 2014 at 14 urban sites and tested with
all 3 tests: the tuberculin skin test (TST) and two
interferon-gamma release assays (Quantiferonw) TB
Gold In-Tube (QFT) using .0.35 IU/ml as positive,
and T-SPOT.TB (TSPOT) using .8 spots as positive).
Latent class analysis (LCA), a statistical technique
suitable for estimating group membership in the absence
of a gold standard (e.g., presence/absence of true LTBI),
was used to 1) identify subgroups where test performance varied substantially and 2) compare LTBI
prevalence estimates, sensitivities, and specificities for
U.S.-born and foreign-born, HIV-infected and HIVuninfected and two age categories (,5 years and 75
years old).
Results: The 11 386 participants included 9057 (80%)
foreign-born .5 years old, 514 (4.5%) foreign-born ,5
years old, 1461 (13%) HIV-infected, and 354 (2.5%)
members of other high-risk groups. Test characteristics
differed significantly (P , 0.001 for all comparisons)
among key sub-groups (Table). QFT was consistently
most sensitive in all sub-groups and TSPOT most
specific, but TSPOT had low sensitivity in children ,5
years old and in HIV-infected persons. When classifying
borderline TSPOT results (57 spots) as positive, overall
results remained the same. TST was sensitive in all groups
except HIV-infected and U.S.-born persons, but had low
specificity in foreign-born persons.
Conclusions: Based on LCA, QFT and TSPOT performed
comparably in most high-risk groups and had greater
specificity than TST in foreign-born persons, particularly
among young children. However, TSPOT was considerably less sensitive in children ,5 years old and in HIVinfected persons. QFT may be a better choice for testing
these vulnerable sub-groups.
S280
symptom during the course of treatment, 94.6% vs.

88.8%, P 0.02.
Conclusions: INH-RPT is a viable alternative to INH
monotherapy for treating LTBI in the BOP with
significant practical advantages (reducing DOT doses
from 76 to 12). The treatment completion rate was high
with low rates of adverse reactions. The shorter
treatment duration might facilitate treatment of short
stay inmates. Based on the pilot evaluation, INH-RPT is
now the standard recommended regimen for LTBI
treatment in the BOP.
OA-398-05 High completion rate for 12 weekly

doses of isoniazid and rifapentine as treatment
for latent tuberculosis infection in the Bureau
of Prisons, USA
OA-399-05 Implementation of QuantiFERONW

TB Gold In-Tube test for diagnosing latent
tuberculosis among newly diagnosed HIVinfected patients in South Africa
M Lobato,1 S Lang,2 L Sosa,2 S Wheeler,3 S Bur,3

N Kendig3 1U.S. Centers for Disease Control and Prevention,
Atlanta, GA, 2Connecticut Department of Public Health,
Hartford, CT, 3Bureau of Prisons, Washington, DC, USA. Fax:
(1)860 509 7743. e-mail: mark.lobato@ct.gov
J Golub,1 L Lebina,2 C Qomfu,2 S Chon,1 S Cohn,1

K Masonoke,2 E Variava,2,3,4 N Martinson1,2 1Johns
Hopkins University, Baltimore, MD, USA; 2Perinatal HIV
Research Unit, Johannesburg, 3Klerksdorp-Tshepong Hospital
Complex, Klerksdorp, 4North West Department of Health,
Klerksdorp, South Africa. Fax: (1) 443 287 7955. e-mail:
jgolub@jhmi.edu
Background: About 4% of tuberculosis (TB) cases

reported in the United States occur among persons
incarcerated at the time of diagnosis. The Federal Bureau
of Prisons (BOP), which houses over 200 000 inmates, is
in a unique position to provide latent TB infection (LTBI)
treatment to prevent development of TB disease. Past
efforts to treat LTBI with isoniazid (INH) monotherapy
have been hampered by poor treatment completion rates.
Recent studies have shown that 12 once-weekly doses of
INH and rifapentine (RPT) given by directly observed
therapy (DOT) is not inferior to 9-months of INH in
preventing development of TB disease.
Methods: Seven BOP facilities participated in the pilot
implementation of INH-RPT for LTBI treatment. Treatment monitoring data were entered in the BOP Electronic
Medical Record. DOT visits, weekly symptom screens,
and demographic, behavioral, and medical risk factors
were collected for inmates treated during July 2012
February 2015.
Results: A total of 463 inmates started treatment using
INH-RPT. The median age was 36 years (range: 2071
years); 70% were male. The population was 76%
Hispanic, 12% non-Hispanic white, 8% black, 2%
Asian, and 1.5% American Indian. Two thirds of inmates
treated were foreign born and 20% were contacts to a
known TB case. Of 463 total inmates treated, 424
(91.6%) successfully completed treatment. Among the
39 (8%) inmates who stopped treatment, 17 discontinued due to an adverse event, 9 were transferred to
another facility or were released from BOP custody
before completion, 8 refused further treatment, and 5 did
not complete due to provider error. Signs or symptoms
associated with treatment discontinuation included
nausea/ vomiting, headache, abdominal pain, elevated
serum liver transaminases, and fever. Five patients who
stopped treatment due to symptoms successfully completed using either INH or rifampin. Inmates with no
symptom complaints were more likely to complete
treatment than those who experienced at least one
Background: Tuberculosis (TB) remains the leading killer

among people living with HIV (PLWH) in South Africa.
Recent guidelines state that all newly diagnosed HIVinfected patients in South Africa should, following
screening for active TB disease, receive a tuberculin skin
test (TST), and a determination of the duration of
recommended isoniazid preventive therapy (IPT) is
dependent upon the TST result and current antiretroviral
(ART) status. Evidence strongly suggests that implementation of TST in HIV clinics is suboptimal throughout the
country.
Methods: We are conducting a cluster-randomized trial
to compare the impact on IPT uptake of a strategy to
collect blood for QuantiFERONwTB Gold In-Tube
(QGIT) at time of CD4 blood draw for newly diagnosed
HIV-infected patients, with the current standard of care provision of TST. Fourteen public HIV clinics in the
North West Province were randomized to either QGIT or
TST only. Consenting patients were interviewed immediately following HIV diagnosis and baseline medical
record abstractions were conducted within 3 weeks of
diagnosis. Here we present characteristics of the first 443
enrolled in the study, and 213 QGIT results.
Results: Enrollment at the first TST clinic began in
November 2014 and the first QGIT clinic in January
2015. Through March 31 2015, 411 patients eligible for
QGIT or TST were enrolled; 225 in QGIT clinics and 186
in TST clinics. Median age was 33 (IQR 26-41), 64%
were female and median CD4 count was 303 cells/mm3
(IQR 161-476). Among females, 33% were pregnant at
time of HIV diagnosis. Among 225 QGIT eligible, 12
(5%) did not provide blood for testing. Among 213 with
QGIT results, 89 (42%) were positive, 90 (42%) were
negative and 34 (16%) were indeterminate. Excluding
indeterminate results, 50% were positive and 50% were
negative. Indeterminate rate ranged from 7% to 21%
across clinics. Patients with positive and negative QGIT
results did not differ by age, sex or smoking or alcohol

use. Median CD4 counts for QGIT positive, negative and
indeterminates were 318, 259 and 160, respectively.
Conclusion: Approximately half of all newly diagnosed
HIV-infected patients with a QGIT result were positive,
though 15% had an indeterminate result. Though
operational comparisons to TST cannot yet be made in
this cluster randomized trial, our preliminary results
suggest that implementation of QGIT into public HIV
clinics in South Africa may be feasible, but indeterminates are a potential limitation.
S281
matory test results. If the correlation between initial and

confirmatory test was high, outcomes were very similar.
If uncorrelated, confirmatory testing even at a threshold
of 0.35 IU/ml reduced the ten-year false positive
percentage to 1.2% (Figure).
Conclusion: Using QFT for annual serial testing among
HCWs in a low-incidence setting may lead to tremendous
over-diagnosis of conversions. Using a higher cutoff for a
QFT conversion or confirming positive results with a
second test may reduce the false positive rate.
OA-400-05 Expected outcomes of serial testing

with interferon-gamma release assays in North
American health care workers: a Markov model
A Zwerling,1 M Moses,1 A Cattamanchi,2 C Denkinger,3
N Banaei,4 S Kik,5 D Dowdy,1 M Pai6 1Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD,
2
University of California San Francisco, San Francisco, CA,
USA; 3FIND, Geneva, Switzerland; 4Stanford University,
Stanford, CA, USA; 5KNCV, The Hague, Netherlands; 6McGill
University, Montreal, QC, Canada. e-mail:
alice.zwerling@mail.mcgill.ca
Background: Interferon-gamma release assays (IGRAs)

have been adopted as an alternative to the tuberculin skin
test (TST) for serial testing in healthcare workers
(HCWs) in low-incidence settings. IGRAs are known to
have multiple sources of variation, and show high rates
of conversions and reversions. The net effect of such
variability on serial testing remains unknown.
Design/Methods: We used a micro-simulation Markov
model to project diagnostic outcomes in a simulated
HCW cohort under different serial screening strategies
after accounting for the major sources of variability in
serial QuantiFERONw-TB Gold In-Tube (QFT) results.
An initial QFT value was assigned using cross-sectional
data; at each subsequent year (in a 10 year cycle) a new
QFT value was calculated based on 1) random intraindividual variation, 2) TST-induced boosting, 3) variation in blood volume and 4) pre-analytical delay.
Distributions of each source of variability were derived
through a systematic review and from best available
evidence. We then projected ten-year outcomes and
required test volumes under different cutoffs for defining
a positive test and under different assumptions about
whether initial positive results required confirmation
with another positive QFT result before initiating
preventive therapy.
Results:In a high-risk setting (1.0% annual risk of TB
infection), a strategy in which a negative to positive
change with threshold QFT value of 70.35 IU/ml would
lead to 34.1% of the population testing positive by the
end of 10 years, of whom only 8.4% were truly infected
at the time of their positive test. Increasing the cutoff for
a positive repeat test to 1.0 IU/ml reduced the proportion
of the population testing false-positive in ten years from
25.7% to 2.3% (Figure), but at the expense of reducing
the proportion of true TB infections diagnosed from 72%
to 60%. The value of a confirmatory test depended
strongly on the correlation between initial and confir-
OA-401-05 National roll-out of LTBI screening in

England
M Loutet,1 D Zenner1 1Public Health England, London, UK.
e-mail: miranda.loutet@phe.gov.uk
Background and challenges to implementation: Latent

TB infection (LTBI) screening and treatment is an
essential tool to support the WHO Post-2015 Global
TB Strategy. The Collaborative TB Strategy for England
2015-2020 between Public Health England (PHE) and
the English National Health Service (NHS) aims to make
significant advances in TB control in England. National
systematic implementation of new entrant LTBI screening is a fundamental component in order to achieve the
strategy ambitions. The aim of this paper is to describe
initial outcomes of this programme.
Intervention or response: All migrants, aged 16-35 years
who have entered the UK from a high incidence country
(7150/100 000 and sub-Saharan Africa) within the last
5 years and had been previously living in that high
incidence country for 76 months are eligible for
screening. Individuals are identified for screening at GP
registration, screened in primary care and treated in
primary or secondary care. PHE collects data on LTBI
screening, including the social and disease risk factors for
TB, and the screening and treatment process in order to
allow for monitoring and evaluation of the programme.
S282
Results and lessons learnt: Initial results are available

from the London borough of Newham (TB rate of 113.7/
100 000) which has been screening for LTBI since July
2014 in line with the national strategy guidelines. 61
General Practices in Newham have identified 5683
eligible individuals, of which 3272 (57.6%) had IGRA
(Quantiferon) tests. 866 of 3272 (26.5%) have had a
positive result. Additionally, this programme has the
potential to identify active TB cases early. In Newham, 8
active TB cases were detected and referred to chest clinic.
Conclusions and key recommendations: Newham illustrates the feasibility of LTBI screening among areas of
high TB incidence, where General Practitioners are
engaged to screen all those eligible. However, it is
difficult to know how treatment uptake will vary across
England. This programme has the potential to have an
enormous impact on TB rates in the UK. The data will
ultimately be linked to notified active TB cases to provide
evidence of the effectiveness of LTBI screening to reduce
the number of active TB cases in a non-study setting.
OA-402-05 A population-based cohort study of

trends in BCG effectiveness against
tuberculosis with time since vaccination in
Norway
P Nguipdop-Djomo,1,2,3 E Heldal,1 K Rnning,1
I Cappelen,1 L Rodrigues,2 I Abubakar,3 P Mangtani2,3
1
Norwegian Institute of Public Health, Oslo, Norway; 2
London School of Hygiene & Tropical Medicine, London,
3
Public Health England, London, UK. e-mail:
patrick.nguipdop-djomo@lshtm.ac.uk
Background: Little is known on how BCG vaccine

protection against tuberculosis (TB) changes with time
after vaccination, particularly beyond 10 years.
Design/Methods: To assess long-term BCG effectiveness
(VE) against tuberculosis, we obtained historical Tuberculin Skin Test (TST) and BCG information from
participants of the last round 1962-1975 of a Norway
mandatory mass TB screening and BCG vaccination
programme, and linked it to National TB register, the
National Population and Housing Censuses (1960/1970)
for socio-economic position (SEP) and the Population
Register for emigrations and deaths. Norwegian-born
TST negative subjects (eligible for BCG vaccination)
were followed-up from the last screening to their first
episode of TB or censoring. Using Cox regressions, we
estimated VE for up to 40 years since BCG, adjusted for
age as a time-varying effect, calendar time, county-level
TB rates and demographic and SEP indicators.
Results: Follow-up was on average over 40 years, for 83
421 unvaccinated and 297 905 BCG vaccinated subjects,
with 260 first TB episodes. Tuberculosis rates were 3.3
per 100 000 person-years in unvaccinated and 1.3 per
100 000 person-years in vaccinated subjects. Adjusted
average VE over the 40 years follow-up was 49%
(95%CI: 26% to 65%). After controlling for confounding and excluding TB in the first- two years after
screening, VE was estimated for the following periods
after vaccination: up to 9 years, 61%(95%CI: 24% to
80%), 10-19 years, 58%(27% to 76%), 20-29,
38%(32%to71%), and 30 to 40, 42% (24to73%).

Specific estimates against pulmonary TB for the same
time intervals were respectively 67% (27% to 85%),
63% (32% to 80%), 50% (19% to 79%) and 40%
(46% to 76%).
Conclusion: Our findings are consistent with a longerlasting BCG protection with average VE of 50% over 40
years, with waning in time and estimated protection of
40% 30 to 40 years after vaccination. BCG may be more
cost-effective than previously estimated, with fewer
people needed to vaccinate per case prevented. Consideration should be given to the interaction of BCG with
new TB vaccine candidate.
14. From specimens to outcome: new

information for molecular diagnosis
OA-403-05 Diagnostic accuracy of the XpertW
MTB/RIF cycle threshold level to predict smear
positivity in respiratory samples: a systematic
review and meta-analys
B Lange,1,2 M E Balcells,3 R Blakemore,4,5 N Lemaitre,6
J J Palacios,7 J Teo,8 G Theron,9 K Kranzer10 1Centre for
Infectious Diseases/Centre of Chronic Immunodeficiency,
Freiburg, 2University Hospital Freiburg, Freiburg, Germany;
3
Pontificia Universidad Catolica de Chile, Santiago De Chile,
Chile; 4Division of Infectious Diseases, Newark, NJ, 5New
Jersey Medical School, Newark, NJ, USA; 6Universite de
Lille-Nord de France, Lille, France; 7Hospital Universitario
Central de Asturias, Oviedo, Spain; 8National University
Hospital, Singapore, Singapore; 9University of Cape Town,
Cape Town, South Africa; 10London School of Hygiene &
Tropical Medicine, London, UK. Fax: (49) 761 2701 8190.
e-mail: Berit.Lange@uniklinik-freiburg.de
Background: Following its global scale-up, Xpertw

MTB/RIF is now the most widely used molecular assay
for the rapid diagnosis of tuberculosis (TB). The number
of PCR cycles after which detectable product is generated
(cycle threshold value (CT)) has been shown to correlate
with mycobacillary burden as measured by time to
culture positivity using liquid culture systems and smear
microscopy. The aim of this study is to investigate the
association between Xpert MTB/RIF CT values and
smear status through a systematic review of the available
literature, and meta-analysis to establish a receiveroperating-characteristic (ROC) curve between CT values
and bacillary load.
Methods: Medline, Embase, Global Health, Current
Controlled Trials and the Cochrane database of systematic reviews was searched using a compound search
strategy. Studies including pulmonary samples and Xpert
MTB/RIF, smear microscopy and culture results were
potentially eligible. Studies on the association between
CT values and smear status or days to culture positivity
were included in the descriptive review. Authors of
potentially eligible studies were asked to provide
individual level data for meta-analysis. Mean CT value
difference between smear positive and negative samples
was calculated as well as ROC curves adjusting for

clustering by study and by patient.
Results: From 498 identified citations 74 were deemed
potentially relevant, 10 studies were included in the
descriptive review. Eight studies provided individual level
data including 4437 respiratory samples from 2631
patients. A total of 968 patients had positive Xpert MTB/
RIF results for at least one respiratory sample, 630
patients had positive results from two samples and 2
patients had positive results from 3 samples. The mean
CT value difference between smear positive and smear
negative samples was 7.6 (95%CI 6.9-8.2). ROC
analysis revealed an AUC of 0.85 (95%CI 0.82-0.87).
A CT value cut-off of 27.7 results in a sensitivity of 85%
(CI95 83-87) and specificity of 67% (95%CI 66-77) and
a CT value cut-off of 32 results in a sensitivity of 95%
(95%CI94-96) and a specificity of 35% (CI95% 30-41)
for smear-positive samples. Stratification by HIV led to a
slightly decreased sensitivity for HIV-positive patients
(see the Table).
Conclusion: This review confirms the strong association
between Xpert MTB/RIF CT values and smear status.
Thus CT value result may be a good replacement for
smear status of respiratory samples in certain contexts.
Table Area under the curve (AUC), sensitivity and specificity
with 95% confidence intervals (95%CI) for smear positivity
with different CT value cut-offs in the meta-analysis on
individual level data derived from 8 different studies
Cut
off
(CT SensiAUC* value) tivity CI95%
HIV positive
(216 samples
of 151
patients)
HIV negative
(611 samples
of 423
patients)
All (1574
samples of
968 patients)
Specificity IC95%
0.85 25.3
(0.77 27.7
0.92) 32
67
79
93
5875
7085
8896
88
79
49
7495
6588
3663
0.84 25.3
(0.78 27.7
0.89) 32
73
83
94
6977
7986
9297
81
62
33
7188
5171
2443
0.85 25.3
(0.82 27.7
0.87) 32
75
85
95
73 78
83 87
94 96
83
67
35
7687
66 77
3041
* Accounting for clustering by patient and by study.
Calculated with robust standard errors to account for clustering by patient.
A sensitivity analysis disregarding the two biggest studies with 208 samples
yielded an AUC of 0.84 (CI95 0.780.9)
Only 827 samples had information on HIV available, AUC in those with
HIV status available was 0.84 (CI95 0.800.88). 720 samples had no
information on HIV available, AUC for those was 0.87 (CI95% 0.840.90)
S283
OA-404-05 Implementation of XpertW MTB/RIF

under Indias National TB Control Programme
S Dhawan,1 S Mannan,2 A Sangwan,3 S Chadha,1
A Sreenivas,2 C N Paramasivan,4 K S Sachdeva,5
N Kulshreshta5 1International Union Against Tuberculosis
and Lung Disease South-East Asia Office, New Delhi, 2World
Health Organization, New Delhi, 3Clinton Health Access
Initiative, New Delhi, 4Foundation for Innovative New
Diagnostics, New Delhi, 5Central TB Division, Ministry of
Health and Family Welfare, New Delhi, India. e-mail:
shikha.dhawan@gmail.com
Background: As feasibility of deployment of Xpertw

MTB/RIF was successful, Indias National TB Control
Programme scaled up implementation of Xpert. In 2014,
the NTP had 89 Xpert sites across India. The programme
has the policy to use Xpert for diagnosis of presumptive
drug resistance cases and has prioritized its use in people
living with HIV/AIDS and paediatrics. We assessed the
implementation of Xpert MTB/RIF under programmatic
conditions.
Design/Methods: Data on Xpert tests was collected in
2014 from 76 functional Xpert sites across 27 provinces
on monthly basis using a standard tool consisting of
information on laboratory performance.
Results: Of the 112 273 tests done across 76 Xpert sites,
there were 1413 (1.3%) referrals from private sector, 10
952 (9.8%) PLHIV and 11 240 (10%) paediatric. 105
243 (93.7%) valid tests were obtained on first test. These
included bacteriologically confirmed tuberculosis in 63
624 samples (positivity 56.7%, 6.6% HIV co-infected,
2% pediatric), Rif resistance was detected in 11 709
(18.4%) samples (3.3% HIV co-infected, 1.9 % pediatric). M. tuberculosis was not detected in 41 619 (37.1%)
samples. Of the initial 8370 (7.4%) test failures, repeat
testing was done on 6156 (73.8%) by retesting same or
second available sample in which additional 3196
(2.8%) TB cases were obtained including 552 (0.5%)
Rif resistance. Out of all those who underwent tests, 35
889 (69.1%) TB and 5429 (46.4%) Rif resistant TB
patients were started on treatment as reported by sites.
Conclusion: Information on treatment initiation on
diagnosed TB patients was not uniform and consistent
at some sites. This highlights the need to ensure better
coordination between the diagnostic laboratories and TB
program managers at sub-district level for strengthening
overall tuberculosis care cascade. Although utilization of
Xpert is poor for testing PLHIV and paediatric in 2014,
NTPs ongoing Xpert projects for paediatric and ART
Centres will show improvement in the coming years. 13
sites were non-functional in 2014 due to supply chain
management issues of cartridges, module failures and
lack of trained manpower. Overcoming programmatic
and operational challenges in implementing and maintaining the Xpert machines is essential for optimal
utilization of these near point of care molecular
diagnostic. For early diagnosis by Xpert and early
treatment initiation, NTP needs to innovate in public
private partnerships and involve different stakeholders.
S284
OA-405-05 Triplicate versus single sputum

samples improves sensitivity for efficacy
determination in studies of anti-tuberculosis
regimens
1
A Diacon, E Van Brakel, N Lounis, P Meyvisch,

B Van Baelen,2 T De Marez,3 E Jenkins,4 B Dannemann3
1
Faculty of Medicine and Health Sciences, Stellenbosch
University, Cape Town, South Africa; 2Janssen Research and
Development, Beerse, Belgium; 3Janssen-R&D, LLC, Titusville,
NJ, USA; 4Janssen UK, High Wycombe, UK. e-mail:
ahd@sun.ac.za
Background: Sputum culture conversion from positive to

negative remains the standard endpoint in studies of antiTB regimens. Most protocols require a single sputum to
be collected for culture at fixed time points. In study
TMC207-C208 the treatment efficacy for multi-drug
resistant TB was assessed with triplicate sputum at each
visit over a 120-week study period. We investigated
whether the aggregate result from triplicate sputum
differed from the result of the first sputum collected at
each visit. Method This analysis included 160 participants who collected at least one triplicate (730 minutes
between sputum). Each sputum was submitted separately
for liquid culture in the Bactec e MGIT960 e system.
Visits were performed pre-treatment and thereafter at
weekly, 2-, 4- and 12-weekly intervals up to 8, 24, 36 and
120 weeks, respectively. Single sputum were scored as
positive, negative or contaminated. Triplicates were
scored as positive if 71 sputum of the triplicate was
positive, as negative if 71 sputum was negative and none
was positive, and as contaminated if none were valid
(positive or negative).
Results First, second and third sputum in 3467 triplicates
had comparable rates of positive, negative and contaminated scores (averages of 34.6%, 55.1% and 10.3%,
respectively). Rates of positive, negative and contaminated scores for first and triplicate sputum were 34.3%
and 42.3%, 54.3% and 55.1%, and 11.3% and 2.6%,
respectively. The changes were mainly due to invalid first
sputum becoming negative triplicates and negative first
sputum becoming positive triplicates. The first and the
first and second sputum combined were positive in
81.4% and 94.0% of 1467 positive triplicates, and
negative in 90.4% and 98.2% of 1911 negative
triplicates.
Conclusions The collection of triplicate instead of single
sputum increased the proportion of visits with positive
sputum culture results by 8%, and reduced the rate of
visits with invalid results from 11.3% to 2.6%. A greater
proportion of valid sputum results means fewer unscheduled visits for repeat sputum collection and/or fewer data
points unavailable for efficacy analysis. In comparative
studies, a lower rate of undetected positive visits is likely
to improve a studys power to detect a difference between
regimens. This can reduce the required number of study
participants and might offset the extra cost of triplicate
sputum culture.
OA-406-05 The impact of routine XpertW MTB/

RIF on empirical TB treatment in Cape Town,
South Africa
S Hermans,1,2 J Caldwell,3 R Kaplan,2 R Wood2
Amsterdam, Netherlands, 2Desmond Tutu HIV Centre,
University of Cape Town, Cape Town, 3City of Cape Town,
Cape Town, South Africa. e-mail:
sabine.hermans@hiv-research.org.za
1
Background: Xpertw MTB/RIF (XP) has been projected

to increase tuberculosis (TB) case detection by 30%.
However four large randomised trials on the use of XP in
sub-Saharan Africa did not show an impact on the
number of cases diagnosed, which has been attributed to
a replacement of empirical treatment by microbiologically confirmed diagnosis. The impact of XP on
treatment decisions in a routine care setting is not
known. Cape Town started roll-out of XP in August
2011 which was completed in June 2013. We aimed to
investigate the impact of XP in a routine care setting on
empirical TB treatment stratified by HIV status.
Design/Methods: In a historical cohort study we
compared all adult pulmonary TB (PTB) cases in a 1
year period before (2010) and after (2014) roll-out of
Xpert MTB/RIF in Cape Town. We utilised data from the
City of Cape Town electronic TB register to calculate the
proportion of microbiologically confirmed TB cases
(either XP, smear or culture). Patients with more than
one positive test result were presumed to be diagnosed on
the basis of either XP, smear or culture, in this order. All
results were stratified by HIV status. We used multivariable logistic regression analysis to determine risk factors
of empirical treatment in HIV-positive PTB cases.
Results: In 2010, 21 562 PTB cases were diagnosed

versus 19 260 in 2014. An XP result was recorded in
88% of PTB cases in 2014. Overall, the proportion of
microbiologically confirmed PTB increased from 74% to

82% (P , 0.001). This increased from 81% to 90% in
HIV-negative PTB and from 66% to 74% in HIVpositive PTB (both P , 0.001). Treatment decisions
moved from being mainly smear-based to mainly XPbased irrespective of HIV status (see Figure). In HIVpositive cases, reliance on TB culture reduced substantially but empirical TB treatment still remained high at
26%. Empirical treatment in HIV-positives was associated with older age (per 10 year increase, adjusted OR
[aOR] 1.15 [95%CI 1.09-1.21], P , 0.001) and lower
CD4 count (per 50 cells decrease, aOR 0.98 [95%CI
0.97-0.99], P 0.002).
Conclusion: Routine use of XP in a setting with a high
burden of TB and HIV did not lead to increased numbers
of pulmonary TB diagnoses, but increased microbiological confirmation by 8%. More than a quarter of HIVassociated pulmonary TB cases were still treated
empirically, highlighting the remaining difficulties in
diagnosing TB in the HIV-infected population.
S285
Results: We analyzed 146 sputum samples treated either

with or without CPC. CPC pretreated samples showed a
higher culture yield compared to non-treated sputum
samples across all decontamination times: 90% versus
50% with a decontamination time of 20 min (P ,
0.0001, Fishers exact test), 94% versus 68% with 15
min (P , 0.0001), and 94% versus 73% with 10 min (P
0.012). The quantitative culture grading was consistently
higher in CPC pretreated compared to non-treated
samples (Table). The proportion of contaminated cultures was less frequent in CPC pretreated samples across
all decontamination times (range 2-6%), compared to
non-treated samples (range 15-16%). Of 50 CPC
pretreated samples with a positive culture, all samples
were also Xpert MTB/RIF positive (50/50).
Conclusion:Preservation of sputum samples with CPC
with subsequent decontamination for 15 minis a simple
and inexpensive method for TB laboratories in lowincome countriesto facilitate the transport logistics. It
increases culture yield while decreasing contamination
rate, and importantly pretreated samples can also be used
for Xpert MTB/RIF.
OA-407-05 Preservation of sputum samples

with cetylpridinium chloride for tuberculosis
cultures and XpertW MTB/RIF in low-income
countries
H Hiza,1 B Doulla,2 M Sasamalo,1 F Mhimbira,1 J Hella,1
L Jugheli,3 L Fenner1,3 1Ifakara Health Institute, Bagamoyo,
2
Ministry of Health, Dar Es Salaam, Tanzania; 3Swiss Tropical
and Public Health Institute, Basel, Switzerland. e-mail:
hhiza@ihi.or.tz
Background: Preservation of sputum samples during

transport and culture contamination with environmental
bacteria is a major challenge in tuberculosis (TB)
laboratories in hot and humid climate zones. Cetylpridinium hloride (CPC) is known to reduce the commensal
flora in sputum samples.However, there is only limited
information from evaluation studies in sub-Saharan
Africa.
Objectives: We aimed 1) to study the effect of CPC
preservation on culture results depending on the duration
of decontamination; and 2) to investigate the performance of Xpertw MTB/RIF in CPC pretreated samples.
Methods: Consecutive sputum samples (at least 4ml)
from smear-positive TB patients were collected at a TB
Clinic in Tanzania, and transported in cooled boxes to a
BSL2 TB Laboratory. Samples were equally split in two
aliquots, one sample was treated with CPC (1:1 ratio of
sputum and CPC), and stored at ambient temperature
(258C) for seven days until processed. Samples were
decontaminated using 1% NaOH decontamination for
20, 15 or 10 min. The second aliquot was immediately
processed using the routine1.5% NaLC-NAOH decontamination for 20 min. All decontaminated samples were
subsequently cultured on Lowenstein-Jensen
solid media
at 378C for 8weeks. Cultures were read using the WHO

grading system and confirmed by Ziehl-Neelsen staining.
In addition, 50CPC treated samples were tested by Xpert
MTB/RIF.
OA-408-05 Development of a multiplex realtime PCR assay for detection of Mycobacterium

tuberculosis from sputum collected in
PrimeStore MTMW
L Daum,1 S A Worthy,1 J D Rodriguez,1 R Peters,2
B Fourie,2 G Fischer1 1Longhorn Vaccines & Diagnostics, San
Antonio, TX, USA; 2University of Pretoria, Pretoria, South
Africa. e-mail: longhorn-vandd@sbcglobal.net
Background: We recently reported a sensitive and timely

Mycobacterium tuberculosis (M. tuberculosis) real-time
PrimeMix PCR detection assay targeting the multi-copy
IS6110 genomic element (Daum et al., IJTLD, May,
2015). However, some M. tuberculosis strains are known
to be IS6110 deficient and therefore undetected by assays
that exploit this target. Up to 56% of M. tuberculosis
strains collected from India were shown to contain 5 or
less copies of the IS6110 element. In one report from
Mexico 20.8% of M. tuberculosis isolates lacked IS6110
(Lopez-Alvarex et al., BMC Microbiology, 2010).
Accurate and timely M. tuberculosis detection is the first
and most critical step for effective M M. tuberculosis
therapeutic measures. Aim: A multiplex real-time PCR
assay for the detection of IS6110 and IS1081 multi copy
M M. tuberculosis segments was developed into a single
step PrimeMixw PCR blend. The limit of detection using
quantified M. tuberculosis was determined and multiplex
S286
results were compared to a real-time PCR uniplex assay.

In addition, this multiplex PCR assay was evaluated on a
panel of specimens collected in South Africa.
Methods: A real-time multiplex assay targeting the M.
tuberculosis IS6110 and 1081 multi copy targets was
developed from conserved regions obtained from multiple sequence alignments of M. tuberculosis strains in
GenBank. For limit of detection, a quantified stock
culture of H37Rv (106 CFU/ml) was serially diluted and
placed into PrimeStorew for real-time PCR detection.
Additionally, an IPC assay targeting a non-specific DNA
fragment present in a clinical collection medium (PrimeStorew) was utilized for monitoring specimen integrity from collection to detection. A random sample of
positive and negative clinical specimens was utilized
consisting of sputum collected and shipped at ambient
temperature in PrimeStore MTMw from rural sites in
Africa.
Results: The multiplex M. tuberculosis assay was
sensitive across five 10-fold serial dilutions of quantified
M. tuberculosis genomic DNA and exhibited no cross
reactivity to a reference panel of non- M. tuberculosis
complex organisms. The multiplex M. tuberculosis assay
correctly identified 30/30 specimens and was comparable
to a real-time M. tuberculosis uniplex equivalent.
Additionally, IPC present in PrimeStorew was detected
in all specimens.
Conclusions: The multiplex M. tuberculosis assay
targeting both IS6110 and IS1081 is sensitive and
specific; exhibiting equivalent results compared to a
single PCR target. This assay is a sensitive dual target
method for M. tuberculosis detection, particularly in
specimens from geographical areas where M. tuberculosis may lack the IS6110 target.
OA-409-05 The effect of automated nucleic acid

amplification assays on mortality in routine
care settings: meta-analysis of individual
participant data
A Khaki,1 G Di Tanna,2 K Fielding,2 G Theron,3 M Nicol,3
K Dheda,3 G Churchyard,4 J Metcalfe1 1University of
California San Francisco, San Francisco, CA, USA; 2London
School of Hygiene & Tropical Medicine, London, UK;
3
University of Cape Town, Cape Town, 4The Aurum Institute,
Johannesburg, South Africa. e-mail: ali.khaki@ucsf.edu
Background: Xpert MTB/RIF (Xpert), an automated

nucleic acid amplification test, is now widely implemented within national tuberculosis programs in over 115
countries. While diagnostic accuracy is well documented,
anticipated improvements in patient outcomes have not
been described.
Design/Methods: We performed an individual participant data (IPD) meta-analysis of four randomized
controlled trials (RCTs) from the Southern African
region investigating the impact on early mortality among
tuberculosis suspects screened using Xpert compared
with sputum smear microscopy. We utilized both oneand two-stage methods to account for individual and
cluster-level trial designs, and Huber-White sandwich
estimators for multiple centers. Analyses were adjusted

for age and gender, and were stratified by HIV infection.
Results: In four trials involving 8,567 patients Xpert was
not associated with a decrease in overall 3-month
mortality (odds ratio [OR] 0.91 [95%CI 0.68-1.20], P
0.50) or in HIV-positive persons (OR 0.86 [95%CI
0.66-1.13], P 0.29). Similar results were found for 6month mortality (three RCTs, n 8,142). In time-toevent analysis Xpert was not associated with a decreased
hazard of death among all (hazard ratio [HR] 0.82
[95%CI 0.64-1.05], P 0.11) or HIV-negative participants (HR 0.81 [0.46-1.41], P 0.45), but was
associated with decreased mortality among HIV-positive
participants (HR 0.75 [95%CI .59-0.96], P 0.02).
Heterogeneity among trials was minimal (Chi-square
test, P 0.55, I20.0%).
Conclusion: In this large multi-country IPD metaanalysis, Xpert was associated with decreased mortality
among HIV-positive patients. Although our preliminary
analyses are potentially limited by informative censoring
and/or unmeasured confounding, our study represents,
to our knowledge, the first randomized evidence that
novel TB diagnostics decrease risk of death.
OA-410-05 A pilot proficiency testing program

for XpertW MTB/RIF using dried tube
specimens
K Klein,1 K Degruy,2 C Hatcher,2 H Alexander2 1U.S.
USA. e-mail: xim0@cdc.gov
Background: Use of the Cepheid Xpert MTB/RIF assay

has expanded rapidly since the World Health Organizations initial recommendation in 2010. However, proficiency testing (PT) material for this assay is not widely
available. The U.S. Centers for Disease Control and
Prevention (CDC) piloted a voluntary, cost-free PT
program using Dried Tube Specimens (DTS) to assist
testing sites in monitoring the performance of Xpert
MTB/RIF testing.
Design/Methods: Inactivated DTS were prepared at
CDC using selected isolates of rifampicin (RIF)-susceptible Mycobacterium tuberculosis (M. tuberculosis), RIFresistant M. tuberculosis, and non-tuberculous mycobacteria. Panels composed of 5 DTS samples per round
were shipped at ambient temperature to designated
country coordinators and then distributed to enrolled
testing sites. Participating sites conducted Xpert MTB/
RIF testing per enclosed instructions and submitted test
results to the country coordinator using a standardized
electronic form. Country coordinators collated and
submitted results to CDC for review. Participating sites
were provided unscored PT reports after each round.
Between March and December 2013, 4 PT rounds were
conducted.
Results: Enrollment increased from 103 sites from 6
countries in the first round (March 2013) to 222 sites
from 10 countries in the fourth round (December 2013),
for a total of 648 panels provided.
S287
Results were returned electronically submitted for 475

(73.3%) or 648 panels. The mean proportion of
responding sites achieving 100% concordance with
expected results over all rounds was 66.7% (range
62.3-70.8%). Among 163 sites with any discordant
result, the number of discordant results per site in a single
panel ranged from 1 to 5. Unsuccessful tests (i.e., error,
invalid, or no result) were responsible for 78 (42.6%) of
the total 183 discordant results in all 4 rounds. Of the
105 discordant test results, 39 (37.1%) were false
negative M. tuberculosis, 37 (35.2%) were indeterminate
RIF resistance, 11 (10.5%) were false negative RIF
resistance, 11 (10.5%) were false positive RIF resistance,
and 7 (6.7%) were false positive M. tuberculosis results.
Conclusion: DTS PT panels are critical for monitoring
Xpert MTB/RIF testing performance as part of a
comprehensive quality assurance program that includes
routine quality indicator monitoring, supervisory visits,
and when necessary, corrective action. To facilitate
quality improvement, future PT rounds will include
performance scores and follow-up for those sites not
achieving satisfactory scores.
working facility help in strengthening the treatment and

the treatment outcome. This found very much useful for
the patients as well as the various stakeholders involved
in TB care.
Results and lessons learnt: The study suggests that peer
networking and connectivity with the care providers
enhance confidence in patients and the DOT providers
and this help in minimizing the default. The problems
could be resolved faster and with clarity and this make
patients on a better side in their ability to resolve
treatment blocks. The communication facility at times of
adverse effect helps in managing them more effectively
and disseminating the solutions to larger mass.
Conclusions and key recommendations: Peer networking
and opportunity to network with the treatment stakeholders can build confidence among patients and can
strengthen the TB and the ultimate treatment outcome.
Mobile phones could be a good way to network faster
with all stakeholders with clarity of information care
process. The care outcomes could be strengthened by
strengthening the patient connectivity with all stakeholders. Technology can facilitate a cost effective
networking.
15. When waves trump paper: the

triumph of e-technology
OA-412-05 A sustainable model of effective

private provider engagement via e-health and
free TB drugs, Mehsana District, Gujarat, India,
2014
OA-411-05 Patient networking for TB Care

using mobile technology: care of I Decide
mApplication
T Kizhakkeparampil Thomas,1 S Kumar,2 J Yakubu,2
S Pulibanti2 1Olive Touch Health Care Services, Thrissur,
2
School of Public Health, SRM University, Chennai, India.
e-mail: ceomarg@gmail.com
P Dave,1 P Nimavat,1 K Patel,1 B Vadera,2 K Rade,2

P Dewan,3 A Sreenivas,2 N Kulshreshta4 1Ministry of
Health & Family Welfare, Gandhinagar, 2World Health
Organization Country Office for India, New Delhi, 3Bill &
Malinda Gates Foundation, New Delhi, 4Ministry of Health &
Family Welfare, New Delhi, India. e-mail:
dr.vadera@gmail.com
Background and challenges to implementation: Patient

Knowledge and ability to communicate with care
providers has been found to be important in enabling
patients to continue the treatment without default. It has
been found gaps in the knowledge flow regarding disease
and treatment which work as barrier in treatment.
Patient to Patient Learning is an important element in
treatment of TB. The drop outs is an important problem
even today and the social stigma on the disease continue
to be an important problem. Thus we have piloted a
project were peer networking and learning is activated
with the help of mobile phone. This project as part of a
bigger project developed a mobile application which
enables patients to interact and network with their peers
and care providers to resolve many of the health issues.
Intervention or response: We have developed a mobile
application which enables learning, interactions and
problem solving with the help of mobile phones. The
application facilitate patient networking, getting feedback from the providers on the various adverse health
effects and raising questions about various problems that
patients have during the treatment. Mobile phone is
provided with application installed to 100 TB patients
who were defaulters to find out if the knowledge
dissemination and mobile real-time tracking and net-
Background: Private providers (PP) in India manage 1 in

8 TB patients in the world, and are believed to account
for 1 million of the 3 million missing TB case
notifications globally. Prior solutions have yielded
modest, incremental increases in TB case notification.
We developed a model to leverage e-health advances and
using free anti-TB drug vouchers to attract PP TB
notifications, reduce patient out-of-pocket expenses,
and improve the quality of TB care at scale.
Intervention: Gujarat State has a large PP presence, with
.18,000 members registered with the Indian Medical
Association. In largely rural Mehsana district (pop.
2.1m), all PP and chemists were mapped & prioritized
for engagement based on speciality and TB drug sales. A
call centre receives TB notification and issues immediate
e-vouchers for anti-TB drugs as per national standards,
which patients redeem at private chemists. Notified
patients are visited by program staff for contact
assessment and offer of HIV & diabetes comorbidity
screening. Patients receive adherence support, reminders,
self-reporting via call centre, and active follow-up by call
centre and then program staff of non-reporting or missed
refills.
Results: From 1 August 14 to 31 March 15, out of 440
total PP, 180 were targeted and 81 (45%) notified;
S288
similarly, of 159 chemists who stocked anti-TB drugs, 80

(50%) redeemed e-vouchers. Total TB case notification
rate (annualized) doubled, relative to the same quarter
one year earlier. This included 1493 from public sector,
and 1582 from PP; public TB notification remained
stable. In eight months, 4098 patient months of vouchers
(USD$30 thousand) were issued and redeemed. Initial
treatment cohort (n 220) show 79% patients completed treatment or continued on treatment at end of 6
months under PP.
Conclusions: This unique model demonstrated that in
rural areas with well-staffed TB programs, PPs in India
can be engaged effectively at scale, with e-health support
and modest additional resources. With notification,
treatment and support as per national standards can be
extended to patients. Similarly, PP quality can be
monitored and improved. Needed enhancements include
modern adherence monitoring and free TB diagnostic
services for PP.
OA-413-05 E- and m-Health solutions for

national TB programmes to track missing cases
S Achanta,1 J Jaju,1 C Chatla,1 R Samyukta,2 M Parmar,1
KS Sachdeva,3 A Sreenivas1 1World Health Organization
Country Office for India, New Delhi, 2State TB Cell,
Hyderabad, 3Central TB Division, New Delhi, India. e-mail:
achantas@rntcp.org
Background: Pre-treatment loss to follow-up (PLFU) of

TB patients, hinders tuberculosis control efforts globally.
Cohort analysis of cases detected by Revised National TB
Control Programme (RNTCP) in Telangana, India show
that nearly 2000 smear positive cases appear in Lab
registers consistently quarterly but are not recorded in
TB patient registers, a log of patients registered for
treatment under the programme. Feedback on treatment
initiation is either missed or delayed of patients who are
referred for treatment to other TB units after diagnosis,
often outside the district or State. Current monitoring
practices fail to trace these patients and the Lab
diagnostic cohort is neither validated nor reported. In
view of the global concern of missed cases from India
and Governments mandate to notify all diagnosed and
treated TB cases; it is significant that programme
addresses challenges associated with notifying diagnosed
cases, tracing PLFU patients, and initiating early
treatment. To address these challenges, an e- and mHealth module has been devised by RNTCP in Hyderabad, Telangana to 1) identify the diagnostic cohort 2)
decrease PLFU during the TB diagnostic pathway.
Intervention: Lab technicians enter smear results and
demographic details of smear positive TB cases into
electronic-Lab registers which are linked through a
central server. Feedback on the date of starting treatment
is exchanged electronically and auto SMS/IVRS system
tracks them for early treatment initiation. Data on
patient referrals and their feedback is compared from
30 microscopic centers in Hyderabad, before and after
this intervention
Results: There were 750 smear positive patients diagnosed during April-June 2013 and 708 patients during
April-June 2014, which could be reported as a diagnostic

cohort of TB patients. Feedback on treatment initiation
of referred out patients was received for 424 of 689 (61
%) in 2013 and 563 of 645 (87 %) in 2014 after
introducing e-Lab register. Electronic feedbacks were
obtained within a range of 2-7 days. Pending feedbacks
and eventual PLFUs dropped from 39 % to 13 %
Conclusions: The e-Lab register and auto SMS/IVRS
system improves patient tracking, reduces PLFU and
identifies a diagnostic cohort for TB case reporting.
Innovative technological solutions are needed to track
missing TB cases and improve TB case notification
Table: Number of feedbacks received of patients referred out
for treatment initiation and registration before and after
introduction of the e-Lab register in 30 Microscopy centers of
Hyderabad, Telangana
Patients
Before Intervention
April-June (2013)
After Intervention
April-June (2014)
750
708
689
645
Sputum positive
patients diagnosed
in the microscopy
centers (A)
Sputum positive
patients of (A)
referred out (B)
Patients of (B)
referred for
treatment to TB
units within the
same district
Patients of (B)
referred for
treatment to other
districts
Patients of (B)
referred for
treatment to other
States
Total
Referred
n
Feedback
received
n (%)
Referred
n
Feedback
received
n (%)
451
262 (70)
405
348 (92)
231
162(58)
233
215 (86)
7
689
0
424 (61)
7
645
0
563 (87)
TB tuberculosis.
OA-414-05 99DOTS: monitoring and improving

TB medication adherence using mobile phones
and augmented packaging
A Cross,1 M Kumar,2 P Soren,2 K Rade,3 A Sreenivas,3
S Kumar,3,4 B Thomas,4, P Dewan5 1Microsoft Research
India, Bangalore, 2Innovators In Health, Patna, 3World Health
Organization, Country Office for India, Delhi, India; 4Arcady
Group, Richmond, VA, USA; 5Melinda Gates Foundation,
Delhi, India. e-mail: across@microsoft.com
Background and challenges to implementation: A critical

challenge in tuberculosis treatment is to ensure that
patients adhere to the full course of medication. In India,
use of Directly Observed Therapy (DOT) in the public
sector has been associated with high treatment success.
However, direct observation introduces challenges for
patients and providers, who need to undertake frequent
travel throughout the course of treatment. It is also
challenging for program managers to detect and respond
to missed doses in an accurate and timely way.
Intervention or response: A low-cost approach for

monitoring and improving TB medication adherence,
called 99DOTS, was deployed in a public-sector program in Samastipur district, Bihar, India. Each anti-TB
blister pack is wrapped in a custom envelope (pictured),
which includes a hidden phone number that is visible
only when a dose is dispensed. After taking the
medication, patients make a free call to hidden phone
number, yielding high confidence that the dose was inhand and has been taken. Initially, patients undergo
normal DOTS supervision and providers assist in making
the call. Over time, adherent patients can take custody of
their drugs and make calls with lesser supervision, while
still permitting remote monitoring of each dose. In
addition to compiling a real-time dosing history for each
patient, 99DOTS enables a wide range of reminders,
alerts and follow-ups.
Results and lessons learnt: From August, 2014 to
January, 2015, 49 patients initiated treatment under
99DOTS. Of these, 31 showed 100% adherence, 16
patients showed 85-99% adherence and 2 patients
showed 75-85% adherence. A total of 1791 doses were
logged via calls to 99DOTS. Most doses were observed
by providers; however, small-scale experience with selfadministration (5 patients) and family observation (6
patients) showed adherence of 94% and 91%, respectively.
Conclusions and key recommendations: 99DOTS empowers patients, providers, and treatment programs,
supporting health system efficacy and efficiency. Adherent patients can take independent custody of their drugs
and confirm consumption without physical observation,
which simultaneously reduces the burden on providers.
Program managers gain real-time information on adherence, enabling more customized and more efficient
patient supervision and support. 99DOTS may be
especially applicable as India switches from thriceweekly to daily regimens, as it enables daily monitoring
without daily observation.
S289
OA-415-05 GxAlert SMS improves DR-TB

patient enrollment and management in Nigeria
K Jimoh Agbaiyero,1,2 L Ekbladh,1,2 M Benezet,2 J Takle,3
C Macek3 1Abt-Health Finance and Governance Project,
Abuja, Nigeria; 2Abt Associates, Bethesda, MD, 3SystemOne,
Boston, MA, USA. e-mail: kehindeagbaiyero@yahoo.com
Background and challenges to Intervention: Nigeria is a

WHO high-burden drug-resistant TB (DR-TB) country
that relies on GeneXpert MTB/RIF machines to diagnosis DR-TB at multiple clinical locations. However,
reporting TB testing results is a lengthy process due to
a continued reliance on paper records and slow data
transit systems. Quality of results data is often poor and
subsequent program management decisions are not
always timely or focused on priority needs.
Intervention: In pursuit of cost-effective technologies to
improve healthcare responses to DR-TB by the Nigerian
Tuberculosis and Leprosy Control Program (NTBLCP),
Abt Associates and SystemOne developed an innovative
mobile-based solution that sends GeneXpert diagnostic
results to key health system actors instantly. GxAlert is
configured on GeneXpert systems by installing a modem
from a local telecom that sends encrypted data sent to the
secure web-based GxAlert database in real time. The
system then sends the results in an SMS alert to program
decision makers at the state and national TB program,
shortening the new-case reporting period from months to
seconds.
Results and lesson Learnt: The proportion of DR-TB
patients enrolled for treatment based on GxAlert
messages received from 35 GeneXpert facilities jumped
to 85% in March, 2015 from only 20% enrolled in April
2014. SMS or text message alerts are sent to TB State
program manager and National Program DR-TB enrollment officer upon GeneXpert confirmation of a M.
tuberculosis Rif case to speed treatment initiation.
Weekly reports of all new TB cases are both emailed and
sent by SMS to local health officials to ensure better
connection between diagnosis, enrollment and treatment. Faster, better quality data on MDR-TB portion of
total TB indicators are now reported with minimal error
or effort.
Conclusions and key Recommendations: The use of
GxAlert SMS notification of GeneXpert testing suggests
a scalable model for sustainability: Installation is done
once and in-country. Technology is kept simple as local
telecom modems are readily available and affordable for
connectivity. In Nigeria, GxAlert has the potential to
strengthen surveillance of DR-TB, TB in children,
speeding response and improving programmatic decision
making for faster enrollment of patients in treatment
programs. Based on these results, a national rollout of
GxAlert was formally launched by the honorable
Minister of Health of Nigeria on World TB Day 2015.
S290
OA-416-05 57 Zone: using social media in China

to empower TB patients for treatment success
X Lin,1 X Xu,2 Z Yu,1 W Zhang,1 Z Huang,3 l Ling,2 G Nie,1
A Innes4 1Yunnan Center for Disease Control and
Prevention, Kunming, Yunnan Province, 2FHI 360 China,
Kunming, Yunnan Province, 357 Zone, Kunming, Yunnan
Province, China; 4FHI 360 Asia Pacific Regional Office,
Bangkok, Thailand. e-mail: ainnes@fhi360.org
Background: Chinas Yunnan Province has a high TB

burden with increasing numbers of complicated retreatment and drug-resistant cases. Interviews with TB
patients in Yunnan show that not knowing the correct
treatment course and poor compliance are common
barriers to treatment success. Effective patient-doctor
communication for treatment adherence is limited by
understaffed TB clinics. At the same time, many TB
patients are open to taking responsibility for their disease
and are motivated to be cured.
Intervention: In December 2013, multi-drug resistant
(MDR) TB patients in Yunnan established a social media
group called 57 Zone using QQ, a popular social
network and mobile app in China with over 830 million
users. Through 57 Zone, patients and their family
caregivers have instant, virtual access to peer educators
and clinicians, asking questions about TB medications,
adverse drug effects, proper nutrition, and psychosocial
concerns. QQ managers were identified at the launch of
57 Zone and given clear roles and responsibilities to
deliver online services in a friendly, timely and professional manner. These QQ managers disseminate key TB
messages daily, follow online discussions, and organize
activities to encourage experience sharing among patients.
Results: As of March 2015, 57 Zone had more than 300
members, 90% of whom are current or former patients
and family caregivers. More than 100 members actively
participate in discussion each day, with conversations
covering tips on navigating the TB system; coordination
of follow-up visits with TB doctors and face-to-face peer
counselling at TB hospitals; and emergency fund raising
for patients who need money to continue treatment. 57
Zone patients express a strong sense of community,
connectivity, and moral support for each other. Improved
patient-doctor relationships have also been formed
through monthly 1-hour consultation sessions and
information exchange online. TB-HIV patients now
participate in 57 Zone, and the social network has also
expanded to include members in rural Zhenxiong
County (which has the highest TB notification rates in
Yunnan) and Xinjiang Province (with the highest TB
prevalence in China).
Conclusions: 57 Zone is among the first social networks
in China created by patients for TB and MDR-TB. Social
networking can give patients instant, virtual access to
education and support, a powerful adjunct to routine
communication during follow-up visits.
OA-417-05 Tuberculosis case finding at

pharmacies in Tanzania using trained
pharmacists and an electronically monitored
referral system
F Mhimbira,1,2,3 J Hella,1,2,3 B Mutayoba,4 B Doulla,4
E Mahongo,3 J Minde,4 L Fenner1,2,3 1Swiss Tropical and
Public Health Institute, Basel, 2University of Basel, Basel,
Switzerland; 3Ifakara Health Institute, Bagamoyo, 4Ministry of
Health and Social Welfare, Dar Es Salaam, Tanzania. e-mail:
fmhimbira@gmail.com
Background: Introducing new tuberculosis (TB) diagnostics is critical, but additional interventions are needed
to bring patients suspected with TB to care. There is
evidence that presumptive TB patients use un-prescribed
antibiotics to relieve their symptoms. We aimed to set up
and evaluate a referral system at pharmacies to detect TB
cases who were missed or delayed by the routine health
care system.
Methods: We established an electronic referral system
using android tablet computers at five pharmacies in a
densely populated area of the Ilala District, Dar es
Salaam (Buguruni Health Center). Pharmacists were
trained to refer any customers who reported coughing
(any duration) or any other TB related symptom.
Anonymous referral cards were given and an automatic
follow-up SMS reminder were sent to customers who did
not report to a within three and five days. TB Diagnosis
was either done by sputum microscopy, chest X-ray or
based on clinical findings. In addition, a control
population of consecutive TB patients diagnosed through
the routine care system was also recruited. The study
started in November 2014 and will last for 12 months.
Results: Until March 2015, the pharmacies referred 118
presumptive TB patients. The median age was 37 years
(Interquartile range [IQR]: 28-43), and 82 (70%) were
men. Apart from coughing, 75 (63.6%) patients reported
excessive night sweats, and 69 (58.5%) weight loss. The
most commonly purchased antibiotics at the pharmacies
were amoxycillin (52 customers, 44 %,) and ampiclox in
36 (31%) customers. Overall, 103 (87%) of the referred
customers reached the TB diagnostic center, of these 89
(86%) came within 3 days. Out of 103, 77 (74%) had TB
investigation results, of which 29 (37%) were diagnosed
with TB and started on treatment. Smear microscopy
diagnosed 16 (55%) of TB patients, and the remaining 12
(41%) were diagnosed based on clinical or radiological
features. Coughing of 4 weeks or more was reported in
26 (90%) of the TB patients. Compared to TB patients
referred by pharmacies, routinely diagnosed TB patients
tended to have longer range of duration of cough (1-16
vs. 0-50 weeks).
Conclusion: We successfully demonstrated the feasibility
that presumptive TB cases visiting pharmacies to
purchase antibiotics can be referred to a TB diagnostic
center. If this low-cost case finding strategy could be
scaled-up, it may have a huge impact on TB control by
increasing early TB diagnosis and thus reduce transmission in the community.
S291
remaining 121 (48%) of unmatched patients were

entered into the application, but either leaked, contaminated, inadequate or otherwise untestable specimens,
which were rejected at the lab.
Conclusions: These data demonstrate that the time from
TB suspicion to treatment initiation can be reduced to 5
days or less for MDR-TB patients using a package of
mobile applications seamlessly communicating through
an integrated platform with the NHLS system. Matching
of results between the lab and the application was perfect
when barcodes were recorded within the lab system and
when specimens were of adequate quality for processing.
Integrating technologic advances with skilled human
resources can increase efficiency and improve time to
treatment initiation.
16. From bench to treatment

OA-418-05 Novel package of integrated mobile
applications reduce time to treatment
initiation for patients with MDR-TB
J Farley,1,2 J Mckenzie-White,3,4 S Seiguer,4 M Naicker,1
L Isherwood,5 F Olsen,5 S Candy,5 W Stevens5 1Jhpiego
South Africa, Pretoria, South Africa;2Johns Hopkins University
School of Nursing, Baltimore, MD, 3Johns Hopkins University,
Baltimore, MD, 4Emocha Inc., Baltimore, MD, USA; 5National
Health Laboratory System, Johannesburg, South Africa. Fax:
(1) 410 955 7733. e-mail: jfarley1@jhu.edu
Background: Like rapid diagnostic technology, mobile

health (m-health) has the power to revolutionize TB care.
We set out to test whether the time from TB suspicion to
treatment initiation could be reduced to 5 days or less to
meet the National Strategic Plan for TB-HIV in South
Africa utilizing an m-health application-based platform.
We describe the initial implementation of a package of
mobile applications, developed by emocha Mobile
Health, Inc., that is designed to register, link, refer and
document initiation.
Design/Methods: Prospective, descriptive study of rapid
identification and linkage to care for TB suspects
registered between March 20, 2015 and April 20, 2015
in Ugu District, KwaZulu Natal, South Africa. Specimen
data and results matching from the National Health
Laboratory System (NHLS) were completed using both
barcodes from clinical specimens and South African
identification number.
Results: We enrolled 586 TB suspects through the
application. Of these, 336 (57%) were matched with
NHLS results.
Results were available via the application for the
ordering clinics in less than 40 hours. Thirty-three
(10%) were positive for drug-susceptible TB; 3(0.9%)
were identified as rifampicin-resistant patients and 16
(5%) were inconclusive and required repeat testing. For
MDR-TB patients, the average time from registration in
the application to initial contact by the tracer was 1.65
days and time from registration to treatment initiation
was 3.55 days. Of the 250 suspects unmatched with
NHLS results, 129 patients (51.6%) did not have a
barcode entered into the NHLS data system. The
OA-419-05 Fluoroquinolone use delays

tuberculosis treatment despite immediate
mycobacteriological tests
J-Y Wang,1 C-H Lee2 1National Taiwan University Hospital,
Taipei, 2Wanfang Hospital and School of Medicine, Taipei
Medical University, Taipei, Taiwan. Fax: (886) 2 2358 2867.
e-mail: jywang@ntu.edu.tw
Background: Empirical use of a fluoroquinolone (FQ) for

pneumonia can mask tuberculosis (TB). This study aimed
to investigate if immediate mycobacteriology study,
including mycobacterial culture and Mycobacterium
tuberculosis-nucleic acid amplification test (MTBNAA) can reduce delays in anti-TB treatment.
Methods: Adult pulmonary TB patients presenting as
pneumonia were identified from the National Health
Insurance Research Database of Taiwan. Those with a
prescription of FQ 77 days on the first consult for
pneumonia were classified as the FQ group. Factors
influencing the interval between pneumonia and anti-TB
treatment were evaluated using linear regression analysis.
Results: Of 16 683 patients identified, 2051 (12.3%)
were in the FQ group. The time to anti-TB treatment was
longer in the FQ group than in the non-FQ group
(63.0647.1 vs. 49.4647.6 days, P , 0.001). By
multivariate analysis, prescription of FQ 77 days was
associated with delayed anti-TB treatment (beta: 16.50
[14.39-18.61]). In sensitivity analyses, prescription of FQ
77 days remained a significant factor in (1) patients with
mycobacteriology study done on the first consult for
pneumonia, (2) out-patients without co-morbidity, and
(3) patients meeting both criteria.
Conclusion: Empiric use of FQ for 77 days delays antiTB treatment even when mycobacteriology study is
prescribed on the first visit for pneumonia.
S292
a cough (any duration), while 6 (60%) of them had at

least three symptoms of PTB. None of the identified cases
reported history of contact with a PTB patient. Comorbidity was observed in two tuberculosis cases, one
was HIV positive and one had diabetes mellitus.
Table History of other chronic illness and Xpert positivity
History of other chronic illnesses
Xpert
result
Positive
Negative
3
(10%)
3
Total
OA-420-05 In-patient wards as a source of

missed pulmonary tuberculosis cases: a case
study in Ethiopia
D Assefa Lemma,1 G W Desalgn,1 E Klinkenberg1,2 1KNCV
Tuberculosis Foundation, Addis Ababa, Ethiopia;
2
RESEARCH, KNCV Tuberculosis Foundation, Haigue,
Netherlands. e-mail: dawita8246@gmail.com
Background: Ethiopia is one of the ten countries

contributing to the 4United States Agency for International Development 4% of estimated missed cases
globally in 2013. The national TB program enhanced
community TB service implementation and scaled up the
use of new technologies and tools to ensure that people
with TB are promptly identified and treated. However,
there might be areas where cases of tuberculosis are being
missed due to low degree of clinical suspicion or
diagnostic difficulties, for example in the hospital wards
which may illustrates possible problems of case finding.
Our aim was to assess whether there are undiagnosed
pulmonary TB (PTB) cases in the inpatient wards of
hospitals.
Design/Methods: A cross sectional design was employed
to review medical records and conduct structured
interviews with medical inpatients from June 10 Sept
30, 2014. All patients without TB diagnosis in the inpatient medical wards of one large hospital were evaluated for signs and symptoms of PTB and for all one
morning sputum sample was collected for smear microscopy (LED) and molecular testing (GeneXpertw).
Multivariate analysis was conducted to assess factors
associated with missed TB cases.
Results: A total of 300 inpatients (40.6% female) were
enrolled in the study. The median duration of sickness was
30 days [IQR 14 - 240]. The median BMI was 21.5 [IQR
20 - 22.6]. Of the inpatients, 102 (34%) had no documented HIV status, while 37/198 (18.7%) were HIV
positive with a median CD4 count of 176 [IQR 52 - 400].
A total of 10 (3.3%) PTB cases were identified as M.
tuberculosis by GeneXpert of which one was smear positive. This is more than 10 times higher than in the general
population. And, of the identified cases 8 (80%) reported
No history
Chronic Diabetes HIV of chronic
bronchitis Mellitus positive illness
other
1
(10%)
37
(12%)
38
1
(10%)
30
(10%)
31
8
(80%)
131
(45%)
139
Total
10
89
(30.7%)
89
290
300
(100%)
Conclusion: All identified PTB cases in the inpatient

ward had at least one common symptom of PTB while
their medical record showed no evaluation for PTB. The
index of suspicion by clinicians should be improved.
This, together with the use of new technologies could
assist the tuberculosis case finding strategy.
OA-421-05 Diagnostic barriers among newly

diagnosed pulmonary TB patients in
Kyrgyzstan: patient delay
C Zhumalieva,1 A Estebesova,2,3 M Ryspekova,1
D Madybaeva,4 A Tursynbayeva,5,6 N Usenbaev,1
C Kamarli,7 S Verver8 1Department of State Disease
Prevention and Sanitary-Epidemiological Surveillance,
Bishkek, 2KNCV Tuberculosis Foundation Country Office,
Bishkek, Kyrgyz Republic; 3Open Society Foundations, Public
Health Program, New York, NY, USA; 4AFEW in Kyrgyzstan,
Bishkek, Kyrgyz Republic; 5KNCV Tuberculosis Foundation
Regional Office, Almaty, 6Central Clinical Hospital of the
Medicine Department of the President Administration
Affairs, Almaty, Kazakhstan; 74United States Agency for
International Development, Health and Education Office,
Bishkek, Kyrgyz Republic; 8KNCV Tuberculosis Foundation,
The Hague, Netherlands. e-mail:
aida.estebesova@gmail.com
Background: Kyrgyzstan has a tuberculosis (TB) notification rate of 102/100 000 in 2013, and WHO estimates
the TB incidence 141/100 000 population (WHO, 2012).
Very little is known about delay in diagnosis and risk
factors related to patients in the country and the Central
Asia region. The objective of this study was to assess
patient related factors contributing to delay in presenting
to the health system.
Design/Methods: We included all new adult (.18 years)
TB patients, both smear positive and negative, drugsensitive and drug-resistant, starting treatment in selected sites in Kyrgyzstan during AprilJune 2014. Two
largest urban areas and four rayons in the North and
South of the country were selected for the study. Data
was collected from medical records and structured
interviews were conducted with TB patients.
Results: Among 416 patients fulfilling the selection
criteria 383 were interviewed (92%). Delay between
first onset of symptoms and first visit to facility varied
between 0 and 207 days, with a median and mean of 28

and 36 days, respectively. Only half the patients visited a
state facility as a first step while almost one third (29%)
visited a pharmacy first. 70% of patients had more than
14 days delay between onset of symptoms and first visit
to a state facility. The following factors were significantly
associated with delay over 14 days in multivariate
analysis: being migrant (OR 2.6, 95%CI 1.3-5.3), living
in rural area (OR 2.5, 95%CI 1.3-4.6) vs. urban, having
any symptom (OR 3.7, 95%CI 1.7-7.5) vs. no TB
symptoms and having no knowledge of symptoms (OR
1.9, 95%CI 1.2-3.3). Non-symptomatic patients had
probably less delay since they were detected by screening.
Conclusion: A mean patient delay of one month is
unacceptably high and interventions should be designed
to improve care seeking behavior in the country,
especially among migrants and patients in rural areas.
Advocacy on TB symptoms is needed. Interventions are
needed to involve pharmacies in referral for TB care.
OA-422-05 Genetic determinants of the

pharmacokinetic variability of isoniazid and
rifampicin in Malawian adults with pulmonary
tuberculosis
A McCallum,1,2,3 A Schipani,2 A Owen,2
H Mwandumba,1,4,5 S Ward,5 S Khoo,2 G R Davies,2,3
D Sloan1,4,5,6 1Department of Medicine, College of Medicine,
University of Malawi, Blantyre, Malawi; 2Institute of
Translational Medicine, University of Liverpool, Liverpool,
3
Institute of Infection and Global Health, University of
Liverpool, Liverpool, UK; 4Malawi-Liverpool-Wellcome Trust
Clinical Research Programme, College of Medicine,
University of Malawi, Blantyre, Malawi; 5Liverpool School of
Tropical Medicine, Liverpool, 6Liverpool Heart and Chest
Hospital, Liverpool, UK. e-mail:
andrewmccallum@doctors.org.uk
Background: Anti-tuberculosis therapy (ATT) has concentration-dependent activity against Mycobacterium

tuberculosis. Inter-individual variability in the pharmacokinetics of ATT may be associated with emergence of
drug resistance. We evaluated the contribution of host
genetic factors to this variability.
Design/Methods: 174 Malawian adults with pulmonary
tuberculosis were assessed for single nucleotide polymorphisms (SNPs) in genes previously reported to have
functional significance for the clearance of isoniazid and
rifampicin; NAT2, SLCO1B1, AADAC and CES1. 14-21
days after commencing first-line ATT, patients returned
for plasma sampling at 2 and 6 hours post-dose. Drug
concentrations were measured by liquid chromatography
/ tandem mass spectrometry and non-compartmental
techniques were used to calculate the area under the
concentration-time curve (AUC). The effect of individual
genotypes and inferred NAT2 acetylator phenotype on
isoniazid and rifampicin exposure was assessed by
multivariate linear regression.
Results: Plasma AUCs for isoniazid and rifampicin varied
up to 20-fold between patients, despite weight adjusted
dosing. Fast NAT2 acetylator phenotype (by a 6-SNP
panel or single rs1495741 tag-SNP) was associated with
S293
a significant reduction in isoniazid exposure (P , 0.05).

The tag-SNP was in a linkage disequilibrium block with
the 6 NAT2 SNPs assessed (D0.62). However, genetic
variability at either of two SNPs of the SLCO1B1 gene
did not explain any of the variability in rifampicin AUCs.
This contrasts with other sub-Saharan African studies, in
which variant alleles on rs4149032 significantly reduced
rifampicin exposure. Rs4149032 variant allele frequency
was also much lower (0.32) in this Malawian population
than previously described from South Africa (0.70).
AADAC and CES1 SNPs did not explain inter-individual
variability in AUCs for either drug.
Conclusion: The finding that assignment of NAT2
phenotype by a single tag-SNP is associated with
isoniazid exposure may be used to simplify future
genomic analysis of clinical pharmacology studies. The
absence of a relationship between SLCO1B1 mutations
and inter-individual rifampicin variability in Malawi
suggests that pharmacogenetic determinants of rifampicin exposure may vary between African populations.
OA-423-05 Assessment of the sustained antituberculosis activity of clofazimine after

stopping treatment in mice
R Swanson,1,2 N Ammerman,1,3 J Adamson,1 C Moodley,1
A Dorasamy,1 S Modley,1 L Bester,4 J Grosset1,3
1
Durban, 2University of KwaZulu-Natal, Durban, South Africa;
3
MD, USA; 4University of KwaZulu-Natal, Westville, South
Africa. e-mail: nicole.ammerman@gmail.com
Background: In recent years, the anti-leprosy drug

clofazimine has shown promising activity for the
treatment of tuberculosis (TB), having demonstrated
both potent bactericidal and treatment-shortening activity in different experimental models and in patients.
Clofazimine has peculiar pharmacokinetic and pharmacodynamic properties, including a very long half-life,
massive accumulation in tissues and lack of early
bactericidal activity. Interestingly, clofazimine can remain in the body for weeks after stopping treatment at
levels above the minimum inhibitory concentration
(MIC) for Mycobacterium tuberculosis. Based on this,
we hypothesized that clofazimine contributes sustained
antimicrobial activity after the cessation of TB treatment.
Design/Methods: To address our hypothesis, we conducted an experimental chemotherapy study in a mouse
model of TB. BALB/c mice were aerosol-infected with M.
tuberculosis and treated with the first-line anti-TB drug
regimen with or without clofazimine for 1 or 2 months.
After stopping treatment, sustained antimicrobial activity was evaluated by monitoring bacterial re-growth in
the lungs and spleen up to 8 weeks post-treatment. The
clofazimine concentration in the serum, lungs and spleen
was determined at each time point.
Results: Clofazimine levels in the serum remained above
the MIC for M. tuberculosis for 2 or 4 weeks after
stopping treatment that was administered for 1 or 2
months, respectively. In mice treated with the clofazimine-containing regimen, the bacterial regrowth in the
S294
tissues was delayed for at least 2 or 4 weeks after

stopping the 1- or 2-month treatment regimens, respectively. In mice that did not receive clofazimine, bacterial
regrowth began immediately after stopping treatment.
Conclusion: Clofazimine does contribute antimicrobial
activity after treatment cessation. This is likely due to the
constant release of bound clofazimine from the tissues
into the serum, where free clofazimine can exert
antimicrobial activity. Thus clofazimines unusual pharmacokinetics may in fact be beneficial features of the
drug, although more work is needed to optimize both the
dose and duration of its use in TB treatment.
OA-424-05 Pharmacokinetics of first-line antituberculosis drugs in children with tuberculosis

with and without HIV coinfection
A Kwara,1,2 A Enimil,3,4 H Yang,5 F Gallani,1,2
A Dompreh,3 A Orstin,3 L Wiesner,6 S Antwi3,4 1The
Miriam Hospital, Providence, RI, 2Warren Alpert Medical
School of Brown University, Providence, RI, USA; 3Komfo
Anokye Teaching Hospital, Kumasi, 4Kwame Nkrumah
University of Science and Technology, Kumasi, Ghana;
5
University of Rochester School of Medicine and Dentistry,
Rochester, NY, USA; 6University of Cape Town, Cape Town,
akwara@lifespan.org
Background: Given the concern that the high frequency

of low plasma peak concentration (Cmax) of the
antituberculosis (anti-TB) drugs among children may be
associated with poor treatment outcomes, the World
Health Organization (WHO) recently recommended
higher dosages of the first-line anti-TB drugs for children.
We investigated the pharmacokinetics (PK) of the firstline anti-TB drugs in children who were mostly
prescribed revised dosages.
Design/Methods: Children aged 3 months to 14 years old
with tuberculosis on first-line therapy for at least 4 weeks
had blood samples collected at pre-dose, 1, 2, 4, and 8
hours post-dose. Drug concentrations were determined
by validated liquid chromatography mass spectrometry
methods and PK parameters calculated using noncompartmental analysis. Factors associated with Cmax
below the target range of each drug were examined
using univariate and multivariate analyses.
Results: Of the 62 children, 32 (51.6%) were male, 29
(46.8%) were younger than 5 years old and 28 (45.2%)
had HIV coinfection. The median (IQR) plasma Cmax
was 4.8 (3.7 6.4) lg/mL for isoniazid, 6.3 (3.5 8.8) lg/
mL for rifampin, 28.6 (21.8 35.6) lg/mL for
pyrazinamide and 1.9 (0.9 3.1) lg/mL for ethambutol.
Compared with published therapeutic ranges, low
plasma Cmax was observed in 10/62 (16.1%), 40/62
(64.5%), 10/59 (16.9%) and 30/59 (50.8%) for isoniazid, rifampin, pyrazinamide and ethambutol, respectively. None of the factors evaluated significantly influenced
isoniazid Cmax. Lower rifampin dose was associated
with rifampin Cmax , 8 lg/ml. Younger age, lower body
weight, shorter height, lower weight-for-age z-score,
lower height-for-age z-score, and smaller head circumference were associated with pyrazinamide Cmax , 20
lg/ml and ethambutol Cmax , 2 lg/ml. Children with

HIV coinfection were also more likely to have low
ethambutol Cmax. In the multivariate analysis, younger
age (odds ratio, 1.23; P 0.023) and lower rifampin dose
(odds ratio, 1.28; P 0.004) were associated with low
rifampin Cmax. Shorter height (odds ratio, 1.05; P
0.011) was associated with low pyrazinamide Cmax.
Shorter height (odds ratio, 1.04; P 0.009) was
associated with low ethambutol Cmax.
Conclusion: Despite the use of revised drug dosages, low
rifampin and ethambutol Cmax were common among
Ghanaian children with TB. These findings suggest that
higher rifampin and ethambutol dosages than currently
recommended are needed to achieve target Cmax in
children.
OA-425-05 Tuberculosis among individuals

presenting with community-acquired
pneumonia to an emergency room at a referral
hospital in Gaborone, Botswana
J Gersh,1 Z Feldman,2 E Greenberger,1 H Friedman,3,4
A Chandra,5 T Lere,6 M Haas,3,4,7,8 A Ho-Foster3,4
1
University of Washington, Seattle, WA, 2University of
Colorado, Boulder, CO, USA; 3Botswana U Penn Partnership,
Gaborone, Botswana; 4University of Pennsylvania,
Philadelphia, PA, USA; 5University of Botswana, Gaborone,
6
Ministry of Health, Gaborone, Botswana, 7University of
Colorado, Denver, CO, 8Denver Public Health, Denver, CO,
USA. e-mail: jillgersh@gmail.com
Background: Delays in diagnosing Tuberculosis (TB) are

associated with increased transmission and worsened
outcomes. Overlap in the clinical presentation of TB with
conditions such as community-acquired pneumonia
(CAP) can contribute to diagnostic delays. Identifying
opportunities to diagnose TB earlier has the potential to
improve outcomes. Our primary objective was to
determine the frequency of TB among patients presenting
with CAP at a large referral hospital in Gaborone,
Botswana.
Methods: We performed a retrospective cohort study of
adults presenting with confirmed(defined as new chest
radiograph findings) or presumed (no chest radiograph
obtained) CAP from April 2010-October 2011 to the
Emergency Department (ED). This cohort was matched
to the National Botswana Tuberculosis Registry to
identify individuals subsequently diagnosed with TB.
We assessed demographic characteristics, time to TB
diagnosis, clinical outcomes and performed logistic
regressions to identify factors associated with a diagnosis
of TB.
Results: We identified 1305 individuals presenting with
CAP and of those, TB was subsequently diagnosed in 64
(4.9%). 63% (n 40) were diagnosed with TB within 15
days of their presentation for CAP and 88% (n 56)
presented within 90 days. The median time to TB
diagnosis after CAP presentation was 11 days (IQR 543 days). 39% were AFB sputum smear positive, 27%
were smear negative and 34% were unknown. 48% of
those later diagnosed with TB were admitted to the
hospital, compared to 56% without TB (P 0.286). HIV
status could be determined in 84% of those with TB

compared to 68% of those without TB (P 0.005). Of
these, 87% were HIV-TB and 87% were HIV positive
without TB (P 0.94). There were no differences in vital
signs, respiratory symptoms, weight loss, night sweats, or
history of TB. Older individuals were significantly less
likely to be diagnosed with TB within 90 days [OR for
one year increase in age 0.977, 95%CI (0.957, 0.997)].
Individuals with known HIV status were more likely to
be diagnosed with TB in 90 days than those with
unknown HIV status [OR 3.36, 95%CI (1.51, 7.48)].
Conclusion: Subsequent presentation with TB is common
among individuals presenting with CAP at our ED,
suggesting that TB may be present at the time of CAP
presentation. Given the lack of distinguishing clinical
factors between pulmonary TB and CAP, all adults
presenting with CAP should be evaluated for active TB in
Botswana.
S295
being exchanged more freely from labs to many health

information systems. Lessons: Plan and implement at the
national level. Connectivity presents immediate results
and reduces commodity waste.
Conclusions and key recommendations: In addition to
improving the linkage to care, this approach demonstrates an ability to reduce diagnostic errors and stockouts, yielding considerable return on investment for a
relatively modest investment in connectivity. Donors and
MOHs should consider building connectivity into
implementation plans.
OA-426-05 Rapid turnaround leads to better

care: designing country solutions for effective
data exchange
J Takle1 1Global Connectivity, Inc., Boston, MA, USA. e-mail:
jeff.takle@globalconnectivity.co
Background and challenges to implementation: Historically, lab services have been a major bottleneck due to
slow information transfer from the labs to the broader
healthcare system. Connectivity solutions for the GeneXpert and other diagnostic devices hold the potential to
exchange diagnostic data directly into the health system
within a few seconds of diagnosis, rather than days or
weeks later, and studies suggest this rapid turnaround of
data leads to higher enrollment rates onto care.
Implementation poses new challenges with management
of modems, SIMs, data plans, and related services.
Challenges persist with inconsistent power and network
availability, and limited computer skills.
Intervention or response: Solutions for connectivity
emerged from 2012-2015 that enable the GeneXpertw
and devices from Alere, Abbott, and others to rapidly
transmit results. They rely on simple USB modems and
SMS/2G/3G networks to move information and have
largely been designed to operate in the developing world
context. These systems can exchange results into EMRs,
LMIS, and other databases. They can auto-alert patients
and doctors via SMS text when results are ready; they can
auto-populate the existing Excel-based reports used by
NTLPs, donors, and projects; and, they enable alerts
when machines exceed acceptable error rates. Connectivity decisions are increasingly made at the NTLP level,
which streamlines implementation and enables more
efficient procurement, cartridge management, and module replacement.
Results and lessons learnt: Connectivity solutions are in
place today at over 500 labs in 23 countries, covering .5
million patient results, and those networks continue to
grow rapidly. More than 100 000 text message alerts are
helping enroll patients into appropriate care. Errors are
measured and error rates are dropping as a result. Data is
17. TB & NCD co-morbidity

OA-427-05 Glycemic control and risk of
tuberculosis among diabetic patients: a cohort
study
P-H Lee,1 H Fu,2 T Lai,3 H-H Lin2 1Centers for Disease
Control, Taipei, 2Institute of Epidemiology and Preventive
Medicine, College of Public Health, National Taiwan
University, Taipei, 3Department of Medical Research,
Mennonite Christian Hospital, Hualien, Taiwan. e-mail:
leepinhui@gmail.com
Background: Diabetes increases the risk of developing

active tuberculosis (TB). However, it remains unclear
whether improving glycemic control in diabetic patients
could lower the risk among these populations.
Methods: We conducted a cohort study of participants
from the community-based integrated health screening
survey in northern Taiwan from 2005 to 2008. The
baseline information of demographics, behavioral risk
factors, and fasting plasma glucose was obtained from
the survey. The baseline status of diabetes was determined through linking the health survey database to the
national health insurance database using the individual
unique identifier. Diabetes was defined by the prescription of oral hypoglycemic agents or insulin for more than
28 days. The cohort was followed up for the development of active tuberculosis, which was defined by the
presence of TB-related ICD-9-CM code (010-018) plus
the prescription of 72 anti-TB drugs for 28 days or
longer. We divided the population into three groups
based on diabetes status and the level of glycemic
S296
control: (1) no diabetes, (2) diabetes with good control:

fasting plasma glucose 6130 mg/dL, and (3) diabetes
with poor control: fasting plasma glucose .130 mg/dL.
We used Cox proportional hazards model to estimate the
hazard ratio (HR) and 95% confidence interval (CI) for
the association between different glycemic control group
and incident TB.
Results: Among the 116 893 eligible participants, 388
incident TB cases developed during follow-up. The TB
incidence per 100 000 person-years was 68.6 (95%CI:
68.683.7) in those without diabetes, 106.6 (50.8
162.5) in diabetic patients with good control, and
169.9 (125.4214.4) in diabetic patients with poor
control. After adjusting for age, sex, tobacco smoking,
alcohol use, body mass index, frequency of outpatient
visit utilization, and medical comorbidities, we found
that diabetic patients with poor control had significantly
increased risk of TB (adjusted HR1.80, 95%CI: 1.34
2.40) than those without diabetes. On the other hand,
good glycemic control in diabetic patients lowered the
risk of TB (adjusted HR0.85 compared with nondiabetic individuals, 95%CI: 0.501.47).
Conclusions: Fasting plasma glucose predicted the risk of
TB regardless of diabetes status. Good glycemic control
could potentially reverse the risk of TB among diabetic
patients.
recorded. We analyzed data collected using Chi-square

test and independent T-test (mean comparison test).
Results: From a total of 253 eligible TB patients screened
for DM, we found 52 (20.6%) with TB-DM, of whom,
41 (78.8%) were previously known DM and 11 (21.2%)
were newly diagnosed. TB-DM patients were significantly older than TB-only patients (mean 51.2 vs. 38.8
years; P , 0.001), had a higher mean body mass index
(BMI) (22.2 vs. 18.1; P , 0.001) and were more likely to
have hypertension (30.1% vs. 6.0% P , 0.001).
However, fewer TB-DM patients presented with anemia
(55.1% vs. 72.5% P 0.018). GeneXpertw MTB/
RIFwas examined in 6 TB-DM patients and 21 TB-only
patients. Rifampicin resistance was found to be higher in
TB-DM patients (33.3% vs. 19.1%; P 0.46), but not
statistically significant.
Conclusion: There was a greater proportion of known
DM found, therefore physicians should be recommended
to ask history of DM to every TB patients. Screening for
DM should be prioritized for pulmonary TB patients
who are older, have higher BMI, non-anemic, and with
hypertension prior to the treatment of TB.
OA-429-05 High burdens of medical comorbidity and drug toxicity on the Category II
retreatment regimen in Malawi
R Koesoemadinata,1 S Mcallister,2 N Rahmadika,1

P Santoso,1 R Ruslami,1 P Hill,2 R Van Crevel,3
B. Alisjahbana1 1Faculty of Medicine Universitas
Padjadjaran, Bandung, Indonesia; 2University of Otago,
Dunedin, New Zealand; 3Radboud University, Nijmegen,
Netherlands. e-mail: craspati@yahoo.com
D Cohen,1,2 K Mbendera,3 W Malwafu,4 S Greenwood,5

E Joekes,2,5 L Corbett,1,6 G R Davies,7 S B Squire2 1Malawi
Liverpool Wellcome Clinical Research Programme, Blantyre,
Malawi; 2Liverpool School of Tropical Medicine, Liverpool,
UK; 3Malawi National Tuberculosis Control Program,
Lilongwe, 4College of Medicine, University of Malawi,
Blantyre, Malawi; 5Royal Liverpool University Hospital,
Liverpool, 6London School of Tropical Medicine & Hygiene,
London, 7University of Liverpool, Liverpool, UK. e-mail:
danbcohen@yahoo.com
Background: Diabetes Mellitus (DM) is a known risk

factor for tuberculosis (TB) disease and also treatment
failure of TB. Screening for DM among TB has been
recommended by The WHO. However, in resource
limited settings, screening may not be feasible for every
TB patient. This study aims to define characteristics of
TB patients that have a greater association with DM to
allow for more directed screening.
Design/Methods: As part of the TANDEM program
(www.tandem-fp7.eu) on TB and DM in different
countries, a cross sectional study was done between
February-December 2014. Presumptive TB cases were
recruited from DOTS Clinic in Hasan Sadikin Hospital
and 30 Primary Health Clinics in Bandung City,
Indonesia. Patients underwent history taking, physical
examination, chest X-ray, and sputum examination for
smear and culture for Mycobacterium tuberculosis.
Those who were found to have active pulmonary TB
were eligible for inclusion and screened for DM using a
structured questionnaire and laboratory HbA1c test.
Patients without history of DM with two laboratory
HbA1c results .6.5% were defined as definite DM,
while for patients with history of DM, laboratory HbA1c
test only done once and their DM medication were
Background: The WHO category II retreatment regimen

is widely used but poorly evaluated. Data are lacking
about the safety of the drug regimen, and outcomes are
consistently reported to be poor with treatment success
rates varying from 27 83%. However, reasons for poor
clinical outcomes are unclear particularly in settings
where the prevalence of drug resistance is low. The aim of
this study was to investigate if any clinical factors are
associated with unsuccessful outcome in patients being
retreated for TB.
Design/Methods: A prospective cohort of patients
starting standard WHO Category II retreatment regimen
was recruited at a central hospital in Blantyre, Malawi.
Patients were evaluated for respiratory co-morbidity (CT
thorax and pulmonary function testing); anaemia; renal
impairment; diabetes; and ART failure. Systematic
screening for drug toxicity included baseline and
month-2 audiolology, renal function, and clinical evaluation.
Results: Of the 158 patients who were recruited, 57%
had TB confirmed on sputum smear, culture or GXP. The
prevalence of medical co-morbidities was high: 131/158
(83%) of patients were HIV positive, of whom 96 (73%)
were already on ART. Of the 63 patients on ART for
OA-428-05 Screening for diabetes mellitus in

tuberculosis patients: who should be
prioritized?
.year, 24 (38%) had an HIV viral load .5000 copies/

ml. Chronic lung disease was found in 88% on CT
thorax, including scarring in 80%, bronchiectasis in
61%, COPD in 22%, and destroyed lung in 19%.
Spirometry revealed restrictive deficit (FVC ,80%
predicted) in 60%, and obstructive deficit (FEV1
,70% predicted) in 7% of patients. Anaemia (Hb
,10g/dL) and renal impairment (CrCl ,90%) were also
common (34% and 45% respectively). Only 3% had
Hba1C .6.5%. During the first 2 months of treatment
ototoxicity (change of 710db at 2 frequencies or 720db
at 1 frequency) was seen in 32%. Nephrotoxicity (25%
decrease in CrCl or 1.5 fold increase in Cr) was seen in
15%. The rate of self-reported peripheral neuropathy
was 40%. Drug induced liver injury developed in 4%,
and 1 patient had rash requiring treatment interruption.
Conclusion: The burdens of co-morbidity and drug
toxicity in this cohort of patients on Category II regimen
were extremely high. If outcomes are to be improved in
TB retreatment, there is an urgent need to address the
impact of other co-morbid medical conditions including
chronic lung disease and ART failure, and the serious
toxicities associated with treatment such as renal injury
and ototoxicity.
OA-430-05 Tuberculosis and diabetes in

Southern Uganda
S Annette1 1RESEARCH, Mulago-Mbarara Teaching
Hospital, Kampala, Uganda. e-mail: daizyjulian@yahoo.com
Background: To determine the impact of diabetes on the

rates of tuberculosis in a region where both diseases are
prevalent.
Design/Methods: Data from a population-based cohort
of patients with pulmonary tuberculosis undergoing
clinical and mycobacteriologic evaluation (isolation,
identification, drug-susceptibility testing, and IS6110based genotyping and spoligotyping) were linked to the
2010 National Health Survey (ENSA2010), a national
probabilistic, polystage, stratified, cluster household
survey of the civilian, noninstitutionalized population
of Uganda.
Results: From March 2010 to March 2012, 581 patients
with Mycobacterium tuberculosis culture and fingerprint
were diagnosed, 29.6% of whom had been diagnosed
previously with diabetes by a physician. According to the
ENSA2010, the estimated prevalence of diabetes in the
study area was 5.3% (95%CI 4.16.5). The estimated
rates of tuberculosis for the study area were greater for
patients with diabetes than for nondiabetic individuals
(209.5 vs. 30.7 per 100 000 person-years, P , 0.0001).
Conclusion: In this setting, the rate of tuberculosis was
increased 6.8-fold (95%CI 5.78.2, P , 0.0001) in
patients with diabetes due to increases in both reactivated and recently transmitted infection. Comorbidity with
diabetes may increase tuberculosis rates as much as
coinfection with human immunodeficiency virus (HIV),
with important implications for the allocation of health
care resources.
S297
OA-431-05 Diabetes in patients with

tuberculosis in Africa: to test or not to test?
N Boillat Blanco,1,2,3 L T Minja,3 C Schindler,1 R Kaushik,4
S Gagneux,1 C Daubenberger,1 K Reither,1
N Probst-Hensch1 1Swiss Tropical and Public Health Institute,
Basel, 2University and University Hopital, Lausanne,
Switzerland; 3Ifakara Health Institute, Dar Es Salaam, 4Hindu
Mandal Hospital, Dar Es Salaam, Tanzania. e-mail:
noemie.boillat@chuv.ch
Background: Diabetes mellitus (DM) increases the risk of

active tuberculosis (TB) while, on the other side, TB, as
an infectious disease, tends to lead to hyperglycemia. It is
therefore important to identify DM in TB patients but
the most efficient algorithm for DM screening among TB
patients, particularly in low resource settings, has to be
established. We prospectively compared healthy controls
with new TB patients over 5 months follow-up using the
3 recommended DM screening strategies in Dar es
Salaam, Tanzania.
Design/Methods: Consecutively recruited adults with
new active TB were included in a longitudinal casecontrol study between July 2012 and June 2014. Healthy
volunteers matched to TB patients for sex and age, free of
acute infection and without past history of TB, were
enrolled in the same recruitment area. Screening for HIV
was done at enrolment. All participants underwent three
tests for the diagnosis of DM (fasting capillary glucose
(FCG), 2-hours capillary glucose (2h-CG) and glycated
hemoglobin (HbA1c; Roche Cobas Integra)) at enrolment and after at least five months of TB treatment. The
association between TB and DM was assessed using
logistic regression after adjustment for age, HIV status
and socioeconomical status.
Results: Overall, 539 TB patients and 496 healthy
controls were included. At enrolment, DM prevalence
was significantly higher in TB cases compared to controls
using any DM diagnostic test and irrespective of HIV
status (Figure). None of the patients newly diagnosed
with DM was treated. A significant association between
newly diagnosed DM and TB was no longer found after 5
months of TB treatment with any of the screening tests
(enrolment vs. follow up (adjusted (a)OR(95%CI): FCG
4.7(1.1-20.7) vs. 1.7(0.3-11); 2-h CG 1.42(0.6-3.7)) vs.
0.5(0.1-1.7); HbA1c 3.9(1.5-10) vs. 1.2(0.4-3.6)) (Figure). DM and pre-DM diagnosed at enrolment with FCG
and 2-h CG were associated with adverse TB outcome
(lost to follow up-death-failure), aOR (95%CI) 2.3(14.9) and 2.3(1.2-4.3), respectively.
Conclusion: In this African setting with low DM
prevalence, hyperglycemia seems to be the result rather
than the cause of TB, as it was no longer present after 5
months of TB treatment. While screening for an
underlying DM, particularly with HbA1c, at the time
of TB diagnosis has no clinical utility, testing for the
presence of transient hyperglycemia with FCG may be
relevant for the success of TB treatment.
S298
[634.6] (P 0.02), despite the majority of DM/TB

having cavitary disease.
Conclusions: Early therapeutic drug monitoring for DM/
TB may shorten the time to sputum culture conversion
compared to age, gender and disease matched non-DM/
TB patients. These findings warrant further controlled
study.
OA-433-05 Associations between tuberculosis

and hyperglycaemia in Zambia, Southern
Africa
OA-432-05 Sputum culture conversion of

diabetes/tuberculosis patients compared to
age, sex and disease-matched patients without
diabetes in Virginia
S Heysell,1 D Staley,2 D Dodge,2 S Keller,2 E Houpt,1
J Moore2 1University of Virginia, Charlottesville, VA, 2Virginia
Department of Health, Richmond, VA, USA. e-mail:
scott.heysell@gmail.com
Background: Diabetes (DM) complicates TB treatment

including a prolonged time to sputum culture conversion
and increased risk of death compared to non-DM/TB
patients. Since 2013 in Virginia, recommendations have
been in place to perform early therapeutic drug
monitoring and dose correction for isoniazid and
rifampin after 2 weeks of therapy for DM/TB patients.
Further effort has been made to screen those without
known DM by HgbA1c.
Methods: A retrospective study of the state TB registry
was performed for patients initiating treatment in 20132014. DM/TB patients undergoing early TDM were
compared to non-DM/TB patients by 1:1 matching on
age (/- 10 years), gender and sputum smear status.
Subjects were included if their pretreatment sputum
culture was positive and excluded if they had extrapulmonary TB, other immunosuppressing conditions,
HgbA1c 76.5% (if not previously known to have
DM), or any first-line drug resistance. DM/TB and
matched non-DM/TB patients were then assessed for the
time to sputum culture conversion.
Results: 373 patients started treatment, of which 60
were DM/TB. 19 DM/TB patients met study criteria,
including 10 (53%) men and 16 (84%) smear positive,
and were compared to 19 matched non-DM/TB
patients. DM/TB patients mean age was 61.5 years
[613.6] compared to non-DM/TB controls of 58.7 [6
15.7] (P 0.57). In DM/TB, the isoniazid mean C2hr
(daily dosing) was 2.5 lg/ml [61.8] and 12 (66%) were
in range for dose increase, rifampin mean C2hr was 8.4
lg/ml [66.8] and 8 (47%) in range for dose increase,
and rifabutin C2hr of 0.57 lg/ml in one subject (within
the expected range thus no dose increase). Mean time to
sputum culture conversion in DM/TB patients was 36.0
days [621.9] compared to non-DM /TB of 58.8 days
S L Bailey,1,2 S Floyd,2,3 J Yudkin,4 P Godfrey-Faussett,2

H Ayles1,2 1Zambart, Lusaka, Zambia, 2LSHTM TB Centre and
Department of Clinical Research, London School of Hygiene
& Tropical Medicine, London, 3LSHTM TB Centre and
Department of Infectious Disease Epidemiology, London
School of Hygiene & Tropical Medicine, London, 4University
College London, London, UK. e-mail:
slbailey@doctors.org.uk
Background: Hyperglycaemia is known to be associated

with active tuberculosis (TB), though to date few relevant
studies have originated from Southern Africa. The
strength of the association may differ in Southern Africa
due to high HIV prevalence and a high prevalence of
poorly-controlled diabetes mellitus (DM). A set of
studies in Lusaka, Zambia aim to determine the
association between hyperglycaemia and active TB in
this location; determine if HIV modifies this association;
determine the contributions of DM and stress-induced
hyperglycaemia to hyperglycaemia among newly-diagnosed TB cases, and determine the diagnostic accuracy
for DM of measures of hyperglycaemia used at the time
of TB diagnosis compared to a gold standard.
Methods: A case-control study among adult TB cases and
population controls in Lusaka defines active TB disease
as the outcome and hyperglycaemia (random blood
glucose (RBG) 711.1mmol/L) as the exposure of
interest. HIV status is defined by serological result.
Multivariable logistic regression is used to explore
associations. A study of diagnostic accuracy compares
RBG, fasting blood glucose (FBG) and glycated haemoglobin (HbA1c) measured at the time of TB diagnosis to
the gold standard for DM diagnosis, defined as FBG
measured 12 weeks after TB treatment commencement,
after stabilisation of TB disease.
Results: The prevalence of hyperglycaemia among 3131
TB cases was 1.3% and among 6977 controls was 1.5%.
Overall, the adjusted odds of active TB was 1.42 (95%CI
0.78-2.58) times higher among those with hyperglycaemia compared to those without. The corresponding
adjusted odds ratio among those with HIV was 5.34
(95%CI 1.21-23.62, P 0.03) and among those without
HIV was 1.00 (95%CI 0.49-2.02). Among 362 participants of the diagnostic accuracy study, only 46%, 50%
and 38% of those with hyperglycaemia measured by
RBG, FBG and HbA1c respectively at the time of TB
diagnosis were confirmed to have DM. RBG, FBG and
HbA1c showed similar test sensitivity, specificity and
positive and negative predictive values for the diagnosis
of DM at the time of TB diagnosis.
Conclusions: Overall, no evidence of association between hyperglycaemia and active TB was found in this
study population, though among those with HIV there
was strong evidence of an association. A relatively low
proportion of hyperglycaemia measured at the time of
TB diagnosis was due to DM. HbA1c showed no greater
test accuracy for DM diagnosis at the time of TB
diagnosis than FBG or RBG.
OA-434-05 Exploring the association between

TB and diabetes
F Pearson,1 M Pearce,2 R Mcnally,2 N Unwin,3 J Critchley1
1
St Georges University of London, London, 2Newcastle
University, Newcastle-Upon-Tyne, UK; 3University of West
Indies, Cave Hill, Barbados. e-mail: fpearson@sgul.ac.uk
Background: Many studies have found an increased risk

of PTB amongst those with Diabetes Mellitus (DM),
however, evidence on whether the association is specific
to disease sub-types or is bi-directional remains sparse.
This study assesses rates of TB, PTB and EPTB amongst
those with DM, Type 1 DM (T1DM) and Type 2 DM
(T2DM) comparative to those without. It also assesses
the converse (DM rates among those with prior TB
episodes) to investigate association bi-directionality.
Methods: Retrospective cohort analyses were completed
using UK primary care data collated prospectively
between 2003 and 2009 in The Health Improvement
Network database. Individuals in the dataset were
classified as either exposed to or unexposed to TB,
PTB, EPTB, DM, T1DM or T2DM. The incidence rate
ratio (IRR) and 95% confidence intervals (95% CI) were
then calculated amongst each exposure group for
outcomes of interest (TB, PTB, EPTB or DM, T1DM,
T2DM) using negative binomial regression.
Results: TB risk was raised amongst individuals with DM
(IRR 1.50 (95%CI 1.27-1.76) P value , 0.001), T1DM
(IRR 1.46 (95%CI 1.10-1.92) P-value 0.008) or T2DM
(IRR 1.54 (95%CI 1.30-1.82) P value , 0.001)
compared to those without. The risk of EPTB amongst
those with T1DM was also increased compared to those
without (IRR 2.09 (95%CI 1.19-3.66), P value 0.010).
DM risk was raised amongst those who had a prior
episode of TB (IRR 5.65 (95%CI 5.19-6.16) P value ,
0.001), PTB (IRR 5.74 (95%CI 5.08-6.50) P value ,
0.001) or EPTB (IRR 4.66 (95%CI 3.94-5.51) P value ,
0.001) compared to those with no TB history. T1DM risk
was raised amongst those with prior TB (any sub-type)
(IRR 5.49 (95%CI 5.02-6.02) P value ,0.001) or EPTB
(IRR 0.84 (95%CI 0.35-2.03) P value ,0.001) compared to those with no TB history. T2DM risk was raised
amongst those who had prior TB (IRR 2.21 (95%CI
1.68-2.91) P value ,0.001), PTB (IRR 5.38 (95%CI
4.73-6.12) P value , 0.001) or EPTB (IRR 4.36 (95%CI
3.65-5.22) P value ,0.001) compared to those with no
TB history.
Discussion TB risk was increased in those with DM and
DM risk raised substantially amongst those who had a
prior TB episode. We are not aware of other studies with
sufficient power to assess the risk of DM among those
S299
with prior TB. As DM prevalence increases, consideration of the specificity and directionality of these
associations will become important to improve disease
control and TB treatment outcomes amongst those with
co-morbid disease.
18. TB in children: diagnosis, prevention

and best practice
OA-435-05 Underdetection of TB exposure
among children admitted to a tertiary hospital
in Botswana
T David,1,2 A Steenhoff,1,3,4,5 L Mazhani,1 A Ho-Foster,3,4
T Mazhani,1,2 T Arscott-Mills3,5,4 1University of Botswana,
Gaborone, 2Botswana Ministry of Health, Gaborone,
3
Botswana-U Penn Partnership, Gaborone, Botswana;
4
University of Pennsylvania Perelman School of Medicine,
Philadelphia, PA, 5The Childrens Hospital of Philadelphia,
Philadelphia, PA, USA. e-mail: pascaldav@yahoo.co.uk
Background: In a high tuberculosis (TB) setting, such as

Botswana, an important risk factor to assess in each child
is whether they have a TB contact at home. We aimed
to determine the proportion of TB exposure detection by
clinicians as compared to the study team, in children
admitted to Princess Marina Hospital (PMH), in
Gaborone Botswana. In secondary analysis, we sought
to determine the difference between TB exposure
detection in two time periods the pre-intervention
and post-intervention phases of the study.
Design/Methods: This was a single center, prospective
cross sectional study. Participants were children (0-13
years) admitted to pediatric medical, surgical or any
neonatal ward at PMH during the pre-intervention and
the post-intervention phases of the study. The intervention was a feedback session to clinicians which gave them
results of their TB exposure direction compared to the
study and an educational handout describing how to
adequately assess TB exposure. An investigator interviewed caregivers using a standardized form to determine whether a child had been exposed to a TB contact.
To determine TB exposure detection in by clinicians,
the investigator then reviewed the childs in-hospital
medical record to assess whether a TB household
exposure was recorded in the medical records. The
proportion of TB exposure detected was compared
between clinicians and the study team using McNemars chi2 test for proportions.
Results: A total of 736 (81%) and 170 (19%) participants were enrolled in the pre and post-intervention
phases respectively. Overall, the study detected TB
exposure in 39/906 (4%) of participants compared to
15/906 (1.7%) by clinicians, P , 0.001. Pre-intervention, TB exposure was detected in 22/736 (3.0%)
participants by the study compared to 6/736 (0.9%) by
clinicians (P , 0.0001); a 3-fold difference. Postintervention, TB exposure was detected in 17/170
(10%) participants by the study compared to 9/170
(5%) by clinicians (P 0.005); a 2-fold difference.
S300
Conclusion: TB exposure detection was lower in

clinicians in both phases of the study confirming that
this important exposure is being missed in this setting. TB
exposure detection may be improved by implementing
standardized tools such as was used by the study team to
screen for TB.
OA-436-05 Focused Point-of-Care Ultrasound

(FASH) for diagnosis of childhood tuberculosis
1,2,3
S Belard,
E Banderker,4 W Isaacs,3 L Bateman,3
J Munro,3 T Heller,2 M Grobusch,2 H Zar3
1
Charite-Universit
atsmedizin
Berlin, Berlin, Germany;
2
Academic Medical Centre, University of Amsterdam,
Amsterdam, Netherlands; 3Red Cross War Memorial
Childrens Hospital, and MRC Unit on Child & Adolescent
Health, Cape Town, 4Red Cross War Memorial Childrens
Hospital, Cape Town, South Africa. e-mail:
sabinebelard@yahoo.de
Background: Pulmonary tuberculosis (PTB) is the commonest presentation of TB disease in children. The
incidence of extrapulmonary disease (EPTB) in children
presenting with symptoms of PTB has not been well
studied. This study aimed to investigate focused point-ofcare ultrasound for diagnosis of EPTB (FASH) in children
presenting with suspected PTB.
Design/Methods: Children with suspected PTB enrolled
into a prospective cohort study of novel diagnostics in
Cape Town, South Africa, underwent a clinicianperformed FASH examination to detect pericardial,
pleural or ascitic effusion, abdominal lymphadenopathy,
or focal hepatic or splenic lesions. Ultrasound scans were
reviewed by a senior pediatric radiologist. Children were
categorized as confirmed TB (microbiologically diagnosed), possible TB (clinical diagnosis only) or NOT TB
(improvement on follow-up in the absence of TB
treatment). The primary objective was to describe FASH
findings of EPTB. Funding: NIH RO1 HD058971, MRC
South Africa.
Results: 125 enrolled children with suspected PTB
underwent FASH; of these 34 (27%), 70 (56%) and 21
(17%) had confirmed, possible, and no TB, respectively;
23 (18%) children were HIV-infected and the median age
was 43 months (IQR 21; 88). Agreement on the
sonographic diagnosis between the clinician performing
FASH and the radiologist reviewing the scans was 97%.
On presentation 38 (30%) children had at least one
FASH finding compatible with active EPTB; 5 (4%)
children had multiple FASH findings. Within the
confirmed, possible and NO TB categories, 10/34
(29%), 25/70 (36%), and 3/21 (14%) children had at
least one FASH finding, respectively. The presence of
FASH findings was lower in HIV-infected compared to
HIV-uninfected children (4/23 (17%) versus 34/102
(33%), respectively). Pleural effusion was most common
(20 (16%)), followed by abdominal lymphadenopathy
(15 (12%)), pericardial effusion (7 (6%)), focal splenic
lesions or ascites (4 (3%), each), and focal hepatic lesions
(1 (,1%)).
Conclusion: Around a third of children with confirmed
or possible PTB had sonographic features of EPTB.
FASH is a promising, quick, non-invasive, radiation-free

imaging tool for visualizing features of EPTB at the
point-of-care with potential to improve diagnosis and
management of childhood TB.
OA-437-05 Piloting upfront XpertW MTB/RIF

testing on various specimens under
programmatic conditions for the diagnosis of
TB and DR-TB in a paediatric population
N Raizada,1 K S Sachdeva,2 S Swaminathan,3
A Sreenivas,4 S Kulsange,1 C Boehme,1 C Paramasivan1
1
FIND, Geneva, Switzerland; 2Central TB Division, New Delhi,
3
NIRT, Chennai, 4WHO, New Delhi, India. e-mail:
drneerajraizada@gmail.com
Background: India accounts for one-fifth of the global

(Tuberculosis) TB incidence. While the exact burden of
childhood TB is not known, TB remains one of the
leading causes of childhood mortality in India. Bacteriological confirmation of TB in children is challenging due
to difficulty in obtaining quality specimens, in the
absence of which diagnosis is largely based on clinical
judgement. Xpertw MTB/RIF (here after referred as
Xpert), a tool with a quick turn-around time, which
simultaneously detects TB and rifampicin resistance,
offers a promising solution to these TB diagnostic
challenges. In line with 2013 WHO recommendations a
pilot project was undertaken under the Revised National
TB Control Programme of India (RNTCP) in four major
cities offering upfront Xpert testing to all types of
presumptive paediatric TB and DR-TB cases. The
objectives of the project were to assess feasibility aspects
of routine upfront Xpert testing for paediatric TB
diagnosis, evaluate Xpert performance on different types
of paediatric specimen and assess the diagnostic yield of
Xpert on different types of specimen.
Methods: Upfront Xpert testing was offered to all
paediatric (0-14 years) presumptive TB cases (both
pulmonary and extra-pulmonary) seeking care at public
and private health facilities in the project areas covering
4 major cities of India.
Results: Overall 8370 paediatric presumptive TB &

presumptive DR-TB patients were tested during 8
months of project implementation. Total 9149 specimens

were tested, of which 4445 (48.6%) were non-respiratory specimens. Overall, Xpert gave 9083 (99.3 %) valid
results. Of the 8143 presumptive TB cases enrolled, 517
(6.3%, CI 5.8-6.9) were bacteriologically confirmed. TB
detection rates were two fold higher on Xpert as
compared to smear microscopy. Further, a total of 60
rifampicin resistant TB cases were detected under the
project.
Conclusion: The pilot demonstrated the feasibility of
extending Xpert testing to non-sputum specimens from
children with a very high proportion of interpretable
results. With more than a two fold increase in TB case
detection over smear microscopy and detection of
significant numbers of rifampicin resistant TB cases,
the study demonstrates the utility of offering upfront
Xpert testing to paediatric presumptive TB and DR-TB
patients under programmatic conditions.
S301
culture positive for Mycobacterium tuberculosis; 60/172

(35%) Xpert positive and 12/172 (7%) smear positive
for acid-fast bacilli. 11/81 (14%) had INH and/or RIF
resistance. Xpert remained positive in 22% and 15% of
children at months 1 and 2 respectively, compared to 6%
and 2% for culture (Figure).
Conclusion: We observed a rapid decline in the proportion of children with positive bacteriology in the first
month after treatment initiation. Xpert remained positive for longer than culture, in both DS and drug-resistant
cases. The clinical relevance of persistent positive Xpert
results in children receiving TB treatment up to 10
months requires further study. The use of clinical and
bacteriological markers of response to therapy requires
investigation in children, especially in view of planned
paediatric TB treatment shortening trials.
OA-438-05 Bacteriological response to

treatment in children with intrathoracic
tuberculosis
E Walters,1 M Van Der Zalm,1 A-M Demers,1 C Bosch,1
H S Schaaf,1 M Palmer,1 R Gie,2 A Hesseling1 1Desmond
Tutu TB Centre, Stellenbosch University, Cape Town,
2
Tygerberg Hospital, Cape Town, South Africa. Fax: (27) 021
938 9719. e-mail: ewal@sun.ac.za
Background: Children with drug-susceptible (DS) intrathoracic tuberculosis (TB) generally respond well to 6
months standard WHO-recommended therapy. However, assessing response to treatment is largely based on
clinical improvement in the absence of bacteriological
confirmation in paucibacillary disease. The bacteriological response to treatment in children with confirmed TB
has not been characterized.
Design/Methods: We enrolled children ,13 years of age
routinely presenting with suspected intrathoracic TB to
Tygerberg and Karl Bremer hospitals, Cape Town, South
Africa, from April 2012 - April 2015. Children were
eligible if 71 of prolonged cough/wheeze, fever or poor
growth, or any duration of cough with 1 of a) close
contact with known TB index case, b) reactive Mantoux,
or c) chest radiograph compatible with intrathoracic TB
were present. Study investigations included a minimum
of 2 respiratory samples (gastric aspirate, sputum,
induced sputum or nasopharyngeal aspirate) and one
stool sample for smear microscopy, liquid mycobacterial
culture and Xpert MTB/RIF. Hain MTB DRplus on
positive cultures was done to determine susceptibility to
INH and RIF. TB investigations from other routinely
collected samples were also included. In children with
Xpert or culture confirmed TB, respiratory sampling was
repeated at 1, 2 and 6 months following treatment
initiation.
Results: Culture results were available for 391/421
children (median age 15 months); 52 (13%) HIV-infected
and 200 (51%) male. 172 (44%) children started TB
treatment (clinical decision to treat). 81/172 (47%) were
confirmed by either culture or Xpert: 68/172 (40%) were
OA-439-05 Implementation of a multi-level

intervention to improve isoniazid preventive
therapy for child TB contacts in South Africa
K Du Preez,1 L Du Plessis,1 J Caldwell,2 V Azevedo,2
E Mhlope,3 A Hesseling1 1University of Stellenbosch, Cape
Town, 2City Health, Cape Town, 3United States Agency for
International Development TB Program, Johannesburg,
South Africa. e-mail: lienkiduplessis@gmail.com
Background: Isoniazid preventive therapy (IPT) is

effective in preventing tuberculosis (TB) in children
following TB exposure. Despite global recommendations, gaps persist at multiple levels of the IPT delivery
cascade. We previously (2010) showed a screening rate of
46%, IPT initiation rate of 58% and completion rate of
13% in children , 5 years of age.
Design/Methods: We assessed the impact of a multi-level
intervention to improve IPT delivery in 3 health clinics in
Khayelitsha, Cape Town, during 2013/4. Clinics were
selected by stratified random selection according to TB
caseload. The intervention included 1) HCW training, 2)
supported implementation of paper-based IPT registers,
and 3) client focussed health education. Audits for preand post-intervention periods (Quarter 1, 2013/4) were
completed to measure the impact.
Results: A total of 264 infectious adult TB cases (preaudit) and 306 (post-audit) were identified from the
electronic TB register; 248 (94%) and 297 (97%) folders
were available for review, respectively. During the pre-
S302
audit, 236/248 (95%) TB cases had documentation of

child contacts; only 47/72 (65%) of documented eligible
child contacts were screened; 39/47 (83%) of screened
contacts started IPT/TB treatment. Table 1 describes
comparisons between pre- and post-audit for key
variables, overall and by clinic. Following the intervention, names of child contacts were more likely to be
recorded in the adult folder [OR 4.9 (95%CI 0.9-49.8), P
0.032], and contacts were more likely to be started on
treatment after screening [OR 6.1 (95%CI 1.1-60.5) P
0.015]. During Q1 2014, 121 contacts were documented
to start IPT in the IPT registers, an additional 63 (52%)
to the 58 identified through the audit. The additional
children were linked to extra-pulmonary exposures (n
11), to exposures proceeding the audit period (n 17),
and to source cases from other clinics (n 35).
Conclusion: Our pre-intervention data already indicate
remarkable improvement in IPT delivery since previous
audits through all levels of the IPT delivery cascade,
likely due to emerging prioritisation of IPT by routine
services. A multi-level intervention resulted in substantial
additional improvements in IPT delivery. A large
remaining gap is the drop-out between contact identification and screening, requiring a better understanding of
processes at this level. Insight into this key area at the
community level could inform more targeted interventions in future.
OA-440-05 Diagnostic yield of XpertW MTB/RIF

assay and Mycobacterium tuberculosis culture
on respiratory and non-respiratory specimens
among Kenyan children
S Rinn,1,2 E Click,3 K McCarthy,3 W Mchembere,4
E O Okeyo,4 J Orwa,4 S Musau,4 K Cain3,4 1Harvard Medical
School, Boston, MA, 2Boston Childrens Hospital, Boston,
MA, 3U.S. Centers for Disease Control and Prevention Division of TB Elimination, Atlanta, GA, USA. 4Kenya Medical
Research Institute - Center for Global Health Research,
Kisumu, Kenya. e-mail: rinn.song@childrens.harvard.edu
Background: Microbiological confirmation of tuberculosis (TB) among young children is challenging because
obtaining the required specimens (gastric aspirate or
induced sputum) is invasive and requires specifically
trained personnel. Even when collected, diagnostic yield
of culture for Mycobacterium tuberculosis is relatively
low. The objective of our study was to assess the yield of
M. tuberculosis culture and the Xpertw MTB/RIF assay
on multiple respiratory and non-respiratory specimens
among children with high clinical suspicion for TB.
Methods: In an ongoing study in Kisumu, Kenya,
children under 5 years suspected of having TB were
enrolled based on history of prolonged symptoms despite
standard therapy (cough .4 weeks, fever .1 week,
malnutrition . 3 weeks) and presence of parenchymal
abnormalities on chest radiograph. Per enrolled child,
two of the following specimens were tested by Mycobacteria Growth Indicator Tube culture for M. tuberculosis and Xpert MTB/RIF assay: nasopharyngeal aspirate, induced sputum, gastric aspirate, string test, stool,
and urine; one blood specimen was tested by TB culture.
Results: As of March 2015, among 189 enrolled children

(median age 27 months, interquartile range 14-45
months; 25% co-infected with HIV) with final culture
results available, 24/189 (12.7%) children had confirmed TB based on at least one specimen positive on
Xpert MTB/RIF or culture. Among the 24 children with
confirmed TB, the number with at least one culture or
one Xpert MTB/RIF positive by specimen type was as
shown in the table. The combination of 2 nasopharyngeal aspirates and 2 stool specimens diagnosed 18/24
(71%) children using culture and 15/24 (63%) children
using Xpert MTB/RIF.
Conclusions: In this cohort, nasopharyngeal aspirates
and stool show promise as alternative specimens to
gastric aspiration or induced sputum collection for
diagnosis of TB in young children using culture and
Xpert MTB/RIF. Neither specimen requires hospitalization and both could be obtained at lower levels of the
health care system in high-burden settings to improve TB
diagnosis in young children.
Table:
Positive TB Culture
Positive Xpert MTB/RIF
Specimen type
Gastric aspirate
Nasopharyngeal
aspirate
Induced sputum
String test
Stool
Urine
Blood
78
18/23
43
10/23
67
63
40
22
9
6
16/24
12/19
8/20
5/22
2/22
1/18
46
42
35
48
19
N/A
11/24
9/21
8/23
11/23
4/21
N/A
OA-441-05 Risk factors for tuberculosis disease

in pediatric TB contacts identified during a TB
Reach contact tracing project in Swaziland
P Swamy,1 R Golin,1,2 P Ustero,1,2 K Ngo,1 J Glickman,2
B Mzileni,2 M Hlatshwayo,2 A Mandalakas1 1Global TB
Program, Department of Pediatrics, Baylor College of
Medicine and TX Childrens Hospital, Houston, TX, USA;
2
Baylor College of Medicine Childrens
FoundationSwaziland, Mbabane, Swaziland. e-mail:
swamy@bcm.edu
Background: Early detection of pediatric tuberculosis

(TB) continues to be a significant challenge in TB-HIV
high-burdened settings like Swaziland. Swaziland has a
TB case detection rate of 38% with 15% child TB case
notification. As limited evidence exists, identification of
risk factors for TB disease among child contacts could
inform preventive interventions in Swaziland and similar
high burden settings.
Design/Methods: Supported via WHO TB REACH
funding, Baylor College of Medicine Childrens Foundation Swaziland (BCMCF-SD) implemented standardized contact tracing in partnership with 7 local TB
clinics. Cough monitors performed contact tracing of
index cases (IC) receiving anti-TB treatment (ATT). The
aim of this intervention was to improve TB case detection
and promote Isoniazid preventive therapy (IPT) uptake.
While adult contacts were assessed, our intervention
prioritized child contacts (,15 years of age per WHO

reporting standards) and HIV-affected households. Contacts were screened for TB symptoms, HIV status,
relationship and sleep proximity to IC, household death,
and miner in the home. Data was collected regarding
diagnostic testing (GeneXpert, smear, culture, CXR) and
treatment initiation (ATT or IPT). Utilizing this dataset,
our study aimed to focus on child TB contacts as often
pediatric TB is underreported. This program provides a
unique opportunity to evaluate risk factors associated
with TB disease in child contacts living in a high TB-HIV
burdened, resource limited setting.
Results: Of 2589 ICs, 1547 reported a child contact,
yielding a total of 4438 child contacts (2.86 child
contacts/IC). 67 cases of child TB were diagnosed
(1510 cases/100 000 child contacts) of which 17 had
bacteriologic confirmation, and 50 clinically diagnosed.
TB was diagnosed 8 times more in child contacts with a
positive TB symptom screen (RR7.90, CI 4.15-15.04)
and 9 times more in HIV-infected contacts (RR8.97, CI
5.29-15.20). Child contacts were likelier to have TB if
they were ,5 years of age, child of an IC, and shared a
bed with an IC. TB was not associated with contacts
gender, recent household death, and miner in the home.
Conclusion: High TB incidence among child contacts
emphasizes the need for comprehensive contact tracing
programs. TB symptom screening and reported HIVinfection are strong predictors of TB among child
contacts. If resource-limited, the cost-effectiveness of
contact tracing programs may be improved by prioritizing HIV-affected households with younger children.
S303
develop active disseminated disease compared to adults

following infection with TB. In addition, the paucibacillary nature of childhood TB makes confirmation of
disease difficult. Though availability of new technologies
like GeneXpert MTB/RIF and updated guidelines from
the World Health Organization (WHO) have improved
diagnosis and treatment of pediatric TB cases, many
obstacles still exist in effectively implementing recommended services. Given the complexity of diagnosing
childhood TB, it is important to evaluate the feasibility of
implementing WHO guidelines and recommendations
related to TB care particularly in resource poor settings.
Intervention: The Global TB Program of TX Childrens
Hospital and Baylor College of Medicine has developed a
comprehensive approach to integrating, monitoring and
evaluating TB best practices within the Baylor International Pediatric AIDS Initiative (BIPAI) network of
pediatric HIV clinics that spans six sub-Saharan African
countries.
Results: Fourteen TB indicators were identified and
incorporated into the standardized medical record. The
indicators were stratified by age and treatment history
and cover a broad range of TB care components including
preventive and diagnostic services, drug resistance, and
treatment in relation to HIV. To improve interpretation of
data and the efficiency of operations, a graphical tool was
developed for the generated TB indicator data. Each
indicator was visually represented by bar, line or pie
charts. On a quarterly bases each clinical site is able to
generate summary statistics with accompanying visual
aids independently to evaluate their performance and
make appropriate quality improvement decisions.
Conclusions: The identification and implementation of
these 14 indicators required clinical site buy-in, regular
training, and reconciling differing country guidelines and
current professional guidelines. After addressing barriers
in implementation and quality data entry, clinical sites
are better able to make practical interpretations of WHO
TB guidelines for program monitoring and evaluation.
19. Tobacco: economics and illicit trade
OA-442-05 The power of visual data across a

multinational integrated pediatric HIVtuberculosis program
R Puttagunta,1,2,3 K Ngo,1,2,3 A Mandalakas1,2,3 1Baylor
College of Medicine, Houston, TX, 2TX Childrens Hospital,
Houston, TX, 3Global TB Program, Baylor College of Medicine
and TX Childrens Hospital, Houston, TX, USA. e-mail:
rputtagunta@gmail.com
Background: The global tuberculosis (TB) burden from

pediatric cases (,15 years) is estimated to be 4-20%.
Among people living with HIV, TB accounts for almost a
quarter of all deaths, yet there is a paucity of data related
to TB-HIV co-infection in children. Given their developing immune system, children are more likely to
OA-443-05 Preference for buying cigarettes in

packs or loose: a cross-sectional study among
male smokers in Chennai city
P Saravanan,1 M K Parangimalai Diwakar2 1Sri
Venkateswara Dental College and Hospital, Chenani, 2Ragas
Dental College and Hospital, Chennai, India. Fax: (91) 442
453 6059. e-mail: poorniis@yahoo.com
Introduction: Amongst the 1.1 billion people who smoke

worldwide, 182 million live in India, representing the
second largest group of smokers in the world after China.
India is a party to the World Health Organizations
Framework Convention on Tobacco Control, which
states that countries shall endeavor to prohibit sales of
loose cigarettes, which are more affordable for minors.
The Government of India, in the proposed amendments of
Cigarette and Other Tobacco Products Bill 2015 (COTPA) has also suggested the same. With conflicting evidence
S304
of the effectiveness of ban on sale of loose cigarettes as a

tobacco-use control advocacy, this study was contemplated to assess the preferences of male smokers in
Chennai over buying cigarettes in packs or in loose.
Methods: The present study was done among current
male smokers of cigarette who were above 18 years of
age and who reported to the dental outpatient department and consented to participate in this study. Data
collection was done using a single-examiner administered closed ended questionnaire. The questionnaire
consisted of two parts the first part gathered
information regarding their smoking dependence based
on the Fagerstrom Test for Nicotine Dependence
Questionnaire and the second part dealt with questions
regarding their preference of buying cigarettes, the most
common place where they bought the cigarette and also
the places where they preferred to smoke.
Results: 73.4% of the male current smokers who
participated in this study were low or moderately
dependent. 94.5% of the participants preferred buying
cigarettes in loose mostly in singles (67.3%), twos
(23%) and more than twos (4.5%). Those who preferred
buying in packs cited reasons that they bought in pack
due to their travel plans or if they felt difficulty in the
availability of cigarette brand of their choice. Almost all
the participants (98.3%) preferred buying cigarettes in
the street vendors rather than supermarkets and also tend
to smoke the cigarette in and around the same area of
purchase. Most of them seem to associate their smoking
habit with consumption of snacks and beverages.
Conclusion: Though a ban of the sale of loose cigarettes
would nevertheless discourage smoking; adequate implementation of this ban by authorities would pose
challenges in the form cigarette companies manufacturing packets with fewer cigarettes and the street vendors
still trying to illegally sell loose cigarettes at a premium.
OA-444-05 Raising taxes on tobacco products

by effective advocacy with government
officials
R Mahajan1 1Government of India, Mumbai, India. e-mail:
drrupeshmahajan@gmail.com
Background and challenges to implementation: As per

WHO, increasing the retail price of tobacco products
through higher taxes is the single most effective way to
decrease consumption and encourage tobacco users to
quit. Comparative to other states of India, in Maharashtra taxes on tobacco products was less (Cigarettes 20% & Bidi - 5%). In year 2012, state government
increased the tax on Bidi to 12.5% but in couple of days
that tax was rolled back to 5% again due to the political
pressure and Bidi industry lobbing. Before the 2013 state
budget session, under the National Tobacco Control
Program (NTCP), initiative was taken to raise the taxes
on tobacco products through the affective advocacy with
the State Finance department and the Finance minister.
Intervention or response: To raise tobacco products tax
in the state, initiative was launched under NTCP,
involving civil society. In the month October 2012, State
consultant approached to Chief Secretary of the state and

requested to form State Level Coordination Committee
as per the guidelines which was formed by the
government in November 2012. In the month of
December, WHO India office organised two day
conference on tobacco taxation. Using this opportunity,
NTCP State consultant had meeting with principal
secretary Finance and tobacco taxation was discussed
with him. After the SLCC formation, SLCCs first
meeting was called in January 2013, tobacco taxation
issue was discussed in detail in the meeting and taxes in
the other state was brought in the notice of Principal
Secretary, Finance and Chief Secretary. In the SLCCs
review meeting, meeting Chief Secretary again asked the
finance to consider the raise in the taxes in coming
budget session. Simultaneously, civil society organised
meetings with the State Finance Minister focusing the
agenda of tax raise, who responded positively. State
Finance Minister was very keen on banning the Gutkha
in the state and after the advocacy he was very much
interested in tobacco products tax raise.
Results and lessons learnt: In state budget 2013, taxes on
Cigarettes gone up to 25% from the existing 20% and on
Bidi gone up to 12.5% from the existing 5% and
financed department also promised about increasing
taxes regularly every year on tobacco products. Sensitization is the key.
Conclusions and key recommendations: Sensitising
higher government officials with eeffective advocacy
can yield the positive results in tobacco control.
Involvement of stakeholders is important.
OA-445-05 Advocacy at the top level resulted in

tax increase on tobacco products in Madhya
Pradesh
B Sharma1 1Free Launcer, Indore, India. e-mail:
silkdrops@rediffmail.com

consumption is quite prevalent in Madhya Pradesh state
of India, 40 percent people consume tobacco in one form
or other in the state. A large population consuming
tobacco belongs to below poverty line. In a state like
Madhya Pradesh where infant mortality and maternal
mortality rate is quite high it was difficult to convince
policy makers to take initiative for tobacco control as
most of them were not considering tobacco control as an
issue ,for them tobacco control can be done only by
awareness. Studies shows that tax increase is the most
effective way for tobacco control.
Intervention or response: Continues sensitization of
policy makers like finance minister,finance secretary,health secretary,commissioner food and drug,nodal
officer was done at state level in which a factsheet on
tax on tobacco products in other states was shared with
them.The issue of tax increase on tobacco products was
also raised during state tobacco control committee
meeting. One to one contact was made with elected
representatives to convince them about the importance of
tax increase. Chief Minister and Health Minister were
motivated to give their message for tobacco control

which has resulted in increased political will and getting
support of elected representatives and other important
stakeholders.Governor was also oriented on various
issues of tobacco control.Other important stakeholders
were involved through their capacity building.
Results and lessons learnt: Major decision was taken by
state government and Value Added Tax (VAT) on all
forms of tobacco product increased from 13 to 27%. The
VAT amendment bill was passed by state assembly on
March 2014 and notification passed in April 2014.
Conclusions and key recommendations: Topllevel advocacy is very effective in bringing major policy changes for
tobacco control in the state. Regular Capacity building
and awareness of elected representatives and other
important stakeholders is important in increasing tax
on tobacco products.
OA-446-05 Indias single cigarette economy: a

back of the envelope assessment of its volume
and size
R Thakur,1 P Lal,2 A Pandey2 1World Health
Organization-Randstad, Shimla, 2International Union Against
Tuberculosis and Lung Disease South East Asia Office, New
Delhi, India. Fax: (91) 177 262 7601. e-mail:
ravindermph@gmail.com
Background: Sale of single cigarettes is an important

factor to the spread of the tobacco epidemic globally.
Single cigarette sale is associated with greater accessibility to minors and youth as it lowers immediate cost of
purchase. It also provides perverse benefits for vendors
and lowers the operating margins for cigarette companies. Singles stand to undermine public health effect of
tax and distort fiscal policy and administration of tax.
Methods: In February 2014, a survey identified three
prominent vendors of 10 cities. Through participant
observation and end of day interviews, the sale of
cigarettes in pack or single sticks, brands, their price
and volumes sold in a single day were recorded and
validated. Percentages and means were used to analyse
and aggregate the survey data and simple back-of-theenvelope methods were used to arrive at broad results.
Results: There are currently no legal provisions that
restrict sale of single cigarettes in India and other
developing countries. We estimate that about 89.7% of
all cigarettes are sold as single sticks. Depending on the
segment (size, and filter or non-filter) retailers have
margins per single cigarette sold within every city. There
are small variations in the price of sale of pack and
singles across the country. The net worth of the single
cigarette market is Rs 35.9 billion (or 0.598 bn USD, 1
USD 60). The single cigarette economy is nearly 30
percent of the revenues from cigarette taxes (Rs 121.33
billion or USD 2.02 bn in 2012-13). This is the price
which the consumers pay but is not captured through tax
and therefore pervades in an informal economy. The
widespread existence of the single cigarette market
buffers for any significant increase tax which is passed
S305
on to the consumer which defeats the public health goals

and fiscal rationale of taxing cigarettes.
Conclusion: Tracking price of single cigarette is an
efficient measure of raising taxes on cigarettes (in the
absence of any other rationale) on a sustained annual
basis by the Government of India, and can help achieve
the desired goals of garnering revenues and meeting
public health objectives. We repeated the survey in 2015
and found that the volumes and revenues have increase.
We are in the process of data validation and will present
the trends in sale of single cigarette in India (2013-14 and
2014-15). We will also deliberate upon practical
implications of excise and VAT increase on retail price
of cigarettes.
OA-447-05 Socio-economic and environmental

determinants of tobacco products
consumption: a case study of Pakistan
N Arshad,1 B Wasim,2 A K Chandio3 1The Network for
Consumer Protection, Islamabad, 2Federal Buearu of
Statistics, Islamabad, 3Ministry of Health, Islamabad,
Pakistan. e-mail: naureenarshad@hotmail.com
Background: Pakistan is the sixth most populous country

with highest tobacco consumption among South East
Asian region. Every year around 110 000 tobacco related
deaths occur in Pakistan. 40 % of male and 9% of female
are smoker and this number is growing. The most
prevalent mode of tobacco use is cigarette smoking
whereas other products of tobacco like Paan, Gutka and
Naswar are very eminent. So far this is the only study
which has used Household level data from HIES (201011) to calculate the percentage of tobacco consumption
expenditure at national level. The objective of the study
is to investigate the demographic and social determinants
of tobacco use and calculated the percentage of tobacco
consumption expenditure.
Design/Methods: Secondary data analysis of HIES
(2010-11). Dependent variables includes Tobacco consumption expenditure and Demographic and dummy
variables were created to assess the impact of tobacco
consumption at the household level. Statistical analysis
of the data was done in SPSS 16. Descriptive and multiple
regression analysis was done by using least square model.
Results: Percentage expenditure on tobacco consumption
in rural areas is 1.36 % and 0.96% in urban settings. The
highest consumption expenditure of 2.98% is observed
in rural Sindh, whereas the lowest 0.52% is observed in
urban areas of Khyber Pakhtunkhwa province. Expenditure on smoke-able products is higher than non-smoke
able tobacco product in Pakistan. Spending on tobacco is
higher in HH with average age group equal or more than
35 years i.e. 1.44%. Tobacco consumption expenditure
is more in male and divorced individuals. Higher tobacco
spending observed in HH with three or more adult male
members, more employed males and HH with per capita
income ranging from 3500 to 5000 per month. Multiple
regression analysis showed significant negative correlation with average years of schooling and per capita
income. Lack of environmental factors such as poor
drinking and toilet facility and non-availability of gas
S306
and electricity have significant negative association with

tobacco consumption expenditure
Conclusion: Higher HH spending per month is observed
in males in rural areas, illiterate, divorced and with poor
living condition. The tobacco control efforts should be
targeted towards low socioeconomic, rural and marginalized communities.
OA-448-05 A closer look at illicit white

cigarettes
H Ross,1 N Vellios,1 K Clegg Smith,2 J Cohen2 1University of
Cape Town, Cape Town, South Africa; 2Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, USA.
e-mail: nicolevellios@gmail.com
Objective: To better understand the phenomenon of Illicit

White cigarettes: how they are defined, both by the
tobacco industry and other sources, their presence at various markets, brand proliferation, prices, manufacturers,
location of manufacturing facilities, trademark ownership and compliance with tax and warning labels laws.
Methods: A review of scientific articles, news articles,
online trademark registries and customs and tobacco
industry documents was the primary data source for the
definitions, market presence, brand proliferation, manufacturers and trade mark owners. The data from the
John Hopkins Bloomberg School of Public Healths
Tobacco Pack Surveillance System (TPackSS) provided
additional information on the manufacturers, prices and
compliance with local tax and warning label laws
Results: We identified 82 Illicit White brands and 54
Illicit White manufacturers that operate at least 83
production facilities. A specific Illicit White brand is
generally sold in a country either legally or illegally.
Among the 35 Illicit White brands in the TPackSS
database, two thirds had neither the correct health
warning nor the required tax stamp at least in one
country, and about 87% of illegal Illicit Whites were
purchased outside large retailers. Both government
agencies and Transnational Tobacco Companies (TTCs)
report an upward trend in the market presence of illegal
Illicit Whites, particularly in the European Union. One
third of the manufacturers are located in the Free Zones
of Russia, Cyprus and UAE. Some TTCs are also involved
in the production of Illicit White cigarette brands. The
trademark ownership is very complex with multiple
owners dividing the global market into geographical
segments and registering different features of a brand.
Conclusions: Similar to TTC brands, Illicit White brands
are sold both legally and illegally, and a portion of them
are distributed without paying the proper taxes. Illicit
cigarette trade is not going to be solved by eliminating
Illicit Whites, because the majority of illicit cigarette still
consist of TTCs brands. The illicit cigarette trade can be
addressed by adopting the WHO Protocol to Eliminate
Illicit Trade in Tobacco Products and implementing the
global tracking and tracing system.
OA-449-05 Pricing strategies for electronic

nicotine delivery systems: gateway or
deterrent?
C Cherukupalli1 1Johns Hopkins University, Baltimore, MD,
USA. e-mail: rajeevrc2063@yahoo.com
Background: As an evolving and growing product

category with a projected US$ 2 billion global market,
ENDS pose challenges for regulation and taxation. The
main argument to tax e-cigarettes is that higher prices
deter use. The main argument against taxing them is that
if e-cigarettes help current smokers quit smoking, higher
prices discourage substitution and experimentation
especially relevant considering the tripling of use by
teenagers in the US within two years. Pricing data on
ENDS, and comparisons to combustible cigarettes have
both been lacking in discussions of recommendations on
tax and price regulations.
Design/Methods: A systematic review of internet retail
sites and manufacturers websites was done to establish
product categories and price points at which ENDS are
sold. Company and industry analyst reports were
analyzed to understand pricing strategies and cost
structures. Retail prices were also compared with
wholesale prices listed on a prominent wholesale trade
website.
Results: Retail prices for ENDS in the United States in
2015 varied fundamentally by whether ENDS products
are disposable (US $ 2 to US $11) or refillable ($ 25 to $
200), a much larger price variation than for combustible
traditional cigarettes (US$ 4.75 to US$ 12.50 in 201).
Price listings were frequently found to be accompanied
by rebates and discounts of up to 50% over listed prices.
Wholesale prices for importers of products were up to
350% cheaper than retail prices, indicating a considerable markup on product cost.
Conclusion: Refillable ENDS have both a fixed cost and
a variable cost associated with repeated use and a
significant premium over traditional cigarettes. By
contrast, disposable e-cigarettes including brand
variants of refillable ENDS are available at prices
comparable to cigarette pack prices. Four phenomena the marketing of starter kits, presence of disposable
variants of refillable brands, large discounts, and
comparisons of lifecycle cost savings of ENDS and
traditional cigarettes underscore active price competition and company strategies to encourage experimentation by potential users. Globally, a large supplier base
with low unit prices suggests that price competition can
keep ENDS cheap in several markets. This points to a
future role of pricing regulations as as an important
complement to regulating ENDS.
OA-450-05 Smokeless tobacco consumption

among adolescents: a global epidemiological
perspective
I Agaku,1 L Ayo-Yusuf2 1U.S. Centers for Disease Control
and Prevention, Atlanta, GA, USA; 2Sefako Makgatho Health
Sciences University, Pretoria, South Africa. e-mail:
WGN9@CDC.GOV
Background: Smokeless tobacco (ST) products can lead

to tobacco addiction, disease, and death. The cultural
differences in global adult tobacco use suggest that
adolescents tobacco use may also vary across cultures.
We assessed ST use among adolescents in 57 countries
during 2007-2010 using school-based surveys.
Design/Methods: Data were analyzed for 57 countries
for which nationally representative data were available,
using the 2007-2010 Global Youth Tobacco Surveys (for
55 countries); the 2009 US National Youth Tobacco
Survey, and the 2008-2009 Canada Youth Smoking
Survey. Number of countries by World Health Organization (WHO) region were: Africa, AFRO (n 11);
Eastern Mediterranean, EMRO (n 9); Europe, EURO
(n 11); the Americas, AMRO (n 14); South-East Asia,
SEAR (n 8), and Western Pacific, WPRO (n 4).
Analyses were restricted to persons aged 13-15 years in
each country. Current ST use was defined as past 30-day
use of chewing tobacco, pinch, snuff, dip, or snus. Data
were weighted and analyzed by sex, WHO region and
World Bank national income categories. Cross-country
dispersion in ST use was assessed with the Coefficient of
variation (CV).
Results: ST use varied widely among countries
(CV0.62); overall prevalence ranged from 1.1% in
Montenegro to 16.4% in Congo. Prevalence by WHO
region was: AFRO (median 7.4%; range5.4% to
16.4%); EMRO (median 5.1%; range1.6% to
12.6%); EURO (median 2.0%; range1.1% to
6.9%); AMRO (median 5.3%; range1.8% to
9.8%); SEAR (median 6.3%; range2.8%, to 9.4%),
and WPRO (median 5.1%; range2.1% to 8.7%).
Prevalence by national income category was: low-income
countries (median 6.2%; range2.5% to 16.4%),
lower-middle-income countries (median 6.2%;
range2.0% to 14.4%), upper-middle-income countries
(median 4.7%; range1.1% to 11.3%), and highincome countries (median 3.4%; range1.8% to
9.8%). Overall ST prevalence exceeded 10% in 5
countries 4 in AFRO and 1 in EMRO regions. Girls
ST use equaled or exceeded that of boys in 10 low and
middle income countries.
Conclusion: While the limited number of countries
assessed may not be representative of each region, these
findings suggest that region-specific patterns of ST use
exist, with higher ST use observed in low and middle
income countries, particularly in the AFRO and SEAR
regions. Evidence-based interventions outlined in the
WHO MPOWER package, such as raising tobacco taxes,
and warning about the dangers of tobacco use, may
prevent and reduce ST use among adolescents.
S307
20. IAP, pneumonia, asthma and chronic

lung disease
OA-451-05 Does Ringland improve air quality
and health in Antwerp? Health impact
assessment of a predicted air pollution change
by covering the Antwerp Ring Raod
D Van Brusselen1 1University of Ghent, Ghent, Belgium.
e-mail: daan.vanbrusselen@hotmail.com

Antwerp, the plans for an expansion of the existing ring
road have initiated widespread resistance over a 10-year
period. This resistance raised awareness about air
quality, gave rise to detailed air quality research in the
city, and stimulated bottom-up development of alternative solutions. The thinktank Ringland, a collective of
urban planners and scientists, advocates covering the
entire ring road. The Ringland plan involves turning the
Ring essentially into a tunnel, in order to reduce noise
and air pollution, creating 384 hectares of open space on
top of the ring, available for development of public parks
and buildings. We conducted a health impact assessment
(HIA) to quantify long term differences in mortality and
cardiovascular/pulmonary outcomes of the predicted
PM2.5 and NO2 reduction in an Immision Frequency
Distribution Model (IFDM) street canyon model of
Ringland.
Intervention or response: The HIA consisted of three
steps: 1) Calculation of an air pollution reduction
scenario, using an IFDM-street canyon model. 2) Review
of the medical literature and selection of the health
outcomes to be evaluated. 3) Calculation of the health
impact.
S308
Results and lessons learnt: A clinically relevant reduction

of PM2.5 and NO2 was calculated within a 1500 meter
buffer of the Ring Road. The calculated reduction of
PM2.5 was associated with an expected annual decrease
of 21 deaths (CI 7 to 41) in the population of Antwerp.
1710 (-570) life years could be gained by Ringland. In
430 schools in the 1500 meter buffer around the Ring
Road a significant improvement of lung function was
calculated among children aged 10 to 18 years by the
reduction of NO2 and PM2.5. Differences in lung cancer
mortality and the incidence of acute myocardial infarction were significant, but low for the calculated air
quality improvement.
Conclusions and key recommendations: Ringland could
significantly improve air quality. The decrease of PM2.5
was associated with a relevant mortality reduction and
an increase in life expectancy. The decrease of NO2 and
PM2.5 was associated with an improved lung function
development in school children, which is linked to long
term respiratory and other health benefits. The impact of
air quality improvement on cardiovascular health has to
be evaluated in combination with noise reduction and the
effect of green space.
OA-452-05 Change over time in peak

expiratory flow among workers exposed to
polyvinyl chloride dust
H Lawin,1 P Ayelo,1 V Hinson,1 J Kagima,2 B Fayomi1
1
University of Abomey Calavi, Cotonou, Benin; 2Egerton
University, Nakuru, Kenya. e-mail: hervelawin@yahoo.fr
Background: Occupational exposure to polyvinyl chloride (PVC) may cause lung disease, including occupational asthma according to anecdotal reports.This study
is to evaluate the variation of PEF over a workweek in
nonsmoking workers and without previous asthma
exposed to PVC dust.
among 42 operators with exposure to PVC dust (filling
hoppers to feed extrusion machines) and 23 employees
without exposure to PVC dust in a plant producing PVC
pipes in West Africa. A pre-tested questionnaire was
administered and PEF was measured using a portable
peak flow meter after a day off (day 1), day 3 and at the
end of the week (day 6).
Results: The two groups did not differ by age or BMI.
Dyspnea was more prevalent in exposed workers (52%)
than controls (13%) (P 0.002). PEF decreased more
significantly in exposed workers than controls (8% vs.
3% on day 3 and 10% vs. 5% on day 6 , both P
0.004).The exposure duration did not affect the PEF
variability in the exposed groups.
Conclusion: The decrease of PEF over the workweek in
workers exposed to PVC dust is consistent with
occupational asthma although standard measures to
diagnose occupational asthma were not used. This result
reinforces the needs to prevent excessive exposure to
PVC dust.
OA-453-05 Asthma control rates in Burundi: a

preliminary study of 183 cases
F Ndikumwenayo1 1Centre Hospitalier Universitaire de
Bujumbura, Bujumbura, Burundi. Fax: (257) 22 23 22 68.
e-mail: ndkmwnyfrancois@yahoo.com
Background: Level of asthma control is not known in

Burundi. A first study was conducted for determining the
level of asthma control and risk factors of asthma
symptoms in adult patients according to GINA classification.
Methods: A prospective, descriptive and preliminary
study was done on asthma patients seen in University
Hospital center of Bujumbura during 12 months. GINA
criteria were used for determining level control of asthma
and risk factors.
Results: It was included 183 asthmatic patients, the
average age was 45.39 6 7.19 years, with 2.3 of sex
ratio. Those patients were not different from age or
gender (p60.10). The beginning of asthma symptoms
was under 50 years of age in 85.39% (p6 0.001),
particularly in women group (p 6 0.001). According comorbidities, allergic rhinitis was found in 60.7%, GERD
in 18%, HIV in 6%, and diabetes mellitus in 4.4%.
Asthma was controlled in 12.6%, partly controlled in
35% and uncontrolled in 52.4%. Factors of the lack of
asthma control were: unconscious therapeutic noncompliance (RR 1.78, 95%CI [1.11 to 2.86]), non-using of
inhalator spacer (RR: 12; 95%CI [6.45 to 22.34]) and
lack of basic treatment (RR 2.26; 95%CI [1.06 to 5.06]),
inadequate patient education and fears about side effect.
Aggravating factors were viral infection, tobacco smoke,
occupational and domestic environment.
Conclusion: Control asthma is a common target.
Therefore, many patients are looking forward to
improving their asthma control. Many actions may be
done for relieving asthma in Burundi. Key words:
Asthma, level, control
OA-454-05 Sequelae following successful

treatment of multidrug-resistant tuberculosis
patients under the national programme in
India
R Singla,1 M Mallick,1 N Singla,1 S Munjal1 1National
India. Fax: (91) 1 126 517 834. e-mail:
drrupaksingla@yahoo.com
Background: India has highest number of multi-drug

resistant tuberculosis (MDR-TB) patients in the world.
However, there is paucity of data about the residual
sequalae at the successful completion of treatment of
MDR-TB patients under national programme. The
current study was aimed to evaluate occurrence and
severity of residual sequalae among MDR-TB patients
following successful completion of their treatment.
Design/Methods: 46 MDR-TB patients were sequentially
enrolled in Delhi, India at the completion of their
successful two years of treatment under the national
programme. They were evaluated clinically; radiologically by X-ray chest; and by complete pulmonary
function tests (PFT) including spirometry, lung volume

and diffusion capacity measurements. The quality of life
(QOL) was evaluated by Seattle Obstructive Lung
Disease Questionnaire (SOLDQ).
Results: 44/46 (95.6%) patients were left with persistent
symptoms with 40/44 (90.9%) among them having
dyspnoea as a prominent symptom. 38/46 (82.6%)
patients had residual advanced lesions on chest X-ray.
Out of 46 patients, 42 could perform PFT. Among these
41/42 (97.6%) patients were left with abnormal pulmonary functions; 19/42 (45.2%) patients had severe
impairment in PFT as per European Respiratory Society
guidelines; 27/42 (64.3%) had mixed obstructive and
restrictive pattern; 4/42 (9.5%) had pure obstructive
pattern; 24/42 (57%) patients had reduced diffusion
capacity. With increase in severity of chest X-ray
abnormalities there was increase in pulmonary function
impairment and decrease in QOL score, especially
related to physical function. Patients had residual
reduced mean SOLDQ scores: Physical function 49%,
Emotional factor 51.6%, Coping skills 60% and
Treatment satisfaction 52.1%.
Conclusion: Under the national programme for the
management of MDR-TB in India, following successful
bacteriological cure of MDR-TB patients, significant
number of patients under study had residual clinical
symptoms, radiological sequalae and impaired lung
functions leading to poor QOL. In addition to achieving
bacteriological cure, there is a need under national
programme to focus on management of post treatment
sequalae and improving quality of life of the successfully
treated MDR-TB patients.
S309
considerably smaller, allowing easy infiltration into the

lungs. Only particles of 1.114 to 2.458 lm aerodynamic
size are therefore presented here
Results: Preliminary results from 3 patients and 2
volunteers are presented. The Figure shows a compelling
association between increases in CO2 concentration and
expired particle load. This relationship was stronger for
smaller particles, becoming random with larger particles
(.2.5 lm) owing to low numbers. Further study subjects
will allow for more refined analysis of factors associated
with this observed relationship.
Conclusion: Preliminary results show evidence to support the utility of carbon dioxide monitoring as a
surrogate for rebreathed air volume. Final results will
open up a new agenda for more detailed investigations
into the infectivity of the air we breathe.
Figure The association between carbon dioxide concentration and expired particles for both untreated, active
PTB patients (TB1 to 3) and TB screen-negative
volunteers (V1 and 2) in a clean room environment.
Only particles in the aerodynamic size range 1.114
2.458 lm are presented here.
OA-455-05 Is monitoring CO2 an acceptable

surrogate for rebreathed air volume
assessment?
C Morrow,1 W Bryden,2 C Call,2 R Wood1 1University of
Cape Town, Cape Town, South Africa, 2ZeteoTech, Paris,
France, e-mail: carl.morrow@hiv-research.org.za
Aim: To determine the validity of using CO2 monitoring

as a surrogate for breathed air. This relationship has been
queried in the context of personal CO2 monitoring and
the development of Rebreathed Air Volume assessment.
Methods: CO2 release and particle accumulation from
untreated patients actively infected with pulmonary
tuberculosis (PTB) and TB screen-negative volunteers
were individually monitored in a 1.45 m3 clean room.
Participants were dressed in Tyvec suits and booties to
reduce nonrespiratory particle sources. After participants
entered the room, the air was cleaned by HEPA filtration
at 300 Air Changes per Hour (ACH). Participants then
sat in the room under conditions of 0.1 ACH and
breathed normally (occasional coughing was allowed as
needed) for the first of 30 minutes or until the
surrounding air reached 4000 parts per million above
ambient CO2 concentration. During this time CO2
concentration and aerodynamic particle size were
monitored at 10 second intervals. Actual size of TB
bacilli is approximately 2 lm, but the aerodynamic size is
OA-456-05 High burden of respiratory

morbidity among individuals treated for
recurrent tuberculosis in Southern Africa
J Metcalfe,1 S Makumbirofa,2 R Penaloza,1 T Vu,1
P Hopewell1 1University of California San Francisco School
of Medicine, San Francisco, CA, USA; 2Biomedical Research
and Training Institute, Harare, Zimbabwe. Fax: (1 415 695
1500. e-mail: john.metcalfe@ucsf.edu
Background: The natural history of patients successfully

treated for tuberculosis (TB) in high HIV-burden
countries is poorly characterized.
Design/Methods: We recruited and followed consecutive
individuals with history of prior TB treatment recruited
from outpatient clinics in Harare, Zimbabwe. At 12-18
months post-treatment completion, we collected detailed
demographic, socioeconomic, respiratory symptom,
HIV, microbiologic, and radiographic data; we performed spirometry, endurance shuttle walk testing
(ESWT), and accelerometry.
Results: From November 2011 to August 2014, 447
individuals were recruited at time of TB treatment
initiation and followed for a median of 1.1 years (IQR
S310
0.98-1.9 years). Median age of the cohort was 37 years

(IQR 29-44), 59% were male, and 70% were HIVpositive; 25% had initial culture-confirmed TB, 19%
had culture-confirmed rifampicin-resistant TB, and 56%
had no microbiologic evidence of TB. At final follow-up,
11% had died and 6% were lost to follow-up. Among
those undergoing respiratory studies (n 183), 61%
reported continued respiratory symptoms (score . 5 on
CAT), 67% had obstruction or restriction on spirometry,
11% desaturated below 90% on ESWT. Mean (SD) for
duration walked during ESWT was 10.4 minutes (SD 2.5
minutes). Overall, 69% had abnormalities on chest
radiograph, with 23% having abnormalities across
25% of lung fields or greater.
Conclusion: Despite successful treatment, considerable
respiratory morbidity exists among TB and DR-TB
patients; in particular among HIV-positive participants
and those without microbiologic confirmation of TB.
were supported by at least one source document. All

digital images of the health passports and CAPS stickers
were clear and legible.
Conclusion: EDC using eCRFs on smartphones is
feasible in rural Malawi. This methodology has facilitated high quality data collection. Capturing digital
images of source documents in the field allows SDV to be
conducted efficiently and remotely without requiring
additional field visits. We recommend these approaches
for future work in similar settings. Acknowledgements
CAPS is funded by the Medical Research Council,
Wellcome Trust and UKAid (Joint Global Health Trials
scheme).
OA-457-05 The Cooking and Pneumonia Study:

successful implementation of electronic data
capture in rural Malawi with source document
verification from the UK
W Weston,1 K Mortimer,1 J Smedley1 1Liverpool School of
Tropical Medicine, Liverpool, UK. e-mail:
williamweston@gmail.com
Background: The Cooking and Pneumonia Study (CAPS)

is a village-level cluster randomised controlled trial
assessing the effectiveness of an advanced cook-stove
intervention in preventing pneumonia in children under 5
in rural Malawi (www.capstudy.org). Each child is
visited at their home every 3 months to collect
information using an electronic Case Report Form
(eCRF) on a smartphone. Here we report our experience
of electronic data capture (EDC) in rural Malawi with
remote Source Document Verification (SDV) (the process
of evaluating the consistency of the data in the eCRF with
the source data in the source documents) in the UK.
Design/Methods: Every 3 months fieldworkers capture
digital images of the health passport (a booklet owned by
the child where health workers make a record of any
health related event) of each CAPS participant (n 9289)
using the camera in the smartphone and enter data from
the health passport (including a study-specific CAPS
sticker shown in the Figure) into an eCRF. If an episode
of pneumonia is recorded in the health passport, the
fieldworkers transcribe the data from the CAPS sticker
(severity, symptoms and signs, and outcome) into the
eCRF. We conducted SDV of pneumonia data (occurrence, severity and 8 clinical indicators) recorded in the
eCRF, the health passport and the CAPS sticker remotely
from the UK. Microsoft Access was used to visualise each
recorded pneumonia episode with the data from the
eCRF and digital images of the health passport and CAPS
sticker.
Results: 585 episodes of pneumonia had been recorded in
the master database by 1/2/2014. Of these, 538 had a
finding of pneumonia in the health passport and 240 in
the CAPS sticker. 547 (94%) episodes of pneumonia
OA-458-05 Obliterative bronchiolitis: a

common cause of chronic lung disease in older
adolescents with vertically acquired HIV
J Rylance,1 S Desai,2 A Nair,2 G Mchugh,3 J Metcalfe,4
H Mujuru,5 K Kranzer,6 R Ferrand3,6 1Liverpool School of
Tropical Medicine, Liverpool, 2Kings College Hospital NHS
Foundation Trust, London, UK; 3Biomedical Research and
Training Institute, Harare, Zimbabwe; 4University of California
San Francisco, San Fransisco, CA, USA; 5University of
Zimbabwe, Harare, Zimbabwe; 6London School of Tropical
Medicine and Hygiene, London, UK. e-mail:
jamie.rylance@lstmed.ac.uk
Background: Vertically HIV-infected children have a

high burden of chronic lung disease that is not wellcharacterised. We prospectively investigated the spectrum of HRCT abnormalities and lung function indices in
HIV-infected children with (non-tuberculous) chronic
respiratory disease.
Methods: HIV infected older children and adolescents
(aged 6-16 years) on antiretroviral therapy for at least 6
months were systematically recruited from an HIV clinic
in Harare, Zimbabwe. Assessment included clinical
examination, spirometry and chest radiography (CXR).
Poster discussion sessions, Saturday, 5 December
Patient with previous Pneumocystis jirovecii pneumonia

and asthma were excluded. Subjects with chronic
symptoms, CXR or spirometric abnormalities were
referred for high resolution CT (HRCT).
Results: Thirty-nine participants (female14 [36%];
median age11.8 [range 6.8 16.4 years] and median
age at HIV diagnosis6.1 [IQR 3.7 8.6] years) had
HRCT, spirometry and clinical assessment. The median
CD4 count was 354 (IQR 200-717) at diagnosis and 641
(IQR 413-775) at enrolment. Thirty patients produced
satisfactory spirometric traces. Ten (26%) patients
desaturated on exercise. Fifteen (38.4%) patients had
been treated for tuberculosis. HRCT was abnormal in 34
(87.2%). Decreased attenuation (consistent with constrictive obliterative bronchiolitis [OB]) was present in
22/39 (56%), strongly associated with reduced FEV1
(mean z-score difference 1.10, P , 0.001) and associated
with chronic cough (but no other symptom). Bronchiectasis (seen in 13/39 (33%) patients) was associated with
breathlessness, cough as well as reduced FEV1. Thinwalled cysts (possibly indicating lymphoid interstitial
pneumonia [LIP] in 1/39 [3%]) and emphysema (10/39
[26%]) were present in in the minority. There was no
linkage between HRCT patterns and presence of hypoxia
and CD4 count at diagnosis / enrolment.
Conclusion: Chronic lung disease in children with
vertically-transmitted HIV appears to be caused by
constrictive obliterative bronchiolitis in over 50%; the
pathogenesis of OB warrants further study. Pulmonary
air cysts and ground glass opacification (consistent with
LIP) are uncommon.

26. TART: TB risk and impact of ART
PC-914-05 Determining factors associated with
occurrence of tuberculosis among adults living
with HIV after initiation of antiretroviral
treatment in Addis Abeba
K Kibret,1 Alem Yalew2 1Wollega University, Nekemte,
2
Addis Ababa University, Addis Ababa, Ethiopia. e-mail:
ktwu27@gmail.com
Background: Tuberculosis (TB) is a leading morbidity

and mortality, and the first presenting sign in majority of
people living with Human Immune deficiency Virus
(PLWH). Determinants of active TB among HIV patients
on anti retroviral treatment (ART) are not well described
in resource limited settings. The aim of this study was to
assess determinant factors for the occurrence of TB
among people living with HIV after ART initiation in
public hospitals and health centers in Addis Ababa,
Ethiopia.
Design/Methods: A case control study was conducted
from December 2011 to February 2012 in 2 public
hospitals and 13 health centers in Addis Ababa. The
study population consisted of 204 cases and 409
controls. Cases were adult people living with HIV who
S311
developed TB after ART initiation and controls were

adult people living with HIV who did not develop TB
after ART initiation. An interviewer administered
structured questionnaire was used to collect information
Results: After adjustment for potential confounders,
presence of isoniazid prophylaxis (adjusted odd ratio
[AOR] 0.35, 95% confidence interval [CI] 0.125, 0.69)
and cotrimoxazole prophylaxis (AOR 0.19; 95%CI:
0.06, 0.62) had protective benefit against risk of TB. In
contrary, bedridden (AOR 9.36; 95%CI: 3.39, 25.85),
having World Health Organization (WHO) clinical stage
III/IV (AOR 3.40; 95%CI: 1.69, 6.87) and hemoglobin
level , 10 mg/dl (AOR 7.43; 95%CI; 3.04, 18.31) at
enrollment to ART care were predictors for increased
risk of tuberculosis in PLWH after ART initiation
Conclusion: Increasing coverage of isoniazid preventive
therapy and cotrimoxazole preventive therapy reduced
risk of TB among HIV patients who started treatment.
All PLWH should be screened for TB, but for patients
who have advanced disease condition (WHO clinical
stage III/IV, being bedridden and having hemoglobin
level , 10 mg/dl) intensified screening is highly
recommended during treatment follow up
PC-915-05 High rate of late incident

tuberculosis among HIV-infected patients
taking long-term antiretroviral therapy in
Western India
A Dravid,1 M Kulkarni,1 C Saraf,2 S Kore,2 P Bhagwat3
1
Ruby Hall Clinic, Pune, 2Precision Diagnostics and
Biosciences, Pune, 3Qua Blu Consultancy Services, Pune,
India. e-mail: ameet.dravid@gmail.com
Background: India has a high tuberculosis (TB) and HIV

burden.As of 2014, 800 000 patients are on antiretroviral therapy(ART) in India.Due to easy access to
ART,India has reported declining HIV incidence rate.However data on incident TB especially late incident
TB(TB developing . 3 months after ART initiation) and
its association with baseline and treatment related risk
factors is sparse.
Methods: In an observational cohort of 1425 HIV
infected patients following up at a private HIV clinic in
Pune, western India between February 2009 and
February 2015 with at least 3 months of follow-up on
ART, TB screening (WHO TB symptom score) was done
at every clinic visit.Sputum smear examination and
radiological investigations were done in case of positive
symptoms.Demographic data, CD4 count and Plasma
HIV viral load was collected at entry and yearly
thereafter. Isoniazid preventive therapy (IPT) was given
to eligible patients.Incidence of late onset TB and 95%
Poisson CIs were estimated. Cases of paradoxical
worsening of baseline TB were excluded. Risk factors
for late onset TB were identified using univariable and
multivariable Poisson regression analysis. Multivariable
model was adjusted for age, sex, baseline TB, baseline
CD4, baseline haemoglobin (Hb) and virologic status
(virologic success/failure).
S312
Results: 1425 HIV patients (34.3% females) were on

ART with median age 40 yrs, median baseline CD4 count
154 cells/mm3 and median (IQR) follow-up on ART
3(1.25-5) years. 36.7% patients had baseline TB. 78%
patients were on Non-nucleoside reverse transcriptase
inhibitor (NNRTI) based ART while 22% were on
protease inhibitor (PI) based ART. 25% patients were
given IPT with ART. Overall incidence of early incident
TB (within 3 months of ART initiation) and late incident
TB(95% CI) was 5.9 episodes/100 person years(4.6-7.1)
and 2.6 episodes/100 person years(1.77-3.41) respectively.126 episodes of late TB were reported among 118
patients with median time to TB(IQR) of 2.38 (1.783.75) years. 13/126(10%) episodes of late TB were due to
unmasking immune reconstitution inflammatory syndrome (IRIS).51.6% patients were virologic failures on
ART at the time of late TB while 48.4% were virologic
successes. 56% of incident cases were confirmed on TB
culture while drug sensitivity testing was performed in
15%. Age, male sex and baseline TB were not associated
with late TB. CD4 count, 200 cells/mm3 (P 0.001)
and baseline Hb, 10 g/dl (P 0.03) were significantly
associated with late incident TB. Virologic failure on
ART increased incidence of late TB (10.99 versus 1.34
episodes/100 person years, P 0.0001). Use of IPT and
ART together were associated with lowest TB incidence
of 0.087 episodes/100 person years. There were 3 deaths
because of late incident TB.
Conclusions: Incidence of late TB in patients on ART
(2.6 episodes/100 person years) is higher compared to
national average of 0.8 episodes/1000 person years in
HIV negative population and calls for measures like
combining IPT with ART in HIV patients without active
TB,HIV viral load monitoring on ART,immediate switch
to virologically suppressive ART on diagnosis of virologic failure and initiation of ART at higher CD4 counts.
program was rolled out in 2004 and new ART treatment

guidelines were introduced in 2013. The TB incidence
rate estimates were calculated using Poisson assumptions. Univariable and multivariable Poisson regression
analysis was done to assess independent risk factors for
Incident TB.
Results: At baseline median age was10 years. A total of
321 children were put on anti TB treatment. Between
2000 and 2014, TB incidence among 6017 HIV infected
children ,18 years was 10.8/1000 person-years (95%CI,
9.7 12.0). Compared to children with WHO Stage I
HIV disease, children with WHO Class Stage IV were at
greater risk for TB (aIRR 6.2, 95%CI: 4.29.2 ;P ,
0.001). TB incidence was also lower for those on ART at
registration vs. those who were not (2.7 and 13.1;
respectively aIRR 0.30, 95%CI: 0.17-0.51; P , 0.001).
At CD4 count of ,250 the TB incidence was 13.9/1000
py with aIRR of 1.47 (CI 1.0 2.12; P , 0.05). KaplanMeier estimates of time to TB show that over 10% of our
pediatric population contracted TB during the study
period. Median time to TB was 3.61 years and IQR was
(2.25-4.96) after registration.
Conclusion: The incidence of TB in children seeking care
at our ART center is higher than other South Asian
countries in a setting of higher background burden of TB
disease. Our study results show that 1 in every 10 HIV
infected child developed active TB disease over time.
Frequent TB screening and preventive therapy should be
considered for those most at risk of developing TB
disease in order to reduce incidence, which for this
population includes those with low CD4 cell counts, not
on ART treatment, and at a late clinical stage classification.
PC-916-05 Incidence and risk factors for

tuberculosis among HIV-infected children in
India
J Kiarie,1 D Mibei,1 B Langat,1 A Ronoh,1 D Nyangahu,1

R Muthee,1 E Mailu1 1Kenyan Ministry of Health, Nairobi,
Kenya. e-mail: jckiarie@gmail.com
A Kinikar,1 S Khadse,1 A Deluca,2 N Gupte,2 D Kadam,1

A Chandanwale,1 A Gupta,3 R Bollinger3 1Byramjee
Jeejeebhoy Government Medical College & Sassoon General
Hospital, Pune, 2Byramjee Jeejeebhoy Government Medical
College-JHU Clinical Trials Unit, Pune, India; 3Johns Hopkins
University, Baltimore, MD, USA. e-mail:
aarti.kinikar63@gmail.com
Background: Tuberculosis is a disease of public health

importance that contributes the high burden in terms of
morbidity and mortality in Kenya. It is established that
HIV/AIDS is the leading driver of Tuberculosis. It has
been noted though that there is a mismatch between the
trends of TB and HIV. HIV alone therefore might not
explain the currents and pattern of TB distribution. The
main objective of this study was to determine other risk
factors for TB disease in the different Counties in Kenya.
Findings from this study will trigger the initiation of
appropriate evidence-based County specific interventions
Design/Methods: Secondary data was used in this study.
Key variables that have been proposed as possible risk
factors for TB from other studies were obtained from
each County and analyzed using regression and ANOVA
models using Ms Excel and SPSS software. There was
clustering of the Counties according to the case
notification level into two categories; high and low
which was then used to compare predictors based on the
Background: The risk for TB among HIV infected

children may vary based on local community burden of
TB and their access to HIV treatment. India has the
largest burden of TB and a national program for HIV
treatment of children. Defining the incidence and risk
factors of TB among HIV-infected Indian children will
help optimize programmatic guidelines and objectives to
reduce TB-associated morbidity and mortality.
Design/Methods: This was a retrospective observational
study, with data drawn from a large cohort of HIV
infected children seeking care at an ART center from
January 2000 to December 2014 at Sassoon General
Hospital in Pune, India. Free ART under the National
PC-917-05 What drives the TB epidemic in

Kenya: county TB profile
two categories. The percentage of urban population in a

county was use to measure urbanization and area per
facility in square kilometers was used to estimate facility
accessibility. A multiple linear regression was then used
to identify factors that could be associated with CNR
variations in different counties
Results: Counties in Kenya have significant variation in
HIV prevalence and TB CNR. Results showed that in
addition to HIV prevalence (P 0.00), Urbanization (P
0.007), HDI, Population density (P 0.041) and Poverty
(P 0.003) were significant predictors. Facility accessibility (P 0.008) was significantly associated with TB
CNR on its own but it was not found to be a predictor.
Poverty was found to be negatively associated with TB
CNR. Population per health workforce, number of TB
facilities, literacy level and fully immunized level were
not significantly associated with TB CNR. The variation
of HIV prevalence, the various factors and TB CNR in
the various Counties shows that different factors are Key
in the different Counties
Conclusion: There is need to align efforts in the fight
against TB to all risk factors if the fight against TB is to
be won in Kenya and further dissect the approaches
according to risks by county. Also, further research needs
to be done to determine the risk factors for TB in
Counties with high poverty levels.
PC-918-05 Epidemiology of tuberculosis among

HIV-positives in Nepal
S Sah,1 G Bhatta,2 S Verma3 1National Tuberculosis Center,
Thimi, Bhaktapur, 2RESEARCH, SAARC Tuberculosis and HIV/
AIDS Center, Thimi, Bhaktapur, 3SAARC Tuberculosis Center,
Thimi, Bhaktapur, Nepal. Fax: (977) 1 663 0706. e-mail:
sujitmph12@gmail.com
Background: Globally, tuberculosis (TB) is the most

common opportunistic infection and leading cause of
mortality in people living with the human immunodeficiency virus (PLHIV), contributing 1 in 5 of these deaths.
Tuberculosis is a public health problem of importance in
Nepal with an estimated incidence of 156 per 100 000
population. The estimated prevalence of HIV among
people aged 15-49 years was 0.23% in 2013. The HIV
epidemic is concentrated to key populations. People
living with HIV are 29 times more likely to develop TB
disease as people without HIV. TB-HIV activities had
been implemented in 43 districts out of total 75 districts
in Nepal. However, the proportion of TB patients in
Nepal who had an HIV test result recorded in the TB
register dropped from 25% in 2012 to 11% in 2013.
Therefore, this study aims to find out TB epidemiology
among People Living with HIV (PLHIV) in Nepal.
Design/Methods: This is the first an institutional based
prospective study. This survey was conducted at 6 sites
representing all five regions of Nepal from July 2012 to
February 2013. The calculated sample size was 400
PLHIV registered in Pre/ART centers.The statistical test
analysis was done in accordance with the distribution of
data.
Results: The study reveals that 11.5% prevalence of TB
was found in total examined HIV clients. Highest
S313
prevalence rate of TB among HIV clients was found in

Far-West Region (28.8%) and Eastern Region (23.1%)
where migration and refugee problem is prevailing. The
prevalence of TB was found higher in age group of 35-65
years and above. The prevalence of TB among enrolled
HIV clients was comparatively high in male (14.4%)
than in females (8.4%). TB prevalence was also found
higher in illiterate clients in comparison to higher
educated. Prevalence of TB among migrants was higher
(14.6%) followed by spouse of migrants (11.5%) and
IDUs (11.1%). The prevalence of TB was found 10.7
percent among HIV clients enrolled for ART/PART. The
study reveals that 5.4%, 3.1% and 3.1% were pulmonary positive; pulmonary negative and extra pulmonary
TB respectively.
Conclusion: The prevalence of TB was found significantly higher in age group of 35-65 years, high risk and
illiterate groups of HIV clients. Therefore, strategy
should be revised to focus on targeted interventions for
migrants, spouse of migrants and IDUs and illiterate
community people to avoid TB co-infection in Nepal.
PC-919-05 Antiretroviral therapy scale-up for

TB reduction in rural Swaziland: how far will it
take us?
B Kerschberger,1 I Zabsonre,1 G Mchunu,2 M M Win,1
I Ciglenecki,3 L Rusike-Pasipamire,1 A Telnov,3
`
W Sikhondze2 1Medecins
Sans Frontieres
(Operational
Centre Geneva), Mbabane, 2National TB Control Programme,
Sans
Ministry of Health, Manzini, Swaziland; 3Medecins
`
Frontieres
(Operational Centre Geneva), Geneva,
Switzerland. e-mail: benadi11@yahoo.de
Background: Timely antiretroviral therapy (ART) reduces the TB burden in people living with HIV (PLHIV). In
Swaziland, integration of TB-HIV collaborative activities
has been strengthened since 2009 resulting in a decline in
TB notifications since 2010. Modelling studies, however,
predict a rebound in population TB incidence in high
HIV prevalence settings when decline in HIV incidence is
delayed and immune-virological responses are not
sustained. It remains uncertain whether and when this
rebound effect will occur under routine program
conditions.
Design/Methods: We triangulated TB and HIV programming data from the rural Shiselweni region (population
210 000) in Swaziland, from 2009-2014. Overall and
age-specific first line TB drugs notifications rates were
described to elaborate signs of TB case rebound. Annual
population ART coverage among PLHIV and population
viral load (VL) suppression rates (defined as estimated
population VL ,1,000 copies/ml among PLHIV) were
calculated for the same time periods. Denominators for
coverage and notification estimates were obtained from
projections of the 2007 population census.
Results: Overall annual TB notifications declined steadily from 1304 to 424/100 000 between 2009- 2013, and
were 408/100 000 in 2014. The annual decline in TB
notifications ranged from 174 to 298/100 000
between 2009-2013 (P , 0.01), and was reduced to
16/100 000 between 2013-2014 (P 0.58). TB
S314
notifications declined in all age-groups from 2009-2014,

but increased in the 20-29 (516 to 577/100 000; P 0.07)
and in the 30-39 (1029 to 1185/100 000; P , 0.01) yrs
old between 2013-2014. ART coverage in PLHIV
increased from 19% in 2009 to 47% in 2013, and was
53% in 2014. It was lowest for the 30-39 yrs old (45%)
and above average in the 20-29 yrs old (63%). For the
same time periods, estimated population level VL
suppression increased from 24% (2009) to 49% (2013)
and was 55% in 2014.
Conclusion: Rapid ART expansion and increasing
population level virologically response may have contributed to a steady decline in TB notifications. Yet an
overall flattening of annual TB notifications was
recorded at approximate 50% ART coverage, and
annual TB notifications started to increase in the 20-39
yrs old for the first time since 2009. Yet, it remains
uncertain whether this constitutes the beginning of
rebound in TB notifications. Modelling studies suggest
universal ART expansion strategies (treatment as prevention) for TB control in high HIV prevalence settings
to avoid TB incidence rebound.
PC-920-05 Seasonal variations in tuberculosis

diagnosis in antiretroviral treatment
programmes in South Africa, Zambia and
Zimbabwe
M Ballif,1 S Reid,2,6 M Zwahlen,1 K Stinson,3,8 L Fenner,1,4
M Fox,5 H Prozesky,7 C Chimbetete,9 M Egger1 1University
of Bern, Bern, Switzerland; 2University of North Carolina at
Chapel Hill & Centre for Infectious Disease Research in
Zambia, Chapel Hill, NC, USA, 3University of Cape Town &
`
Medecins
Sans Frontieres,
Khayelitsha, Cape Town, South
Africa; 4University of Basel, Basel, Switzerland; 5Boston
University, Boston, MA, USA; 6Centre for Infectious Disease
Research in Zambia, Lusaka, Zambia; 7University of
Stellenbosch & Tygerberg Academic Hospital, Cape Town,
8
`
Medecins
Sans Frontieres,
Khayelitsha, South Africa;
9
Newlands Clinic, Harare, Zimbabwe. e-mail:
mballif@ispm.unibe.ch
Background: Variations in tuberculosis (TB) diagnosis

across the year have been attributed to seasonal effects,
winter being associated with indoor crowding, poorer
access to health care, and vitamin D deficiency due to
reduced sun exposure. We investigated seasonal trends in
pulmonary TB (PTB) diagnosis at five antiretroviral
treatment (ART) programs from three countries in
Southern Africa, with different climates depending on
their latitude.
Design/Methods: We included HIV-infected adults (.15
years) enrolled between January 2004 and December
2012 in five ART programs of the International
epidemiological Database to Evaluate AIDS (IeDEA)
consortium: Khayelitsha, Thembalethu, and Tygerberg
(South Africa), CIDRZ (Zambia), Newlands (Zimbabwe). We analyzed crude monthly numbers of PTB
diagnosis and of patients starting ART. As control, we
used Kaposi sarcoma (KS) diagnosis, for which no
seasonal effect was expected. We accounted for trends
over the study period by calculating deviations of
monthly counts to yearly averages. We assessed associations between monthly deviations in number of ART
enrollments, PTB diagnosis and KS diagnosis by calculating correlation coefficients.
Results: Average monthly PTB diagnosis counts were 42
(range 3-76) in South Africa, 308 (271-354) in Zambia
and 6 (1-9) in Zimbabwe. Monthly variations in PTB
diagnosis were seen at all sites (Figure). The strongest
fluctuations were seen in Zimbabwe with PTB diagnosis
peaks in October and troughs in December (52% and
76% of monthly average, respectively). Sites in South
Africa also showed marked PTB diagnosis drops in
December (up to 54% of monthly average). The largest
site, CIDRZ, showed milder PTB diagnosis variations.
Fluctuations in PTB diagnosis paralleled changes in ART
enrollment, with recurrent patterns across 2004-2012.
Correlations between deviations of monthly PTB diagnosis and ART start counts from yearly averages
confirmed this parallelism (coefficient range: 0.69-0.95,
P , 0.0001 for all sites). Numbers of KS diagnosis were
low at all sites, but with marked December drops, as
observed for ART enrollment and PTB diagnosis.
Conclusion: Monthly variations in PTB diagnosis at ART
programs in Southern Africa were observed regardless of
climate, likely reflecting fluctuations in ART enrollment
rather than seasonal effects. This was further supported
by the similar patterns of variations seen for KS
diagnosis. In South Africa and Zimbabwe, pronounced
PTB diagnosis drops in December may be associated with
reduced clinical activities during summer vacations.
PC-921-05 Incidence and predictors of

tuberculosis among HIV-infected adults after
initiation of antiretroviral therapy, Nigeria,
20042012
I Pathmanathan,1 E K Dokubo,1 S Ray,1 D Onotu,2
S Odafe,2 I Dalhatu,2 H Debem,2 S Agolory1 1U.S. Centers
for Disease Control and Prevention, Atlanta, GA, USA, 2U.S.
Centers for Disease Control and Prevention, Abuja, Nigeria.
e-mail: ydi6@cdc.gov
Background: Nigeria had the most AIDS-related deaths

worldwide in 2013 (210 000); 39% were associated with
tuberculosis (TB). The World Health Organization
(WHO) recommends intensified TB case-finding and
treatment before treating HIV with antiretroviral therapy (ART) to prevent TB-associated morbidity and
mortality, yet TB diagnosis in immunocompromised
patients is challenging. We aimed to estimate incidence
and characterize factors associated with TB after ART
initiation in Nigeria.
Design/Methods: We analyzed retrospective cohort data
from a nationally representative sample of adult (715
years) ART patients. Data were abstracted from 3496
patient records at 35 sites selected using probabilityproportional-to-size sampling. Analyses were weighted
and controlled for complex survey design. Subgroup
domain analyses were performed on patients without
baseline TB disease. A Cox proportional hazard model
was used to assess factors associated with TB incidence.
Results: At ART initiation, 3350 patients (95.8%,

95%CI: 94.6%96.9%) were not receiving TB treatment; of these, mean age was 34.9, 66% were female,
and median CD4 count was 161/mL. TB incidence was
0.48 per 100 person-years (56 cases), significantly higher
for patients with baseline CD4 ,50/mL (adjusted hazard
ratio [AHR]: 4.9, 95% CI: 1.5-15.8) compared with
CD4 .200/mL, and marginally higher for patients
bedridden at baseline (AHR: 4.8, 95% CI: 0.924.7)
compared with being asymptomatic. No significant
associations were observed for age, weight, employment,
marital status, education, or WHO HIV clinical stage.
Conclusion: Incidence of TB after ART initiation was
associated with markers of advanced HIV infection such
as low CD4 and poor functional status. Study results
reinforce the benefit of early ART initiation, and
highlight the need for intensified TB case-finding in
patients initiating ART at advanced stages of HIV
disease.
PC-922-05 The impact of AIDS treatment on

tuberculosis outcomes at the national level in
South Africa
Z Mclaren,1 A Sharp,1 E Brouwer,1 A Nanoo2 1University of
Michigan, Ann Arbor, MI, USA; 2National Health Laboratory
Service, Johannesburg, South Africa. e-mail:
zmclaren@umich.edu
Background: Fueled by the HIV/AIDS epidemic, TB has

been the leading natural cause of death in South Africa
for over a decade. In 2004, the South African government
initiated the national rollout of anti-retroviral therapy
for AIDS that resulted in a population-level life
expectancy increase of 11.3 years. The ART rollout
provides a unique and unprecedented opportunity to
examine the population-level impact of ART on TBrelated outcomes. This study performs a longitudinal
analysis that follows the evolution of outcomes before
and after the introduction of ART. This is the first study
to capture national-level spillover effects of ART on TB
incidence for those who were not accessing ART,
including HIV-negative individuals.
Design/Methods: We use data from the National Health
Laboratory Service database on tuberculosis tests performed on patients in public health facilities for the
period January 2004 - December 2011, which includes
over 30 million test records from over 10 million
patients. Culture-confirmed TB-positive cases are based
on the presence of at least one positive result from TB
culture or molecular testing with Xpertw MTB/RIF or
line probe assays. ART rollout data come from the
Department of Health. Regressions include facility-level
fixed effects to control for time-invariant confounders.
Results: The overall number patients per facility rise in a
linear trend, with an uptick around the time of ART
accreditation at the facility. We find an increase in the
number of TB tests occuring more than 90 days from the
initial testing visit, which is consistent with retaining
patients in TB care for longer. There is an overall increase
in TB cases in the period before ART, but it levels off at
S315
the time ART becomes available locally. The leveling-off

of TB cases occurs despite the concurrent shift away from
smear testing, which reduces the underdetection of HIVpositive TB cases. Our results are consistent with
patterns of increased testing and retention of HIVpositive patients when ART becomes available.
Conclusion: Our results capture the important spillover
benefit of ART on TB prevalence, which reduces
morbidity and mortality for HIV-positive and HIVnegative South Africans alike. Improved integration
between HIV and TB health programs should take
spillovers into account to improve the design and costeffectiveness of health policy.
PC-923-05 Characterizing non-IRIS TB

treatment among cART patients in Zambia
P Katayamoyo,1 N Chishinga,1 J Mukundu,1 L Nyirenda,1
P Kasonde,1 A Mwango,1,2 M Welsh1 1FHI360, Lusaka,
2
Zambia Ministry of Health, Lusaka, Zambia. Fax: (26) 021
125 7329. e-mail: pkatayamoyo@fhi360.org
Background: The development of TB among patients

receiving combination antiretroviral therapy (cART) is
fairly well documented. TB disease among patients on
cART may be a result of poor ART adherence or HIV
drug resistance that renders HIV-infected patients
susceptible to TB, but generally literature may not be
very clear on why this is so. This study aims to better
understand and hopefully support effective interventions
for TB treatment among patients receiving cART.
Design/Methods: We conducted a retrospective crosssectional study in 12 high volume ART sites in the
northern parts of Zambia between October, 2011 and
September, 2012. We considered patients aged 15 years
and older that developed TB after 3 months of starting
cART and excluded those with immune reconstitution
inflammatory syndrome (IRIS) TB. Data were extracted
from patient medical charts, TB cards, TB and ART
registers that included age, sex, baseline CD4 count and
type of TB (pulmonary or extra pulmonary TB). Linear
regression was used to determine the association between
age of patients and the timing of TB treatment initiation
after 3 months of receiving cART.
S316
Results: There were 177 patients on cART for more than

3 months that were started on TB treatment. 96 (54.2%)
were males and 153 (86.4%) had pulmonary TB. The
median age was 38.4 years [interquartile range (IQR):
33.446.1]. The median baseline CD4 count at start of
cART was 236 cells/lL (IQR: 147326) and the median
duration to starting non IRIS TB treatment was 23.8
months (IQR: 10.645.9). We found that for such
patients on cART for more than 3 months, every
additional year in age showed duration to starting TB
treatment increased by an average of 0.42 months (Coeff.
0.42, P 0.022) (Figure).
Conclusion: Our results suggest that younger patients on
cART were more likely to be at early risk of TB disease
and starting treatment compared to older patients.
Prospective research is recommended to better understand the implications of these findings on the control of
TB disease among younger HIV-TB co-infected patients
receiving cART.
27. Drug resistant TB: outcomes

PC-925-05 Shortening length of treatment in
MDR-TB patients using a six-drug antituberculosis regimen
Results: 1) In 251 MDR-TB patients, the rate of acquired

drug resistance was more than 86%, the numbers of drug
resistance were more than 3.6, the course was more than
3.1years, the BMI was less than 20.6, the extent of
pulmonary lesion was more than 3.5 lung fields, on
average. 2) In three groups of the MDR-TB patients, the
sputum culture conversion were 74%, 64% and 81% in
the end of intensive phase and were 78%, 79% and 75%
in the end of continuation phase respectively (P . 0.05).
3) In the end of continuation phase, the pulmonary lesion
were improved and/or absorbed in 74%, 79% and 78%
of MDR-TB patients of three groups, respectively (P .
0.05). 4) There were 68 cases occurring adverse reaction
during the treatment, of which the rate of liver injury
were highest among three groups and respectively 8%,
13% and 9% (P . 0.05). 5) In three groups, the
successful rate were 60%, 62% and 58%, mortality rate
were 6%, 8% and 3%, failed rate were 19%, 14% and
17%, respectively (P . 0.05).
Conclusion: The effect and outcome of three treatment
regimens were similar. It is suggested that two regimens
composed of 4core drugs and 2 companion drugs can
shorten the length of treatment to 18 months for MDRTB patients.
PC-926-05 Residual efficacy of levofloxacin

versus moxifloxacin against fluoroquinoloneresistant M. tuberculosis murine infection
Q Li,1 X Jiang,1 K Yang,2 S Tang,1 J Q Liang,3 K Kan,4 L Li,1

l H Qiu5 1Beijing Chest Hospital, Capital Medical University,
Beijing, 2Hunan Tuberculosis Hospital, Changsha, 3The 309th
Hospital of PLA, Beijing, 4Anhui Chest Hospital, Hefei,
5
Shandong Chest Hospital, Jinan, China. e-mail:
lq0703@hotmail.com
T Maitre,1,2 A Chauffour,1,2 C Bernard,1,2,3 A Aubry,1,2,3

Universite Pierre et
N Veziris1,2,3 1Sorbonne Universites,
Marie Curie, Paris 6, Paris, 2Institut National de la Sante et de
la Recherche Medicale,
Paris, 3Assistance Publique-Hopitaux
de Paris, Paris, France. Fax: (33) 1 4216 2127. e-mail:
nicolas.veziris@upmc.fr
Background: Although the course of 24 month of

treatment can ensure the treatment success for the
MDR-TB patients, longer treatment can bring the
damage to the body and the economic burden in the
MDR-TB patients so that their compliance is low and
loss rate is high. The aim of research is to observe the
effect and outcome of treatment using the regimens
which was composed of six anti-tuberculosis drugs and
shortened the length of treatment to 18 months in the
MDR-TB patients.
Methods: The 251 MDR-TB patients were diagnosed
and included in this study respectively in ten hospitals of
China between July 2009 and July 2010. We formulated
three regimens composed of 3~4 core drugs and 2
companion drugs as follows: 6ZAmMfxXY/
18ZMfxXY(the group 1), 6RfbZAmMfxXY/
1 2 R f b Z M f x X Y ( t h e g r o u p 2 A ) a n d
6ClrZAmMfxXY/12ClrZMfxXY(the group 2B), of
which XY were two companion drugs selected as DST of
the MDR-TB patients. 251 MDR-TB patients were
randomly divided into three groups to be treated
respectively using three regimens above-mentioned. The
bacteriology, chest radiograph, blood, urine and so on
were tested for the MDR-TB patients during and the end
of course so as to evaluate the effect and outcome of
treatment.
Background: Preliminary works, in a murine model of

tuberculosis, have shown that moxifloxacin retains
partial activity against fluoroquinolone resistant mutants. Levofloxacin has proven to be equivalent to
moxifloxacin against drug susceptible tuberculosis and
to have a better safety profile. In particular, levofloxacin
prolongs less QT interval than moxifloxacin which
makes it the preferred fluoroquinolone for combination
with new drugs such as bedaquiline or delamanid. Our
obective is to compare the in vivo activity of levofloxacin
and moxifloxacin against M. tuberculosis strains harboring DNA gyrase mutations with various levels of
fluoroquinolone resistance.
Design/Methods: We selected in vivo isogenic fluoroquinolone resistant mutants of M. tuberculosis H37Rv. A
total of 120, 5 weeks old, BALB/c female mice were
intravenously infected with 106 M. tuberculosis GyrA
(A90V) or GyrB (E540A and A543V) mutants. Treatment for one month started one day after infection with
66 mg/kg moxifloxacin or 50 mg/kg levofloxacin given
orally every 6 hours, 5 days-a-week. Antibiotic dosings
were chosen in order to obtain pharmacokinetic parameters equipotent to 800 mg/day of moxifloxacin and
1000 mg/day of levofloxacin in humans.
Results: Moxifloxacin prevented mortality of mice
infected with E540A, A543V GyrB mutants (MIC 0.5
mcg/ml) and GyrA A90V mutant (MIC 2 mcg/ml).

Levofloxacin prevented mortality of mice infected with
E540A GyrB mutant (MIC 0.5 mcg/ml), reduced
mortality of mice infected with GyrB A543V mutant
(MIC 1 mcg/ml) but not of mice infected with GyrA
A90V mutant (MIC 4 mcg/ml). CFU are pending.
Conclusion: High-dose levofloxacin is less active than
high dose moxifloxacin against M. tuberculosis fluoroquinolone resistant mutants. Moxifloxacin should be
preferred against XDR tuberculosis.
S317
treatment and nearly half of them amplified to R

resistance. Appropriate DST guided treatment regimen
of mono/poly resistant TB cases and daily FDC based
regimen for drug sensitive TB cases need to be
systematically introduced under RNTCP with capacity
building and resource allocations to generate evidence on
improvement in treatment outcomes for all forms of TB
and averting amplification of DR TB. This will have a
crucial role in achieving the goals of End TB Strategy.
PC-927-05 Amplification to rifampicin

resistance by not addressing isoniazid mono/
polyresistance under Indias MDR-TB
programme
M Parmar,1 K S Sachdeva,2 S Balakrishnan,1
A Amar Shah,1,2 M Ghedia,1,2 F Sayyed Imran,1,2
R Ramachandran,1 A Sreenivas1 1World Health
Organization Country Office for India, New Delhi, 2Central TB
Division, Ministry of Health and Family Welfare, New Delhi,
India. e-mail: parmarm@searo.who.int
Background: Globally, India has worlds highest burden

of multidrug-resistant tuberculosis (MDR-TB) with
~3% in new, 12-17% in previously-treated TB cases.
Among the later, pooled results of various sub-national
DRS showed that other than R, any resistance to H (6080%) was invariably associated with any poly-resistance
to first line drugs (SH-23%, SHE-6%, HE-2%). Rapid
scale-up of Programmatic Management of Drug resistant
Tuberculosis (PMDT) continues in India with priority
accorded to diagnosis and management of MDR-TB and
XDR-TB. Early cohorts (2007-2010) of patients failing
first line treatment and offered phenotypic DST (HRES)
in 19 labs and LPA (HR) from 2009, revealed ~20%
patients with mono/poly resistance to first line drugs
other than R and were treated with standard first line
intermittent regimen without modification under Indias
Revised National TB Control Programme (RNTCP). We
studied its impact on treatment outcomes and resulting
amplification of drug resistance.
Methods: Retrospective cohort analysis of treatment
outcomes among TB cases with mono/poly resistance to
first line drugs other than R, registered under RNTCP,
treated with standard programme regimen and results of
subsequent DST in cases who failed treatment, was done.
Data on initial DST patterns, TB treatment outcomes and
repeat DST in failures of standarf programme regimen
were extracted from programme records. Data was
cleaned and validated before analysis.
Results: 8848 TB patients with mono/poly resistance to
first line drugs were identified from lab records, whose
treatment outcomes on standard programme regimen
and subsequent DSTs could be traced. Overall failure
rate to standard first line regimen ranged from 24-67%,
of whom 35-64% amplified to R resistance on repeat
DST with not much variation among new and previously
treated patients. (Table 1).
Conclusion: Half of the patients with mono/poly drug
resistance to first line drugs other than R, treated with
standard first line intermittent regimen underfailed
PC-928-05 High efficacy of the Peruvian XDR-TB

regimen for culture conversion in
programmatic conditions
V Alarcon,1,2 D Vargas,3 J Cabrera,1,2 D Vela Trejo,1,2
R Espinoza,1,2 J Diaz,1,2 J Cornejo,1,2 A Mendoza,1
A Valentina Antonieta,1 A Mendoza-Ticona2 1Estrategia
y Control de la Tuberculosis,
Sanitaria Nacional de Prevencion
Lima, 2Peruvian National Tuberculosis Programme, Lima,
3
Hospital Nacional Hipolito

Unanue, Lima, Peru. e-mail:
mendozalberto@hotmail.com
Background: XDR-TB cases have a high rate of fatality.

Peru reports around 90 cases every year. In 2012,
according to WHO guidelines, Peru began to treat
XDR-TB with WHOs fifth group drugs: linezolid,
imipenem/clavulanic acid completed by thioridazine
and susceptible drugs according to a proportion of DST
assay and moxifloxacin and capreomycin. Objective: To
evaluate the efficacy of XDR-TB regimen for culture
conversion in XDR-patients in their first six months of
therapy.
Methods: Between May 2012 and December 2014, 56
XDR-TB patients (older than 15 years of age, culturepositive at diagnosis, with at least six months of
treatment with monthly culture results) were evaluated
through operational information. The average time and
the proportion of culture conversion during the first six
months were determined.
Results: Out of 56 patients, 52% were male; the age
average was 34 years. The average for culture conversion
was 2.1 6 1.2 months. At the first month, 41% patients
converted their culture and 100% during the first five
months, (Figure).
Conclusion: The regimen based on linezolid, carbapenem/clavulanic acid and thioridazine was highly effective
for culture conversion in XDR-TB patients in Peru
during their first six months of treatment. The introduction of new drugs such as bedaquiline must be evaluated
according relation between security and efficacy.
S318
PC-929-05 Isoniazid, even at low dose, exerts

some anti-tuberculosis activity against
isoniazid-resistant Mycobacterium
tuberculosis in mice
R Swanson,1,2 N Ammerman,1,3 C Moodley,1
A Dorasamy,1 A Tapley,1,4,5 J Grosset,1,3 C Rodrigues,6
D Almeida1,3 1KwaZulu-Natal Research Institute for
Tuberculosis and HIV, Durban, 2University of KwaZulu-Natal,
Durban, South Africa; 3Johns Hopkins University School of
Medicine, Baltimore, MD, 4Howard Hughes Medical Institute,
Chevy Chase, MD, 5University of California San Francisco
School of Medicine, San Francisco, California, USA; 6P.D.
Hinduja National Hospital, Mumbai, India. e-mail:
nicole.ammerman@gmail.com
Background: Multidrug resistant tuberculosis (MDRTB) is defined as TB due to Mycobacterium tuberculosis

that is resistant to both isoniazid (INH) and rifampicin,
key first-line drugs in TB treatment. Resistance to INH
can occur at relatively low levels such that the minimum
inhibitory concentration (MIC) for M. tuberculosis is
below the steady-state plasma concentration of INH
after a high dose (15-20 mg/kg daily compared to the
standard dosing of 5 mg/kg/day). Thus, high-dose INH is
often included as part of a second-line regimen for the
treatment of MDR-TB, despite that its effectiveness has
not been clearly demonstrated either clinically or
experimentally. The objective of this work was to
evaluate the impact of low-, standard-, and high-dose
INH in a mouse model of INH-resistant TB.
Design/Methods: BALB/c mice were infected with the
following M. tuberculosis strains: wild-type, INHsusceptible H37Rv (MIC: 0.03 lg/ml); INH-resistant
(low-level) H37Rv with inhA promoter mutation (MIC:
0.25-0.5 lg/ml); INH- resistant (high level) H37Rv with
katG mutation (MIC: 8 lg/ml); and INH-resistant
(intermediate level) LH4, a clinical isolate with the KatG
S315T mutation (MIC: 2 lg/ml). Mice were treated for 4
weeks with one of the following regimens: (i) no drug
control; (ii) INH at the standard therapeutic dose of 10
mg/kg (comparable to the 5 mg/kg dose in humans); (iii)
INH at 5 mg/kg (half of standard dose); and (iv) INH at
20 mg/kg (double standard dose). Bacterial colony counts
in the lungs were monitored throughout treatment.
Results: All untreated mice infected with any strain of M.
tuberculosis died within 1-2 weeks after treatment
initiation. In mice infected with the INH-susceptible
strain, INH at any dose had activity: static at 5mg/kg and

cidal at 10 or 20 mg/kg. In mice infected with INHresistant strains, INH at any dose prevented death during
the 4 weeks of the study. Regardless of the level of INH
resistance, INH at 5 or 10 mg/kg exhibited bacteriostatic
activity and INH at 20 mg/kg exhibited limited
bactericidal activity.
Conclusion: INH administered at low-, standard- or
high-dose showed some benefit to mice infected with
INH-resistant strains of M. tuberculosis. Based on this
experiment, there appears to be some merit to administering INH, not necessarily at high-dose, in the treatment
of MDR-TB, although further studies are needed to
evaluate the activity of INH as part of a second-line
regimen for MDR-TB treatment.
PC-930-05 Two-year survival in patients with

XDR-TB from two provinces in South Africa
C Kvasnovsky,1 M Van Der Walt2 1University of Maryland,
Baltimore, MD, USA; 2Medical Research Council of South
Africa, Pretoria, South Africa. e-mail: ckvasno@gmail.com
Background: When extensive drug-resistant tuberculosis

(XDR TB) was first described in 2006, outcomes were
typified by high mortality. More favorable outcomes
have been achieved, but mostly in settings with few HIV
infected patients. We have a longitudinal dataset of XDR
TB patients from the Eastern Cape (EC) and KwaZuluNatal (KZN). The aim of this analysis is to determine 2year survival and factors predictive of successful outcomes
Design/Methods: A retrospective cohort study of patients initiating treatment for XDR TB in these provinces
from 2006-2008. HIV infected patients had access to
ARVs. Data were collected from patient records. Patients
with at least 2 consecutive negative sputum had achieved
culture conversion. HIV patients not yet on ARVs. At
onset of treatment were initiated on ARVs. At the end of
the fourth month and defined as not on ARVs. Effective
treatments were when a patient received at least 4 drugs
to which their TB could be considered susceptible.
Results: 355 patients initiated XDR TB treatment, of
which 62% were HIV-infected at treatment start, among
which 59% had access to ARVs at onset. Many
demographic and baseline clinical factors differed
significantly between HIV-negative patients, HIV-positive on ART and HIV-positive not on ART. 50% of
patients were smear-positive at onset, but significantly
more so of HIV-positives on ART (P 0.02). Patients
were resistant to a median of 5 (range 4-6) drugs, and
HIV uninfected patients significantly less (P 0.005).
31% patients received effective treatment, among which
the HIV-positives on ART significantly more (P 0.008).
After two years of treatment, 22% had achieved culture
conversion, 27% patients were alive and still receiving
treatment, and 10% had a favorable outcome. On
multivariate analysis, predictors of favorable treatment
outcomes were smear negative status at treatment start
(hazard ratio [HR] 2.92, P 0.004) and weight .50kg
(HR 4.11, P , 0.001). HIV-negative patients had similar
outcomes as HIV-positive patients on ARVs. HIVpositive patients on ARVs were more likely to have
favorable treatment outcomes than those patients not on
ARVs. (HR 2.94, P 0.05).
Conclusions: Even though many patients achieve culture
conversion, 2 year survival and favorable outcomes are
low. Early identification of and suitable interventions are
needed for patients failing treatment to result in
improved outcomes.
PC-931-05 Polymorphisms in the vitamin D

receptor gene reduce the rate of sputum
culture conversion in multidrug-resistant
tuberculosis patients in South Africa
M Magee,1 Y Sun,2 S Shah,2,3 S Allana,2 J Brust,4 Q Hui,2
T Mthiyane,5 N Gandhi2 1Division of Epidemiology and
Biostatistics, School of Public Health, Georgia State
University, Atlanta, GA, 2Department of Epidemiology, Rollins
School of Public Health, Emory University, Atlanta, GA, 3U.S.
4
Divisions of General Internal Medicine and Infectious
Diseases, Albert Einstein College of Medicine, Bronx, NY,
USA; 5University of KwaZulu-Natal, Durban, South Africa.
e-mail: mjmagee@gsu.edu
Background: Vitamin D is an anti-inflammatory regulator of the T-helper 1 response to tuberculosis (TB) and
also mediates the induction of the antimicrobial peptide
cathelicidin. Deficiency of 25-hydroxyvitamin D and
single nucleotide polymorphisms (SNPs) in the vitamin D
receptor (VDR) gene may increase the risk of TB disease
and decrease culture conversion rates in drug susceptible
TB. Whether these VDR SNPs are found in African
populations or impact multidrug-resistant (MDR) TB
treatment has not been determined. We aimed to
determine if SNPs in the VDR gene were associated with
sputum culture conversion among a cohort of MDR TB
patients in South Africa.
Design/Methods: We conducted a prospective cohort
study of adult MDR TB patients receiving second-line TB
treatment in KwaZulu-Natal province. Subjects had
monthly sputum cultures performed. In a subset of
participants, whole blood samples were obtained for
genomic analyses. Genomic DNA was extracted and
genotyped with Affymetrix Axiom Pan-African Array.
Standard quality control procedures were applied including sex-mismatch, 95% call rate, Hardy-Weinberg
Equilibrium, outlier of population structure, and minor
allele frequencies higher than 5%. Generalized linear
Poisson and Cox proportional models were used to
determine the association between VDR SNPs (additive
effect) and the rate of culture conversion.
Results: Genomic analyses were performed on 96 MDR
TB subjects enrolled in the sub-study; 60% were female
and median age was 35 years (interquartile range [IQR]
30-42). Smoking was reported by 21% of subjects and
most subjects had HIV (81%), were smear negative
(57%), and had cavitary disease (56%). Overall, 92
(97%) subjects converted cultures to negative, with
median time to culture conversion of 57 days (IQR 17114). Of 121 VDR SNPs examined, 8 were significantly
S319
associated (P , 0.01) with both days to culture

conversion and rate of sputum conversion in multivariable analyses. Each additional risk allele on SNP
rs74085240 delayed culture conversion significantly
(adjusted hazard ratio 0.42, 95% confidence interval
0.22-0.78).
Conclusion: Polymorphisms in the VDR gene were
associated with time to sputum culture conversion in
MDR TB patients in this high HIV prevalence setting in
South Africa. Future studies are needed to investigate the
functional roles of these VDR polymorphisms and their
implications for treatment.
28. Drug-resistant TB: epidemiology and

transmission
PC-932-05 Primary transmission of drugresistant tuberculosis in South Fly District,
Western Province, Papua New Guinea
S Hiasihri,1 A Honjepari,1 A Aia,2 L John,2 E Lavu,3
C Coulter,4 G Chan,5 D Paison2 1Provincial Health Office,
Daru, 2National Department of Health, Port Moresby,
3
Central Public Health Laboratory, Port Moresby, Papua New
Guinea; 4Queensland Mycobacterial Reference Laboratory,
Brisbane, Queensland, 5Burnet Institute, Melbourne, Victoria,
Australia. e-mail: suman.majumdar@burnet.edu.au
Background: Tuberculosis is a major public health issue

in Western Province PNG. In 2013 the case notification
rate for TB was 912 per 100 000 and it was the leading
cause of hospital admissions and deaths. TB has placed
an immense burden on a struggling health system.
Adding to the challenge is the emergency and transmission of MDR-TB and XDR-TB.
Design/Methods: A retrospective cohort study using
routine program data from the PMDT registers was
conducted. Cases that were TB treatment nave were
identified as possible cases of primary transmission.
Demographic and clinical details were collected. The
DST results for MDR-TB and XDR-TB cases were
reviewed to identify implications for programmatic
management.
Results: There have been 216 cases of MDR-TB enrolled
in care from January 2012 to December 2014. The data
from the TB cohort at Daru General Hospital shows that
the proportion of treatment-nave cases enrolled in
MDR-TB care increased from 31% in 2012 to over
50% in 2014. Twelve case of XDR-TB have been
identified since 2012. Drug-susceptibility testing on
available samples with rifampicin resistance revealed
high rates of resistance to several TB drugs: pyrazinamide
(56.6%), streptomycin (96%), fluoroquinolones (15%),
ethionomide (99%). Demographic, clinical characteristics and interim cohort outcomes will be described.
Conclusion: Evidence of primary transmission of DRTB, inclduing XDR-TB strains is of concern in South Fly
District, PNG. Western Province is in the process of
scaling up its TB program in partnership with technical
S320
partners and a national level taskforce to implement an

accelerated response to the challenges of TB and DR-TB.
PC-933-05 Molecular characterization of MDRTB and XDR-TB strains among tuberculosis

patients in Chad
G W Ossoga,1,2 A Ba Diallo,2 R Ngandolo,1 A Thiam,2
F Gehre,3 I Ndoye,4 B De Jong,5 A Gaye-Diallo2
1
Veterinary and Zootechnical Research Laboratory of Farcha,
Ndjamena, Chad; 2Bacteriology and Virology Laboratory.
CHNU Le Dantec, Dakar, Senegal, 3Medical Research Council
Unit, Fajara, Banjul, Gambia; 4Biology Plant Department,
Science and Technology Faculty, University Cheikh Anta Diop
of Dakar, Dakar, Senegal; 5Institute of Tropical Medicine,
Antwerp, Belgium. e-mail: ossogagedeon@gmail.com
Background: Tuberculosis is a public health concern

worldwide particularly in Chad where culture and
resistance testing were not performed regularly. We
genotyped Mycobacterium tuberculosis complex strains
from patients who presented with pulmonary tuberculosis from different cities in the country and performed
resistance tests to anti-tuberculosis drugs.
Design/Methods: Sputum from different cities in Chad
were collected at different microscopy centers using
cetyl-pyridinium chloride between 2007 and 2012.
Genetic characterization was applied using standard
spoligotyping and MIRU-VNTR after culture based on
Lowenstein
Jensen media. We used the Genotype M.
TUBERCULOSIS DRplus v.2 and the BACTEC MGIT

960 for resistance testing.
Results: The study population of 486 New and 107
retreatment cases included 224 men (72%) and 87
women (28%). The age of patients was ranged from 12
to 70 years. 311 (52.5%) strains were genotyped
successfully out of which 304 (97.8%) were modern
sub-lineages and 7 (2.25%) ancestral sub-lineages.
Resistance tests among new cases gave 1.6% strains
mono-resistant to rifampicin, 4.5% strains resistant to
isoniazid and 0.96% MDR strains. For patients in
retreatment, 3.8% strains were mono-resistant to rifampicin, 9.0% to isoniazid, 3.5% were MDR and 0.6%
were XDR.
Conclusion: For the first time in Chad we identified
XDR-TB strains and a rate of 4.5% MDR-TB strains.
These resistant strains were found to be associated with
the CAM family.
PC-934-05 The timeline of developing MDR-TB

among contacts of MDR-TB patients, 20092012, Taiwan
P-C Chan,1,2,3 C H Liu,1,3 Y H Huang,1 H Y Tsai,1 K-F
Wang,1 P-H Lee,1 C-H Chen1 1Division of HIV/AIDS and TB,
Taipei City, 2Institute of Epidemiology and Preventive
Medicine, Taipei City, 3Department of Pediatrics, Taipei City,
Taiwan. e-mail: pcanita.tw@cdc.gov.tw
Background and challenges to implementation: Despite

slowly increasing coverage of multidrug-resistant tuberculosis (MDR-TB) treatment, contacts of MDR-TB
patients rarely receive preventive therapy. The manage-
ment of contacts of MDR-TB patients is challenging as

the evidence base recommending best practices is very
limited. Follow-up studies for development of MDR-TB
among contacts of MDR-TB patients are scarce. Since
April 2009, sending clinically reported MDR-TB strains
to the National Mycobacteria Laboratory for confirmation before registry in Taiwan has been required. Our aim
is to determine the timeline, rate and risk factors of
development of MDR-TB among contacts of MDR-TB
patients.
Intervention or response: We conducted a retrospective
cohort study among contacts whose index cases were
MDR-TB patients registered on the National web-based
TB Registry between April 2009 and December 2012. All
contacts were cross-linked to the National web-based TB
Registry in April, 2015 to identify whether the contact
developed active TB or MDR-TB during follow-up. We
used the sputum collection date of an index patient
related to a contact as the starting point for calculating
the time needed for MDR-TB to develop among contacts.
v2 test was used to determine risk factors for developing
MDR-TB among contacts.
Results and lessons learnt: A total of 565 MDR-TB
patients were registered during the study period and 550
patients (97%) had at least one contact registered in the
system. During follow-up, 60 contacts (0.9%) developed
active TB and 17 (0.3%) developed MDR-TB out of the
6568 contacts. Figure shows the progression of active TB
among contacts; 4 contacts developed MDR-TB beyond
a 2-year- follow-up (24%). After a 24 month-follow-up
of contacts after index patient sputum collection, six
contacts (46%) developed MDR-TB at 0-6 months; one
contact (8%) at 7-12 months; one contact (8%) at 13-18
months; and five contacts (38%) at 19-24 months.
Contact sharing a household with index patients were
2.8 (95% confidence interval: 1.1-7.4) times more likely
to develop MDR-TB compared to those not sharing a
household with index patients. Contacts of index
patients with smear-positive or cavitation were 5.2
(1.5-18.1) and 3.1 (1.2-8.1) times more likely to develop
MDR-TB compared to those without.
PC-935-05 Whole-genome sequencing to

delineate transmission of multidrug-resistant
tuberculosis in Shanghai, China: a crosssectional study
S321
urgently needed to prevent the transmission of MDRTB in China.
C Yang,1 T Luo,1 X Shen,2 J Wu,2 J Mei,2 K Deriemer,3

Z Yuan,2 Q Gao1 1Key Laboratory of Medical Molecular
Virology of Ministries of Education and Health, Institutes of
Biomedical Sciences and Institute of Medical Microbiology,
Shanghai Medical College, Fudan University, Shanghai,
2
Department of TB Control, Shanghai Municipal Center for
Disease Control and Prevention, Shanghai; 3School of
Medicine, University of California, Davis, Davis, CA, USA. Fax:
(86) 021-5423-7195. e-mail: yangstoptb@gmail.com
Background: Multidrug-resistant (MDR) tuberculosis

(TB) is a threat to tuberculosis elimination. MDR-TB
can be acquired during inappropriate treatment and case
management, or can be caused by person-to-person
transmission and infection with a MDR strain. We
investigated the magnitude of and risk factors for MDRTB cases arising from recent transmission of MDR strain
in China.
Design/Methods: We identified MDR TB notified in
Shanghai from Jan 1, 2009 to Dec 31, 2012. We used
variable number of tandem repeats (VNTR) genotyping
of clinical isolates of Mycobacterium tuberculosis to
identify clusters of TB cases with identical genotypes,
followed by whole genome sequencing (WGS) of the
VNTR-clustered isolates for higher resolution. We
estimated the proportion of TB cases in clusters defined
by WGS, calculated the pairwise genomic distances
between isolates within each cluster, and used logistic
regression to identify the risk factors associated with
clustering. We constructed and interpreted network
diagrams of the clusters derived from genotypic and
epidemiological data.
Results: There were 343 patients diagnosed with MDR
TB, of whom 324 (94.5%) had VNTR genotyping
profiles. There were 125 (38.6%) TB cases in 45 VNTR
clusters, and 122 clustered isolates were successfully
sequenced. The genetic distance between the isolates of
clustered cases ranged from 0-109 single nucleotide
polymorphisms (SNPs). There were 38 clusters with 2-10
TB cases whose 103 (84.4%) isolates differed by 12 or
fewer SNPs. Epidemiological links were identified in
56.5% of clustered TB cases, and the isolates from the
epidemiologically linked TB cases were separated by 12
SNPs or less. Being a local resident versus an urban
migrant (OR, 5.32, 95%CI, 3.08-9.20) and having a
delayed diagnosis of at least two months (2.57, 95%CI,
1.29-5.11) were the risk factors that were significantly
associated with WGS clustering. Residential communities were the most common setting of MDR TB
transmission, and clustered individuals lived a relatively
short distances from each other. The majority (73.7%,
28/38) of clustered MDR-TB isolates harbored compensatory mutations in the rpoA, rpoB and rpoC genes.
Conclusion: Recent transmission of MDR TB in the
study population was high. Multiple strategies, including
rapid diagnosis and effective case management, are
Figure Detailed investigation of cluster nine. Time of onset of symptoms,

diagnosis, and treatment (A). Epidemiological network (B). Genetic
distance and drug-related resistance mutations of WGS-based clusters.
Each circle represents clustered patients who were infected with isolates
separated by no SNPs; the size of circle indicates the number of patients;
Black dots are added when patients within circles are separated by more
than one SNP; Colored labels on the solid line represent drug-related
resistant mutations during the transmission chains, and label within a
circle represents mutations emerged within patients (C).
PC-937-05 Mycobacterium tuberculosis lineage

is more strongly associated with MDR-TB
infection than treatment history in the
Republic of Georgia
N Bablishvili,1 N Tukvadze,1 N Bzekalava,1 S Sengstake,2
I Bergval,2 S Alba,2 R Anthony,2 R Aspindzelashvili1
1
National Center for Tuberculosis and Lung Diseases, Tbilisi,
Georgia; 2Royal Tropical Institute (KIT), Amsterdam,
Netherlands. Fax: (995) 322 910 467. e-mail:
marikushane@yahoo.com
Background: A history of Tuberculosis (TB) treatment

increases the risk of multidrug resistant (MDR) TB.
MDR-TB may result from acquired resistance or
reinfection with resistant bacteria. In the Republic of
Georgia and other settings with a high burden of MDRTB the Mycobacterium tuberculosis Beijing lineage is
associated with MDR-TB. Typing and lineage identification of clinical isolates is usually performed retrospectively, if at all, and often reported apart from patient and
drug susceptibility data. Consequently the association
between M. Tuberculosis lineage and the MDR-TB
epidemic is not fully explored.
Design/Methods: Patient treatment history, DST profiles
and M. Tuberculosis genotype were determined for 399
diagnostic pulmonary TB samples collected at the
National Center for Tuberculosis and Lung Diseases in
Tbilisi, Georgia in 2012 and 2013. Association between
MDR-TB, patient treatment history and strain characteristics were examined by univariate and multivariate
Results: Out of 74 M(X)DR TB samples identified, 63
(85.1%) were assigned to the Beijing lineage and within
these group previous treatment was reported for 38
S322
(51.4%). The Euro-American lineage dominated 248 of

399 (62.2%) strains and the Beijing lineage accounted
for 140 of 399 (35.1%) strains. Cluster analysis of all
399 strains revealed two major M(X)DR-TB clusters one
of which was formed entirely of Beijing M(X)DR strains.
In total 42/74 (56.8%) M(X)DR-TB strains were part of
a cluster. In univariate analyses individuals infected with
a Beijing strain had 25-fold higher odds of being MDRTB than individuals infected with a Euro-American
strain; retreatment patients had 4-fold higher odds of
being infected with an MDR-TB. Multivariate analysis
confirmed that the effects of Beijing strain and retreatment were independent.
Conclusion: In our results infection with the M.
Tuberculosis Beijing lineage is more strongly associated
with multidrug resistance than treatment history. Furthermore, majority of M(X)DR cases were clustered,
suggesting that the driving force of the MDR-TB
epidemic in this setting is transmission. Strengthening
Infection Control measures that interrupt transmission
chains will therefore be expected to strongly impact the
incidence of MDR-TB. M. Tuberculosis genotyping
when combined with patient and DST information can
provide clues to the dynamics of MDR-TB transmission
and this type of analysis would be interesting to perform
in other settings where associations may be different.
PC-938-05 Drug resistance patterns in a

northeast Indian state through a line probe
assay laboratory functioning in PPM mode
M Mawlong,1 K Ranee,1 G Singh,2 M Puthenchirayil,2
V Sameer,3,4 T Thomas,3,4 P Varghese2 1Nazareth Hospital,
Shillong, 2Catholic Bishop Conference of India - Coalition for
AIDS and Related Diseases, Delhi, 3The Catholic Health
Association of India, Secunderabad, 4Catholic Health
Association of India, Secunderabad, India. e-mail:
Background and challenges to implementation: Access to

drug susceptibility testing (DST) is limited in many
regions in India and the transport of samples to regional
centres implies a diagnostic and treatment delay which
contributes to the rising burden of MDR TB. Early
diagnosis of RIF mono-resistance and multidrug resistance is paramount to reduce treatment failure, provide a
targeted intervention and reduce initial loss to follow up.
Public private mix initiatives provide a sustainable model
to expand the reach of DST in a universally accessible
manner with minimal delay.
Intervention or response: The LPA laboratory at Nazareth Hospital, Shillong is a certified state level referral
centre catering to all 7 districts in the state with a
combined population of 2.65 million. The laboratory
was established in the PPM mode without external
funding. Resistance patterns were evaluated among
samples over a period of two years from March 2013
to March 2015.
Results and lessons learnt: A total of 690 samples were
processed in a mean duration of 3.75 6 1.69 days
ranging from a minimum of one day to a maximum of 20
days. 453 (65.65%) men and 237 (34.35%) of women
were tested with a mean age of 34.49 6 14.37 years. 223

(32.32%) of the samples exhibited multidrug resistance,
47 (6.81%) exhibited INH mono-resistance and 34
(4.93%) exhibited RIF mono-resistance. 297 (43.04%)
were sensitive to both drugs and results were not
available for 89 (12.9%) samples. Females were significantly more likely to present with MDR (P value
0.02).
Conclusions and key recommendations: For every two
patients with MDR, one patient presents with INH or
RIF mono-resistance. 10% of the sample could not be
processed and some of these may have been atypical
Mycobacteria. LPA provides a sensitive, cost effective
approach to detect both MDR and INH or RIF monoresistance which can be rolled out as a PPM initiative.
PC-939-05 Drug-resistant tuberculosis in

Turkmenistan: results of the first nationwide
survey
M Durdyeva,1 S Tomasova,2 B Karriyeva,3,4 M Dara,5
H Hoffmann,6 M Zignol,7 P De Colombani,8 A Dadu8
1
Medical University of Turkmenistan, Ashgabat, 2National
Center of Infectious Diseases, Ashgabat, Turkmenistan;
3
World Health Organization Temporary Adviser, Yerevan,
Armenia; 4World Health Organization, Ashgabat,
Turkmenistan; 5World Health Organization, Brussels,
Belgium; 6Institute of Microbiology & Laboratory Medicine,
Gauting, Germany; 7World Health Organization, Geneva,
Switzerland; 8World Health Organization, Regional Office for
Europe, Copenhagen, Denmark. e-mail: dad@euro.who.int
Background: The first countrywide anti-tuberculosis

(TB) drug resistance survey in Turkmenistan was carried
out from August 2012 to February 2013.
Design/Methods: To investigate the levels and patterns of
resistance to first-line anti-TB drugs among new and
previously treated pulmonary TB cases and explore the
risk factors for multidrug-resistant (MDR) TB. Crosssectional study with sampling of all 64 (100%) TB
diagnostic units in the country. All consecutive patients
meeting the specific criteria for survey eligibility were
enrolled into the study during the data intake period.
Sputum samples were transported to the National TB
Reference Laboratory for bacteriological culture and
drug susceptibility testing to first-line anti-TB drugs.
Results: Of the 731 patients enrolled, 561 were newlydiagnosed and 170 previously treated TB patients. MDRTB was found in 13.9% (95%CI: 11.1-17.0) among the
new patients and 37.6% (95%CI: 30.3-45.4) among the
previously treated patients. History of previous anti-TB
treatment was a risk factor for MDR-TB (OR: 3.66;
95%CI: 2.455.46). None of the socio-behavioral and
demographic factors explored were found associated
with MDR-TB.
Conclusion: This is the first countrywide survey of antiTB drug resistance ever conducted in Turkmenistan. It
was implemented according to international standards
and provides valuable information for the future
planning of effective programmatic management of drug
resistant TB in the country.
PC-940-05 TB drug-resistant patterns of new

multidrug-resistant TB cases in Moscow
S Borisov,1 S Safonova,1 E Belilovsky,1 I Danilova1
1
Moscow Scientific and Clinical Center for TB Control,
Moscow, Russian Federation. Fax: (7) 495 964 86 37.
e-mail: belilovs.ky@gmail.com
Background: Information about TB drug-resistant patterns for different groups of patients (data about TB
drugs, for which the strain of M. tuberculosis was
resistant) is needed for planning and coordination of
activities for choice and prescription of treatment
regimens and chemotherapy control for multidrugresistant TB (MDR TB) patients, including extensively
drug-resistant (XDR TB) cases.
Design/Methods: Data of TB drug resistant (DR) for 82
new (NC) and 126 re-treatment (RTC) MDR-TB
patients registered in 2014 in Moscow, Russia, were
analyzed. According to city TB control regulations all
MDR TB patients were tested on drug susceptibility to
the first line TB drugs (FLD), fluoroquinolone (FqR-TB)
and injectable second-line drugs (ISLDR-TB).
Results: NC and RTC included 70.7% and 77.0% men
and age medians of 40.5 (IQR31.5-56) and 40.0
(IQR33.5-50), accordingly, P . 0.05. RTC patients
more often than NC demonstrated resistance to all FLD,
including streptomycin (S) and ethambutol (E): 38.1%
vs. 24.4%, OR1.2 (95%CI: 1.01-1.5). XDR-TB was
registered among 20.7% (12.6-31.1) NC with MDR-TB
and among 23.0% (16.0-31.4) RTC, P .0.05. PreXDR-TB (f), which was defined as MDR-TB plus FqRTB cases, was more common among RTC (19.1%) than
NC (8.5%): OR1.13 (1.01-1.3). Pre-XDR-TB (i),
which was defined as MDR-TB plus ISLDR-TB cases,
was slightly more common among RTC (22.2%, 15.330.5) than NC (19.5%, 11.6-29.4), P .0.05. There was
no statistical differences for RTC and NC in DR to
levofloxacin (12.7% and 13.4%), and moxifloxacin
(8.7% and 7.3%, accordingly). At the same time,
41.3% of RTC patients had TB with resistance to
ofloxacine vs. 24.4% for NC: OR1.3 (1.1-1.6). There
was no statistical differences for RTC and NC in DR to
amikacin (7.9% and 14.6%), capreomycin (13.5% and
6.1%, accordingly) and kanamycin, for which the very
high level of DR was found for both groups of patients:
40.8% (34.1-47.9).
Conclusion: Variety of DR patterns among new and
retreatment patients confirms significant role of drug
susceptibility testing (DST) data for the assignment of
successful regimens of MDR TB chemotherapy based on
as DST results as well empirical data.
S323
PC-941-05 Ongoing multidrug-resistant

tuberculosis transmission in Ghanzi district,
Botswana, 20122014
C Modongo,1 D Mbatshi,1 A Finlay,2,3 R Boyd,2,3
J Oeltmann,3 E Click,3 P Moonan,3 N Zetola1 1Botswana
University of Pennsylvania Partnership, Gaborone, 2U.S.
Centers for Disease Control and Prevention, Gaborone,
Botswana; 3U.S. Centers for Disease Control and Prevention,
Atlanta, GA, USA. e-mail: afinlay@cdc.gov
Background: Ghanzi, a rural district in Botswana, is

characterized by a low population density (0.4 persons/
km2), a high burden of tuberculosis (TB) (.1000/100
000 population), and high prevalence of HIV among TB
patients (36%). In 2008, the prevalence of MDR-TB
among new and previously treated patients was 2.5%
and 6.6% respectively. The Kopanyo study was designed
to use molecular epidemiologic methods combined with
geospatial and demographic information to identify
previously unrecognized TB transmission among social
networks and to identify geographic hot-spots for
transmission in Botswana.
Methods: From Aug 2012 to Dec 2104, Mycobacterium
tuberculosis isolates collected from culture-positive
pulmonary TB cases reported in Ghanzi underwent
drug-susceptibility testing of rifampicin, isoniazid, ethambutol and streptomycin and 24-locus mycobacterial
interspersed repetitive unit-variable number tandem
repeat (MIRU-VNTR) analysis. TB genotype clusters
were defined as 72 case-patients with matching 24-locus
MIRU-VNTR results. The number of clusters, cluster
size and the proportion of patients in a cluster were
analysed. Patients were interviewed and their clinical
records reviewed to identify potential epidemiologic
linkages.
Results: A total of 314 cases were reported in Ghanzi,
including 11 (3.5%) with multidrug-resistant (MDR)
TB. Genotyping results were available for 288 participants. In total, 86 different genotypes were identified,
including 27 clusters. Median cluster size was 14
(interquartile range: 376). The proportion of patients
included in genotype clusters was 169/288 (59%). One
cluster of 9 patients included 9 (82%) of the 11 MDR-TB
cases diagnosed. This was the only cluster with MDR-TB
cases. Among the MDR-TB genotype cluster members, 7
(78%) belonged to the same minority ethnic group
(Basarwa). Epidemiologic linkages between MDR-TB
cluster members were established for 7 (78%) cases and 5
(71%) of them had no prior TB treatment. All 7 were
from two communities and 4 (29%) had non-household,
community-based linkages.
Conclusion: These results suggest potential MDR-TB
transmission in Ghanzi. A robust public health response
of community-based targeted testing for active TB and
treatment is underway to identify potential undiagnosed
secondary cases, including long-term follow-up of
exposed contacts to monitor progression to disease.
S324
29. LTBI testing, perceptions and

epidemiology
based on the result of the QTF-GIT and the assessment of

the cost-effectiveness relation of using QTF-GIT among
the PLHIV.
PC-942-05 Tuberculin skin testing and

QuantiFERON TB Gold In-Tube to indicate the
treatment of latent tuberculosis infection in
people living with HIV
PC-943-05 Optimization of latent tuberculosis

screening for new refugees in Canada
J Souza,1 M D S Evangelista,1,2 J P Toledo,2 F Dockhorn,2

A Trajman3,4 1University of Brasilia, Brasilia, Distrito Federal,
2
Ministry of Health, Brasilia, Distrito Federal, 3Federal
University of Rio de Janeiro, Rio De Janeiro, Rio de Janeiro,
Brazil; 4Montreal Chest Institute, McGill University, Montreal,
Quebec, Canada. e-mail: josianemaria@unb.br
Background: Human immunodeficiency virus (HIV) is

one of the risk factors for infection and development of
tuberculosis (TB) in people living with HIV (PLHIV),
hence the importance of early identification of Latent TB
infection (LTBI) and prophylactic treatment in this
population. We analysed the use of the tuberculin skin
test (TST) and QuantiFERON-TB Gold in-Tube (QTFGIT) to identify latent tuberculosis infection in people
living with HIV and the outcome of cases treated with
isoniazid.
Method: The sample consisted of 300 PLHIV selected in
eight outpatient sexually-transmitted disease public
clinics in Federal District were submitted to QFT-GIT
and TST, between 2011 and 2013, with follow-up until
May 2014. A positive result of either test was considered
as LTBI. The association of independent variables and a
positive result was estimated by the adjusted odds ratio
(OR) in a multivariate model using logistic regression.
Wilcoxons non-parametric test was used.
Results: The mean CD4 was 477.5cells/mm3. Eighteen
patients (6%, 95%CI: 3.6%-9.3%) had LTBI, of whom 4
(1.3%, 95%CI: 0.04%-2.63%) had only a positive TST,
8 (2.7%, 95%CI: 0.8%-4.5%) had only a QFT-GIT
positive test and 6 (2%, 95%CI 0.4%-3.6%) had
positive results for both tests. This represents an 81.8%
relative increase in LTBI detection when QFT-GIT is
added to TST. The concordance between both tests was
96% (k 0.48). In the follow-up, one patient was
excluded who revealed a history of TB, leaving 17
patients with LTBI (5.1%), 12 (70.6%) of whom
received isoniazid preventive therapy (IPT)-none of them
became ill due to TB. The mean protection after the IPT
was 495.4 days for QTF-GIT and 540 days for
TST75mm. Two cases of pulmonary TB were identified
among the patients with negative results on the QTF-GIT
and TST.
Conclusion: The results revealed a low prevalence of
LTBI based on the results of the TST and QTF-GIT
among the PLHIV with mild immunosuppression in a
scenario of low TB load. The QFT-GIT alone was more
effective to detect LTBI than TST alone and had an 81%
added value as an add-on sequential test. There was
protection of IPT by the results of the two tests, however
studies are needed to assess other risk factors and
biomarkers. In addition, a longer follow-up period is
needed to assess the durability of isoniazid protection
L Delmar,1 D Fisher,1 N Hajjar,2 J Jarand,1 S Field,1

H Gardiner3 1University of Calgary, Calgary, AB, 2Refugee
Health Clinic, Calgary, AB, 3Tuberculosis Services, Calgary,
AB, Canada. e-mail: ljdelmar@ucalgary.ca

refugees arriving in Canada from tuberculosis (TB)
endemic countries experience a higher prevalence of
latent tuberculosis infection (LTBI) and are at a higher
risk for TB reactivation than other migrant groups. The
pre-arrival immigration medical examination aims to
detect cases of infectious pulmonary TB disease but does
not include screening for LTBI. Canadian guidelines
recommend targeted screening and preventive treatment
for LTBI as a strategy for TB control and prevention in
new refugees. There is a joint partnership for LTBI
screening and treatment between TB Services and the
Refugee Health Clinic (RHC) in Calgary, Alberta.
Intervention or response: Patients were screened at RHC
with the tuberculin skin test (TST). Those with a positive
TST (10mm cut-off) received a confirmatory interferon
gamma release assay (IGRA) test to improve diagnostic
specificity and chest X-ray (CXR) to rule out pulmonary
TB prior to physician assessment at Calgary TB Services.
Only those with positive IGRA findings or abnormal
CXR were clinically assessed at TB Services. A retrospective chart review of refugees screened at RHC from
November 2013 to December 2014 was performed to
investigate the process.
Results and lessons learnt: A total of 476 patients were

screened with the TST at RHC and 91 were positive with
the 10mm cut-off. Eighty of those patients (80/91)
completed an IGRA test which confirmed 60 cases of
LTBI at a concordance rate of 75% (60/80). At TB
Services, 51 patients were offered a 3-month regimen of
isoniazid/rifampin and the rate of treatment completion
was 82% (42/51). The median time from arrival in

Canada to confirmed LTBI diagnosis was 113 days,
which is within the 2-year period of increased risk for TB
reactivation. Throughout the process there was patient
drop-off in that 5% (24/476) did not return for the TST
reading, 12% (11/91) did not complete the IGRA test,
8% (5/60) did not follow up at TB Services and 12% (6/
51) did not complete treatment because they were lost to
follow-up or non-compliant.
Conclusions and key recommendations: The investigation of the LTBI screening process in Calgary revealed a
high rate of LTBI treatment completion in a population
that has been historically hard to reach. Addressing the
areas of patient drop-off throughout the screening
process would improve the rate of detected LTBI and
increase opportunity for TB prevention and elimination.
PC-944-05 Risk predictors for developing

tuberculosis among adult tuberculosis contacts
P-C Chan,1,2,3 Y H Huang,1 M-J Lu,1 P-H Lee,1 C B Hsu,1
K-F Wang,1 C-H Chen1 1Centers for Disease Control, Taipei
City, 2National Taiwan University, College of Public Health,
Taipei City, 3National Taiwan University, College of Medicine,
National Taiwan University Hospital,, Taipei City, Taiwan.
e-mail: pcanita.tw@cdc.gov.tw
Background: Tuberculin skin test (TST) is not routinely

recommended for tuberculosis (TB) contacts born before
January 1, 1986 in Taiwan due to the concern of bacille
Calmette-Guerin (BCG) immunization beyond infancy
and isoniazid preventive therapy (IPT) related hepatotoxicity. A pilot program launched during 2010 2012
provided both TST and QuantiFERONw-TB Gold InTube (QFT-IT) test to contacts born before January 1,
1986. This study is meant to evaluate the risk predictors
for developing TB among these contacts.
Design/Methods: 2203 contacts aged 29 years or older,
who received both tests in the pilot program in eleven
townships, were enrolled for model derivation. All
contacts were cross-linked to the National web-based
TB Registry on April 10, 2015 to identify if the contact
developed active TB during follow-up. Contagiousness
of TB index cases, comorbidities of contacts, results of
TST and QFT-IT, exposure periods, and relationships
between index cases and contacts were used to generate a
Cox proportional hazards model. We used the notification date of an index case related to a contact as the
starting point. The endpoint of event-free time could be
earlier than the follow-up end date, including: the date
on which the contact was notified as an active TB case
(event), the date of the commencement of IPT in a
contact (censored), or the date of mortality due to any
cause (censored).
Results: Of 2203 contacts who completed both TST and
QFT-IT, 415 (18.8%) were double-positive (TST7
10mm and QFT-IT7 0.35 IU/mL) and 194 (47%) of
these completed LTBI treatment. Amongst those who
received or did not receive IPT, only exposure period was
statistically significant (P 0.042, v2). The number
needed to treat for contacts with double-positive results
was 43. At the end of follow-up, 52 contacts died and 8
S325
TB cases developed among contacts: 7 from the contacts

with double-positive results and one from the contacts
with double-negative results developing culture-confirmed TB at three months after contact investigation
(TB risk: 0.032 vs. 0.0006, P , 0.001 for Poisson
analysis). The table shows a final multivariate Cox
regression analysis. The Contacts with double-positive
results had a TB risk of 1690/100 000 compared to 60/
100 000 for those with single-positive or double-negative
results (adjusted hazard ratio [aHR]: 50.9, 95% confidence interval [CI]: 5.9-439.8). Household contacts had
a TB risk of 640/100 000 compared to 90/100 000 for
non-household contacts (aHR: 9.9, 95%CI: 1.1-93.0).
Conclusion: Among contacts who received BCG immunization beyond infancy, the predictors for developing
active TB were double-positivity tested by both TST and
QFT-IT and the sharing of a household between index
cases and contacts. For adult contacts of contagious TB
patients, LTBI screening by using TST and QFT-IT in
conjunction with LTBI treatment could be helpful for
achieving global TB control goals for 2035.
PC-945-05 Therapeutic education in latent

tuberculous infection: analysis of patients and
relatives needs and proposal of a list of
competences to acquire
S Le,1 A Artebasse-Medjahed,2 A Rouault,1 C Brahmy,2
A Bozize,2 C Guillet Caruba,3 O Bourdon,1
F Doucet-Populaire3 1Assistance Publique Hopitaux

Paris,
Universite Paris 13, Paris, France,
Hopital Robert Debre,
2
Centre de Lutte Antituberculeuse des Hauts de Seine,
Nanterre, 3Assistance Publique Hopitaux

Paris, Hopital
ere,
`
Antoine Becl
Universite Paris Sud, Clamart, France.
e-mail: florence.doucet-populaire@abc.aphp.fr
Background: Prevention of Tuberculosis (TB) has major

public health implications. We focus on latent tuberculosis infection (LTBI) because its treatment is a key
component for controlling and reducing TB disease
incidence. For LTBI, starting treatment and following it
through is difficult to understand for asymptomatic
patients. The patients acceptance of LTBI treatment is
often influenced by the initial approach of the health care
provider. French Speaking Pneumology Society recommends therapeutic patient education (TPE) for these
patients, however there is no competence guidelines. Our
objective was to identify patients and their relatives
educational needs and to establish a list of competences
that patients have to acquire during this TPE program for
LTBI.
Design/Methods: Interview guidelines for patients and
health care professionals were written following literature analysis. On the basis of these guidelines, semistructured interviews were conducted until data saturation was reached. Twelve patients (7 males, 5 females,
and aged 14 to 49) from a fight against TB center near
Paris (Centre de Lutte Anti-Tuberculeuse des Hauts-deSeine (CLAT)) and 14 health professionals (3 lung
specialists, 4 pediatricians, 4 nurses, 1 pharmacist, 1
psychologist and 1 anthropologist) were interviewed.
S326
Results: From these interviews, we identified 32 learning

objectives belonging to one of the following topics:
knowledge of the disease, medication management and
monitoring and psychological aspects. For each learning
objective, competences to acquire were defined. The
medication management turns out to be the major
difficulty, particularly fasting constraint and adherence
to the full treatment course.
Conclusion: On the basis of this competences list, the
CLAT will be able to create a TPE program for LTBI
treatment. Several modules on demand will be offered
and taught by a professional trained for TPE, according
to the educational diagnostic. They will consist of face to
face individual, or for families, collective interviews for
about 30-45 minutes. These interviews will take place
just after a medical consultation, a blood test or
according to the needs expressed by patients. An
evaluation session will be planned at the end of the
program. The aim of this program is to promote
treatment adherence. Moreover, the patients having
competences will be a vector for knowledge and
competences passing on.
PC-946-05 Correlates of adherence to latent

tuberculous infection therapy in children
A Cruz,1 J R Starke1 1Baylor College of Medicine, Houston,
TX, USA. Fax: (1) 832 825 1182. e-mail: acruz@bcm.edu
Background: Therapy for latent tuberculosis infection

(LTBI) is safe and efficacious in children; efficiency is
limited by poor adherence. The objective was to
determine associations between LTBI completion and
demographics and mode of LTBI diagnosis.
of children 6 18-years-old seen for LTBI at a TB clinic in
a low-incidence nation. Isoniazid (9-months: 9H),
rifampicin (4R), or isoniazid/rifapentine (3HP) were
used as directly-observed preventive therapy (DOPT) or
administered by the family (self-administered).
Results: 684 children (50% Hispanic, 26% Asian, 14%
Black, 10% white) with LTBI had outcomes known;
69% were diagnosed by tuberculin skin test (TST), 17%
by interferon gamma release assay (IGRA), and 14% by
both TST and IGRA. 36% were identified via contact
investigations. 83% completed therapy (82% receiving
9H, 79% receiving 4R, 100% receiving 3HP). On
univariate analyses, completion was associated with
younger age (7.1 vs. 8.3y, P 0.03), use of DOPT vs.
self-administered therapy (94% vs. 67%, odds ratio
(OR) 8.1; CI: 5-13), diagnosis by TST vs. IGRA (85% vs.
71%; OR: 2.25, CI: 1.41-3.6), and use of 3HP vs.
isoniazid (100% vs. 82%, P 0.002) on univariate
analyses. When self-administration regimens were compared, rifampin was more likely to result in completion
than isoniazid (80% vs. 56%, OR: 2.6, CI: 1.46-4.64).
On multivariate analyses, only the use of DOPT (OR:
9.7; CI: 5.6-16.8), diagnosis by TST (OR 1.6; CI: 1.22.3), and use of 3HP (OR 2.2; CI 1.3-3.6) were
associated with LTBI completion. Subgroup Overall
Completed (n 565) Defaulted (n 119) OR [CI] 9H,
self 149 (22%) 84 (56%) 65 (44%) 1 9H, DOPT 331

(48%) 309 (93%) 22 (7%) 10.9 [6.3-18.7] 4R, DOPT 16
(2%) 16 (100%) 0 P , 0.001 4R, Self 108 (16%) 86
(80%) 22 (20%) 3 [1.7-5.3] 3HP 44 (6%) 44 (100%) 0 P
, 0.001 Self-administered 260 (38%) 173 (67%) 87
(33%) 1 DOPT 395 (58%) 372 (94%) 23 (6%) 8. [513.3].
Conclusion: Demographic factors were unassociated
with LTBI adherence. Short-course therapy, and therapy
via DOPT, were the only variables associated with
treatment completion.
PC-947-05 Effect of age and gender on risk of

latent tuberculous infection in household
contacts: Kampala, Uganda and Vitoria, Brazil
1
M Gaeddert,1 E Jones-Lopez,
J Ellner,1 N Hochberg,1,2
2
2
T Tsacogianis, L White, R Dietze,3 I Ayakaka4 1Boston
Medical Center, Boston, MA, 2Boston University, Boston, MA,
USA; 3Universidade Federal do Espirito Santo, Vitoria, Esprito
Santo, Brazil; 4Makerere University, Kampala, Uganda.
e-mail: mary.gaeddert@bmc.org
Background: Although there is a male predominance in

tuberculosis disease (TB) cases worldwide, there are
limited data on whether such differences are seen in
latent TB infection (LTBI). We investigated whether there
were differences in LTBI by age and sex in two settings.
Design/Methods: We analyzed data from studies in
Vitoria,
Brazil and Kampala, Uganda that enrolled
household contacts (HHC) of pulmonary TB index cases

(IC). We collected data on IC and HHC demographics.
Tuberculin skin tests (TST) were administered at baseline
and 6-8 weeks later if negative at baseline. LTBI was
defined as a TST of 710mm at either time point. P values
were calculated using logistic regression fit with generalized estimating equations.
Results: In Brazil, 423/939 (45.1%) of HHC were male,

and the median age was 21.0 years (range 0.3-87.0); in
Uganda, 194/442 (43.9%) were male, and the median
age was 13.0 years (range 0.1-76.4). In Brazil, 61.9% of
males and 69.4% of females had LTBI; in Uganda 77.3%
of males and 82.7% of females had LTBI. The LTBI
prevalence in Brazil was significantly different between
age categories for each gender: 0-4 years (44.9% males,
64% females), 5-14 years (63.4%, 63.2%), 15-44 years
(71.5%, 76.4%), and 45 (47.2%, 72.6%; P 0.01 for
males and 0.01 for females). In Uganda, the differences
were not significant for males but were for females: 0-4
years (75.0% males, 70.7% females), 5-14 years (72.6%,
80.7%), 15-44 years (86.3%, 89.4%), and 45 (70.0%,
99.4%; P 0.1 for males, 0.03 for females). The
differences between genders were not significant at either
site. In multivariate models that controlled for factors
related to IC disease severity (e.g., smear grade) and
HHC factors (e.g., sleeping proximity to IC), we found
that Brazilian females age 15-44 were 2.8 times (and
males 2.0 times) as likely to have LTBI as those of the
same gender aged 5-14 years (P , .0001 and P .004). In
Uganda, males age 15-44 years were 2.4 times as likely to
have LTBI as those age 5-14 years (P .04); the difference
was not significant in females.
Conclusions: We found notable differences in the age and
gender of HHC with LTBI in Brazil and Uganda. The
increased odds of infection among adult Brazilian
women compared to younger ages (and compared to
men) may result from increased biological susceptibility
or infection outside the household. Between-site differences in LTBI may reflect differential risk of community
infection between Brazil and Uganda.
S327
contacts, 127 (40.3%) were male, the median age was 23

years (IQR 15-39), and 63 (20.0%) were spouses of the
index case. Of the 96 (30%) household contacts with
LTBI, 31 (32.3%) were male, and the median age was 27
years (IQR 16-47). The overall prevalence ratio comparing LTBI in male to female household contacts is 0.73
(0.96 for ages 615, 0.68 for ages 16-45, and 0.61 for
ages 46-65). Household contacts who were female
spouses of index cases were more likely to have LTBI
than male spouses of index cases (24 [36.9%] vs. 2
[6.5%]; P 0.0017).
Conclusion: Preliminary analyses of these data show that
the male household contacts were less likely to develop
LTBI than female household contacts, and that these
differences become more apparent with increasing age.
Additional analyses are needed to control for potential
confounders including infectiousness of index case,
sleeping proximity to index case, co-morbidities, and
nutritional status. Mechanistic studies are needed to
understand why a female predominance is seen among
LTBI cases and how this relates to the male predominance seen among TB cases.
PC-948-05 Age and gender distribution of

latent tuberculous infection cases in a
household contact study, India
S Sarkar,1 P Fernandes,2 S Lakshminarayanan,1
R Kubiak,3 R Horsburgh,3 R Thanjavur,1 J Ellner,3
N Hochberg3 1Jawaharlal Institute of Postgraduate Medical
Education and Research, Pondicherry, India; 2Boston
Childrens Hospital, Boston, MA, 3Boston Medical Center,
Boston, MA, USA. e-mail:
priyanka.fernandes@childrens.harvard.edu
Background: Worldwide, there is a clear gender differential in tuberculosis disease with up to seven times more
cases occurring among men than women in parts of
India. What is unknown is the degree to which men are
more likely to also develop latent tuberculosis infection
(LTBI) and the age at which gender differences become
apparent.
Design/Methods: Data come from the Indo-US Regional
Prospective Observational Research for Tuberculosis
(RePORT) cohort in Pondicherry and Tamil Nadu in
southern India. This prospective cohort follows newly
identified culture-confirmed pulmonary TB cases
through treatment completion. Household contacts (7
6 years of age) are enrolled within 8 weeks of index case
enrollment, socio-demographic data are collected and
tuberculin skin testing (TST) is performed. We analyzed
the age and sex distribution of the LTBI cases, with LTBI
defined as TST induration 710mm. The v 2 test was used
for statistical comparisons.
Results: Data on 315 household contacts of 94 index
cases were analyzed. Among index cases, 72 (76.6%)
were male, and the median age was 33 years (interquartile range [IQR] 17-45). The median duration of
cough in index cases was 4 weeks (IQR 2-4), and 22
(23%) had a concentrated sputum smear grade of 3 (12
[13%] were negative, 10 [11%] were scanty, 28 [30%]
were 1, and 22 [23%] were 2). Of the household
PC-949-05 Patient perceptions prior to LTBI

screening at primary care in Newham, London
S Ikram,1 H Kunst1 1Barts Health NHS Trust, London, UK.
e-mail: sikram26@yahoo.com
Background: The London borough of Newham has the

highest rate of tuberculosis (TB) in the UK. In 2013, 335
Newham patients were treated for TB a rate of 107 per
100 000 people; 87 % of patients diagnosed were born
outside of the UK, with the vast majority of cases in
patients of Indian sub-continent origin. In Newham,
screening for latent TB was started in 2014 with the aim
of identifying and treating people with latent TB to
prevent development of active disease and spread of TB.
The initiative has been incorporated into the new patient
check on new patient registeration at a general practice
surgery. Eligible patients are identified via a serious of
questions including country of birth; if eligible they are
consented and given an information leaflet about TB (in
English). Screening is then done in the form of an
interferon gamma assay (IGRA) blood test.
Design/Methods: A pilot questionnaire was distributed
over a three week period to patients at a large general
practice surgery within the borough. All patients
applicable for being screened for latent TB were asked
S328
to complete an anonymous questionnaire. The aim was

to gain an insight into the demographics, knowledge and
understandingof the targeted Newham patient population regarding latent TB infection. The questionnaire was
adapted from that used by Goswami et al.
Results: The questionnaire was fully completed by 12
patients. 11 of the 12 patients had been born in the
Indian sub-continent. 6 different languages were stated
as being the primary language of the patients.None of the
patients believed they could have latent TB, and all of
them would tell their family if the result of the IGRA was
positive. 75% of patients believed that latent TB
infection could make them unwell, however all respondents thought TB infection could be cured. 3 patients did
not understand the term latent or dormant TB. Only 2
patients were concerned that if the IGRA was positive,
others would assume that they had infectious, active TB.
Conclusion: The results show that the screening programme is covering patients from countries of origin
known to present with TB in Newham. That patients
would inform family members whatever the IGRA result,
indicates an openness to discuss TB. However further
information including an information leaflet in a number
of languages on active and latent TB and their
differences, as well as the potential treatment and side
effects, is needed at the new patient check stage.
30. Collaborative services for TB and

diabetes
PC-950-05 A case-control study of the
differences in serum IFN-c and IL-10 levels in
tuberculosis-diabetes comorbidity patients and
their controls
L Yuan,1 X Xiao,1 C Chen,2 Q Shu,3 W L Jiang,1 B Xu,1
Q Zhao1,2 1Fudan University, Shanghai, 2Center for Disease
Control and Prevention of Jiangsu Province, Nanjing, 3Center
for Disease Control and Prevention of Jiangxi Province,
Nanchang, China. Fax: (86) 215 423 7335. e-mail:
zhaoqi@shmu.edu.cn
Background: The burden of dDiabetes mellitus (DM)

and tuberculosis (TB) comorbidity has emerged as a
global public health priority, especially in low and
middle-income countries. Effects of diabetes on innate
and adaptive immunity, which are potentially relevant to
tuberculosis defense, have been identified, but have not
been verified in the interaction of these two disease
states. This study aims to describe serum IFN-c and IL-10
levels among DM-TB comorbidity patients, pure TB
(PTB) patients, pure DM (PDM) patients and healthy
controls, and to reflect the influence of coincident
diabetes on immune function in tuberculosis patients.
Methods: A case-control study based on the screening of
DM in TB patients was carried out in 4 counties in
eastern China. All registered active TB patients during
April 2013 to March 2014 were screened by fasting
blood glucose (FBG) test. Those who had an FGB equal
to 126 mg/dl or higher were referred for glycolated
hemoglobin A1C (HbA1C) test for the confirmation of

DM. Cases in this study were all diagnosed DM-TB
patients, whereas control groups were PTB patients,
PDM patients and healthy controls matching with gender
and age. Serum levels of IFN-c and IL-10 were measured
by avidinbiotincomplex enzyme-linked immunosorbent
assay.
Results: In total, 71 DM-TB comorbidity patients, 71
PTB patients, 71 PDM patients and 71 health controls
were recruited during the study period. The geometric
mean of serum IFN-c and IL-10 levels of participants
were 6.01 and 7.36 pg/ml, respectively; and there were
statistical differences among the four groups (P , 0.001).
The IFN-c level was higher in PTB patients compared
with healthy controls (geometric mean of 8.59 vs. 3.52
pg/ml); whereas it was lower in DM-TB patients
compared with PTB patients (geometric mean of 7.88
vs. 8.59 pg/ml). An opposite trend was indicated on
levels of IL-10. It increased in turn among healthy
controls, PTB patients and DM-TB patients, with
geometric mean at 4.85, 9.03 and 14.40 pg/ml,
respectively.
Conclusions: DM-TB patients had lower level of IFN-c
and higher level of IL-10 compared with PTB patients.
Findings from this study suggested that comorbidity with
diabetes may have negative effects on the immune
function of TB patients, which might affect the outcome
of TB treatment.
PC-951-05 Integrated care for tuberculosis, HIV

and diabetes in four public hospitals in two
regions of Ethiopia
D J Dare,1 N Demmelash,1 T Anteneh,1 I Jemal
Abdulahi,2 D Habte,1 M Aseresa Melese,1 P G Suarez,3
G Sangiwa3 1Management Sciences for Health, HEAL TB
Project, Addis Ababa, 2Oromia Regional Health Bureau, Addis
Ababa, Ethiopia; 3Management Sciences for Health, Center
for Health Services, Arlington, VA, USA. Fax: (251) 11 372
4473. e-mail: degujerene@gmail.com
Introduction: There is limited experience with integrated

management of TB, HIV and diabetes mellitus (DM).
Our objective was to develop and test an integrated
model of care for TB, HIV and DM in a high TB/ HIV/
DM burden setting with limited resources. We report
preliminary results from an ongoing pilot work from four
public hospitals in Ethiopia.
Interventions: First, we oriented clinicians working in
TB, antiretroviral therapy (ART), and DM clinics on
how to identify and manage/refer clients for the three
diseases; and provided minor supplies. Clinicians in ART
clinics screened patients both for DM and TB; those in
TB clinics for HIV and TB; and patients attending DM
clinics were screened for TB using symptom-based
checklists adapted from the national guidelines. Diagnosis of DM was confirmed if random blood glucose (RBG)
was .200 mg/dl or if patient was known diabetic. Where
the patient could afford, routine chest x-ray was
encouraged. Patients with presumptive TB were further
evaluated and sputum microscopy or GeneXpert was
done. HIV testing and counseling was done as per the
national standard. We collected and analyzed quantitative and qualitative data acquired during monthly
mentoring visits.
Results and lessons learnt: We screened 1156 patients
including 268 in TB, 525 in HIV, and 363 in DM clinics.
Of 268 TB patients screened, 8 (3%) had DM of which
six were known diabetic patients while two were newly
detected. Also, 15.1% were co-infected with HIV. Of 525
patients receiving chronic HIV care, 12 (2.3%) had DM
of which nine were newly detected. Further, 54 (10.3%)
were currently co-infected with TB. Of 363 DM patients
screened for TB, eight had previous history of TB and
four (1.1%) had active TB at the time of screening (2
currently under treatment and 2 newly diagnosed).
Limited experience with symptom-based TB screening
and lack of recording and reporting tools were the major
challenges encountered in the DM clinics. High patient
load was the greatest challenge in the ART clinics while
TB clinics did not report any significant challenge. The
integrated approach was well received by program
managers and clinicians alike.
Conclusion: The yield of TB among DM patients was
about three times the prevalence in the general population. Three-fourths of the DM cases among HIV infected
patients were newly diagnosed. The screening helped
identify a quarter of DM patients detected in TB clinics.
The integrated approach should be implemented at wider
scale.
PC-952-05 Prevalence of diabetes mellitus

among TB patients and impact of diabetes on
treatment outcome: a field-based study under
the programme setting in India
S329
Results: Prevalence of DM among TB patients was

reported to be 10%, 5.3% and 7.9% in 2013, 2014 and
2015 respectively. Out of total 48 co-morbid patients, 33
patients had self-reporting history of DM, whereas 15
were newly diagnosed. Most patients were of senior age
groups (Mean age; 58 years, range; 40-75 years). 17
(46%) patients under the study were retreatment cases
(treatment failure and relapse). Mean random blood
sugar level among all cases at the time of diagnosis, end
of IP and end of CP was 182 mg/dl, 139 mg/dl and
120mg/dl respectively. Treatment outcome of patients as
per the program guidelines was cured in 17 (46%)
patients and treatment completed in 10 (27%)
patients. Four (11%) patients defaulted the treatment,
while four (11%) other patients had failure outcome.
Two patients died while on treatment.
Conclusion: The study highlights evidence that DM
makes a substantial contribution to TB incidence and
increases the morbidity and mortality. Geriatric population is more vulnerable to this co-morbidity. This study
emphasizes on the need to raise awareness and policy
change on screening for DM in persons with TB and
effective anti-diabetic treatment to TB patients with comorbidity. There is a need to develop a mechanism for
coordination between ongoing NCD control program
and TB control program in India.
PC-953-05 TB and DM linkage campaigns: how

beneficial is screening for people attending
camps?
R Thakur,1 N Ahmed2 1World Health Organization

Randstad, Shimla, 2O/o CMO Nahan, Sirmour, India. Fax:
(91) 0177 262 7601. e-mail: ravindermph@gmail.com
A Madhab,1 S Chadha,2 S Burugina Nagaraja,3

S Satapathy2 1Jagran Pehel, New Delhi, 2International Union
Office, New Delhi, 3Employees State Insurance-Post
Graduate Institute of Medical Sciences and Research &
Employees State Insurance Corporation Medical College,
Bangalore, India. e-mail: Sachi.Satapathy@theunion.org
Background: India is facing a dual burden of merging

epidemic of diabetes mellitus (DM) and tuberculosis
(TB); there were an estimated 65.1 million people with
DM and 2.3 million TB patients living in India in 2013.
Routine screening of TB patients for DM under
programme setting may be an opportunity for its early
diagnosis and improved management and might improve
TB treatment outcomes. Current study was conducted to
assess the co-existence of DM and TB in persons with
established TB and treatment outcome under programme
settings.
Design/Methods: Cohorts of notified TB patients in
2013 (3rd and 4th quarters), 2014 (all quarters) and
2015 (1st quarters) in Pounta Sahib TB Unit (Pop: 78
000) in District Sirmour of Himachal Pradesh in India
were tested for routine blood sugar using semi-autoanalyser at the time of diagnosis. TB patients with diabetes
received anti-diabetic medication from the physicians
and ATT under the program. Blood sugar examination
was repeated at the end of intensive phase (IP) and end of
continuous phase (CP) of ATT. Patients treatment
outcome was compared with level of blood sugar level.
Background: In India, tuberculosis (TB) continues to be a

major public health problem. The looming epidemic of
diabetes mellitus (DM) is presumed to be detrimental to
the ongoing TB control efforts in the country. The
National TB Control programme has a policy of
screening all TB patients for diabetes at districts where
non-communicable disease programme is implemented.
Jagran Pehel and The Union has launched a project
supported by WDF and State Health Soiety on enhancing
awareness about TB and DM linkages in the community
at non-NCD implementing districts from the state of
Bihar. We conducted a study to determine the usefulness
of screening the people for TB and diabetes attending the
campaigns.
Design/Methods: The project is being implemented in
Bhojpur districts with population of 27.20 lakhs since
September 2014; By March 2015 we have been able to
complete activities in the districts. The campaign consists
two types of activities a) IEC through displaying the
posters, pamphlets, wall painting and audio-visuals
through mobile vans at villages on prefixed days while
the trained personnel at the camps encourage and
educate the people to get screened for TB and DM, b)
S330
followed by screening camps for DM-TB the following

day. The details of the patients and their results were
electronically entered in Epidata data entry format.
Results: A total of 14 434 people were screened for TB
and DM during the period. Among them 502 (3.5%)
people had suffered from TB and 92 (18.3%) were
having diabetes. About 898 (6.2%) of patients were
suffering from diabetes and 271 (1.9%) of them were on
diabetic treatment; on screening them for TB, 432
(48.2%) patients were found to be sputum smear
positive.
Conclusion: Screening of people in TB and DM linkages
campaign is found to be beneficial. Nearly one in five TB
patients had diabetes and nearly one in two diabetic
patients had tuberculosis. The National TB programme
should emphasize on TB-DM linkages through extensive
Advocacy, communication and social mobilization
(ACSM) activities.
PC-954-05 Addressing TB-DM comorbidity in an

urban setup: experience from Delhi, India
A Khanna,1 P Kumar,2 S Chandra,3 B Neeti4 1State TB
Officer, Directorate of Health Services, Delhi, New Delhi,
2
District TB Officer, Municipal Corporaton Delhi, New Delhi,
3
Office of the World Health Organization Representative to
India, New Delhi, 4State TB Office, Directorate of Health
Services, Government of National Capital Territory of Delhi
(GNCTD), New Delhi, India. e-mail: chandras@rntcp.org

Tuberculosis (TB) and Diabetes Mellitus (DM) dual
burden screening framework of World Health Organization endorses that bidirectional screening for both the
disease control programs should be delivered within the
context of general health systems, with zero cost entailed
by the patients. Medical experts have opined that about
10% of TB cases globally are linked to diabetes. Our aim
was to assess the TB-DM screening package drafted by
Delhi State in context with the above recommendations,
along with assessment of its operational feasibility
amidst busy working schedule of TB workers in an
urban set up.
Intervention or response: Collaborative screening package of TB and DM was rolled out pan Delhi in January
2015. The bidirectional screening framework was in
accordance with the Government of India DM screening
guidelines. It differed from the Revised National TB
Control Program-screening of TB patients for DM
framework- as it enrolled registered TB patients more
than 30 years age and offered them fasting blood glucose
(FBG) test at point of entry into RNTCP (thus omitting
DM screening for all registered TB patients and random
blood glucose testing protocol).
Results and lessons learnt: From the preliminary reports
received for the first reporting quarter of 2015, out of the
2274 registered TB patients who were more than 30
years age, 1715 (75%) were screened for diabetes. In
them 225 (13%) were diagnosed as diabetes and were
referred to the linked DM treatment centre for their
treatment management. Among the 143 (66%) men TB
patients with DM, 79% were New TB patients whereas
21% gave prior history of TB disease. The FBG test was

available at government peripheral health institutes with
the link DM treatment sites in the vicinity of the TB
treatment centers. All services were offered free of cost to
the patients.
Conclusions and key recommendations: Bidirectional
screening especially for DM in registered TB patients
with inclusion of age more than 30 years and FBG at
point of entry with support from general health system
increases the yield of diabetes screening and makes it
operationally feasible for patients to undertake the FBG
test amidst their busy work schedule in an urban set up.
PC-955-05 High rates of previously known but

poorly controlled diabetes among TB patients
in Lima, Peru
C Ugarte-Gil,1 R Van Crevel,2 F Pearson,3 D Moore4
1
Universidad Peruana Cayetano Heredia, Lima, 2Radboud
University Medical Center, Nijmegen, Netherlands, 3St
Georges University, London, 4London School of Hygiene &
Tropical Medicine, London, UK. e-mail:
cesar.ugarte@upch.pe
Background: The prevalence of diabetes mellitus (DM) is

increasing in low and middle-income countries and is a
risk factor for worse tuberculosis (TB) treatment
outcomes. Although evidence concerning the TB-DM
association is increasing, there are gaps in our knowledge
of prevention, treatment and management strategies in
patients with these 2 comorbidities. TANDEM (www.
tandem-fp7.eu) is a multi-country TB-DM project with
field sites in Indonesia, Peru, Romania and South Africa.
We present preliminary data on screening TB patients for
DM in Lima, Peru.
Design/Methods: Adult patients recently diagnosed with
TB were enrolled in 3 DOTS primary health clinics in
Lima, and were screened for DM (using HbA1c
measured in an accredited laboratory) as a part of the
TANDEM project (Jan 2014 Mar 2015). DM, TB and
anthropometric information were recorded. Data captured electronically in a central online database (REDCap), and quality assurance programmes created and run
regularly to check for missing and out of range data.
Results: 225 TB patients were enrolled, 119 (52.9%)
were male and the median age of 34 years (IQR: 23-45).
Overall, 21 (9.3%) participants had HbA1c7 6.5%,
only one was younger than 35 years old (34yo). 20/111
(18.0%) participants older than 35 years old had HbA1c
7 6.5%. Among the group over 35 years, mean age was
48.4 (SD: 11.6), 53 (47.8%) were male, 32 (28.8%) had
a previous history of TB, median BMI was 22.8 (IQR
20.4-25.8) and median waist circumference was 83
(IQR: 75-90). The median HbA1c was 5.4 (IQR 5.15.9) with a maximum value of 17.1. 16 (14.4%) had
been previously diagnosed with DM before presenting
for TB treatment and only 4 (3.6%) didnt know their
DM status. Among the 20 participants over 35 years old
with HbA1c7 6.5%, mean age was 52.5 (SD: 9.4), 8
(40%) were male; median BMI was 22.9 (IQR: 21.627.2) and median HbA1c was 10.7 (IQR: 9.8-14.1). Half
were diagnosed over 12 months before starting TB
treatment and 11/20 (55%) were taking anti-DM

medication (insulin and/or metformin) at the time of
enrolment. 4/20 (20%) reported a previous TB episode.
Conclusions: Among TB patients in Peru, DM is
associated with older age, female gender and recurrent
TB. The high proportion of previously known, yet poorly
controlled DM underlines challenges in chronic care for
DM in LMIC, which also impact TB control. More study
is needed to develop optimal and sustainable preventive
and therapeutic strategies for both diseases.
PC-956-05 Mitochondrial mutations in diabetes

and isoniazid treatment of tuberculosis
L Wangh,1 A Osborne,1 J A Sanchez,1 K C Hayes1
1
Brandeis University, Waltham, MA, USA. e-mail:
wangh@brandeis.edu
Background: Diabetics are three times more likely to

develop active TB than non-diabetics, and having
diabetes adversely affects TB treatment and prognosis.
Mitochondrial dysfunction is associated with both
diabetes and isoniazid (INH) treatment of TB. It is
therefore possible that increased mitochondrial mutagenesis due to drug toxicity may exacerbate development
and progression of diabetes and, conversely, that
diabetes-dependent mitochondrial mutagenesis may exacerbate the side effects of TB drug treatment. We are
examining these complementary hypotheses in cultured
human liver cells and in Nile rats (Arvicanthis niloticus),
a model system that develops diabetes in response to a
high carbohydrate diet. We are also attempting to
mitigate mitochondrial DNA damage in these systems.
Design/Methods: Non-diabetic and diabetic Nile Rats
were scored for disease progression and severity by
measurement of random blood glucose and oral glucose
tolerance. Disease onset was significantly accelerated by
feeding a carbohydrate-rich diet to a diabetes-prone
strain of rats. Ingestion of palm fruit juice (PFJ), a watersoluble extract from the kernel of the oil palm (Elaeis
guineensis), delayed the onset of diabetes. HepG2 cells
that activate INH were treated for 30 days, with and
without PFJ. Liver cells from age-matched diabetic and
healthy rats, and cultured human cells were screened for
mutations in multiple regions of the rat and human
mitochondrial genome (mtDNA) using LATE-PCR and
Lights-On/Lights-Off probes.
Results: Diabetic animals had twice the number of
random mtDNA mutations as healthy animals. Mutations accumulated before onset of diabetes symptoms
and palm fruit juice reduced the number of mutations in
diabetic rats. Mutational load did not correlate with the
severity of diabetes. Chronic treatment of HepG2 cells
with INH increased mutations in the HV2 and ND3
targets as compared to untreated cells. Preliminary
experiments suggest that PFJ co-treatment may reduce
these additional mutations.
Conclusion: Mitochondrial mutations due to diabetes
and INH therapy raises serious concerns about possible
long term consequences for diabetic individuals, particularly those on INH therapeutically or prophylactically
S331
for many months. The potential benefits of PFJ to

mitigate mtDNA damage warrant further investigation.
Diabetes and drug treatment can be combined in the
future using the Nile Rat model system. Supported by a
grant from the MPOB.
PC-957-05 Chronic cavitary pulmonary

aspergillosis is a common complication of
pulmonary tuberculosis: results of a crosssectional survey
I Page,1,2,3 N Onyachi,4 C Opira,5 J Opwonya,6
S Hosmane,1 S Sawyer,1 R Sawyer,1 M Richardson,1,2,3
D Denning1,2,3 1University Hospital of South Manchester,
Manchester, 2University of Manchester, Mancheseter,
3
Manchester Academic Health Science Centre, Manchester,
UK; 4Gulu Regional Referral Hospital, Gulu, 5St. Marys
Hospital, Lacor, Gulu, 6Gulu District Health Office, Gulu,
Uganda. e-mail: iain.page@manchester.ac.uk
Background: CCPA complicates pulmonary tuberculosis,

with a 5-year mortality of up to 85%, but is treatable
with itraconazole or surgery. The estimated global
prevalence of CCPA secondary to tuberculosis is 0.8 to
1.3 million cases.
Design/Methods: We conducted a cross-sectional survey
in Gulu, Uganda. Adults who completed treatment for
pulmonary tuberculosis in the last 7 years were recruited.
All underwent clinical assessment and chest X-ray.
Aspergillus-specific IgG was measured by Siemens
Immulite, which has a sensitivity of 96% and specificity
of 98% for the diagnosis of CCPA. This assessment was
repeated in a resurvey two years later. CT scan was also
performed in those with positive serology or chest X-ray
changes suggestive of CCPA. GeneXpert TB PCR testing
was performed on those with productive cough. CCPA
was diagnosed in patients with ALL of the following; 1
over one month cough or haemoptysis, 2 raised levels
of Aspergillus-specific IgG and 3 paracavitary fibrosis
or aspergilloma or progressive cavitation on imaging.
Simple aspergilloma was diagnosed in patients with
aspergilloma and positive serology, but no symptoms.
Results: 400 patients were recruited between October
2012 and February 2013. 200 (50%) were HIV positive.
Median age was 42 years (range 16-83). 39% of patients
were female. Median CD4 count in those with HIV was
415 cells/lL (range 0-1400). 284 patients were resurveyed between October 2014 and February 2015.
Frequency of disease is shown in the table. Condition
Cases (n 284) CCPA 20 (7%) Simple aspergilloma 4
(1.5%) All chronic pulmonary aspergillosis 24 (8.5%)
HIV co-infection had no significant impact on the
frequency of CCPA. Three cases of recurrent pulmonary
tuberculosis were identified, none of which had CCPA.
Conclusion: CCPA is a frequent complication of pulmonary tuberculosis. This data supports the predicted
prevalence of up to 1.3 millions cases worldwide. This
represents a significant global public health problem,
which is currently neglected. Patients with CCPA will
meet the WHO diagnostic criteria for smear-negative
tuberculosis. Many cases of recurrent smear-negative
tuberculosis are probably CCPA. Accurate diagnosis of
S332
CCPA in this group would avoid unnecessary and

potentially harmful TB treatment and also significantly
reduce the reported global burden of smear-negative
tuberculosis. Further research is now needed to develop
appropriate CCPA screening and treatment strategies for
use in resource poor settings.
31. Spatial epidemiology of tuberculosis

PC-958-05 A spatial analysis of gender
differentials in tuberculosis in the Amazonas
State, Brazil, 2013
A Belchior,1 M Yamamura,1 M Popolin,1 M Santos Neto,2
M C C Garcia,1 A Queiroz,1 A C V Ramos,1 R Arcencio1
1
Nursing School of Ribeirao Preto, University of Sao Paulo,
Ribeirao Preto, Sao Paulo, 2Federal University of Maranhao,
Imperatriz, Maranhao, Brazil. e-mail:
aylanabelchior14@gmail.com
Background: Considering that gender inequalities may

influence the access to health services and delay in
tuberculosis (TB) diagnosis, it was necessary to identify
TB incidence according to gender in the state of
Amazonas, Brazil.
tive smear), and also, those recorded in 2013 and

reference variables of gender, age, diagnosis month, State
of notification, State of residence and municipality of
notification. It was possible to estimate female-male
incidence per municipality and choropleth maps development through the ArcGIS software, version 10.1.
Results: General TB incidence ranged from 0 to 189.48
cases per 100 000 inhabitants, the female rate ranged
from 0 to 202.74 cases per 100 000 female inhabitants
and male incidence from 0 to 177.86 cases per 100 000
male inhabitants. The area with the highest incidence
rate (Figure ) includes 12 municipalities, and five of
them: Amatura, Canutama, Ipixuna, Sao Gabriel da
Cachoeira and Tabatinga integrate the Arco Norte of the
Brazilian international border. A distinctive feature of
this region is the presence of indigenous people and their
intense mobility in the border region. In addition, three
out of six municipalities: Manaus, Sao Gabriel da
Cachoeira and Tabatinga, which are quite necessary for
TB control in the Amazon, integrate the group with the
highest incidence of TB in women.
Conclusion: Differentiated gender approach was not a
priority of the studies conducted in Brazil. However,
findings indicate that gender differences can relate to the
municipalities geographic location that may influence
womens accessibility to diagnosis and treatment.
PC-959-05 Spatial distribution of tuberculosis

cases in Sao Carlos, Sao Paulo, Brazil, between
2008 and 2013
S Protti,1 A Pereira Biffi Fusco,1
Alecrim,1 A Alessandra Dos Reis,1
T Ferraz De Araujo
R Arcencio,2 P F P Pedro Fredemir Palha,2 M Yamamura2
1
Federal University of Sao Carlos, Sao Carlos, SP, 2University
of Sao Paulo at Ribeirao Preto College of Nursing, Ribeirao
Preto, SP, Brazil. Fax: (55)16 3322 5265. e-mail:
tatienf@gmail.com
Design/Methods: Descriptive and ecological study that

considered data from the Information System of Notifiable Diseases (SINAN) of tuberculosis cases (new
extrapulmonary and pulmonary tuberculosis with posi-
Background: Tuberculosis is an important public health

problem in Brazil, with an estimated incidence of 42
cases per 100 000 inhabitants. Among the various
strategies launched to monitor the evolution of the
disease in the country, the spatial analysis can highlight
the main areas of infection and even identify populations
at risk of disease.
Design/Methods: This study analyzed the spatial distribution of tuberculosis in Sao Carlos (Sao Paulo state,
Brazil). An exploratory study of tuberculosis cases
registered in the Information System of the TB-WEB,
between the years 2008 and 2013, was performed. The
data were geocoded in TerraView software, version
4.2.2, and the interpolation statistical technique used
was the Kernel estimative, considering a band of 1000
meters. We used the software Statistica, version 10.1, for
bivariate statistical analyzes and the software ArcGIS,
version 10.2, for the construction of thematic maps.
Results: We identified 315 cases of tuberculosis, with 290
cases geocoded. Most of the identified cases were male (n
224; 70.89%) with a median age of 40 years. The
predominant clinical type was pulmonary (n 257;
81.33%). The spatial distribution of tuberculosis in Sao
S333
Carlos showed up in a heterogeneous manner, tending to

concentrate on specific areas of the city, important in
epidemiological terms.
Conclusion: The study highlighted the most problematic
areas in relation to tuberculosis, which require further
attention, planning and investment by the healthcare
management systems. In this sense, respiratory symptoms research programs; early case detection; increased
treatment adherence; as well as epidemiological investigation of contacts, should be prioritized in the areas
identified as a risk.
therefore, vary among subpopulations depending on the

regional epidemiology. This cluster analyses approach
can be implemented with open source software.
PC-960-05 An approach for spatial cluster

identification of tuberculosis using surveillance
data
Background: Geospatial Delhi Limited (GSDL) is the

custodian of the affluent database prepared during the
Delhi State Spatial Data Infrastructure (DSSDI) project.
The project has bestowed Delhi with an extremely rich
database with the entire infrastructure and facilities of
the city mapped using locality and sub-locality information in Geographic Information System (GIS). To utilise
this vast geo-knowledge resource and opportunity within
GSDL, it is envisioned to link up the this database with
enhanced Nikshay (e-Nikshay), a case-based web application, jointly developed by National Information
Centre (NIC) and Revised National Tuberculosis Control
Programme (RNTCP) for real time plotting of TB
patients location coordinates in GIS platform.
Intervention or response: Mapped locality and sublocality coordinates of entire Delhi state in the GeoWeb Portal were extracted in excel sheet and transferred
electronically to Delhi State TB Control Programme for
establishing linkages between e-Nikshay and GSDL
database. Each locality and sub-locality area coordinates
were filtered from the excel sheet, marked, validated and
linked to the respective RNTCP districts and tuberculosis
units (TUs) with codes, as unique identifiers, by the
RNTCP supervisory staff in the districts. NIC add
locality and sub-locality information in Nikshay TB
patient registration form for registering a TB case having
a valid Delhi residential address in e-Nikshay.
Results and lessons learnt: 3500 Delhi locality and sublocality coordinates of GSDL Geo-Web Services were
segregated and aligned to 25 RNTCP implementing
districts and 38 TUs. This was done for the purpose of
preparing an interface to plot the point of patient
address, registered in e-Nikshay, in GSDL database.
Conclusions and key recommendations: A patient
coordinates as a point in locality and sub-locality data
of GSDL database is plotted during real time entry of TB
cases in e-Nikshay to know the geographic distribution
of TB patients. Creation of Geospatial datasets for the
TB control program and linking it with the patient
management information system in e-Nikshay can assist
in identifying unserved geographical areas where TB
treatment and diagnosis facilities are insufficient.
D Roth,1 S Mak,1 M Otterstatter,1 J Wong,1 V Cook,1

J Johnston1 1British Columbia Centre for Disease Control,
Vancouver, BC, Canada. e-mail: david.roth@bccdc.ca
Background: Tuberculosis (TB) incidence is decreasing in

many countries, and an increase in the heterogeneity of
TB is expected as rates fall. Clustering will not occur
equally in all subpopulations, and not all clusters require
the same public health response. This necessitates the
development of approaches for the timely identification
and prioritization of disease clusters using surveillance
data. Here we present a simple approach using open
source software, and use British Columbia (BC) as a case
study for its practical application.
Design/Methods: All cases meeting the national TB case
definition from 1990-2013 are geocoded using 3 digit
postal code of residence. Clustering in TB incidence is
compared: 1) across the province, and then at finer
resolution, 2) between Canadian (CB) and foreign-born
(FB) cases, and 3) across 3 time periods (1990-1997,
1998-2005, 2006-2013). Gini coefficients are used to
quantify and compare the degree of global clustering of
cases in the BC population. SaTScan, a spatial scanning
window statistic, is used to identify statistically significant disease clusters, with results overlaid on provincial
maps of TB rates produced with Geographic Information
Systems mapping. Finally, epidemiological reports for
key clusters are produced and reviewed to prioritize
outreach, control, and education. All analyses are done
using freely available software.
Results: The degree of provincial clustering remains
stable over 3 8-year periods (Gini coefficient: 1990970.57, 1998-20050.54, 2006-130.50). However,
CB cases show a higher degree of clustering than FB
cases, especially in the latter period (Gini coefficient:
Can. Born 0.64, FB0.33). Incidence mapping and
SaTScan analysis reveal a geographically large stable
cluster for FB individuals, with smaller and less stable
clusters seen in CB populations.
Conclusion: The use of incidence mapping and cluster
detection methods together can achieve improved
guidance for prioritizing control and resource allocation.
While the degree of clustering in BC has remained stable
over time, clustering does differ between FB and CB
individuals. The usefulness of the cluster analysis may,
PC-961-05 Creation of geospatial datasets for

the TB control programme and linking it with
the patient management information system
A Khanna,1 D Kundu,2 K Rade,2 T Bhowmick3
1
Directorate of Health Services, Delhi, 2World Health
Organization Country Office for India, Delhi, 3Geospatial
Delhi Limited, a Government. of NCT of Delhi Company,
Delhi, India. e-mail: debashishkundu@yahoo.com
S334
PC-962-05 Spatial and temporal TB notification

rates among districts in Malawi: how should
we interpret this?
B G Belaineh,1,2 I Dambe,1 K Mbendera,1 F Kassa,1
A A Mwenye,1 J Mpunga1 1National Tuberculosis
Programme, Ministry of Health, Lilongwe, 2International
Training and Education Center for Health Malawi, Lilongwe,
Malawi. e-mail: belaineh@gmail.com
Background: According to the 2014 national tuberculosis prvalence survey in Malawi , the prevalence of TB was
451/100 000.With the current trend in TB notification
combined with high burden of TB in spite of expansion
of DOTs ,dwelling on passive appproach only will
maintain the status quo. Hence it is appropriate to
implement active case finding in high burden geographicall defined areas and high risk groups.For this,
prevalence survey would not help much as it is not
designed to provide subnational estimate. Prevalence
survey at sub national level requires substantial resources. Hence, use of routine data for decision making seems
to be a readily available source .The purpose of this
analysis is to show the variation in Tuberculosis
notification rate across districts and to review trend
and pattern of distribution on notification.
Methods Districts routinely collect data and compile
notificatrion data every quarter. The routine data are
aggregated in the data management unit of National
Tuberculosis. Program .TB case notification date for
2002, 2008, and 2014 were used.TB case notiifcation
rates were cacluated for all forms of TB of each
respective year and projected mid year populaton of the
districts was used. Epi info 3.5.4 was used to do the
mapping using shape filesfor Malawi. Result: The overall
notification rate was 104/100 000 in 2014. In 2002 the
notification rate was 202 and declined to 181/100 000 by
2008 (12 %) and in year 2014 rate of decline was 42%
from 2008. Districts with relatively high rate of
notification were able to maintain their rank among all
other districts. Muanza, Blantyre, Nsanje, Chiradzulu
and Lilongwe had the highest notification rates consistently. Dowa, Chitipa and Nchisi had lower notification
rate consistently (see Figure).
Conclusion: Routinely collected data shows that TB case

notification rates are heterogeneous. Although the trend
in notification rate declined over the last years, those
districts with relatively high rate of TB notification tend
to continue with relative high rate when compared others
while those with low notification tend to maintain their
low level despite an increased coverage of DOTs service.

Hence, the attribution to health system variables is not
likely to explain all the difference. The notification data
provides reliable information that would help to
differentiate districts by disease burden and administer
package of intervention that ensure efficiency and
effectiveness.
PC-963-05 Early success in community contact

investigations within a 50 metre radius of an
index tuberculosis case in urban Pakistan: a
before and after intervention
R K Fatima,1 E Qadeer,1 A Yaqoob,1 A Kumar,2
S Majumdar,3 H Shewade,2 M Haq1 1National Tuberculosis
Control Programme, Pakistan, Islamabad, Pakistan;
2
South-East Asia Regional Office, New Delhi, India; 3Centre for
International Health, Burnet Institute, Vic, Australia. Fax:
(92) 51 843 081. e-mail: drraziafatima@gmail.com
Background: Evidence for effect of active case finding on

TB epidemiology is weak as there is dearth of studies
with before and after comparison with control group as
well as long term follow up. National TB Program of
Pakistan adopted a unique and innovative intervention to
expand contact investigation beyond household.
Design/Methods: To determine the impact of contact
investigation beyond household on TB case detection and
case notification rate (CNR) in NTP of Pakistan. Before
and after intervention study with control districts were
done, in which selected districts were predominantly
urban with high concentration of slums as intervention
(n 4) and control (n 4) areas which had comparable
CNR at baseline. July 2012 June 2013 and July 2013 June 2014 was the pre-intervention and intervention
period respectively. While passive case finding and
household contact investigation was routinely done in
all districts, contact investigation beyond household was
done in intervention districts: all people staying within a
radius of 50 metres (using GIS) from the household of
smear positive TB case were screened for tuberculosis
and those with presumptive TB were investigated using
smear microscopy and as a part of intervention, first time
in Pakistan, Xpert MTB/RIF test was applied for smear
negatives. All the diagnosed TB cases were linked to TB
treatment and care.
Results: In intervention districts, out of total 14 086
infective TB cases notified during intervention period,
community contact investigation was done in 5469
(38.8%) cases. Of total bacteriologically confirmed TB
notified, 9.5% were detected by the intervention, of
which 73% were by contact investigation beyond
household. TB CNR per 100 000 populations attributable to contact investigation beyond household (difference in change of TB CNR in intervention and control
districts during pre-intervention and intervention period)
was 25 and 13 for all forms of TB and bacteriological
positive TB respectively.
Conclusion: Community contact investigation beyond
household not only detected additional TB cases but also
increased TB CNR over a period of one year. However,
further long term assessments of the same are required to

ascertain sustainment of increase in CNR.
S335
py; those found positive are registered. However, those

found negative are subject to Gene Xpert. Gene Xpert
positive cases with or without rifampicin resistance are
also registered for treatment at the respective sites. TB
presumptive ageds less than 15 years identified are
referred to child TB managing sites for diagnosis and
treatment and are followed up.
Results: Until 7 April 2015, a total of 617 642
individuals were screened, among them 74 329 are
household and 543 313 are community based, total TB
presumptive found 19 257 (household4326,
community14 931), in which smear positive cases are
2988 (community2254, Household734), gene expert
positive cases are 271, MDR cases are 39, child referred
are 4875 and total registered cases are 3406.
Conclusion: Contact screening among community is
proved to be effective in addition to household screening
for early case detection and to limit spread of disease.
PC-965-05 Spatial distribution and cluster

analysis of extensively drug-resistant
tuberculosis in KwaZulu-Natal, South Africa
PC-964-05 Innovative GIS based screening for

tuberculosis contacts in household and
community contacts in Pakistan: how
effective?
1
1 1
R K Fatima, E Qadeer, M Haq, A Yaqoob RESEARCH,

National TB Control Programme, Pakistan, Islamabad,
Pakistan. Fax: (92) 51 843 8081. e-mail:
drraziafatima@gmail.com
Background: Pakistan ranks fourth globally among the

22 high TB burden countries with 58% of case detection
rate and It accounts for 61% of the TB burden in the
Eastern Mediterranean Region (WHO). There is a need
to enhance TB case finding to meet MDG targets.
Design/Methods: Active case finding through contact
screening using GIS software in 3 cities in Punjab
Province i.e., Lahore, Faisalabad and Rawalpindi and
the Capital Territory utilizing all SS notified cases as
index cases registered at BMUs. The study aims to
determine the effectiveness of community and household
contact tracing through outreach using GIS. Project
funded through TB Reach wave 3 is focusing on contact
screening is being carried out within household and a
diameter of 100m around each index case using GIS
Technology and therefore strengthen current contact
screening strategy adopted by NTP. Household contacts,
i.e. those normally resident or sharing the same airspace,
are verbally screened initially, following by a widening
circle screening of close community contacts. The
coordinates of the household are entered into a GIS
database via a mobile phone link. There are 50 field
officers trained for this project. All the TB presumptive
found from the screening are subject to smear microsco-
T Kapwata,1 N Morris,1 T Mthiyane,1,2 P Mpangase,1,2

P Moodley,2 J Brust,3 N Gandhi,4 S Shah4,5 1South Africa
Medical Research Council, Durban, 2University of
KwaZulu-Natal, Durban, South Africa; 3Albert Einstein
College of Medicine and Montefiore Medical Center, Bronx,
NY, 4Emory University Rollins School of Public Health, Altanta,
Georgia, 5U.S. Centers for Disease Control and Prevention,
Atlanta, GA, USA. e-mail: saritashah2@gmail.com
Background: KwaZulu-Natal province, South Africa,

has among the highest burden of XDR-TB worldwide
with 1545 cases diagnosed and a prevalence 3.1/100 000
population in 2012. The majority of cases occur due to
transmission. Poor access to health facilities can be a
barrier to timely diagnosis and treatment of TB, which
can contribute to ongoing transmission. The Transmission of HIV-Associated XDR-TB (TRAX) study recruited
adults and children diagnosed with XDR-TB from 2010
2014. Our analysis assessed the geographic distribution
of participants around health facilities.
Design/Methods: We calculated distance and time from
participants home to the closest hospital, clinic and
community health centre [CHC]) using tools within
ArcGIS Network analyst. Speed of travel was assigned to
the various road classes based on KwaZulu-Natal
Department of Transport regulations.
Results were compared to guidelines for the provision of
social facilities in South Africa: 5km to a clinic or CHC
and 30km to a hospital. Cluster analysis was conducted
with SaTScan software using the purely spatial scan
statistic and Poisson probability model, with various
maximum cluster sizes specified (ranging from 2050%
of the dataset).
Results: We included 321 participants for whom geospatial data were available. Cases were distributed
throughout all 11 districts in KwaZulu-Natal province.
The mean participant distance to the closest hospital,
clinic and CHC was 12, 3.5 and 43 kilometres,
respectively. Mean travel times were 16, 4 and 109
S336
minutes, respectively. Participants in rural areas were

more likely to exceed the recommended distance to a
clinic or a hospital (Figure). Preliminary cluster analysis
produced several clusters, the most significant of which
was located in the rural Msinga area of Umzinyathi
District (P , 0.00001). The occurrence of these clusters
appears related to greater distance from a health facility,
mainly clinics (Figure).
Conclusion: Although XDR-TB cases are distributed
through the province, geospatial analysis identified
possible areas of high transmission. The average travel
distance of the participants to hospitals and clinics falls
within the CSIR recommended distance; however additional analysis is needed to confirm the correlation
between distance from health facilities and clustering.
New, rapid molecular tests for drug resistance that are
closer to patients may reduce delays in diagnosis and
treatment, and reduce transmission.
32. The game of numbers: prevalence

and incidence of tuberculosis
PC-966-05 Prevalence and incidence of smearpositive pulmonary tuberculosis in the Hetosa
District of Arsi Zone, Oromia Regional State,
Central Ethiopia
S Hamusse,1,2 M Demissie,3 B Lindtjorn2 1Oromia Regional
Health Bureau, Addis Ababa, Ethiopia; 2Centre for
International Health, University of Bergen, Bergen, Norway;
3
Addis Continental Institute of Public Health, Addis Ababa,
Ethiopia. e-mail: shalohamusse@gmail.com
Background: Understanding the local epdiemiology of

tuberculosis (TB) is important to design targeted
intervetnions tailored to the specific needs of the districts,
as there is huge heterogeniety among districts in terms of
TB burden. Community-based survey to estimate the
prevalence and incidence of TB could serve as an

alternative option. Our objective was to estimate the
prevalence and incidence of TB at a district in Oromia
region of Ethiopia.
Methods: We used a population-based cross-sectional
survey design at baseline which was followed by
prospective cohort study deign. We included all individuals aged 715 years, permanent residents and temporary
visitors who arrived 15 days before the study time in as
participants. Face to face interviews using standardized
TB screening questionnaire were used to identify
participants with persistent cough of more than two
weeks. Further, spot and morning sputum samples were
collected for standard sputum smear microscopy and
culture.
Results: We interviewed 33 073 individuals at baseline
survey and 32 800 apparently healthy individuals aged
715 years in the follow up study, and identified 1042
and 468 participants with chronic cough at baseline and
during follow up respectively who provided two sputum
samples for acid fast bacilli (AFB). Of these, we identified
36 smear and 43 culture positive individuals in the
baseline survey, and 70 smear positive individuals during
a follow up period of over one year. The prevalence of
smear and culture postive TB was 109 and 130 per 105
population respectively, while the incidence was 213/105
person-year. Forty-five PTB cases were diagnosed
through passive case finding while 36 were identified
through active case finding, implying that 44.4% of
PTB cases were undiagnosed and untreated in the
community. Family history of TB contact was independently associated with high risk of TB infection both at
baseline (AOR, 3.14, 95%CI: 1.47-6.70) and during
follow up (aIRR 9.27, 95%CI: 7.38-11.30).
Conclusion: The burden of bacteriologically confirmed
TB was high in the study area, and about 44.4% of the
batceriologically confirmed cases were undiagnosed
through the passive approach. Family history of TB
contact was an independent risk factor for TB infection.
This demonstrates the need to improve TB case findings
and intensify contact tracing among household members
of TB to detect and treat the undetected PTB cases at
household level.
PC-967-05 Prevalence of Mycobacterium

africanum in Ghana over an 8-year period
D Yeboah-Manu,1 A Asante-Poku,1 I D Otchere,1
S Osei-Wusu,1 E Sarpong,1 S Gagneux2 1Noguchi
Memorial Institute for Medical Research, University of Ghana,
Legon, Accra, Ghana; 2Swiss Tropical and Public Health
Institute, Basel, Switzerland. Fax: (233) 3021 502 182.
e-mail: dyeboah-manu@noguchi.ug.edu.gh
Background: Mycobacterium tuberculosis complex (M.

tuberculosis c) has two species adapted to humans even
though there are occasional cross-species infections
especially with M. bovis. M. tuberculosis sensu stricto
(M. tuberculosis ss) is perceived to have competitive
advantage over M. africanum (M. afric) and may
gradually out-compete M. afric. Our objective was to
analyze the distribution of these species among TB cases

over 8-year period.
Design/Methods: Mycobacterial isolates obtained from
sputum were confirmed as M. tuberculosis c by IS6110
PCR and further characterized by spoligotyping and
LSPs (RDs 9, 4, 12,702,711) assays. The distribution of
the species isolated was stratified both in space and in
time for analyses.
Results: Among 1926 M. tuberculosis c isolates obtained
from 2007-2014, 1516 (78.7%), 400 (20.8%) and 10
(0.8%) were identified as M. tuberculosis ss, M. afric and
M. bovis respectively. The proportion of M. afric in the
same period were 22.2%, 15.4%,10.7%, 23.1%,
11.5%, 24.6%, 23.0% and 18.5% respectively. M. bovis
was isolated in 2012 to 2014 as 1.0%, 0.4% and 2.0%
respectively when Northern Ghana (NG) was included in
the study. Comparing the isolates from 2012 to 2014, we
found M. bovis to be significantly higher in NG (2.6% of
151) than Southern Ghana (SG; 0.5%) (P 0.020).
Among the M. afric isolates, 58.7% and 41.3% of the SG
as compared to 48.1% and 51.9% of the NG isolates
were lineage 5 and lineage 6 respectively. The most
dominant spoligotype of M. tuberculosis ss in SG was SIT
61 (42.9%) as compared to SIT 53 (28.7%) in NG.
Lineage 6 spoligotype SIT 181 and 326 were the most
prevalent in south where as SIT 181 (6.7%) dominate the
north.
Conclusion: M. africanum contributes significantly to TB
in Ghana as its prevalence over the years remains fairly
constant.
PC-968-05 Findings from the First Malawi TB

Prevalence Survey
1
1 1
R Banda, A Munthali, J Mpunga Ministry of Health,

Lilongwe, 2University of Malawi, Zomba, Malawi. e-mail:
rhobanda@gmail.com
Background: Malawi has for a long, time used estimates

provided by WHO to determine its TB burden and
monitor TB control progress. It was among countries
recommended by WHO to carry out a TB prevalence
survey by 2015 to better estimate the burden TB. The
first Malawi nationwide Tuberculosis prevalence survey
was conducted in Malawi (2013- 2014). The overall goal
was to determine the prevalence of pulmonary tuberculosis (PTB) among persons aged 15 years and older.
Design/Methods: A multistage cluster-sampled crosssectional survey was implemented nationwide. All
eligible persons aged at least 15 years were screened for
TB symptoms the key one of which was a cough of two
weeks or more. A chest X-ray (CXR) was then done on
all clients in the field. Persons who screened positive for
symptoms and/or chest X-ray were asked to submit spot
and morning (2) sputum specimens for smear microscopy
and culture. These were examined at the National
Tuberculosis Reference Laboratory (NTRL). Microscopy
was done using fluorescent technique and a Gene xpert
performed on all smear positive samples. A bacteriologically-confirmed case was defined as an eligible partic-
S337
ipant with M. tuberculosis confirmation by gene-Xpert

or culture.
Results: Of the 60 944 individuals enumerated in the
census, 39 020 were eligible to participate in the survey
from which a sample of 31 579 (80.9%) was drawn.
Chest X-ray was performed among 31 561 participants.
A total of 3432 (10.9%) participants were eligible for
sputum examination. The provisional TB prevalence of
bacteriologically-confirmed TB in those 15 years and
above is estimated to be 451/100 000 (95%CI 308-594).
The provisional adjusted prevalence for all age groups is
286/100 000 which 2 time higher than WHO estimate in
2014. Prevalence increased with age, was higher in men
at 543/100 000 (95%CI 327-759) than in women at 374/
100 000 (95%CI 245-504). Also, the results indicate that
TB prevalence is high among urban residents and in
adults aged .55 compared to prevalence in the general
adult population (451/100 000).
Conclusion: There is need for the Malawi NTP to
strengthen its TB control strategies more specifically
towards improving case detection and paying special
attention to the selected population groups that indicate
a high risk of TB.
PC-969-05 Life-time tuberculosis incidence in

Cape Town exceeds that in New York City in the
early 20th century and continues to increase
S Hermans,1,2 R Wood2 1Amsterdam Institute for Global
Health and Development, Amsterdam, Netherlands;
2
Desmond Tutu HIV Centre, Institute of Infectious Disease
and Molecular Medicine, Cape Town, South Africa. e-mail:
sabine.hermans@hiv-research.org.za
Background: Cape Town reports the second highest TB

burden in South Africa, third of the TB high burden
countries. At the start of the 20th century, Cape Town
and New York City had similar TB notification rates.
Despite employing similar TB control strategies, TB rates
in Cape Town are currently almost 100-fold those of
New York. High quality long term historical data on TB
notifications from Cape Town and New York allowed
for a comparison between the life-time TB risks of birth
cohorts in the two settings, to enable insight into the
cumulative impact of TB control strategies and the
advent of the HIV epidemic.
Design/Methods: Annual age-stratified TB notifications
and mid-year population estimates for 1930-2010 (Cape
Town) and 1910-2010 (New York) were used to
calculate age-stratified TB incidence rates per 100 000
population. Age cohort analysis techniques were used to
calculate cumulative incidence rates by decade.
Results: Over the last century, cumulative life-time TB
incidence in Cape Town and New York showed
contrasting trajectories: a continuous decline in New
York versus a continuous increase in Cape Town
(Figure). In 1930 life-time TB risk was 8 fold higher in
Cape Town (25.4%) than New York (3.2%). Since then,
cumulative TB incidence has been rising steadily in Cape
Town with more TB disease occurring at an increasingly
younger age. Age-specific cumulative incidence remained
S338
high before the advent of the HIV epidemic, after which it

increased by 30-60% from adolescence onward (ages
affected by generalized HIV epidemic indicated by
marker).
Conclusion: The life-time TB risk in Cape Town prior to
HIV remained very high and much greater than in New
York City during the same time. The life-time risk of over
25% rather than 10% suggests multiple infections in this
setting. The HIV epidemic led to further increases in the
cumulative TB incidence which, in addition to faster
disease progression among those infected with HIV,
could also be due to growing numbers of retreatment
disease or to an increasing force of infection. TB control
in Cape Town will only be achieved by addressing the
drivers underlying the persistent TB epidemic in addition
to tackling HIV.
PC-970-05 Prevalence and incidence of

tuberculosis among presumptive TB patients in
Bagamoyo, Tanzania: a prospective cohort
study
F Mhimbira,1,2,3 K Ibrahim,1,2,3 J Hella,1,2,3 F E Haraka,3
E Mfinanga,3 L Jugheli,1,2,3 F Lwilla,3 K Reither1,2,3 1Swiss
Tropical and Public Health Institute, Basel, 2University of
Basel, Basel, Switzerland; 3Ifakara Health Institute,
Bagamoyo, Tanzania. e-mail: fmhimbira@gmail.com
Background: The burden of tuberculosis (TB) has

escalated with the emergence of HIV in Sub-Saharan
Africa. Identification of presumptive TB patients and
subsequent early TB diagnosis and treatment is the
hallmark of TB control. The aim of the study was to
determine the prevalence and incidence of TB among
presumptive TB patients followed for 18 months.
Methods: Presumptive TB patients presenting with
cough and other TB related symptoms were enrolled
and followed-up at Ifakara Health Institute TB research
from clinic September 2011 to September 2013. The

patients were referred from Bagamoyo District Hospital
and a diagnostic cohort study. Time of diagnosis refers to
the date of microbiological confirmation by either
positive smear microscopy or mycobacterium culture or
molecular genotyping. Prevalence was defined as TB
diagnosed at recruitment or within the following three
months. Incidence was defined by any TB diagnosed after
three months until 18 months of follow up. Descriptive,
logistic regression models, poisson and cox proportional
hazards are used.
Results: We analysed 764 presumptive TB patients with
median age of 39 years (Interquartile range [IQR]: 3150), females were 394 (51.6%) and 348 (45.5%) were
HIV infected. HIV infection was higher in women than
men (216 [54.8%] vs. 132 [35.7%], P value , 0.001).
Apart from cough the commonest reported symptoms
were fever in 518 (67.8%) patients and un-intentional
weight loss 463(60.6%). Prevalent TB was in 152 (20%)
among presumptive TB patients. Prevalent TB was
associated with unit increase in age in years [adjusted
Odds Ratio [aOR] 0.98, 95% Confidence Interval
[95%CI]: 0.97-0.99, P value 0.015) and Body Mass
Index (BMI, kg/m2) (aOR 0.92. 95%CI: 0.87-0.97, P
value 0.003). Excessive night sweat and un-intentional
weight significantly predicted TB at recruitment. Additional 19 TB patients were diagnosed between 3 and 18
months of follow-up. The incident rate was 320 per
person-month. The incident rate in HIV positives and
HIV negative was 338 per person-month and 148 per
person-month respectively. The hazard ratio for incident
TB for HIV (Hazard Ratio [HR]: 2.60, 0.92-7.40, P
value 0.071).
Conclusion: The TB prevalence among presumptive TB
patients was relatively high in the described setting.
Furthermore, a substantial number of patients developed
TB between three and eighteen months of follow-up,
which needs to be considered in follow up strategies for
presumptive TB patients.
PC-971-05 Prevalent tuberculosis detected by

active case finding among adults in the
community in Ca Mau, Viet Nam
T A Nguyen,1 P Nguyen,1 N V Nhung,2 B H Nguyen,2
K H Tran,3 J Ho,1,4 G J Fox,1 G Marks1,4 1Woolcock Institute
of Medical Research, Ha Noi, 2National Tuberculosis
Programme, Ha Noi, 3Ca Mau Centre for Social Disease
Prevention, Ca Mau, Viet Nam, 4University of New South
Wales, Sydney, NSW, Australia. e-mail:
thuanh.nguyen@sydney.edu.au
Background: Interventions that improve detection and

treatment of people with tuberculosis (TB) could be an
important part of strategies to reduce the prevalence of
TB in high-burden settings. We are conducting a cluster
randomized trial of community-wide active case finding
for TB among adults in the rural province of Ca Mau in
Viet Nam. Here, we aimed to determine the prevalence of
Xpert positive tuberculosis in the general community,
using of active case finding in the intervention clusters
during the first year of the study.
Design/Methods: All eligible and consented individuals
aged 15 and above in 57 randomly selected subcommunes in Ca Mau province were asked to produce
a single spontaneous sputum sample for Xpert MTB/RIF
testing, regardless of symptoms. Participants whose
sputum was Xpert MTB positive were invited to have a
plain chest X-ray and to submit two additional sputum
samples for confirmatory liquid culture.
Results: The eligible adult population in the selected subcommunes was 46,694. Of these, 39,403 (84%) agreed
to be interviewed (screened) and 21 155 (45% of the
eligible population and 54% of those interviewed)
produced a sputum specimen that was suitable for Xpert
testing. Overall, the sputum of 169 people was positive
for Xpert MTB. The crude prevalence was 429/100 000
adults screened (95%CI: 369 - 499). Among those with a
positive Xpert MTB test result, 54% had very low
levels of M. tuberculosis detected. Among 140 Xpert
MTB positive participants with at least one finalized
mycobacterial culture, 83 (59%) were culture positive
for M. tuberculosis. Cultures for a further 22 (16%) grew
non-tuberculous mycobacteria or were contaminated
and the remaining 35 (25%) were culture negative.
Among the Xpert MTB positive / culture negative
participants, 21 (60%) had abnormal chest X-rays.
Conclusion: The prevalence of pulmonary TB detected
by screening using Xpert MTB/RIF test in the general
population was high. The discrepancy between Xpert
and culture results requires further study. Support:
NHMRC Project grant (#1045236).
S339
PC-972-05 Gender ratios in tuberculosis

prevalence and patient diagnostic rates in lowand middle-income countries: a systematic
review and meta-analysis
K Horton,1 P Macpherson,2,3 R Houben,1 R White,1
L Corbett1,4 1London School of Hygiene & Tropical Medicine,
London, 2University of Liverpool, Liverpool, 3Liverpool School
of Tropical Medicine, Liverpool, UK;
4
Malawi-Liverpool-Wellcome Trust Clinical Research
Programme, Blantyre, Malawi. e-mail:
katherine.horton@lshtm.ac.uk
Background: In 2012, the male-to-female (M:F) case

notification ratio for smear-positive pulmonary tuberculosis (TB) was 1.9:1. However, a substantial proportion
of TB cases remain undiagnosed or unreported, especially in high burden and low resource settings. This study
set out to systematically investigate whether observed
M:F case notification ratios reflect true population
gender differences in disease prevalence and diagnostic
rates.
Design/Methods: Studies describing TB prevalence surveys in nationally-representative and sub-national adult
populations (age 715 years) in low- and middle-income
countries published between January 1993 and August
2014 were identified through searches of PubMed,
Embase, Global Health, the Cochrane Database of
Systematic Reviews and the WHO Global TB Report
2014. Gender-specific prevalence of bacteriologicallyconfirmed (culture- and/or smear-positive) TB was
calculated or extracted, and M:F ratios were compared
across regions, TB prevalence and HIV prevalence.
Patient diagnostic rate (PDR) was calculated as smearpositive TB case notifications per 100 000 divided by
bacteriologically-confirmed TB prevalence per 100 000,
matching on country and study year(s). Risk of bias was
assessed and study quality was ranked using the GRADE
system. Heterogeneity was assessed, and random-effects
meta-analyses were performed.
Results: Sixty-eight surveys with 2.9 million participants
in 30 countries were identified (31 in Africa, 20 in
Southeast Asia, 11 in the Western Pacific, three each in
the Americas and the Eastern Mediterranean), with 24
nationally representative surveys and 44 at a subnational level. Among 37 surveys with available data,
the overall random-effects weighted M:F prevalence
ratio was 2.19 (95%CI: 1.80-2.66; range 0.31 to
11.16). There was consistent male excess in TB
prevalence in surveys from all regions except the
Americas. The excess of male prevalent cases remained
when stratified by TB prevalence and HIV prevalence
and in studies with low (n 14) or moderate (n 15) risk
of bias. M:F ratios in PDR ranged from 0.39 to 2.92 and
were less than one in 23 of 30 surveys with available
data.
Conclusions: TB prevalence is significantly higher among
men than women in low- and middle-income countries.
Gender ratios in PDR indicate that men are disadvantaged in accessing TB care.
S340
33. Molecules, polymorphisms and

resistance to TB
PC-973-05 Standardized mycobacterial
interspersed repetitive units-variable number
of tandem repeat genotyping of
Mycobacterium tuberculosis in Taiwan
P-C Chuang,1 R Jou1,2 1Centers for Disease Control, Taipei,
2
Background and aim: The IS6110 restriction fragment

length polymorphism (RFLP) was adopted as a standard
genotyping method for outbreak investigations of
tuberculosis (TB) in Taiwan. It is time-consuming and
labor-intensive. Even though 24-loci mycobacterial
interspersed repetitive units-variable number of tandem
repeat (MIRU-VNTR) typing were recommended for
genotyping of Mycobacterium tuberculosis complex (M.
tuberculosis c), the discriminatory power for the
prevalent Beijing genotypes in Taiwan is un-satisfactory.
In this study, we optimized loci combination and
standardized the MIRU-VNTR to improve the turnaround time and genotyping capacity.
Methods: Twenty-nine MIRU-VNTR loci were evaluated using 457 M. tuberculosis c isolates collected from
January 2013 to June 2014. Of the 457 isolates, 83
(18.2%) were confirmed outbreak isolates belonging to
32 episodes based on results of the IS6110 RFLP
fingerprinting. The optimal loci combination of MIRUVNTR was evaluated and the discriminative power was
assessed by parallel comparisons with results of the
IS6110 RFLP. Spacer oligonucleotide typing (spoligotyping) was performed with a commercial kit.
Results: The MIRU(10) containing 10 loci, QUB3232,
VNTR3820, QUB2163b, QUB18, Mtub21, MIRU26,
QUB-26, VNTR4120, Mtub04, and MIRU39, were
suggested as an optimal combination. Together with
spoligotyping, the MIRU(10) typing revealed results with
the highest discriminative power that is compatible with
the IS6110 RFLP fingerprinting. However, the RFLP
fingerprints of 5 East-African-Indian (EAI) as well as 5
Beijing genotypes were not able to be distinguished even
using 29-loci MIRU-VNTR. Our results demonstrated
that not only Beijing but also EAI strains had high
similarity of MIRU genotypes in Taiwan.
Conclusion: The MIRU(10) typing combined with the
spoligotyping was proposed as a new standard genotyping of M. tuberculosis in Taiwan. Nevertheless, through
epidemiological survey is strongly suggested while the
same Beijing and EAI spoligotypes were obtained in
outbreak investigations.
PC-974-05 Polymorphisms in the STAT4 gene

and tuberculosis susceptibility in Chinese Han
people
Q Wu,1 Y Ji,1 C Wu,1 Y Wang,1 Q He1 1Sichuan University,
Chengdu, China. e-mail: 76201313@qq.com
Background: The association between genetic factors

and the development of tuberculosis (TB) has obtained
much attention. The signal transducer and activator of
transcription 4 (STAT4) gene encodes a special transcriptional factor that transmits signals induced by
several cytokines which play critical roles in the
development of autoimmune and chronic inflammation
diseases. It can transduces signals induced by cytokines
comprising I interferon (IFN), interleukin-12 (IL-12),
and interleukin-23 (IL-23) in monocytes and T cells. IL12/IFN-c axis plays a key role for the elimination and
control of M. tuberculosis. Here, we performed a
correlation analysis between STAT4 polymorphisms
and TB.
Methods: A total of 636 TB cases and 608 healthy
controls were studied to examine the presence of three
SNPs (rs7574865, rs4853542, and rs1031509) in the
STAT4 locus. The STAT4 SNPs genotyping was performed with the Sequenom MassARRAY SNP genotyping platform (Sequenom, San Diego, CA, USA) . Odds
ratiods (OR) and 95% confidence intervals (CIs) were
calculated using the multivariable logistic regression
analysis. The online software SHEsis program was
utilized to analyze genotyping data.
Results: Out of three SNPs tested in the sample, one
STAT4 SNP (rs4853542) was related to TB. Compared
with the genotype AA, individuals carrying AG had a
significantly increased risk of TB (P 0.02, OR1.15,
95%CI: 1.02-1.30). An allele model revealed that G
allele had a 33% increased risk of TB compare with A
allele (P 0.01, OR1.33, 95%CI: 1.06-1.66). We found
the haplotype GAT significantly referred to protection
against TB (OR0.562, 95%CI: 0.33-0.96, P 0.03).
Conclusions: In the present study, the rs4853542
polymorphism of STAT4 was suggested to increase risk
of TB in Chinese Han people. A study with larger
numbers of subjects will be required to confirm our
results.
PC-975-05 Genetic polymorphisms of

glutathione S-transferase P1 and the incidence
of anti-tuberculosis drug-induced hepatitis in a
Chinese cohort
J He,1 Q Wu,1 Y Wang,1 C Wu,1 Y Ji1 1Quan Wu, Chengdu,
China. e-mail: 18980602293@qq.com
Background: Anti-tuberculosis drug induced hepatotoxicity (ATDH) is one of the most common adverse effects
associated with TB therapy. The GST genes code for a
superfamily of enzymes that participant in the phase-II
drug metabolism. Several studies in animal models
demonstrated the important role of GST in the prevention of chemically induced hepatotoxicity. This study
aimed at investigating the relationship between variants
and haplotypes of GSTP1 in TB patients among Chinese

population.
Design/Methods: In a prospective study, 247 tuberculosis patients were followed up for three months after
initiating anti-tuberculosis therapy. ATDH was defined
as an ALT more than twice the upper limit of normal or a
combined increase in AST and total bilirubin, and one of
them was more than two-times over the upper limit of
normal AST or bilirubin value. GSTP1 SNPs were
genotyped with the MassARRAY platform. The associations were analyzed by the v2 test and also binary
logistic regression analysis adjusting for confounding
factors, such as age, sex, body mass index (BMI) and
drinking history. The online software SHEsis program
was utilized to analyze genotyping data.
Results: A total of 247 newly diagnosed TB patients were
recruited in this study. Twenty-four cases of ATDH were
diagnosed and were considered as experimental group
and the rest of subjects (223 cases) were considered as the
control group during 3 months follow-up. After correction for potential confounding factors, significant differences were found for rs1695 under an allele model
(OR4.436, 95%CI: 1.159-16.984; P 0.03) and a
recessive model (OR0.24, 95%CI: 0.064-0.898; P
0.034). Week significant differences were also found for
rs4147581 under an allele model (OR2.346, 95%CI:
1.011-5.442; P 0.047). We did not find significant
difference in haplotype distribution between the experimental group and the control group in our study.
Conclusion: Our results support the GSTP1 polymorphisms as predictors of ATDH in Chinese Han population treated with anti-tuberculosis drugs.
PC-976-05 Key population for tuberculosis

intensified case finding identified through
national TB prevalence survey
D B Lolong,1 O S Simamarta,1 B Dwihardiani,2 M Farid,3
l Pangaribuan,1 A V Siagian1 1Ministry of Health Republic
Indonesia, Jakarta Pusat, 2World Health Organization,
Indonesia Country Office, Jakarta Pusat, 3Tuberculosis
Operational Research Group, Jakarta Pusat, Indonesia.
e-mail: bintari_dwihardiani@yahoo.com
Background: Tuberculosis (TB) is a major public health

problem in Indonesia. The gap between the prevalence of
bacteriologically confirmed TB found during National
TB Prevalence Survey 2013 (NPS) and current case
notification is still very large. Only a small proportion of
confirmed cases were under treatment. Intensified case
finding (ICF) through active screening is one of the
proposed strategies to reduce TB incidence and address
the detection and notification gaps. Identification of key
population groups is needed to improve the effectiveness
of ICF.
Design/Methods: Analysis of data obtained from NPS
2013 was conducted. Demographic information, TB
history, TB contact history, smoking, were analyzed to
find the association with TB cases found from NPS using
logistic regression. Prevalence by sex, age group, region,
and region and urban/rural stratification was calculated
S341
according to WHO standard method to calculate

prevalence
Results: Nationwide survey with 67 946 participants was
conducted from April 2013 to June 2014. There were 15
467 (23%) participants with chronic cough or haemoptysis or TB suggested thorax x-ray photo. There were 426
bacteriologically TB confirmed cases detected. Male sex,
old age (. 55 years old), previous TB treatment history,
existing TB contacts were significantly associated with
higher TB confirmed cases. However, estimated prevalent cases are still high among productive age. There is no
significant difference of TB prevalence among regions
and urban-rural living place. Participants with previous
TB treatment history and existed TB contacts had 6.1
(95%CI 4.7-7.8) and 1.8 (95%CI 1.2-2.7) times higher
odds to get TB respectively compared to those without
the factors.
Conclusion: Concentrating ICF in the group of people
with previous TB history, having TB contacts, and in old
age and male population may give significant yield for
case finding in Indonesia. Further evaluation of their
yields, cost effectiveness, and feasibility are needed
before incorporating this into policy. As TB burden is
still high in productive age population, overall TB
control strategies should be improved to treat more cases
earlier.
PC-977-05 Trends of fluoroquinolone resistance

in Mycobacterium tuberculosis mirror
resistance trends among community
pathogens
S Shakoor,1 J Farooqi,1 K Jabeen,1 F Malik,1 R Hasan1
1
Aga Khan University, Karachi, Pakistan. e-mail:
sadia.shakoor@aku.edu
Background: Fluoroquinolone (FQ) resistance is widely

recognized among M. tuberculosis strains from high
burden countries and recent literature shows that it
impacts development of treatment regimens. Increase in
M. tuberculosis FQ resistance trends is also documented
and is related to the unchecked and inappropriate use of
FQs in the population. We hypothesized that the
development and spread of resistance in other community acquired (CA) bacteria can be used to predict
resistance among M. tuberculosis strains in high burden
regions. As a first step we have correlated resistance
trends among these CA-bacteria with those in M.
tuberculosis isolated at a clinical laboratory in Karachi,
Pakistan.
Design/Methods: Laboratory records were retrieved and
duplicate removed for FQ resistance reported among
pathogens of interest for the years 2010-2014. These
comprised community (and not hospital or long-term
facility) culture isolates of S. pneumoniae, and H.
influenzae (HI) among respiratory pathogens and Salmonella Typhi and Paratyphi, Shigella spp, and Vibrio
cholerae among diarrheal pathogens, where the use of
FQ is common. Data was analysed in MS Excel and bar
charts were created. M. tuberculosis data was analysed
for FQ resistance in the same years and proportion of
S342
resistant M. tuberculosis in each year was mapped

graphically to visually contrast resistance trends among
CA bacteria with M. tuberculosis. Coefficient of regression (R2) values and v2 values for trends were calculated
in MS Excel.
Results: FQ resistance in HI increased consistently from
9.1% in 2010 to 23.4% in 2014 (linear R20.9071). FQ
resistance in Shigella spp increased from no detectable
resistance in 2010 to 20.9% FQ resistance (linear R2
0.9066). M. tuberculosis FQ resistance also increased
from 27.2% in 2010 to 33.3% in 2014. Linear regression
explained only a small proportion of changes in these
trends (R20.26). However, the increase in resistance is
significant (P , 0.0001; v2 for trend 95.2423). On
visual examination of trends, the trend of resistance
development mirrors more closely that in HI, a respiratory pathogen (Figure). HI resistance trends predicted FQ
resistance in M. tuberculosis (v2 for linear trends
2.7808; P 0.09; non-zero slope).
Conclusion: Variance in M. tuberculosis FQ resistance
may be affected by changes in testing methods and FQs
tested. The evolution of resistance in M. tuberculosis is
slower than seen in Shigella spp, but resembles development of resistance in HI. Similarity between resistance
trends in M. tuberculosis and HI suggests common
driving factors; possibly use of FQ in undiagnosed M.
tuberculosis patients. Rapid development of resistance in
these community pathogens is alarming. Increasing FQ
consumption data calls for an urgent public health action
to enforce responsible use of FQs in Pakistan.
PC-978-05 Key genotypes of M. tuberculosis

responsible for the spread of multi- and
extensively drug-resistant disease in the Sakha
Republic, Yakutia
growing threat of multidrug-resistant (MDR) and

extensively drug-resistant (XDR) TB, study of the
population structure of M.tuberculosis (common for
Yakutia becomes especially important.
Design/Methods: M.tuberculosis isolates from 343
patients with TB were studied by 24-locus MIRU-VNTR
genotyping, using MIRU-VNTRplus and SITVIT open
databases for strain identification. Drug resistance (DR)
was determined using automated BACTEC-960 system
based on liquid media, and classical solid culture
method.
Results: The following strain families were identified:
Beijing (158; 46.1%), Orphan (57; 16.6%), S (39;
11.4%), T/H (27; 7.9%), Ural (22; 6.4%), Haarlem
(15; 4.4%), LAM (16; 4.6%), Uganda (9; 2.6%). The 3
families accounting for 74.1% of all the cases were
studied in more detail. Beijing genotype was detected in
158 patients, of them 67.1% were male, 54.6% were
urban citizens, and the majority (77.8%) were young or
middle-aged. Aboriginal residents made 70.7%. DR
cases made 68.3%, monoresistance was detected in
1.3%, polyresistance in 1.9%, MDR in 26.6%, and
XDR in 1.9% of cases. Next in the incidence was the
Orphan genotype. It was detected in 57 patients (57.9%
male, 68.2% urban residents, 43.9% young/middleaged, 52.6% older than 45 years, 2/3 were aboriginal
residents). DR was present in 64.9% of cases, with
monoresistance in 7.0%, polyresistance in 8.8%, MDR
in 12.3%, and XDR in 7.0% of cases. S genotype was
determined in 39 patients (69.2% male, 2/3 urban), who
were mostly (79.5%) young/middle-aged, and predominantly aboriginal (81.2%). DR was determined only in
7.7% of cases, with polyresistance in 7.7%, high
incidence of MDR (74.3%), and XDR in 10.3%.
Conclusion: This was the first study on genotyping
M.tuberculosis in Yakutia, and it showed the predominance of the Beijing, Orphan, and S families, and these
same strains were responsible for the spread of M/XDR
in Yakutia. Notably, severe types of DR were observed
more often in S family (2.9 and 4.3 times more often,
than in Beijing family, respectively). The S family was of
particular interest, as in the past, it was found only in
singular cases in Russia, and its current global incidence
is highly sporadic. In the future, more thorough
molecular epidemiological monitoring of the TB causative agent will be needed.
PC-979-05 WGS reveals genetic heterogeneity

and suggests the role of selective bottleneck in
defining the population structure of
M K Vinokurova,1 N E Evdokimova,1 G I Alekseeva,1

A F Kravchenko,1 E D Savilov,2 O B Ogarkov,2
S Zhdanova2 1Research & Practice Center for Tuberculosis of
the Sakha Republic (Yakutia), Yakutsk, 2Scientific Center for
the Family Health and Human Reproduction Problems,
Irkutsk, Russian Federation. e-mail: mkvin61@mail.ru
M De Vos,1 P Black,1 E Louw,1 R Van Der Merwe,1

S Sampson,1 P Van Helden,1 R M Warren,1 P Arnab2
1
Stellenbosch University, Cape Town, South Africa; 2King
Abdullah University of Science and Technology, Thuwal,
Saudi Arabia. e-mail: margab@sun.ac.za
Background: Sakha Republic (Yakutia) is a northern

territory with an unfavorable epidemiological situation
for tuberculosis (TB). In view of this, and considering the
Background: Whole genome sequencing (WGS) has

revolutionised the interrogation of mycobacterial genomes. Recent studies have reported conflicting findings
on the genomic stability of M. tuberculosis during the

evolution of drug resistance. Some studies have demonstrated genomic stability, while more recent reports have
shown genomic instability with the emergence of
additional genetic variants independent of those observed in drug target genes conferring resistance. The
identification of sub-populations within the context of
WGS is dependent on the read frequency cut-off values
used in the standard variant filtering approach. This
study aimed to define a reliable cut-off for identification
of low frequency sequence variants and to subsequently
investigate genetic heterogeneity and the evolution of
drug resistance in M. tuberculosis.
Design/Methods: Single colony forming units were
selected from 14 M. tuberculosis rifampicin monoresistant clinical isolates. Two of these patients had
follow up isolates which were shown to be multi-drug
resistant (MDR) by routine drug susceptibility testing
(DST). Genomic DNA was isolated from single colonies
for each rifampicin mono-resistant M. tuberculosis
isolate and the primary cultures of the paired M.
tuberculosis isolates demonstrating intra-patient evolution of isoniazid resistance. The whole genomes of the M.
tuberculosis isolates were sequenced using either the
Illumina MiSeq or Illumina HiSeq platforms.
Results and discussion: Our data enabled us to define a
read frequency cut off of 30% to accurately detect
heterogeneous variants. Using this cut-off we demonstrated a high rate of genetic diversity between single
colonies isolated from one population, showing that by
using the current sequencing technology, single colonies
are not a true reflection of the diversity within a whole
population and vice versa. In addition, our read
frequency filtering approach was used to investigate the
evolution of isoniazid resistance in M. tuberculosis
clinical isolates within two patients. We found that
numerous heterogeneous variants emerge and disappear
during the evolution of isoniazid resistance within
individual patients. Our findings suggest that the
isoniazid selective pressure imposed an evolutionary
bottleneck, dramatically reducing genetic diversity within the mycobacterial population. This was then followed
by the subsequent accumulation of new variants in the
mycobacterial population.
PC-980-05 Treatment-based risk factors for

tuberculosis drug resistance in Bolivia
A Flynn,1 J Carlos Mollinedo,2 C L Daley,3 A Faino,3
R Flick,1 E Casso,2 M Camacho De Colque2 1University of
Colorado School of Medicine, Aurora, CO, USA; 2National
Tuberculosis Reference Laboratory, La Paz, Bolivia; 3National
Jewish Health, Denver, CO, USA. e-mail:
andrew.flynn@ucdenver.edu
Background: Understanding the epidemiology of drug

resistant TB (DRTB) in Bolivia is important for
continued progress in TB control. The National TB
Laboratory performs drug susceptibility testing (DST) on
patients with risk factors identified by the National TB
Programme (NTP).
S343
Methods: This is a retrospective analysis of patients

diagnosed with DRTB between 2006 and 2013 with risk
factors for DRTB. Data collected by the Departmental
and National TB Laboratories were evaluated for drug
resistance by risk factor: 1) previous treatmentmore
than thirty days of Category I therapy, 2) lack of clinical
response or sputum conversion at two months of
Category I therapy, or 3) treatment failureCategory I
therapy completion with persistent positive sputum
smear. Odds ratios were calculated using SAS version
9.4.
Results: 3791 culture-confirmed patients with risk
factors for DRTB received DST from 2006-2013:
treatment failures exhibited higher MDR incidence than
patients with lack of therapy response at two months
(35% vs. 18%, OR2.3, CI1.7-3.0) and those with lack
of response at two months exhibited higher MDR
incidence than patients treated previously (18% vs.
8%, OR2.7, CI2.2-3.4). Neither isoniazid nor rifampicin monoresistance were significantly associated with
treatment risk factors.
Conclusions: NTP-defined risk factors impart significant
risk for MDR in Bolivia, suggesting risk factor criteria
appropriately select patients with higher probability of
this resistance pattern. Treatment failure was the greatest
risk factor for MDR.
34. Thinking broadly to advance the END

TB Strategy
PC-981-05 Tuberculosis drug kit
implementation eases drug supply
management in two regions of Ethiopia
M Weldeselassie,1 J Degu,1 H Tesfaye Molamo,2
Y Yiegezu Zegiorgis,3 Y Anteneh Kassie,1 S Negash,1
M Aseresa Melese,1 P G Suarez4 1Management Sciences
for Health, Help Ethiopia Address the Low Tuberculosis
Performance (HEAL TB) Project, Addis Ababa, 2Federal
Ministry of Health, Addis Ababa, 3Pharmaceuticals Fund &
Supply Agency, Addis Ababa, Ethiopia: 4Management
Sciences for Health, Center for Health Services, Arlington,
VA, USA. e-mail: mlegesse@msh.org
Background and challenges to implementation: To

strengthen TB supplies management, the Ethiopian
Ministry of Health (MoH) started TB kit implementation
as of July 2012. The USAID-funded HEAL TB Project
and MoH supported the implementation of the kits in
two regions of Ethiopia. Our objective was to determine
the contribution of TB kit implementation to improving
drug supply management in selected health facilities
(HFs) in the project regions.
conducted to estimate the effect of TB kit on easing drug
management. We included 98 HFs in the study (49 from
each of kit implementing and non-kit implementing
sites). Trained data collectors interviewed 98 HF TB
focal persons. HF registers were reviewed in the kit
implementing and comparison HFs to check for patients
S344
interruption of anti-TB medications. Qualitative data

was analyzed thematically.
Results and lessons learnt: Ninety four percent of the kit
implementing HFs (n 49) preferred kit over the loose
form of drugs due to its building confidence to patients
(57.1%), less preparation time for dose administration
(69.4%), ease of recording & reporting (75.5%),
treatment monitoring (65.3%), logistics (34.7%), and
quantification (38.8%). Moreover, a 7% net increase on
the number of patients who do not interrupt their
treatment for at least a day was recorded in the TB kits
implementation HFs as compared to the comparison
group. However, 32.6% HFs cited that they have faced
workload due to kit implementation and 42.8% reported
shortage of space to store kits. TB kit implementation at
health posts (HPs) was only 14% in the kit implementing
zones. The reported reasons being TB focal persons not
comfortable to give TB drug kits to HPs because of fear
of mismanagement and Health Extension Workers
would not be always available in the HPs to give the
treatment to patients. Additional barriers to kit implementation included absence of sensitization, careful
planning and coordination. Regular reporting and
requisition, intrinsic nature of kit form to build
confidence and ease of drug management were the
factors that facilitated kit implementation.
Conclusions and key recommendations: TB patient Kit
implementation eases drug supply management. Creating awareness on reconstituting of TB patient kits,
providing of shelves for TB clinics, strengthening
planning and coordination at each level of the supply
chain can further improve the kit implementation
process.
PC-982-05 TB lacks space inside the Parliament

of India: demands for a call for action to
involve more parliamentarians both inside and
outside Parliament
S Nayak,1 S Chadha,1 R Pathak,2 V Ghule,2 R Dayal,3
O Bera,1 P Mishra,3 A Das,1 P Mishra3 1International Union
Office, New Delhi, 2World Health Organization Country
Office, India, Ranchi, Jharkhand, 3State TB Cell, Ranchi,
Jharkhand, India. Fax: (91) 114 605 4430. e-mail:
snayak@theunion.org
Background: India, the highest TB burden country of the

globe accounts for 26% TB cases of earth with an
estimated incidence of 2.1 million & 2.6 million
prevalence per annum. India put ~1.4 million TB patient
on treatment in 2013. 1/3rd of cases are still missing from
the national notification system. Challenges of M/XDR
TB, Pediatric TB, quality of care & private notifications
emerge every day/year. Political & admin commitment
have been kept in priority in TB Prog., so also in the
recent End TB strategy. The objective of this paper is to
analyze, how TB as an issue has penetrated into both
Upper (Rajya Sabha) & Lower House (Lok Sabha) of
Indian Parliament. Design & Method Questions raised in
both Rajya Sabha (RS) & Lok Sabha (LS) of Indian
Parliament were selected for analysis. Questions (Qs)
relating to health & TB in total 15 sessions of the 15th LS
(1 Jun 2009 to 21 Feb 2014) represented by 545 elected
members & 4 sessions (230 to 233) held in 2014 of RS
represented by 250 members were reviewed from sites of
both houses. All Qs (starred & unstarred) raised in both
houses were studied under 12 categories: 1.TB Epidemiology, 2.National Prog, Diagnosis & Treatment, 3.MDR/
XDR/TDR, 4.Global TB Report, 5.Corporate, 6.Drugs/
Vaccine, 7.HIV-TB, 8.TB in Women & Children,
9.Policy, 10.Research, 11.External Funding & 12.TB &
Nutrition.
Results: 79 401 Qs asked in 5yr term of 15th LS that
include 4283(5.4%) Qs on health related issues. Among
health questions, TB related was 1.2% (52 Qs) that
include 12(3%) starred questions of total 405 starred Qs
on health. Similarly, 10 025 Qs raised in RS, include 556
(5.5%) on health. The proportion of TB Qs among
health is 3% (17). Out of 795 parliamentarians of both
houses, 61(7.6%) raised Qs on TB. Repeat Qs asked by
8(1%) parliam entarians. Review of 69 Qs of both
houses reveals, 26% (18) are on National Prog.,
Diagnosis & Treatment, followed by TB Epidemiology
20% (14), M/XDR 15% (10), Drugs & Vaccine 12% (8).
Issues relating to other 8 categories are below 5%. Only
2 Qs raised in RS relating to TB & Nutrition and 2 Qs
each on Policy & Research.
Conclusion: TB issues have some space inside the Indian
Parliament. However, that is meagre in comparison to
other health topics in both inside the Parliament and also
among the Parliamentarians. At this hour of enforcing
Post 2015 End TB strategy, the component of Political
Commitment need to be emphasized & re-emphasized
both in and out of the Parliament to substantiate the
commitment of India.
S345
PC-983-05 The need to establish regulatory

framework for TB notification
PC-984-05 The role of medical colleges in

tuberculosis notification
K Chopra,1 D Kundu,2 A Khanna,3 N Sharma,4 S Saini,5

TJ Padmini3 1New Delhi Tuberculosis Centre, Delhi, 2World
Health Organizaion Country Office for India, Delhi,
3
Directorate of Health Services, Delhi, 4South Delhi Municipal
Corporation, Delhi, 5Jag Pravesh Chandra Hospital, Delhi,
India. e-mail: debashishkundu@yahoo.com
S Chandra,1 N Sharma,2 K Chopra,3,4 A Khanna,3

TJ Padmini3 1World Health Organization, Country Office for
India, New Delhi, 2Maulana Azad Medical College, New
Delhi, 3Directorate of Health Services, State TB Office, New
Delhi, 4New Delhi Tuberculosis Centre, New Delhi, India.
e-mail: chandras@rntcp.org
Background: An important component of the WHO End

TB Strategy is to have regulatory frameworks for case
notification. Government of India had declared TB as a
notifiable disease. There are many regulatory acts in the
country wherein reporting a notifiable and dangerous
disease is compulsory. For instance, in the Delhi
Municipal Corporation (DMC) Act, TB is defined as a
dangerous disease and the DMC act also mandates
reporting under section 371 to the Municipal Health
Officer. Also, not reporting a notifiable disease is a
violation of Indian Medical Council (Professional
conduct, Etiquette and Ethics) Regulations, 2002 under
regulation 7.14. However, there is no standard regulatory framework for TB notification resulting in weak
notification from the private sector.
Intervention or response: Delhi State TB Control
Programme had taken initiatives for improving notification of TB cases from the private sector in 2014. Key
steps taken were to constitute a state level TB notification
committee to oversee the progress of TB Notification
efforts in the state and direct one to one sensitization of
PPs (in single PPs clinic, corporate hospitals and
laboratories) by the state notification teams with the
help of available tools for sensitising the private
practitioners (PP) on TB notification - TB Notification
Government Order, Guidance Tool for TB Notification
and Standards of TB Care in India. State TB notification
committee convened 10 focussed action oriented meetings with all stakeholders for strengthening TB notification in the state.
Results and lessons learnt: TB notification cases from the
private sector rose from 200 (2013) to 4000 (2014). 6%
(100/1800) of registered private health facilities in
Nikshay notified TB cases in 2014, with a marginally
rise from 4% in 2013. More than 90% of private health
facilities in Delhi are still not notifying to the public
health authorities.
Conclusions and key recommendations: Active state level
initiatives have led to increase in TB case notification, but
not much improvement is seen in number of PPs
notifying the cases. Lack of regulatory measures at
policy level and lack of awareness among private health
care providers are major challenges in TB notification
implementation. Public health efforts thus need to be
aligned through a framework of complementary measures, both regulatory and enabling, that promote an
adequate level of vigilance. The notification of TB should
be made compulsory at all levels.
Background and challenges to implementation: Medical

Colleges are the hub for referral and tertiary care services
and cater to a large medical outdoor which is not bound
by State boundary. Notification of TB through college
window is of paramount importance as majority of TB
patients get referred for treatment outside the State,
which seldom gets reflected as disease burden in program
records. In order to effectively plan any public health
intervention, it is vital to have a complete burden
estimate by notifying all the TB cases. It is imperative
then that the referred patients do not get missed and get
notified under the program umbrella.
Intervention or response: In 2014, Delhi State took the
initiative to notify TB patients not only from private
sector but also from the bustling outdoors of all its eight
government Medical Colleges. Under the Revised National Tuberculosis Control Program, NIKSHAY online
web based system is used for TB notification. As there
was no separate portal for Medical College notification
under NIKSHAY, each department of Medical College
was registered as an independent referring health facility
with their own web space for TB notification. This
helped track patient referrals from each department and
made the departments accountable for the patient
referrals and feedback. Those patients were omitted
from the notification portal whose feedback of treatment
initiation had been received by the Medical College. This
measure was adopted to avoid record duplication.
Results and lessons learnt: Out of the 11 906 TB cases
diagnosed in 2014 by the eight Medical Colleges in
Delhi, 11 158 were referred for treatment outside the
State and feedback was received for 8433 (76%)
patients. 1195 (48%) of the remaining 2725 TB patients
whose feedback was not received, were notified in the
NIKSHAY under the designated referring health facility
portal created for each Medical College department. The
entries for later half of 2014 are ongoing. Without this
domain in the notification portal, these patients would
have been lost for treatment follow up and missed in the
community.
Conclusions and key recommendations: Medical Colleges play a pivotal role in TB notification as each year
considerable load of TB patients get referred from
Medical College across the country. Separate notification
domain for individual departments of Medical College
ensures that the referred patient gets notified under the
program umbrella and does not get missed in the
community.
S346
PC-985-05 Targets vs. impacts: understanding

the impact of the Global Funds performance
based funding modality in the Nepalese health
care sector
K Dahal,1,2 R Khatri,1 S Khanal,1 S C Baral,1 I Harper1,3
1
Health Research and Social Development Forum,
Kathmandu, 2Tribhuvan University, Kathmandu, Nepal;
3
University of Edinburgh, Edinburgh, UK. Fax: (977) 1 410
2016. e-mail: kapildahal@gmail.com
Background: The Global Fund to Fight AIDS, Tuberculosis and Malaria (GFATM) is distinct among funding
agencies because of its lack of country presence and
performance based financial disbursement mechanism. It
operates through distributing funds to the Primary
Recipients (PRs), which can be either government entities
or organizations from the non-governmental sector.
These then disperse resources to their implementing
partners, the Sub-Recipients (SRs) who implement
programmes. Outputs, calculated against proposed
targets, are measured through specific recording and
reporting, monitoring and evaluation systems.
Design/Methods: The data for this study was collected
between 2013 and 2015. Extensive in-depth interviews
were carried out with those responsible for implementing
the TB and HIV components of GFTAM work in
government departments, with PRs, SRs and people
living with HIV and AIDS (PLHAs) and their networks,
from the central to peripheral levels in Nepal. Participant
observation of different meetings, workshops and
interactions were also another crucial means of data
collection.
Results: Despite its claim of being a simple funding
disbursement mechanism, GFATM has contributed to
increasing bureaucratization in the health sector. There is
a disjuncture between the government budget release
through the Red Book, which records all government
planned activities and the GFATM form of disbursement.
In addition, reporting against outputs has become more
focused on targets at the expense of evaluating programmatic impact. This has, ultimately, made the stakeholders
involved in the GFATM process focus on measuring
activities, in order to reach targets so that further funds
can be released. In such a situation, the problem is further
compounded by delayed disbursement in an unstable
political context and the highly politicized health service
delivery environment.
Conclusion: Greater awareness of the unintended difficulties of managing GFATM projects is required to
appreciate the entanglement of GFATM projects with
Government and NGOs. Appreciation of this can be
achieved through qualitative research performed with
those involved in running the programmes.
PC-986-05 Pakistans National TB Control

Programs response to the End TB Strategy
E Qadeer,1 B Khan,1 N Safdar,2 S-E Ottmani,3 M Ahmad1
National TB Control Program, Islamabad, 2Social & Health
Inequalities Network, Islamabad, Pakistan; 3Independent
Consultant, Geneva, Switzerland. Fax: (92) 843 8081.
e-mail: drbasharatjaved01@gmail.com
1
Background: A vibrant National TB Control Program in

Pakistan since 2001 has provided free of cost standardized TB diagnosis and treatment to about 2.5 million
sensitive and six thousand resistant cases in the country.
A nationwide population-based TB prevalence survey
(2010-11) and drug resistance survey (2102-13) reported
an estimated TB prevalence of 342/100 000 populations
and drug resistance of 3.7% among new and 18.1%
among previous treated TB case. The NTP Pakistan,
being responsive to End TB Strategy, imminent Sustainable Development Goals and growing TB burden has
given its strategic vision and securing resources to reduce
TB incidence.
Intervention: The NTP Pakistan in 2014 developed a
costed National Strategic Plan (NSP) Vision 2020 0
envisaging The End TB Strategy with full expression of
demand worth US$ 876M. The NSP included innovative
interventions to find missed cases through systematic
screening, maximizing public sector investments and
accountability, universal access to susceptible and drug
resistance TB, addressing TB-HIV and other co-morbidities and prioritizing research. In addition, NTP Pakistan
developed concept note for New Funding Model of
Global Fund for a period 2015/17, mainly to address
missed TB cases and managing DR-TB cases.
Results: The milestones of End TB Strategy were adapted

as goal of NSP Vision 2020 i.e., to reduce 50% the
prevalence of TB by 2025 in comparison to 2012 in
Pakistan and reduce new TB cases at 5% annually from
2018. This implies that in 2018 the NTP Pakistan will be
notifying 420,000 cases (71%) from current 316,544
(62%) at an incremental effect of about 2.5% annual
increase in case notification (Figure). It is expected that
social determinants will also impact the TB incidence by
2018 as Gross Domestic Product (GDP) shows a positive
correlation since the inception of TB program (2001GDP 72.3, 2013-GDP 237). The investment through
NSP will also help reducing by at least 5% per year by

2020 the prevalence of DR-TB among TB patients who
have never received any TB treatment and to strengthen
programmatic and operational management capacity of
the TB Control Program while enhancing public sector
support by 2020.
Conclusion: The NTP Pakistan based on the NSP Vision
2020 has secured a grant of US$ 125M from Global Fund
up to 2017.The TB control program is also developing
plans to secure public sector support to address the
funding gaps to achieve its targets in pursuance of The
End TB strategy.
PC-987-05 Social drivers of TB in the context of

an increasing burden of non-communicable
diseases: evidence from South Africa
P Naidoo1 1University of the Western Cape, Cape Town,
South Africa. e-mail: pnaidoo@hsrc.ac.za
Background and challenges to implementation: Deaths

due to non-communicable diseases (NCDs) are expected
to increase by 17% over the next 10 years with the
largest increase of 24% expected to be in Africa. In South
Africa (SA) this rise in the NCDs cluster which includes
mental disorders, such as depression and anxiety, is
exacerbated by the existing burden of HIV and TB and is
the second leading cause of death after HIV. In 2009, 9%
of individuals 30 years and older were estimated to have
Type2-diabetes mellitus, representing approximately 2
million people.
Intervention or response: There is also evidence to show
that in a country such as SA there are numerous social
drivers and risk factors that are common to both the
NCDs cluster and communicable diseases, such as TB.
Despite South Africa being categorized as a middleincome country by the World Bank, there are glaring
levels of social and economic inequalities, which are
associated with poor health outcomes and ill-health. On
closer examination the South African disease burden
profile resembles that of many low-income countries.
Results and lessons learnt: At population level there was
a 9.6% prevalence of diabetes reported in the 1st South
African National Health and Nutrition Examination
Survey [SANHANES I]. Adding to this burden of disease
is the fact that the country has 28% of the worlds HIV
positive adult TB cases and the worlds highest TB
incidence. Diabetes predisposes patients to developing
TB with indications that diabetes is associated with a 3fold incident risk of TB. Moreover, the mediating effect
of common mental disorders (CMDs), especially depression and anxiety, on the health outcome of individuals
with diabetes, HIV and TB is poorly characterized given
the fact that there is an 11% life-time prevalence of
depression in low and middle income countries, including SA. Data from the SANHANES I, indicated that at
least 28.4% of the South African population 15 years
and older have symptoms of anxiety and depression.
Conclusions and key recommendations: Given the
epidemiological pattern of the NCDs cluster, HIV-TB
diseases in SA it is evident that the country is facing a
S347
multiple burden of disease. To address the disease burden

it is also important to understand the social epidemiological profile of the country. Public Health Policy
adjustments need to be made for both NCDs and HIVTB given their increasing co-existence.
PC-988-05 Use of regulatory action in

tuberculosis control in New York City, 20112013
J Burzynski,1 R Oken,1 D Nilsen,1 M Robinson,1
R Acevedo,1 J Martinez,1 M Macaraig1 1New York City
Department of Health and Mental Hygiene, Queens, NY,
USA. e-mail: mmacarai@health.nyc.gov
Background: The tuberculosis (TB) resurgence in New

York City (NYC) in the early 1980s to 1990s led to a
series of enhanced TB control measures, including
revising the NYC Health Code to allow for regulatory
action against non-adherent infectious TB patients.
Regulatory action is only permissible if the patient is a
threat to public health and attempts to achieve voluntary
adherence have failed or were considered and ruled out.
After over two decades of decline in TB incidence, the use
and impact of regulatory actions in NYCs TB control
program were examined.
Design/Methods: All Mycobacterium tuberculosis culture-positive pulmonary patients reported to the NYC
Department of Health and Mental Hygiene between
January 1, 2011 and December 31, 2013 were obtained
from the TB registry. Patients referred for regulatory
action were identified using referral log sheets. Demographic, clinical, and social characteristics of patients
were examined. The proportion of patients with an order
for TB examination, treatment, and detention was
calculated. Characteristics and treatment outcomes of
patients issued an order were compared to those who
were not. Pearsons v2 was used to compare proportions
and the Wilcoxon rank-sum test to compare medians.
Results: During this period, there were 1542 culture
positive pulmonary TB patients reported in NYC, of
whom 146 (9%) were referred for possible regulatory
action. Sixty-four (44%) referrals were resolved without
regulatory action. Of the 82 patients who received an
order, 52 (64%) were issued an order to attend
outpatient examination or complete treatment, primarily
under directly observed therapy; 10 (12%) an order to be
detained until no longer infectious; and 20 (24%) an
order to be detained until completion of treatment.
Compared to patients who were not issued an order,
those with an order were more likely to have TB with
drug resistance, including multidrug-resistant TB (27%
vs. 11%, P 0.02), and report homelessness (26% vs.
11%, P 0.03) and drug use (35% vs. 14%, P , 0.01).
Treatment completion did not differ significantly between the two groups (82% vs. 85%, P 0.65).
Conclusion: Regulatory action in NYC is used sparingly
and only after efforts have been made to achieve
voluntary compliance or where voluntary compliance is
not feasible. Orders were more likely issued to patients at
high risk for non-adherence. Regulatory measures
S348
continue to play an important role in NYCs TB control

efforts.
PC-989-05 Generating sub-national estimates

of the effectiveness of tuberculosis policy
interventions to meet the post-2015 global TB
targets
Z Mclaren,1 K Schnippel2 1University of Michigan, Ann
Arbor, MI, USA; 2Right to Care, Johannesburg, South Africa.
e-mail: zmclaren@umich.edu
Background: The WHO Global TB Programme strategy

sets ambitious targets for the reduction in TB cases and
TB deaths post-2015. Modeling has been conducted at
the national level to inform the choice of interventions,
scale and resource allocation to support country program
planning. This study leverages the wealth of data
available at the sub-national level in South Africa to
guide region-specific policy that is targeted to areas
where it meets local needs of populations and is likely to
be most effective and cost-effective. Tuberculosis has
been the leading cause of death in South Africa for the
past decade and there is an urgent need for effective subnational policy to address the epidemic. There are
important implications from this study for program
implementation by WHO GTB, STOP TB Partnership,
Global Fund, UNAIDS and other key global stakeholders.
Laboratory Service database on TB diagnostic tests
performed on patients in public health facilities and
Census data from 2011 to calculate the burden of TB
cases per capita by 226 local municipalities. We examine
factors related to (1) early diagnosis of tuberculosis
including universal drug-susceptibility testing and (2)
treatment of all people with tuberculosis including drugresistant tuberculosis, and patient support. Using regression analysis, we evaluate the relationship between the
TB burden and measures of the poverty rate, access to
primary health care, and clinician compliance with
current diagnosis guidelines.
Results: Local municipalities with a greater fraction of
urban residents and higher rates of household access to a
cell phone have higher tuberculosis rates per capita.
Counter-intuitively local municipalities with lower unemployment rates have higher TB rates per capita, which
may be driven by urban residency and better access to
health facilities.
Conclusion: This analysis focuses on population characteristics and service-delivery factors that influence the
cost-effectiveness of WHO interventions. It highlights
factors that are most associated with high TB burdens,
and would therefore need to play a large role in any
effective targeting strategy. Allocating additional resources for sub-national analysis will conserve resources
with improved program cost-effectiveness over time.
35. Financial impact of TB

PC-990-05 BRICS: cooperation and innovation
towards tuberculosis elimination
C Oliveira Dantas,1 P Bartholomay Oliveira,1 A Torrens,1
D Pelissari,1 F Moherdaui,1 D Barreira1 1Ministry of Health,
Braslia, Distrito Federal, Brazil. e-mail:
cintia.dantas@saude.gov.br

BRICS countries label refers to a select group of 5 large,
developing countries (Brazil, Russia, India, China and
South Africa). These countries are promising emerging
markets, leading the world in public policies. In terms of
health, tuberculosis (TB) is a common threat. The World
Health Organization (WHO) estimates that Brazil,
Russia, India, China and South Africa account for almost
50% of all worlds reported cases, adding up to 2 830
807 TB cases in 2013 and a high amount of patients in
these countries did not have access to free of charge and
quality-assured treatment.
Intervention or response: Considering that TB is a
poverty related disease and aligned with the Tuberculosis
prevention, care and control after 2015, BRICS came
together on an innovative proposal involving universal
access to First Line TB Drugs (FLD) to BRICS countries
and, gradually to 83 low and middle income countries
selected based on data from the World Bank and the
WHO tuberculosis report.
Results and lessons learnt: Data assessment was conducted to understand the real number of cases not
covered by the public sector, as well the costs dimensions
to provide universal access to FLD within the BRICS, and
later, to low and middle income countries. In 2013,
BRICS notified 46.4% of the worlds TB cases. Currently
the gap among cases not covered by public sector is
almost 880,000 patients, representing almost 30% of all
TB cases notified by BRICS. Brazil estimative treatment
expenditure for FLD was approximately US$ 24.00 per
patient in 2013. To cover the treatment for those cases
within the BRICS, it is estimated that US$ 18 702 027. is
needed. In addition, for that same year, low and middle
income countries reported 41.2% (2 511 734). Considering BRICS TB cases, in total, 88% (5 342 541) of all
worlds tuberculosis cases will be covered by the
proposal. In terms of budget, it is estimated that US$
128 220 984. is needed, considering all TB notified cases
in BRICS and other selected countries. The sharing of this
amount will respect BRICS consensus.
Conclusion and recommendations: BRICS are currently
negotiating the next steps of the Tuberculosis Cooperation Plan. In addition, a Commitment Term must be put
in place, in which recipient countries will grant their
former TB budget addressed to drugs to be applied on
another TB need lab strengthening, second line drugs,
early diagnosis investments, R&D, and so forth.
PC-991-05 Multi-channel fundraising

mechanism demonstrated to reduce the
economic burden of tuberculosis patients in
China
1 1
l Wang Chinese Center for Disease Control and

Prevention, Beijing, China. e-mail: quyan@chinatb.org
Background and challenges to implementation: As part

of massive scale-up of Nation TB control Program in
2004 in China, policy of free TB screening examinations
and free anti-TB drugs and follow-up examinations for
all confirmed TB cases was implemented nationwide. In
spite of this progress, however, the epidemic remains a
serious public health issue and is compounded by the
emergence of drug resistance. Based on the ongoing
strengthening TB control strategy in China, innovative
financing models should be established for diagnosis,
treatment and case management of TB patients. Chinese
National Health and Family Planning Commission
(NHFPC) and Bill & Melinda Gates Foundation
cooperated to explore the feasibility of the multi-channel
fundraising mechanism for TB and MDR-TB control
since 2009.
Intervention or response: Local governments in Zhenjiang city of Jiangsu province, Yichang city of Hubei
province, and Hanzhong city of Shannxi province as well
as 26 counties and districts affiliated to the 3 cities,
representing different geographical and economic situation in the east, middle and west regions of China TB
control were advocated to demonstrate financing and
reimbursement policy for the costs of essential services
through local urban residents medical insurance, new
rural cooperative medical insurance, civil affairs aid and
related charitable institutions, expected the rifampicin
susceptible TB patients paid no more than 30% of the
total cost out-ofpocket cost in the service package and
rifampin resistant TB patients no more than 10%.
Results and lessons learnt: In the project cities, the new
financing and reimbursement policy developed and
demonstrated since 2013. As results, all Rifampin
Sensitive TB patients just paid 20% costs of the services
received in Zhenjiang and Yichang and 30% in
Hanzhong respectively, while the remaining cost were
covered by medical insurances. For the Rifampin
Resistant TB patients, the service costs were 100% in
Zhenjiang and Hanzhong and 90% in Yichang covered
by medical insurances (urban/rural), local civil affairs aid
and charitable organizations.
Conclusions and key recommendations: The multichannel fundraising mechanism demonstrated in project
cities dramatically reduced the economic burden of both
Rifampin Sensitive/Resistant TB patients, which proved
the feasibility of fundraising mechanism for TB treatment in different regions of China.
S349
PC-992-05 Out-of-pocket expenditure on

private health care providers among TB
patients in Salem district, Kerala, India
M Arulanantham,1 C Joltin,1 V Sameer,1 P Bhat,1
T Thomas1 1The Catholic Health Association of India,
Secunderabad, India. e-mail:
Background: The initial point of contact for seeking

health care is a private service provider among more than
80% of the population in India and TB is no exception to
this norm. This is despite the free DOTS treatment
provided through the public health system. The current
study was conducted to assess out of pocket expenditure
on private healthcare providers among TB patients in
Salem district, Kerala, India.
Design/Methods: The study was conducted in Salem
District of Tamil Nadu, South India. A convenient
sample of 50 patients who were attending a Government
Treatment center and receiving Category-I (new cases)
treatment were selected and interviewed with a structured questionnaire. Expenditure patterns among patients who had initially contacted a private service
provider were assessed. Consent was taken from District
TB officer and the patients.
Results: Only 10% of the sample respondents were
aware of free treatment and directly went to government
health service. 20% of the TB patients had spent nearly
Rs.100 000 in private treatment. 20% of the TB patient
had spent Rs.10 000 to Rs. 25 000. 45% of the TB
patients had spent Rs.2000 to Rs.5000. The remaining
5% had spent less than Rs.1000 in private health
facilities. 90% of the patients belonged to low socioeconomic status. 92% of the patients were satisfied with
the treatment and government health facilities.
Conclusion: The main reason for seeking private
facilities is lack of knowledge on TB and free DOTS
treatment. Dissemination of key messages on TB and free
treatment is the only way to reach the poor people to stop
their economic downfall and expenditure on private
providers for tuberculosis despite free treatment being
available.
PC-993-05 Rural India paying more towards

direct and indirect costs for pre-treatment
evaluation of TB: expenditure analysis in rural
and urban Pune
O Bera,1 M Chandge,2 S Kamble,3 B Pawar3 1The Union
South-East Asia Office, New Delhi, 2District TB Office,
Mumbai, 3State TB Office, Pune, India. e-mail:
oprakash@theunion.org
Background: Patients cost are, in part a function of the

structure of health system, with pre diagnostic costs
reflecting accessibility of diagnosis. The present study
was undertaken to estimate overall (direct and indirect)
cost incurred by patients in rural and urban areas for pretreatment evaluation of TB.
Intervention: Cross sectional questionnaire based study
was conducted among new adult smear positive pulmonary TB patients during the period Jan Dec 2014 in
S350
Pune. New sputum smear positive TB patients having

HIV negative status having completed intensive phase of
treatment were included. Pune urban has 6 tuberculosis
units (TU) having population of 1.2 million and rural
part has 12 TUs with population of 4.7 million. Out of
the shortlisted TB patients from the TB register 120 TB
patients each from rural and urban area were randomly
selected, including 20% non-response rate. Information
regarding direct cost (medical and non-medical) and
indirect cost was estimated for two periods (1) Before
treatment (two months prior to start of TB treatment) (2)
During Intensive phase (two month treatment period) (3)
During Continuation phase. All calculations are in terms
of Indian Rupees (Rs). Result: Of the 240 TB patients,
205 were interviewed. 15 were excluded as outliers based
on interview. Analysis was done for 190 patients [93 and
97 from urban and rural respectively]. 79% patients
belonged to age group of 15- 54 years. Study group
included 59.4% males and 40.6% females. About 51%
in urban had per capita income of less than Rs 2250 and
56% in rural had per capita income of less than Rs 1100.
During pre-treatment phase, it was observed that 91.8 %
and 98% of rural patients are paying towards direct
medical and non-medical cost in comparison to 83.9%
and 92.5% urban patients in the same head respectively.
Also rural patients are paying more towards indirect
cost; 89.7% rural in comparison to 54.8% urban
patients. During treatment phase, rural patients (32%)
are paying more towards indirect cost as compared to
urban patients (24.7%).
Conclusion: Low awareness among patients about
disease, free TB diagnosis and treatment have forced
them to go on a spiral shopping trend for treatment
resulting in increased pre-treatment cost in rural areas .
Although there have been multiple commitments from
national and state governments, still large scale public
health programmes are not reachable to common man.
Table Pre-treatment costs to Urban and Rural TB patients
Cost Incurred
Nil cost
n (%)
6 Rs
1000
n (%)
. Rs
1000
n (%)
Total
n (%)
Pre-treatment cost to Rural TB patients (n 97)

Direct Medical Cost
8(8.2) 44 (45.4) 45 (46.4) 89 (91.8)
Direct non-medical
cost
2(2)
33 (34) 62 (64) 95 (98)
Indirect cost
10(10.3) 36 (37.1) 51 (52.6) 87 (89.7)
Pre-treatment cost to Urban TB patients (n
Direct Medical Cost
15(16.1) 41(44.1)
Direct non-medical
cost
7 (7.5) 56 (60.2)
Indirect cost
42 (45.2) 36 (38.7)
93)
37(39.8) 78 (83.9)
30 (32.3) 86 (92.5)
15 (16.1) 51 (54.8)
PC-994-05 The role of the Public-Private Mix

team is key to the success of PPM
implementation in East Jakarta: from theory to
practice
Y N Sumarli,1 I Kurniaty,2 Y Tahir,3 I Ibnu,4 D Setyawati,5
Y I Hiola,6 M R Christian7 1Cipinang Detention Center,
Jakarta 2East Jakarta District Health Office, Jakarta, 3KNCV
Tuberculosis, Jakarta, 4Cipto Mangunkusumo Hospital,
Jakarta, 5Family Health International 360, Jakarta, 6District
Office Ministry Law and Human Rights, Jakarta, 7World
Health Organization Country Office Jakarta, Indonesia. Fax:
(62) 218 591 1415. e-mail: dr.yulius.st@gmail.com
Background and challenges to implementation: National

health survey in 2010 showed that 63.9 % tuberculosis
patients seekstreatment to hospitals. The data from MoH
in 2009 only 17% of the hospitals implement DOTS
optimally of which only 28% hospital follow the
treatment guideline, and only 40% hospital do reporting
and recording timely. East Jakarta city has 571 health
care facilities, 88 primary HCF, 22 hospitals, 4 prisons,
20 private clinics, and hundreds of private practitioners.
Therefore engaging hospital and private sectors sector is
of utmost priority for Tuberculosis (TB) Control
Program in East Jakarta
Intervention or response: To response the problem,
Public Private Mix team in east Jakarta was established
in August 2013 The team consists of representative from
doctor association, hospitals, prisons, district health
office, provincial office of MoLHR, and NGOs. The
aims of the establishment PPM team was to find and treat
TB patient, provide access to quality TB services through
ISTC implementation, thus reducing TB morbidity and
mortality in east Jakarta and prevent MDR. The
expected roles are to act as a counterpart from DoH to
evaluate and monitor the process of PPM implementation, to raise the teams awareness and belonging to the
program, to have a uniform system and understanding of
reporting recording, diagnosis, treatment, and case
management, and to create linkage among the health
care facilities in east Jakarta
Results and lessons learnt: By the end of 2014 all Prisons
in East Jakarta, 7 private practitioners, and 4 clinics have
implemented DOTS and submit report in timely manners
and contribute 46 % of TB case finding in East
Jakarta.The team had also facilitated the establishment
of East Jakarta local TB movement (GERDUDA), and
MoU signing among stakeholders to fight against
Tuberculosis in East Jakarta. Series of advocacy to the
east Jakarta Mayor have been conducted to achieve the
mayor commitment. A circulating letter to all HCF in
East Jakarta emphasizing the importance of TB prevention and control has been issued by the local government.
Conclusions and key recommendations: The establishment of PPM team plays an important role in engaging
more sectors in PPM. The team can build the linkage and
enhance the partnership. It can also help to advocate and
strengthen DOTS implementation. Therefore it needs to
be promoted and strengthened continuously.
S351
PC-995-05 Engagement of pharmacists to

ensure TB commodity security for public sector
TB patients in Kenya
PC-996-05 Previous tuberculosis experience

and its impact on household income: a
historical analysis, South Korea, 19802012
H Kalili,1 R Muthoka,2 B Mungai,1 E Masini2 1Centre for

Health Solutions Kenya, Nairobi, 2Ministry of Health, Nairobi,
Kenya. e-mail: hellenben@hotmail.com
C Hongjo,1 H Chung2 1Korean Institute of Tuberculosis,

Osong, 2Korea University, Seoul, Republic of Korea. e-mail:
hongjochoi@gmail.com
Background and challenges to implementation: TB is the

4th leading cause of death in Kenya, with 89 170 cases
notified in 2013. An uninterrupted supply of TB drugs is
imperative for a successful TB national program.
Although pharmacists play a crucial role in TB treatment, their involvement in the national TB control
program (NTP) has been sub-optimal. Since 2006,
management of TB drugs in NTP was performed by an
epidemiologist. The program experienced delayed deliveries, stock-outs, expiries and inaccurate forecasting. In
response, the Ministry of Health (MOH) in 2008
recruited a pharmacist to NTP, which resulted in
improved planning for future procurements, resource
mobilization and central distribution of drugs. The
distribution system beyond the county level however
remained weak with poor commodity management
practices. County TB officers in their role as commodity
managers had limited capacity in pharmaceutical management.
Intervention or response: In 2010, NTP recruited 2
pharmacists and established a national commodity
security committee. NTP shifted roles and responsibilities of county TB commodity managers from TB officers
to pharmacists. Between 2010 and 2011, 400 pharmacists received training on commodity management and
pharmacovigilance. In 2014, 2 workshops for 47 county
pharmacists were conducted on forecasting and quantification including use of QuanTB tool. County pharmacists now participate in quarterly and annual performance review meetings, where they present data on
current stock. They also attend national quantification
workshops.
Results and lessons learnt: 47 county budgets were
developed to justify allocation of funds. As a result, 8
county governments allocated 850 000 USD for drug
procurement. The increased involvement of pharmacists
in management of TB drugs at all levels of health care has
improved efficiency throughout the system. At county
level, sharing of supply chain information has enabled
the identification of gaps, and informed reverse logistics.
256 TB drug distribution points were established in
counties.
Conclusions and key recommendations: Active engagement of pharmacists at national and county levels,
contributes to the success and sustainability of effective
TB pharmaceutical management. There is need for
regular dialogue and capacity enhancement for increased
support and commitment. The impact achieved in Kenya
suggests that this model could be applicable in high TB
burden countries.
Background: A myriad of studies has focused on the

pathway from social factors to tuberculosis (TB), but no
studies have investigated the reverse mechanism from TB
to social outcomes. The study aims to investigate
relationship between social outcomes measured by
household income and the previous TB experience
stratified by paternal education attainment in South
Korea. We focused on finding the historical trends of TB
and a shift or a break time point of its trends.
Design/Methods: We used the Korean National Health
and Nutrition Examination Survey from 2007 to 2012.
Four exposure groups were constructed with lifetime TB
experience (TB or non-TB) and the level of paternal
education (High-Edu or Low-Edu): TB with LowEdu, TB with High-Edu, non-TB with Low-Edu and nonTB with High-Edu. Outcome measure was a quartile of
current household income (1 to 4, 4: the highest income
group). A multinomial logistic regression model, adjusting for age, gender, occupation and education, was used.
In addition, we applied recursive regression model
suggested by Isaac & Griffin (1989). In the first analysis,
TB cases being diagnosed from 1980 to 2012 and all nonTB cases were included. We dropped the first year in the
second analysis and subsequently dropped the previous
year in the next analysis. The last analysis included TB
cases being diagnosed from 2003 to 2012 and all non-TB
cases. Consequently, different point estimates were
calculated at each time period from 1980-2012 to
2003-2012.
Results: In the first analysis (1980-2012), the TB with
Low-Edu group was likely to be presented as the lowest
current household income compared to the non-TB with
High-Edu group (coefficient: 0.966, 95%CI: 0.631.30,
P value: , 0.001). In the last analysis (2003-2012), the
magnitude of trend was strengthened (coefficient: 1.512,
95%CI: 0.732.31, P value: ,0.001). The trend was not
considerably changed during a study period from 19802012 to 19972012 (coefficient: 0.993), but its magnitude started to become stronger after dropping the year
of 1997.
Conclusion: The study revealed that previous TB
experience in lower paternal education group is associated with current lower household income. The magnitude of trend has become stronger since the year of 1998
that was hit by economic crisis in South Korea. Thus, the
year of 1998 would be considered as a shift or break
point that helps better understanding of social outcomes
generated by previous TB experience. TB requires at least
more than 6 months to be treated and it may cause
financial difficulties. This study suggests three plausible
pathways to explain these disparities among the TB
with Low-Edu group since 1998. Firstly, social resilience to overcome financial difficulties derived from the
diseases may be weakened after the Korean economic
S352
crisis in 1998. Secondly, individuals with low-income

were diagnosed TB more than those with high-income
since 1998, so the result may reflect the reverse
mechanism from low income to more TB experiences.
Lastly, interactions between social outcomes and TB may
simultaneously occurred right after the Korean economic
crisis. Therefore, the recursive regression model would
be a useful way to investigate the relationship between
health and social outcomes and to study social impacts
on specific diseases.
four studies of robust design aiming at measuring effect

of incentives on case detection or on return rates for skin
test reading showing varying levels of effectiveness. A
good number of additional studies described the use of
incentives but only as a supportive strategy in a wider
intervention. A preliminary analysis shows great variety
in types of incentives used (from flat rate to results based
payments) and in their operationalization. While some
authors strongly believed that incentives were crucial in
achieving the program goals others expressed concerns
on costs implications and sustainability. Only few
described assumptions underpinning design and context
necessary to fully appreciate operational challenges.
Conclusion: While incentives for patients and providers
are frequently used in programs aiming at increasing TB
case detection, little is known on their effectiveness and
on specific conditions for successful implementation.
There is an urgent need to supplement anecdotal
information currently used by proponents and opponents
with robust evidence, a more contextualized description
of lessons learned and better research on the dynamics of
organizations including human motivation.
PC-998-05 Impact of providing financial

incentive to tuberculosis patients on treatment
outcomes in a low-resource setting:
implications for the End TB strategy
PC-997-05 The use of incentives to accelerate

tuberculosis case notification: current available
knowledge reviewed
L Blok,1 M Tijm,1 O Ramis,2 M Bakker1 1Royal Tropical
Institute, Amsterdam, Netherlands; 2Mott MacDonald,
London, UK. Fax: (31) 20 568 8444. e-mail: l.blok@kit.nl
Background: Incentives are frequently used in public

health interventions to influence behavior of patients or
providers. Also in tuberculosis control we observe
increased interest in incentives as a supportive strategy
to enhance tuberculosis control. While there are studies
on the effectiveness of incentives to support treatment
completion and improve treatment outcomes, little is
known regarding its effectiveness in increasing case
notification. Our aim is to review current knowledge
on use of incentives for patients or providers in increasing
detection of active TB.
Design/Methods: We performed a systematic review of
studies published between 1995 and Nov 2014 using
Pubmed, EMBASE and Cochrane, supplemented with
studies found through snowballing and relevant website
searches. Search terms included a combination of a term
related to tuberculosis AND a term related to motivation/incentives/enabler AND a term related to diagnosis/
case-detection. Abstracts and shortlisted full articles
were reviewed by two reviewers independently and data
were extracted using a standard template.
Results: The search identified 1095 articles, 178 of which
were identified as possibly relevant for our study
questions and their full text was read. We found only
K Ukwaja,1 I Alobu,2 M Gidado,3 J Onazi3 1Department of

Ebonyi State, 2National Tuberculosis and Leprosy Control
Programme, Ministry of Health, Abakaliki, Ebonyi State,
3
KNCV Tuberculosis Foundation/TB CARE I Project, Abuja,
Background: Poverty undermines adherence to tuberculosis treatment and patients cost of care is one of the
major barriers against tuberculosis care. Previous studies
have shown that these costs are catastrophic to households and may discourage tuberculosis patients from
initiating or continuing care. The aim of this study was to
evaluate the feasibility and effectiveness of delivering
economic support to patients with tuberculosis in a highburden setting with significant resource limitations.
Design/Methods: A before-and-after implementation
design was employed. The study was carried-out in a
large rural health facility in Ebonyi State, Nigeria. We
prospectively enrolled all registered tuberculosis patients
in the facility during April to June 2014 into the
intervention arm. Then, all registered tuberculosis
patients during January to March 2014 were used as
the control arm. Patients in the intervention arm were
offered a monthly financial incentive of N2500 (US$15)
until the completion of their treatment. Patients in the
control arm received the usual care. Logistic regression
analysis was used to identify the determinants of
treatment success.
Results: A total of 294 tuberculosis patients were
registered; (173 and 121 in the control and intervention
arms respectively). The patients did not differ according
to their demographic or clinical characteristics (P .
0.05). Treatment success rate was 104/121 (86.0%) in
the intervention, and 123/173 (71.1%) in the control

period (p 0.003). The proportion of patients who were
lost-to-follow-up significantly decreased during the
intervention period (20.2% vs. 5.0%; P , 0.001). There
were no differences in death rates (6.6% vs. 5.8%; p
0.8) or treatment failure rates (1.7% vs. 2.9%; P 0.5)
between patients in the intervention and control arms
respectively. Also, among smear-positive tuberculosis
patients, there was no difference in the rate of sputumsmear conversion after the intensive phase of treatment
in the intervention and control arms respectively (88.1%
vs. 91.5%; P 0.5). Independent determinants of
treatment success were; female gender (adjusted odds
ratio [aOR] 1.9), HIV-negative status (aOR 2.5), and
receiving financial incentives (aOR 2.3).
Conclusion: Financial incentives proved to be effective in
improving treatment success and reducing loss-to-follow-up among tuberculosis patients in Nigeria. The
findings suggest that providing financial incentives can
be utilized to enhance case management of tuberculosis
in under-resourced populations. These observations
should inform policy on the End-TB strategies.
S353
and slightly inverted: from a 12% annual decrease in the

pre-crisis period to a 0.5% increase in the post-crisis
period. Also, in this population the inequality by income
is much stronger, as migrants living in lowest income
neighbourhoods have an incidence 3 times greater than
that of the richest neighbourhoods. In indigenous
population, the economic crisis has not affected the
downward trend in incidence, but they are also affected
by the effect of income inequality: a low-income
neighbourhood has a 34% higher incidence than a high
income neighbourhood.
Conclusion: The downward trend in the incidence of TB
has suffered a clear slowdown in the migrant population
since the beginning of the crisis and the cuts in social
protection policies. In addition, the highly unequal
distribution of TB by income level leads to differences
between neighbourhoods comparable to those occurring
between Western countries and low income countries.
Measures to protect the migrant community are essential
to reduce inequalities and the incidence of TB in
Barcelona.
PC-999-05 Impact of economic crisis on the

incidence of tuberculosis in the city of
Barcelona
A Prats-Uribe,1,2 J P Millet,2,3 A Orcau Palau,2,3
J A Cayla` 2,3 1Unitat Docent de Medicina Preventiva i Salut
Publica
PSMar-UPF-ASPB, Barcelona, 2Epidemiology Service.
Public Health Agency of Barcelona, Barcelona, 3Ciber De
Epidemiologia y Salud Publica,

Barcelona, Spain. e-mail:
albert.pratsu@gmail.com
Background: There is plenty of evidence of economic

crises affecting health in several aspects, including
communicable diseases. The actual economic crisis has
affected the city of Barcelona incrementing inequalities
and may have influenced the tendency of tuberculosis
(TB) incidence, especially for the more disadvantaged
groups.
Design/Methods: Population-based incidence study. All
TB cases living in Barcelona who started treatment
between 2003 and 2013 were studied. A descriptive and
analytical study of socio-demographic, epidemiological
and clinical features of the cases was conducted,
comparing pre and post-crisis periods (2003-08 vs.
2009-13). The annual incidence was calculated for each
neighbourhood globally and by migrant status. A panel
negative binomial regression model was adjusted to
examine differences in the trends of these incidences
before and after the crisis, and by the neighbourhood
family income.
Results: There is a global trend towards reduction of the
TB incidence citywide, with no overall effect of the crisis
on it. Differences in the incidence of TB were found in
relation with family income level, where lower income
neighbourhoods had 80% more TB incidence than the
less deprived. The separate analysis of migrants and
native population shows a clear difference in the
behaviour of trends: The previous downward trend of
TB incidence in migrant population has been stopped
PC-1000-05 In Peruvian shantytowns, the

financial impact of TB-related and catastrophic
costs has not changed over the past decade,
despite changes in TB rates
T Wingfield,1,2,3 M Tovar,1,4 D Huff,1,5 D Boccia,1,6
R Montoya,1 E Ramos,4 J Lewis,1,6 C Evans1,2,4 1Innovacion

Por la Salud Y Desarrollo (IPSYD), Lima, Peru;, 2Imperial
College London, London, 3North Manchester General
Hospital, Manchester, UK; 4Universidad Peruana Cayetano
Heredia, Lima, Peru; 5Tulane University, Tulane, LA, USA;
6
e-mail: tom.wingfield@ifhad.org
Background: Catastrophic costs incurred during TB

illness are associated with adverse treatment outcome
and thus may negatively effect TB control. The World
Health Organisation (WHO) has mandated that no TBaffected family should incur catastrophic costs. In order
to inform the design and implementation of a complex
social protection intervention to control TB, we measured catastrophic costs of free TB treatment and
compared the results to costs data from the same study
site in 2004.
Design/Methods: Between February and August 2014,
consecutive TB patients diagnosed by the Peruvian
S354
National TB Program in 32 contiguous shantytown

communities of north Lima, Peru, were invited to
participate in the study. Healthy controls were randomly-selected and recruited concurrently. Patient households direct out-of-pocket expenses, lost income, and
total costs were recorded throughout treatment. Catastrophic costs were defined as total costs of 720% of the
annual income of the same household, as previously
defined in the study setting and endorsed by the WHO. In
addition, a composite poverty score in arbitrary units
was derived by principal component analysis from 13
stable socioeconomic variables.
Results: 312 TB patients were invited to participate, of
whom 282 were recruited and had costs and data
available for analysis. TB patient households were poorer
than healthy control households (40% [95%CI34-45]
versus 27% [95%CI22-32] in the lowest poverty tercile,
P , 0.002). Comparing 2014 and 2004 costs data, it was
found that: pretreatment total costs as a proportion of
mean household monthly incomes were lower in 2014
(0.51 [95%CI0.4-0.6] versus 1.1 [95%CI1.0-1.2], P
, 0.0001) despite symptom duration prior to diagnosis
being the same (median 30 days); the proportion of
patient households incurring catastrophic costs (44%
[95%CI38-50] versus 39% [95%CI36-43], P 0.1,
Figure) and total costs as a proportion of mean
household monthly incomes during the entire illness
(2.2 [95%CI1.9-2.4] vs. 2.3 [95%CI2.2-2.4], P 0.1)
were similar (Figure).
Conclusion: The proportional decrease in pretreatment
total costs between 2004 and 2014 require further
exploration and may relate to increased costs of TBrelated travel and food, better coverage of national
health insurance, or improved healthcare access. However, this decrease was offset by a proportional increase
in total costs during treatment. Thus, in these Peruvian
shantytowns, the financial impact of having TB disease
has not diminished over the last 10 years despite a
decrease in TB case notification rates. This may hamper
further TB control, especially amongst high-risk patients
such as those from the poorest households or with MDRTB.
36. Molecular diagnostics: pitfalls and

improvements
PC-1001-05 XpertW MTB/RIF for the diagnosis
of smear-negative pulmonary TB in Ethiopia:
can bleach concentration increase detection of
M. tuberculosis?
M T Jano,1 D Beyene,1 G A Ayana1 1Jimma University,
Jimma, Ethiopia. e-mail: mulualemt.tadesse@gmail.com
Background: Recently the number of smear-negative

cases is more than the smear positive pulmonary TB
cases. This is a peculiar picture seen in Ethiopia for over a
decade. The national TB Prevalence survey in 2010/11
showed that smear negative cases accounted for 57% of
culture positive cases. This indicates the need for more
sensitive and specific diagnostics for improving the
diagnosis of smear negative TB cases. The GeneXpert
MTB/RIF assay has been endorsed by the WHO as a
replacement for sputum smear microscopy. However,
there is no report regarding the performance of bleach
concentrated sputum for the diagnosis of smear negative
TB by Xpert.
Design/Methods: Two sputum specimens (spot and
morning specimen) were collected from each of patients
with suspected smear negative TB. The spot sputum was
analysed for culture (Lowenstein
Jensen & Mycobacteria
Growth Indicator Tube 960) and direct Xpert test. The

morning sputum was processed for direct and bleach
concentrated Xpert tests. The sensitivity and specificity
of the Xpert with 95%CI was calculated by using culture
as reference method.
Results: Out of 185 patients with suspected smear
negative TB, M. tuberculosis was confirmed in 16
(8.6%) cases by the composite reference standard (L-J
and/or MGIT 960). The detection rate of direct Xpert
was 5.4 % (10/185) on spot and 6.5% (12/185) on
morning specimen while the detection rate increased to
11.4% (21/185) by bleach concentrated Xpert test. More
than half of Xpert positive cases on bleach treated
sputum specimen were culture negative. Direct Xpert
had sensitivity of 50.0% and specificity of 97.6%. The
sensitivity and specificity was 56.2% and 93% respectively by bleach concentrated Xpert test. The sensitivity
of direct Xpert test on spot and morning sputum was
similar (50.0%). Xpert on bleach treated sputum had
4.9% incremental yield than the direct one. All Xpert
positive cases were sensitive for rifampicin.
Conclusion: A significant proportion of patients with
suspected smear negative TB missed on conventional
Ziehl-Neelsen method were detected on direct and
bleach concentrated Xpert test. Bleach concentration
method for Xpert had relatively higher sensitivity than
the direct one and can potentially improve the diagnosis
of smear negative TB. The use of only one sputum sample
could be sufficient for TB diagnosis with Xpert test.
PC-1002-05 XpertW MTB/RIF (CBNAAT): an

initial analysis of errors, indeterminates and
invalid results
P Desikan,1 S Dhawan,2 A Pauranik,1 S Mirza,1
S Mannan,3 A Sangwan,4 N Kulshreshta,5 M Pandey1
1
National Reference Laboratory, Bhopal Memorial Hospital &
Research Centre, Bhopal, 2International Union Against
Tuberculosis and Lung Disease / Central TB Division, New
Delhi, 3World Health Organization Country Office for India,
New Delhi, 4CHAI/Central TB Division, New Delhi, 5Central TB
Division, New Delhi, India. e-mail:
prabhadesikan@yahoo.com
Background: Indias NTP has deployed 119 Xpert MTB/

RIF machines. Xpert MTB/RIF is a sophisticated
instrument & test failures can occur. Identifying causes
for test failures can help institute corrective measures. We
assess the performance of Xpert MTB/RIF at NRL,
Bhopal Memorial Hospital & Research Centre
(BMHRC), India.
Design/Methods: Retrospective analysis of report failures generated from Xpert M MTB/RIF software &
causes for test failures (indeterminate, no results, invalids
and errors) from April 2014-March 2015 at NRL,
BMHRC.
Results: Of 1052 samples tested on Xpert, 735 were
pulmonary and 317 non-pulmonary. There were 392
valid results from pulmonary& 263 from non-pulmonary samples. Initial test failures in pulmonary samples
included 3.13% (23/735) indeterminates, 2.17% (16/
735) no results, 2.17% (16/735) invalids & 11.7% (86/
735) errors. Initial failures from non pulmonary samples
included 2.52% (8/317) indeterminates, 1.57% (5/317)
no results, 0.31% (1/317) invalids & 12.6% (40/317)
errors. Reasons for test failures included inadequate
sample processing in 91/126-(72.2%) samples, {73.2%
(63/86) pulmonary & 70% (28/40) non-pulmonary}.
Equipment malfunction led to 4.6% (4/86) failures in
pulmonary samples only. 52.3% (11/21) tests failed due
to power failure {(62.5% (10/16) pulmonary, 20%(1/5)
non-pulmonary)}; Xpert-computer communication error
led to 13.5% (17/126) failures {9.3%(08/86) pulmonary
& 22.5% (9/40)non-pulmonary}; & temperature related
errors caused 2.38%(3/126) tests to fail{2.32% (2/86)
pulmonary & 2.5 % (1/40) non-pulmonary}. To resolve
failures, the available 119 pulmonary samples were
retested. Initial retesting resolved 78 & next retesting
resolved 11. 39 non pulmonary samples were retested.
Initial retesting resolved 33 & next retesting resolved
none.
Conclusion: Overall test failures were low {18.53%
(195/1052) initial, 0.66% (7/1052) final}. Commonest
reason was inadequate sample processing. Appropriate
sample viscosity & homogeneity can avoid this. The NTP
can highlight importance of sample preparation in the
guidance document. Regular maintenance can prevent
equipment malfunction & Xpert-computer communication errors. Invalids can be reduced by avoiding PCR
inhibitors in samples. Since many test failures were
resolved & additional TB cases identified by retesting the
2nd sample, the NTP should retain the policy of
S355
collecting 2 samples for diagnosis by Xpert to ensure

that an additional sample is readily available for
retesting.
PC-1003-05 Rapid TB diagnosis using

GeneXpertW: promising yet fragile?
N Rachmayanti,1 R Chandra,1 B Sonata,1
C Widaningrum,2 D A Riana,3 A Rukmana4 1KNCV
Tuberculosis Foundation, Jakarta, 2National Tuberculosis
Control Program, Jakarta, 3Directorate of Medical Services
Support, Jakarta, 4National TB Reference Laboratory
Microbiology UI, Jakarta, Indonesia. Fax: (62) 218 379
3353. e-mail: novia.rachmayanti@kncvtbc.org
Background and challenges to implementation: To

increase MDR TB case detection, since 2012 Indonesia
has been implementing GeneXpert as a rapid TB
diagnostic tool. WHO recommendation in 2013, has
encouraged the country to expand the diagnostic
services. Under support from USAID and Global Fund,
41 GeneXpert 4 modules has operated and has conducted 13,858 tests by the end of 2014. Besides the benefits,
the expansion of service also bring challenges to the
country to be dealt with. One of them is the increase of
error result which has been notified since the second year
of implementation, particularly sites with high work
load.
Intervention or response: A total 41 sites at 28 out of 34
provinces has been fully operating in 2014. Each site has
been trained centrally and locally for basic knowledge
about processing sample and maintenance, and supervised 3-6 months after installation. National TB Program, BPPM, National Reference Lab, and TB CARE I/
KNCV analyzed the data of GeneXpert reports (41 sites,
2 0 1 2 - 2 0 1 4 ) ,
s u b m i t t e d
t o
genexpert.indonesia@gmail.com, and conducted supervisions on selected sites for a closer look to the error
cause.
Results and lessons learnt: Data 2014 showed the error
rate reached 5.3% (740/13 858), higher than tolerated
level ,5%. Since there is no real time reporting system in
the country, only 487 out of 740 errors are notified with
error code. Data suggested the errors were more likely
because of cartridge related issues (error code 5006
(38%), 5007 (22%), 5011 (21%)) rather than sample
processing errors (error code 2008 (8%)) or hardware
failures (6%). Machines have been operating for two
years tend to have higher module failure rate. Despite of
all machines are under warranty, 9 out of 164 modules
have been replaced due to failure. By comparing the same
sites that were installed in 2012, the calibration failure
rate increased from 0% (0/24) in 2013 to 41.6% (10/24)
in 2014. Cartridge related issues increased significantly
on Jul (7%) and Aug (10%) 2014. The cartridges were
the same batch procured in one shipment on Nov 2013 (8
months storage). Error rates decrease on Sept (6%), Oct
(4%), Nov (2%), and Dec (4%) 2014. The cartridges
were a new batch came to the country in separate
shipments with 4 months interval. The storage had met
the standard, but errors happened on Jul-Aug 2014
slightly indicate was caused by a long storage.
S356
Conclusions and key recommendations: Expansion of

GeneXpert service should be in line with a good
monitoring system, including test results monitoring,
error results, and logistics management. Procurement
with at least 3 years warranty package is recommended
to cope up with the problems.
PC-1005-05 Effect of sputum quality on

microscopy and XpertW MTB/RIF results in the
detection of Mycobacterium tuberculosis
PC-1004-05 The outcomes of XpertW MTB/RIF

test scale-up in Kenya: machine utilization
rates since installation
Background: The common problem in management of TB

is misdiagnosis of cases. Determining the sputum quality
in relationship to bacillary yield is important in enhancing
prompt detection of cases and preventing wasting of
laboratory resources. In this study we compared the effect
of sputum quality in diagnosis of TB using microscopy
and Xpertw MTB/RIF (GeneXpert).
Design/Methods: Spot and morning sputum specimens
were collected from persons presumed to have pulmonary TB enrolled in nine East Africa Public Health
Laboratory Networking (EAPHLN) Project study sites in
Kenya. The specimens were appropriately packaged and
transported to KEMRI Mycobacteriology research laboratory where they were macroscopically characterized
into muco-purulent; mucoid, salivary or blood stained.
Each specimen was processed for microscopy (Ziehl
Neelsen and fluorescent), GeneXpert testing and culture.
Results: A total of 839 sputum specimens with microscopy, GeneXpert and culture results were used to
determine the diagnostic performance of Ziehl Neelsen,
fluorescent and GeneXpert. The highest sensitivity was
obtained with muco-purulent sputum (92.9%, 92.9%,
and 87.5% respectively) compared to salivary sputum
(44.2%, 53.5% and 76.7% respectfully). Diagnostic
performance of ZN, FM on salivary specimens indicated
higher sensitivity in HIV negative (52.4(95%CI:31.073.8)% and 66.7(95%CI: 46.5-86.9)% respectively)
though not significant than HIV positive
(42.9(95%CI:17.0-68.8)% and 42.9(95%CI:17.068.8)% respectively). Using salivary sputum, GeneXpert
sensitivity was higher in HIV positive (85.7%; 95%CI,
67.4-100 %) than HIV negative (71.4%; 95%CI, 52.190.7) and significantly different compared to ZN
microscopy.
Conclusion: The significance difference in the sensitivity
of salivary specimens between ZN microscopy and
GeneXpert in HIV positives emphasizes the importance
of advocating for use of good quality sputum specimens
for detection of M. tuberculosis especially when using
microscopy compared to GeneXpert.
J Ogoro,1 E Masini,1 M Kamene,1 T Miura2 1National TB

Program Kenya, Nairobi, 2Japan International Cooporation,
Nairobi, Kenya. Fax: (254) 271 3198. e-mail:
jeergo@yahoo.com
Background: The Ministry of Health Kenyan through the

National Tuberculosis Leprosy and Lung Disease Program has been rolling out the Xpert MTB/RIF assay since
2011. The NTP through various support has a total of 70
machines whose implementation has been in a phased
manner. Each of the consignment had one year worth of
cartridges, calibration kits and one year warrant. In
2014, Antiretroviral Therapy guidelines and the GeneXpert algorithm were revised in which Xpert MTB/RIF test
was recommended as the first test for diagnosis of
tuberculosis among Persons Living with Human Immunodeficiency virus and children ,15 years. The rollout
was to improve timely detection of TB and identify
rifampicin (RIF) resistance so that they may initiate
appropriate treatment as soon as possible.We describe
herein the utilization outcomes of the 70 machines two
years since arrival and installation.
Design/Methods: Before the procurement was initiated,
facilities were assessed for their preparedness to meet set
criteria. The sites were then identified based on available
previous years data of HIV and TB caseloads. Prior to
machine delivery and installation, priority health Care
workers were trained on testing procedures and clinical
algorithm for Xpert MTB/RIF test. At installation, each
machine was given a target of processing at least 8
specimens per day. The facilities were monitored every
month to track utilization electronically and through
submission of reports from facilities based on installation
dates, expected tests and actual tests done.
Results: Of the 70 machines, 63 (90%) had data
uploaded into the TB program reporting system. As of
April 2015, there were 45 246 specimens processed
within the 70 facilities with tuberculosis identification
rate of 23% as seen in table below: Of the 63 facilities,
only 15 (23.8%) machines were able to process on
average 4 to 10 specimens per day. The usage of (31%) of
the machines was between 25% and 45%, the remaining
67% had utilization below 25%.
Conclusion: Overall, the machines were under-utilized
although the general trend shows the longer the live
period, the better the utilization results except. The error
rates were below 5% implying good laboratory training
and practices.
F Orina,1 M Mwangi,2 W Githui1 1Center for Respiratory

Disease Research, Kenya Medical Research Institute, Nairobi,
2
Center for Public Health Research, Kenya Medical Research
Institute, Nairobi, Kenya. e-mail: orinafred@gmail.com
PC-1006-05 Usefulness of the XpertW MTB/RIF

assay in gastric aspirates for the diagnosis of
tuberculosis in adult patients
S Tahseen,1 C Schmotzer,2 F M Khanzada,1 A Hussain,1
E Qadeer,3 W Aslam1 1National TB Reference Laboratory,
Islamabad, 2Rawalpindi Leprosy Hospital, Islamabad,
3
National Tuberculosis Control Programme, Islamabad,
Pakistan. e-mail: sabira.tahseen@gmail.com
Background: Pakistan is 4th high burden country for

Tuberculosis and MDR-TB. In Pakistan Xpert MTB/RIF
assay is recommended for diagnosis of drug resistant TB
in patient at risk and for diagnosis of TB in vulnerable

groups or those having difficulty in diagnosis. Diagnosis
of tuberculosis is challenging in adults patient who are
unable to expectorate sputum or good quality sputum.
Usefulness of Xpert MTB/RIF assay was evaluated in
gastric aspirates for diagnosis of Tuberculosis in adults
Design/Methods: Retrospective cross sectional study.
Adult patient who had gastric aspiration (GA) performed
for diagnosis of PTB based on X-rays finding suggestive
of tuberculosis but sputum smear was either negative or
not performed because of difficulty in expectoration
were included. Children under 15yrs and adults patient
reported AFB smears positive on sputum prior to gastric
aspiration were excluded. AFB microscopy and Xpert
MTB/RIF assay was performed on sediment of gastric
aspirate in National TB Reference laboratory
Results: GA (GA) from 489 adult patients received in

2013-14 were evaluated, 68 (13.9%) of these patient
were reported AFB smear negative on sputum and others
had difficulty in expectorating good quality sputum.
92.6% (453) of these samples were received from single
TB care facility in city. Of all included 254 were females
and 235 male and mean age was 38.7 yrs. 399 patients
(81.6%) were at risk of MDR because of history of
previous TB treatment (389) or contact with MDR
patient and 18.4 %( 90) had no known risk of MDR. GA
of 115 (23.5%) patients were positive on AFB microscopy whereas M. tuberculosis was detected on Xpert
MTB/RIF assay in 152(31.3%) cases. Use of Xpert
provided bacteriological diagnosis in 47 cases reported
negative on AFB microscopy and in 8/115 cases reported
AFB positive, M. tuberculosis was not detected on Xpert.
Error or no result was reported in 3.7 %( 18) of total
tested. GA of 18.9% of new and 24.6% of previously
treated were positive on AFB microscopy compared to
27.8% and 31.8% on Xpert MTB/RIF assay. RIF
resistance was detected in 8.0% of new cases and
20.5% of patient with history of previous treatment (P
0.141). There was no significant difference in RIF
resistance (P 0.229) in smear positive (21.0%) and
smear negative cases (12.8%)
Conclusion: Use of Xpert MTB/RIF assay in Gastric
aspirate is recommended for diagnosis of tuberculosis in
adult patients at risk of MDR who encounter difficulty in
expectorating good quality sputum.
S357
PC-1007-05 Comparison of the XpertW MTB/RIF

assay in bronchoalveolar lavage and gastric
lavage samples
A Kumar,1 Ravi Mahat1 1Jinnah Posrgraduate Medical
Centre, Karachi, Pakistan. e-mail:
ashokpanjwani2002@hotmail.com
Background: About 50% suspected cases of pulmonaryt

tuberculosis do not present sputum. For diagnosis of
these cases either bronchoscopy with BAL, Induced
sputum or Gastric lavage is used. Bronchoscopy with
BAL requires special facilities, trained staffs and is not
accessible in many regions because of the limited
resources. Additionally, due to its invasive nature many
patients may not prefer and tolerate this. Although yield
of induced sputum has favorable results but its use
requires isolation rooms with negative pressure which
may not be available in many health care facilities.The
GL method is preferred in diagnosis of TB in children
who swallow their sputum and cannot expectorate.
Many studies has shown good result of GL for AFB
smear and culture. In the centers where who has provided
GeneXpert machine, it can be very helpful to make
bacteriological diagnosis within 2 hours.
Objective: To compare the yield from gastric lavage (GL)
and Broncho alveolar lavage (BAL) samples in adult
patient suspected case of TB but not producing sputum
Design/Methods: Total 80 consecutive adults with
suspected case of tuberculosis who were not producing
sputum were recruited. 72 patients prospectively subjected to one gastric lavage and followed by Bronchoalveolar lavage on the same morning. The collected
samples were subjected to GeneXpert MTB/RIF Assay.
Results: Sample of 72 patients were collected. Mean age
was 38.6618.3. 41(56.9%) were male and 31(43.1%)
were female. History of TB contact was present in
34.8%. Total 37 (51.4%) patients had GeneXpert MTB/
RIF positive on BAL and/or GL samples. The GeneXpert
MTB/RIF of BAL fluid was positive on 35(48.6%),
which was not significantly greater than that for
specimens from GL i.e. 28 (38.9%) (P . 0.05). In 26
(36.1%) cases GeneXpert MTB/RIF was positive both in
both BAL and GA samples.
Conclusion: This study showed that the yield of
GeneXpert MTB/RIF in GL was equally effective to
BAL to detect Mycobacterium tuberculosis complex.
Patients who cant produce sputum, GL can be a good
alternative to BAL to detect Mycobacterium tuberculosis
complex in resource poor areas and patient who doesnt
tolerate Bronchoscopy.
S358
PC-1008-05 The use of national XpertW MTB/

RIFs cycle threshold as an audit indicator for
program and laboratory performance
L Scott,1 K Schnippel,2 S Molapo,3 R Berhanu,2 L Berrie,3
A Van Rie,4 W Stevens1,3 1University of the Witwatersrand,
Johannesburg, 2Right to Care, Johannesburg, 3National
Health Laboratory Service, Johannesburg, South Africa;
4
University of Antwerp, Antwerp, Belgium. e-mail:
lesley.scott@nhls.ac.za
Background: The South African National Priority

Program has performed .5.6million Xpert MTB/RIF
tests since 2011 in 207 smear microscopy centres. All
GeneXpert instruments are connected by a laboratory
information system (LIS) and all test results collected in a
central data warehouse. This data not only contains the
qualitative result [positive or negative for Mycobacterium tuberculosis complex (M. tuberculosis c) and
rifampicin (RIF) status (sensitive, resistant or indeterminate)] and pathology review, but also the quantitative
molecular characteristics of each test in the form of the
cycle threshold (Ct) value for each molecular beacon
hybridized (probe A-E) to rpoB region and the internal
control. We investigated the utility of these Ct values.
Design/Methods: Ct values from 24 months (January
2013 December 2014) across all nine provinces from
Xpert MTB/RIF version G4 were analysed in Stata v13.
Results: A total of 4 034 831 Xpert tests were
successfully reported, with 11.3% detecting M. tuberculosis c and 452 368 (99.3%) generating Ct values in the
LIS. Rifampicin resistance was detected in 6.6% (n 29
876), with a frequency of the molecular probes (dropout
or delayed hybridization) reporting Rif resistance as:
E:531bp (55.7%), D:526bp (18.6%), B:516bp (18.2%),
A (6.0%), C(1.5%). This frequency pattern varied across
provinces: in the Eastern Cape E:(48.8%) and B:(31.1%)
while in the Western Cape E:(68.6%) and B:(9.72%).
Amongst RIF resistant tests, the ratio of delayed probe
(DCt.4) to drop-out probes was 1:4.76 (IQR 4.19,
5.35), with increased variability in delayed probes. This
ratio is lower than the 1:9 relationship previously
reported. Probe B accounted for 32.9% of delayed
hybridizations results. There was also variability in the
Rif indeterminate rate (n 11 265, 2.4% [IQR 1.7%2.9%]) by week over the two years. The average national
Ct (mycobacterial burden) for probe C was 22.6
(95%CI: 22.56, 22.64), with a significantly lower disease
burden in Gauteng Ct23.0 (22.9, 27.38) and significantly higher burden in the Western Cape Ct21.8
(21.68, 21.89).
Conclusion: The molecular characteristics of Xpert
MTB/RIF showed geographic variability of M. tuberculosis c, mycobacterial burden (Ct value) and assay
performance. Using the LIS, these audit indicators can
be observed in near real time down to sub-district level
and assay lot number which may assist in improving
laboratory algorithms, program monitoring and national
surveillance.
37. Rapid diagnosis the world over

PC-1009-05 Evaluation of the use of XpertW
MTB/RIF to diagnose and treat MDR-TB under
routine program conditions in Botswana
R Boyd,1,2,3 L Kuate,4,5 T Kuebrich,1 C Modongo,6 T Lere,4
B Kgwaadira,4 H Cox,2 A Finlay1,3 1Centers for Disease
Control Botswana, Gaborone, Botswana; 2University of Cape
Town, Cape Town, South Africa; 3Division of Tuberculosis
Elimination, Atlanta, GA, USA; 4Ministry of Health, National
Tuberculosis Program, Gaborone, 5University of Botswana,
Gaborone, 6Botswana University of Pennsylvania Partnership,
Gaborone, Botswana. e-mail: icg7@cdc.gov
Background: The Botswana MoH revised tuberculosis

(TB) guidelines in 2011, incorporating Xpertw MTB/RIF
(Xpert) for diagnosis of TB among people living with
HIV and those at risk of multidrug-resistant (MDR) TB;
34 Xpert devices were installed from 10/2012 to 12/2014
in 26/28 districts. WHO recommends that people with
rifampicin resistance (RR) detected by Xpert initiate
category IV MDR-TB treatment while waiting for
confirmatory drug-susceptibility testing (DST). An assessment was conducted to determine the rate of RR-TB
and how patients with positive or indeterminate RR
Xpert test results were managed.
Methods: Xpert test results were extracted from archived
files covering the period from 10/2012 12/2014.
Patients with a positive RR test result were identified
and patient lab and clinical data was collected from the
National TB Reference Laboratory and MDR TB
treatment facilities to evaluate patient management.
Results: Data were obtained from 31 of 34 Xpert sites;
three sites were unable to archive data and were
excluded. A total of 9086 Xpert tests were conducted;
1147 (12.6%) detected Mycobacterium tuberculosis (M.
tuberculosis), 86 (7.5%) detected RR from 76 patients
and 48 (4.2%) had an indeterminate RR result from 47
patients (see Figure). Overall, 68/76 (89.5%) patients
with an RR test result initiated Category IV treatment
and 49/76 (64.5%) patients had confirmatory DST; 4/49
had discordant genotypic and phenotypic results. The
median time from the Xpert RR test result to treatment
initiation was 6 days (IQR: 2-16 days); 78.5% were
initiated within 20 days. Median time to phenotypic DST
was 79 days (IQR; 50-114 days). Five (6.6%) patients
with an RR result were not registered with the MDR-TB
program. Of the 47 people with RR indeterminate
results, 20 (43%) had a second Xpert test in which no
RR was detected.
Conclusions: Improvements are needed to follow-up
patients with RR indeterminate results and confirm
diagnosis. Most patients with RR detected by Xpert
rapidly started Category IV treatment in Botswana;
however confirmatory DST was not always conducted
and is important for patient management. The national
TB diagnostic algorithm should be updated to guide
clinicians on interpretation and management of patients
with discordant genotypic and phenotypic results. This
analysis enabled the program to identify previously
unknown patients with an RR test result; the program
is now working to locate and enroll these patients into

care.
PC-1010-05 Cost-effectiveness of the XpertW

MTB/RIF assay for tuberculosis diagnosis in
Brazil
M F T Pinto,1 R Steffen,2 F Cobelens,3 S Van Den Hof,3
A Entringer,1 A Trajman2,4 1Oswaldo Cruz Foundation, Rio
De Janeiro, 2Rio de Janeiro Federal University, Rio de Janeiro,
Brazil; 3KNCV Tuberculosis Foundation, The Hague,
QC, Canada.
Netherlands; 4McGill University, Montreal,
e-mail: ricardo.steffen@gmail.com
Background: The Xpert MTB/RIF assay is being

implemented as a substitute for sputum smear microscopy (SSM) in many low- and high-tuberculosis burden
countries, including Brazil, a country with low multidrug
resistance and moderate HIV co-infection rates.
Design/Methods: We conducted a cost-effectiveness
analysis considering a hypothetical cohort of 100 000
individuals, stratified by HIV status, suspected of having
TB and going through all the TB diagnostic procedures
recommended by the Brazilian National Guidelines. We
estimated the incremental cost-effectiveness ratio (ICER)
of one Xpert MTB/RIF as a substitute for two SSM for
diagnosing susceptible tuberculosis from the Brazilian
National TB Program (NTP) perspective. The main
endpoint was the number of bacteriologically confirmed
TB cases detected. The secondary endpoint was the
number of false-positive TB diagnoses avoided. Because
of the low MDR-TB prevalence in Brazil, we focused this
analysis on first-line drug treatment only. Costs for
confirming each additional case and for avoiding
empirical treatment from false-positive diagnoses were
estimated. The model was parameterized using data on
bacteriological confirmation and clinical diagnosis
among presumptive TB patients derived from the
pragmatic trial of Xpert MTB/RIF in Brazil, and using
TB treatment outcomes from the National TB Information System.
Results: The ICER was US$ 437 for each additional TB
bacteriological confirmation and US$ 23 489 for each
false tuberculosis diagnosis averted assuming an 80%
specificity of clinical diagnosis using both strategies. The
S359
model was highly sensitive to changes in the specificity of

clinical diagnosis following laboratory testing, when
increasing the specificity of clinical diagnosis after an
Xpert MTB/RIF by 1 percent-point steps, the newer
technology became dominant at a specificity of 83%.
The two-way sensitivity analyse showed that Xpert
MTB/RIF would remain dominant for avoiding false
diagnoses if the specificity of clinical diagnosis after a
negative Xpert MTB/RIF was 73% higher than that
after two negative SSM.
Conclusion: Although the NTP has no threshold for costeffectiveness, our model can support decision-makers in
Brazil and other countries with low resistance rates. A
potential economical benefit was shown in case physicians rely more on a negative Xpert MTB/RIF result.
PC-1011-05 Modeling demonstrates that

targeted use of XpertW MTB/RIF is a costeffective and affordable option for
tuberculosis diagnosis in Addis Ababa,
Ethiopia
A Bika,1 D F Adhana,2 D Assefa Lemma,2 E Klinkenberg,2
S Blanco,3 I Langley4 1Addis Ababa City Admin Health
Bureau, Addis Ababa, 2KNCV Tuberculosis Foundation, Addis
Ababab, 3Addis Ababa University, Addis Ababa, Ethiopia;
4
Liverpool school of Tropical Medicine, Liverpool, UK. e-mail:
atabrish@gmail.com
Background: To reduce the global tuberculosis (TB)

burden, active disease must be diagnosed quickly and
accurately. In Ethiopia, however, TB diagnosis mainly
relies on spot-morning-spot sputum smear analysis using
Ziehl-Neelsen (ZN) staining techniques, which fails to
identify some people who have TB. So the objective of
our study was to project the effects of alternative
diagnostic algorithms on patient, health system and
costs, and identify the most applicable algorithm.
Design/Methods: An observational quantitative modeling framework was applied using the Virtual Implementation approach. The model was designed to represent
the operational and epidemiological context of Addis
Ababa. We compared the base diagnostic algorithm with
eight different alternative algorithms of light emitting
diode (LED) fluorescence microscopy and Xpert MTB/
RIF.
Results: If society were willing to pay up to the average
per-capita GDP (US$630 for Ethiopia) for each averted
disability-adjusted life-year (DALY), our results suggest
that three of the modeled algorithms are cost-effective,
and would have important patient, health system and
population level effects in the context of Addis Ababa:
Full rollout of Xpert MTB/RIF would avert 91170
DALYs (95% credible interval [CrI] 54 888.3 127
447.5) with an additional health system cost of US$ 11.6
million over the next 10 years. The incremental costeffectiveness ratio (ICER) is $370.1 per DALY averted.
Same day LED combined with Xpert MTB/RIF targeted
to HIV positive and High multi drug resistance (MDR)
risk groups would avert 73600 DALYs( 95% CrI
48372.8 - 99213.6) with an additional cost of US$5.1
S360
million over the next 10 years. The ICER is $168.8per

DALY averted. The third alternative is same-day LED
fluorescence microscopy at ICER of $12.6 per DALY
averted. The other options would produce DALY gains at
a higher incremental cost and are dominated by other
options in the context of Addis Ababa.
Conclusion: Full rollout of Xpert MTB/RIF is predicted
as the best alternative to substantially reduce the TB
burden of Addis Ababa and is cost effective. However,
the additional financial burden would be substantial and
beyond the budget of the TB control program. So
targeted use of Xpert MTB/RIF for HIV positive and
High MDR risk groups with same-day LED fluorescence
microscopy for other presumptive TB cases would be a
good affordable alternative in area where financial
resources are limited.
PC-1012-05 Comparative performance and

reproducibility of XpertW in the diagnosis of TB
in different geographical settings in Kenya
W Githui,1 M Mwangi,2 F Orina1 1Center for Respiratory
Disease Research, Kenya Medical Research Institute, Nairobi,
2
Center for Public Health Research, Kenya Medical Research
Institute, Nairobi, Kenya. Fax: (254) 202 729 308. e-mail:
wgithui@yahoo.com
Background: Overall performance of GeneXpert for

diagnosis of pulmonary tuberculosis has recently been
documented in Kenya and is comparable with previous
findings from a few studies in African. To our knowledge, no study has compared performance and reproducibility of GeneXpert results in different geographical
settings within the same country. Our study set to
undertake to determine these aspects.
Design/Methods: A cross-sectional study was conducted
between February 2013 and December 2014 in four
selected public health facilities namely Busia, Kitale,
Machakos and Malindi. People with presumptive TB
aged 18 years and above were consecutively recruitment
at each site. Spot and morning sputum were collected. At
the study site, a proportion of each specimen was
processed for GeneXpert. The remainder of the specimen
was shipped to the TB research laboratory at the Kenya
Medical Research Institute (KEMRI), Nairobi, where
specimens were processed for GeneXpert and culture in
accordance with standard procedures. Lowenstein
Jen
sen (LJ) culture was used as a goldstandard.
Results: A total of 1576 specimens with culture results
for M. tuberculosis were included in the analysis.
Performance of GeneXpert across different geographical
location in Kenya was comparable. At KEMRI, sensitivity of GeneXpert ranged from 84.8(95%CI: 72.6-97.1%)
for Busia to 91.1(95% CI: 84.8-97.4%) for Kitale. At the
sites, sensitivity of GeneXpert was 82.4(95%CI: 69.194.9%) for Busia and 91.3(95% CI: 85.9-96.7%) for
Kitale. Using homogeneous sample done at the sites and
KEMRI, a high level of reproducibility of GeneXpert was
observed at KEMRI. However, sensitivity of GeneXpert
varied, insignificantly between sites ranging from
66.7(95% CI: 13.4-100%) in Busia to 100(95% CI: N/
A%) in Machakos, all were reproduced at KEMRI.

Specificity was also comparable and reproducible in all
the sites including KEMRI.
Conclusion: Overall performance of GeneXpert in the
four sites remained constant and comparable with recent
findings from Kenya. Comparable variations in sensitivity observed in different sites were also reproduced at
KEMRI, indicating that if GeneXpert is performed
meticulously, results are reproducible regardless of
geographical locations. Further studies involving more
sites are required to ascertain these findings.
PC-1013-05 Countrywide roll-out of XpertW

MTB/RIF in Swaziland: the first three years of
implementation
W Sikhondze,1 D Khumalo,1,2 T Dlamini,1 G Maphalala,3
H Albert,4 J Wambugu,4 S S Ade,5 A D Harries5 1National
Tuberculosis Control Programme, Manzini, 2National TB
Reference Laboratory, Mbabane, 3Swaziland Health
Laboratory Services, Mbabane, Swaziland; 4Foundation for
Innovative New Diagnostics, Geneva, Switzerland;
5
Paris, France. e-mail: welile.sikhondze@gmail.com
Setting: All 18 public health laboratories in Swaziland

that had Xpert MTB/RIF machines installed as part of
countrywide roll-out between June 2011 and June 2014.
Objective: To evaluate the utilization and functionality of
Xpert MTB/RIF from 2011 to mid-2014
Design: Descriptive study of Xpert MTB/RIF implementation using routinely collected data.
Results: There were 48 829 Xpert tests conducted. Of
these, 93% were successful of which 14% detected
Mycobacterium tuberculosis with 12% showing rifampicin resistance. The commonest cause of unsuccessful
tests was an Error result (62%). Similar findings were
obtained in government-supported and partner-supported laboratories. Annual utilization of Xpert MTB/RIF
improved from 51% of maximum capacity in 2011 and
2012 to 74% in 2013 and 2014. A monitoring and
supervision exercise of all Xpert testing sites in 2014
showed a generally good performance with over 50% of
laboratories achieving a score 780% on most components. However, poor scores were obtained with equipment use and maintenance (6% achieving a score of
780%), internal audit (19% achieving a score of
780%) and process control (25% achieving a score of
780%).
Conclusion: Countrywide roll-out of Xpert MTB/RIF in
Swaziland has been successful although some operational issues have been identified and need to be resolved.
PC-1014-05 Use of XpertW MTB/RIF to diagnose

TB in reference-level mycobacteriology
laboratory in Port-au-Prince, Haiti
A Duncan,1 M Mabou,2 P Christos,1 D Fitzgerald,1 W
Morose,3 P Joseph,2 J W Pape,1,2 O Ocheretina1,2 1Weill
Cornell Medical College, New York, NY, USA; 2Les Centres
GHESKIO, Port-au-Prince, 3National Tuberculosis Program,
Port-au-Prince, Haiti. e-mail: ayd2001@med.cornell.edu
Background/Objectives: The traditional diagnostic algorithm for tuberculosis in Haiti relies on smear microscopy of three sputum samples. This study aims to assess the
performance and feasibility of a diagnostic algorithm
including the Xpert MTB/RIF assay compared to smear
microscopy for patients with clinically suspected TB in
Haiti. Design: This is a retrospective analysis of data
collected by GHESKIO laboratory in Port-au-Prince,
Haiti. Three sputum samples were collected using a
spot-morning-spot protocol. Between June 2011 and
July 2012 all three samples from 4043 patients were
analyzed by smear microscopy. Between September 2012
and February 2013, microscopy for the morning sample
from 1650 patients was substituted with the Xpert MTB/
RIF assay.
Results: 1) The morning sample has a higher diagnostic

value than spot samples (91.4% positivity rate compared
to 81.8% and 86.8% for first and second spot samples,
respectively), 2) a single Xpert MTB/RIF test detected
more TB cases in HIV-positive patients than smear
microscopy of three samples (17.2% vs. 13.9%, P
0.06), and detected as many TB cases in HIV-negative
S361
patients as smear microscopy of three samples (22.1%

vs. 23.0%), 3) after implementation of Xpert MTB/RIF,
the diagnostic yield of smear microscopy for a third
sputum sample was reduced from 2.29% to 0% in the
HIV-negative group and from 6.59% to 3.45% in the
HIV-positive one, and 4) reduction in the laboratory
workload compensated for the higher cost of the
GeneXpert reagents.
Conclusions: A TB diagnostic algorithm with one Xpert
MTB/RIF test and one AFB smear on two collected
sputum samples results in a higher diagnostic yield in
HIV-positive patients compared to the traditional three
sputum smear diagnostic algorithm. Routine use of the
Xpert MTB/RIF assay to diagnose TB is feasible in a
resource-limited setting like Haiti.
PC-1015-05 GeneXpertW shortens time to

treatment initiation for Latvian MDR-TB cases
H Stagg,1 P White,2,3 V Riekstina,4 A Cirule,4 J Brown,1,5,8
G Dravniece,6,7 E Adam,7 C Jackson1 1University College
London, London, 2Imperial College School of Public Health,
London, 3Public Health England, London, UK; 4Centre of TB
and Lung Diseases, Riga East University Hospital, Riga, Latvia;
5
Royal Free London NHS Foundation Trust, London, UK;
6
KNCV Tuberculosis Foundation, The Hague, Netherlands;
7
Otsuka SA, Geneva, Switzerland; 8Centre for Respiratory
Medicine, Royal Free London NHS Foundation Trust, London,
UK. e-mail: h.stagg@ucl.ac.uk
Background: Rapid diagnosis of multidrug resistant

tuberculosis (MDR-TB) is vital for the prompt initiation
of appropriate treatment. In 2010 GeneXpert MTB/RIF
was recommended by the World Health Organization
and adopted by Latvia for individuals considered at high
MDR-TB risk. We sought to establish the impact of
GeneXpert on time to treatment initiation in Latvia.
Methods: The 2009-2012 Latvian MDR-TB treatment
cohort was evaluated, excluding cases with solely
extrapulmonary TB and prison cases. Use of GeneXpert
was categorised as either a binary (not used (baseline);
used) or ternary variable (not used (baseline); cases not
rifampicin resistant (RIF-R) TB by GeneXpert; cases
RIF-R TB by GeneXpert). Time (in days) between
presentation to hospital and MDR-TB treatment initiation was calculated. Descriptive analyses were undertaken, followed by univariate Poisson regression. Multivariable Poisson models adjusting for confounding and
effect modification were built.
Results: 398 cases were available for analysis; 10 were
excluded due to missing dates and 6 more due to being
solely extrapulmonary. 259 did not have GeneXpert
(100%, 84%, 46%, 40% of cases from 2009-12,
respectively), 15 were not identified as MDR-TB by
GeneXpert (11 TB not detected, 3 RIF sensitive, 1
sensitivity undetermined), 108 were RIF-R TB by
GeneXpert. The median time to treatment initiation
was 27 days (95% confidence interval 22-30) for cases
where GeneXpert was not used and 5 days (3-7) for cases
with GeneXpert. The univariate rate ratio (RR) for the
association between the use of GeneXpert and time to
treatment initiation was 3.05 (2.40-3.87), P value ,0
S362
.001 and the multivariable RR (controlling for gender,

age, region, site and HIV) 3.40 (2.52-4.60), ,0.001.
Results: controlling for year (excluding 112 cases from
2009 when GeneXpert was not used in Latvia) were
similar. The median time to treatment initiation was 58
days (7-99) when GeneXpert did not detect RIF-R TB
and 4 days (2-6) when it did. Univariate RRs using no
GeneXpert as the baseline were 0.63 (0.37-1.07) and
6.52 (5.09-8.36), respectively (P value 0.01). A multivariable model gave similar results (0.55 (0.32-0.94),
6.26 (4.96-7.89); ,0.001).
Conclusions: The use of GeneXpert in Latvia shortened
the time to treatment initiation for MDR-TB cases. Cases
not diagnosed as RIF-R TB by GeneXpert had similar
times to treatment initiation as those for whom
GeneXpert was not used. This project was supported
through a grant by Otsuka SA.
PC-1016-05 Use of GeneXpertW to support fully

ambulatory treatment of TB in Kyrgyzstan
M Joncevska,1,2 G Kurbaniyazova,3 S Kalon,3
G Kalmambetova,4 T Mohr1 1Project HOPE, Almaty,
Kazakhstan; 2Project HOPE, Skopje, Macedonia, Yugoslav
Rep; 3Project HOPE, Bishkek, 4National TB Institute, Bishkek,
Kyrgyz Republic. Fax: (389) 2270 0601. e-mail:
joncevs.kam@yahoo.com
Background and challenges to implementation: Kyrgyzstan is a high TB and MDR-TB prevalence country in
Central Asia, with TB case notification rate of 128 per
100 000 population, reporting one of the highest MDRTB rates globally. According to the data presented from
the National Drug Resistance Survey, 26.4% of newly
diagnosed TB cases and 51.6 % of previously treated
patients have MDR-TB. The current approach in TB
treatment includes hospitalization of all TB patients
during the intensive phase of treatment and full hospital
treatment for MDR-TB.
Intervention or response: In order to support the ongoing
Health Sector Reform and rationalization of health care
services, the NTP of Kyrgyzstan developed a plan for
reducing the number of beds in inpatient TB facilities, by
introducing full ambulatory treatment of TB. Xpert M.
tuberculosis Rif testing was introduced to support the
new policy for ambulatory treatment, integrated in the
Primary Health Care level laboratory services to make
the rapid diagnosis and timely start of treatment
available for TB patients close to their place of living.
The new policy was introduced through a pilot project,
implemented within the USAID Quality Health Care
Project, during the period 2012 2014. Issyk Atta
district, located in the Southeast of Chui Province, was
selected as a pilot implementation site. GeneXpert
platform was placed in the district policlinic, providing
access to M. tuberculosis Rif testing for 3 districts and
one city covering a population of 270 000 people.
Results and lessons learnt: During the period 2012-2014,
a total of 1819 M. tuberculosis Rif tests were done and
461 TB cases were confirmed as M. tuberculosis . 140
cases among them were Rif resistant. Following the NTP
diagnostic algorithm, it was possible to immediately start
the treatment for TB and MDR-TB. Ambulatory

treatment was started for 148 TB patients, including 32
resistant cases. Based on the results from Issyk Atta, the
ambulatory model was rolled-out to other districts in
Chui Oblast.
Conclusions and key recommendations: 1. Initial results
from the pilot project support the NTP plan for
ambulatory treatment of TB, followed by the nationwide
scale-up of Xpert MTB/RIF testing. 2. Further downsizing of existing inpatient TB facilities is a process, which
requires step wise approach, leading to full outpatient
treatment for a larger number of TB patients. 3. A critical
issue for countrywide implementation of Xpert MTB/
RIF would be development or reliable sputum transportation mechanisms
38. Phenotypes and genotypes:

deciphering resistance development
PC-1017-05 Molecular and phenotypic
detection of fluoroquinolone resistance
A Aubry1 1National Reference Center for Mycobacteria,
Paris, France. Fax: (33) 1 45 82 75 77. e-mail:
alexandra.aubry@upmc.fr
As a consequence of the use of fluoroquinolones (FQ),

resistance to FQ has emerged, leading to nearly
untreatable, extensively drug-resistant tuberculosis. M.
tuberculosis conventional phenotypic susceptibility testing is cumbersome and takes several weeks. The
recommended critical concentration, for FQ susceptibility testing, defined by WHO as the ability of the bacilli to
grow on medium containing 2 or 4 mg/l ofloxacin is to a
large extent based on consensus rather than scientific
evidence. Moreover, from a medical point of view, using
a single breakpoint does not allow delineating the
different levels of resistance, thus impeding therapeutic
decision about whether maintaining or not FQ depending on the resistance level. Therefore, molecular assays
allowing rapid detection of resistance to antituberculous
agents are of major interest in routine practice. For FQ,
these assays, detecting mutations within the DNA gyrase
Quinolone Resistance Determining Regions, may be able
to predict FQ level of resistance if properly used. We will
review the advantages and drawbacks of the available
phenotypic and genotypic methods for FQ resistance in
M. tuberculosis, in regard with the laboratory capacities.
PC-1018-05 Slow recovery of Mycobacterium

africanum-infected patients post-treatment in
associated with high content of persistor-like
bacilli in sputum
L Tientcheu,1 A Bell,2 N J Garton,2 J Sutherland,1
H Dockrell,3 B Kampmann,1 M R Barer2 1Medical Research
Council Unit The Gambia, Banjul, Gambia; 2University of
Leicester, Leicester, 3London School of Hygiene & Tropical
Medicine, London, UK. Fax: (220) 449 5442. e-mail:
ltientcheu@mrc.gm
Background: Mycobacterium africanum (M. africanum)

causes 40% of all tuberculosis (TB) cases in The Gambia.
It grows more slowly in culture, has similar transmission
capacity but slower progression to active disease
compared with Mycobacterium tuberculosis (M. tuberculosis). Our previous work comparing M. africanum
versus M. tuberculosis-infected patients before and
following anti-TB treatment showed a similar profile of
clinical and immunological parameters pre-treatment,
but significant differences in body mass index, chest Xray scores and antigen-specific immune responses between M. tuberculosis and M. africanum-infected patients post-treatment. We hypothesise that these differences could either be due to significant differences
between the strains in relation to drug resistance or the
percentage of persister-like lipid bodies positive bacilli.
Design/Methods: To explore these hypotheses, drug
susceptibility testing of pre-treatment sputum isolates
was carried out using the BACTEC MGIT 960 method,
and the content of acid-fast bacilli (AFB) and lipid bodypositive bacilli in sputum were analysed using Auramine
and LipidTOX Red neutral lipid staining respectively.
Results: Low levels of drug resistance were observed in
both groups. Two MDR (resistant at least to rifampicin
and isoniazid) and three isoniazid mono-resistant cases
were seen among M. tuberculosis-infected patients, while
there were three streptomycin mono-resistant Mafinfected patients. Although AFB-positive bacilli percentages were similar between the groups (P 0.821), lipid
body-positive bacilli percentages were significantly
higher in M. africanum (n 24) compared to M.
tuberculosis (n 36)-sputum (P 0.0059). In addition,
bacilliary length was significantly higher in M. africanum
than M. tuberculosis-sputum (P 0.0007).
Conclusion: The slow recovery of M. africanum relative
to M. tuberculosis-infected patients post-treatment might
be at least partially associated with the persistence of
drug tolerant fat and lazy bacilli.
PC-1019-05 Clinico-immunological diagnostic

criteria for the progression of multidrugresistant tuberculosis
A Alenova,1 E Berikova,1 B Toksanbaeva,1 D Sabazova1
1
National Center for Tuberculosis Problems, Almaty,
Kazakhstan. e-mail: arike.alenova@mail.ru
Background: Despite a decrease of incidence of tuberculosis and its mortality rate in Kazakhstan, there is a
growth of the progressive forms of the multidrugresistant tuberculosis (MDR-TB). The objective of this
S363
study is to make a comparative analysis of immunity of

the lung tuberculosis patients with the preserved
sensitivity to the anti-TB drugs (group A) and with the
progressive MDR-TB (group B).
Design/Methods: CD3, CD4, CD8, CD16, CD95; IL-1,
IL-2, IL-4 IL-8, IFN-g and antituberculous antibody
levels have been studied in ELISA on the peripheral blood
of 62 A-group patients and 50 B-group patients.
Results: It has been observed that the patients of group B
had the following difference with the patients of group A:
1) Group B had a reasonably more frequent reduction of
CD4 as compared to group A (24.462.3% versus 42.4 6
3.1% in group A) and of CD16 (5.1 6 0.3% vs. 10.8 6
0.5% in group A). 2) Group B had a reasonably more
frequent increase of CD95 lymphocytes (49.9 6 3.2% vs.
28.1 6 2.2% in group A) and of antituberculous
antibodies (0.75 6 0.09 absorbance units [au] vs. 0.29
6 0.05 au). It was defined that the deficiency of IL-2 of
less than 165.0 pcg/ml and of IFN-g of less than 15.1 pcg/
ml causes the phagocytosis and macrophage depression
and leads to a severe specific process in the lungs.
Conclusion: It is found that one of the tuberculosis
progression reasons is a serious disorder of the immune
system, namely the balance change of CD4, CD16 and
CD95 lymphocytes and the lymphocyte-derived IL-2 and
IFN-g cytokines.
PC-1020-05 To ascertain the proportion of drugresistant tuberculosis among people living

with HIV in Abia State
O Okorie,1 M Gidado,2 C Osakwe,3 E Ubochioma,2
V Ibeziako,2 G Akang,2 R Eneogu2 1Tuberculosis Control
Program, Abia, 2Tuberculosis Unit, Abuja, 3Tuberculosis Unit,
Enugu, Nigeria. e-mail: onphil200@yahoo.co.uk

emergence of drug resistance Tuberculosis (TB) is a
major public health challenge and threatens progress
made in global TB control. The 2012 National Tuberculosis drug resistant prevalence survey conducted in
Nigeria showed a prevalence of 2.9% among new TB
cases and 14.3% for retreatment TB cases. Drug
resistance TB arises due to improper use of antibiotics
in chemotherapy of drug-susceptible TB. This is as a
result of, administration of improper treatment regimens
and failure to ensure that patients complete the whole
course of treatment. Diagnosis of TB among PLWHIV is
difficult and problematic especially those with low CD4
count. Consequently there is under diagnosis and over
diagnosis of HIV associated Tuberculosis. With the
advent of Xpert MTB/RIF, diagnosis of HIV associated
TB and Rifampicin resistant TB is more expedient and
within reach. It is against this backdrop that the State
control program proposed a study to ascertain the
proportion of TB among PLWHIV with Xpert MTB/
RIF assay.
Intervention or response: All the presumptive drug
resistant TB cases and symptomatic PLWHIV were
examined with GeneXpert and recorded in DR-TB
register.
S364
Results and lessons learnt: A total of 1392 Drug resistant

presumptive cases screened with gene xpert machine in
2014, 17 % ( 240) were M. tuberculosis positive and 2
%( 31) were rifampicin resistant. More so out of 416 TB
symptomatic HIV patients screened, 4% were rifampicin
resistant
Conclusions and key recommendations: The result
depicts that PLWHIV has 50% risk of developing
resistant tuberculosis when compared to the general
population
confirm our hypothesis that recognizable strain characteristics can predict which strains will not grow using
routine DST recommendations.
PC-1022-05 Resistance pattern, prevalence and

predictors of fluoroquinolone resistance
among multidrug-resistant tuberculosis
patients
N Ahmad,1 A H Khan,1 S A S Sulaiman,1 A Javaid2
Universiti Sains Malaysia, Pulau Pinang, Malaysia;
2
Postgraduate Medical Institute, Peshawar, Pakistan. e-mail:
nafeesuob@gmail.com
1
PC-1021-05 Morphological characteristics of

Mycobacterium tuberculosis strains failing to
grow on 7H10 agar proportion drug
J Farooqi,1 K Jabeen,1 S Shafiq,1 S Ali,1 R Hasan1 1Aga
Khan University, Karachi, Pakistan. e-mail:
joveria.farooqi@aku.edu
Background: Drug susceptibility testing (DST) of M.

tuberculosis in the Aga Khan University (AKU) Clinical
Laboratory is performed by agar proportion method on
Middlebrook 7H10 agar after preparing a 1:10 dilution
of the organism suspension as recommended by CLSI. In
our experience certain strains of M. tuberculosis fail to
grow on DST when tested with 1:10 dilution of the
inoculum, delaying the DST report. In this study we aim
to describe the strains which do not grow on DST with
1:10 dilution but grow well without diluting the
inoculum, so that they may be identified before DST is
set up, preventing delayed reports.
Design/Methods: DST was performed on M. tuberculosis
strains obtained from culture on LJ, and MGIT
subculture on 7H10 agar as recommended by CLSI
(M24-A2 2011). Briefly, a 0.5 McFarland (McF)
organism suspension was made and diluted 1:10 times
before inoculating 0.1 ml onto 7H10 agar 4 quadrant
plates with a drug free quadrant for growth control.
Strains which failed to grow in the growth control were
chosen to be re-tested in parallel with a 1 McF
suspension, with and without 1:10 dilution and the
results recorded. Treatment history and description of
morphology on LJ and 7H10 agar for these strains were
also noted.
Results: Out of 149 strains tested in January and
February 2015, 14 M. tuberculosis strains failed to grow
on routine DST and were tested in parallel with and
without dilution of 1 McF organism suspension. All 14
strains grew on undiluted DST, while10 strains did not
grow on 1:10 dilution. Details of treatment and colony
morphology are given in table below: DST result (14
cases) Treated (history available in 7 cases) 7H10
(morphology recorded in 7 cases) LJ (morphology
recorded in 14 cases) Dry and rough Wrinkled and
raised Dry Hard/ wrinkled Growth on dilution 2 1 0 1 3
No growth on dilution 5 3 3 3 7
Conclusion: M. tuberculosis strains that fail to grow on
DST by dilution were more likely to have hard, wrinkled
colonies and isolated from patients on ATT. With an
adequate sample size, the results of this study may
Background: Fluoroquinolones form the backbone of

MDR-TB treatment regimen. Despite MDR-TB high
burden country, little is known about drug resistance
pattern, and prevalence and predictors of fluoroquinolones resistance in MDR-TB patients from Pakistan.
Design/Methods: This was a cross-sectional study
conducted at programmatic management unit of drug
resistant TB (PMDT), Lady Reading Hospital Peshawar,
Pakistan. Two hundred and forty three newly diagnosed
MDR-TB patients consecutively enrolled for treatment at
study site from January 1, 2012 to July 28, 2013 were
included in the study. A standardized data collection
form was used to collect patients socio-demographic,
microbiological and clinical data. SPSS 16 was used for
data analysis.
Results: Majority patients were females (55.6%), Pakistanis (91.8%), aged 21-40 years (46.1%), baseline
weight .40kg (62.1%), rural residents (77%), previously treated for TB (91.4%), non-smokers (87.2%),
previously treated for TB at a public facility (54.7%)
and had no history of SLD use (95.1%). Only 39.5%
patients had cavities on baseline chest X-ray and
concurrent co-morbidity (14.4%). High degree of drug
resistance (median 5 drugs, range 2-8) was observed.
High proportion of patients was resistant to all five firstline anti-TB drugs (62.6%) and second line drugs
(55.1%). Among three first-line anti TB drugs, the rate
of resistance was highest for pyrazinamide (97.5%)
followed by ethambutol (81.1%) and streptomycin
(70%). Resistance to ofloxacin was observed in 52.7%
patients. Upon multivariate analysis, previous TB treatment at private (OR 1.953, P 0.034) and public
private mix (PPM) sectors (OR 2.824, P 0.046) were
predictors of fluoroquinolones resistance.
Conclusion: The high degree of drug resistance observed
particularly to fluoroquinolones is alarming. Irrational
prescriptions, frequent use of fluoroquinolones for
respiratory tract infections other than TB and nonprescription sale of antibiotics are the probable reasons
for high fluoroquinolones resistant Mycobacterium
tuberculosis. Adoption of more restrictive policies to
discourage the use of fluoroquinolones, and training
doctors in both private and PPM sectors to diagnose and
manage TB in line with national guidelines is necessary
for halting further increase in fluoroquinolones resistance.
PC-1023-05 Profil de resistance

aux
antituberculeux de souches provenant de
patients presentant
une tuberculose
pulmonaire ou extra-pulmonaire en 2012
1
A Ba Diallo, A Thiam, M Y Fall Niang,

A Ndiaye Diawara,1 S Lo,2 A Gaye Diallo1 1Laboratoire de
Bacteriologie
Virologie, Unite de Mycobacteries
du CHU
Aristide Le Dantec, Dakar, 2UFR Sciences de la Sante,

Universite Gaston Berger, Saint Louis, Senegal. Fax: (221)
338 213 825. e-mail: choux_ba@yahoo.fr
Introduction: Letude de la sensibilite aux antituberculeux est importante dans le cadre de leradication de la
tuberculose surtout pour les pays qui ne disposent que du
niveau central pour faire ce test. Le diagnostic de la
tuberculose dans les pays en voie de developpement est
base sur la microscopie et très souvent le patient est mis
sous traitement sans que la nature de la souche ne soit
connue. Ce travail avait pour but de determiner les
profils de resistance parmi des patients nouvellement
infectes ou en retraitement recus pour diagnostic au
laboratoire hospitalier du CHU A. Le Dantec.
Methodologie:Les e chantillons e taient decontamines par
la methode NALC puis mis en culture sur milieu
Lohenstein Jensen et Bactec MGIT 960. Les tests de
sensibilite e taient effectues sur milieu liquide MGIT960
avec les molecules de première et seconde ligne
Resultats: Les patients comportaient 149 nouveaux cas et
6 cas de retraitements. Ils e taient ag
es de 14 a` 82 ans avec
une mediane de 33,5. Les hommes representaient 62%
versus 37%pour les femmes et le statut HIV e tait de
10,3%. La proportion de ceux qui presentaient une
tuberculose pulmonaire e tait de 87.09% contre 12.6%
de tuberculose extra pulmonaire. Les souches e taient
isolees de prelèvement extra pulmonaires comprenant
des Pus, Liquide pleural, LCR, Liquide articulaire, Urines
et de prelèvements dorigine pulmonaire. La resistance
aux antituberculeux parmi les nouveaux cas e tait
respectivement pour la Streptomycine de 13.4%,
(20.8%) pour lIsoniazide, (12.08%) pour la Rifampicine et lEthambutol (8.7 %). Tandis que les patients en
retraitement avaient 83.3% de resistance a` la Streptomycine et isoniazide et 50% de resistance a` la Rifampicine et Ethambutol. La mono resistance e tait respectivement chez les nouveaux patients de 6,04%, 7,3%, 0,6%
et 1,3% respectivement pour S, I, R, E. Les cas de
multiresistante parmi les nouveaux cas e taient de 7%
pour HR, 1,34% pour HRE et HRS et 2,68% pour
HRES. Aucun cas de souches ultraresistantes navait e te
detecte parmi celles testees.
Conclusion: La problematique des patients nouvellement
infectes qui sont mis sous traitement sans le profil de
sensibilite de la souche causale doit e te mis en exergue
puisque cela pourrait encore accroitre les risques
dapparition de souches multiresistantes aux antituberculeux disponibles.
S365
PC-1024-05 Feasibility of molecular testing for

drug resistance as part of a TB clinical trial
screening protocol
P Shete,1,2 H Nguyen,3 N Hung,4 T L Luu,5 N Hung,4
M Tran,2 H Phan,6 N Phuong,4 A Cattamanchi,1,2
P Nahid1,2 1San Francisco General Hospital, San Francisco,
CA, 2University of California, San Francisco, San Francisco,
CA, USA; 3Hanoi Lung Hospital, Hanoi, 4National Lung
Hospital, Hanoi, 5Hanoi Department of Health, Hanoi, Viet
Nam. Fax: (1) 415 206 1551. e-mail: pbshete@gmail.com
Background: Late exclusion of TB clinical trial participants due to drug resistance is a costly and resourceconsuming process. Sample size calculations for clinical
trials need to be inflated in anticipation of late
exclusions. Molecular testing of drug resistance directly
on sputum has the potential to address these challenges.
As part of a CDC-TB Trials Consortium (TBTC) study in
Hanoi, Viet Nam, participants were screened for
enrollment using standard interview, sputum smears
and culture, and Hain Genotypew MTB-DRplus (Hain
Lifescience, Nehren Germany), a molecular test that
detects the presence of M. tuberculosis (TB) as well as
isoniazid (INH) and rifampin (RMP) resistance. We
sought to assess the feasibility and performance of using
MTB-DRplus directly on sputum as a screening tool for
clinical trial enrollment.
Design/Methods: This cross sectional study was conducted during consecutive screening of 64 patients for
enrollment into TBTC Study S29X between February
and October 2012. MTB-DRplus was performed directly
on smear positive sputum specimens within 24 hours of
collection in addition to standard enrollment procedures.
We calculated the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of
MTB-DRplus using results of drug susceptibility testing
(DST) on Lowenstein-Jensen (LJ) media as the reference
standard.
Results: Results were missing for eight patients (six for M
MTB-DRplus and two for LJ DST), two patients had
indeterminate MTB-DRplus results and one patient
tested negative for M. tuberculosis. Among the remaining 51 patients, MTB-DRplus identified 9 of 12 with
INH-resistant TB by LJ DST (sensitivity 75%, 95%CI
43-94%), all of whom had mutations in katG. MTBDRplus was negative for INH-resistance in all 39
patients with INH susceptible TB by LJ DST (specificity
100%, 95%CI 91-100%). MTB-DRplus results were
available within 36 hours, enabling all 9 patients with
positive molecular results for INH resistance to be
excluded from the clinical trial prior to randomization
and instead be referred for treatment by the Viet Nam
National TB Programme.
Conclusion: Use of MTB-DRplus directly on sputum is a
feasible component of the screening process for TB
clinical trials. Diagnostic performance characteristics are
similar to that reported in non-trials settings. Further
studies are needed to quantify the time and cost savings
of routine use of MTB-DRplus during eligibility screening for clinical trials focused on drug-susceptible TB.
S366
Table Performance characteristics of MTBDRplus in S29X as

compared to LJ- DST
Rifampin
Isoniazid
MDR
Sensitivity
(95% CI)
Specificity
(95% CI)
PPV
NPV
3/3, 100%
(31-100)
9/12 75%
(43-94)
2/2 100%
(19-100)
48/48, 100%
(93-100)
39/39 100%
(91-100)
49/49 100%
(93-100)
100%
(31-100)
100%
(66-100)
100%
(19-100)
100%
(31-100)
93%
(80-98)
100%
(93-100)
Conclusion: The results of the present study suggest that

early marriage, womens malnutrition, lack of financial
autonomy and living in crowded families has independent significant positive association with prevalence of
TB among women in Afghanistan. Efforts should be
centered on the enforcement of legislation to discourage
early marriage, removing the barriers for womens
financial autonomy and increasing the quality of living
conditions in order to combat the spread of tuberculosis
in Afghanistan.
39. Tuberculosis drug resistance and

susceptibility testing in a nutshell
PC-1025-05 Determinants of pulmonary
tuberculosis among women in Afghanistan
S D Mahmoodi,1 M Seddiq,1 M R Aloudal,2 M Mashal,1
S S Mohammad,1 G Qader3 1Minstry of Public Health,
Kabul, 2World Health Organization Country Office, Kabul,
3
Challenge TB, Managment Sciences for Health, Kabul,
Afghanistan. e-mail: ntp.mahmoodi@gmail.com
Background: Afghanistan is one of the 22 TB highburden countries and TB is a major public health
problem in the country. WHO estimates that approximately 61,000 all types of TB cases occur every year with
incidence rate of 189/100 000 population per year. More
female cases (particularly, aged 15 to 45 years old)
compared to males are seen for any form of TB infection
(female to male ratio 2:1). Causes of this predominance
are not known and this important information gap has
been repeatedly highlighted. This study aims at filling
this void by examining the likely risk factors including
socio-cultural, nutrition, economic, environmental, and
health seeking behavioral factors that may be underlying
this high burden of TB among women.
Methodology: A case-control study was conducted
whereby a sample of new smear positive pulmonary TB
cases among adult females was enrolled from a randomly
selected sample of health facilities. An equal number of
adult males diagnosed with smear positive pulmonary TB
living in the same community was selected as first
controls and an equal number of adult males diagnosed
with smear positive pulmonary TB from a different
community were selected as the second control.
Results: Women who were younger than 18 years at the
time of their first marriage (early marriage) had 25 times
higher risk of getting TB compared to those married at 31
years or older. Severely malnourished women were 2.72
times more at risk of getting TB compared to wellnourished people. Women living in families where five or
more people sleep in the same room during the night
reported higher risk of getting tuberculosis compared to
those where less than five people slept under the same
roof. The study indicated that women who use private
health service providers have higher proportion of
tuberculosis compared to those who used public health
services.
PC-1026-05 Evaluation of the BACTEC MGIT 960

system for bedaquiline susceptibility testing of
G Torrea,1 N Coeck,1 C Desmaretz,1 T Van De Parre,1
T Van Poucke,1 N Lounis,2 B C De Jong,1 L Rigouts1
1
Prince Leopold Institute of Tropical Medicine, Antwerp,
2
Janssen Infectious Diseases, Beerse, Belgium. Fax: (32) 3
247 63 33. e-mail: gtorrea@itg.be
Background: After the approval of Bedaquiline (BDQ) by

the U.S. Food and Drug Administration (FDA) the
compassionate use of the drug and expanded access
programs have been established in many countries,
creating the need to monitor the susceptibility pattern
of strains isolated from patients receiving BDQ. The lack
of bacteriological data lead us to evaluate the feasibility
and accuracy of Minimal Inhibitory Concentration
(MIC) determination of BDQ in BACTEC MGIT 960
versus the Middlebrook 7H11 (M7H11) method.
Design/Methods: In a first phase the wild-type MIC
distribution was determined to establish the Epidemiological cut-off (ECOFF) or breakpoint to declare a strain
as resistant to BDQ, using 47 strains from BDQtreatment-nave patients both in MGIT960 and on
M7H11 agar. In the second phase the accuracy of
MGIT960 was evaluated versus M7H11 using 74
S367
probable susceptible and 18 probable resistant strains.

Repeatability and reproducibility of MGIT960 were
assessed using 5 strains showing different BDQ resistance
levels.
Results: The wild-type MIC distribution ranged from
60.03 mg/L to 1.00 mg/L in MGIT960 and from
60.008 mg/L to 0.25 mg/L on M7H11 with suggested
ECOFFs of 1.00 mg/L and 0.25 mg/L respectively.
Applying these ECOFFs, the probable susceptible and
resistant strains were distinguishable by both methods.
However, only a twofold increased MIC was observed
for one of the resistant strains compared to the ECOFF
both in MGIT960 and on M7H11. Inter-method
agreement to classify the isolates was excellent (100%).
All replicates in the repeatability and reproducibility
experiments fell within the normal range.
Conclusion: MGIT960 proved to be highly accurate and
reproducible relative to M7H11 to determine BDQ MIC.
The small margin between the suggested ECOFF and the
lowest MIC of the mutant strains risks making both
methods prone to discordant results. Further validation
in clinical settings linking with treatment outcome data is
needed.
moreover, turn an increasing MDR epidemic into a

diminishing one. While improved second-line treatment
and private sector engagement remain important factors
in the control of TB in India, our results illustrate the
potential impact of new and emerging technologies that
enable widespread, timely DST.
PC-1027-05 The potential impact of up-front

drug susceptibility testing on Indias MDR-TB
epidemic
J Farooqi,1 K Jabeen,1 S Shafiq,1 R Hasan1 1Aga Khan

University, Karachi, Pakistan. e-mail: joveria.farooqi@aku.edu
K S Sachdeva,1 N Raizada,2 A Sreenivas,3 C Denkinger,2

C Paramasivan,2 S Kulsange,2 C Boehme,2 N
Arinaminpathy4 1Central TB Division, New Delhi, India;
2
Foundation for Innovative New Diagnostics, Geneva,
Switzerland; 3World Health Organization Country Office,
New Delhi, India; 4Imperial College, London, UK. e-mail:
drneerajraizada@gmail.com
Background: In India, as elsewhere, multi-drug resistance (MDR) poses a serious challenge in the control of
tuberculosis (TB). The End TB strategy, recently approved by the world health assembly, aims to reduce TB
deaths by 95% and new cases by 90% between 2015 and
2035. A key pillar of this approach is early diagnosis of
tuberculosis, including universal drug susceptibility
testing (DST), and systematic screening of high-risk
groups. Despite limitations of current laboratory assays,
universal access to rapid DST could become more
feasible with the advent of new and emerging technologies. Here we use a mathematical model of TB
transmission, calibrated to the TB epidemic in India, to
explore the potential impact of a major national scale-up
of rapid DST. Towards this we take GeneXpert
technology as an example that could enable scale-up of
rapid DST, drawing from findings of recent multi-centric
Xpert MTB/RIF demonstration study conducted in India
to construct such a model.
Results and Conclusion: We find that widespread,
public-sector DST, with Xpert MTB/RIF appropriately
linked with treatment, could substantially impact MDRTB in India: 75% access amongst all cases being
diagnosed for TB could avert over 180 000 cases of
MDR-TB (95%CI 44187 317077 cases) between 2015
and 2025. Sufficiently wide deployment of Xpert could,
PC-1028-05 Linezolid susceptibility of XDR-TB

strains from Pakistan tested on the MGIT 960
system
Background: According to WHO 2014 data, Pakistan

has the highest MDR-TB burden in the Eastern
Mediterranean Region and increasing frequency of
XDR-TB amongst MDR strains is a major concern. In
2012 W.H.O. Global TB program released updated
critical concentration for linezolid amongst Group-5
anti-tuberculous agents (ATT) and recommended
MGIT960 as the testing method. In this study we report
the results of linezolid drug susceptibility testing (DST)
on MGIT960 in XDR-TB strains isolated in our
laboratory.
Design/Methods: Nineteen XDR-M. tuberculosis strains
isolated from specimens received at Aga Khan Clinical
Laboratories from July 2014 to January 2015 were tested
against linezolid 1 mcg/ml on Bactec MGIT960 system.
DST results for other first and second-line ATT were also
recorded. Frequencies of resistance were calculated for
each drug amongst the linezolid susceptible strains.
Results: Eighteen out of 19 XDR-TB isolates were found
to be susceptible to linezolid. The linezolid resistant
strain was susceptible only to ethionamide 5 mcg/ml and
resistant to all first-line (streptomycin, ethambutol and
pyrazinamide) and second-line (amikacin, kanamycin,
capreomycin and ofloxacin) drugs tested. Resistance
rates of first- and second-line amongst linezolid sensitive
strains were: 89% (16) streptomycin, 77% (14) ethambutol, 94% (17) pyrazinamide, 61% (11) amikacin, 67%
(12) kanamycin, 100% (18) capreomycin and 39% (7)
ethionamide.
Conclusion: The resistance against linezolid amongst 19
XDR- M. tuberculosis strains from Pakistan was found
low (5%) while resistance against ethionamide was 39%.
This suggests that linezolid and ethionamide may be used
S368
for treatment of XDR-TB in addition to other Group-5

ATT while awaiting susceptibility results.
Test, Xpert MTB/RIF and MODS whether samples were

collected pre-treatment or during therapy.
PC-1029-05 Comparison of the MDR/XDR-TB

Colour Test, XpertW MTB/RIF, and MODS assays
to diagnose rifampicin-resistant TB
M Tovar,1,2, E Ramos,1 T Valencia,1 S Datta,1.2
R Montoya,2 R Montoya,2 T Wingfield,1,2 A Valencia,3
R Patrocinio,1 C Evans1,2,4 1 IFHAD Innovation for Health
and Development, Lima, 2IPSYD Innovacion Por la Salud Y el
Desarollo, Lima, 3DIRESA, Regional National Tuberculosis
Program, Callao, Peru; 4Infectious Diseases & Immunity,
Imperial College London, and Wellcome Trust Imperial
College Centre for Global Health Research, London, UK.
e-mail: marco.tovar@upch.pe
Background: Multidrug resistant tuberculosis (MDRTB) is challenging TB control. Numerous assays have
been developed to screen for MDR-TB. Despite this,
MDR-TB rates are increasing and globally only 45% of
people with MDR TB are estimated to be diagnosed. The
Peruvian NTP uses Microscopic-Observation DrugSusceptibility assay (MODS) as a universal MDR-TB
screening test in patients diagnosed with TB disease.
WHO endorses use of Xpert MTB/RIFw test for rapid
DST in patients at high risk of MDR-TB. The MDR/
XDR-TB Colour Test is a novel non-commercial direct
DST that allows MDR-TB diagnosis and screening for
extensively drug-resistant TB (XDR-TB).
Objective: To compare the concordance of rifampin
resistance testing in three DST assays: MDR/XDRTB
Colour Test, Xpert MTB/RIF and MODS.
Design/Methods: From February 2014 through October
2014, consecutive sputum samples of patients diagnosed
at regional NTP clinics in shantytown communities of
north Lima, Peru, were processed in the Research and
Development Laboratory of Universidad Peruana Cayetano Heredia (LID). All patients signed a written,
informed consent form. All samples were tested for
rifampicin resistance concurrently by the three DST
methods: MDR/XDR-TB Colour Test, Xpert MTB/RIF
and MODS. Each assay was performed by staff who
were blinded to the results of other assays. Concordance
between the test results was analysed as percentage
agreement compared by the z-test of proportions and
McNemars test.
Results: 762 samples were received and processed during
the study period. 283/762 (37%) were culture positive
and had an interpretable DST result in all 3 assays. The
percentage agreement between MODS and MDR/XDRTB Colour Test, MODS and Xpert MTB/RIF, Xpert
MTB/RIF and MDR/XDR-TB Colour Test was: 98.6%,
98.2%, and 98.2%, respectively for all samples. When
this analysis was repeated for the 170 sputum samples
collected pre-treatment there was no significant change
in the agreement between the three DST assays: 99.4%,
98.2%, and 97.6%, respectively. There were no statistically significant differences between the DST results for
any samples groups.
Conclusion: There is high concordance between rifampicin resistance testing results for MDR/XDR-TB Colour
PC-1030-05 Exploring drug resistant

tuberculosis profiles in the West Coast region
of the Western Cape Province, South Africa
P Tshavhungwe,1 R M Warren,1 M Buckley,2 T Dolby,3
P Van Helden,1 J Simpson,3 K Jacobson,2 E Streicher1
1
Biomedical Sciences, Stellenbosch University, Tygerberg,
South Africa; 2Boston University, Boston, MA, USA; 3National
Health Laboratory Services, Cape Town, South Africa. Fax:
(27) 938 9863. e-mail: phophi.tshavhungwe@gmail.com
Background: The West Coast, a rural district in the

Western Cape Province, has one of the highest reported
tuberculosis (TB) prevalence in the Province of 1225/100
000 population. Previous molecular epidemiological
studies have demonstrated that the dynamics of TB
transmission vary geographically. The dynamics governing TB transmission in this region are unknown,
specifically that of drug resistance TB. The aim of this
study was to describe the DR-TB epidemic and identify
transmission hotspots within the West Coast region of
the Western Cape, South Africa.
Design/Methods: We performed a retrospective cohort
study, selecting one isolate from each of the patients
diagnosed with drug resistant TB from 2008-2012 in the
West Coast District. Isolates were received from all
health facilities in the West Coast and incidence of
disease as well as distribution of strains was described for
sub-districts based on these locations. We genotypically
characterised the isolates by the internationally standardised spoligotyping method and by Sanger sequencing
drug resistance conferring genes.
Results: Isolates from 611 patients were genotyped and
grouped according to their spoligotype pattern. Spoligotyping data revealed that X-family is the most dominant
strain family (33 %), followed by the Beijing family (21
%). This is in contrast to previous studies in the rest of
the Western Cape Province that found the Beijing strain
family to be predominant in the province. Within the XFamily, 62 % of strains were MDR, 18 % pre-XDR, and
3% XDR. The X-family is the predominant strain in 3 of

the 5 sub-districts in the region.
Conclusion: Even though spoligotyping is considered to
have a low discriminatory power and could overestimate
the extent of transmission, this study suggest an epidemic
spread of MDR strains in the West Coast region of the
Western Cape Province. Our findings also highlight the
regional variation of outbreaks and the need for
molecular epidemiologic studies in various regions to
tailor interventions to curb TB and drug resistant TB
spread.
PC-1031-05 Direct comparison of

spontaneously expectorated sputum with
sputum expectorated following mouth wash
with water or with chlorhexidine
C Kleinhans,1 M Hanekom,1,2 S Friedrich,2 A Venter,1,2
A Diacon1,2 1TASK Applied Science, Bellville, 2Stellenbosch
University, Tygerberg, South Africa. Fax: (27) 219 389 317.
e-mail: hanekom@sun.ac.za
Background: Mycobacterial sputum cultures are critical

to score conversion from positive to negative in clinical
trials of anti-tuberculosis regimens. Contaminated cultures do not allow clear scoring. We evaluated whether
mouth wash with water or with antibacterial agents can
reduce contaminated sputum culture results.
Design/Methods: Between December 2013 and February
2015 we asked Xpert MTB/RIF positive, sputum smear
microscopy positive or negative, untreated HIV-infected
adults to collect sputum A) expectorated spontaneouly B)
following a mouth wash with water, and C) following
mouth wash with chlorhexidine solution (60 seconds,
Corsodyl, South Africa) in that sequence with at least 30
minutes in between collections, once before initiation of
anti-tuberculosis treatment and at 3, 7, 14, 28, 35, 56,
112 days thereafter. Sputum was transported to a central
laboratory on the same day and processed for liquid
culture in duplicate in routine fashion (MGIT960).
Results of each sputum were reported as positive,
negative or contaminated.
Results: We collected 93 sputum triplicates from 18
initially smear-positive (n 14) and smear-negative (n
4) patients. No adverse events were observed. A Spontaneous B - Water C - Chlorhexidine Positive 68
73.1% 83 89.3% 78 89.3% Negative 18 19.4% 8 8.6%
15 16.1% Contaminated 7 7.5% 2 2.2% 0 0%. The odds
for obtaining a valid result (not contaminated) were
higher with a supervised mouth wash than without (P ,
0.01). Mouth wash with water yielded the most positive
cultures whereas mouth wash with chlorhexidine eliminated contamination, but those differences were not
signifcant.
Conclusion: Water or a chlorhexidine mouth wash
before sputum collection lowers the contamination rate
in liquid culture. A greater number of samples is needed
to determine whether water or chlorhexidine are more
effective.
S369
PC-1032-05 Molecular testing for tuberculosis

in the United States: implications for national
surveillance
L Armstrong,1 T Dalton,1 R Oatis,2 J Higashi,3 S Allen,4
M Pecha,5 A Starks1 1U.S. Centers for Disease Control and
Prevention, Atlanta, GA, 2State of Maryland Department of
Health and Mental Hygiene, Baltimore, MD, 3San Francisco
Department of Public Health, San Francisco, CA,
4
Washington State Department of Health, Seattle, WA,
5
Public Health-Seattle & King County, Seattle, WA, USA. Fax:
(1) 404 639 8959. e-mail: larmstrong@cdc.gov
Background: Rapid molecular tests for the detection of

mutations associated with resistance to anti-TB drugs in
M. tuberculosis are now available worldwide, but their
use varies widely. The U.S. National Tuberculosis
Surveillance System (NTSS) cannot currently incorporate
results of these newer molecular tests, instead, relying on
documenting drug resistance or susceptibility to anti-TB
drugs using conventional growth-based drug susceptibility testing (DST).
Design/Methods: Confirmed tuberculosis (TB) cases
were selected for further study if they were reported to
NTSS as resistant to at least one of four first-line anti-TB
drugs, isoniazid, rifampin, pyrazinamide, or ethambutol.
Between the years 20112012, cases were abstracted for
the states of Georgia (61 cases) and Maryland (27 cases);
20102013 for San Francisco County, California (42
cases); and 20112013 for Santa Clara County, California (80 cases); and 20092013 for Washington state
(53 cases from King County and 46 cases from the
remaining counties). All growth-based DST and molecular test results were abstracted from laboratory reports
in patient medical records.
Results: A total of 309 drug-resistant TB cases were
reported among project sites; 77 (25%) with molecular
test results recorded in the medical record of which 57
(74%) had results detecting mutations associated with
drug resistance. Most isolates were tested for genes
associated with rifampin (67) and isoniazid (68) resistance (Table). Two cases had an isolate with a silent
mutation (one in pncA and one in embB). Six cases had
mutations associated with drug resistance without
comparable growth-based DST results, including four
cases with resistance to second-line injectable drugs
(tlyA) or to fluoroquinolones (FQ) (gyrA), drug classes
that define extensively drug-resistant TB. These cases
were reported in NTSS as resistant to first line drugs, but
either susceptible or not tested to second-line drugs. One
case had results indicating resistance to FQ by growthbased DST and a gyrA mutation, but the growth-based
FQ resistance was not reported to NTSS.
Conclusion: State and local TB programs have access to
molecular tests for identifying mutations associated with
drug resistance and sometimes, molecular results do not
match those from growth-based assays. The inability of
NTSS to capture these results has negative implications
for complete and accurate reporting of drug resistant TB
cases.
S370
Includes eis (1 case) and tlyA (2 cases)
Results: RODAC stamping could detect as little as three

bacteria on slides stamped either 5 or 60 minutes after
inoculation. No contamination was detected. PMAbased PCR, tested in parallel, did not pass validation.
Mycobacteria were still detected after disinfection with
ethanol 70% (applied 5 or 60 minutes after BCG
inoculation). In the BSL-3, from 201 RODAC-stamped
surfaces, M. tuberculosis was detected on four: three
inside a BSC - on a tube cap and on operators gloves and one outside, on an operators gown.
Conclusions: We show that RODAC plates, but not
PMA-PCR, detect mycobacteria at low numbers of
organisms, and can be applied to assess mycobacterial
contamination on surfaces in TB environments. We also
show that 70% ethanol does not reliably sterilize
mycobacteria when applied for 5 minutes to a dried
surface, and that a class II biosafety cabinet may be
inadequate for strict containment of mycobacteria
during manipulation of M. tuberculosis cultures.
40. From safety to supply-chain: lets talk

lab management
PC-1034-05 Distribution and availability of

essential tuberculosis diagnostic items in
Amhara region, Ethiopia
Table
Molecular
test done
Result
Yes, n 77 Mutation
(%) (25)
No, n 232 No muta(%) (75)
tion
Silent
Total
309 Total
(100)
Mutations
associated
with
iso- rpoB pncA embB gyrA
niazid* n
n
n
n
n (%) (%) (%) (%) (%)
53
(78)
15
(22)
68
26
7
(39) (32)
41
14
(61) (64)
1
(5)
67
22
9
(60)
5
(33)
1
(7)
15
Mutations
associated
with
2nd-line
injectables
n (%)
7
3 (7)
(18)
33 38 (93)
(83)
40
41
*Includes katG (32 cases), inhA (5 cases), katG/inhA dual mutations (10
cases), katG/ahpC dual mutation (1 case) and 5 cases with unspecified
mutations associated with isoniazid.
PC-1033-05 Containment of Mycobacterium

tuberculosis in a Class II B biosafety cabinet
during routine operations
G Daneau,1 E Nduwamahoro,1 K Fissette,1
2
P Rudelsheim,
D Van Soolingen,3 B De Jong,1,4,5
1,6 1
L Rigouts
Institute of Tropical Medicine, Antwerp,
2
Perseus BVBA, Zwijnaarde, Belgium; 3National Institute for
Public Health and the Environment, Bilthoven, Netherlands;
4
Medical Research Council Unit, Fajara, Gambia; 5New York
University, New York, NY, USA; 6University of Antwerp,
Antwerp, Belgium. Fax: (32) 3247 6333. e-mail:
gdaneau@itg.be
Background: Guidelines for the manipulation of Mycobacterium tuberculosis (M. tuberculosis) cultures require
a Biosafety-Level 3 (BSL-3) infrastructure and accompanying code of conduct, however the evidence on which
these are based is limited. In this study, we aim to validate
and apply detection-methods for viable mycobacteria
from surfaces in a BSL-3 TB lab.
Methods: We evaluated phenotypic (RODAC plates) and
molecular (propidum monoazide (PMA)-based PCR)
approaches for the detection of viable mycobacteria, as
well as the effect of 70% ethanol applied for 5 minutes
for mycobacterial sterilisation, in a BSL-3 lab with a class
II biosafety cabinet (BSC). For validation of the method,
recovery of serial dilutions of M. bovis BCG from glass
slides was measured. We stamped surfaces in and around
the BSC after different technicians had manipulated high
bacterial load suspensions from liquid or solid medium
for routine drug susceptibility testing (1mg/ml, i.e.
approximately 106 colony forming units/ml), without
prior alert on time of stamping. Technicians had been
trained on biosafety and specific techniques according to
the good laboratory practices in place in our ISO15189accredited lab.
M Sinishaw,1 G Gebregergs,2 M Balew1 1Bahir Dar

Regional Health Research Laboratory Center, Bahir Dar,
2
Bahir Dar University, Bahir Dar, Ethiopia. Fax: (251) 226
6701. e-mail: mulusewalemneh@yahoo.com
Background and challenges to implementation: Adequate supplies of tuberculosis laboratory reagents and
consumables are necessary for tuberculosis diagnosis and
monitoring of treatment response. This study was
initiated to assess the distribution and stock levels of
laboratory reagents and consumables used in tuberculosis control in public health centers of Amhara region,
Ethiopia.
conducted in 82 health centers providing sputum
microscopy services from April 28 to May 26, 2014.
Stock levels were calculated and distribution of reagents
and consumables assessed.
Results and lessons learnt: Thirty three (40.2%) health
centers were under stocked for at least one of the key
items for tuberculosis diagnosis at the time of visit.
Fifteen (18.3%) health centers had no stocks of at least
one of the key laboratory items (9 (11%) methylene blue,
9 (11%) carbol fuchsin, 7 (8.5%) acid alcohol and 3
(3.7%) sputum cups). Of the 82 health centers, 77
(93.9%) did not fulfill the criteria for effective distribution of tuberculosis laboratory reagents and consumables.
Conclusions and key recommendations: There were
many health centers that had no or only low stocks of
key tuberculosis laboratory reagents and consumables as
a result of ineffective distribution system. This seriously
hinders TB control in the area disrupting early diagnosis
and treatment follow up services. It is necessary to
strengthen supply chain management to ensure uninterrupted TB diagnostic service.
PC-1035-05 Trends in the rate of follow up

sputum smear examination and conversion
rate among smear-positive pulmonary TB
patients in two regions of Ethiopia
J Seid,1 N Demmelash,1 W Gebrekirstos,2 F Belachew,3
Z Dememew,1 D Habte,1 M Aseresa Melese,1 P G Suarez4
1
Management Sciences for Health, Help Ethiopia Address the
Low Tuberculosis Performance (HEAL TB) Project, Addis
Ababa, 2Amhara Regional Reference Laboratory, Bahar Dar,
3
Oromia Regional Reference Laboratory, Addis Ababa,
Ethiopia; 4Management Sciences for Health, Center for
Health Services, Arlington, VA, USA. e-mail:
jseidmsh@gmail.com
Background and challenges to implementation: WHO

recommends 2nd, 5th and 6th month follow up smear
microscopy for sputum smear positive (SS) patients.
Smear conversion at these periods is used as an indicator
for good adherence and strong directly observed TB
treatment practice. Patients who failed to convert at these
points are also likely cases of MDR TB and require close
follow-up. At the beginning of the project, the complete
examination rate of SS pulmonary TB patients was
below the standard.
Intervention or response: USAID funded MSH/HEAL TB
project supported Amhara and Oromia regional health
bureaus of Ethiopia through various interventions. Heal
TB trained staffs from health facilities on management of
TB patients using the national guideline. The project
supported the lab diagnostic and follow-up capacities of
the health facilities by providing microscopes, trained
laboratory professionals, and established a decentralized
external quality assurance system. Standard of Care
(SOC) monitoring was introduced to help district TB
focal persons monitor each TB patient status including
the status of sputum follow-up tests. Sputum follow up
test monitoring tool/dashboard was also developed to be
used by health facility TB focal persons.
Results and lessons learnt: We analyzed the routine data
from health facilities in 10 zones that were included
under the project support during the initial phase of
implementation. The number of SS patients who gave
sputum at the end of intensive phase has increased from
80 % (95%CI: 78% - 82%) at baseline to 96 % (95%CI:
95% - 97%) in 2014. It shows that only 4% of smear
positive TB patients did not get follow up test at the end
of intensive phase. In the last three years, smear
conversion rate of SS TB patients at the end of 2nd
month (intensive phase) has improved by 17 % [from
74% (95%CI: 71% - 76%) in 2011 to 91% (95%CI:
90%-92%) in 2014]. Similarly the sputum follow up at
the end of the 5th month increased from 64% to 90% (P
, 0.001) and the cure rate increased from 66 to 88% (P
, 0.001)[Table 1]. For those who failed to convert at
2nd, 5th and 6th month, sputum was sent for culture and
DST or GeneXpert to rule out MDR TB.
Conclusions and key recommendations: District level
monitoring by health managers and local capacity
building of laboratory and TB clinic staff plays a huge
role towards ensuring quality service and treatment
outcome. Indicator assisted close holistic support helps in
S371
improving the performance of TB treatment follow up

practice.
Table Trend of sputum follow-up at the end of intensive phase
and smear conversion, Phase I zones of Amhara and Oromia
regions, 2011-2014
Period
Baseline
Year I/2012
Year II/2013
Year III/2014
Smear conversion
rate at end of
intensive phase
(%)1
Examined rate
at 5th month
(%)
Cure rate
(%)
74
82
89
91
64
80
90
90
66
77
84
88
1
The percentage shows that patients who converted to smear negative. The
remaining could be still smear positive or sputum is not examined
PC-1036-05 The effect of triage in improving

sputum-positive case detection in crowded
out-patient settings: experience of 164 primary
health centres in Bihar
S Mugudalabetta,1 A K Pandey,1 L V Ramana Rao,1
S K Muthusamy1 1Damien Foundation India Trust, Chennai,
India. Fax: (91) 044 2836 0496. e-mail:
admin@damienfoundation.in
Background and challenges to implementation: In 2013,

out of the estimated global annual incidence of 8.6
million TB cases, 2.3 million were in India. Bihar is
second most populous state in India; Revised National
TB Control Programme was implemented in the state in
2005. One of the major challenges faced by the state in
TB control is poor case detection. Though it is estimated
203 TB cases of all types per 100 000 population, the
State is not able to achieve even 35% of the target set by
the programme. One of the main reasons was poor health
infrastructure; each primary health centre covers a
population range from 150 000 to 200 000. Health
facilities are generally overcrowded with patients,
doctors may not have enough time to take complete
history from the patients attending outpatient clinics and
many TB symptomatic cases can be miss the opportunity
to undergo screening for TB.
Intervention or response: Damien Foundation India
Trust is supporting Revised National TB Control
Programme in 15 districts with less than 30 New Sputum
Positive case notification per 100 000 population. One of
the strategies to improve case notification is to promote
triage approach for identification of TB symptomatic in
crowded outpatient clinics at Primary Health Centres.
Out of 293 microscopy centres in 15 districts, 209 DMCs
were selected to implement triage approach based on
high patient load. Lab Technicians were trained to visit
the waiting area every hour and segregate patients with
cough for sputum examination with the approval of
doctor. By the end of 2014, complete data was available
from 164 (78%) DMCs. The data was also collected for
2012 and 2013 from same microscopy centres for
comparison.
Results and lessons learnt: There is significant improvement seen both in screening of TB symptomatic and
sputum positive case detection in 2014 compared to
S372
previous two years. It was observed that screening of TB

symptomatic increased by 18% and 25% in 2014
compared to 2012 and 2013.Sputum positive case
detection increased by 30% and 37% in 2014 compared
to 2012 and 2013.
Conclusions and key recommendations: Implementation
of triage approach can improve both screening of TB
symptomatic and sputum positive case detection in
crowded outpatient clinics.
PC-1037-05 The impact of seasonal factors on

TB treatment loss to follow-up
A Mordovin,1 C Acosta,2 D Pashkevich,1 S Sterlikov,3
I Danilova,4 L Migliorini,2 M Dara2 1World Health
Organization, Moscow, Russian Federation; 2World Health
Organization, Copenhagen, Denmark; 3Federal Research
Institute for Health Organization and Informatics, Moscow,
4
Moscow City Scientific and Practical Center for TB control,
Moscow, Russian Federation. Fax: (495) 787 2119. e-mail:
a.mordovin@whorussia.org
Background: The Russian Federation is a country with

difficult climatic conditions caused by extreme continental climate. We hypothesized the possible impact of
seasonal factors on TB treatment loss to follow-up
(LFU).
Design/Methods: We examined the proportion of LFU
patients during the intensive phase of treatment and
during the entire treatment course. In total, data was
available on 19 923 new cases, including 5259 cases
notified in quarter I (January-March), 2012, 5223 cases
notified in quarter II (April-June), 2012, 4515 cases
notified in quarter III (July-September), 2012, and 4926
cases notified in quarter IV (October-December), 2012.
We determined Relative risk (RR) of LFU taking into
account statistical significance (P ,0.05).
Results: For those patients treated during the intensive
phase of treatment, there was a significant decrease in the
RR of LFU among patients registered in Q1 compared
toQ2 (RR0.77) and Q3 (RR0.76); there was a
significant decrease in the RR among patients registered
in Q4compared to Q2 (RR0.75) and Q3 (RR0.74).
During the whole course of treatment there was a
significant decrease in the RR among patients registered
in Q4 in comparison with Q1 (RR 0.75), Q2 (RR
0.70) and Q3 (RR 0.85) (P ,0.05 for all).
Conclusion: Seasonal factors influence the frequency of
LFU in the intensive phase and entire course of
treatment. A decrease in the LFU RR during the intensive
phase occurs at the beginning treatment in the cold
season, when smear-positive patients start treatment in a
hospital. A decrease in RR LFU over the entire course of
treatment is associated with the continuation phase of
treatment in the cold season, when patients have less
opportunity to participate in seasonal work, which
increases dependence on food packages, if issued. Since
the RR of treatment LFU in the warm time of the year vs.
the cold time is 1.2-1.4, it is advisable to reallocate
resources accordingly.
PC-1039-05 The threat of human resource

shortage to quality laboratory practice in Africa
T Maruta1 1African Society for Laboratory Medcine, Addis
Ababa, Ethiopia. e-mail: talkmoremaruta@gmail.com
Background: The shortage of health workers has been

linked to poor health indicators with Africa estimated to
have 2.3 healthcare workers per 1000 population. Of the
57 countries with critical shortage of healthcare workers,
36 are in Africa and as much as a 140% increase in
healthcare professionals is required for them to achieve
some of the MDG targets. Over the years, the role of the
laboratory in the delivery of healthcare has increasingly
been recognized and consequently its efficiency, costeffectiveness, and quality of service has to be high. To
ensure quality, laboratories implementing a quality
management system (QMS). Implementing a QMS
requires personnel that have a specific set of skills and
knowledge about international standards like the ISO
15189 and their requirements. Given the already existing
workforce shortage, limited personnel with competency
in QMS has threatened the investments in laboratory
system strengthening and quality of laboratory service.
Design/Methods: The ASLM developed a key strategic
goal on workforce development; Strengthen the laboratory workforce in Africa to achieve Millennium
Development Goals for Health by (1) Training and
certifying 30 000 laboratory professionals and clinicians,
(2) Advocating for national laboratory professional
regulatory councils and (3) Developing standardized
frameworks for workforce development, retention, and
improving and maintaining pre-service training capacity.
Results: To date, ASLM has trained 30 Laboratory
professionals in research and publication, 48 in laboratory mentorship, 60 in Laboratory quality management
systems, 24 in accreditation preparedness, 88 certified as
WHO auditors. 2 pre-service curricula have been
reviewed and updated, 1 laboratory regulatory body
established and 14 MOUs signed with laboratory
associations
Conclusion: The availability of expert human resources
continues to be a threat to provision of quality laboratory
services. ASLM continues to provide advocacy for
investment in building human resource capacity through
its 2020 Strategic goal on workforce development.
PC-1040-05 Risk of acute lower respiratory tract

infections in adults associated with household
air pollution from solid fuel use: a systematic
review
H Jary,1 D Havens,1 K Mortimer,1 D Pope,2 N Bruce2
1
Liverpool School of Tropical Medicine, Liverpool, 2University
of Liverpool, Liverpool, UK. e-mail:
hannah.jary@liverpool.ac.uk
Background: Acute lower respiratory infections (ALRI)

in adults cause a large burden of mortality worldwide,
particularly in developing countries. Exposure to household air pollution (HAP) is also high in many developing
countries, as 2.8 billion people rely on burning solid fuels
for essential domestic energy. HAP exposure is known to
S373
be a risk factor for childhood ALRI, but the association

in adults has not been confirmed and quantified. A
systematic review in 2009 found just three studies
investigating this relationship with inconclusive findings.
We aim to update this systematic review from 2010
onwards to compile the evidence, for the purpose of
informing future public health policy decisions regarding
HAP interventions strategies.
Design/Methods: Searches were made of seven English
language databases, using search terms that covered all
sources of HAP and different forms of ALRI. Relevant
studies were selected using a protocol driven process, and
data was extracted and quality assessed using pre-defined
forms.
Results: From 9238 titles, 4 studies were selected,
including three cross-sectional studies and one observational natural experiment. All studies had methodological limitations and their results may have been affected
by systematic biases. Significant methodological heterogeneity was seen between the four studies and so it was
inappropriate to conduct a meta-analysis for calculation
of an overall risk estimate. Of the studies providing a risk
estimate, one found an increased risk of bronchitis with
increased CO exposure (unadjusted odds ratio (OR) 1.66
(95% confidence interval (CI) 1.15-2.40) but the other
two found no significant effect of reported biomass use
or exposure on the risk of ALRI/bronchitis (OR 0.91,
95%CI 0.71-1.15 and OR 2.40, 95%CI 0.86-6.72).
Conclusion: The findings of this systematic review
indicate that there is little robust evidence available and
existing studies do not provide conclusive evidence
regarding the effects of HAP on adult ALRI. Further
prospective studies examining this relationship are
warranted. To ensure comprehensive results, this systematic review is being extended to include non-English
language databases and grey literature from 1960: final
results expected by mid-2015. A thorough analysis of the
existing evidence is essential to understand the contribution of HAP to the high burden of disease caused by adult
ALRI, to facilitate targeted public health interventions if
a need is identified.
munity involvement and assessing the needs of TB

patients is essential for universal access to TB care.
Intervention or response: BRAC design peripheral
laboratory facilities with solar system in very remote
areas and outreach sputum collection centers to improve
access to diagnosis. BRACs frontline health workers
(Shasthya Sebika) disseminate TB messages in the
community and refer TB presumptive for sputum
examination at NTP designated laboratories during
household visits and group meeting. Those persistent
TB symptoms and negative smear referred to higher level
health facilities for X-ray, FNAC and biopsy. Treatment
for all TB patients is initiated by graduate medical
doctors and daily DOT is ensured by Shasthya Sabika.
Results and lessons learnt: Additional 14,037 outreach
sputum collection centers were held between January
2012 to December 2014 and established referral linkage
with 4,967 additional graduate medical practitioners. A
total of 11 peripheral laboratories with solar system were
expanded to cover hard to reach areas. Significant
number of DOT is ensured by these Shasthya Shebikas
in BRAC covered area, service intensification result
improved case notifications from 91 703 in 2011 to 1 24
286 in 2014. In 2013, among 1 01 555 pulmonary
patients, 505 (0.5%) were lost to follow up and out of
1481 retreatment cases, lost to follow up 16 (1.1%) only,
and treatment success rate reached 95%. Case notification of NSP in hard to reach areas also increased from 73/
100 000 population in 2011 to 78 in 2014.
Conclusions and key recommendations: Patient centered
approach by involving community people for showed
better result in increased case notification, maintain high
cure rate, reduce lost to follow up and DOTS expansion.
41. Stepping up patient access and

support
Background and challenges to implementation: National

TB control programme (NTP) initiated programmatic
management of DR-TB (PMDT) in 2008 with 24 months
standard MDR-TB treatment regimen and BRAC is
major implementing partner of NTP. NTP developed
PMDT guidelines in 2009 and updated in 2013, and
patient treated according to guideline. Shasthya Shebikas
are front line health care providers of BRAC, ensures
daily intake of medicine during ambulatory phase of
treatment. They play significant role in community base
MDR-TB management.
Intervention or response: Since 2013 BRAC has been
providing diagnostic support to identify MDR-TB from
presumptive, nutritional and other support to improve
treatment adherence and outcome. Moreover, from 2013
NTP initiated community based programmatic management of DR-TB, treatment modality changes from 6-8
PC-1041-05 Enhancing accessibility and

intensifying TB interventions with patientcentered care: BRAC experience
S Munim,1 S Islam,1 M Siddiqui,1 M Akramul Islam,1
M Rahman,2 K Uddin1 1BRAC, Dhaka, 2National
Tuberculosis Programme, Dhaka, Bangladesh. Fax: (88)02
8823542. e-mail: munim.s@brac.net
Background and challenges to implementation: Ever

since 1984 BRAC has started TB control programme in
Bangladesh and it has been waging this battle against
tuberculosis, currently it serving 93 million populations.
To improve case notification and overcome unfavorable
treatment outcome, the organization introduced Com-
PC-1042-05 Community based MDR-TB

management: BRAC experience from
Bangladesh
S Munim,1 S Islam,1 M Basher,1 M Siddiqui,1
M Akramul Islam,1 M M Rahman,2 S Sultana3 1BRAC,
Dhaka, 2National Tuberculosis Programme, Dhaka, 3World
Health Organization, Dhaka, Bangladesh. Fax: (88) 02 882
3542. e-mail: munim.s@brac.net
S374
months to 1-2 months hospitalization and patients

discharge to community from hospital. Outpatient DRTB team trained and ready to provide community based
DOT. Shasthya Shebikas receive basic orientation on
MDR-TB management at community level, ensures daily
intake of medicine in ambulatory phase and appropriate
referral if any adverse drugs reaction occurs.
Results and lessons learnt: National DR TB patient
enrolment increased of from 376 in 2012 to 717 in 2014
almost double. In 2012 total 336 patients have received
ambulatory phases treatment at community level in
BRAC supported areas. Out of 336 patients, 238 were
successfully completed treatment in the community.
Treatment success rate increased 71% in 2012 cohort
from 68% in 2010 cohort and lost to follow up rate
decreased to 15% in 2012 from 27% in 2008.
Conclusions and key recommendations: Communitybased care is built on three tiers support: clinical care,
socio-economic and psycho-emotional. Community
based PMDT improves case notification and care for
ensuring treatment adherence and outcome.
PC-1043-05 Improved treatment adherence

among patients by making them aware of their
rights and responsibilities through the TB
charter
T Laha,1 S Diwakar,1 V Sameer,1 P Bhat,1 T Thomas1 1The
Catholic Health Association of India, Secunderabad, India.
e-mail: sameer-valsangkar@chai-india.org
Background and challenges to implementation: Incomplete adherence to treatment has been identified as a
major obstacle to the control of the disease. Tuberculosis
(TB) is nearly always curable if patients are treated with
effective, uninterrupted antituberculous therapy. Adherence to treatment is critical for cure of individual
patients, controlling spread of infection and minimizing
the development of drug resistance. This abstract makes
an attempt to enumerate the barriers of treatment
adherence and modes to encourage adherence through
sensitizing patients on their rights and responsibilities.
Intervention or response: In the districts of Pakur,
Jharkhand, India, TB patients were made aware about
their Rights & Responsibilities through project Axshya,
a Global Fund supported initiative. Through the project,
trainings and Mid-media activities were organized for TB
Patients on their Rights & Responsibilities and 180 TB
Patients were made aware regarding the TB charter.
Patients were identified from villages where completion
rates were low and selection of patients was done
through the help of TB Forum members involving local
formal leaders.
Results and lessons learnt: Out of 180 TB Patients, 82%
informed about difficulties in receiving medicine during
the period of migration and they also informed that most
of DOTS provider not ready to give medicine in advance
before their departure to village. During their migration
patients have to depend on Private Practitioner of the
area and became irregular in taking medicine. Most of
trained TB Patients started demanding necessary refer-
rals before their migration following the sensitization

with the TB charter.
Conclusions and key recommendations: Migration is a
norm for better opportunities in Pakur district of India
and this leads to interruption of treatment. Establishment
of proper referral system in consultation with respective
DOTS providers is required. There is a need for advocacy
with policy makers for ensuring their regular treatment
during the period of migration.
PC-1045-05 Experiences of delivering long-term

injectables during TB treatment in Malawi
D Cohen,1,2 M Phiri,1 H Banda,2,3 S B Squire,2
N Desmond,1,2 I Namakhoma3 1Malawi Liverpool
Wellcome Clinical Research Programme, Blantyre, Malawi;
2
Liverpool School of Tropical Medicine, Liverpool, UK;
3
REACH Trust, Lilongwe, Malawi. e-mail:
danbcohen@yahoo.com
Background: Patients on treatment for recurrent or

multidrug-resistant tuberculosis require long courses of
injectable anti-tuberculous agents. There are various
options for the daily delivery of injections, but each has
major disadvantages. In order to investigate the feasibility of a novel model of care delivery, a clinical trial (The
TB-RROC Study) was conducted in patients receiving
Category II retreatment regimen at two central hospitals
in Malawi. Hospital-based care was compared to a
community-based approach in which guardians were
trained to deliver injections to patients at home. In order
to comprehensively evaluate this intervention, it is
important to understand the experiences of those
involved.
Design/Methods: A qualitative evaluation of the TBRROC intervention was conducted, using phenomenographic methods to describe understanding and experiences of TB retreatment. TB-RROC trial participants
were purposively sampled, and in-depth interviews were
conducted with patients and guardians in both arms of
the trial. Key informant interviews and observations in
the TB wards and in patients homes were also
conducted. Thematic content analysis was used to derive
analytical themes.
Results: 14 patients and 12 guardians were interviewed.
9 key informants including household members, potential trial recruits who declined to participate, nurses and
TB officers were also interviewed. Three key themes
relating to TB retreatment emerged: medical experiences
(including symptoms, treatment, and HIV); the effects of
the physical environment (particular distress caused by
conditions on the ward, and disruption to daily routines
and livelihoods); and trust (in other people, the
community as a whole and in the health system).
Experiences were affected by the nature of a persons
prior role in their community and resulted in a range of
emotional responses. Patients and guardians in the
community benefited from better environment, social
interactions and financial stability. Concerns were
expressed about the potential for patients health or
relationships to be adversely affected in the community.
These potential concerns were very rarely realised, and
overall participants described overwhelmingly positive

experiences of community-based care.
Conclusion: Patients and guardians have a range of
medical, social, financial and emotional experiences
during TB retreatment. Overall, community-based care
offered many advantages over hospital-based care.
PC-1046-05 Knowledge, attitudes and practices

about home-based TB treatment support by
community care workers in a South African
urban informal settlement
T Collings,1 A Von Delft,1,2 G Dhlomo,1
B Khatry-Chhetry,1 N Koen,1 K Wilson1 1University of Cape
Town, Cape Town, 2TB Proof, Cape Town, South Africa.
e-mail: vuzumsi@gmail.com
Problem: Tuberculosis (TB) rates in urban informal

settlements in South Africa are amongst the highest in
the world. Despite a renewed State focus on patient
centred care, the optional utilisation of home-based TB
treatment support (HTS) provided by Community Care
Workers (CCWs) remains low.
Methods: A cross-sectional study was used to assess the
knowledge, attitudes and health practices of community
members in Langa, Cape Town, towards TB management and CCWs, and explore associations with uptake
of HTS. An adapted KAP questionnaire was administered by medical students to community members with
current and previous TB (TB exposed) and no history of
prior disease. Participants were recruited from the Langa
Community Health Centre and during home visits.
Results: 31.4% of the total 172 participants expressed a
preference for HTS delivered by CCWs, but initial
knowledge about what CCWs were was low (49.4%).
Respondents interviewed at home (n 99) were more
likely to prefer HTS (PR1.75, P 0.02) compared to
clinic respondents. There were no significant associations
between demographic and socio-economic factors, HIV,
previous TB, or levels of knowledge about CCWs or TB,
and choice of HTS versus clinic-based support (CTS).
The majority of current TB patients preferred HTS
(52.6%, P 0.034). Main reasons for preferring CTS
were self-sufficiency and being able to travel (21.5%)
and the belief in better services (15.7%). HTS was
favoured if too old or frail (14.0%) or due to convenience
(11.1%). Barriers to HTS included perceived stigma
associated with CCWs (5.2%) and a lack of confidentiality (4.7%). The TB exposed group (n 53) had better
knowledge about TB (PR1.5; P 0.03) and CCWs
(PR1.4; P 0.02) and had a greater preference for
CCWs to wear civilian clothes (43.4% vs. 20.5%, P
0.001). 34.3% of all respondents felt people with TB
were treated differently by the community.
Conclusions: Despite the high burden of TB, preference
for HTS remains low. Increased uptake among current
TB patients is encouraging. Generally low knowledge
about CCWs suggests more resources should be focused
on promoting the benefits of and trust in patient centred
care delivered by this cadre of workers. High selfreported HIV co-infection rates (70.0%) support current
S375
efforts to integrate TB-HIV services. Subgroup analyses

point to the complex interplay between TB and HIV
exposure and perceptions of stigma, which requires
careful exploration to render community-based services
more acceptable.
PC-1047-05 Community involvement role in

MDR-TB treatment in Tajikistan
A Ismayilov,1 A Trusov,2 O Norov,1 P. Tabarov,1
M Sianozova,3 S Odinaeva1 1Project HOPE, Dushanbe,
Tajikistan; 2Project HOPE, Millwood, VA, USA; 3Project HOPE,
Yerevan, Armenia. Fax: (992) 227 3646. e-mail:
aismayilov@projecthope.org
Background and challenges to implementation: Tajikistan is among the 27 high burden multidrug-resistant
tuberculosis (MDR-TB) countries in the world. It has one
of the highest estimated TB cases (all forms) 142 (range
70-239) per 100 000 population incident number in the
WHO European Region. According to the latest drug
resistance survey in 2011, the prevalence of MDR is 13%
among new and 54% among previously treated TB cases
(WHO report, 2013). It has been a challenge to optimally
manage MDR-TB treatment in civilian population living
in remote and impenetrable districts.
Intervention or response: Medical and lab records were
retrospectively analyzed for interim treatment results of
two groups of MDR-TB patients enrolled since the third
quarter of 2013. One group of patients (99) has been
receiving the standard treatment for all Tajikistan
settings, while the second group (114) has been
supported by social workers and volunteers.
Results and lessons learnt: As the patients from cohort
have not yet completed MDR-TB treatment, analyses
were based on six months culture conversion and default
rates. Culture conversion at six months in the group with
community involvement was almost two times higher
than conversion of MDR-TB patients under standard
treatment. Groups / Cohort (3q 2013 3q 2014) Culture
conversion at six months Default (including transfer out)
Negative Positive Not known (including died) Standard
Treatment / 99 45 (45.5%) 14 (14.2%) 27 (27.3%) 13
(13.1 %) Community DOT support 88 (77.2%) 8 (7%)
10 (8.8%) 8 (7%) Moreover, under the community
surveillance group, there were no defaults during this
period and no cases without lab results.
involvement in DOT has greatly positively influenced
both culture conversion and default rates. Community
DOT is proving to be an essential tool in achieving
MDR-TB treatment success. It can be assured by
involving the Community, particularly social workers
and volunteers into TB control at household levels. NTP
in Tajikistan should seriously consider expanding this
example throughout the country.
S376
PC-1048-05 Adherence to TB treatment among

patients on supervised treatment approach in
Tororo District
PC-1049-05 Implementation of tuberculosis

directly observed treatment in Kampala,
Uganda
L Kwagonza,1 M Ediau,1 G Okure,1 D Tuhebwe,1

D Okumu,2 E Buregyeya1 1Makerere University, Kampala,
2
Tororo District, Tororo, Uganda. e-mail:
kwagonzaleo@gmail.com
H Sempeera,1 E Buregyeya,1 A Kawooya3,4 1Makerere

University School of Public Health, Kampala, 2International
Health Sciences University, Kampala, Uganda. e-mail:
hassardsempra@gmail.com
Background: Supervision of TB patients during the

treatment course is widely advocated for and forms part
of the World Health Organizations Directly Observed
Therapy, (DOT) strategy. The supervised treatment is
aimed to ensure adherence to drugs as prescribed as also
reduces chances of TB drug resistance The aim of this
study was to assess the level and factors associated with
adherence to TB treatment in patients enrolled for TB
supervised treatment in Tororo District.
Design/Methods: A cross sectional study was conducted
among 105 TB patients enrolled on supervised treatment
approach. Descriptive statistics and logistic regressions
were done to determine the level and factors associated
with adherence to TB treatment. Adherence was defined
as taking the medication at prescribed time
Results: The mean age of the respondents was 44 years
(SD 1.3) ranging from 18 to 80 years. Eighty three
percent (83.0%; 84/105) of the respondents were found
to be adherent to TB treatment in this study. Among the
adherent respondents: 75.0% (63/84) were female, and
76.6% (64/84) had attained at least primary level of
education. 81.8 % had chronic illnesses. In addition, out
of the respondents who adhered to TB treatment; 81.8%
(68/84) were HIV infected, many (93.2%) reported that
they had disclosed their TB status to their partners and
71.6% reported having been supported by their family
members, and 61.9% reported travelling a distance of
more than 5km to Health facilities. Married clients were
more likely to adhere to TB treatment AOR5.6, 95%CI
(1.68 18.6). Much as the study participants were
registered as supervised treatment clients, only 34.3%
(36/105) reported to have been supervised during
treatment. Those who had a treatment supporter as part
of supervision were more likely to adhere to TB
treatment as compared to those who were not supervised
during the treatment course OR 5.01, 95%CI (1.05124.009).
Conclusion: Adherence to TB treatment was moderate,
but still below 100% target adherence. Being married
and having a supervisor or treatment supporter were
associated with good adherence to TB treatment. In
order to improve on TB treatment outcomes, particular
attention should be put on clients who are single, due to
reduced opportunity of being reminded to take their
medication. TB control programmes should ensure that
TB clients receive appropriate supervision as suggested
by the Direct Observed Therapy short strategy.
Background: Due to increasing slum population and

HIV/AIDS, Kampala city (KC) is now responsible for
about 20% of the Ugandas tuberculosis (TB) burden. In
2012, KC reported an increase (29%) of patients on TB
directly observed treatment (DOT). Though DOT
enhances treatment adherence and completion, KC
registered a treatment completion of 45% and a default
rate of 17%. This study assessed the implementation of
TB DOT in Kampala.
Design/Methods: A cross-sectional study of quantitative
and qualitative data collection was carried out in 2014 in
Rubaga division (RD). Rubaga was chosen because in
2012 it reported the highest (70%) proportion of patients
on TB DOT. Data was collected from division KC health
facilities. 201 new TB patients on treatment for less than
16 weeks were recruited. Questionnaire and key informant interview guide were used to collect data from
patients and health workers respectively. Stata version 12
was used whereas qualitative data was analysed manually.
Results: 66% of patients reported being DOT. Majority
(82%) of the treatment supporters were relative/family
members. Lack of caretakers was the main reason for not
being on TB DOT (n 44, 77%). 26% of all patients
missed doses for at least one week. Lack of transport to
collect medication and stock outs were the main reason
for missing doses. In the adjusted regression analysis
being married and positive attitude towards TB DOT
were significantly associated with being on TB DOT,
adj.PRR1.4(1.1-1.6), adj.PRR2.4 (1.6-3.6). Lack of
independent community-based treatment supporters was
the main barrier to the implementation of TB DOT in
RD.
Conclusion: Over half of the patients were on DOT.
Treatment supporters are mainly family members. A
quarter of all patients did not adhere on treatment. The
division lacks independent community-based treatment
observers to support the poor and those without
caretakers.
PC-1050-05 Equitable delivery and outcomes of

tuberculosis treatment and care in the North
West of England: analysis of North West TB
Cohort Audit data
P Macpherson,1,2,3 B Squire,2 P Cleary,4 C Wake,2 K Dee,4
M Woodhead5,6 D Sloan,2 On Behalf of the North West
TB Cohort2 1University of Liverpool, Liverpool, 2Liverpool
School of Tropical Medicine, Liverpool, 3Cheshire and
Merseyside Public Health England, Liverpool, 4Public Health
England Field Epidemiology Service North West, Liverpool,
5
Central Manchester University Hospitals NHS Foundation
Trust, Manchester, 6University of Manchester, Manchester,
UK. e-mail: p.macpherson@liverpool.ac.uk
Background: Individuals with TB in the UK predominately come from the most socioeconomically deprived
groups. However, the extent to which deprivation affects
access to care and TB outcomes is unknown.
Design/Methods: Since 2011, the North West TB Cohort
Audit has undertaken quarterly review of treatment and
care outcomes against consensus-defined standards for
all individuals notified with TB in the north west of
England; the first region of the UK to undertake cohort
audit across such a large geographical footprint. Socioeconomic deprivation groups were constructed using
quintiles of individuals Index of Multiple Deprivation
(IMD) scores measured at lower super output area of
residence. Logistic regression and Cox proportional
hazard models investigated associations between socioeconomic deprivation group and adverse TB care
outcomes.
Results: 1831 individuals notified with TB between 2011
and 2014 were included in this analysis. 62% (1131/
1831) came from the most deprived national quintile,
and socioeconomic deprivation group was associated
with younger age group (P , 0.001), non-UK born status
(P , 0.001) and having any social risk factors (P
0.061). In single variable analysis, greater socioeconomic
deprivation was not significantly associated with adverse
TB care outcomes, including delay between symptom
onset and treatment initiation (hazard ratio for most
deprived vs. least deprived group: 0.91, 95%CI: 0.711.17); completion of standardized risk assessment to
identify complex needs (odds ratio [OR]: 0.66, 95%CI:
0.24-1.86); having at least 90% of close contacts
assessed (OR: 2.86, 95%CI: 0.79-10.31); all child
contacts being assessed (OR: 1.30, 95%CI: 0.29-5.76);
being offered an HIV test (OR: 0.70, 95%CI: 0.33-1.45);
or completing treatment within 1 year in fully-sensitive
cases (OR: 0.66, 95%CI: 0.43-1.01). On multivariable
analysis, there remained no significant associations
between socioeconomic deprivation and adverse TB care
outcomes.
Conclusion: Despite high levels of socioeconomic deprivation in this population, across a range of TB care
standards, TB patients in the most deprived group had
similar care to more affluent individuals, suggesting that
access to, and delivery of TB care in the North West of
England is equitable. The extent to which the cohort
audit process contributes to, and sustains this standard of
care deserves further study.
S377
42. Inside the belly of the beast: TB in

correctional settings
PC-1051-05 The impact of XpertW MTB/RIF and
WHO-reviewed case definition on the number
of TB relapse cases registered in the Azerbaijan
penitentiary system
E Gurbanova,1 K Blondal,2 R Mekhdiyev,1 R Tahirli,3
F Huseynov,1 N Kerimova,4 F Mirzayev,5 A Altraja6 1Main
Medical Department of the Ministry of Justice, Baku,
Azerbaijan; 2Department of Communicable Disease
Prevention and Control, Reykjavik Health Care Services,
Reykjavik, Iceland; 3Laboratory of the Specialized Treatment
Institution for Detainees with Tuberculosis, Baku, 4Project
Implementation Unit, Main Medical Department of the
Ministry of Justice, Baku, Azerbaijan; 5Laboratories,
Diagnostics & Drug Resistance Unit, World Health
Organization, Geneva, Switzerland; 6Department of
Pulmonary Medicine, University of Tartu, Tartu, Estonia.
e-mail: e.gurbanova@prisonhealth.az
Background: The Main Medical Department of Azerbaijan Ministry of Justice implements an exemplary TB
Control programme in the penitentiary system (PS)
providing up-to-date diagnostics and effective treatment
for inmates with both susceptible and drug-resistant TB
in the Special Treatment Institution (STI). The STI also
houses a modern quality-assured reference laboratory
that performs all WHO-recommended diagnostics of TB.
In 2013, Xpert MTB/RIF assay was introduced as a
routine diagnostic algorithm in STI along with adopting
the reporting system to the Reviewed WHO Case
Definitions Framework.
Objectives: To assess the impact of introduction of Xpert
MTB/RIF assay to the diagnostic algorithm of TB and the
impact of the use of the reviewed WHO definition of TB
cases on the number of registered relapse cases in
Azerbaijan PS. Design and methods: A cohort of all
new and previously treated cases with pulmonary and
extra-pulmonary TB that started treatment with the first
line (FLD) and second line drugs (SLD) in the PS of
Azerbaijan during 01.01.200713.02.2015 was included. The cohort was divided into 2 groups: Group 1: cases
enrolled during 01.01.2007-31.12.2012 and Group 2:
cases enrolled during 01.01.2013-13.02.2015 (i.e. before
and after the Introduction of Xpert MTB/RIF to the
diagnostic algorithm of TB and use of the new WHO
Case Definition Framework (Table). Logistic regression
analysis was used to determine the association of
registered relapse cases with the introduction of Xpert
MTB/RIF and use of the new definition.
Results: In the Group 2, 31% (57) out of all 180 relapse
cases were due to diagnosis with Xpert MTB/RIF only (n
27) and new WHO case definition introduction (n
30). Among patients diagnosed with Xpert MTB/RIF
only, the average duration between the end of most
recent treatment course and the recurrent episode was 31
months. The number of relapse cases significantly
increased in the overall cohort: OR 1.5 (95%CI 1.21.8, P , 0.001) and in patients on treatment with FLD:
OR 1.5 (95%CI 1.2-1.9, P , 0.001). There was no
S378
significant increase of relapse cases among the patients

treated with SLD.
Conclusions: Introduction of Xpert MTB/RIF to the
diagnostic algorithm of TB and use of the new WHO
Case Definition Framework significantly increase the
number of TB cases registered as relapse in the
Azerbaijan PS.
smear positive, 12% new smear negative, 14% new extra

pulmonary and 7% were retreatment cases. Screening
procedure should be in placed at the time of entry of
inmate into the prison. Difficulty in securing follow up of
TB cases who are released/ transferred before completion
of their treatment.
Conclusions and key recommendations: Early detection
of TB among prison inmates is crucial to control TB
transmission and to have favorable treatment outcome.
High transfer out and unknown treatment outcome
causes low success rate. So, strong external referral linkage is essential to strengthen activity in special situation.
PC-1053-05 Effectiveness of an entry screening

programme in the penitentiary sector in
Georgia
N Lomtadze,1 Z Avaliani,1 M Madzgharashvili,1
U Nanava1 1National Center for Tuberculosis and Lung
Diseases, Tbilisi, Georgia. Fax: (995) 322 911 941. e-mail:
nlomtadze@gmail.com
PC-1052-05 Tuberculosis control services in

prisons: patient care and referral in Bangladesh
K Uddin,1 M Siddiqui,1 D-U Mesbah,1 S Islam,1
K Khatun,1 A Islam,1 F Khanam2 1BRAC, Dhaka, 2National
Tuberculosis Programme, Dhaka, Bangladesh. Fax: (88)
02988 8026. e-mail: shayla.dhaka@gmail.com
Background and challenges to implementation: Prison

inmates are one of the high risk groups to develop
infectious diseases like tuberculosis. Infected prison
inmates spread the disease to others very easily and
rapidly. BRAC, besides its community based TB services,
started to work with prison inmates in 2004 covering 7
prisons and gradually expands the program in 41 prisons
by 2011under the stewardship of National TB Control
Program (NTP). Entry examination, symptomatic screening and diagnosis of all forms of TB cases is a challenges
to combat the deadly disease in this difficult setting. The
objective of the study is to identify the tuberculosis status
among prisoners in BRAC supported areas.
Intervention or response: To identify TB presumptive in
prison, BRAC staff goes to prison everyday where
laboratory facilities are available and two or three times
in a month where laboratory facilities are not available.
Prison authority also informs BRAC staff when new
inmate enters the prison with sign symptom of TB.
Diagnosed TB case is registered and managed following
NTP guideline. Treatment initiated by medical officer
and DOT is ensured by prison pharmacist. BRAC staff
follows up patients weekly or fortnightly. Patients are
transferred to other DOTS center with necessary
documents, if needed, to continue the treatment.
Results and lessons learnt: In 2013, a total 148 cases
were diagnosed in 41 prisons. Among them, 68% were
new smear positive, 16 % new smear negative , 11% new
extra pulmonary and 5% retreatment cases. Treatment
success rate was 87% which is a bit lower than other
population. Transfer rate from prison to prison and to
outside the prison was as high as 12%. In 2014, a total
180 patients were diagnosed. Among them, 67% new
Background: TB in prisons has been a public health

problem in Georgia, which is one among high X/MDRTB burden countries globally. During 2010-2012 the TB
notification rates per 100 000 inmates in Georgian
prisons was 30 times higher compared to civilian sector,
overcrowding being the major contributor. Since 2011,
the Entry Screening of every prisoner is performed using
a standardized questionnaire, followed by sputum smear
microscopy for those considered to have presumptive TB.
An operational research aiming at assessing the validity
and overall yield of entry screening within penitentiary
sector has been conducted.
Design/Methods: A retrospective analysis of completed
entry screening questionnaires merged with data for
each screened inmate regarding TB diagnosis was
performed. The validity analysis of the screening
questionnaire by assessing sensitivity, specificity, PPV
and NPV for assessment of presumptive TB case was
performed. The three year and annual yields of TB cases
were assessed as proportion of cases identified due to
screening among all TB cases detected within the
corresponding period of time in the Prison sector.
Results: Overall out of 32 152 performed entry
screenings, 1559 (5%) were assessed as presumptive
TB case and remaining 30 593 (95%) as non-presumptive TB case. Out of presumptive TB cases 102 (6.5%)
were diagnosed to have TB, while in 1457 (93.5%) TB
was not confirmed, neither bacteriologically or clinically.
Out of screened non-presumptive cases TB was never
detected among 30 192 (98.7%) and TB was detected in
401 (1.3%) cases by intervention other than screening,
such as self reporting with symptoms. The overall odds of
having TB diagnosis for those who screened positive on
screening questionnaire is 5.2 times greater compared to
odds of having TB among those who screened negative
using questionnaire with strong statistical significance:
OR5.2; 95% Confidence Interval (CI)4.19-6.57; Chi2
(Mantel-Haenszel) v2 value ,0.0000001. The sensitivity
and specificity of the screening questionnaire were 20%
S379
and 95%, while PPV and NPV were 6.5% and 99% respectively. The yield of entry screening in detecting TB
cases in prison in during the period of 2011-2013 was
5%.
Conclusion: Entry Screening with questionnaire that
scores the patients based on symptoms, BMI and past TB
history, proves to have high specificity but, low
sensitivity, thus making the tool very effective method
to rule out TB. Also overall yield of TB cases due to entry
screening was 5%, which is in line with screening
performance data internationally and in the range to
prove the cost effectiveness.
PC-1054-05 RNTCP in prisons: special ACSM

inputs in jails saved lives of vulnerable inmates
in Chhattisgarh State, India
G Mallick,1 M R Aggarwal,2 S Chadha1 1International
Union Against Tuberculosis and Lung Disease South-East Asia
Office, New Delhi, 2Directorate of Health Services
Chhattisgarh, Raipur, India. e-mail: gmallick@theunion.org
Background and challenges to implementation: More

than 9.8 million people are detained in penitentiary
services around the world. TB in prisons is a major public
health problem in many settings, particularly in India
with high incidence. Concomitant conditions, particularly HIV infection, injecting drug use, poor nutritional
status, overcrowding and inadequate or inaccessible
medical care put inmates at risk of infection and rapid
progression from latent TB infection to TB disease. The
situation is worsened by the emergence and spread of
MDR- and XDR-TB Many hard-core criminals including
Left-wing ultra prisoners have posed perennial problems
where 16 199/6482 inmates (three times of the capacity)
are housed in 28/30 different high security jails in
Chhattisgarh which is the most vulnerable Naxalinfested state in India
Intervention or response:The five incarnations that
factors TB in the prisons Prisons receive TB, Prisons
concentrate TB, Prisons disseminate TB, Prisons make
TB worse and Prisons export TB were prioritised in all
28 jails where intensified ACSM activities were conducted. Educational sessions like talks, videos, flipcharts,
various educational materials, contests, question-andanswer games were done during the year. Few inmates
were engaged as peer educators and treatment supporters
who identified presumptive TB suspects. Regular sputum
collection/test camps, close monitoring of diagnosed
cases and their treatment has added credibility to the
program outcome. Two groups of medicines first-line
anti-TB drugs and second-line parenteral agent (injectable anti-TB drugs) kanamycin, amikacin, capreomycin
were used for TB and MDR cases respectively
Results and lessons learnt: 92 extensive ACSM events for
16 199 prisoners and 735 prison staff were conducted in
28 jails of the state during 2014. 124 diagnosed TB
patients (96 positives, 28 negatives) and 3 MDRs were
put on treatment from sputum examination of 1348
presumptive TB suspects that added about 1% to the
states 154 868 sputum examined during the year
Table Revised National TB Control Program outcomes in

different jails in Chhattisgarh state for 2014
Sl.No.
1
Name of Jail
No. of
TB
Under
sensiMDR*
lock- With tiza- Spu- paCaup
DMC tion tum tients
pac- (as on facil- mtgs. exam on
ity Jan15) ity
held done Rx
Central Jail,
1190
Raipur
2
District Jail,
90
Dhamtari
3
District Jail,
70
Mahasamund
4
Sub Jail, Baloda
70
Bazar
5
Sub Jail,
50
Gariyaband
6
Central Jail,
639
Jagdalpur
7
District Jail,
65
Kanker
8
District Jail,
150
Dantewada
9
Sub Jail, Sukma
50
10
Sub Jail,
50
Narayanpur
11
Central Jail,
1028
Bilaspur
12
District Jail,
225
Raigarh
13
District Jail,
110
Korba
14
District Jail,
70
Janjgir
15
Sub Jail, Pendra
50
Road
16
Sub Jail,
50
Katghora
17
Sub Jail,
50
Sarangarh
18
Sub Jail, Sakti
100
19
Central Jail,
975
Ambikapur
20
District Jail,
148
Baikuntpur
21
District Jail,
230
Jashpur
22
Sub Jail,
130
Ramanujganj
23
Sub Jail, Surajpur 90
24
Sub Jail,
50
Manendragarh
25
Central Jail,
396
Durg
26
District Jail,
116
Rajnandgaon
27
Sub Jail, Sanjari
70
Balod
28
Sub Jail,
70
Dongargarh
29
Sub Jail,
50
Bemetara
30
Sub Jail,
50
Kawardha
Total
6482
2804 DMC
106 121*
144 ASDTC
42
246 ASDTC
37
240 ASDTC
20
80 ASDTC
18
82 251*
361 ASDTC
32
547 ASDTC
87
1591 DMC
0
0
Non-functional
Non-functional
2925 DMC
96 141*
575 ASDTC
77
265 ASDTC
52
179 ASDTC
49
92 ASDTC
15
172 ASDTC
41
55 ASDTC
32
174 ASDTC
1933 ASDTC
2
4
51
59
2
7
152 ASDTC
48
400 ASDTC
62
323 ASDTC
45
222 ASDTC
134 ASDTC
3
2
38
35
2
1
1763 ASDTC
57
299 ASDTC
68
129 ASDTC
22
85 ASDTC
21
176 ASDTC
34
133 ASDTC
22
92
1348
16199
124
(96/28)
3
Source: State Prison Office, Raipur; Central Jail 5, District Jail 10, Sub Jail
15, Non-functional Jail 2
*3 MDR patients were diagnosed and put on treatment one each from
Raipur, Jagdalpur and Bilaspur Central Jails
TB tuberculosis; MDR multidrug-resistant; ASDTC Attached to
District/sub-District Tuberculosis Unit.
S380
Conclusions and key recommendations: Special ACSM

efforts addressed early diagnosis and adherence to
medication for 127 TB patients. Measures to reduce
overcrowding and improved living conditions, integrating prison and national TB control to ensure treatment
completion in the incarcerated population need to be
viewed critically. A complementary political commitment must be built to establish and sustain effective TB
control strategies in prisons
PC-1055-05 XpertW MTB/RIF improves yield of

bacteriologically confirmed TB in correctional
services in Mozambique
A Mondlane,1 I Manhica,1 I Chirrime,1 C Mutaquiha,1
K Alberto,1 S Hassamo,1 L Nhocuana,1 H Hamene1
1
Ministry of Health, Maputo, Mozambique. e-mail:
angelanvm1@gmail.com
Background: Mozambique prevails as one of 22 high

burden countries for TB, with 37% case detection rate.
The National TB control program (NPTC) faces
challenges in the timely identification of TB, due to
HIV prevalence (11.5% in 2009) and access to health
services. TB-HIV co-infection is high, at 52% in 2014.
The challenges are magnified for correctional services
inmates who often face overcrowding, understaffing
(health), high turnover, and therefore, passive case
finding and inadequate follow-up of cases. A study
conducted in correctional services in 2013 using smear
microcopy, revealed 1.5% smear positive TB point
prevalence and 24% HIV prevalence. In 2014, 2% of
the 13 000 correctional inmates in the system were
smear-positive TB by routine symptom screening. Current screening programs in correctional facilities involves
a quarterly general TB awareness talk with verbal
invitations for symptomatic inmates to provide sputum
that is tested by microscopy.
Design/Methods: To estimate the proportion of undetected bacteriologically confirmed TB cases by current
screening strategy in correctional services; To evaluate
the yield of GeneXpert over routine practice (smear
microscopy); smptoms based screening was conducted in
5 correctional facilities located in South, Center and
North region of the country, namely, Maputo, Sofala,
Tete, Zambezia and Nampula, from 23 March 2015 to
10 April 2015. Every correctional inmate was screened
for cough, fever, weight loss, night sweats of at least oneweek duration. Onsite sputum collection was undertaken
after instructions. HIV testing was offered for inmates
with presumptive TB. Patients on TB treatment were not
screened. Samples were analyzed using Xpert/MTB/RIF.
Yields from this intervention were compared to yields
from routine passive first quarter 2015 data (21
December to 20 March), to determine the proportion
of bacteriologically confirmed cases missed by current
approaches. Smear microscopy may under-diagnose and
over-diagnose TB when used in active case finding in
Southern Africa.
Results: From 5279 inmates in correctional services,
4969 (96.9%), were screened for TB. From the consent-
ing participants, 18.6% (927) were symptomatic. From

927 symptomatic inmates, were Xpert positive (2.7%).
315 were HIV Positive. Routine TB screening and passive
case detection with smear had detected 13 TB cases
(36%) in first quarte.
Conclusion: Active case finding and xpert MTB/RIF
increases case detection by 63%.
PC-1056-05 High yield of mass screening for TB

in police detention cells in Namibia
N Ruswa,1,2 A Evard,3 K Amufufu,3 F Mavhunga,1
S S Nghoshi,3 J Ndahapele3 1Ministry of Health and Social
Services, Windhoek, 2KNCV Tuberculosis Foundation,
Windhoek, 3Ministry of Health and Social Services, Opuwo,
Namibia. Fax: (264 )61 252 740. e-mail:
ncruswa@yahoo.com
Background and challenges to implementation: A third

of the nine million global cases of TB in 2013 were
missed by health systems. In 2013, Namibia reported
about 10 610 cases of TB, and it is estimated that
Namibia missed 34% of its cases. The burden of TB in
prisons in Namibia is unknown, and there are no
dedicated health services for inmates in police detention
cells. People living in the congregated setting such as
prisons and police detention cells are at an increased risk
of being infected with TB because these settings have
poor ventilation, are often over-crowded, and have no
system for screening inmates. Inmates in police cells are
incarcerated for between a few days to a few years while
awaiting trial. We present the results of a mass screening
exercise for TB performed in a small town (Opuwo) in
Namibia in March 2015.
Intervention or response: We used an adapted symptomatic screening tool from the Namibian TB guidelines,
focusing on five questions (cough, weight loss, night
sweats, fever, swollen lymph nodes) to interview inmates
in the police cells in Opuwo. Two sputum specimens
(spot-morning) were collected from participants who
reported any one of the 5 symptoms. The specimens were
sent for testing with smear microscopy and molecular
testing with Xpert MTB/RIF at the district laboratory.
Results and lessons learnt: All 50 (100%) inmates, and
43 (36%) police officers were screened. 39 (78%) of
inmates reported symptoms of TB and all of them had
sputum tested. 3 (8%) were found to have active TB by
smear and/or Xpert. 17 (34%) inmates knew their HIV
status, of which 2 (12%) were HIV positive. None of the
inmates or police officers were on TB treatment at the
time of screening, but 5 (10%) of inmates and (1) 2% of
officers had been treated for TB before. No TB case was
identified among the police officers.
Conclusions and key recommendations: The prevalence
of bacteriologically positive active TB among the inmates
in Opuwo police holding cell is 8%. This is very high,
considering none of these inmates were on treatment or
had presented to health facilities. We recommend
establishment mass screening of inmates for TB in police
holding cells at regular intervals in Namibia.
PC-1057-05 TB-HIV integration in Gauteng

correctional facilities
E Zishiri,1 G Lekubu,2 A Makhubela,2 L Page-Shipp,1
S Charalambous,1,3 G Gresak,1 V Chihota1,3 1The Aurum
Institute, Johannesburg, 2Department of Correctional
Services, Johannesburg, 3University of the Witwatersrand,
Johannesburg, South Africa. e-mail:
ezishiri@auruminstitute.org
Background: South African national policy recommends

integrating TB and HIV service delivery to improve
health outcomes for co-infected patients. The Department of Correctional Services has taken up this
recommendation and assimilated the TB-HIV integration
model in their health centres. We describe health
outcomes among inmates started on TB treatment in
four correctional centres in Gauteng.
Methods: We conducted a retrospective record review of
routine TB programme data for inmates started on TB
treatment between April 2013-June 2014 in Modderbee,
Leeuwkop, Krugersdorp and Boksburg correctional
centres, Gauteng, South Africa. We collected data on
demographic characteristics, details of TB diagnosis
(method of TB detection, episode type, site of disease),
TB treatment start date, treatment outcome and HIV
care and treatment (date of HIV test, CD4 count testing,
ART and Cotrimoxazole start dates).
Results: 186 inmates were started on TB treatment across
the four centres (median age 33 years, all male, 5.9% retreatment cases). HIV Counselling and Testing (HCT)
was offered to 182 (97.8%) inmates, there was no
documentation for 4 cases. 48 (25.8%) were known HIV
positive at time of diagnosis and 134 were offered HIV
testing. 129/134 (96.3%) agreed to undergo HIV testing
resulting in 81/129 (62.8%) testing HIV positive. Overall
129 (69%) TB cases were co-infected with HIV. 104/129
(80.6%) co-infected inmates had a CD4 count documented (median CD4 count 119 cells/ll [IQR 41-303
cells/ll]). Among 129 TB cases co-infected with HIV, 36
(27.9%) were initiated on ART prior to starting TB
treatment, 51 (39.5%) after starting TB treatment and 32
(24.8%) were documented as not having been initiated
on ART and 10 (7.8 %) had no record of ART initiation.
Among the 51 TB cases initiated on ART after starting
TB treatment, 30 (58.8%) were initiated within two
months of starting TB treatment. Among the 32 cases
documented as not having initiated ART, 17 (53.1%) had
either been released or transferred out.
Conclusion: There is evidence of TB and HIV integration
in correctional centres. Of the inmates started on TB
treatment a large proportion were offered HCT and there
was good uptake of HIV testing. The levels of knowledge
of HIV status and the proportion starting ART are
similar to the general population. The importance of TBHIV integration is reinforced by the high co-infection
rate. ART initiation remains a challenge especially where
inmates are either released or transferred out.
S381
PC-1058-05 Systematic review of the

epidemiology of and programmatic response
to TB in prisons and inmates in South Africa
F K Mukinda,1 H Mahomed1,2 1Stellenbosch University,
Cape Town, 2Western Cape Government, Cape Town, South
Africa. e-mail: liesl.nicol@gmail.com
Background: South Africas total prison population of

156 370 offenders is the largest in Africa and ranks ninth
in the world. The South African prison population rate,
estimated from a national population of 53.1 million in
August 2013, was 294 per 100 000. The occupancy level
of correctional facilities based on the official capacity
was 131.7%. South Africas prisons are overcrowded
and represent a high-risk environment for tuberculosis
(TB) transmission. Overcrowding, insufficient or lack of
proper ventilation, poor nutrition, limited access to clean
water, lack of TB preventive measures and failure to
supervise and ensure adherence to treatment are major
contributing factors to the occurrence of TB in prisons,
especially in Sub-Saharan Africa, where high rates of TB
in prison populations are reported. With the aim of
relevant research on the epidemiology of and programmatic response to TB in South African inmates and
prisons.
Design/Methods: A comprehensive search for both
published and unpublished research was conducted in
April 2015. Based on specific inclusion criteria, research
investigating TB incidence, prevalence, risk factors,
prevention, treatment and care in South African inmates
and prisons was included in this systematic review.
Results: Three studies met the inclusion criteria of this
review. Two studies reported on TB prevalence and one
study reported on HIV-related health outcomes of
inmates accessing antiretroviral therapy (ART). There
are no studies investigating the programmatic response
to TB in prison inmates. There are no studies investigating the occurrence of TB in prison employees (one study
included prison employees but did not provide outcome
data for the employees).
Conclusion: Inadequate TB management in prison will
not only reduce the rate of cure, but will also be a risk for
the emergence of drug-resistant strains that may affect
other prisoners, and be spread through prison staff,
visitors and via former inmates to the community. In this
way, TB in prison becomes a public health threat. It is
important that we determine accurately what the rates of
TB are in prison, what interventions are effective in
reducing the risk of TB transmission in prison settings as
well as effective programmatic response to TB in South
African prisons.
S382
PC-1059-05 Tuberculosis control among prison

population: what is the best approach?
H Kanyerere,1 L C Chisuwo,1 A Dimba,1 R Banda,1
J Mpunga,1 L Kanyenda,2 H Banda,3 K Kamoto4 1Ministry
of Health, National Tuberculosis Programme, Lilongwe, 2
United States Agency for International Development TB
CARE II Project, Lilongwe, 3Research for Equity and
Community Health Trust, Lilongwe, 4Malawi Prison Health
Services, Lilongwe, Malawi. Fax: (1) 301 941 8427. e-mail:
asmitharthur@urc-chs.com
Background: Infectious diseases like TB pose serious

threats in Malawian prisons and the world at large; and
present significant challenges for prison and public
health authorities. Since 1995 Malawi National TB
Programme (NTP) has been collaborating with Prison
Health Services in order to improve TB control in
Malawian Prisons.
Intervention: The Malawi NTP developed a policy for TB
control in prisons in 2007 and had put in place TB
surveillance mechanism in all prisons. However, TB case
detection in prisons remains low. As part of World TB
Day (2015) commemoration activities, the NTP decided
to commemorate the day by conducting mass TB
screening at Mzimba Prison (medium size prison). The
screening process included a standardised questionnaire,
chest X-rays for everybody and sputum submission for
those with TB related symptoms and/or abnormal chest
X-rays. TB screening was conducted within the prison
premises from 21st to 24rd March 2015. Those eligible
for sputum examination were requested to submit spot
and morning sputum specimens. All sputum specimens
were examined using fluorescent technique on iLED
microscopy and GeneXpert machines. All M. Tuberculosis positives on GeneXpert, Rif resistant and indeterminate results were repeated in order to confirm the
results. All microscopy positive were reread by the
second and third technicians blindly to confirm the
results.
Results: During the 4 days of screening, there were 611
prisoners and 85 members of staff. 608 prisoners
(99.5%; 604 males and 4 females) and 29 members of
staff (34%; 20 males and 9 females) were screened for
TB. 192 prisoners were eligible for sputum examination
and 172 (89%) submitted sputum. Of these172 who
submitted sputum, 22 were HIV positive, 114 were HIV
negative and 36 had unknown HIV status. Microscopy
was done on all the 172 prisoners who submitted
sputum. 161 had both microscopy and GeneXpert done.
71 prisoners (41%) were bacteriologically diagnosed
with TB (7 were sputum smear microscopy positive only;
59 were positive on GeneXpert alone and 5 were positive
to both microscopy and GeneXpert). Out of those that
were GeneXpert positive, 5 prisoners had rifampicin
resistance.
Conclusions and recommendations: The TB notification
rate at Mzimba prison is very high at (11 620/100 000
population). Mass TB screening on regular basis could
arrest the spread of tuberculosis in Malawi prisons. More
efforts need to be made to do a similar intervention in all
the prisons in Malawi.
43. The urban-rural divide: different

places, similar issues
PC-1060-05 Tuberculosis control in the state of
Sao Paulo, Brazil: big cities and urban risk
groups
V M Galesi,1 V Souza Pinto,1 S Fukasava,1 N K Komatsu,1
N Matsuda De Lima1 1Tuberculosis Division, Sao Paulo State
Health Secretary, Sao Paulo, SP, Brazil. Fax: (55) 11 3082
2772. e-mail: veragalesi@uol.com.br
Background and challenges to implementation: Sao

Paulo State is situated in the Southeast of Brazil with
an estimated population of 44 million inhabitants in
2014 distributed in 645 cities. It was prioritized for this
evaluation a big city with more than 1.2 million
inhabitants and one megalopolis with more than 10
million inhabitants. In 2014 the state had about 19 000
TB new cases with an incidence rate (IR) of 37.4/100 000
inhabitants.
Intervention or response: Using data of TB Information
System of Sao Paulo State (TBWEB) from Sao Paulo
State, Sao Paulo municipality (SPO) and Guarulhos
municipality (GRU) from 2006 to 2014 with the purpose
to describe the disease occurrence and activities developed to TB control in the following risk groups: people
living with HIV/Aids (PLWHA), inmates, homeless and
Latin American immigrants.
Results and lessons learnt: In 2014 SPO and GRU
discovered 5508 and 435 new TB cases with an IR of
46.3 and 33.2/100 000 inhabitants, respectively. Related
to PLWHA were 1035 TB cases in SPO and 40 TB cases
in GRU with 85.0% and 90.0%, respectively of HIV
testing. Related to inmates and homeless were 395 and

645 in SPO and 96 and 8 in GRU, respectively in 2013.
On Figure 1 is shown the percentage of TB risk groups in
the total of TB cases in each municipality. Concerning
activities developed, in GRU was implanted the PAHO
Project TB in Great Cities in the Health Districts of
Pimentas and Jurema (2013) with the following activities: geoprocessing with mapping of community resources; DOTS training; Cultural & Health Fair for the
Hispanic People aiming integration and matricial implantation of all TB and HIV cases referring to
psychosocial care; awarding of health care facilities;
implantation of rapid molecular test with 5379 done
detecting 259 TB cases and 9 Rifampicin resistance.SPO
works with the TB contingency actions for the cure
project (PACTU) aiming the development of unique
therapeutic project for the homeless with TB establishing
intra and inter sector articulation intending social
inclusion and develops activities according WHO recommendation in HIV polices. SPO done 14 143
molecular tests detecting 915 TB cases and 22 Rifampicin resistances.
Conclusions and key recommendations: TB control in
great urban centers requires an ongoing evaluation
process which includes analysis of active case finding
and treatment cohort according risk groups, allowing to
review strategies triggered from an initial situation
analysis.
S383
estimated incidence of TB in all forms is 131 per 100

000 population and the cohorts in 2008 were 70%
successful treatment, 11% of default, 3% failures, 4% of
deaths and 12% not evaluated. There were not infections
Control in Health Care Setting.
Intervention or response: Improve scope of bacteriological diagnosis and TB treatment under DOTS in the
community, a network of Community Health Workers
(CHW) is formed, on diagnosis, treatment and monitoring, initially with 25 CHW, who were monitored by 7
nurses who were moving between urban and rural, where
the Program control TB had identified TB cases in
previous years. CHW went where TB cases came
Spuctum Smear Frotis Positive SSF() with difficult
geographical access. CHW identified respiratory symptomatic, taking sputum samples were extended and
smear reading, report it to the health authorities and
identify a community leader. We set up support networks
for drugs supply with the help of the community, who
also took part in the cases registration and supervision,
which were also linked to the bacilloscopy carried out by
CHW. Bacilloscopy was after confirmed by the network
laboratories and the Public Health Laboratory. Patients
residing in poverty belts surrounding urban areas also
benefited from this project. The Infection Control Plan
was implemented in healthcare institutions and was
successful in 34 health centers in 34 out of the 46
municipalities. In 2013 recommendations for administrative, environmental and personal control measures
were incorporated.
Results and lessons learnt: 53% of the cases were male,
34% from indigenous peoples, 54% African descendants, 9% had some degree of disability, 49% were
householder who brings to the family income, 36% had
steady work, the average income is 200 USD. 27%
received state subsidies, 15% reported that he had forced
displacement. A 94% success rate was reported in new
sputum smear positive () cases within the DOTS
strategy in the community in 2012.
Conclusions and key recommendations: The communitybased DOTS contributed to increase success in treatment
by 11%. In 46 municipalities the success rate reached
81%. In Choco and Narino
success rates of 88% and
94% respectively were reported
PC-1061-05 Community interventions and

patient-centered activities on the pacific coast
of Colombia
PC-1062-05 TB literate villages: is this the way

to garner politico-administrative support and
empower marginal and vulnerable
communities?
B Gutierrez,1 D Rodriguez,1 N Gil,1 E Varela,2 A Palacios,3

1 1Organizacion Internacional para las
M Rondon
Migraciones, Bogota, 2Instituto Departamental de Salud de
Quibdo,
Pasto, 3Secretaria de Salud del Choco,
Narino,
Colombia. e-mail: davidalejorod@yahoo.es
O Bera,1 M Chandge,2 S Kamble,3 S Nayak,1 S Chadha,1

B Pawar3 1International Union Against Tuberculosis and
Lung Disease South-East Asia Office, New Delhi, 2District TB
Office, Mumbai, 3State TB Office, Pune, India. e-mail:

highest incidences are in dispersed rural populations,
inaccessible, with high levels of Unmet Basic Needs. The
46 priority municipalities with a majority are Africandescent and indigenous communities. In 2009, the
Background: With low TB knowledge in communities

and sub-optimal political and administrative commitment followed by lack of proper integration with general
health system case detection, adherence and treatment
outcomes remains low. A multipronged advocacy,
S384
communication and social mobilization approach is

desirable but difficulties associated with planning and
field level implementation remains a major hindrance to
achieve desired awareness level and community ownership. The objective of the intervention study was to assess
commitment and awareness level of tuberculosis and
impact of ACSM intervention package amongst local
leaders and community members of 25 villages in Latur
district, Maharashtra.
Intervention: Latur district has a population of 2.6
million with more than 700 villages having Gram
Panchayats constituted by elected people representatives
and public servant. At the onset in Jan 2014, 5 low
performing designated microscopy centres (DMC) were
identified within a low performing tuberculosis unit Ausa
with regards to case detection and treatment adherence.
Mapping was done and 10 villages were finalised which
were located more than 10 kilometres from DMCs. It
was followed by interventions which lasted for around 57 days within a village which were conducted by trained
community volunteers and communication materials
from Project Axshya. At first, political representatives
were sensitized on symptoms, diagnosis and treatment of
TB followed by two to three community meetings at
common areas within the village and at least one school
children sensitization meeting. Also intensive house to
house communication campaign was conducted followed by presumptive TB referral and symptomatic
sputum collection and transport. Two oil paintings were
painted at prime locations within the village and 10
running lines were painted throughout the village.
Baseline KAP evaluations and end line evaluations was
conducted after an observatory period of 6 months
among new random batches of 500 villagers and 10
political representatives with semi structured, pre tested
questionnaire in local language .
Results: A total of 8000 village members and 50 people
representatives were sensitized on TB symptoms, diagnosis and treatment during the entire campaign within 25
villages. Of the 500 villagers interviewed, in the base line
survey only 35.6% had heard of TB as compared to
88.4% at end of 6 months. Significant (P , .05) increase
in knowledge was observed at end of 6 months in
variables such as cough of two week as major symptom
(90%), sputum testing as method of diagnosis (69.9%),
spread of TB (55.9) high chance of acquiring TB in HIV
patients (44.5%), contact examination (63%) free
treatment at all government health facilities (92.1%),
DOTS as preferred treatment (90%), TB is curable
(95.2%). A positive change of attitude regarding
decreased stigma associated with TB (20.1%), increased
sharing of TB Status (23%) and increased acceptance of
TB treatment at Government health facility and increased practice of covering face while coughing or
sneezing (25.3%). Amongst 10 political leaders, there
was significant increase in knowledge with regards to
cough of two weeks as main symptom (72%), sputum
testing (71%), free DOTS provision (79%), free treatment (83%) and TB is curable (92%). Also there was
commitment of assured DOT provider within each
village from the local leaders.
Conclusions: Political and administrative commitment

remains one of the most important pillars for TB control.
The study has brought into light the importance of
decentralizing TB education and screening activities at
village level. With the vision of reaching the unreached,
TB literate villages has provided a common platform for
roping in politicians, government officers and local
people for the cause of TB control.
Table Comparative analysis of baseline, end line KAP survey
among villagers.
Variables
Baseline
KAP Study
(Total500)
n (%)
End line
KAP Study
(Total500)
n (%)
178 (35.6)
442 (88.4)
54 (30.3)
401 (90.7)
30 (17)
247 (55.9)
32 (30.4)
309 (69.9)
9 (5.1)
197 (44.5))
Knowledge
Heard of TB
*Cough of 2 weeks as symptom
of TB
*TB spreads when a person
having TB cough or sneezes
*Sputum testing as method of
diagnosis
*HIV increases the chances of
having TB
* Knowledge of TB contact
examination
*DOTS as TB treatment
*TB treatment better in private
health facility
*TB Treatment free in nearest
government health facility.
124 (69.7)
227 (51.4)
101 (56.7)
407 (92.1)
Attitude
* TB is curable
*TB creates stigma
*TB status sharing
114 (64)
102 (57.3)
39 (21.9)
421 (95.2)
89 (20.1)
102 (23.1)
27 (15.2)
112 (25.3)
Practice
*Cover your face with cloth
while coughing
43 (24.2)
11 (6.2)
278 (63)
398 (90)
*These questions are only for those who have heard of TB
PC-1063-05 Training TB patients: a strategy for

a better treatment adherence
G Martelli,1 M Tadolini,1 E Vanino,1 P Viale,1
G Lombardi,2 S Ianniruberto,1 P Dalmonte,2
A Gramegna1 1Infectious Diseases Unit, S. Orsola-Malpighi
University Hospital, Bologna, 2Microbiology Unit, S.
Orsola-Malpighi University Hospital, Bologna, Italy. Fax: (39)
0516 363 802. e-mail: giuliamartelli84@gmail.com
Tuberculosis (TB) remains a major public health concern

predominantly affecting low and middle-income countries. However in high-income countries TB epidemiology is characterized by a low rate of transmission in
native population and higher incidence in certain
vulnerable groups, especially migrants. Therefore current
public health strategies must address these targeted
groups. The metropolitan area of Bologna, a city in
northern Italy, is a low TB incidence setting, but has a
high rate of immigration. The Infectious Diseases clinic
of St.Orsola-Malpighi University Hospital notifies
around 90 new TB patients per year and nearly 75% of
them are foreign born. The majority is initially admitted
and started on treatment and then followed-up at outpatient level. At the first post-discharge visit patients are
interviewed about compliance to treatment: several

mistakes had been observed (i.e. wrong type or number
of pills taken, wrong timing) especially among migrants,
mostly due to language barriers or illiteracy. It must be
noted that direct observed therapy (D.O.T.) is usually not
implemented in western countries, except in very limited
cases and treatment is self-administer at home. In
addition, in Italy 4 drugs fixed-dose combination only
is available posing challenges in all those cases where
standard treatment cannot be utilized (due to side effects
or resistance). From the beginning of 2014 we introduced
an educational project involving both nurses and
doctors aimed to improve the adherence to anti-TB
treatment (ATT). During their stay in the ward, all TB
patients started on treatment receive the boxes of oral
drugs and a leaflet (see fig.1) outlining the time and the
numbers of pills they must take. Every day the nurse
instead of merely distributing the pills, asks the patients
to self-administer the drugs under direct observation and
corrects potential mistakes. Before being discharged
patients need to demonstrate to be able to take the
medication correctly and autonomously. At the postdischarge follow-up, adherence to treatment at home is
verified with a standardized questionnaire. After one
year of this experience we observed a reduction of
proportion of patients making mistakes in taking ATT
from 16.2% to 7.5%. Treatment adherence is critical to
guarantee the control of TB. Simple tools like the
example described might prevent lack of adherence and
reduce onset of resistance, especially among vulnerable
groups.
S385
PC-1064-05 Effectiveness of innovative TB case

finding strategies to reach the unreached slum
population in the urban areas in Nepal
S C Baral,1 P Shrestha,1 S Khanal,1 S Kandel,1 M Pandey,1
B Lamichhane,2 M Puri1 1Health Research and Social
Development Forum (HERD), Kathmandu, 2National
Tuberculosis Centre, Bhaktapur, Nepal. Fax: (977) 1 410
2016. e-mail: sushilbaral@hotmail.com
Background and challenges to implementation: Low case

detection among poor and marginalised, particularly the
urban poor is one of the primary strategic challenges
faced by NTP. The urban poor comprise 25.2% of the
total urban population of Nepal (UN, Habitat, 2005). Of
the 25.2% of the urban poor, 60.7% population reside in
slums and 7% is from marginalized group (Lumanti,
2001). The poor socio-economic conditions, health
seeking behaviour and accessibility to health services of
the urban poor make them vulnerable population for TB.
Intervention or response: HERD under TB REACH
implemented intensified case finding activities using
using GeneXpert MTB/RIF in 22 districts of Nepal. To
reach the slum groups, door-to-door screening for cough
more than 2 weeks was done by urban health volunteers.
Identified presumptive TB patients were brought to
mobile clinics or static clinic and were evaluated for TB
following the NTP algorithm (annex . Sputum samples of
those presumptive TB patients not visiting clinics were
collected by volunteers during home visits and submitted
in the clinics. Besides these, various awareness about
mobile clinics were conducted in urban areas.
Results and lessons learnt: Over the period of one year, a
total of 74 187 people were screened in the slum
population of which 4982 (7%) were symptomatic.
Out of those symptomatic 4016 (81%) were microscopy
tested (using both Xpert and microscope). Of those
tested microscopically 1509 (30%) were tested using
gene expert (30%). In total, 277 (6%) were identified as
positive cases. Inadequate participation in the screening
camps, demand for direct benefits for attending the
camps and unavailability of people at home for screening
due to their working hours were some of the challenges
faced during the screening process. Also, due to low
awareness amongst the people in the slum, they neglected
the symptoms of TB and were reluctant to give their
sputum for testing
Conclusions and key recommendations: Mobilization of
the urban health volunteers and door to door screening in
the urban slums though comes with some challenges is an
effective strategy to reach and improve early diagnosis of
TB in the urban slum population.
S386
PC-1065-05 Profile and determinants of

unsuccessful tuberculosis outcome in rural
Nigeria: implications for tuberculosis control
K Ukwaja,1 S Oshi,2 I Alobu,3 D Oshi2 1Department of
Ebonyi State, 2Centre for Development and Reproductive
Health, Enugu, 3National Tuberculosis and Leprosy Control
Programme, Ministry of Health, Abakaliki, Ebonyi State,
Background: Rural residence can be a marker for poverty

and, thus vulnerability to tuberculosis (TB). Previous
studies demonstrate that rural TB patients often do not
recognize TB symptoms. This has resulted in prolonged
delays in seeking care for TB as well as inability to
complete treatment once started. The aim of this study
was to investigate the treatment outcomes of TB and the
determinants of adverse outcomes in rural Nigeria.
Design/Methods: Adult rural TB patients treated during
2011 and 2012 in two healthcare facilities were
retrospectively reviewed. Determinants of unsuccessful
outcomes were analyzed using a multivariable logistic
regression analysis.
Results: Of 1180 rural TB patients enrolled, 494 (41.9%)
were females. The overall treatment success rate was 893
(75.7%), death, loss-to-follow-up, and treatment failure
rates were 10.9%, 8.5%, and 1.5% respectively.
Independent predictors for unsuccessful outcomes were;
public (urban) facility-care (adjusted odds ratio [aOR]
2.9), smear-negative TB (aOR 1.3) and extrapulmonary
TB (aOR 2.7), HIV co-infection (aOR 2.1), and receiving
the 8-month regimen (aOR 1.3).
Conclusion: Treatment success among rural TB patients
in Nigeria is low and receiving care at an urban public
facility, smear-negative or extrapulmonary TB, HIV coinfection, and receiving the eight-month regimen were
predictors for unsuccessful outcomes. We recommend
that: 1) urgent measures should be adopted to reduce
loss-to-follow-up and deaths among TB patients especially TB-HIV patients, 2) there is need to further expand
quality TB education, services and TB-HIV collaborative
activities in rural Nigeria, 3) targeted interventions to
reduce unsuccessful outcomes for patients in the highrisk groups should be implemented, and 4) further
studies to identify additional determinants of outcomes
including socioeconomic and clinical factors should be
conducted; and the impact of socio-economic interventions on outcomes for rural TB patients assessed.
PC-1066-05 Improving outcomes in a remote,

low-resource tuberculosis DOTS program, East
New Britain Province, Papua New Guinea,
2012-2013
D Ulaias,1,2 M OReilly,2 B Pavlin2,3 1Nonga Hospital,
Nonga, 2Field Epidemiology Training Programme Papua New
Guinea, Port Moresby, 3World Health Organization, Port
Moresby, Papua New Guinea. e-mail: moreilly24@yahoo.com
Background and challenges to implementation: Papua

New Guinea has the fourth highest TB prevalence in
WHOs Western Pacific Region. Multi-drug resistant TB
(MDR-TB) is an emerging problem in PNG. Directly

Observed Treatment Strategy (DOTS) was initiated in
quarter four of 2011 at Nonga Hospital in East New
Britain Province. Nonga is a remote, low resource
setting. At the time of initial roll out of DOTS, many
elements of the TB program, including contact tracing
and data management, were chaotic. Early post-DOTS
performance fell short of expectations: in 2012, cure rate
was at 28%, defaulter rate was unknown, and only
49.8% of sputum specimens were processed by the
laboratory within a week. New case detection identified
only a fraction of the estimated number of new TB cases
within the catchment area. We undertook a study to
assess trends in TB program performance, designed an
intervention, and reassessed performance in order to
evaluate the intervention.
Intervention: Our interventions were designed in fall
2012, and implemented from the end of 2012 throughout
2013. Interventions included a local radio TB awareness
campaign, training for health care workers on TB
screening, and development of weekly rosters to facilitate
case follow-ups and contact tracing. We identified
strategic areas of the hospital to provide ongoing
reminders to staff about TB program referrals and
protocols. We trained TB program staff in patient
documentation. We provided feedback to laboratory
staff on sputum processing performance. The lab
assigned one technician to our program.
Results and lessons learnt: In 2013, post-intervention, TB
detection increased from 6 to 18 cases per quarter. The
cure rate increased from 28.0% to 50.0%; successful
sputum diagnosis improved from 49.8% to 98.0%; the
mortality declined from 9.9% to 2.0%. The defaulter
rate was unchanged: 52.0% to 50.0%. The radio
campaign first aired during the second quarter of 2013
was replayed throughout the remainder of 2013.
Conclusions and key recommendations: Post interventions, TB cure rate, new case detections, and successful
sputum diagnosis improved, although cure rate still
lagged national target of 80%. Mortality rate declined.
Defaulter rate remains unchanged, which is concerning
given the emerging MDR-TB. Further work is required to
understand and address the high defaulter rates and to
confirm programmatic gains. Our next step is to assess
MDR-TB prevalence in our setting.
PC-1067-05 Role of community-driven mobile

clinics in the detection and management of
smear-tositive tuberculosis in underserved
areas in Sudan
A Ali1 1Geisinger Medical Center, Danville, PA, USA. e-mail:
abbaskhid@hotmail.com
Background and challenges to implementation: Res Sea

State in Sudan is a marginalized state with extreme
poverty and high prevalence of tuberculosis. The
community participation to curtail TB in Sudan is limited
Intervention or response: This project is funded by
Khider Ali Mustafa organization, a philanthropic organization established to mobilize local resources to
support TB patients and remove the associated stigma

(www.khiderali.org). A truck is modified to work as a
mobile clinic, equipped with a laboratory, a portable Xray machine, a digitilzer and DICOM software to utilize
the new technology of tele-radiology (Fig 1 and 2). The
objective is to have a free walk in mobile clinic with no
direct or indirect fees. Red Sea State has ten districts. The
first campaign covered Dordaib district.
Results and lessons learnt: Mobile walk in clinics were
held in the eight villages of the district. A total of 364
patients were examined by the physician, 111 ZN smears
were done and 33 chest x rays were done. All patients
received free treatment. One patient turned out to be
smear positive. The patient received free treatment and
monthly supplies of flour, sugar, tea and cocking oil. The
local medical assistant reports monthly to the organization about the patients health.
Conclusions and key recommendations: The mobile
clinic is able to bring the health service to the underserved
TB patients at their homes and encouraged them to step
forward without being stigmatized. The organization
will continue to cover all the state districts without
additional costs to the government or international
community. All smear positive patients will receive
treatment and will be incentivized by nutritional support.
44. Paediatric TB: MDR

PC-1068-05 Drug-resistant tuberculosis among
extra-pulmonary samples in children at a
national reference laboratory in India
A K Verma,1 G Kumar,1 J Arora,1 R Singhal,1 M Bhalla,1
V Myneedu,1 R Sarin1 1National Institute of Tuberculosis and
akv_680@yahoo.co.in
Background and challenges to implementation: Tuberculosis (TB) in children has been less discussed and
extrapulmonary tuberculosis (EPTB) is rarely addressed
in the public health literature as a childhood problem. In
this study we aimed to estimate the drug resistant
mycobacterial strains among EPTB samples in children
at a tertiary care referral centre of India retrospectively.
Intervention or response: EPTB samples from 109
children (615 years) suspected of TB were received in
the laboratory for culture and drug susceptibility testing
(DST). Laboratory data was analysed retrospectively for
a period from April 2012 to December 2014. Samples
were processed and cultured in Mycobacterial Growth
Indicator Tube (MGITe). All identified positive cultures
were subjected to DST against streptomycin; 1lg/ml,
isoniazid 0.1lg/ml, rifampicin 1lg/ml and ethambutol;
1lg/ml as per manufacturers recommendations.
Results and lessons learnt: The distribution of various
extrapulmonary samples was: pleural fluid (62;
56.88%), lymph-node aspirate (34; 31.19%), tracheal
aspirate (5; 4.58%), synovial fluid (1; 0.91%), urine (6;
5.50%) and cerebrospinal fluid (1; 0.91%). The samples
were categorised as failure (62), new cases (19), relapse
S387
(15), follow up (07), defaulter (04), and chronic (02). The

mean age of male child in this study was 11.28 (6SD
3.63) and female child was 11.99 (6SD 3.29). Of 109
cultures, 34 (31.19%) were positive for Mycobacterium
tuberculosis complex, 72 (66.05%) were negative for
mycobacterium, and 3 (2.75%) were non tuberculous
mycobacterium. Drug sensitivity results showed that 7/
109 (15.6%) strains were sensitive to all 1st line drug and
23 (21.11%) were multidrug resistant and (3/109) were
monoresistant to isoniazid.
showed the stunningly high occurrence of EPTB in
children being isolated at the referral centre. It may be a
hidden contributory factor in the transmission of drug
resistant strains and reason for partial effectiveness of
tuberculosis control program especially in developing
countries.
PC-1069-05 Drug resistant Mycobacterium

tuberculosis among pulmonary samples from
children at a reference hospital
J Arora,1 A K Verma,1 G Kumar,1 M Bhalla,1 R Singhal,1
R Sarin,1 V Myneedu1 1National Institute of Tuberculosis and
arojyoti@gmail.com
Background: 246 children who attended the outpatient

department (OPD) during the study period of 2011-2013
and whose sputum sample was received at Department of
Microbiology for liquid culture and DST were analyzed
retrospectively.
Design/Methods: The mean age of the children included
in the study was 12.94 years with majority of children,
213/246 (86.6%) in the age group of 11-15 years. There
was high prevalence of disease in females (73.98%; P
value 6 0.0001 ) with male to female sex-ratio 64/182 or
0.35. Samples received were processed by NaLC-NaOH
method and cultured using MGIT 960 system (Middlebrook 7H9 broth-based Mycobacteria Growth Indicator
Tubes; Becton Dickinson, Sparks, MD) in a BSL III
laboratory. Drug susceptibility to four first line drugs was
performed on 184 positive cultures by MGIT 960 liquid
culture system at the following critical concentrations:
rifampicin 1.0 lg/ml, isoniazid 0.1 lg/ml, streptomycin
1.0 lg/ml, ethambutol 5.0 lg/ml as per manufacturers
instructions. 54 MDR isolates were also subjected to
DST for four second line drugs at the following critical
concentrations: kanamycin 2.5 lg /ml, amikacin 1.0 lg /
ml, capreomycin 2.5 lg /ml, ofloxacin 2.0 lg /ml.
Results: Ziehl Neelsen smear microscopy was positive
for acid fast bacilli (AFB) in 155/264 (58.7%) whereas
culture was positive for 193 (78.4%) samples. DST was
conducted for 184 isolates of which 46 (25%) isolates
were sensitive to all four first line drugs. 137 (74.4%)
isolates were resistant to at least one drug and 116 (63%)
were resistant to both isoniazid (H) and rifampicin (R);
i.e. multidrug resistant [MDR]). Isolates of 54 MDR
patients were also subjected to second line drugs as
requested by clinicians. 16 (29.62%) were sensitive to all
four drugs. 29 (53.7%) isolates were resistant to
S388
ofloxacin alone. 3(5.55%) patients showed resistance to

one of the aminoglyside . 6 (11%) patients were resistant
to both ofloxacin and all the aminoglycosides (XDR).
Conclusion: Since childhood tuberculosis is a measure of
recent transmission of M. tuberculosis and an indicator
of the effectiveness of TB control programmes, high rate
of drug resistant tuberculosis in the present study is
alarming. Large scale surveillance studies are needed to
estimate the prevalence of drug resistance among
children. Active tracing and screening of household
contacts at high risk would allow children with disease to
receive a diagnosis earlier.
Conclusion: The proportion of INH resistance among

children is increasing. However, year-to-year variation in
proportions may indicate underutilization of laboratory
services to diagnose tuberculosis among children and
therefore may be an underestimation of true population
proportions. High INH resistance demonstrates the need
to explore performance of alternative regimens to treat
latent TB infection among children in high burden
settings.
PC-1070-05 Isoniazid monoresistance among

culture-positive Mycobacterium tuberculosis
from children aged 014 years in Pakistan:
implications for preventive therapy
S Shakoor,1 F Malik,1 R Hasan1 1Aga Khan University,
Karachi, Pakistan. e-mail: sadia.shakoor@aku.edu
Background: In Pakistan, where annually 275 new cases

of tuberculosis (TB) occur per 100 000 population,
children are at high risk for infection. Isoniazid
preventive therapy (IPT) is recommended by the national
childhood TB guidelines to arrest progression of latent to
active TB. However, children infected with isoniazid
(INH)-resistant TB may fail IPT. We present INH
monoresistance rates among Mycobacterium tuberculosis (M. tuberculosis) isolates cultured from children 0-14
years, during 2003-2012 in Pakistan.
Design/Methods: Data from the Aga Khan University
(AKU) laboratory was retrieved and analysed in MS
Excel. The AKU mycobacteriology laboratory is a
technical partner of Pakistan National TB Program and
part of the WHO Supra-national Laboratory Network
for tuberculosis. We receive samples from all over
Pakistan through over 200 collection units. Duplicates
in the data were removed and resistance rates and
proportions were calculated. v2 values for trends were
calculated with the EpiTools software (epitools.ausvet.
com.au).
Results: From 2003 to 2012, 2841 cultures were received
from children aged between 0-14 years. Of these, 27% (n
767) were culture-positive for M. tuberculosis. INH
monoresistance was found in 42.9% (n 329) of these
M. tuberculosis strains. Of INH monoresistant M.
tuberculosis, 88.8% (n 292) strains were pulmonary
while 11.2% (n 37) were extrapulmonary. Analysis of
trends in isoniazid monoresistance among all M.
tuberculosis isolates from children demonstrated a
significant upward linear trend (v2 5.7911; P
0.016) from 2003 to 2012. Analysis of trends in INH
monoresistance among extrapulmonary isolates also
demonstrated an increase in resistance (v214.4742; P
0.0001); however, trends showed a non-significant
downward slope for pulmonary isolates (v20.0022; P
0.96). Figure 1 shows proportions and regression lines
from 2003 to 2012 for INH resistance among pulmonary, extrapulmonary and total pediatric isolates.
PC-1071-05 To treat or not to treat? Outcomes

of an observation programme for children with
a DR-TB diagnosis not started on treatment in
South Africa
M Loveday,1 V B Sunkari,2 G Osburn,2 I Master,2 J Brust3
South African Medical Research Council, Cape Town, 2King
Dinuzulu Hospital, Durban, South Africa; 3Montefiore
Medical Center & Albert Einstein College of Medicine, New
York, NY, USA. Fax: (27) 865 126 019. e-mail:
marian.loveday@mrc.ac.za
Background: KwaZulu-Natal South Africa has a high

burden of drug-resistant TB (DR-TB). Children diagnosed with DR-TB are referred to the specialist DR-TB
hospital in the province for treatment initiation. A
number of children referred to this hospital with a
laboratory diagnosis of DR-TB are clinically well, but
because DR-TB therapy causes frequent and serious
adverse events, some children are not started on
treatment but enrolled in an observation programme.
The decision to defer therapy is based on several criteria:
chest X-ray, HIV status, body mass index (BMI),
contacts, previous history of TB and clinical signs and
symptoms of TB. We reviewed the outcome of children
enrolled in this programme after 2 year of follow-up.
Methods: Between January and December 2011, 43
children diagnosed with DR-TB were enrolled in the
observation programme. We reviewed medical records
for baseline demographic and medical characteristics
including HIV status and previous TB history. We also
collected clinical data from the 2 year follow-up period
including results of sputum culture and drug susceptibility testing, weight, signs and symptoms of TB and details
of DR-TB treatment if it was commenced.
Results: Of the 43 children enrolled in our study, the

median age was 3 years (IQR 1.5, 6.5) and 17 (40%)
were female (Table). Twenty-nine (68%) of the children
had MDR-TB; 9 (21%) mono-resistant TB of whom 7
(16%) were rifampicin resistant only; 1 (2%) pre-XDRTB and 4 (9%) XDR-TB. Of the entry criteria for the
observation programme, all children (100%) had a clear
chest X-ray and were asymptomatic, 83% were HIV
negative and 7% had a close contact with MDR-TB
(Table). The median length of time in the observation
programme was 549 days (IQR 159, 622). At the final
evaluation, 34 (80%) of the children were well, 7(16%)
had been lost to follow up, 1 child (2%) had started DRTB treatment and 1 child had died of causes unrelated to
TB.
Conclusions: This exploratory study showed that 80% of
the children enrolled in the observation programme did
not need DR-TB treatment, as they were well at the end
of the observation programme. We are currently
exploring reasons as to why these children did not need
treatment including whether contamination in the
laboratory or transient secretion had a role. In addition
we are refining the selection criteria for entry into the
programme.
S389
academic hospitals. The purpose of this investigation is

to compare a cohort of children and adolescents treated
in the DR-TB hospital with those treated in academic
hospitals to inform practices for inclusion of all children
in surveillance systems.
Design/Methods: Retrospective record review of children and adolescents enrolled for DR-TB treatment from
2005-2010 at the specialized DR hospital with those
treated in an academic hospital. Data sources were
individual patient files, site- specific and provincial
registers. Demographic, treatment, clinical and drugresistant variables were collected. Patients included in the
project had a culture-confirmed diagnosis of drugresistant tuberculosis and received treatment with
second-line drugs between 2005 and 2010.
Results: Of 320 children, 53% were treated in academic
hospitals, 65% of children with a known HIV status
were infected. Children treated in the academic hospital
were significantly younger (range 0.5-7.3) compared to
the DR hospitals (4.5-14.4), and had worse outcomes
among which more had unknown outcomes. Signs and
symptoms differed significantly between the two groups.
Diagnosis of DR TB takes place at primary health care
level, but we could not determine why some children
were referred to the provincial DR hospital and others to
the academic hospital.
Conclusion: the epidemiology, response to treatment and
clinical presentation of children treated outside of the
national DR hospital network differ significantly from
those included in the network. Provinces should seek
mechanisms to integrate surveillance and treatment with
all facilities treating children, and have dedicated
facilities for treating children not suitable for DR
hospitals.
PC-1073-05 Presentation and outcome of

children with culture-confirmed isoniazidresistant, rifampicin-susceptible tuberculosis
PC-1072-05 Children treated for drugtuberculosis in academic hospitals differ from

those treated in the provincial referral hospital,
Gauteng Province, South Africa
M Van Der Walt,1 J Lancaster,1 N Ndjeka,2 L Erasmus,3
S Masuku,1 R Odendaal1 1Medical Research Council, Cape
Town, 2South Africa National Department of Health, Pretoria,
3
National Health Laboratory Services, Pretoria, South Africa.
Fax: (27) 866 193 784. e-mail: karen.shean@gmail.com
Background: Background: Children with drug-resistant

tuberculosis (DR-TB) in South Africa should be referred
to specialized DR-TB hospitals for treatment. However,
not all DR TB hospitals have facilities to accommodate
babies or very young children, and little is known about
the epidemiology and treatment of them elsewhere. In
Gauteng the DR hospital only admit children .2 years,
while it is known that some children are treated in
A Garcia-Prats,1 L Du Plessis,1 W Burger,2 H R Draper,1

J Seddon,3 K Zimri,1 A Hesseling,1 H S Schaaf1 1Desmond
Tutu TB Centre, Stellenbosch University, Tygerberg,
2
Brewelskloof Hospital, Western Cape Department of Health,
Worcester, South Africa; 3Imperial College London, London,
UK. e-mail: garciaprats@sun.ac.za
Background: Isoniazid (INH)-resistant/rifampicin (RIF)susceptible (HRRS) tuberculosis (TB) is the most

common form of drug-resistant TB. However, the clinical
presentation, treatment regimens and outcomes are
poorly described in children. Children with HRRS-TB
may be at risk of poor outcomes, given frequent use of a
3-drug intensive phase (INH, RIF, pyrazinamide [PZA]),
which may be inadequate. Based on limited evidence,
current guidelines suggest adding a fluoroquinolone in
cases of severe disease.
Design/Methods: A retrospective cohort study of children ,14 years of age with culture-confirmed HRRS-TB
was conducted at 3 hospitals in the Western Cape, South
Africa from 2006-2012. Baseline demographic and
clinical characteristics were extracted from medical
S390
records; treatment and outcomes are described where

available.
Results: 72 children were included (median age 4.1 years;
range 0.1-14.2); 42% were male. 12 (17%) were HIVinfected; 7 (10%) had unknown HIV-status. Only 44
(61%) had a documented TB source case; of the 21
source cases with available drug susceptibility test
results, 10 had HRRSTB, 7 had multidrug-resistant
TB, and the rest had other resistance patterns. 15
children (21%) had previous TB episodes; outcomes
from previous episodes included 4 treatment failures and
2 lost to follow-up (LTFU). Thirty-four (47%) children
had pulmonary TB (PTB) only, 26 (36%) had extrapulmonary TB (EPTB) only, and 12 (17%) had both PTB
and EPTB. 54 (75%) children had severe TB using a
standard classification. 46 children had information on
treatment regimen and outcome; 41 (89%) had successful outcome while 2 were LTFU, 1 died, 1 failed
treatment and acquired RIF resistance, and 1 transferred
care. The median treatment duration was 11.5 months
(IQR 9.0-13.0); 13 received INH/RIF/PZA for .1 month
including the child with acquired RIF resistance, while 39
received a fluoroquinolone for the majority of their
treatment.
Conclusions: Few children with HRRS-TB had documented exposure to a confirmed HRRS-TB source case,
highlighting the importance of bacteriological confirmation in children. Treatment outcomes were good despite a
high proportion with severe disease, possibly related to
prolonged treatment regimens and frequent inclusion of
a fluoroquinolone. The impact of HRRS-TB in children
deserves additional investigation, especially as the
increasingly utilized Xpert MTB/RIF tests for RIF
resistance only. Optimal regimens need to be formally
evaluated in children.
PC-1074-05 Assessment of treatment practices

for pediatric drug-resistant tuberculosis in
three provinces of South Africa
S Smith,1 L Erasmus,2 B Moore,1 M Van Der Walt,3
R Odendaal,3 N Ismail,2,4 N Ndjeka,5 S Morris1 1US Centers
for Disease Control and Prevention, Atlanta, GA, USA;
2
National Health Laboratory Service, Johannesburg, 3South
African Medical Research Council, Pretoria, 4University of
Pretoria, Pretoria, 5South Africa National Department of
Health, Pretoria, South Africa. Fax: (1) 404 718 8308.
e-mail: uyi7@cdc.gov
Background: The aim of this study was to assess

effectiveness of key drug groups and regimens used to
treat pediatric drug-resistant tuberculosis (DR TB).
Design/Methods: This was a secondary analysis of a
retrospective cohort study of children (,13 years) and
adolescents (13-17 years) diagnosed with and treated for
DR TB, 2005-2008, at DR TB hospitals in three South
African provinces 1. A drug was defined as effective if
drug susceptibility test (DST) results did not show
resistance to the drug. Considering cross resistance, a
drug was defined as effective if DST results did not show
resistance to any drug in the drug group. Resistance
patterns and treatment outcomes have been previously
defined 1,2. Multivariate logistic regression was used to

estimate odds ratios (aOR) and 95% confidence intervals
(CI) of treatment success adjusting for age, sex, province,
HIV status, and weight category.
Results: Of 399 eligible children and adolescents,
patients median age was 11 years (interquartile range
6-15), 46% were male, 49% were HIV-positive, and
51% were underweight. Of 363 patients with DST
results, 83% had MDR, 9% had XDR, 5% had pre-XDR
and 4% had other DR TB. Of 359 with a treatment
outcome, 77% of patients were successfully treated while
23% had poor outcomes (death or treatment failure). In
multivariate analysis, patients receiving 45 effective
drugs (aOR7.6; 95%CI 3.4-17.3) and those receiving
68 effective drugs (aOR8.6; 95%CI 3.3-22.2) had
greater odds of treatment success than those receiving 1
3 effective drugs. However, treatment outcomes for those
receiving 68 effective drugs was not statistically
different from 45 effective drugs. Those receiving an
effective second-line injectable and those receiving an
effective fluoroquinolone had 6.4-fold (95%CI 2.2-18.4)
and 4.4-fold (95%CI 1.3-14.9) increased odds of
treatment success, respectively. Considering cross resistance, those receiving an effective second-line injectable
and those receiving an effective fluoroquinolone had
greater odds of treatment success (aOR8.9; 95%CI 3.423.1 and aOR11.4; 95%CI 3.9-32.9, respectively).
Conclusion: Children with DR TB who received an
effective second-line injectable, effective fluoroquinolone, or 74 effective drugs had greater odds of treatment
success. Susceptibility to first- and second-line drugs
should be determined quickly for children with DR TB in
order to develop a regimen containing 74 effective
drugs.
References
1 Moore B K, Anyalechi E, van der Walt M, et al. Epidemiology of drugresistant tuberculosis among children and adolescents in South Africa
20052010. Int J Tuberc Lung Dis 2015; 19: 663669.
2 World Health Organization. Guidelines for the Programmatic
Management of Drug-Resistant Tuberculosis: Emergency Update
2008. Geneva, Switzerland: WHO, 2008.
Table Effective treatment and treatment outcomes of pediatric

DR TB in three provinces of South Africa
Effective treatment
Each additional effective drug
13 effective drugs
45 effective drugs
68 effective drugs
Effective SLI
Effective SLI (excluding
potential cross-resistance)
Effective FQ
Effective FQ (excluding
potential cross-resistance)
Treatment
success vs.
death/failure
n/N
aOR* (95%CI)
351
49/351
196/351
106/351
281/302
1.69 (1.332.14)
1.0 reference
7.64 (3.3717.31)
8.58 (3.3222.17)
6.37 (2.2118.36)
269/302
303/320
8.90 (3.4323.11)
4.35 (1.2714.93)
293/320
11.35 (3.9232.87)
*Multivariate models contained age, sex, HIV status, province, and disease
severity.
SLI second-line injectable; FQ fluoroquinolone aOR adjusted odds
ratio.
PC-1075-05 Predictors of unfavorable outcomes

among children and adolescents with drugresistant tuberculosis in three provinces in
South Africa
1
D Burton, M Pinto, M Van Der Walt, S Smith,

L Erasmus,3 B Moore1 1U.S. Centers for Disease Control and
Prevention, Atlanta, GA, USA; 2South African Medical
Research Council, Cape Town, 3National Institute for
Communicable Diseases, National Health Laboratory Service,
Johannesburg, South Africa. e-mail: dburton@cdc.gov
Background: There is a high burden of drug-resistant

tuberculosis (DR-TB) among children (,13 years) and
adolescents (13 to ,18 years) in some settings in South
Africa, with potential for high morbidity and mortality.
Identifying predictors of unfavorable outcomes among
pediatric patients with DR-TB may lead to opportunities
to improve patient care.
Design/Methods: We conducted a retrospective review of
medical records at designated multidrug-resistant
(MDR) TB treatment facilities and identified 487
children and adolescents treated for any form of DRTB in three provinces in South Africa from 2005-2008.
Data were abstracted on demographics, clinical presentation, treatment course and outcomes. Completion of
DR-TB treatment or documented cure were considered
favorable outcomes, whereas death or documented
treatment failure were considered unfavorable outcomes.
Characteristics independently associated with unfavorable outcomes of DR-TB treatment were examined using
multivariate logistic regression modeling, controlling for
province and calendar year. Patients with unknown
outcomes (n 96) were analyzed separately as an
alternative category of unfavorable outcomes.
Results: Of 361 children and adolescents with DR-TB for
whom treatment outcome was known, 277 (76.7%) had
favorable outcomes and 84 (23.3%) had unfavorable
outcomes, including 80 deaths. In multivariable analysis,
factors associated with increased odds of an unfavorable
outcome included male gender (adjusted odds ratio
[aOR] 4.0, 95% confidence interval [CI] 1.8-8.9), HIV
infection (aOR 3.8, 95%CI 1.3-10.7), being severely
underweight (aOR 6.5, 95%CI 2.7-15.8), and having
extensively drug-resistant TB (aOR 10.5, 95%CI 3.432.5) compared to MDR TB. Presenting complaints
associated with increased odds of an unfavorable
outcome were non-specific constitutional symptoms
(aOR 5.9, 95%CI 2.2-16.2), neurologic symptoms
(aOR 5.2, 95%CI 1.1-24.5), and cough (aOR 4.0,
95%CI 1.6-9.6). Children aged 8 to 12 years had lower
odds of an unfavorable outcome (aOR 0.25, 95%CI 0.10.7) compared to adolescents aged 13-17 years. Only
drug resistance was significantly associated with unknown outcome (aOR 5.1, 95%CI 1.9-13.6 for mono- or
poly-resistance compared to MDR).
Conclusion: A large proportion of children and adolescents treated for DR-TB in our study sites in South Africa
had unfavorable outcomes. Close monitoring should be
provided for pediatric DR-TB patients at increased risk
of unfavorable outcomes.
S391
PC-1076-05 Treating children for drug-resistant

tuberculosis in Tajikistan with Group 5
medications
A Swaminathan,1 J Seddon,2 P Du Cros,3 S Quinnell,1
O Bobokhojaev,4 K West,3 Z Dusmatova,4 J Achar3
1
`
Medecins
Sans Frontieres,
Dushanbe, Tajikistan; 2Imperial
`
Sans Frontieres,
London, UK;
College, London, 3Medecins
4
National Tuberculosis Programme, Ministry of Health and
Social Protection, Dushanbe, Tajikistan. e-mail:
philipp.ducros@london.msf.org
Background: Treatment of drug-resistant tuberculosis

(DR-TB) in children is challenging. Besides limited
treatment options, the side-effect profile of many drugs,
in particular the group-5 anti-TB drugs (like linezolid,
clofazimine, amoxicillin/clavulanate and clarithromycin), are not fully known when used for a long term in
children for pre extensively drug resistant (pre-XDR) and
extensively drug resistant (XDR) tuberculosis. MSF runs
a comprehensive paediatric tuberculosis programme in
Tajikistan in coordination with the Ministry of Health
(MoH) that since 2011 has enrolled and treated 45 DRTB patients, including children and their household
contacts.
Design/Methods: We conducted a retrospective review of
eight children with multidrug resistant (MDR) TB and
resistance to one fluoroquinolone or at least one
injectable (pre-XDRTB); or XDRTB (MDRresistance
to one fluoroquinolone and 71 injectable) who were
treated between October 2012 and January 2015 with
71of the four group 5 drugs available in the programme:
linezolid (Lzd), amoxicillin-clavulanate (Amox/Clav),
clofazimine (Cfz) and clarithromycin (Clr); in addition
to first and second-line drugs. We examined interim
outcomes (sputum conversion; or good clinical response
in case of extrapulmonary TB at 6 months), final
outcomes (if attained) and adverse effects.
Results: One child was cured, six achieved favorable
interim outcomes and one died due to extensive disease
not responding to treatment. Adverse effects attributable
to Lzd occurred in one child warranting stoppage of the
drug; there were no significant adverse effects of the
other group 5 drugs.
Conclusion: Our early results suggest that group 5 antiTB drugs, when used appropriately, are safe and may be
effective for pre-XDR and XDRTB in children. With
proper monitoring and aggressive management of
adverse effects, the safety profile of these drugs might
be acceptable, even in long-term use; however, current
evidence remains limited. Choice of a drug depends on
availability in the country, paediatric-friendly formulations, efficacy, safety, co-morbidities, drug-drug interactions and facilities for monitoring adverse events. With
the advent of new anti-TB drugs there is a need for
increased evidence and guidelines regarding appropriate
selection and careful use of these drugs in children.
S392
45. Paediatric TB: training, BCG and

pharmacokinetics
PC-1077-05 Capacity building program for
doctors and health care workers increased the
child TB case detection rate in Bangladesh
S Ahmed,1 A L Kabir,2 R Amin,3 A Mollah,3,4
M Shahidullah,5 A H Khan6 1Shaheed Suhrawardy Medical
College, Dhaka, 2Sir Salimullah Medical College, Dhaka,
3
Dhaka Shishu Hospital, Dhaka, 4Dhaka Medical College,
Dhaka, 5Bangabandhu Sheikh Mujib Medical University,
Dhaka, 6Director General Health Services, Dhaka,
Bangladesh. e-mail: shakildr@gmail.com
Background and Challenges: Diagnosing child TB is a

challenge in 22 HBCs including Bangladesh. WHO
estimates the case detection rate should be 6-10% of
total TB cases. Recent publication shows it 4-22%. In
Bangladesh, case detection was 3.1% in 2011, 2.85% in
2012 and it was 2.74% in 2013. One of the important
reasons is lack of clinical skills of doctors and awareness
of health care workers on child TB. To build capacity of
these stakeholders (adults are difficult learner), no andragogy training modules were available. We planned to
build capacity of doctors and Health Care Workers
(HCW) of Dhaka Division (population 47 428 421) of
122 sub-district district hospitals, 17 district and medical
colleges in 2013-2014.
Intervention: To develop capacity among these groups of
stakeholder, we developed generic interactive training
modules, training video, flip chart and other training aids
on Child TB. By using these tools, a capacity building
program was conducted between November 2013 and
July 2014. This was done with support from USAID under
NTP-Bangladesh with Bangladesh Pediatric Association
as technical & implementing partner. After the training,
data were collected on child TB from NTP source on cases
up to 31st December 2014. Data of 2014 were compared
with 2010 to 2013 of Dhaka division and other divisions.
Results and lessons learnt: After development of 39
facilitators, 1181 doctors and 8345 HCW have been
trained. Pre- and post-test analysis showed significant (P
, 0.001) improvement of knowledge of the participant
irrespective of age, qualification and portfolio of doctors.
Trend of case detection of Dhaka division showed
significant increase in Dhaka division: 2010- 1849
(3.26%), 2011-2119 (3.74%), 2012- 2187 (3.60%),
2013- 2299 (3.84%) and 2014-2881(4.33%). Age
aggregated data shows increase in cases between 0-4
years increased by 156% (456/291) between 2014/2013,
while it was 116% (291/251) between 2013/2012 while
it was 121% and 104% between 5-15 yrs. No significant
increase was noticed in 6 other divisions of Bangladesh
except Chittagong Division.
Conclusions: The generic training modules made a
significant change of knowledge across age, qualification
and portfolio of doctors. Since no other intervention was
undertaken during this period in Dhaka Division, it can
be assumed that the increase in the case detection rate
was due to this capacity development program. This can
be implemented in other divisions of Bangladesh and

probably in other HBCs.
PC-1078-05 Assessment of risk and benefit of

BCG vaccine program in Taiwan
Y-J Shih,1 H-H Lin1 1National Taiwan University, Taipei,
Taiwan. e-mail: diana820111@gmail.com
Background: Recent active surveillance revealed a high

incidence of Bacille Calmette-Guerin (BCG) associated
osteomyelitis (2-9 per 100 000 per year) in Taiwan. On
the other hand, the incidence of tuberculosis (TB)
meningitis and miliary TB among those under five was
low (0.1-0.2 per 100 000 per year) under the continued
decline of TB incidence nationwide. We assessed the risk
(increase of under five TB meningitis and miliary TB) and
benefit (decrease of BCG-related osteomyelitis) of
stopping BCG vaccination in Taiwan.
Methods: We obtained the incidence rates of TB and
BCG-related osteomyelitis under five years old from the
notification system and active surveillance program of
Taiwan CDC. Assuming BCG can prevent 80% of under
five TB meningitis and miliary TB, we projected the
(counterfactual) incidence of TB meningitis and miliary
TB if the vaccine program was stopped completely. We
further estimated the incidence of TB meningitis and
miliary TB at the county level by assuming a constant
ratio between incidence of all TB and incidence of under
five TB meningitis and miliary TB across all counties. We
then projected the (counterfactual) incidence of TB
meningitis and miliary TB if the vaccine program was
stopped in low risk counties (those with all-age TB
incidence ,50 per 100 000).
Results: For the 2008 birth cohort, if BCG had been
stopped completely, the five-year cumulative incidence of
TB meningitis and miliary TB would have increased from
0.5 to 2.7 per 100 000, while the incidence of TB
osteomyelitis would have decreased from 5.9 to 0 per
100 000. The number needed to vaccine to prevent one
case of TB meningitis or miliary TB was 46 892, and the
number needed to harm to cause one case of BCGosteomyelitis was 17052. If BCG had been removed only
from low risk counties, the five-year incidence of TB
meningitis and miliary TB would have increased from 0.5
to 1.6 per 100 000, while the incidence of TB
osteomyelitis would have decreased from 5.9 to 2.0 per
100 000. The number needed to vaccine to prevent one
case of TB meningitis or miliary TB was 34273.
Conclusion: The decision to continue or discontinue
BCG vaccination should consider and balance the benefit
and risk of BCG vaccine.
PC-1079-05 The use of ethionamide and

prothionamide in childhood tuberculosis
S Thee,1,2 A Garcia-Prats,2 P Donald,2 A Hesseling,2
Universitatsmedizin,
H S Schaaf2 1Charite,
Berlin, Germany;
2
Desmond Tutu TB Centre, Stellenbosch University, Cape
Town, South Africa. e-mail: stephanie.thee@charite.de
Background: Ethionamide (ETH) and prothionamide

(PTH) are important components of treatment regimens
for multidrug-resistant tuberculosis (MDR-TB) and

disseminated forms of drug-susceptible (DS)-TB, however there is limited specific guidance for their use in
children. In order to inform treatment recommendations
for ETH and PTH, and to highlight knowledge gaps for
priority research in children, we performed a scoping
review to assess the existing evidence for ETH and PTH
use in the treatment of paediatric TB, with a focus on
pharmacokinetic (PK) and safety data.
Table Literature review on ethionamide adverse events in
children with tuberculosis
Study
TB
disease
Number
De Voogd, PTB
1963
LN-TB
59
Freour
1965
PTB
LN-TB
30
Guimard,
1966
PTB
LN-TB
107
Donald,
1987
TBM
Thee
2011
MDR-TB
Seddon
2014
MDR-TB
Van Toorn, TBM

2014
56
137
137
184
Dose
Adverse effects
ETH 10No severe

30mg
hepatotoxicity
p.o./
rectally
ETH 15
In 3 (10%) children
(12-35mg/
mild GIkg) rectally
intolerance
ETH
In 7 (6.5%) children
20mg/kg
ETH stopped:
p.o.
skin reaction and
fever (n1),
severe GIintolerance (n6)
ETH
46 (82%) children
20mg/kg
had grade 1 or 2,
p.o.
and3 (5.3%)
grade 3 elevated
liver enzymes
ETH 1579/137 (58%)
20mg/kg
children with
abnormal thyorid
function tests;
higher risk in
HIV-infected
children and
concomittant
treatment with
PAS; 18 (7.5%)
children received
thyroxine
supplementation
ETH 15Thyroxine
20mg/kg
supplementation
in 32 (23.4%)
children based on
elevated thyroidstimulating
hormone and low
free thyroxine
levels
ETH
8 (5.6%) children
20mg/kg
with grade 3
p.o.
hepatotoxicity,
following
interruption and
normalisation of
liver enzymes
original regimen
was restarted
without
reccurrence
PTB pulmonary tuberculosis, LN-TB lymph node TB, TBM tuberculous

TB, MDR-TB multi-drug resistant TB
GI gastrointestinal
Methods: We searched PubMed for English, French,

Spanish, Italian and German language articles, without
S393
date restriction. Abstracts were reviewed and full text

articles retrieved for studies with relevant information.
The reference lists of identified articles were searched for
additional relevant reports.
Results: We included information from 113 articles. ETH
and PTH have significant activity against Mycobacterium tuberculosis, including a bactericidal effect at higher
doses, and prevent drug resistance in companion drugs in
clinical studies. The mode of action and mechanisms
underlying ETH/PTH resistance show that combination
of high-dose isoniazid and ETH provides at least one
active drug for most MDR-TB strains. We identified 6
studies of ETH PK in adults and 2 in children. There were
2 studies of PTH in adults, and no studies in children. In
children, ETH at doses of 15-20mg/kg achieves maximum serum concentrations found in adults following
oral doses associated with good outcome in MDR-TB
treatment and in disseminated DS-TB. Information on
optimal pharmacodynamic targets is lacking. The effect
of crushing/breaking ETH/PTH tablets has not been
assessed, despite a lack of child-friendly formulations.
Seven studies reported on ETH adverse effects in children
(Table), which commonly included dose dependent
gastrointestinal disturbances, mild liver enzymes elevations (5-20% of children), and hypothyroidism during
long-term therapy (.40% of children). These are rarely
life-threatening and seldom necessitate cessation of ETH
therapy. Data on PTH safety in children is limited.
Conclusions: The existing limited data suggests the
recommended dose of ETH is adequate, however
additional data on PK and pharmacodynamic targets
and on safety in children are needed to inform paediatric
TB treatment.
PC-1080-05 Isoniazid hair concentrations in

children with tuberculosis
V Mave,1,2 A Kinikar,1 N Gupte,1,2 A Kagal,1 H Horng,3
G Ramachandran,4 K Dooley,2 A Gupta1,2
1
Byramjee-Jeejeebhoy Medical College Clinical Trials Unit,
Pune, India; 2Johns Hopkins School of Medicine, Baltimore,
MD, 3University of California, San Francisco, CA, USA;
4
National Institute of Research in Tuberculosis, Chennai,
India. e-mail: vidyamave@gmail.com
Background: Measurement of adherence and adequate

exposure to drugs for tuberculosis (TB) are difficult
among children. We propose an innovative non-invasive
therapeutic drug monitoring (TDM) tool among children
with TB by measuring anti-TB treatment (ATT) drug
concentration of isoniazid (INH) in hair samples that
estimates an average level of drug exposure over weeks to
months.
Design/Methods:We prospectively enrolled HIV-infected
and uninfected children ,12 years of age initiated on
thrice weekly ATT drugs including INH at 10mg/kg for 6
months for a new diagnosis of TB. INH hair concentration was measured using liquid chromatography-tandem
mass spectrometry at 1, 2, 4 and 6 months of ATT. The
median (interquartile range (IQR)) isoniazid (INH) levels
in hair was calculated for months on ATT and for
children, ,2 years, 2-,5 years, 5-12years. Univariate
S394
and multivariate random effects models were constructed to assess the difference in INH levels over months on
ATT, sex, and age groups.
Results: Of 38 children enrolled, the median age was 5.3
(IQR, 2 7.5) years; 18 (47%) were female, 3 (12%) had
HIV co-infection and 14 (48%) had pulmonary TB. All
caregivers reported .95% adherence to ATT. The
overall median INH concentration was 8.9 (5.15 15.20) ng/mg in hair and showed higher concentrations
for females at month 1 (P 0.03), a trend for higher
levels at 4 months on ATT (P 0.08), and a trend for
lower concentrations in 2-,5 years age group (p 0.05)
(Table). Figure represents intra-individual INH hair
concentrations among children (spaghetti lines) and
comparisons to cross-sectional adult hair levels (stars).
The multivariate random effects model adjusted for age,
sex and type of TB showed that INH levels at month 4
were significantly higher compared to other months (P
0.002).
Conclusion: INH drug concentrations in hair can be
measured in children on ATT and the concentrations tend
to increase over time. The average concentration and
variability of INH concentration in hair in highlyadherent children getting thrice-weekly ATT is now
established. Therefore measurement of INH hair concentrations may be an innovative TDM tool to assess
drug exposure to ATT.
PC-1081-05 Pharmacokinetics of rifampicin,

isoniazid, pyrazinamide and ethambutol in
South African infants at revised WHOrecommended dosing guidelines
A Bekker,1 H S Schaaf,1 H R Draper,1 L Van Der Laan,1
S Murray,2 L Wiesner,3 H Mcilleron,3 A Hesseling1
1
Stellenbosch University, Cape Town, South Africa; 2Global
Alliance for TB Drug Development, New York, NY, USA;
3
University of Cape Town, Cape Town, South Africa. e-mail:
adrie@sun.ac.za

recommended higher dosing of first-line TB drugs in
2009. Limited pharmacokinetic (PK) data are available

for rifampicin (RMP), isoniazid (INH), pyrazinamide
(PZA) and ethambutol (EMB) in infants (, 12 months)
who have high TB-related morbidity and mortality, and
developmental changes that may influence drug disposition.
Objectives: To determine the PK of RMP, INH, PZA /EMB at doses of 10 (10-15), 15 (10-20), 35 (30-40), and
20 (15-25) mg/kg/day, respectively in infants, and to
describe the related effects of age, weight, prematurity
and ethnicity. Liver toxicity was assessed.
Methods: Intensive PK sampling was conducted in infants
routinely initiating first-line TB therapy (fixed dose
combination tablets) in Cape Town, March 2014 - March
2015. Samples were collected at baseline (0 hours), 1, 2,
4, 6 and 8 hours following dosing. Regulatory approved
single TB drug formulations were used on day of PK
sampling, including two RMP formulations.
Results: Thirty-nine infants (26 male; 29 black) were
included and sampled, 14 (36%) with culture-confirmed
TB. Fifteen (38%) infants were premature (, 37 weeks);
5 (13%) were HIV-infected. The mean corrected age was
6.6 months (SD; 3.3), and mean weight 6.45 kilograms
(SD; 1.67). All 39 were receiving RMP, INH and PZA; 16
(41%) were also on EMB. The median maximum drug
concentration in serum (Cmax) for RMP, INH, PZA and
EMB were 2.17, 8.08, 40.4 and 1.35 lg/ml; concentration-time curve (AUC0-8)9.36, 23.35, 217.21 and 5.61
lg.h/ml; and half-life (t) 1.84, 1.87, 7.59 and 3.04
hours, respectively. Cmax and AUC0-8 differed for two
RMP formulations used. RMP was associated with a
short time to Cmax (Tmax) and t in African infants,
and delayed absorption was noted in HIV-infected
infants. No significant differences due to age, weight,
prematurity or ethnicity were observed for INH, PZA or
EMB. Alanine aminotransferase (ALT) values were
abnormal in 2 infants (one 7 x and the other 2 x elevated)
at PK sampling, which normalized without intervention.
Conclusions: WHO-recommended dosing resulted in
very low RMP exposures in infants, even if dosed at 15
mg/kg/day. PZA and INH concentrations compared well
to proposed adult target concentrations. EMB exposure
was lower than suggested adult target concentrations,
but equivalent to other paediatric studies. The low RMP
concentrations in infants require further investigation,
especially with consideration of future treatment shortening trials in children.
PC-1082-05 Impact of drug-resistant

tuberculosis on the education outcomes of
school children In the Northern Cape, South
Africa
P Pholoholo1 1Department of Health, Kimberley, South
Africa. Fax: (27) 538 300 655. e-mail:
phyllisb018@gmail.com
Background and challenges to implementation: Drug

Resistant TB (DR-TB) has been declared a threat to the
gains made globally in the control and management of
TB. The number of DR-TB patients in the Northern Cape
(NC) has increased three-fold between 2004 and 2014

and school children are reported to be among these cases.
There is uncertainty on access to education support to
school children diagnosed with DR-TB and the impact of
the DR-TB on their education outcomes.
Intervention or response: This is a retrospective, quantitative and partly qualitative study of children diagnosed
with DR-TB in the NC Province between 2012- 2014,
aged 7-18 years and attending school at time of
diagnosis. Existing files and EDRWeb data were used
to extract age, sex and treatment outcomes. The children
were followed to determine whether any education
support was received during treatment period, whether
they returned to school during or after treatment and if
they completed the school grade they were enrolled at the
time of diagnosis.
Results and lessons learnt: Twenty five (25) children were
diagnosed with DR-TB during 2012 - 2014. Three (3)
children were in primary school aged 7-12 years and 20
were in high school aged 13-18 years at the time of
diagnosis. Three (3) children died and 2 were lost to
follow-up. Smear positive sick children were unable to
attend school for up to 6 months. Only children (28.5%)
who received education support in any form passed their
grades at the end of the year. School peers played a
significant role in sharing learning material. More than
70% of the children who returned to school during the
same year had poor end of the year results. A high
proportion (83.3%) of children who were in the last two
years of schooling (grade 11 and 12) at the time of
diagnosis did not complete their high school education.
DR-TB diagnosis, prolonged treatment and poor education support are strongly associated with poor education
outcomes of school children.
Conclusions and key recommendations: Education support to school children diagnosed with DR-TB is
important in ensuring good education outcomes. Peers,
parents and teachers can play a meaningful role in
supporting the infected children during treatment and
integrating them into the school system once they are
declared noninfectious. Education of children can
contribute toward breaking the cycle of poverty associated with high TB prevalence rate.
PC-1083-05 Treatment outcomes and

tolerability of the revised WHO antituberculosis drug dosages among children
living in high HIV prevalence settings
1
M Nansumba., E Kumbakumba, P Orikiriza,

Y Boum,1,2 N Kyomugasho,1 D Nansera,2 J Kiwanuka,2
M Bonnet1,3 1Epicentre Mbarara Research Centre, Mbarara,
2
Mbarara University of Science and Technology, Mbarara,
Uganda; 3IRD LUnite Mixte Internationale 233 Recherches
Translationnelles sur le VIH et les Maladies Infectieuses,
Institut Nationale de la Sante et Recherche Medicale,

Montpellier, France. e-mail:
margaret.nansumba@epicentre.msf.org
Background: In 2010, WHO revised the dosages of antituberculosis drugs for children increasing rifampicin to
15mg/kg, isoniazid to 10mg/kg and pyrazinamide to
S395
35mg/kg. We assessed the treatment outcomes and

safety, particularly liver toxicity of children treated with
the new recommendations.
Design/Methods: Tuberculosis (TB) child suspects (1
month to 14 years) attending Mbarara Regional Referral
Hospital, with confirmed TB microbiologically, or
probable TB based on clinical and/or radiological
grounds, were treated with the new recommended
dosages. There was monitoring of alanine aminotransferase (ALT) at weeks 2, 4 and 8 of treatment. TB relapse
was monitored for after 6 months of treatment.
Results: Between April 2012 and January 2014, out of
the 396 TB suspects enrolled, 144 (36.4%) were started
on treatment. Of these, 18 (12.5%) were TB confirmed
microbiologically and 126 (87.5%) were probable TB.
They were categorized by age as follows; 64 (44.4%), 42
(29.2%), 23 (16.0%) and 15 (10.4%) for ,2, 2-4, 5-10,
and. 10 years respectively. 48 (33.3%) were HIV
positive and 48 (32.6%) were moderately to severely
malnourished (weight for height Z score). Treatment
outcomes at 24 weeks were 116 (80.5%) treatment
success, 3 (2.1%) failures, 4(2.8%) defaulters, 19
(13.2%) deaths and 1(0.7%) transferred-out. There
was no TB relapse. Among deaths, 10 (52.6%) were
severely malnourished and 9 (47.4%) were HIV infected.
Median time to death was 10 days (IQR 10-40). Thirty
(20.1%) children had serious adverse events including 1
grade 3 increased of ALT (5xULN) at week 8 which
resolved .Median (IQR) ALT values at baseline, weeks 2,
4 and 8 of treatment were 24.5 (16-39), 26 (18-38), 28.5
(21-40) and 27 (19-38) UI/L respectively.
Conclusion: Despite increase in pill burden due the
increased of the dosage of anti-tuberculosis drugs, the
new TB recommended drug dosages did not result in high
proportion of defaulters and treatment was well tolerated. More than 80% of children achieved treatment
success. However, the high proportion of deaths,
especially among malnourished and HIV TB co-infected
children are a serious concern. It highlights the importance of early and better management of comorbidities in
children with Tuberculosis in high HIV prevalence
settings.
PC-1084-05 A theory-informed approach to

identifying barriers to use of guidelines for
diagnosis of childhood tuberculosis in health
centers in Kampala, Uganda
A Katamba,1 J Ggita,2 I Ayakaka,2 P Turimumahoro,2
F Mugabe,3 M Sekadde,3 L Davis,4 A Cattamanchi5
1
2
Makerere University-University of California Research
Collaboration, Kampala, 3National Tuberculosis and Leprosy
Control Program, Kampala, Uganda; 4Yale School of Public
Health, New Haven, CT, 5Curry International Tuberculosis
Center, San Francisco, CA, USA. e-mail:
akatamba@yahoo.com
Background: Tuberculosis (TB) in children remains

under-diagnosed in high burden countries. We conducted
a mixed methods study at primary health centers in
Uganda to 1) Evaluate the number and proportion of
S396
pediatric TB diagnoses and 2) Evaluate barriers to

implementation of evidence-based guidelines (Union
Desk Guide) for pediatric TB evaluation.
Design/Methods: The study took place at 5 public,
primary health centers in Kampala, Uganda. We abstracted data from TB laboratory and treatment registers
to ascertain the number of childhood and total number of
TB cases identified in 2013 and 2014. We then conducted
a qualitative study informed by the Behavior Change
Wheel, a synthesis of 19 behavior change frameworks, to
elicit health center staff perspectives on barriers to
implementation of childhood TB diagnosis guidelines.
Thematic analysis of de-identified focus group discussion
(FGD) and interview transcripts was performed to
identify barriers related to capability, opportunity and
motivation.
Results: In 2013 and 2014, 3.7% (27/739) and 3.0% (47/
1545), respectively, of all new smear-positive TB cases
were among children ,15 years. The 56 health workers
who participated in 4 FGDs and 16 in-depth interviews
included 10 medical officers, 18 clinical officers, 21
nursing officers, 1 laboratory technologist, 4 counselors
and 2 community health workers. Barriers were primarily in the following theoretical domains: 1) Psychological
capability: lack of awareness of clinical algorithms for
childhood TB diagnosis; 2) Physical opportunity: lack of
tests/equipment, high staff workload, lack of visual aids
on childhood TB diagnosis; and 3) Reflective motivation:
beliefs that clinically diagnosing TB should be done by a
specialist and TB medications could make children suffer.
Because of these factors, clinicians consider TB mainly in
children with prolonged cough, and refer most children
who cannot produce sputum to the National Referral
Hospital.
Conclusion: TB in children is under-diagnosed at primary
health centers in Uganda. In addition to education and
training, our findings indicate that successful Desk Guide
implementation will require additional interventions that
use communication to stimulate action (persuasion);
create an expectation of reward or cost; make more
resources available for childhood TB diagnosis; model
positive examples of desired childhood TB evaluation
practices and outcomes; and/or utilize technology or
other means to reduce workload (enablement).
PC-1085-05 Four-SNP NAT2 genotyping panel

identifies slow but not rapid isoniazid
metabolizer phenotype in Ghanaian children
with tuberculosis
pharmacokinetics (PK). The 4-SNP NAT2 genotype

panel infers acetylator phenotype with high accuracy
experimentally. This study examined the relationship
between the 4-SNP NAT2 acetylator predicted phenotype status and isoniazid PK in Ghanaian children on
anti-TB treatment.
Design/Methods: PK samples were collected at times 0,
1, 2, 4, and 8 hours post-dosing in children with TB on
standard first-line antituberculosis therapy for at least 4
weeks. Drug concentrations were determined by validated methods and PK parameters calculated using noncompartmental analysis. Genotyping for NAT2 SNPs
rs1801279 (191G.A), rs1801280 (341T.C),
rs1799930 (590G.A) and rs1799931 (857G.A) was
performed using validated TaqManw real-time PCR
assays. Samples homozygous wild-type for all SNPs were
classified as rapid, those heterozygous for any one of the
SNPs were classified intermediate, and those homozygous variant for one or more SNPs or heterozygous for
two or more SNPs were classified slow acetylator
genotypes. One-way ANOVA on ranks was used to
compare PK parameters between the three genotypic
groups. Where significant (P , 0.05), post hoc multiple
pairwise comparison analysis was performed to identify
groups that were significantly different from each other.
Results: Of 59 patients with PK and genotype data
available, 5 (8.5%) had rapid, 28 (47.4%) intermediate
and 26 (44.1%) had slow acetylator genotypes. There
were no differences in baseline characteristics and
isoniazid dose between the genotypic groups. Isoniazid
area under the concentration-time curve time 0-infinity
(AUC0-) was 27.3 vs. 12.5 vs. 15.4 lg *hr/mL, half-life
was 2.8 vs. 1.5 vs. 1.7 hours, and apparent oral clearance
(CL/F) was 5.1 vs. 10.3 vs. 14.6 L/hr for patients with
slow, intermediate and rapid acetylator genotype, respectively (P , 0.001). A significant difference in AUC0, half-life and CL/F was observed between slow and
intermediate acetylator genotypes, but there were no
differences between slow and rapid or intermediate and
rapid acetylator genotypes. Half-life was also different
between slow and rapid genotypes.
Conclusion: Acetylator status inferred by the 4-SNP
NAT2 genotyping panel identified children with slow
acetylator phenotype. This economical genotyping panel
could be used to identify children at risk of reduced
isoniazid elimination and high plasma exposure.
PC-1086-05 Do childhood TB guidelines provide

optimal guidance to providers of child health
services in the identification and management
of childhood TB?
X Tang,1 A Dompreh,2 J Zhou,1,3 A Topletz,1

E Ahwireng,2,4 S Antwi,2,4 M Court,5 A Kwara1,6 1Brown
University, Providence, RI, USA; 2Komfo Anokye Teaching
Hospital, Kumasi, Ghana; 3Hunan Normal University,
Changsha, China; 4Kwame Nkrumah University of Science
and Technology, Kumasi, Ghana; 55Washington State
University, Pullman, WA, 6The Miriam Hospital, Providence,
RI, USA. e-mail: xiaoli_tang@brown.edu
R Thetard,1 C Lijinsky,2 C Powell,2 K Sawyer1

1
Management Sciences for Health, Arlington, VA, 2Office of
Sustainable Development, Bureau for Africa, United States
Agency for International Development, Washington DC,
USA. e-mail: rthetard@msh.org
Background: Isoniazid is primarily metabolized by

hepatic N-acetyltransferase type 2 (NAT-2) and polymorphisms in the NAT2 gene influences isoniazid
Background: WHOs 2014 Guidance for National

Tuberculosis Programs on the Management of Tuberculosis in Children notes that many countries developed
national policies and strategies to address childhood TB.

Practical implementation of these strategies, however, has
not always been achieved. African countries still face
challenges with identifying, diagnosing, and treating
children with TB, as well as with notifying cases to the
National TB Program. Sick children are unlikely to
present to TB clinics, so in addition to following technical
standards, an effective service delivery model would
require organizational structures linking TB and child
health services at community and primary care levels. We
assessed whether Childhood TB Guidelines adequately
guide health workers in the who, what, where and
when required for practical implementation of service
delivery aimed at increasing access by children to TB care.
Design/Methods: Data were extracted from childhood
TB guidelines from 13 English speaking countries in
Africa. Data extraction was guided by an assessment of
technical elements of TB service delivery with the
organizational elements that support effective implementation. Technical questions included diagnostic criteria,
approach to disease confirmation and treatment of TB in
children. Organizational questions explored reference to
integration with maternal and child health services
(IMCI and nutrition programs), incorporation of indications for referral and hospitalization, and description
of roles and responsibilities across all levels of the health
system.
Results: Ten countries describe when a child should be
suspected with TB; thirteen emphasize a systematic
clinical evaluation approach. Six countries include
diagnostic algorithms. In contrast, only one country
incorporates a statement with recommendations for
screening for active TB while two refer to IMCI and
one refers to screening for TB in children with undernutrition. Twelve countries describe the practical aspects
for treatment monitoring while nine describe monitoring
of IPT. Eleven countries incorporate referral guidelines;
six refer to service provision by level of service delivery or
task allocation by cadre
Conclusion: Future versions of guidelines should
strengthen content about the organizational aspects of
service delivery, especially linkages with maternal and
child health services. This could contribute to increased
efficiency and effectiveness of TB care for children.
46. Harmful effects of second hand

smoke
PC-1089-05 Comprehensive assessment and
study on the results of the Tianjin Tobacco
Control Act
G H Jiang,1,2 W Li,1,2 W Zheng,1,2 Z L Xu,1,2 D Z Wang,1,2
Y Pan,2 C F Shen,1,2 X D Xue,1,2 Y Yang,2 W D Shen1,2
1
Tianjin Centers for Disease Control and Prevention, Tianjin,
2
Tianjin Medical University, Tianjin, China. e-mail:
jiangguohongtjcdc@126.com
Background: Tianjin Tobacco Control Regulation was

formally implemented on May 31, 2012. Regulation
S397
covered 13 categories indoor area. To evaluate the

integrated effects of tobacco control by the comparison
of before and after implementing Tianjin Tobacco
Control Act in successive 4 years.
Design/Methods: The case data of myocardial infarction
was collected by Tianjin Surveillance System of New
Case Registry. The representative sample size of indoor
work place and public place where we developed
observation survey and PM2.5 monitoring was selected
with the calculation of The Survey System, and the
representative sample size people was involved in the
survey of interview by Door to Door and Intercept.
Results: Before and after implementing Tianjin Tobacco
Control Act, the post of No Smoking Sign was much
more improved in hospitals, schools, governmental
building and the waiting areas of public transportation.
The great reduction of smoking phenomenon was in the
nine types main public places excepting for hotel and
public bath room. The average level of PM2.5 decreased
about 39% in the main public places and workplaces (P
0.00) . The awareness rate about the harm of smoking
and SHS increased dramatically, such as the awareness
rate about smoking can cause stroke and heart disease
reached 66.2% and 75.0% respectively (25.2% and
58.0% before the implementation of the Regulation).
Second Hand Smoke Exposure declined 18% (P , 0.05)
compare to before implementing ACT in work place and
public place.The cases to hospitals with myocardial
infarction declined with years in the indoor workers
while the cases grow up in the entire population in
Tianjin (b0.0974, P 0.0002;b 0.0267, P 0.0153).
Conclusion: Tianjin Tobacco Control Act decreased the
smoking in public places and has already protected
people from SHS. Once again proves that tobacco related
disease is preventable, controllable. Smoke-free laws can
create a smoke-free environment to reduce the burden of
disease. And it is the effective measures to improve the
public health. It should disseminate the health effect and
social benefits of tobacco control are another way we
should do in the future, so that the Act would be
implemented comprehensively.
S398
PC-1090-05 Maternal active and passive

tobacco smoking during pregnancy: a
retrospective cohort study
V Vivilaki,1 A Diamanti,1,2 M Tzeli,1 D Bick,3 E Patelarou,3
K Lykeridou,1 P Katsaounou4 1Technological Educational
Institute of Athens, Athens, 2Gaia Maternity Hospital,
Athens, Greece; 3Kings College London, London, UK;
4
Evangelismos Hospital, Athens, Greece. e-mail:
ath.diamanti@gmail.com
Background Active and passive maternal tobacco smoking during pregnancy are recognized as high risk factors
for adverse pregnancy outcomes. There is good evidence
that change in maternal health behavior can improve
pregnancy outcomes of either stopping or reducing the
amount of smoking a woman does in pregnancy. Both
active and passive maternal smoking are generally held
responsible for a number of adverse outcomes.
Methods: 337 women were invited to complete a
questionnaire specially developed for the study. Data
on active and passive maternal smoking status in the first,
second and third trimesters of pregnancy, maternal
environmental tobacco smoke exposure at home and
work, pregnant womens experiences and beliefs towards
smoking cessation during pregnancy, were collected
using self-administered questionnaires. Descriptive statistics were used to illustrate the demographic data of
pregnant smokers and quitters in order to reveal their
smoking and passive smoking behaviors. We compared
risk perception, attitude toward smoking during pregnancy, smoking behaviors, and behaviors for avoiding
passive smoking between smokers and quitters using chisquare tests of association and t-test.
Results: Around 3/4 of the women (n 221, 73.7%, n

65, 21.7% of whom were pregnancy quitters) chose not
to smoke during pregnancy. Most pregnant smokers did
not stop smoking because they felt unable to do so (n
78, 55.8%), they did not want to stop (n 76, 25.6%) or
they did not consider it was an important health behavior
intervention (n 28, 9.3%). There were statistically
significant differences between smokers and non-smokers in pregnancy and fetal problems (v211.41; df5; P
, 0.05), newborn problems (v26.41; df2; P , 0.05)
and smoking status of the partner (v214.62; df1; P ,
0.001). The mean EPDS score for smokers was 9.72
(SD6.28; Std Error Mean 0.52) and for quitters 8.04
(SD5.17; Std Error Mean 0.41). Smokers had signifi-
cantly more depressive symptomatology as measured

using the EPDS than quitters.
Conclusion In this sample, many women chose to stop
smoking during pregnancy, however many continued to
be exposed to passive smoking environments. It is
essential that pregnant women, their partners and close
relatives are educated on the health risks of active and
passive smoking to fetal and infant outcomes as well as
the pregnant women themselves. Only smoke-free
environments will sufficiently promote optimal perinatal
health for both the woman and her fetus/newborn.
PC-1091-05 High prevalence of secondhand

smoke: a major challenge to achieve a smokefree community in Shimla
D Singh1 1Indira Gandhi Medical College, Shimla, Shimla,
India. e-mail: dsdhadwal@gmail.com
Background and challenges to implementation: Second

hand smoke (SHS) is also known as environmental
tobacco smoke (ETS).Exposure to SHS leads to increased
risk of contracting lung cancer, coronary heart disease,
lower respiratory tract infections and asthma. Women
and children are the most vulnerable groups to SHS and
yet they are helpless in reducing this risk.
Intervention or response: The present study was done to
assess the smoking and SHS patterns and awareness
regarding ill effects of smoking and SHS amongst people
of Boileauganj, urban field practice area of Department
of Community Medicine, IGMC Shimla.
Results and lessons learnt: In spite of the fact that 77% of
smokers knew about harms of SHS, 70% still reported to
smoke in room at office or home. 58% had heard of ban
on smoking in public places and 53% knew that
Himachal Pradesh has been declared as a no smoke
state. Only 50% of the participants had heard of COTPA
act. All the smokers (100%) were males and admitted to
smoking inside thier homes or inside offices.Only 23%
females knew about harmful effects of SHS though 90%
of them knew about ill effects of active smoking.
Conclusions and key recommendations: Reinforcement
of smoking ban in public places can have a major effect
on reducing SHS. Aggressive media campaigns educating
people about harmful effects of SHS need to be started.
Involvement of local communities, Resident welfare
societies and women organizations in raising awareness
as well as taking stand against SHS in their localities and
homes is the need of the hour. Say No to SHS campaigns
need to be launched with help of prominent female &
youth celebrities.
PC-1092-05 Hoshiarpur, a Total Tobacco Free

City: a district headquarter city with
comprehensive compliance with COTPA 2003
S Surila,1 R Gupta,2 J Singh Bains,3 S Goel4 1Social Network
of Education and Health Awareness, Mohali, 2Directorate of
Health and Family Welfare Punjab, Chandigarh, 3Directorate
of Health and Family Welfare Punjab, Chandigarh,
4
Postgraduate Institute of Medical Education & Research
Chandigarh, Chandigarh, India. e-mail:
surin3006@gmail.com

consumption is very harmful to human body. Indian
Government spends lacs of rupees annually to cure the
diseases caused by it. The Government spends annually
more than one lac crore of INR on the treatment of such
diseases besides the loss of above 8 lacs of precious lives.
To control this, Indian Anti Tobacco Act, COTPA-2003
is in force since Oct. 2008. Objective: Head Quarter City
of Hoshiarpur has been recently identified as to potential
comprehensively implementing COTPA-2003. Hence
study is being done: 1) To measure the level of
compliance of Sec-5, Sec-6a Sec-7, 8 and 9 of the Act
at POS across the municipal limits of the City Hoshiarpur. 2) To measure the level of compliance of Sec-6b of
the Act in educational institutions across the municipal
limits of the City.
Intervention or response: Enforcement of COTPA 2003
through health and family welfare department- tobacco
control cell Punjab at district Hoshiarpur and social
activitist and district administration year 2013-14.
Results and lessons learnt: No Active Smoking in Public
Places and other provisions of Section-4 were observed
in 99% of visited places. Points of Sale complied
98.5% with displaying of Signages and other provisions
under the Sec-6a of the Act. There was no Advertisement
found anywhere in the city except as per Sec-5 of the Act.
Eateries, Hotels and Restaurants and Point of Sale
complied 100%.There was no violation of sec- 7, 8, 9
of the Act with the packages of tobacco products being
sold. No imported tobacco products were being sold
there.
Conclusions and key recommendations: The comprehensive compliance of COTPA-2003 in Section-4, 5, 6a
and 6b, 7, 8 and 9 has been observed is above the level
considered to be the minimum bench mark of declaring
city as A Comprehensive Tobacco Free City. There is
100 % awareness level of the COTPA-2003.
PC-1093-05 Community-led tobacco control

initiative combatting secondhand smoke at
home in a North Eastern State, Tripura, India
G Tripathi,1 A Singh,2 R J Singh,3 P Tripathi4 1International
Union Against Tuberculosis and Lung Disease, South-East
Asia Office, New Delhi, 2Government of India, Agartalla,
3
University of Delhi, New Delhi, India. e-mail:
tripathi.govind@gmail.com
Background and challenges to implementation: Integrating Tobacco control programme with Village Health
Sanitation Nutrition Day (VHSND). A unique initiative
S399
for reducing tobacco use by involving community

through increased level of awareness. VHSND is a
Governmental Flexi Programme across the country being
run for providing health, sanitation and nutrition
services at the doorstep. Tripura State (Population 3.7
million) integrated VHSND with Community Led
Tobacco Control Initiative (CLTCI),a unique intervention at the community level to prepare the people for
quitting tobacco by changing the behavioural aspects.
Intervention or response: Community Led Tobacco
Control Initiative (CLTCI) using VHSND platform run
by Community and Health workers to work against
tobacco use. Here, community gets repeated doses of the
counselling, discussions and reach at cessation centers for
quitting tobacco. As a result, significant reduction
noticed in second hand smoke at households level.
Particularly youth and women groups have played vital
role through direct, peer group/family. A Cross sectional
survey of randomly selected 2856 households in 6
administrative blocks of West Tripura district was done
in the month of Jan-Feb, 2014 by the trained investigators using the pretested checklist. The core parameters of
evaluation were: Presence of some IEC material,
Interaction with children, absence of active smoking,
absence of smoking aids, absence of tobacco litter and
absence of tobacco smell at the house hold level.
Results and lessons learnt: Some IEC material found at
70% places, negative report from children about
smoking at home found around 82 % places, 84.50%
House holds found without active smoking. 82.69%
households found free from Smoking aids like ashtrays,
& lighters. More than 83 % of sampled sites didnt have
any tobacco litter (cigarette butts and bidi ends). Over
85% of sampled sites were having no evidence of recent
smoking smell.
Conclusions and key recommendations: Second hand
prevalence substantially reduced from the level of the
71.10 percent (GATS- 2009-10) to around 30 percent,
which is reduced by 60 percent. District has achieved
high level of success against SHS at household level. A
pro-active district administration approached with community driven initiation and established historic achievement.
PC-1094-05 Prevalence of exposure to smoking

and indoor air pollution among the general
population in 30 districts across 6 states of
India
S Pandurangan,1 S Mohanty1 1International Union Against
Delhi, India. Fax: (91) 11 46 05 40 30. e-mail:
sripriya14@gmail.com
Background: Tuberculosis (TB) is one among the

important public health challenges and accounts for
about 300 000 deaths per year in India. Lack of
awareness and exposure to various risk factors are the
main causes for Tuberculosis. Levels of these risk factors
vary from time to time. Assessing the levels of risk factors
is an important step in addressing the issue in the country.
S400
This study attempts to evaluate the information on

exposure to different risk factors of TB such as smoking,
drinking and living conditions in urban and rural areas.
Methodology: Two rounds of cross sectional community
based survey was conducted adopting multistage systematic random sampling in 30 districts of 6 states of
India. The estimated sample size was 2426 from 30
districts in the 6 states considering 5% change in two
years. A structured pretested questionnaire was administered to capture information on respondents exposure
to different risk factors of TB like smoking, drinking and
living.
Results: The study showed that 45.6 % of target group
exposed to smoke due to unsafe fuel usage for cooking.
Out of them 18 % are from urban and whereas nearly 3
out of 5 rural respondents reported using unsafe fuel for
cooking purpose. Over 85 percent of respondents
reported to have a safe kitchen and the remaining
reported having an unsafe kitchen (defined as a kitchen
located in the living room). A similar proportion (80%)
of respondents reported to be living in safe housing
conditions (defined as house with windows and doors).
About 19.8% of rural and 19 % of urban respondents
reported to be living in a closed living room. Usage of
unsafe heating methods is also seemingly high in rural
areas with a proportion of rural respondents more than
double (11.5%) the proportion reported by urban
respondents (4.7%). Nearly 16 % of respondents said
that they smoke and about 19 % said that they consume
alcohol. A difference of 4-6 percentage points were found
between urban and rural respondents with respect to
smoking and drinking patterns.
Conclusion: This study highlights the risk factors of
tuberculosis that are prevalent in rural and urban areas.
The remedial measures for these risk factors are mainly
socio economical in nature. However some behavioural
change strategies could be initiated on factors such as
usage of unsafe fuel, smoking and consumption of
alcohol. These changes could result in the reduction of
incidence and prevalence of tuberculosis in the country.
PC-1095-05 Reactions to smoke-free policies in

the Republic of Georgia and to messaging
strategies in support and opposition: results
from a national survey
C Berg,1 M Topuridze,2 N Maglakelidze,2 L Sturua,2
M Shishniashvili2 1Emory University, Atlanta, GA, USA;
2
National Centers for Disease Control, Tbilisi, Georgia.
e-mail: carlajberg@gmail.com
Background: Despite the ratification of the Framework

Convention on Tobacco Control, the Republic of
Georgia has limited tobacco control policies in place,
particularly in the area of smoke-free public policies.
This implies a substantial rate of secondhand smoke
exposure (SHSe). Thus, we examined: 1) the adoption
and enforcement of smoke-free policies in personal
settings, 2) reactions toward public smoke-free policies,
and 3) the persuasiveness of messaging strategies related
to smoke-free policies in bars and restaurants among

Georgian adults.
Design/Methods: We conducted a national household
survey of 1163 Georgian adults aged 18-65 years
conducted in Spring 2014. A multi-stage, clustered
sample design was used to produce representative data
with stratification done by region. We assessed the
aforementioned variables and conducted bivariate and
multivariate analyses.
Results: Our sample was on average 42.41 years old
(SD13.58), 51.1% male, 43.2% urban, 65.6% married
or living with a partner, and 53.0% had children in the
home. They had an average of 12.78 years of education
(SD2.85), and 40.0% were employed. Current tobacco
use prevalence was 54.2% among men and 6.5% among
women. Overall, 38.8% indicated that someone smoked
inside their home on a daily basis, and only 14.3% had a
complete smoke-free home policy. However, only a
minority of participants opposed a smoke-free policy in
most public settings. Less than 15% opposed smoke-free
public policies on school grounds (indoor or outdoor), on
college/university campuses, in workplaces or offices, in
public transportation, in taxis, or in vehicles with
children present. The greatest opposition was in relation
to outdoor seating areas of restaurants, bars, pubs, or
clubs (38.7%). Notably, only 33.3% and 35.1% opposed
smoke-free restaurants and bars, pubs, or clubs, respectively. Less than 25% opposed smoke-free policies in
common areas of multiunit housing or individual
apartments. Those who were older and female were
more likely to have smoke-free homes and were more
receptive to public smoke-free policies (P , .01).
Messaging with arguments regarding the health and
economic impact of smoke-free policies were most
persuasive both in support of and opposition to such
policies.
Conclusions: In Georgia, efforts must promote smokefree policies, particularly among male smokers, in order
to reduce SHSe, smoking rates, and ultimately the related
morbidities and mortality.
47. Monitoring the tobacco industry:

promotion and advertisements
PC-1096-05 Tobacco control: a messed up
agenda in India due to conflicts of interest
A Phull1 1Daily Post, Shimla, India. e-mail:
archanaphull@yahoo.com
Background and challenges to implementation: Health is

a priority, but it is often a loser in the tug of war with
tobacco promotion in India. According to WHO
Tobacco kills up to half of its users. Around six million
people are killed each year in the world due to tobacco
use. More than five million of those deaths are the result
of direct tobacco use. The WHO data says that Tobacco
use causes 22% of the global 8.2 million deaths due to
cancer and 71% of the global lung cancer deaths. Nearly
60-70% of all cancer cases in men are due to tobacco.
The scenario is dismal in India as about one million

people die of tobacco use every year in the country. As
per GATS, tobacco users in India are 34.6 % (majority of
them using smokeless tobacco), while the average age of
initiation of tobacco use is17-18 years. Gutkha is found
to be the major tobacco hazard in the country. But all this
does not seem to wake up the countrys policy makers.
And why would it? Conflict of Interest: The visible
conflict of interest within the Indian government on the
issue has led to deliberate official callousness towards the
gaps in tobacco control. The implementation of Cigarettes and other Tobacco Products Act, 2003 (COTPA)
and National Tobacco control programme has not been
effective, all for want of spirited work. India is the third
largest producer of tobacco in the world and over 38
million people in the country are now directly or
indirectly dependent on Tobacco industry for livelihood.
A Tobacco Board, instituted under the Union Commerce
Ministry, has been there since 1975 to promote tobacco
industry. The Tobacco Board is in contrast with the
tobacco control commitment of Government of India.
The controversy over the Parliamentary panel on
subordinate legislation examining the provisions of
COTPA, 2003 corroborates it. The panels conclusion
that there was no need to enhance health warnings on
tobacco products from 40 to 85 % on retail pack size and
the argument by a panel member that there was no
Indian study to link tobacco with cancer is selfexplanatory. Not surprisingly, two Members of (Indian)
Parliament on the panel are directly associated with
Tobacco industry.
Conclusions and key recommendations: Tobacco or
Health? The policy makers in India will to have to set
the priorities to save the generations next from the ills of
tobacco.
PC-1097-05 Assessment and implementation of

prohibition of tobacco advertisement,
promotion and sponsorship at point of sale in
Delhi, India
S Arora1 1Public Health Wing-1, Directorate of Health
Services, Government of Delhi, New Delhi, India. e-mail:
aroradrsk7@yahoo.com
Background information: Tobacco industry interferes

with anti-tobacco activities through POS (point of sale)
advertisement and promotion of tobacco products.
Tobacco used pattern has been shifted from adult to
minor and women group due to aggressive advertisement
strategy of tobacco companies. At present 14% students
(upto age 15 years) currently use any tobacco products
where 4% currently smoke and 12% use tobacco
products other than cigarettes. Chewable tobacco use is
also increasing despite the fact that the gutka has been
banned in Delhi since September 2012. Section 5 of
cigarette and other tobacco products act (COTPA)
prohibits tobacco advertisement, promotion and sponsorship at point of sale.
Methods: Cross-sectional analysis using observation
checklist was conducted in 331 shops in different areas/
S401
districts of Delhi state. Inspection of shops done for

presence of advertisement , its type, size of board,
presence of mandatory board regarding prohibition of
sale of tobacco to minor (less than 18 years of age) &
interviews of tobacco vendors Result:the advertisements
were observed at 79 (24%) places.among them at 32
places big stickers & at 20 places small paper pamphlets
displays observed. At 27 (41%) places big hoarding
boards were displayed, out of which 13 (48%) were
backlit. The size of all 27 boards were more than 60X45
cm. Products display were observed at 72 places.
Mandatory board regarding prohibition of sale of
tobacco to minor (less than 18 years of age) were seen
in 223 (67%) shops. 50 (63%) were aware that it is
illegal.Tobacco companies had distributed these as free
attractive gifts. Strategy adopted: all advertisements
removed on the spot with warnings. 12 court cases
initiated against the repeated offenders. media was used
stratigicaly for masses, tobacco vendor associations
sensitized.Tobacco companies were instructed to remove
all such displays from entire Delhi and stop distributing
these gifts in future, failure will attract strict actions
against them. Meeting with the municipalities, police
department and other stakeholders was conducted and
municipalities and police department were instructed to
remove all such displays from entire Delhi. For awareness among public & tobacco vendors, IEC activites were
performed through different media.
Conclusions and recommendations: Challenges regarding tobacco industry interference needs to be tackled
strategically using all stakeholders with repeated &
extensive awareness along with rigorous enforcement.
PC-1098-05 Pan masala and mouth freshener

surrogates for tobacco product advertisements
in India?
A Mangla1 1Shimla Municipal Corporation, Shimla, India.
e-mail: manglaindia@gmail.com
Background: Article 13 of FCTC and Indian tobacco

control legislation (COTPA) prohibit direct and indirect
advertisement of tobacco products in any media.
However, tobacco industry is circumventing the law
and adopting below the belt strategies; there are several
advertisements of non-tobacco products such as pan
masala (a non-tobacco product containing areca nut) and
mouth fresheners having same brand name and similar
packaging design to tobacco products seen in Indian
print and television media. Current study was carried out
to know whether advertisements in Indian print and
television media were for the intended products or serve
as a surrogate for tobacco products.
Design/Methods: Investigators did a cross sectional
survey using a pre-tested checklist in Shimla city in
April-June, 2014. Three vernacular newspapers were
scanned, selected 30 matches of Indian Premier League
(Cricket) telecasted live on a Sports TV channel were
observed for any advertisement of pan masala and mouth
freshener. Total 286 tobacco shops in Shimla city were
S402
also surveyed for availability of these products in the

market.
Results: Total 13 advertisements (of four brands) of pan
masala and mouth fresheners were observed in the
newspapers. On an average, four advertisements of pan
masala/mouth fresheners (of three brands) were aired in
each of the 30 IPL matches telecasted in a Sports TV
channel. Field survey revealed that the advertised nontobacco products and tobacco products of same brand
name and package design are also available in the
market. Total 223 (78%) shops were selling at least one
brand of pan masala/ mouth fresheners and zarda (raw
tobacco product) of same brand name and package
design. Overall, total 12 brands of pan masala/ mouth
fresheners and zarda (including the advertised brands)
were available in the market.
Conclusion: The advertisements in Indian print and
television media is serving dual purpose. Tobacco
industry is advertising the non-products (pan masala
and mouth fresheners) but also the tobacco product
(zarda) of same brand and packaging design, which
otherwise under the law was totally banned. Pan masala
and mouth fresheners advertisements become surrogate
for tobacco products (zarda) in India. There is an
immediate need for enforcement of complete ban on
such advertisements.
PC-1099-05 The first tobacco advertisementfree point of sale state in India: the Kerala
experience
P Kumar1 1Kerala State Government Health Services,
Trivandrum, Kerala, India. e-mail: aspksct@yahoo.com

Kerala, smoking prevalence was high and stringent
tobacco control was needed for further progress in TB
control since smoking has linkages with TB disease. State
health department carried out vivid activities during
2011 to 2013 to build capacity of department officials
and to sensitize the community to create favourable
environment for smooth implementation of tobacco
control laws. Department mobilized community leaders,
civil society groups, NGO, academia and other government departments. Over 23 500 notices were issued to
remove advertisement boards non complying legal
restrictions at point of sale. But shop owners did not
remove them since tobacco industry gave support. After
gaining political support, two search campaigns, serial
monitoring and legal actions were planned.
Intervention or response: About 3000 health workers
were trained and organized a week long search campaign
in January 2014 to check shops along with community
leaders, civil society representatives and police to remove
illegal boards and issue legal notice. Of the 19 455 shops
checked, illegal advertisement boards were detected in
48.8% shops and all of them were found to be violating
legal specifications. About 65% boards were removed on
the spot and correction notices issued to others. Another
search campaign conducted in May 2014 found that
about 90% of left over boards were still present but
modified by removing or masking all text messages

including brand name. All of them were given legal
notice and conducted monthly monitoring at different
levels. Evaluation was conducted by an NGO through 90
trained volunteers using pretested questionnaire during
November 2014. Study was conducted in 84 wards
selected by stratified random sampling method from the
whole state and collected information from 22 344 point
of sale by observing seven components namely big
boards, promotional materials, display of cost, display
of tobacco products, open sale of single cigarette, supply
of smoking aids and discount sale.
Results and lessons learnt: Overall compliance of all
seven components was 95.3%. Big hoardings were
absent in 98.5% sites and promotional materials were
not seen in 96.7% sites. Discount sale could not be seen
anywhere. Other components were absent in about 90%
sites. There was no significant difference between
districts.
Conclusions: Kerala became the first tobacco advertisement free point of sale state in India in November 2014.
Table District wise compliance of components (%)
DisBig
Promo- play
Hoard- tional
of
District ings materials cost
TVM
KLM
PTA
ALP
KTM
IDK
EKM
TSR
PKD
MLP
KKD
WYD
KNR
KSD
99
99
99
96
97
99
99
99
99
98
97
99
99
100
98
96
98
94
96
98
96
98
97
96
92
99
98
97
97
96
100
98
96
96
94
98
98
96
97
99
96
98
Display
of
tobacco
products
89
87
94
89
92
94
91
95
89
90
86
96
91
90
open
sale Supply
of
of
single smok- Disciga- ing count Overrettes aids sale all
94
93
98
90
94
93
90
95
89
91
85
91
94
88
97
89
97
95
97
91
90
98
94
91
89
93
89
88
100
100
100
100
100
100
100
100
100
100
100
100
100
100
95.9
94.1
97.6
93.7
95.4
96.1
94.5
97.4
95.0
95.2
91.8
97.0
95.4
95.0
PC-1100-05 Compliance level with Section 6A of

the Cigarettes and Other Tobacco Products Act
to protect minors from tobacco in Karnataka,
India: a cross-sectional study
Mahamad P,1 P Poojary1 1State Anti Tobacco
Cell-Karnataka, Bengaluru, India. e-mail:
bimahamad.kar@gmail.com
Background: Indias Tobacco Control Legislation Cigarettes and Other Tobacco Products Act (COTPA),
2003 Section 6 A of COTPA prohibits the sale of
tobacco products to and by minors. Since its inception,
Section 6A has not achieved success in banning the access
of a minor to tobacco products in India. In order to find
the compliance of COTPA , an observation was
conducted regarding inception of section 6 A in
Karnataka.
Design/Methods: In the year 2014, a Cross-Sectional
Observation Study was conducted on compliance mon-
itoring.Out of 30 Districts of Karnataka,08 districts were

selected using a structured, pre-tested checklist-based
assessment method. A total of 3084 Point of Sale (PoS)
were observed at the estimated compliance of 50%,
confidence interval of 95% and at 5% margin of error.
This survey was conducted using the cluster sampling.
Results: Out of 3084 PoS observed, 373 (11.88%) PoS
were found to have displaying a mandatory signages
used to indicate that sale of tobacco products to a
person below the age of eighteen years is a punishable
offence. In about 423 (14.19%) of the PoS, the vendor
was a minor, it was observed that 906 (28.3%) buyers
were minors visibly below the age of 18 years. In very few
cases the vendors enquired the age of the buyers in
borderline cases 184 (5.9% of vendors).This clearly
points out the failure of COTPA in curbing the buy and
sale of tobacco products to and by minors. Hence, the
compliance level of Section 6A of COTPA is fairly low in
Karnataka.
Conclusion: A majority of the PoS are not complying to
the COTPA presently, and those who are following the
norms are not abiding completely to the specifications.
The basic and most primary reason is that people who
are supposed to follow the rules are not even aware of the
presence of such rules and laws. And the vendors are least
bothered about it, as most of them are profit oriented and
no such restrictions are also levied on them during the
time of license sanction. Therefore, Efforts must be made
to make such laws more effective & easy to understand
and interpret the importance of Law by common people
who are essentially the target group of the laws that are
being passed.
PC-1101-05 Countering tobacco industry

interference in government policy on bidi tax
S V Itty1 1Kerala Voluntary Health Services, Kottayam, India.
Fax: (91) 984 781 9080. e-mail: sajuitty@rediffmail.com
Background and challenges to implementation: Beedi are

the popular smoking form of tobacco in the state of
Kerala. According to the GATS data 4.9 % of adult
population are (above 15 years) smoke beedi. Kerala has
the 13th highest beedi smokers in India. Kerala
established the first beedi manufacturing unit under the
co-operate sector with the support of govt. in 1969. This
corporate beedi making unit and trade unions working in
it influenced all government policies on tobacco control
especially on smoking tobacco products. As per the
GATS data in the state 1 53 7936 people are exposed to
beedi smoking. Last four year our advocacy to increase
beedi tax was failed due to the intervention of beedi
industry lobby. As a result beedi had no tax up to 2015
March. This year government first time in history
imposed 14.5% tax on beedi.
Intervention or response: Kerala Voluntary Health
Services made some unique strategies to counter the
tobacco company tactics. The year-long campaign
planned just after the 2014- 15 budget: Systematically
assess the previous years campaigns and identify the
challenges and loop holes. Recruited special team for
S403
study the Beedi company details including their profit,

welfare measurers, promotion tactics, management
system, linkage with political leaders etc. Prepare case
studies on promotional tactics of beedi companies.
Conduct advocacy at political and administrative front.
Mass Action by civil society coalition. Prepare and
conduct follow up campaign. Use media in strategic way.
Results and lessons learnt: The government changed its
non-tax policy on Beedi and imposed 14.5% tax on all
beedies. As a result the cost of Beedi packet rose from
25% to 30%. The point of sale study shows that 15 to
20% reduction in consumption and 4% shift from Beedi
to cigarette.
Conclusions and key recommendations: Kerala VHS will
continue its efforts in this year also, targeting a 40% to
60% tax imposition on Beedi. Long term planning and
focused intervention needed for successful advocacy
campaign.
PC-1102-05 How good is compliance with

legislative provisions on TAPS, protection of
minors and pack warnings in Himachal
Pradesh, India?
R Sharma,1 R Thakur,2 R Badrel,3 R J Singh1 1International
Office, New Delhi; 2Department of Health and Family
Welfare, Government of Himachal Pradesh, Shimla;
3
Himachal Pradesh Voluntary Health Association, Shimla,
Himachal Pradesh, India. e-mail: rsharma@theunion.org
Background: The Northern state of Himachal Pradesh

was declared Smokefree by state Health Minister on 2nd
July 2013 based on a compliance assessment study. The
government and civil society resolved to make the state
fully compliant to all provisions of Cigarette and Other
Tobacco Products Act (COTPA) namely prohibition of
TAPS, prohibition of sale to and by minors and around
100 yards of educational institutes and mandatory
display of pack warnings on all tobacco products. This
paper presents a baseline compliance assessment conducted in 2013 to know the level of compliance before
tobacco control interventions were instituted. An end
line compliance assessment is planned in June 2015 and
the final paper will also outline the progress made in
infrastructure set-up, strengthening of enforcement of
COTPA and variance in level of compliance in 2013 and
2015.
Methodology: The assessment was conducted at all 12
districts and 34 blocks headquarter towns of the state.
The sampling was done using protocols developed by
The Union and its partners. A mapping of Point of Sales
(PoS) and Educational Institutes (EI) was done and
finally 1645 PoS and 693 EI were assessed in the survey.
The investigators were trained by the Union technical
staff for data collection using an observational check list
based on various provisions of COTPA.
Results: Violations of legislative provisions of TAPS ban
was highest in Kangra and Solan and least in Kinnour
district. Product showcase was most common type of
TAPS violation found at 54.8% PoS and promotional
activities were noticed only at 5.2% PoS. The visibility of
S404
tobacco products to minors was highest in Kullu and Una

and lowest in Shimla (50.2%). Majority of PoS boards
were not as per specification of COTPA and violated its
norms in size, colour, text and picture. Tobacco products
were being sold within 100 yards of educational
institutes in 88% of institutes surveyed and enquiry for
proof of age in borderline children was done only in
0.6% cases. The tobacco products without pack warnings were mostly observed at tourist destinations where
illicit products didnt have warnings.
Conclusion: The baseline assessment showed gross
violations of above three provisions of COTPA on TAPS,
protection of children and pack warnings. A comprehensive mechanism was instituted to strengthen enforcement across the state in 2013. The baseline results will be
compared with end line compliance due in June 2015 and
presented in the conference.
PC-1103-05 Point-of-sale tobacco promotion

and risk of smoking amongst youth: a metaanalysis
1
1 1
L Robertson, C Cameron University of Otago, Dunedin,

New Zealand. e-mail: l.robertson@otago.ac.nz
Background: Several countries have now implemented

bans on PoS tobacco promotion, yet this form of
marketing remains extensive in many jurisdictions. Two
systematic reviews have indicated a positive association
between exposure to PoS tobacco promotion and
smoking amongst children and young people. We aimed
to provide an estimate of the effect size of this
association.
Methods: Inclusion criteria were quantitative observational research published in a peer-reviewed journal
between 1 January 1990 and 23 June 2014 that included
self-reported or objective measures of exposure to POS
tobacco promotion, and either behavioural or cognitive
smoking-related outcome measures. Studies were eligible
if the samples included children and adolescents aged 18
years old or younger. We used random effects models to
provide effect size estimates for 1) behavioural outcomes
and 2) cognitive outcomes. Subgroup and sensitivity
analyses were conducted to take into account the extent
of PoS tobacco promotion in the study location, study
quality and sample size.
Results: Thirteen articles were judged to have met the
inclusion criteria; 11 of these reported associations for
behavioural outcomes and 6 reported associations for
cognitive outcomes. Eight studies were conducted in the
US, 3 in Europe, 1 in New Zealand and 1 in Japan. Two
were cohort studies and the remainder were crosssectional. For the eleven studies examining behavioural
outcomes, the pooled odds ratio (OR) was 1.61 (95%
confidence interval (CI): 1.33- 1.96), and for those
examining cognitive outcomes the pooled OR was 1.32
(95%CI: 1.09 1.61). Sub-groups analysed indicated
higher odds ratios in study jurisdictions where the only
form of promotion was the powerwall of tobacco.
Conclusion: The evidence indicates a small to moderate
positive association between exposure to PoS tobacco
promotion and smoking risk amongst children and

youth. This suggests that prohibiting PoS tobacco
promotion would lead to important reductions in youth
smoking at the population level. The powerwall of
tobacco products appears to be particularly salient to
young people where no other form of retail tobacco
promotion is present. Therefore partial restrictions on
PoS promotion are inadequate; policies should prohibit
all forms of retail tobacco advertising and display.
PC-1104-05 Monitoring the tobacco industrys

compliance with the Russian Federal Tobacco
Control law in Krasnoyarsky krai
I Berezhnova,1 E Terskih,2 E Zeynalova,2 D Trufanov3
Moscow, 2Krasnoyarsk Regional Organization Center for
Health Promotion, Krasnoyarsk, 3Center for Sociological
Research Public Opinion Monitoring, Krasnoyarsk, Russian
Federation. Fax: (7) 495 984 7408. e-mail:
iberezhnova@theunion.org
1
Background: Art 5.3 FCTC provisions are well reflected

in the Federal Tobacco Control (TC) Law (Art 8 and 16)
of the Russian Federation though they may not be fully
comprehensive and well enforced by the federal enforcement agencies. Experts of the Center for Sociological
Research Public Opinion monitoring and Krasnoyarsk
regional public organization Center for Health Promotion carried out TC Legislations (No15-FZ) Art. 8 on
Tobacco Industry (TI) transparancy and accountability,
Art. 16 on TAPS and Art. 19 on POS compliance
monitoring from September to December 2014 in the
Krasnoyarsk territory.
Design/Methods: Observation and desk reseach were
used for data collection. Indicators included violations of
TAPS, PoS bans, and sales regulations and others. Areas
of monitoring included urban environment (outdoor
advertising, outdoor events, promotions, contests in
public places, Internet, Social networks, media (after
2013 - films, videos, audio, radio, and TV content) and
press (newspapers, magazines).
Results: The monitoring revealed 191 facts of violation
of the Federal Law 15-FZ Articles 16 and 19 in the
Krasnoyarsk territory. The most frequent violations
included: ban on PoS 55 cases; violations of TAPS
41 (including Internet advertisement); and brand stretching - 30 cases. Forth most common type of violations was
deviation of usage and placement of the trademark 20
cases. And the fifth is violation of the ban on tobacco
companies promotional activities, including lotteries
and competitions, with terms provided for purchase of
cigarettes 20 cases. Monitoring of the official sources of
information revealed that Art. 8 on transparency of
government interaction with the tobacco industry has
not been observed.
Conclusion: Despite adoption of the Federal Tobacco
Control Legislation in Russia, the tobacco industry
continues to apply new and traditional marketing
strategies to promote tobacco consumption in Russia.
Statistical analysis of the monitoring data was conducted
on TI activity levels from September to December 2014
in Krasnoyarsk city and region. The analysis concluded

that a daily level of 1.57 TI activities contained violations
of articles 16 and 19 of the Russian Federal legislation. In
our opinion, these findings demonstrate a high level of
violation of the Tobacco Control Law.
48. Enforcement of smoke-free

legislation
PC-1105-05 Assessment of the establishment of
smoke-free venues in national chronic disease
demonstration areas and tobacco control
activities
S405
(76.9%) out of 39 NDAs provided the information on

the creating of tobacco-free environments outside health
system. The median of the number of tobacco-free
environments outside health system is 12.5 and the
interquartile range is 30.75.
Conclusion: The outcomes showed that, indictors on
tobacco control in the evaluation system applied in
national demonstration area were not specific enough.
Further amendments were needed to complete the
evaluation system in order to achieve the goal of
decreasing the adult smoking prevalence to 25% by
2015.
Y M Bai,1 Y Zhang,1 N Ji,1 F Mao,1 S Deng,2 Y Fan,3

L H Wang,1 J X Ma1 1National Center for Chronic and
Noncommunicable Disease Control and Prevention, Chinese
Center for Disease Control and Prevention, Beijing, 2Xinjiang
Center for Disease Control and Prevention, Urumqi,
3
Nanchang Center for Disease Control and Prevention,
Nanchang, China. Fax: (86) 106 302 2350. e-mail:
bym0057@sina.com
Background: In order to further strengthen NCD

prevention and control work in China, the construction
of national demonstration areas (hereinafter referred as
NDAs) of NCD comprehensive prevention and control
was launched in 2010 by the National Health and Family
Planning Commission. By the end of 2014, 265 districts
or counties have been named NDAs (Figure). In order to
control tobacco use, a program Integrating tobacco
control into the implementation strategy of the National
Plan on NCD Control and Prevention in China was
carried out under the support from The Union, aiming to
improve tobacco control capacity building and tobacco
control policy intervention in the contexts of NDAs.
Objective To collect baseline data and understand the
present tobacco control work in NDAs.
Methods: The data between 2010 to Jun 2011 were
collected from the work reports and the community
diagnosis reports of the 39 NDAs. And the data in 2013
were collected from the questionnaire of the 39 NDAs.
Results:Tobacco control activities In 2013, execution
rates of the MPOWER strategies in the 39 NDAs from
high to low were as below: W (warning about the
dangers of tobacco use) 84.6%, P (protecting people
from secondhand smoke) 82.1%, E(ensuring the ban on
tobacco advertising, promotion and sponsorship)
53.8%, M(monitoring tobacco use and prevention
countermeasures) 48.7% and O (providing services on
smoking cessation ) 48.7%. While the R, which means
raising taxes on tobacco products, is unavailable since it
is beyond certain county or districts power. The coverage
rate of tobacco-free health institutions Among the 39
NDAs, 32 (82.1%) achieved that all health institutions
went smoke free. Geographically, 20 out of 24 NDAs
(83.3%) in the eastern region, 6 out of 8 (75.0%) in the
central region, and 6 out of 7 (85.7%) in the western
region achieved that all health institutions go smoke-free.
Tobacco-free environments outside health system Thirty
PC-1106-05 How capable is the health system

of tobacco control through current policies
G Chauhan1 1 Postgraduate Institute of Medical Education &
Research, Chandigarh, India. e-mail:
drgopal7475@yahoo.co.in

control is a health regulatory issue thus implementation
of tobacco control laws (COTPA) and FCTC framework
policies falls within the domain of health system at
national and sub national level in India. Eventually heath
system functions are preventive, promotive and curative.
The manpower within the health system encompasses the
inherited characters of the above streams. Health system
also regulates prevention of female feticide, food safety
standards, drugs and cosmetics through food safety
officers and drug inspectors but they are not from the
health streams. As tobacco control involves all skills like
search, seizure, enforcement, court cases/prosecutions,
and legal formalities so implementation through current
health system remains a question.
Intervention or response: Police has undefined role in
health regulations in India. On the request of the State
head of the police in the high level meeting in the
Himachal Secretariat on 5.5.09 a separate body was
established with the mandate to enforce all the health
regulations .The body is headed by an senior level
administrator with specialized manpower and legal
experts with capacity to exercise powers over other
departments too .In the last 5 years more than 80 000
violators has been panelized but the health authorities
S406
contribution is less than 5% as law enforcers and the

story is even worst in rest of India.
Results and lessons learnt: Enforcement of the law is the
likely solution of tobacco control India. The Health is the
nodal /regulatory agency for tobacco control with poor
capacity The role of police is minimally defined as
supporter for tobacco control Police has contributed
.95% in enforcement The key strategies like raising tax
on tobacco is beyond preview COTPA and health system
Conclusions and key recommendations: Regulation
means enforcement. Awareness as a part of program
can be well taken care within the health system but
enforcement is beyond its current capacity. Health
regulatory body at national /state levelfor law enforcement is the possible solution subject to the stake of
enforcers in it.
PC-1107-05 Comparitive study on the

enforcement of smoke-free legislation in
Bengaluru Rural District, Karnataka
Mahamad P,1 R J Singh,2 P Poojary,1 J Purushothama1
State Anti Tobacco Cell- Karnataka, Bengaluru,
2
South-East Asia Office, New Delhi, India, e-mail:
bimahamad.kar@gmail.com
Background: Section 4 of Indias tobacco control

legislation Cigarettes and Other Tobacco Products Act
(COTPA), 2003comprehensively prohibits smoking in
public places. A comparative observation was done
regarding inception of section 4 in Bangalore rural
district.
Objectives: The objective of the comparative observation
was to assess the level of compliance to Section 4 of
COTPA in Bangalore Rural and assess the level to which
COTPA has been implemented in Bangalore rural to
declare it as smoke free. Methodology: A comparative
observation was conducted before and after the effective
enforcement of section 4 of COTPA in Bangalore Rural.
An observation checklist was used for compliance
monitoring. A total of 348 public places were observed
at the estimated compliance of 50%, confidence interval
of 95% and at 5% margin of error. This survey was
conducted using the cluster sampling.
Results: 1) Before the Effective implementation of section
4: 4.2% of the public places had the NO SMOKING
signage as per COTPA.In terms of active smoking being
observed at various places, no public smoking was
observed in 73.2% of the public places during the time of
investigation. In about 67.6% of the public places
smoking aids which encourage smoking were not found.
In about 65.8% of the public places visited, there was no
smell of cigarette or bidi suggesting that someone had
smoked recently. 2) After the Effective implementation of
section 4: 87.35% of the public places had the NO
SMOKING signage as per COTPA. In terms of active
smoking being observed at various places, no public
smoking was observed in 87.64% of the public places
during the time of investigation. In about 88.21% of the
public places smoking aids which encourage smoking
were not found. In about 91.95% of the public places

visited, there was no smell of cigarette or bidi suggesting
that someone had smoked recently.
Conclusions: As a result of effective implementation of
Smoke Free Legislation, Display of COTPA Section 4
Signage had been increased up to 87.35%, No smoking
in public places has been increased to 87.64%, Smoking
aids were not found in public places increased from
67.6% to 88.21%,Over all, a majority of the places are
complying to the COTPA presently, and those who are
following the norms are also abiding completely to the
specifications as evident by the results. Approximately,
more than 3 lakh people have been protected from
Second Hand Smoking in Bangalore Rural District.
PC-1108-05 In-school prevention of tobacco use

through life skills development: the super army
model
R Kadam1 1Salaam Bombay Foundation, Mumbai, India. Fax:
(91) 222 034 809. e-mail: devika@salaambombay.org

is the single largest preventable cause of death in the
world today. In India, approximately 14.6% of youth
(13-15 years) use tobacco products. Some children start
using before the age of 10. In response to Indias tobacco
epidemic, Salaam Bombay Foundation has developed the
Super Army School Leadership Programme to prevent
tobacco use and build advocacy skills among children
and youth.
Intervention or response: The Super Army School
Leadership Programme uses a life skills development
model in schools for the 7th 9th grades. The Super
Army includes two modules Awareness (tobacco and its
health effects) and Advocacy (community-based advocacy). Through the modules, the programme develops
life skills including leadership, communication and selfempowerment. The program uses a variety of teaching
methods, including arts and theatre, to develop students
knowledge and skills. In the 8th and 9th grade, students
conduct advocacy with local stakeholders (including
education and police officials) on tobacco control, learn
about government and tools to enact social change. To
date, more than 530 000 students have participated in
the Super Army programme from 230 schools in
Mumbai.
Results and lessons learnt: In 2010, researchers from the
Harvard School of Public Health conducted an evaluation of the Super Army1. They found that Super Army
students were half as likely to report tobacco use - 4.1%
of 8th standard Super Army students, compared to 8.7%
of control school students (OR 0.51). Super Army
students were also significantly more knowledgeable
about tobacco and related legislation and had stronger
life skills.
Conclusions and key recommendations: The Super Army
is an effective model for low-cost tobacco prevention and
life skills development in schools in low- and middleincome countries.
S407
1 Sorensen G, Gupta P C, Nagler E, Viswanath K.

Promoting life skills and preventing tobacco use among
low-income Mumbai youth: effects of Salaam Bombay
Foundation Intervention. PLoS ONE 2012; 7(4):
e34982.
PC-1109-05 District level administration can be

a role model for countries in becoming smokefree
R Choudhary1 1Shikshit Rojgar Kendra Prabandhak Samiti,
Jhunjhunu, Rajasthan, India. e-mail: srkpsjjn@gmail.com
Background and challenges to implementation: India has

been very conscious of the harmful effects of tobacco use,
disease burden and related social and economic costs of
health care. Government of India enacted COTPA 2003
in May 2003 with a view to protect public health by
prohibiting smoking in public places. But, due to high
market integration of tobacco based industries, prevalence of smoking is increasing day by day in all parts of
the country.
Intervention or response: SRKPS initiated the tobacco
control program (Enforcement of section 4, 5 & 6 under
COTPA 2003) in Jhunjhunu district of Rajasthan in the
year 2005 and worked jointly with district administration towards creating a smoke free environment at public
places as well as sensitized the vendors, public place
managers, school/college principal, all the head administrators in hospitals, hotels, government offices, etc.
about the ill effect of smoking and other tobacco
products on the society. For this, one to one meeting
with stakeholders was conducted with all district level
officers (District Collector and CMHOs, District Education Officers, etc.).District level committee has been
formed to take lead role in taking necessary action for
ensuring effective implementation of COTPA. All block
level officers (SDM, BDOs, CDPOs, BEEOs, etc.) and
village level health workers (ANMS, ASHA and AWW),
Educational institution heads, government officials and
other partners were oriented to increase the outreach of
the program up to village level in sustained participatory
mechanism. Also, the implementation mechanism (Challan & violation reporting) has been activated with the
help of district administration. More than 4500 educational institution heads, 5497 NSS students, 580 health
workers and 93 members of NGOs and 888 CBOs were
oriented for making smoke-free society.
Results and lessons learnt: With efforts of SRKPS and
support of District Administration, Jhunjhunu was
declared as the first Smoke-Free district in Rajasthan
and forth in India, on 31st May, 2012. After the
successful intervention in Jhunjhunu, SRKPS with
support of The Union replicated the strategy in other
two districts Jalore and Pali in Rajasthan on 13th
November, 2013. As a result, in a short period, Pali
and Jalore become Smoke-free districts on 26th January,
2015 and many districts are in progress.
Conclusions and key recommendations: If the GOs and
NGOs work jointly with this effort, we can make the
world Smoke-free and Tobacco-free.
PC-1110-05 Does enforcement of smoke-free

rules in hospitality venues harm tourism? A
case study from three tourist states in India
R Thakur,1 P Lal,2 R J Singh2 1World Health
Organization-Randstad, Shimla, 2 International Union
Office, New Delhi, India. Fax: (91) 177 262 7601. e-mail:
ravindermph@gmail.com
Background: Rajasthan, Himachal Pradesh and Goa are

three touristy states in India with millions of tourists
(both domestic and foreigners) influx each year. Ever
since smoke free rules were enacted in 2008 across the
county; these states performed well in law enforcement.
Since 2008, more than 50 000, 15 000 and 2000
offenders including tourists has been fined for smoking
in public places in Himachal Pradesh, Rajasthan and Goa
respectively. Compliance surveys carried out in 2012 and
1st quarter of 2013 demonstrated high level of compliance in all public places including hospitality venues in
all three states. During the process of smoke free rules
implementation in these states, few policy makers and
stakeholders from tourism industry showed were apprehensive of losing tourists and overall business. Tobacco
industry also tried to fuel these concerns on different
platforms.
Design/Methods: To understand whether enforcement of
smoke free rules in these states has any negative impact
on tourist influx; researchers did a literature search and
secondary data analysis by visiting state specific and
Government of India tourism website. Researchers
revisited the state wise compliance surveys and opinion
poll reports and also collated current status of enforcement level.
Results: Compliance among the hospitality venues
(accommodation and eateries) in comparatively is at
par with other public places. Tourists flow in all the three
states was increased. In 2013, total 3.12 million tourist
visited Goa as compared to 2.78 million and 2.67 million
in 2012 and 2011 respectively. Similarly Himachal
Pradesh and Rajasthan witnessed an increase of 7-8%
in tourist influx in 2013 (Himachal Pardesh: 16.1 million
(2012) and 15.0 million (2011) and Rajasthan: 30.1
S408
million (2012) and 28.6 million (2011). There is

consistent increase in foreign tourists in all the three
states in last three years. Opinion poll in Himachal
Pradesh revealed that tourists are also in favour of smoke
free public places.
Conclusion: The apprehension that strict enforcement of
smoke free rules in the hospitality venues is not correct;
quite the contrary, implementation of these laws is often
associated with an increase in the rate of growth of
tourism revenues. This study debunks the tobacco
industry allegation that smoke-free restaurant laws
adversely affect tourism, including international tourism.
In fact tourists are motivated to demand smoke free
public places back home in their states.
PC-1111-05 #Selasatanparokok (Tuesday

without cigarettes)
D Putri Margiani,1 F Prawirakusumah2 1National
Commission on Tobacco Control, Bandung, 2National
Commission onTobacco Control, Bandung, Indonesia. e-mail:
deligiapm@gmail.com

tobacco industries kills about 521 people per day and
recruits 10.000 youth per day ages 13 14 years in
Indonesia1. Advertisements are one of the tactics used to
attract new smokers. * In response to the strong presence
of the tobacco industries in our community, we formed
an organization Bebas Rokok Bandung (Smoke - Free
Bandung) 3 years ago and have held a variety of
campaign. One campaign that has been implemented
successfully is #SelasaTanpaRokok (Tuesday Without
Cigarettes).
Intervention or response: The #SelasaTanpaRokok
campaign became a part of our Mayors weekly schedule, aimd at reaching out to the Bandung community.
#SelasaTanpaRokok has reached 2 735.435 tweets with
the original tweet representation breakdown being
32.80%, retweets consisting of 65.30%, and replies at
1.90%. The #SelasaTanpaRokok campaign started with
the programs targeted at the community. Other days of
the week include a culinary night, English language day,
local language, and car free day for example. Then
Bebas Rokok Bandung advocated to the Mayor to make
Tuesday as #SelasaTanpaRokok. The #SelasaTanpaRokok campaign was adopted by the Government of
Bandung on 6th January 2014 by the Mayors of
Bandung, Ridwan Kamil through his personal twitter
account, according to www.keyhole.co (per 2 month in
februry - march 2015)
Results and lessons learnt: As a result our advocacy and
the Mayors support, #SelasaTanpaRokok has also been
carried out by various agencies in Bandung (Departement
of Education, Departement of Health, Departement of
youth and Sport, and the City Government) which
prohibit the employees from smoking on Tuesday. The
program has also inspired other regions like Garut and
Tasik to adopt one day without cigarettes
Conclusions and key recommendations: Indonesia is the
5th highest user of social media and the most active in the
world. Other advocates can learn #SelasaTanpaRokok

campaign how to engage the power of social media
towards tobacco control goals and thereby influence
stakeholders on policy decisions regulations in their
community.
PC-1112-05 Enforcement of smoke-free laws in

public transport terminals in metropolitan
Manila, Philippines, after the sin tax law
M L Alzona,1 L Wood2 1Metropolitan Manila Development
Authority, Makati City, Philippines; 2 International Union
Against Tuberculosis and Lung Disease, Paris, France. Fax:
(632) 882 0870. e-mail: drlalzona@yahoo.com.ph
Background and challenges to implementation: Raising

taxes is the most effective measure for tobacco control.
The Philippines passed the Sin Tax Law in December,
2012. A year later, the Department of Health announced
a decrease in consumption of cigarettes among the young
and poor. To be most effective, tax increases must be
supported by other comprehensive tobacco control
strategies, including enforcement of smoke free laws.
The Metropolitan Manila Development Authority supported by a grant from Bloomberg Philanthropies
through The Union, launched enforcement of the Smoke
Free Environment policy in public transportation terminals in Metropolitan Manila in July, 2011. From
inception until March 31, 2014, a total of 220, 157
violators have been apprehended for smoking.
Intervention or response: To determine if the Sin Tax Law
(STL) had an impact on smoking in public transportation
terminals, a review of smoking violations from 3
apprehension sites was undertaken. Data was compared
from the period prior to STL (July, 2011 to December,
2012) to period following implementation of STL (July,
2013 to December, 2014). Number of enforcers and
duration of enforcement periods remained the same.
Results and lessons learnt: There was 112% increase in
apprehensions after STL. With its passage, the industry
aggressively advertised. Dont Be a Maybe signs
became prominent at points of sale. Elections inhibited
officials from acting appropriately on TAPS violations.
While STL increased price of cigarettes, they remained
cheap (a premium brand sells for USD 1.34 per pack) and
the industry swamped the market with lower price
brands. Single cigarettes are widely sold and cost from 2
to 6 US cents only. 70% smoking violators in public
transportation terminals are the working class and can
still afford cigarettes after the STL.
Conclusions and key recommendations: Despite tax
increase, smokers continued by switching to lower price
brands. Many found it difficult to quit because of
aggressive advertising and uncontrolled sale by the stick.
Information on harms must be sustained using mass
media for greater reach. Support to quit is needed. SFE
and TAPS enforcement can be strengthened with public
engagement. Monitoring smoking violations in public
places provides valuable insights into the impact of tax
increases on smoking behaviour.
49. Strategies to improve TB-HIV

screening and service integration
PC-1113-05 Enhancing integration of HIV
testing into the TB control program in Taiwan
W-T Hsieh,1 M-Y Chiou,1 P-H Lee,1 P-C Chan,1 S-L Yang,1
C B Hsu,1 C-H Chen1 1Centers for Disease Control, Taipei,
Taiwan. e-mail: bubbles1221toxi@gmail.com
Background: People living with human immunodeficiency virus (PLWH) had 29 times higher risk to develop
active tuberculosis (TB) than non-HIV infected persons,
and TB was the leading cause of death among PLWH.
Collaborative management of TB-HIV had been recommended by WHO that aimed to reduce the burden of TB
and HIV in populations affected by both infections.
However, a previous study revealed that only 17% of all
TB patients received HIV testing in Taiwan. In 2012, the
prevalence of HIV among new TB patients with age of
15-49 years old was 2.42%, which was higher than the
HIV prevalence of general population of the same age
group (0.2%).
Intervention: All newly confirmed TB patients with aged
15-49 years old in TB registry during July 2013 to June
2014 were cross-linked with the national HIV surveillance system by individual identification. If unknown
HIV status was found, the public health staff would
provide counseling and testing, or encourage TB patients
to receive HIV testing in TB care facilities within the
target period (3 months after the date of TB reporting).
We used descriptive analysis to depict the trend of
targeting patients receiving HIV testing, the HIV
prevalence and the result of linking-to-care among young
HIV-TB co-infected patients in our study.
Results: A total of 3596 TB patients aged 15-49 years old

were reported, and 63 HIV co-infected patients had been
identified before TB reporting. We found that 2664 TB
patients (overall 75.4 %) received HIV testing with the
increasing trend of testing proportion during the study
period. The majority of TB patients received HIV testing
at health care facilities (63.4%), and 15 (0.56%) newly
diagnosed HIV- infected patients were identified. The
prevalence of HIV infection among TB patients aged 15-
S409
49 years old was 2.17%, and the coverage of initiating

highly active anti-retroviral therapy (HAART) among
TB-HIV patients was up to 89.7%.
Conclusions: The proportion of HIV testing could be
improved among young TB patients by enhancing TBHIV collaborative management in Taiwan. This intervention identified HIV-positive TB patients who were
unaware of their HIV status and linked or referred to
appropriate medical care. The HIV prevalence among TB
patients was stable after enhancing HIV surveillance. To
improve TB-HIV collaborative management, further
study about the reasons why TB patients refused HIV
testing should be conducted.
PC-1114-05 Progress in TB-HIV collaborative

activities in the two regions of Ethiopia
T Anteneh,1 M Nigussie,2 D Habte,1 D J Dare,1
I J Abdulahi,3 M Aseresa Melese,1 Y Haile,4 P G Suarez5
1
Low Tuberculosis Performance (HEAL TB) Project, Addis
Ababa, 2Amhara Regional Health Bureau, Bahar Dar, 3Oromia
Regional Health Bureau, Addis Ababa, 4United States Agency
for International Development (USAID), Addis Ababa,
Ethiopia; 5Management Sciences for Health, Center for
Health Services, Arlington, VA, USA. e-mail:
tanteneh@msh.org
Background and challenges to implementation: The HIV

epidemic presents a major challenge to the control of
tuberculosis in Ethiopia. To address this challenge, the
Ethiopian Federal Ministry of Health (FMOH) in
collaboration with development partners has been
implementing the WHO recommended TB-HIV collaborative activities for over a decade now. In this paper, we
present progress in key TB-HIV collaborative activities in
two regions of the Ethiopia, which were supported by the
Help Ethiopia Address the Low Performance of TB
(HEAL TB) - five year USAID funded project, led by
Management Sciences for Health (MSH) since July 2011.
Intervention or response: Federal Ministry of Health
with support from HEAL TB provided training to health
care workers on the comprehensive TB-HIV curriculum,
regular mentoring and supportive supervision through
the district health system, and job aids and supplies. We
also ensured adequate supply of test kits, laboratory
supplies and medicines through the government system.
Health facilities that did not meet the minimum service
quality were placed on a performance improvement plan
and closely monitored every quarter to ensure progress.
We present the result of the data collected from health
facilities.
Results and lessons learnt: By June 30, 2014, the
proportion of TB patients tested for HIV increased from
70% in October 2011 to 94% in the first phase
supported 10 zones. Similarly, the testing rate increased
from 85.5% to 94% in the 11 zones supported during the
second phase. Of those tested, 89.0% and 88.4% were
linked to ART services in the phase 1 and 2 health
facilities respectively by the end of June 30, 2014. The
percentage of newly diagnosed TB-HIV co-infected
patients started on ART has reached 79%, from as low
S410
as 35% at baseline. Similarly, CPT uptake rate among

TB-HIV patients improved from 18% to 87.4%. On the
other hand, the proportion of HIV infected TB patients
declined from 11% to 7.5% over the last three years
(Table). During July 2012-June 2013, HIV testing among
TB patients in the project supported zones reached
95.4%.
Conclusions and key recommendations: Capacity-building of health care workers through training and regular
mentoring are effective interventions to improve HIV and
TB services for TB-HIV co-infected patients.
Table TB-HIV program performance in HEAL TB supported
zones, Ethiopia
Project
phase
Phase I
Phase II
Phase I
and II
(Overall)
Total
new
TB
cases
Period
Jul 12Jun 13
Jul 13Jun 14
Jul 11Jun 12
Jul 13Jun 14
Jul 13Jun 14
% TB
patients
%
%
with
started enrolled
HIV % HIV
on
on
result positive CPT
ART
22 778 95.4%
10.4% 87.4% 79.0%
23 437 94.0%
8.0% 86.0% 77.0%
20 538 88.0%
11.0%
NA
NA
17 989 94.0%
7.0% 90.0% 82.0%
41 426 94.0%
7.5% 88.0% 80.0%
PC-1115-05 Road-map to effective communitycentric TB-HIV collaboration: contribution of

the NGO-consortium of India in GF-supported
TB project
1
S Mukhopadhyay, S Cornelius, B Samuel, M Jose,

D Thuralayil Cherian,2 G Karapetyan2 1World Vision India,
New Delhi, India; 2World Vision US, Washington DC, USA.
e-mail: sugata64@gmail.com
Background and challenges to implementation: TB-HIV

collaboration is mostly health-system centric which
means the activities aim generally to medical management of co-infected cases without considering much the
importance of community sensitization and mobilization. The NGO-consortium led by World Vision India
(WV India) made meaningful contribution to bring the
TB-HIV collaborative programs at community level. It
strengthened key linkages which not only enhanced
notification of TB-HIV co-infected cases and TB cases in
HIV high-risk groups but also improved their serviceutilization.
Intervention or response: Under the GF-supported TB
Project (Axshya) WV India and 6 partners established
linkages with 1) 65 district level networks of PLHIV and
2) 450 Targeted
Intervention: (TI) HIV projects in 74 districts of 8 states
of India. Using specially designed training modules, 1617
PLHIV and 374 Program Managers of the TI projects
were sensitized on TB by Project Axshya from Apr13 to
Dec14. They were also assisted in developing and
implementing their organizational TB action plans
including creating case-wise line-lists of the referrals to
TB services and their outcomes. These activities enabled

PLHIV and HIV high risk groups to gain knowledge and
awareness on TB, reduce stigma and myths and
subsequently utilize the services under the national TB
& AIDS programs.
Results and lessons learnt: Between April 2013 to
December 2014 the PLHIV networks identified 986
presumptive TB cases (following WHOs standard
symptom selection criteria) through community level
TB education and active case search, 953 of them tested
in local RNTCP-affiliated laboratories, 83 of them found
to be SS Pulmonary TB and put on DOT. They were also
initiated ART and CPT. The HIV-TI projects referred 504
presumptive TB cases from the high risk groups like
FSWs, IDUs, migrants, MSM/TG and truckers, 493 of
them tested in RNTCP, 23 detected with SS Pulmonary
TB and put on DOT. The project spent around 30 USD to
help the PLHIV networks and TI-HIV projects to detect
one TB case.
Conclusions and key recommendations: The interventions establishing linkages with PLHIV networks and TI
projects created the road-map to effective communitybased TB-HIV collaborative programs where dual
infected, affected and vulnerable community groups
and networks are sensitized, de-stigmatized and mobilized to optimally utilize the existing services with the
help of a NGO-consortium under the GF-supported TB
project.
PC-1116-05 Uptake of partner testing among

notified TB cases in Kenya in 2013
D Nyangahu,1 R Muthee,1 B Langat2 1National
Tuberculosis, Leprosy and Lung Disease Program, Nairobi,
2
The National Treasury, Nairobi, Kenya. e-mail:
drunyaboke@gmail.com
Background: Routine HIV testing for people with

tuberculosis (TB), their partners and families, and
people with symptoms of TB is one of the priorities of
HIV-TB interventions as outlined by PEPFAR. Partner
testing and counseling is one of the critical action steps if
the mileage on PwP interventions has to be attained. The
study sought to describe the uptake of partner testing, its
effect on uptake of other TB-HIV interventions and
treatment outcomes among TB-HIV co-infected patients.
Design/Methods: A retrospective study based on the
Kenya national TB surveillance data of 2013 was
conducted to describe the uptake of partner testing
among TB patients. Descriptive statistics were used and
chi-square tests were applied to establish if there was any
association among the paired variables using SPSS.
Results: Partner testing uptake was 37.1 % and 5 % of
the partners declined the HIV test. The discordance rate
among HIV/TB patients was 22.3 % while among
HIV-/TB patients was 2%. Amongst TB-HIV patients
whose partners were tested, 89.8 % of them were on
ART as compared to 79.7 % of those whose partners had
not been tested. There were better outcomes (treatment
completion) among those who had their partners tested,
and fewer out of controls were recorded from HIVpositive TB patients whose partners had been tested.
Conclusion: HIV Partner testing among TB patients is an
acceptable public health intervention that can roll out in
a TB control program. The partner testing uptake is low
hence it is paramount to enact practical and plausible
measures to mitigate challenges and setbacks to partner
testing. Special interventions need to be tailored to hard
to reach areas, as in addition to having low uptake, the
public sector registered lower performance in comparison to the private sector in these areas. HIV partner
testing can contribute to better linkage to ART and CPT
services. There is need to undertake further research on
provider and client factors that affect uptake of partner
testing.
S411
diagnosis to TB treatment initiation was 14days (1224) and, among 99 TB-HIV patients with available
data, median (IQR) CD4 count was 140cells/mm3 (59323). A greater proportion of patients at CIP HF had
category II TB treatment (19% vs. 10%, P 0.002).
Other characteristics did not differ significantly by study
arm.
Conclusion: Seventy percent of pulmonary TB patients
were clinically diagnosed; 16% were without sputum
smear. Improved clinical and laboratory diagnostic
capacity for TB among PLHIV in Lesotho is urgently
needed.
PC-1117-05 The majority of TB cases among

PLHIV in Lesotho are clinically diagnosed:
START study baseline characteristics
D Oconnor,1,2 K Frederix,1 L Maama-Maime,4 S Saito,1,3
Y Hirsch-Moverman,1,3 B Pitt,1 A Howard1,3 1ICAP at
Columbia University, New York, NY, 2Columbia University,
College of Physicians and Surgeons, New York, NY,
3
Columbia University, Mailman School of Public Health, New
York, NY, USA; 4Lesotho Ministry of Health, Maseru, Lesotho.
e-mail: koenfrederix70@gmail.com
Background: The Start TB patients on ART and Retain

on Treatment (START) Study is an ongoing clusterrandomized trial evaluating the effectiveness, costeffectiveness and acceptability of a combination intervention package (CIP) vs. standard of care (SOC) to
improve early ART initiation and TB treatment success
among adults with TB-HIV in Berea district, Lesotho.
Methods: Twelve health facilities (HF) were randomized
to CIP or SOC, with stratification by HF type (hospital or
health center). A standardized survey was administered
to assess available services prior to CIP implementation.
CIP includes nurse training and mentorship; patient
education and transport reimbursement for patients and
treatment supporters; and village health worker (VHW)
and SMS text message adherence support. After CIP
implementation, routine clinical data were collected for
all new adult TB-HIV patients. Descriptive statistics were
used to characterize baseline facility and participant
characteristics. Chi-square, Fishers Exact, and t-tests
were used to compare characteristics by study arm.
Results: Of the 12 HF, 4 are managed by Government of
Lesotho (GOL), 7 by Christian Health Association of
Lesotho (CHAL), and 1 by Red Cross; 10 are rural and 2
semi-urban. Prior to CIP implementation, 6 HF had
nurses trained in TB-HIV co-treatment, 12 provided
VHW adherence support, 6 provided patient education,
and none provided transport reimbursement or SMS
adherence support. Among 624 new TB-HIV patients
between 4/2013-9/2014, 62% were at CIP HF, mean age
was 39, 57% were male, 15% were factory or
mineworkers, 85% had category I TB treatment, 85%
had pulmonary TB (smear positive 30%, smear negative
54%, no smear 16%) and 21% were on ART at TB
treatment initiation. Median (IQR) time from HIV
PC-1118-05 Improving antiretroviral therapy

uptake for TB-HIV co-infected patients in
Nkangala District, Mpumalanga Province
M Sithole,1 J Zingwari1 1University Research Co LLC,
Pretoria, South Africa. e-mail: jebbyzing@gmail.com
Background: Nkangala District is located in Mpumalanga Province, one of the top four high HIV burden
provinces in South Africa. In 2011, the district reported
5742 new TB cases. 61.5% of the TB patients are coinfected with HIV. South African TB guidelines recommend antiretroviral therapy (ART) for all TB HIV coinfected patients. Baseline assessment conducted in
January- March 2012 in Nkangala district, revealed
low ART uptake among TB-HIV co-infected patients
with 54.8% of TB HIV co-infected patients on HAART.
Poor TB HIV service integration, poor adherence to
guidelines and poor recording and data management
were identified as contributors for poor performance.
Interventions: To address these challenges, the USAID TB
program collaborated with the district health team and
conducted support supervisory visits, focusing on high
case load, poorly performing facilities. Activities conducted during these visits included mentoring and
coaching, data validation and analysis of paper-based
TB registers and patient files and data from the electronic
TB register to identify performance gaps related to ART
initiation. Training was given to health care workers on
S412
quality improvement. Facilities were supported in

developing and implementing quality improvement
plans.
Results and lessons learnt: In Nkangala district, ART
uptake increased from 54.8% in Jan-March 2012 to
73% in Oct-Dec 2014 as shown in figure below.
Integration improved patient management as this encouraged team based approach to TB and HIV services
thus improving quality of services.
Conclusion: Integration of TB and HIV services in
facilities improves the management of TB HIV coinfected patients. This should be supported by comprehensive and integrated supervision support. As facilities
scale up integration of TB and HIV services, there is need
to constantly monitor implementation and assess the
impact of different models using standardised outcome
measures. Data management is critical to the successful
integration of TB and HIV services.
PC-1119-05 Performance of TB screening and

testing in sub-populations of people living
with HIV in Uganda
F Semitala,1,2,3 C Yoon,3,4 J Katende,1,3 A Andama,2,3
P Byanyima,3 I Ayakaka,3 M Kamya,1,2,3 A Cattamanchi3,4
1
Makerere University Joint AIDS Program, Kampala,
2
3
Infectious Diseases Research Collaboration, Kampala,
Uganda; 4San Francisco General Hospital, University of
California San Francisco, San Francisco, CA, USA. e-mail:
semitala@gmail.com
Background: WHO recommends that all PLHIV should

be screened for active tuberculosis (TB) using a symptom-based tool, followed by confirmatory testing if
symptomatic. However, the utility of symptom screening
and confirmatory testing among sub-populations of
PLHIV are unknown. We evaluated the diagnostic
accuracy of the WHO symptom screen, smear microscopy and Xpert MTB/RIF (Xpert) in two important subpopulations: clients new to and clients already enrolled in
HIV/AIDS care.
Design/Methods: We enrolled consecutive PLHIV with
CD4 cell-counts 6350 cells/uL presenting to the Mulago
Immune Suppression Syndrome Clinic (Kampala, Uganda) from July 2013-April 2015 for antiretroviral therapy
(ART) initiation. We administered the WHO symptom
screen and collected sputum for smear microscopy, Xpert
and culture. We calculated the diagnostic accuracy of

symptom screening in all clients and confirmatory tests
(smear and Xpert) among symptomatic clients in
reference to culture. We compared clinical characteristics
and diagnostic accuracy between clients new to and
clients already enrolled in HIV/AIDS care.
Results: Of 665 clients, 232 (35%) were new to HIV/
AIDS care. These clients had significantly lower median
CD4 cell counts (144 vs. 199 cells/uL, P 0.003) and a
higher prevalence of culture-positive TB (24% vs. 9%, P
, 0.001) than clients already enrolled in HIV/AIDS care.
The majority of clients new to (93%) and already
enrolled in (85%) HIV/AIDS care screened positive by
the WHO symptom screen. In both sub-populations, the
WHO symptom screen had high sensitivity (98%
[95%CI 90-100%] and 95% [95%CI 82-99%], respectively) but poor specificity (8% [95%CI 5-14] and 16%
[95%CI 13-20%], respectively) for active TB. The
sensitivity of smear (35%, 95%CI 25-46%) and Xpert
(47%, 95%CI 37-58%) was poor but specificity of both
tests was high (99%, 95%CI 98-99.9%), with no
clinically significant differences between sub-populations.
Conclusion: TB prevalence is high among PLHIV
initiating ART in Kampala, especially among clients
new to HIV/AIDS care. Symptom-based screening results
in nearly all clients requiring confirmatory TB testing.
Neither smear nor Xpert were sufficiently sensitive
confirmatory tests, missing more than half of all TB
cases. Until better screening and rapid confirmatory tests
are available, routine culture should be considered prior
to ART initiation, particularly for all clients new to HIV/
AIDS care
PC-1120-05 Improving antiretroviral therapy

uptake among TB-HIV co-infected clients in
care: lessons from quality improvement
implementation in Kampala, Uganda
A Burua,1 E Mabumba,2 D Lukoye,1 M Kenneth,1
M Ruhweza,1 F Mugabe,2 B Woldemariam,3 P G Suarez4
1
Management Sciences for Health/Track Tuberculosis Activity
Project, Kampala2Uganda National TB and Leprosy Program,
Kampala, Uganda; 3Management Sciences for Health, Addis
Ababa, Ethiopia; 4Management Sciences for Health,
Arlington VA, USA. e-mail: aburua@msh.org
Background and challenges to implementation: HIV and

TB continue to be diseases of public health concern in
Uganda. To decrease the burden of HIV among TB
patients, the national guidelines recommend early
initiation of antiretroviral therapy for TB-HIV coinfected patients. At the end of 2012, ART uptake
among TB-HIV co-infected patients in Kampala was
only 56% far below the national targets of 100%.
Intervention or response: The USAID funded TRACK TB
project provides technical support to Kampala Capital
City Authority in TB control including implementation
of continuous quality improvement approaches in TB
care among other things. An assessment of TB HIV
collaborative activities indicated gaps in provision of
ART to TB-HIV co-infected patients, inadequate pro-
vider capacity to provide ART and incomplete recording.

Five out of the 20 high volume health facilities in
Kampala piloted the Quality Improvement (QI) interventions which includes problem identification, analysis,
team problem solving, implementation of solutions and
monitoring. The Health care providers in these five
facilities were trained and received monthly coaching in
QI methods which included onsite training and mentorship on management of ART and allocation of clinic staff
to harmonize TB and ART registers regularly. The other
facilities received routine mentorships in TB care. The
facilities performance was then measured by tracking
indicator on proportion of TB-HIV co-infected patients
that received ART.
Results and lessons learnt: ART provision for TB-HIV co
infected patients at the 5 pilot facilities improved from
71.4% in February to 100% in December 2014. In
contrast, 8 health facilities comparable with the pilot
facilities that had no QI interventions improved from
67.6% in February to 82.4% in July but later declined to
66.7% in December 2014. Challenges of poor involvement of the facility leadership in QI activities affected
performance and a lesson learnt is that on-site mentorships and QI coaching should target the facility
leadership as well to achieve sustainable improvements.
Conclusions and key recommendations: Mentorship
combined with application of QI methods and engagement of health facility leadership were found to be
feasible for improvement of TB care and yields more
improvement compared to provision of only clinical
mentorship support.
PC-1121-05 Factors related to sub-standard

offer and documentation of HIV testing for
persons with tuberculosis in a low-incidence,
high-resource setting
R Bhavaraju,1,2 M Passannante,2 C Delnevo,3 E Pevzner,4
M J Lewis3 1International Union Against Tuberculosis and
Lung Disease, New York, NY, 2Rutgers University, Newark, NJ,
3
Rutgers University, New Brunswick, NJ, 4US Centers for
Disease Control and Prevention, Atlanta, GA, USA. e-mail:
rbhavaraju@theunion.org
Background: The recommended offer of HIV testing to

persons with TB in the United States is lower than
national standards and is less than in many resource-poor
countries. US patients have high access to testing and the
ability to manage co-infected patients. This study
explored factors related to offer and documentation of
HIV testing in the State of New Jersey.
Design/Methods: Surveillance data (2000-2013) for
9502 persons with TB were analysed looking at age,
sex, substance use, correctional/long term care facility
residence, birthplace, race/ethnicity and year. Using
outcomes of offer and documentation of testing, v2 tests
were done and 2 logistic regression models created,
stratifying by provider type. The latter was based on a
qualitative study by these authors that indicated a
difference in HIV testing exists in patients seeing private
and public health providers.
S413
Results: Less than 60% of patients had a documented

HIV test status. Fifty percent of patients were seen by
public health providers and this increased yearly as did
the offer and documentation of HIV testing (P ,
0.0001). In the 3 strata models, females were less likely
to be offered testing but documentation of HIV testing
status was 17-25% more likely to be available. Birthplace was not a significant predictor in all strata,
however, in the public sector, non-US-born persons were
~30% less likely to be offered HIV testing (OR.70;
95%CI .58-85), but with greater odds of documented
testing status (OR 1.34; 95%CI 1.13-1.60). By age,
persons 65 were most likely to have HIV testing offered
compared to other age groups but presence of documented results were 1.7-3.5 times the odds (95%CI
1.46-4.16) in other age groups. Substance users had
double the odds of HIV testing offered in all strata (OR
2.289-2.61; 95%CI 1.77-4.37). Persons of Black race
also had 2-3 higher odds of HIV testing offered and
documentation compared to White persons in all strata
(OR 2.45-3.06; 95%CI 1.81-3.83); other races (with
the exception of Hispanics under private care) were less
likely to be offered or have documented HIV testing.
Conclusion: Despite access, universal HIV testing rates in
TB patients is low, and varies by provider type and
demographic characteristics. The lack of documented
testing status in a system with advanced disease
surveillance provides an incomplete picture and lost
opportunities for the appropriate care of co-infected
persons. Educational outreach to TB providers is needed.
S414
Abstract presentations, Sunday, 6 December
SUNDAY
6 DECEMBER 2015
e-POSTER SESSIONS
11. m-health solutions
EP-185-06 WelTel LTBI: participant experiences
with a text-messaging intervention to improve
latent tuberculous infection treatment
adherence in Canada
K Smillie,1,2 D Mahal,2 M Van Der Kop,2,3 J Johnston,1,2
R Lester2 1British Columbia Centre for Disease Control,
Vancouver, BC, 2University of British Columbia, Vancouver,
BC, Canada; 3Karolinska Institutet, Stockholm, Sweden.
e-mail: kirsten.smillie@bccdc.ca
Background: Successful treatment of latent tuberculosis

infection (LTBI) is critical to reduce the impact of TB;
however, in North America, fewer than half of individuals starting LTBI treatment complete therapy. A multisite, randomized controlled trial is testing the effectiveness of a text-messaging intervention to improve LTBI
treatment completion in British Columbia, Canada. We
conducted a qualitative study to understand participants
experiences with the intervention.
Objectives: 1) acceptability of the intervention; 2) if the
intervention mitigated barriers to care; and 3) whether it
affects clients perceptions of the healthcare system.
Design/Methods: Participants were recruited from two
urban TB clinics in British Columbia. Individuals were
eligible to participate if they were 719 years of age;
starting LTBI treatment; able to send text messages in
English or have assistance text-messaging; access to a
mobile phone. Interviews were conducted either inperson or over the telephone and were recorded and
transcribed verbatim. Interview data were analyzed using
inductive thematic analysis. Transcripts were read
numerous times by two researchers who then coded
important concepts, events, and experiences using NVivo
9 software.
Results: To date, we have conducted 12 interviews with
participants, eight of whom were women. Recruitment
will continue until data saturation is reached. The
median age of participants was 37 years (range: 23
56). Of the 12 participants, nine had completed LTBI
treatment, while four stopped treatment early due to side
effects. All interviews were conducted in English except
for one, which was conducted in Punjabi. Participants
found the intervention easy to use. The intervention was
appreciated for its simplicity, as it only required basic
English to respond to the open-ended question (Are you
OK?) with a yes/no answer, and the use of any type of
mobile phone. Nearly all participants reported the
weekly text-message helped them engage in treatment,
whether by being reminded of their health, or having

timely access to clinician support. Many participants felt
cared for as a result of the weekly message. Some
participants found the message to be monotonous and
recommended varying or personalizing the outgoing
message.
Conclusion: The weekly SMS intervention was accepted
by a diverse group of participants who found it easy to
use, and appreciated the increased communication and
provision of support from clinicians.
EP-186-06 Real-time monitoring of drugresistant TB cases for better programme

management: e-Smarts in Andhra Pradesh,
India
J Jaju,1 S Achanta,1 J P Carel,1 M Parmar,1 A Sreenivas1
1
World Health Organization Country Office for India, New
Delhi, India. e-mail: achantas@rntcp.org
Background: Programmatic management of Drug Resistant TB in India faced large scale recording and reporting
challenges. There was no real time data exchange
between centralized Culture & Drug Susceptibility
Testing (C&DST) Labs, Drug Resistant TB (DR TB)
centres where treatment is initiated and Districts where
ambulatory treatment is continued. The aggregate
reporting system though electronic was not suitable for
efficient patient wise management or for data analysis. A
set of five treatment cards at various levels were
maintained for each patient to update treatment status
for the two years of MDR-TB treatment. Paper based
system of communication at multiple levels caused delay
in transmission of information, transcription errors and
involved tedious processes to update records and prepare
programme reports
Intervention: In 2012, an electronic web based surveillance system e-Smarts was customized in AP, linking
Labs, DR TB centres and districts for real time data
exchange. Patient wise information of drug resistant TB
suspects, test results, treatment initiation details, ambulatory DOT is entered electronically. Auto generated emails and messages of test results to programme staff, off
line data entry mode, automatic treatment card updation
and technical platform which is compatible with
National e-recording system, Nikshay, are some of its
unique features
Results: e-Smarts is being used by 6 C&DST Labs, 19
DR-TB centres and 24 Districts in erstwhile AP (Telengana and AP) since 2013. Currently, 45546 details of
samples have been entered at C&DST Labs,
32053(70%) test results declared, 3739 started on
treatment and 2286 (61%) treatment cards updated.
Programme staff received 50691 e-mails and 56859
messages.This tool is used for monitoring time to declare
results, time to treatment initiation and regular updating
of treatment cards. It is also used in identifying regional
level gaps in programme implementation. Time taken to
prepare routine programme quarterly reports has decreased from 3 days to a few minutes. Compliance in
reports generated by e-Smarts and by manual prepara-
tion has increased to 95 % in 4th quarter 2014 versus

2nd quarter 2014 when it was 70%.
Conclusions: Data quality and use of data for analysis
has improved. Programme is gradually transitioning
from a paper based system to a paperless electronic
system. Advantages are real-time data sharing at all
levels, reduction of paperwork and decreased errors due
to data transcription
S415
More patients have been enrolled for DR-TB care.

GxAlert strengthened surveillance of DR-TB, TB in
children and TB in the HIV infected, speeding response
and improving programmatic decision making for
enrollment and placement of DR-TB patient on treatment, GxAlert also Prevent cartridge stock-outs and
track usage for accurate ordering. GxAlert has received
adequate government commitment.
Conclusions and key recommendations: Networking all
the rapid GeneXpert diagnostic machines in Nigeria to
GeneXpert, the National TB and Leprosy control
program and private health systems are now able to
report rapid TB and HIV indicators automatically and
send targeted action messages (alerts) by SMS/text and/
or email to health system decision makers, deliver realtime disease surveillance data to the GxAlert database.
Theres ownership and sustainability from the government of Nigeria as the NTP is scaling up to all the
genexpert facilities nationwide.
EP-188-06 Patient incentives, mobile health and

electronic medical records for communitybased DR-TB treatment at Indus Hospital,
Pakistan
EP-187-06 Government commitment in GxAlert

mobile technology solution enhanced DR-TB
management in Nigeria
K Jimoh Agbaiyero,1,2,3 L Ekbladh,1,3 M Benezet,3
G Akang,2 J Takle,4 C Macek4 1Abt-HFG Project, Abuja,
2
National TB and Leprosy Control Program, Abuja, Nigeria;
3
Abt Associates, Bethesda, MD, 4SystemOne, Boston, MA,
USA. e-mail: kehindeagbaiyero@yahoo.com
Background and challenges to intervention: The World

Health Organization rank Nigeria among high DR-TB
burden countries in the world. The country relies on
GeneXpert machine to often diagnosis DR-TB. However,
testing for TB and reporting results is a lengthy process
partly due to a reliance on paper records, overburdened
labs, and slow data transit systems making it difficult for
the Nigeria Tuberculosis Program to account for patients
diagnosed for DR-TB, enroll them for care and management decisions are not timely or focused on priority
needs.
Intervention: GxAlert is configured on GeneXpert
systems by installing a modem from a telecom that gives
access to internet and encrypting the data sent to the
GxAlert database. The system sends GeneXpert MTB/
RIF diagnostic results in real time to a secure web-based
database, shortening a reporting period from months to
mere seconds and enabling better data quality and faster
recruitment of patients into appropriate care. Due to its
success in improving data system and patient management, the government decided to scale up to all facilities
in the country.
Results and Lesson Learnt: GxAlert eliminated the need
for human error in data entry, reduced the lag time, and
helped pinpoint patients that should be place on care.
H Hussain,1,2 A Malik,1,2 N Salahuddin,1 S Khowaja,2

S Butt,1 M Basir,1 U Khan,2 A Khan2 1Indus Hospital,
Karachi, 2Interactive research and Development, Karachi,
Pakistan. e-mail: hamidah.hussain@irdresearch.org

Indus Hospital in Karachi started treating DR-TB
patients in an ambulatory setting in 2008 with the
objective of providing free, comprehensive, community
based management for drug-resistant tuberculosis. In
2010 The Indus Hospital became a Sub-Recipient for GF
and was one of three initial PMDT sites in Pakistan.
Intervention or response: We conducted a cohort based
retrospective review of the DR-TB patient enrolled from
Q4 2008 till Q1 2015. All patients are provided with
consultations, diagnostics; monthly follow-up; psychological counseling; and quality assured second line drugs.
A treatment supporter provides daily DOT at patients
home. Patients are also provided with monthly enablers
in the form of food basket for family and travel cost. The
treatment supporter also receives the same food basket.
Nutritional supplements are provided to patients with
low BMI. Data are collected and maintained in open
source electronic recording and reporting system.
Results and lessons learnt: Of the 754 patients enrolled
with a median age of 30 years (IQR 22-40 years). 50%
were females; 37% of all patient and 23% of all female
were between 15-24 years of age. Majority 653 (87%)
had history of first or second line drugs where as 101
(13%) were new cases. 547 (72%) patients were
confirmed MDR while 119 (16%) patients were MTB
R resistant through Xpert; 12% were XDR, PDR and
mono DR patients. The DST for second line drugs
showed 11.5% of patients resistant to Ethionamide;
23.5% resistance to Fluoroquinolones (FQ); 2.3% were
resistant to any injectable and 3.7% of patients were
S416
resistant to both FQ and any of the injectable. Treatment

outcomes are available for 456 (60%) patients. The over
all treatment success rate for MDR patients was 75%
and 61% for XDR where as yearly cohort wise treatment
success in 2009 was 100 %, 78% in 2010, 74% in 2011
and 79% in 2012. Over all mortality rate was 12%,
default rate of 4% and 6% patients failed treatment.
Conclusions and key recommendations: Ambulatory
care model with patient enablers and efficient tracing
through electronic data management is highly successful
in high burden and low resource setting like Pakistan and
should be scaled up.
worker. The GPS system and audio recording is used only

upon mutual agreement.
Results and lessons learnt: Project managers, and all
interested parties can supervise social workers online
with data on current location, attendance, health
message delivery, and etc. User friendly setup and
functional ease have already attracted NTP and other
TB implementers interest in the country.
Conclusions and key recommendations: Such simplified
accessories may improve managerial issues and highly
increase the probability of properly delivered DOT at the
community level, thus increase the likelihood of treatment success for MDR-TB patients. It is an exemplary
Community and Private business involvement mixture
into Tajik TB control measures. Now Satellite technology
for MDR-TB treatment is being intensively discussed and
planned for use under the New Funding Mechanism of
the GFATM grant.
EP-189-06 Satellite technologies use for MDRTB treatment support

A Ismayilov,1 A Trusov,2 M Sianozova,3 F. Rahimov,1
S. Odinaeva,1 D. Ismatov4 1Project HOPE, Dushanbe,
Tajikistan; 2Project HOPE, Millwood, VA, USA; 3Project HOPE,
Yerevan, Armenia; 4Sarfaroz LLC, Dushanbe, Tajikistan. Fax:
(992) 227 3646. e-mail: aismayilov@projecthope.org
Background and challenges to implementation: Tajikistan is among the 27 high burden multidrug-resistant
tuberculosis (MDR-TB) countries in the world. It has one
of the highest estimated TB cases (all forms) 142 (range
70-239) per 100 000 population incident number in the
WHO European Region. According to the latest drug
resistance survey in 2011, the prevalence of MDR is 13%
among new and 54% among previously treated TB cases
(WHO report, 2013). Sustainable provision of DOT
during the long course of MDR-TB treatment in civilian
populations living in remote and impenetrable districts
remains a challenge.
Intervention or response: In order to pilot opportunities
to strengthen and guarantee DOT for MDR-TB patients,
two visiting social workers from different NGOs were
equipped with and trained to use small donated GPS
tracker devices. The trackers have been used to monitor
DOT of 10 patients in two different districts for three
months. A special geozone (R50m) was appointed for
each patient on the map, as well as a DOT medical unit
where the social worker is collecting Second Line Drugs.
The tracker sends data via GMS and reports to the server
on GPS positioning, which is easily reached by NGO
coordinators and Project HOPE managers. Reports
contain data on the amount of time spent and shows
travel on a map among geozones (Pic 1). The system
allows recording the dialog between patient and social
EP-190-06 Using mobile health to strengthen

community-based management for MDR-TB in
Myanmar
S Aung,1,2 A Innes,2 P Tun,1 N Naung,1KZ Aye,1KS Win3
FHI 360 Myanmar, Yangon, Myanmar; 2FHI 360 Asia Pacific
Regional Office, Bangkok, Thailand; 3Myanmar Medical
Association, Yangon, Myanmar. e-mail: ainnes@fhi360.org
1
Background: An estimated 9000 MDR-TB cases arise

each year in Myanmar. One key limitation for scaling up
MDR-TB treatment is the low capacity and availability
of directly observed therapy (DOT) providers. The
USAID Control and Prevention-Tuberculosis (CAP-TB)
project pilot-tested Myanmars first community DOT
providers for MDR-TB, filling a critical gap for patient
support. To ensure quality care by community DOT
provision, CAP-TB developed DOTsync, a mobile health
application to train and track community volunteers in
their daily work.
Intervention: DOTsync was created using Dimagis open
source CommCare software. The mobile app goes
through simple steps to guide the activity workflow in
the home, providing checklists to review DOT, health
education messages, and audio visual aids for patient
communication. Forty-two community DOT providers
were trained to use DOTsync in Yangon and Mandalay.
Utilization was tracked by monitoring the frequency and
timing of data upload to the cloud. Data are presented
for six months of implementation.
Results: We found that 76% (32/42) of the trained

community DOT providers consistently used CommCare
during their activities, which included direct observation
of MDR-TB therapy, health education, infection control
and side effect checklists, and referrals for treatment
complications or presumptive TB cases among contacts.
After launching DOTsync in select townships, the
number of people reached for health education increased
by 40% (from 324 to 452). At the start of DOTsync, 29
presumptive TB/MDR-TB cases had sputum tested, of
which 28% were smear positive. After DOTsync, 39 of
65 (60%) presumptive cases had sputum tested with a
smear positivity rate of 33%. Users and beneficiaries had
positive feedback from their experience with DOTsyncs
integration into their daily lives.
Conclusions: DOTsync was successfully used to strengthen community-based MDR-TB prevention and patient
support. Compared to pre-DOTsync, the identification
of presumptive TB/MDR-TB cases was more efficient,
with a higher sputum positivity rate. DOT provision was
monitored by tracking the daily data upload to the cloud,
enabling timely follow-up for missed doses and quality
control for community volunteers. The Myanmar NTP is
now using DOTsync in Yangon townships with the
highest MDR-TB prevalence. DOTsyncs longer term
impact on TB/MDR-TB case detection and MDR-TB
treatment success will be evaluated with this scale-up.
EP-191-06 Using mHealth techniques, onsite

coaching and incentives to motivate rural
pharmacies to screen and refer individuals with
TB symptoms in Viet Nam
H Ngo,1 Dat Tran,1 Tuan Nguyen,1 J Neukom,1 M Theuss1
1
Population Services International (PSI/Viet Nam), Hanoi, Viet
Nam. e-mail: nh.ytcc@gmail.com
Background and challenges to implementation: In Viet

Nam, while most investments in TB programming have
focused on public sector channels, a significant proportion of TB care is provided by pharmacies and private
clinics. Data collected in 2013 reflects 80% of individuals with TB symptoms first seek care from pharmacies
or private clinics. In this context, PSI implemented a
program designed to motivate rural pharmacies to screen
and refer individuals with TB symptoms.
Intervention or response: During the first 9 months of
project, PSI signed agreements outlining commitment
from 1,305 rural pharmacies in 3 provinces to screen and
refer customers with TB symptoms. Based on insights
about rural operators, PSI designed onsite coaching and
mHealth tools, print materials and non-monetary incentive packages to encourage and reward them. Face-toface, print and SMS messages consistently emphasized
how TB screening and referral practices could help
improve pharmacies as businesses committed to serving
their communities. A mobile screening app, print
screening guide and a referral card with list of TB care
sites, were designed to motivate pharmacies to comply
with national guidelines. Pharmacy referrals were 1 of 4
channels used to increase TB case detection.
S417
Results and lessons learnt: Between Jul 2014 and Mar

2015, the project screened 4382 individuals in rural
districts of 3 provinces. Among these, 3,882 were
successfully referred for diagnosis at the nearest referral
site and 692 cases of smear positive (SS) TB were
detected. Pharmacies generated 11.3% of the total,
clinics (81.1%), outreach (0.1%) and camps (7.5%).
Results through the pharmacy channel increased steadily
over time, as repeat coverage of signed rural pharmacy
operators increased. Monitoring data highlights the
deeply engrained patterns of TB customer care which
need further efforts to addressincluding sale of
antibiotics prior to screening or referral for TB diagnosisas well as the limitations of some mHealth
techniques in cases where individuals have limited text
messaging capacity.
Conclusions and key recommendations: Multiple points
of contact are required to change pharmacy operator
behaviors related to care for TB suspected customers
including repeat face-to-face coaching, simple prints and
tools, and non-monetary incentives valued by rural
pharmacies. SMS or call-back messaging may be more
effective than mobile screening apps for individuals with
limited text messaging capacity.
EP-192-06 Does the use of electronic health

records improve data capture of TB screening in
an HIV setting? The Maina Soko Military
Hospital case study
J Nikisi,1 G Muyunda,1 T Sajisa,2 K Asiedu1 1Jhpiego-An
affiliate of Johns Hopkins University, Lusaka, 2Ministry of
Defense, Lusaka, Zambia. Fax: (260) 211 253314. e-mail:
Joseph.Nikisi@jhpiego.org
Zambia has an HIV of 13.3% prevalence and faces a high

burden of TB. Approximately 65% of patients with TB
are co-infected with HIV. As a part of integrating TB and
HIV activities, almost 100% of TB patients undergo HIV
testing and counseling and this is well documented.
However, there has been a challenge in documenting TB
screening occurring in HIV settings. As a result, it is
difficult to ascertain the coverage of TB screening among
HIV patients. Zambia has implemented SmartCare, an
Electronic Health Record (EHR), since 2005. The system
is rolled out in 832 of 1800 health facilities. The EHR
covers most clinical areas including HIV care and
treatment and TB services; data is captured during each
patient interaction. Jhpiego supports Maina Soko Military Hospital (MSMH) one of the first sites with
SmartCare. The expectation was the introduction of
EHR would improve data capture and documentation of
TB screening in the HIV setting leading to improved
patient care. HIV care and treatment SmartCare data
from January 2012 to December 2014 was reviewed.
Data was analyzed using the number of patients
commencing ART as the proxy for patients in care and
treatment and compared to the total number of patients
screened for TB during the same period from the same
pool of patients A total of 1283 patients were commenced on ART but only 69 patients were documented
S418
in SmartCare as having received a TB screen (5.7 %).

However, a review of the physical register during the
same time period showed 437 patients were screened for
TB (34%). Implementation of SmartCare did not
improve data capture of TB screening among HIV
patients. In fact, paper documentation had a better data
recording. Challenges exist around staff attitude, acceptance of new technology systems and appreciation of the
importance of timely, accurate, complete and up to date
data for clinical and program decision-making. While
technology can support improved data collection and
quality, if not accompanied by provider orientation and
training, appropriate change in behavior, attitude and
practices the same documentation challenges remain. It is
recommended that the actual coverage of TB screening
among HIV patients be explored. Furthermore the EHR
system should be reviewed to require key data entry by
providers during patient interaction and prompt to
screen for TB. This should include appropriate provider
orientation and training in SmartCare with emphasis on
data quality and use for decision making.
12. Pneumonia and indoor air pollution

EP-194-06 High prevalence of byssinosis among
powerloom workers in Burhanpur, India
A Das,1 B Thapa,1 S Chadha,1 V Lal2 1International Union
Office, New Delhi, 2All India Institute of Medical Sciences,
Jodhpur, India. e-mail: adas@theunion.org
Background: In India, the Power-loom sector provides

employment to 5.74 million persons, accounting for 62
% of the countrys cloth production. The prevalence of
bysinosis is not known amongst the power-loom
workers. The objective of the study was to determine
the prevalence of byssinosis amongst the power-loom
workers and assess the relationship of duration of
exposure with other co-variates.
Design/Methods: This was a cross sectional study
conducted in Mominpura, a muslim-dominated, powerloom dependent administrative ward of Burhanpur in
Madhya Pradesh, India having 2818 adult population.
Semi-structured questionnaire was administered to gather data on patient demographics, duration of exposure to
cotton dust, smoking, and presence of symptoms of
byssinosis in 291 power-loom workers in 2014. Peak
Expiratory Flow Rate (PEFR) was measured and chest Xrays was taken. The outcome measures were, prevalence
of byssinosis, grades (Hunters classification), relationship of duration of exposure-2-10 years (short exposure)
and more than 10 years (prolong exposure) with other
co-variates (age, sex, smoking, grades of byssinosis,
reduction in PEFR, and chest X-ray abnormalities).
Results: The mean age of the participants was 40612.6
years (range 16-75 years). The male to female ratio was
3.5:1 (Male, 225; Female 65), and 77% were married.
The prevalence of byssinosis, based on Hunters classification, was 98% (n 284). Of those, 239 (84%) were
in Grade 0.5, 32 (11%) in Grade 1, 8 (3%) in Grade 2

and 5 (2%) in Grade 3. PEFR was reduced in 44% (n
124). Of those having byssinosis in whom PEFR was
reduced, 81 (29%) had PEFR reduced by more than
50%, and of these, 69 (82%) were in early stage of
byssinosis (Grade 0.5). Among 124 participants in whom
chest X-rays was done, 102 had byssinosis (Grade 0.5)
and only 19% (19) had abnormal chest X-rays. Of 284
who had byssinosis, 44 (15%) and 240 (85%) had short
and prolong exposure to cotton dust, respectively. Older
age group (.35 years), smokers and those who smoked
more than 10 sticks/day, grade 1 byssinosis, and more
than 30% reduction in PEFR had prolonged exposure (P
, 0.05).
Conclusion: Byssinosis was common among powerloom-workers in Burhanpur. Work place safety rules
and regulations need to be enforced in order to tackle this
problem in the study setting. This study could not be
generalized and further study in a nationally representative sample is required to determine the epidemiological
situation of byssinosis in the country.
EP-195-06 Microbiological characteristics and

predictive factors for mortality in pleural
infection: a single-center cohort study in South
Korea
YS Kwon,1 C.K. Park,1 H.J. Shin,1 YI Kim,1 SC Lim1
1
Chonnam National University Hospital, Gwangju, Republic
of Korea. Fax: (82) 622 258 578. e-mail:
yongsu3@gmail.com
Background: Identification and understanding of the

pathogens responsible for pleural infection is critical for
treatment with the proper antibiotics. This study sought
to determine microbiological characteristics of pleural
infection and to identify potential predictive factors
associated with mortality.
Design/Methods: In this retrospective study, we analyzed
patient information from 114 cases of parapneumonic
effusion with a total of 134 microorganisms that were
isolated using two culture systems (standard method and
pairs of aerobic and anaerobic blood culture).
Results: Of all isolated microorganisms, the most
frequently found pathogen was streptococci (35.8%),
followed by staphylococci (19.4%), anaerobes (15.7%),
and gram-negative bacteria (9.0%). Streptococci were
the main microorganisms found in standard culture
(48.1%) and community-acquired infections (58.7%)
and were susceptible to all antimicrobial agents in drug
sensitivity testing, whereas staphylococci were the most
frequently isolated pathogens found in blood culture
(26.8%) and hospital-acquired infections (33.8%), and
were primarily multidrug-resistant (69.2%). Patient
history of smoking, no use of intrapleural fibrinolytics,
surgery, and severity score (CURB-65 7 2) were
significantly associated with 30-day mortality in univariate analysis. In multivariate analysis, smoking (OR
4.051, 95%CI 1.21-13.61, P 0.024), CURB-65 7 2
(OR 3.42, 95%CI 1.15-10.14, P 0.027), and no use
of intrapleural fibrinolytics (OR 8.19, 95%CI 1.02
66.14, P 0.048) were significant predictive factors for

30-day mortality.
Conclusion: Common pathogens of pleural infection in
this study were similar to other studies, which consist of
streptococci, staphylococci, and anaerobes. Patient
history of smoking, CURB-65 7 2, and no use of
intrapleural fibrinolytics as a medical intervention could
be predictive factors for death in pleural infection.
EP-196-06 Study of the clinical characteristics

and putcomes of patients admitted with
exudative pleural effusion: improving clinical
decision making in resource-limited countries
D R Mishra1 1B P Koirala Institute of Health Sciences, Dharan,
Nepal. e-mail: deebyaraj@hotmail.com
Background: Pleural effusion complicates various diseases. In resource limited setting like Nepal, where
Tuberculosis is so prevalent, it is almost obligatory to rely
on clinical and only basic lab parameters to differentiate
causes of exudative pleural effusion and specially
Tuberculous from Non-tuberculous causes. In this
background, we set out to identify if specific clinical
characteristics and basic lab parameters would guide
differentiation of Tuberculous from other causes of
exudative pleural effusion and their effects on the
treatment outcome.
Design/Methods: A retrospective study was performed
on consecutive patients with exudative pleural effusion
admitted in our hospital from April 2013 till July 2014.
Complete medical charts were reviewed and only the
initial pleural fluid examination was recorded. Statistical
data was analysed using SPSS 17 and P ,0.05 considered
significant.
Results: Among 109 patients, 45(41.3%) of the cases
were Tuberculous. Among Non-tubercuolous causes,
there were 29(26.6%) with Parapneumonic effusions
and 23(21.1%) with Malignant Pleural effusions. 60%
of Non-tubercular effusions were Right sided. 82.4% of
the Parapneumonic effusions were small, whereas 39.1%
of the Malignant pleural effusions were large. Compared
to Tubercular pleural effusions, increased age, increased
duration of symptom, history of smoking and increased
pack years statistically favoured a diagnosis of Malignant
pleural effusion, whereas presence of fever, cough and
increased pleural ADA levels favoured Tubercular
pleural effusions. With regards to Parapneumonic
effusions, a shorter duration of symptom, smaller
effusions, higher pleural Neutrophils, lower pleural
lymphocyte neutrophil ratio and lower ADA favoured
the diagnosis as compared to Tubercular pleural
effusions.
Conclusion: There exists important clinical and pleural
biochemical differences between Tubercular and other
major causes of exudative pleural effusions, the appreciation of which aids in improved clinical decision
making with minimal resources and thus has great
economic implications in resource limited settings like
ours.
S419
EP-197-06 Effect of health education on

caregivers knowledge, attitude and practice of
home management of childhood acute
respiratory infections in Issele-Aza
S Olarewaju,1 T Idaboh,2 K Odeyemi2 1Ladoke Akintola
University Teaching Hospital, Ogbomoso, 2College of Health
Sciences, Lagos, Nigeria. e-mail: oolarewaju7@gmail.com
Background and challenges to implementation: Acute

Respiratory Infections (ARIs) among under-five year old
children when managed inappropriately by their caregivers results in preventable mortality and morbidity.
Treatment failure, unnecessary adverse effects of antibiotics used, waste of health-care and family resources, and
an increased emergence of bacterial resistance are some
of the negative consequences of ARI mismanagement.
The assessment of the Knowledge, Attitude and Practices
(KAP) of caregivers of under-fives on childhood ARIs is
important in enabling policy makers plan and provide
specific interventions to reduce mismanagement of ARIs.
This study assesses the effect of health education on
caregivers KAP regarding home management of ARIs in
Issele-Azagba community, Nigeria.
Intervention or response: A quasi-experimental study
carried out among 139 caregivers of children aged 0-59
months in Issele-Azagba community in Nigeria, using
139 caregivers of children also aged 0-59 months in
Ekwuoma community as controls. Multistage random
sampling procedure was used to select the respondents in
both groups. The instrument of the study was semistructured, pre-tested interviewer administered questionnaires adapted from UNICEF (IMCI tool). The intervention was health education based on the Health Belief
Model. Baseline and post intervention data entry and
analysis were carried out using SPSS version 16.
Results and lessons learnt: A total of 278 respondents
participated in the study. The mean age of the respondents was 33.3 6 10.0 years and most of them were
mothers (81.3%). At baseline, all the respondents in the
intervention group and majority (98.6 %) in the control
group had heard of ARIs but their comprehensive
knowledge on its causative agent, mode of prevention
and transmission was low. Likewise, 97.1% and 86.3%
in the intervention and control groups respectively had
negative attitude regarding appropriate management of
ARIs. Inappropriate use of antibiotics was practiced by
75.9% in the intervention group and 56.8% in the
control group. Other inappropriate practices including
rubbing menthol/balm on the nostrils, neck, chest and
sometimes dissolving the menthol/balm in water for the
under-fives to ingest as concoctions were observed in
both groups. Post-intervention, statistically significant
increases were observed from baseline in the proportion
of respondents with a better level of knowledge by 41%
(P , 0.05), positive attitudes by 77%, (P , 0.05) and
better level of practice by 30.9% (P , 0.05) in the
intervention group. There were no statistically significant
positive changes observed in the KAP of the respondents
in the control group
Conclusions and key recommendations: The health
education program was effective at improving the
S420
knowledge and practice, and modifying the attitudes of

caregivers regarding appropriate home management of
childhood ARIs in the intervention group. Strengthening
community health education interventions on appropriate management of ARIs especially targeted at young
mothers are recommended.
EP-198-06 Barriers in adoption of improved

stoves: inquiry among users, non-users and
program managers in Sindh and Punjab
provinces in Pakistan
1
T Jamali, A Shahid, Z Fatmi, A Khoso, M M Kadir

1
Aga Khan University, Karachi, Pakistan. e-mail:
tanzil.jamali@aku.edu
Background: Improved stoves are considered as the

mainstay intervention to reduce household air pollution
due to biomass use. Several small scales stove intervention initiatives have been undertaken in Pakistan.
However, no intervention has seen to sustain. This article
inquires about the barriers of adoption of improved
stoves in two programs in the provinces of Sindh and
Punjab, Pakistan.
Design/Methods: This qualitative study was conducted,
to evaluate two stove intervention programs: one in
Kasur district in Punjab province and the other in
Dhabeji in Sindh province, during May-August, 2014.
Two key-informants, in-depth interviews were carried
out with program managers and six focus group
discussions (3 each from improved and traditional stoves
users) were conducted using semi-structured questionnaire. The study explored the factors facilitating or
inhibiting the adoption of improved stoves.
Results: Of total 48 women participated in the study who
were main cook in the household, 18 were using
improved and 30 were using traditional stoves. Women
used improved stoves perceived that it decreases respiratory symptoms and illnesses, eye discomfort and
prevent blackening of utensils and kitchen. Women were
more satisfied with the traditional stoves, owing to the
decreased amount of fuel and time consumed. Repairing
or constructing a new stove was found easier with those
using traditional stoves. Improved stoves once damaged
or destroyed, women perceive it difficulty in repairing.
Also the chimney is costly and not easily available for
replacement.
Conclusion: Women perceive the health and social
benefits of improved stoves. However, the main barrier
in adoption for improved stoves were time for cooking
and fuel consumption. Fuel-efficient stoves need to be
designed to improve adoption along with strengthening
of the program credit and supply-chain strengthening for
provision of continuous training and material such as
chimney.
EP-199-06 Detecting pneumonia in resourcepoor settings: searching for new devices for
frontline health workers
K Baker,1 A Mucunguzi,2 L Matata,3 M Posada,4
T Memera,5 K Kallander1 1Malaria Consortium, London,
UK; 2Malaria Consortium, Kampala, Uganda; 3Malaria
Consortium, Juba, South Sudan, Republic of; 4Malaria
Consortium, Phnom Penh, Cambodia; 5Malaria Consortium,
Addis Ababa, Ethiopia. e-mail:
k.baker@malariaconsortium.org
Background: Globally, pneumonia remains the leading

infectious cause of death in children under five years.
Diagnosis can be challenging given the current available
tools for frontline health workers in resource-poor
settings. Community health workers (CHWs) and firstlevel health facility workers (FLHFWs) diagnose pneumonia primarily through counting the respiratory rate of
children who have cough or difficulty breathing.
However, counting respiratory rate is often challenging
and misclassification is common. Malaria Consortiums
Pneumonia Diagnostics project is identifying the most
accurate and acceptable respiratory rate timers and pulse
oximeters to support CHWs and FLHFWs in the
detection of pneumonia symptoms in children. The
research project is based in four low-income countries
Cambodia, Ethiopia, South Sudan and Uganda.
Design/Methods: The project combines qualitative and
quantitative methodologies in three phases: device
selection; accuracy evaluation; and field testing. The
device selection phase was based on the following stages:
landscape review; CHW focus group discussions; device
scoring; pile sorting with CHWs and key Ministry of
Health personnel to assess acceptability and scalability;
and laboratory testing for accuracy and robustness.
Results: The landscape review identified 190 respiratory
rate and pulse oximetry devices. Focus group discussions
with CHWs identified 20 product attributes that were
used to score and rank these devices. CHWs in all
countries expressed a felt need for new pneumonia
diagnostic devices. Issues with current devices included
suitability, usability and durability. CHWs most frequently highlighted the need for automation in their ideal
device. Pile sorting of the twelve highest scoring devices
(8 pulse oximeters and 4 respiratory rate counters), by
CHWs and government stakeholders, favoured the
fingertip pulse oximeters due to their ease of use. All
laboratory tested devices passed, except one fingertip
pulse oximeter, which failed to switch on after the tests
were completed. Based on the laboratory test results and
pile sorting exercise, 9 devices proceeded to the accuracy
evaluation phase.
Conclusion: Phase one of the research project indicates
that frontline health workers in resource poor settings
seek new devices to support the detection of pneumonia
in children. For these devices to be acceptable to CHWs
and FLHFWs, manufacturers need to ensure devices are
automated, accurate and easy to use.
EP-200-06 Diagnosis of chronic pulmonary

aspergillosis: conventional and newer tools
M Richardson1 1Mycology Reference Centre Manchester,
Manchester, UK. e-mail:
Malcolm.Richardson@manchester.ac.uk
Chronic pulmonary aspergillosis (CPA) is a slowly

progressive inflammatory destruction of lung tissue due
to Aspergillus infection. It is estimated that the global
burden of CPA is in excess of 3 million individuals. These
patients require special attention because of the possibility of developing subsequent invasive pulmonary
aspergillosis, given their poor health and worsening
immune state. Conventional diagnostic modalities include culture of respiratory secretions, serology (precipitin tests and indirect haemagglutination) and imaging.
The main radiographic features are chronic pulmonary
infiltrates, progressive cavitation, and subsequent aspergilloma formation. Chronic cavitary pulmonary aspergillosis (CCPA), which is synonymous with complex
aspergilloma, shows one or more pre-existing and / or
newly formed pulmonary cavities that may or may not
contain an aspergilloma, and cavity expansion and / or
increasing pericavitary infiltrates. Newer diagnostic
tools include various ELISA formats for detecting
Aspergillus-specific IgG, detection of Aspergillus biomarkers such as galactomannan in serum, sputum and
bronchoalveolar lavage fluid. Other methods that are
being developed include biomarker detection in exhaled
breath condensate, and monitoring of exhaled volatile
organic compounds (extrolites) specific for Aspergillus.
The performance of the Aspergillus Western blot (AspWB) IgG kit (LDBio Diagnostics, Lyon, France), a
recently commercialized immunoblot assay for the
diagnosis of various clinical presentations of chronic
aspergillosis, has recently been reported (Oliva et al., J
Clin Microbiol 2015; 53: 248-254). Serum samples (n
267) were derived from patients with Aspergillus disease,
including 89 cases of chronic pulmonary aspergillosis.
The Asp-WB kit performed well for the diagnosis of CPA
in nonimmunocompromised patients, with an enhanced
standardization and a higher sensitivity than precipitin
tests. In summary, the diagnostic criteria for CPA also
need to exclude infectious (TB, non-TB mycobacterial
infection, histoplasmosis and coccidiomycosis) or noninfectious (lung cancer, rheumatoid arthritis and sarcoidosis) pulmonary disease that can mimic or promote the
occurrence of CPA.
S421
13. Case finding and contact tracing

EP-201-06 Effectiveness of active tuberculosis
screening program for Aboriginal students
living in urban area, 2011-2012
Y-T Liao,1 C B Hsu,1 C-H Yang,1 C-H Chen1 1Centers for
Disease Control, Taipei, Taiwan. Fax: (886) 233 936 149.
e-mail: yuntsan@cdc.gov.tw

Taiwan, people residing in mountainous areas have
much higher incidence of tuberculosis (TB) than general
population. Chest X-ray (CXR) screening program had
been conducted by Taiwan CDC for decades in
mountainous area. Yet most young students originated
in these areas might leave their hometown and study in
urban areas, and became a substantial gap for CXR
screening program. Therefore, CDC coordinates with
educational sector to implement new active screening
strategy aiming at this group.
Intervention or response: The program used free CXR
screening voucher and X-ray patrol truck to provide
service to aboriginal students. In 2011, we first focused
on senior high school and technical college students, and
further expanded to college and university students in
2012.
Results and lessons learnt: There were 2969 and 4741
students received the X-ray screening in 2011 and 2012,
and the participating rates among eligible population
was 34% and 57% respectively. TB detection rate were
134.7/ 100 000 and 105.5/ 100 000 in 2011 and 2012,
which were 6.6 and 6.1 times higher than general
population of the same age in Taiwan. Especially, the
detection rate was highest among senior high school
students participating screening in 2012 (179.2/ 100
000).
Conclusions and key recommendations: Through coordinating with educational sector, CDC provides screening service for 7710 aboriginal students in urban areas
during 2011-2012. We found that, although not living in
mountainous areas, the target population in this program
still provided a TB detection rate as high as people
staying in mountainous areas. According to this, we
modified our policy since 2013 to include this people as
one of target populations in active case finding (ACF)
program performed by local health authorities.
EP-202-06 Contribution of chemists to TB

control in semi-urban areas of West Bengal,
India
B Sengupta,1 A Chakraborty,1 S Sarkar,2 S Haldar3 1Care
India, Kolkata, 2Revised National TB Control Programme,
Kolkata, 3Department of Health and Family Welfare, Kolkata,
India. e-mail: bsengupta@careindia.org
Background and challenges to implementation: Involvement of chemists in TB Control programme in West

Bengal is a pilot initiative by CARE India and supported
by Lilly MDR-TB partnership. The implementation
started during October 2012 in consultation with Bengal
S422
Chemists & Druggist Association (BCDA) and RNTCP

in two districts.
Intervention or response: Based on the lists shared by
BCDA, CARE Field Officer (FO) visited each and every
chemist shop to understand their willingness to participate in the program. The training module for the
community pharmacists which is acknowledged by
CTD was translated into local language and used for
the training. Referral slips and necessary formats were
developed as per need. Till June 2014 a total of 278
chemists have been trained using the module. Project
monitoring data shows that out of these 278 trained
chemists 70 are regularly participating in the program.
Results and lessons learnt: 96% of the chemists have the
correct knowledge on the commonest form of TB and are
aware of how disease spreads. 83 % of the chemists were
able to answer correctly what they mean by DOTS. 63%
chemists have correct knowledge on duration of treatment of newly diagnosed TB patient while the rest gave
incorrect answer. 100% chemist are remains that TB is
curable if the drug taken in regular basis and follow the
regimen. 25% chemists have referred patients suspected
of TB during the last one month. 4% of them have
provided sputum containers 28% of them have used
RNTCP referral slip during referral and 47 % mentioned
that they have referred verbally. 25% chemists got
feedback from the RNTCP officials on the result of
sputum test. 53% chemists shared the information on TB
patients who are taking medicine from private sector.
48% chemists have recorded the information of private
TB patients regularly while 5% of them have kept the
records inconsistently. 20% chemists have stopped to
stock anti TB drug in their shop.
Conclusions and key recommendations: Majority of the
chemists have very good understanding on TB and the
mode of disease transmission. Use of referral slip for
documentation at the DMC and feedback from the
system is a challenge. Sharing information on TB patients
on treatment from private sector is increasing as 53% of
them are sharing the information with RNTCP. Scaling
up of the intervention through other CSOs to improve
the notification of TB cases from private sector is
recommended.
grammatic settings with objective to assess feasibility and

effectiveness of this strategy.
Intervention or response: Out of 27 jails in 21 districts in
Gujarat, cross sectional study was undertaken in 23 jails
in 2014. Team of 1 doctor and 3 paramedical workers
with the help of local jail staff screened all prisoners and
presumptive TB cases were subjected to sputum smear
microscopy and X-ray. The results were entered and
analyzed in MS Excel.
Results and lessons learnt: Total 92% (10 306) prisoners
were screened. 601 (6%) presumptive TB cases were
identified. Of them, 559 (93%) of presumptive TB cases
were subjected to sputum smear microscopy. 6% (36)
cases were screened with X-ray chest. 19(3%) smear
positive TB cases were diagnosed. 1(3%) probable TB
case was identified. All diagnosed patients were put on
treatment.
Conclusions and key recommendations: It is feasible to
organize active case finding session in prisons with
formal communication and established coordination
between Health officials and jail authorities. Sizable i.e.
601 (6%) of the prisoners were identified as presumptive
TB cases. It indicates that case finding efforts may get an
impetus if such sessions are organized in high risk settings
like Prison. Only 6% of the presumptive TB cases
underwent X-ray examination because prisoners who are
presumptive TB cases had to be shifted to facility where
X-ray machine is available. If portable X-ray machine
are made available for such activities it can be made more
effective. 3% (20) TB cases were diagnosed which
indicates that organizing such activities may be effective
in improving case notification under programme. Such
initiatives are important for reaching out missing cases.
Moreover newer molecular diagnostic techniques like
Xpert MTB/RIF may be utilized for such activities which
may further improve yield of these case finding activities.
Programme may undertake such initiatives systematically at periodic intervals in all prisons to ensure early and
prompt diagnosis of TB and prevent further infection in
these congested settings.
EP-203-06 Study of active case finding for

tuberculosis in prisons in Gujarat, India
S Pattanaik,1 A Dutta1 1Asian Institute of Public Health,

Bhubaneswar, India. e-mail: dr.sarthakpattanaik@gmail.com
P Dave,1 S Bharaswadkar,2 A Amar Shah,2 B Vadera,2

H Thanki,2 K Rade,2 S Khaparde,3 A Sreenivas2
1
Commissionerate of Health, Gandhinagar, 2World Health
Organization, Country Office for India, New Delhi, 3Central
TB Division, Ministry of Health & Family Welfare, New Delhi,
India. e-mail: bharaswadkars@tbcindia.nic.in
Background: As per WHO estimates, 3 million TB cases

remain undetected annually; who fail to receive standardized TB care, suffer severely and continue to spread
infection in the community. Mostly, missed cases come
from vulnerable migrant populations of slum-dwellers
and industrial labours, who often fail to access local
National TB Programme (NTP) services, due to unfamiliarity with it or for prohibitive opportunity costs of
accessing these often patient-unfriendly services. Consequently, the presumptive TB patients (PTP) from these
vulnerable communities often present to neighbourhood non-formal healthcare providers (NFHP) for
symptomatic relief, who fail to channelize them to
Background and challenges to implementation: Overcrowding in prisons poses a risk for increased TB
transmission. Indias National Strategic Plan articulates
active case finding in vulnerable and high risk groups like
Prisoners. However no clear guidelines have been issued
to implement this. We conducted systematic screening
and active case finding amongst prisoners under pro-
EP-204-06 TB Reach project in Odisha, India:

intervention to find missed TB cases from
migrant slum dwellers and industrial laborers
appropriate diagnosis and treatment. Our project funded

by TB REACH was planned to detect such missed
cases. We aim to assess the early effects of the
intervention, so that positive learning can be scaled-up.
Intervention: The project operates in urban slums of
Bhubaneswar, the rapidly-growing capital city and the
mining and ferro-alloy-related industrial labour settlements in Jajpur, in Odisha, India. The project targets
0.55 million migrant slum dwellers and industrial
labours, nested within 1.55 million evaluation population, spread across four basic NTP management units.
The project uses a network of sensitized and incentivized
NFHPs for identifying PTPs early; collection and
transportation of sputum by community health workers;
and examining specimens using two modern Xpert MTB/
RIF machines; one installed at public hospital, Bhubaneswar, other at an industry-owned clinic at Jajpur.
Bacve cases are referred promptly to NTP for treatment.
The present assessment study has collected sputum
examination and new smear/Bacve positive case notification data from local NTP for October 2014 to March
2015; and also for corresponding quarters of four
previous years.
Results: The project, starting in November 2014, has
tested 824 specimens using Xpert MTB/RIF of which 173
were Bac positive (4 were Rif resisant). Despite working
for 5 months, in comparison to 280 new smear/Bacve
cases diagnosed between October 2013 and March 2014;
the project helped to increase case detection to 388
between October 2014 and March 2015, in the
evaluation population; an additional contribution of
98 cases.
Conclusions: Projects targeting vulnerable populations
through their neighbourhood providers using convenient patient-friendly diagnostic strategies can detect
missed cases substantially.
S423
EP-205-06 Seeking the lost: Intensified

tuberculosis case finding amongst vulnerable
communities in southern India
C K K Gali,1 T Thomas,1 S Gadala,1 R Ananthakrishnan,2
A Jacob,3 M Das,4 V Sameer,1 P Isaakidis4 1Catholic Health
Association of India, Secunderabad, 2Resource Group for
Education and Advocacy for Community Health, Chennai,
3
South-East Asia Office, New Delhi, 4Medecins

Sans
`
Frontieres,
Mumbai, India. e-mail: gkkr109@gmail.com
Background: Indias Revised National Tuberculosis

Control Programme (RNTCP) primarily relies on passive
case finding to detect new cases of TB. The new sputum
case detection rate in Karnataka, a state in southern India
was 58% in 2013. One Designated Microscopic Centers
(DMC) exists for every 100 000 population in the plains,
and in hilly areas, the state government establishes 1
DMC for a population of 50 000. About 24% of
Karnatakas population is specifically classified by the
Census of India as belonging to lower (scheduled) castes;
7%, to the tribal groups. Low socio-economic status and
marginalization is correlated with TB, requiring interventions to increase TB case detection rates.
Intervention: Project Axshya (through Round 9 of the
Global Fund to Fight AIDS, Tuberculosis and Malaria)
helps expand the reach and effectiveness of RNTCP,
engaging communities to improve TB services for
vulnerable populations. An Intensified Case Finding
(ICF) intervention was initiated under Project Axshya
in 16 districts, covering 56% of Karnatakas population
(34 million), including 54% of its lower caste and 45%
of its tribal population. This study focuses on ICF
activities in two districts (Kolar and Bidar) wherein
areas mapped to have a predominant vulnerable population (20 DMCs, intervention area) received the ICF
intervention whereas the remaining (11 DMCs, nonintervention area) did not. To evaluate the effect of ICF
on TB case detection, we reviewed the routine programme data and compared the change in a) number of
presumptive TB patients examined b) smear-positive
patients detected and treated in the intervention and nonintervention areas before (July-December 2012) and
after (July-December 2013) implementing the intervention.
Results: The number of smear-positive cases detected
increased by 8.8% relative to the pre-intervention period
(July-December 2012) in intervention communities as
compared to an 8.6% decrease in communities without
the ICF intervention. This study showed that an
innovative, community-based, education-cum-case finding intervention increased the number of TB cases
detected and treated and brought TB diagnostic and
treatment services closer to vulnerable communities,
boosting awareness about RNTCP services.
Conclusions: In conclusion, ICF using family counseling
mechanisms and community-based TB screening led to a
modest increase in the number of smear positive TB cases
diagnosed among vulnerable communities in India.
S424
EP-206-06 Model-based evaluation of contact

investigations for tuberculosis in low-burden
settings
EP-207-06 Study of symptom screening for

active case detection in the community using
prevalence survey data
G Guzzetta,1,2 M Ajelli,1 Z Yang,3 L Mukasa,4,5 N Patil,4,5

J Bates,4,5 D Kirschner,6 S Merler1 1Fondazione Bruno
Kessler, Trento, 2TrentoRise, Trento, Italy; 3University of
Michigan, Ann Arbor, MI, 4Arkansas Department of Health,
Little Rock, AR, 5University of Arkansas for Medical Sciences,
Little Rock, AR, 6University of Michigan, Ann Arbor, MI, USA.
e-mail: guzzetta@fbk.eu
K Okada,1 N Yamada,1 S Saint,2 T E Mao2 1Research

Institute of Tuberculosis / Japan Anti-tuberculosis association
(RIT/JATA), Tokyo, Japan; 2National Center for Tuberculosis
and Leprosy Control (CENAT), Phnom Penh, Cambodia.
e-mail: okadak@jata.or.jp
Background: Active screening of contacts of tuberculosis

(TB) patients is one of the main TB control strategies in
low-burden countries. To date, evidence from epidemiological studies is insufficient to prove and quantify the
effectiveness of such programs in reducing incidence in
the general community.
Design/Methods: We propose an individual based model
of TB transmission dynamics occurring in a timeevolving network of contacts, realistically reproducing
socio-demographic characteristics of the resident population of Arkansas, USA in the absence of migration. The
model incorporates TB control interventions implemented by the local Department of Health, which include
contact tracing activities. TB incidence over time and by
age from the period 2001-2011 are used to calibrate the
model, and its predictions are validated against several
independent data sets from Arkansas and the US.
Results: The model predicts that contact tracing prevented about 30% of TB cases and deaths that would
have occurred in the U.S.-born Arkansas population in
2001-2014 if only passive diagnosis were in place.
Inclusion of contacts of sputum smear-negative patients
in the program accounted for about half of avoided cases.
According to the model, only about 5% of all TB cases
from endogenous reactivation occurring over the next 60
years will be prevented by contact tracing; of these, about
half are prevented by successful treatment of latent
infections, and the other half by avoided transmission
due to the early diagnosis afforded of actively found
cases. The model also predicts that improving the
performance of contact investigations to meet targets
set for 2015 by the U.S. Centers for Disease Control has
large potential for further reducing the TB burden.
Conclusion: This work demonstrates the significant
impact of contact tracing to reduce TB burden in high
resource settings and identifies critical aspects (e.g.
extensive inclusion of contacts of smear-negative patients) that may be leveraged to improve program
effectiveness. We also show that the limiting factor for
TB control in the next decades will be endogenous
reactivation due to prevalent latent infections, especially
from older cohorts who experienced epochs of higher TB
transmission. Therefore, in absence of additional interventions, the decline of TB in low burden countries will
be driven mainly by demographic forces.
Background: End TB Strategy aims at a 10% reduction

rate per year between 2016 and 2025 and more efforts
need to be made to accelerate a current decline rate of
1.5%/ year in incidence by optimizing current tools
including active case finding (ACF) especially in high risk
groups of TB. However, one of the drawbacks in ACF is
more prohibitive and resource-consuming than passive
case finding (PCF). Although many countries adopt
prolonged cough (lasting 2-3 weeks) as presumptive TB,
only 30% of bacteriologically positive TB in the
community meet the definition.
Objective: To examine what symptom screening method
is effective in terms of sensitivity and specificity by using
national prevalence survey data obtained from Cambodia, 2011.
Methods: First, we identified TB-related symptoms by
examining the association of seven TB-related symptoms
(sputum, haemoptysis, chest pain, weight loss, fatigue,
fever and night sweat) with being smear-positive or
bacteriologically positive TB by logistic regression
models. Next, we compared sensitivities and specificities
among symptom screening methods which include the
duration of cough (any duration, 7 days or longer, and 14
days or longer) and the number of symptoms selected by
the multivariate logistic regression model, by using
receiver operating characteristic (ROC) curves.
Results: All the seven symptoms were significantly

associated with being TB case by univariate analysis
with P , 0.05. By multivariate analysis, four symptoms
(haemoptysis (OR 1.87 (95%CI: 1.09-3.20), weight
loss (OR 1.42 (95%CI: 1.08-1.85), fatigue (OR 1.29
(95%CI: 0.96-1.72), and night sweat (OR 1.34

(95%CI: 1.02-1.77)) were selected to examine the
sensitivities and the specificities. As a result of comparison in ROC curve, 7-day or longer cough only, or 7-day
or longer cough with one of the four symptoms was
suboptimum for symptom screening to detect bacteriologically positive TB.
Conclusion: The suboptimum symptom screening with
17% of the community people screen-positive provides
50% sensitivity and 84% specificity for bacteriologically
positive TB. For those with screening positive, other
tests, for example, chest radiography examination
followed by Xpert test can be proposed. Further studies
including operational researches are required to examine
the efficiency and the cost effectiveness.
EP-208-06 Prevalence of positive tuberculosis

skin test among health care workers in San
Juan de Lurigancho district, Lima, Peru
J Sedamano,1 C Ugarte-Gil,1 C Zamudio,1 G Soto,2
C Munayco,3 C Seas1,4 1Instituto de Medicina Tropical
General de
Alexander von Humboldt, Lima, 2Direccion
Epidemiologa - Ministerio de Salud, Lima, Peru; 3Uniformed
Services University of Health Sciences, Bethesda, MD, USA;
4
Hospital Nacional Cayetano Heredia, Lima, Peru. e-mail:
juana.sedamano@upch.pe
Background: Identification and treatment of individuals

with Latent Tuberculosis Infection (LTBI) is one of the
biggest challenges for Tuberculosis (TB) control. Health
Care Workers (HCWs) are under high risk of M.
tuberculosis transmission, which varies depending on
the type of facility where HCW work, community TB
prevalence, area of the facility in which HCW work and
effectiveness of preventive measures. In 2012 Peru had an
incidence of 119 TB cases per 100 000 habitants, that
concentrates in urban areas, 54% of TB cases, 82% of
MDR-TB cases, and 89% of XDR-TB cases in Peru were
reported in Lima and Callao. San Juan de Lurigancho
(SJL) district on its own represents 17.1% of cases of TB
in Peru, and presents an incidence above the national
estimate.
Objectives: To estimate the prevalence of positive
Tuberculosis Skin Test (TST) in HCW in SJL and to
investigate factors associated with a positive value of
TST.
Methods: An observational cross-sectional study was
conducted between September 2014 and March 2015
among HCWs working at least 3 months at their primary
care health centers. TST was performed and the result
was read after 48-72 hours. In addition, all participants
completed a questionnaire regarding demographics and
use of TB prevention strategies. We considered 710 mm
of induration as a positive TST.
Results: 248 HCW were enrolled. TST was administered
to 198 participants, of which 9 did not return for the
reading appointment. Of the 50 workers, who did not
undergo the TST, 26 had a previous positive TST, 14 had
a history of active TB and 10 refused the application of
the TST. Among the 248 HCWs, 81.1% were women;
the mean age was 42 years (SD: 11.2). The mean period
S425
of work was 153.2 months (SD: 127.3). The prevalence

of a positive test in 198 participants who received the
TST was 56.6% (95%CI: 49.5%-63.7%). When considering those who had a previous positive TST result and
those with a history of active TB (n 238), the prevalence
was 64.1% (95%CI: 58.0%-74.4%). TST positive
HCWs have longer mean periods of work compared
with TST negative (116.2 months vs. 72.5 months; P ,
0.05). Only being 40 years old or older (adjusted PR: 1.4
95%CI: 1.1-1.7) remained significantly associated with
an increased risk of a positive TST after multivariate
analysis.
Conclusions: The prevalence of positive TST is high in
HCWs in SJL. Our results indicate the need for greater
implementation of preventive measures to avoid TB
acquisition among HCW.
EP-209-06 Contribution of the private sector to

TB treatment outcomes: experiences from
Kampala city, Uganda
D Muhwezi,1 D Lukoye,1 M Ruhweza,1 A Etwom,1
D Okello,2 P G Suarez,1 E Birabwa,3 D Sama1 1TRACK TB
Project, Management Sciences for Health, Kampala,
2
Kampala Capital City Authority, Kampala, 3United States
Agency for International Development (USAID), Kampala,
Uganda. e-mail: dmuhwezi@msh.org
Background and challenges to implementation: Evidence

suggests that failure to engage all health providers in
identification of presumed TB cases hampers case
detection, delays diagnosis, leads to inappropriate or
incomplete treatment leading to increasing drug resistance. Public-Private Mix (PPM) in various countries has
demonstrated improved Tuberculosis case detection and
treatment results. By January 2013, Kampala city had 49
Tuberculosis (TB) diagnostic and treatment Units (DTUs)
which were very few in relation to the population served
limiting access to services. Our aim is to describe the
PPM approach that was adopted in Kampala city to
engage the private sector and the results achieved so far.
Intervention or response: KCCA with support from the
USAID funded Track TB project engaged private healthcare facilities in Kampala that did not previously provide
TB care services. These were selected based on the disease
burden as per patient records in the unit registers for
2013. Key medical staffs from these facilities were then
trained in TB case detection, diagnosis, treatment,
patient follow up and reporting using standard NTLP
tools. The Union/TB Reach project also supported
enrollment of some private facilities to provide services.
Facilities that provided TB care at the time received
mentorship to improve their quality of TB service
delivery. The facilities then initiated the process of
systematic screening of patients at entry points of each
facility especially the outpatient department using the
intensified case finding form. Facilities with laboratory
services went on to analyze sputum while facilities
without sufficient laboratory capacity referred the
presumed TB cases to the nearest established TB
diagnostic units.
S426
Results and lessons learnt: Overall, the number of DTUs

has increased from 49 to 72 by engaging 23 private
health facilities while the number of health workers
trained in TB diagnosis and treatment has increased from
147 to 216 over the same period. TSR improved from
63.1% in the fourth quarter of 2013 to 75.7% in the
fourth quarter of 2014. The loss to follow up has reduced
from 30% to fewer than 5% over the same time period.
Conclusions and key recommendations: The implementation of the PPM for TB control in Kampala improved
case holding as reflected by the improved TSR and
reduced patient losses to follow up. More work is
required to overcome the barriers to implementation and
to make gains in case notification using PPM.
TSR by facility
type (%)
Time Period
Oct - Dec 13
Jan - Mar 14
Apr - Jun 14
Jul - Sep 14
Oct - Dec 14
Average TSR
Lost to follow
up by facility
type (%)
GovernGovernment
NGO Private
ment
NGO Private
64.9
68.1
75.2
80.5
81.2
73.98
74.8
72.8
81.7
83.8
79.6
78.54
63.1
63.7
87.5
85.3
75.7
75.06
30
20.20
20.40
9.40
9.70
7.20
15
30
10.10 2.80
6.30 27.60
9.40 5.60
1.60 6.90
5.20 3
14. Sniffing out TB: from nose to

immunochemistry
also used for whole genome sequencing from aMTB

culture.
Results: TB DNA extraction with PTMAX resulted in
significantly improved detection of mid- and lowpositive samples. qPCR detection from low-positive
samples was 25% with the SOC compared to 87% with
PTMAX. Detection from mid-positive samples was
70% with the SOC versus 100% with PTMAX.
Detection of antibiotic resistance markers was much
improved by the use of PTMAX. PTMAX enabled
100% detection of the rpoB and inhA markers, while the
SOC only achieved 50% and 90% detection respectively.
The SOC method required an average of 20 days culture
growth whereas PTMAX enabled same-day results
without culturing. The PTMAX and SOC extraction
methods both demonstrated 100% detection of aMTB
by the GenoType MTBC line probe assay and the RD4
qPCR assay. The line probe bands obtained from
PTMAX were of equal or greater intensity than those
obtained from the SOC method in 93% of the samples.
Similarly, in the RD4 qPCR assay, PTMAX Ct values
were equivalent or better than SOC Ct values in 87% of
the samples. Whole genome sequencing from a pure
culture of aMTB yielded an assembled genome of highquality reads, which was over 99% match to the
reference M. tuberculosis genome.
Conclusion: prepITMAX significantly improves the
performance of established molecular methods for TB
diagnosis and can shorten the time required to deliver a
clinically relevant result back to the patient.
EP-210-06
a chemical lysis
method to optimize the performance of
molecular TB assays
EP-211-06 OMNIgeneW SPUTUM: a novel

transport chemistry that liquefies and
decontaminates sputum while maintaining
viability of Mycobacterium tuberculosis
J Niles,1 B Ray,1 O De Bruin,1 G Robideau,1 C Kelly-Cirino1

1
DNA Genotek, Ottawa, ON, Canada. e-mail:
cassandra.kelly-cirino@dnagenotek.com
B Ray,1 O De Bruin,1 J Niles,1 G Robideau,1 C Kelly-Cirino1

DNA Genotek, Ottawa, ON, Canada. e-mail:
cassandra.kelly-cirino@dnagenotek.com
Background: Molecular technologies have great potential to increase the sensitivity and accuracy of tuberculosis diagnosis, but their performance hinges on the
quality and quantity of TB DNA obtained from primary
samples. prepITMAX (PTMAX), a simple chemical
lysis method, improves the performance of molecular
assays by increasing the recovery of high-quality TB
DNA from sputum samples and simplifies laboratory
workflows by eliminating the need for bead beating and
complex extraction methods.
Design/Methods: TB-positive sputum, or TB-negative
sputum spiked with attenuated Mycobacterium tuberculosis (aMTB) were processed with NaOH/NALC, or
OMNIgeneQSPUTUM reagent. DNA from NaOH/
NALC treated sediments was extracted by bead-beating
whereas DNA from OMNIgeneQSPUTUM treated
sediments was extracted with PTMAX. Performance
was compared in real-time qPCR targeting the TB RD4
region, pyrosequencing for rifampicin and isoniazid
antibiotic resistance markers, and the Hain Lifesciences
GenoType MTBC line probe assay. PTMAX DNA was
Background: Putrefaction of sputum during transport

and loss of viable Mycobacterium tuberculosis by
NaOH-NALC standard of care (SOC) method are
factors contributing to delay in diagnosis and treatment
of TB. A novel transport chemistry, OMNIgeneSPUTUM (OMSPD), decontaminates sputum while maintaining viable MTB for testing by smear, culture and
molecular diagnostic methods. The long shelf-life of
OMSPD eliminates the need for daily NaOH-NALC
preparation.
Design/Methods: Functional assessment of OMSPD
was performed for (i) long-term sputum transport from
remote sites in Nepal (ii) reduction of culture contamination in Lowenstein-Jenson slants (LJ) and Mycobacterium Growth Indicator Tubes (MGIT) in two countries
and (iii) compatibility with smear microscopy, Cepheid
GeneXpert MTB/RIF, pyrosequencing and line probe
assay (LPA).
Results: In Nepal, M. tuberculosis was detected in 100%
of sputum samples held in OMSPD, compared to 81%
of SOC samples after 2-6 days in transport. In addition,
prepITQMAX:
OMSPD decreased average time-to-positive (TTP) by 5

days on LJ. OMSPD transported sputum showed no
culture contamination compared to 13% for raw
transported sputum. Samples transported in OMSPD
allowed immediate processing upon arrival, while
samples transported according to SOC required
NaOH-NALC processing by the laboratory. Furthermore, OMSPD transported sputum was compatible
with smear microscopy. In a second study, sputum was
treated with either OMSPD or NaOH-NALC for 15
min. Samples treated with OMSPD showed concordance with the SOC in TTP by LJ and MGIT. OMSPD
treated sputum was found to be compatible with
pyrosequencing for antibiotic resistance profiles, LPA,
and GeneXpert MTB/RIF. After treatment with
OMSPD, MTB was detected in 88-93% of samples
using standard GeneXpert protocols, comparable to the
M. tuberculosis detection rate of 90-93% of samples
using SOC.
Conclusion: Our results show that OMSPD enables
long-term transport of sputum without the need for cold
chain stabilization, while maintaining viable M. tuberculosis for smear, culture, and molecular tests. Field use
of OMSPD obviates the need for processing with
NaOH-NALC in the laboratory. OMSPD eliminates
contamination and decreases time required to call TB
samples positive, which will enable quicker initiation of
treatment regimens and improvements in TB diagnostics
to assist high-burden countries to meet the WHO End TB
Strategy goals by 2035.
S427
without using sampling bags, prior to any treatment.

TB was diagnosed by sputum smear microscopy as well
as by culture and clinical follow-up. After collecting the
data in the field, data were loaded on a laptop and later
transferred by internet for statistical analysis in the
Netherlands. After data compression using a PARAFAC/
TUCKER3 algorithm, an artificial neural network was
used to distinguish between sick and healthy individuals.
v2 and Fishers exact test were used to analyse the
association between different parameters and breath test
outcome. Ethics approval from the Gadjah Mada
University Ethics Committee was obtained, and participants gave informed consent.
Results: All (87) subjects completed the study protocol
without significant difficulty. Median age was 45; range:
20-70 years; and median BMI was 20; range: 14.7-32.9
kg/m2 . The eNose device had sensitivity of 0.92
( 9 5 %C I0 . 7 4 - 0 . 9 9 ) , a n d s p e c i f i c i t y o f 0 . 8 8
(95%CI0.76-0.95), positive likelihood ratio of 7.36
(CI: 3.7-14.9), and negative likelihood ratio of 0.09 (CI:
0.02-0.35). Comorbidities, use of antibiotics not active
against TB, prior food intake, and smoking did not have
significant association with the breath test outcome (P .
0.05).
Conclusion: In this development phase, the eNose had
high sensitivity, specificity, and likelihood ratios. It is also
an easy-to-use device. It needs to be further developed in
a validation study that is currently in progress.
EP-212-06 A new tuberculosis screening tool: a

hand-held electronic nose. A cross-sectional
field study in Indonesia
A M Saktiawati,1,2 Y Stienstra,2 Y W Subronto,1
S Sumardi,1 M Bruins,3 J Gerritsen,3 O Akkerman,2
T Van Der Werf2 1Faculty of Medicine, Gadjah Mada
University, Yogyakarta, Indonesia; 2University Medical Center
Groningen, Groningen, 4ENose B.V. (The eNose Company),
Zutphen, Netherlands. e-mail: a.morita_iswari@yahoo.com
Background: Early diagnosis of tuberculosis (TB) is

challenging in low-resourced, highly burdened settings.
An accurate, easy-to-use, point-of-care diagnostic tool is
needed. An earlier release of a neuronal network-based
multi-detector system showed promise but it needed gas
collection in a bag (Bruins M, et al. Tuberculosis (Edinb)
2013; 93: 232238). A next-generation, mass produced
robust electronic nose (eNose) may detect TB with
adequate sensitivity and specificity by analyzing exhaled
breath.
Design/Methods: This study was carried out in three
governmental chest clinics and Sardjito Hospital, Yogyakarta, Indonesia. For this development phase, we
recruited different diagnostic groups of subjects in order
to build pattern recognition techniques. We enrolled 20
TB patients, 20 TB suspects that were later diagnosed
with disease other than TB, 20 patients with a variety of
other pulmonary diseases than TB, 20 healthy matchedcontrols, and 7 patients with Multi Drug Resistant TB.
Subjects had to breathe gently through the eNose
EP-213-06 CD27 expression as a new tool to

distinguish active tuberculosis from LTBI
D Goletti,1 E Petruccioli,1 L Petrone,1 V Vanini,1 G Cuzzi,1
F Palmieri,1 E Girardi1 1Istituto Nazionale per le Malattie
Infettive, Rome, Italy. Fax: (39) 6558 2825. e-mail:
delia.goletti@inmi.it
Background: Control and elimination of the global

tuberculosis (TB) may be enhanced by identifying and
treating latent tuberculosis infection (LTBI). LTBI is
identified by a normal chest X-ray in a healthy subject
with a positive response to immune tests as QuantiFERON TB Gold (QFT) which, unfortunately, does not
distinguish between LTBI and active TB. Recently in
studies from high TB endemic countries the cytometry
evaluation of the down-modulation of the surface
molecule CD27 on TB-specific CD4 T-cells producing
IFNc in peripheral blood mononuclear cells (PBMC) has
been associated to active TB, providing evidence of being
S428
a potential tool to distinguish active TB from LTBI. No

studies have been conducted in Europe and using whole
blood which is an easier sample to obtain compared to
PBMC.
Design/Methods: To validate in a low TB endemic
country as Italy if the down modulation of CD27
evaluated by cytometry in whole blood is a tool to
discriminate, among QFT-IT-positive patients, active TB
from LTBI.62 HIV-uninfected QFT-positive subjects
were enrolled: 15 active TB; 17 cured TB; 30 LTBI.
Whole blood cells were stimulated with RD1 proteins
(ESAT-6 and CFP-10), and as recall antigen with
Cytomegalovirus lysate (CMV). Antigen-specific response was evaluated by cytometry calculating IFNc
production. The data were presented as a RATIO of the
Median Fluorescence Intensity (MFI) of CD27 in the
CD4 T-cells gate over the MFI of CD27 in the gate of
CD4 IFNc T-cells responding to the stimulus.
Results: A higher RATIO to RD1 antigens was found in
active TB compared to cured TB and LTBI (P 0.01 and
P 0.0013 respectively) due to a down modulation of
CD27 specific T-cell in active TB. Conversely no
differences of MFI RATIO were observed in response
to CMV stimulation among the different groups. Based
on the significant difference found in the quantitative
analysis in response to RD1, we performed a ROC
analysis to evaluate the potential of CD27 MFI RATIO
for TB diagnostics. By ROC analysis a significant area
under the curve (AUC) was found when comparing TB
versus LTBI (AUC, 0.86; 95%CI, 0.69-1, P 0.001). For
scoring purposes we found that a cut-off of 3.19
predicted active TB with 75% sensitivity (95%CI,
42.81% to 94.51%) and 100% specificity (95%CI,
79.41% to 100.0%).
Conclusion: This study proposes a new cytometric tool in
whole blood based on the modulation of TB-specific
CD27 CD4 T-cells to distinguish, among QFT-positive
subjects, TB disease from LTBI.
EP-214-06 Rapid detection of viable

Mycobacterium tuberculosis using qPCR with
propidium/ethidium monoazide
A Takaki,1 K Chikamatsu,1 A Aono,1 H Yamada,1,2
K Sakashita,3 L Yi,3 T Matsumaru,1 S Mitarai1,2 1The
Research Institute of Tuberculosis (RIT), Japan
Anti-Tuberculosis Association (JATA), Kiyose, Tokyo,
2
Graduate school of Biomedical Science, Nagasaki University,
Kiyose, Tokyo, Japan. Fax: (81) 42 492 4600. e-mail:
takaki@jata.or.jp
Background: Recently, nucleic acid amplification techniques (NATs) including XpertwMTB/RIF are used
widely for the rapid diagnosis pulmonary tuberculosis
(TB). However, NATs detect all bacterial DNA regardless
of the viability, and cannot be used to monitor the
treatment outcome. Given the long time for TB culture,
to develop rapid methods for quick distinction of viable
M. tuberculosis from dead one is very important for
accurate treatment plans. The qPCR methods with
Propidium/ Ethidium monoazide (PMA/EMA) are widely used for the evaluation of viable microorganisms like
Escherichia coli and Cronobacter sakazakii, but not for

Mycobacterium. To date, no method is available for clear
distinction and accurate quantification of viable M.
tuberculosis. In this study, we try to develop the qPCR
system for accurate quantification of viable M. tuberculosis in clinical specimen.
Design/Methods: PMA/EMA can cross-link to double
stranded DNA of dead/injured bacteria through the
impaired membrane. The cross-linked DNA cannot be
denatured, thus, only live bacteria can be detected using
PMA/EMA-qPCR system. M. tuberculosis H37Rv in the
logarithmic growth phase was killed by boiling 10 min or
10N formalin. The numbers of live cell (with PMA/
EMA) and total cell (no treatment) were quantified by
TaqMan PCR system. To confirm the result, we counted
the bacterial number under microscopy with ZiehlNeelsen staining (total cell count), Auramin O-CTC
double staining and culture (CFU count).
Results: The proportion of viable M. tuberculosis in the
logarithmic growth phase using microscopy and culture
were 67100%. Those of TaqMan PCR with PMA (100
200lM) and EMA (100200lg/ml) were 4154% and
6190%, respectively. The viability measured by AOCTC double staining was low (1548%) compared with
conventional methods. With killed bacteria, PMA/EMAqPCR indicated 1020% and 111% of viablity,
respectively. Therefore, EMA treated specimens were
considered indicating the cell viability correctly.
Conclusion: We considered that EMA-qPCR system
could quantify viable M. tuberculosis more accurately
than PMA-qPCR in our setting. This method could be
applied to evaluate clinical specimens.
EP-215-06 Improved diagnosis of extrapulmonary tuberculosis by antigen detection in

extra-pulmonary samples by
immunochemistry-based assay
M Jrstad,1,2,3 M Marijani,3 M Ali,3 L Sviland,1,2
T Mustafa1,2 1Haukeland University Hospital, Bergen,
2
University of Bergen, Bergen, Norway; 3Mnazi Mmoja
Hospital, Zanzibar, Tanzania. e-mail:
melissa.jorstad@gmail.com
Background: Extrapulmonary tuberculosis (EPTB) constitutes 20-40 % of all tuberculosis (TB) cases and poses
diagnostic challenges due to the paucibacillary nature of
disease, leading to lower sensitivity of routine acid-fast
microscopy and culture methods. Immunochemistry
with an antibody against the Mycobacterium tuberculosis complex specific antigen MPT64 is a robust and easily
applicable method, which has shown higher sensitivity
and specificity in the diagnosis of EPTB. However, the
feasibility of implementation of the assay in the routine
diagnostics and the cost-effectiveness remains to be
evaluated. The aim of this study was to improve the
diagnosis of EPTB by implementation and validation of
the sensitive and specific immunochemistry-based assay
in the routine diagnostics in a resource-constrained
setting.
Design/Methods: The study was conducted at Mnazi

Mmoja Hospital, Zanzibar, from August 2014 till April
2015. Patients suspected of having EPTB as identified by
using standardized clinical diagnostic algorithms were
included. The samples were collected from the respective
sites; by tapping of abdominal-, pleural-, and CSF fluid,
fine needle aspiration (FNA) of lymph nodes and biopsy
of lymph nodes where FNA was inconclusive, and
drainage of abscess material. Samples were subjected to
routine diagnostic work-up; Ziehl-Neelsen (ZN) staining, cytology, histology, culture and immunochemistry
assay and Xpert MTB/RIF. Positive ZN staining, and/or
positive culture, and/or positive Xpert MTB/RIF and/or
response to treatment were used to define a TB case.
Results: There were 83 patients enrolled in the study, 46
males and 37 females, median age 30 years. The sites of
infection were 37 lymph nodes, 24 pleura, 11 abdominal,
9 meninges, 1 spine and 1 osteomyelitis. Based on clinical
criteria and initial investigations, 48 patients were
started on antituberculous therapy. Among these, 7 were
positive with ZN staining, 4 with culture (24 unfinished
results), and 29 were positive with immunochemistry
assay. 28 of these samples were examined with Xpert
MTB/RIF, and M. tuberculosis was detected in 5 samples.
All samples positive with ZN staining, culture, and Xpert
MTB/RIF were also positive with the immunochemistry
assay.
Conclusion: The immunochemistry assay for identification of the antigen MPT64, is implementable in resourceconstrained settings, and may prove a valuable and costeffective test for correct and earlier diagnosis of EPTB.
EP-216-06 Molecular detection of M.

tuberculosis from sputum in a novel transport
medium is not affected by laboratory delay and
ambient temperature
S Omar,1,2 R Peters,2,3 N Ismail,1,2 A W Dreyer,1 P B Fourie2
1
National Institute for Communicable Diseases,
Johannesburg, 2University of Pretoria, Pretoria, 3Anova
Health Institute, Johannesburg, South Africa. e-mail:
shaheedvo@nicd.ac.za
Background: Modern molecular-based approaches to

detection of M. tuberculosis in sputum samples promise
quicker and more accurate detection of cases. However,
processing sputum samples at central diagnostic facilities
provides a diagnostic approach, but requires a safe and
efficient system that is not affected by transport delays
and ambient temperature to be feasible. We evaluated the
technical properties of PrimeStorew Molecular Transport Medium (PS-MTM) for its ability to inactivate
mycobacteria and ensuring stability of DNA over time at
ambient temperatures, and to assess the compatibility of
the transport medium with DNA extraction systems.
Design/Methods: For assessing inactivation ability, a
matrix of high-concentration M. tuberculosis spiked to
proteinacious sputum was inoculated into PS-MTM and
regularly sampled for culturing over time. Compatibility
tests employed QiaAMP DNA, MagNAPure 96 and
Nuclisens EasyMag. Spiked sputum directly to Xpert
S429
MTB/RIF served as control. To show DNA stability,

changes in Ct-values were recorded at 5 time-points over
28 days of positive samples bench-stored in PS-MTM.
For detection of M. tuberculosis in clinical samples, 256
specimens were investigated.
Results: Complete inactivation of M. tuberculosis
occurred within 30 minute of exposure at ratio of 1:2
for sputum to PS-MTM. PS-MTM in automated beadbased extraction systems produced the highest DNA
(estimated lower limits of detection: 169 and 173 CFU/
mL respectively). The Xpert MTB/RIF control was less
sensitive at concentrations below 2500 CFU/ml. No
significant change occurred over time in Ct-values (, 5%
on a mean starting value of 22.47) for PS-MTM samples
over 28 days at ambient temperature. Of 256 clinical
sputum specimens, 10.2% were culture positive and
11.0% positive by real-time PCR (PS-MTM samples).
Against culture, detection of M. tuberculosis from
sputum in PS-MTM had a sensitivity of 77% (CI 95%:
56-91%) and specificity of 96% (CI 95%: 93-98%).
Conclusion: Collecting and transporting sputum from
TB suspects in PS-MTM offer safe transport at ambient
temperature, DNA stability for extended periods without
cooling, specimens directly suitable for molecular testing
and sequencing, and on-shelf long-term molecular
biobanking that could save on storage costs. Detection
of M. tuberculosis positive samples from clinical
specimens with low CFU/ml concentrations is improved
when sputum is collected in PS-MTM.
EP-217-06 Cough sample collection and

processing for the diagnosis of pulmonary TB
W Elmaraachli,1 P Sislian,2 S Chapman,2
R Laniado-Laborin,3 D Martinez-Oceguera,3
N Areli Avila,3 A Catanzaro1 1University of California, San
Diego, San Diego, 2Deton Corp, Pasadena, CA, USA; 3Clinica
y Laboratorio de Tuberculosis, Hospital General Tijuana,
Tijuana, Mexico. e-mail: welmaraachli@ucsd.edu
Background: We propose that the infectious droplets in

cough can be collected and used as a means to diagnose
active TB. Detons patent-pending Cough Collector
(Figure) replaces sputum samples by capturing airborne
bacteria directly from cough for subsequent diagnosis.
The Cough Collector improves TB diagnosis by collecting samples from all patients, including those who
cannot expectorate sputum, thus potentially avoiding
invasive procedures such as bronchoscopy. It is simple,
non-invasive, and can be use in low-resource facilities
with minimal training. Aerosolized M. tuberculosis
captured by Detons Cough Collector can be eluted
directly from the collection surface for diagnostic
analysis. No liquefaction or decontamination of the
sample is necessary thus allowing direct transfer of the
cough sample to an automated diagnostic platform,
eliminating manual steps, and possibly avoiding the
decreased yield associated with the decontamination
procedures that sputum samples are subjected to. .
Design/Methods: Patients suspected of having TB
presenting to Tijuana General Hospital were recruited.
Sputum was collected for acid fast bacillus smear and
S430
culture (BBL MGIT). Cough samples were collected from

each patient on the same day as the sputum sample using
the Cough Collector by having them cough into the
mouthpiece every 10 seconds for 10 minutes. The
collection surface was removed and placed in a microfuge tube; DNA extraction was then performed. TB
DNA was detected by real-time polymerase chain
reaction (rt-PCR) using the AnDiaTec Mycobacterium
tuberculosis Complex kit.
Results: 13 patients were recruited, 9 of whom were
smear positive (and culture-positive) for M. tuberculosis.
The Cough Collector successfully detected M. tuberculosis in 8 of these patients.
Conclusion: The Cough Collector may present a
potentially simpler and more reliable method of sample
collection for the diagnosis of TB through the collection
of cough droplets. More studies with larger numbers of
patients are required.
confounders were assessed in the predictive model and

variables with completeness above 80% and statistical
significance in the bivariate analysis (P , 0.05) were
included in the final model. The association measure
relative risk and confidence interval between exposure to
ART and TB treatment successes in TB-HIV co infected
cases was estimated by Poisson regression with robust
variance model. Data was analyzed in Stata v.12
software.
Results: In 2011, 63.6% (46 865) of TB new cases were
screened to HIV and 7628 (10.3%) had a positive result.
Among TB-HIV coinfected cases only 49.5% (3,502)
were under ART. The cure rate was 63.1% among those
exposed to ART, 27.7 percentage points higher than
among those not exposed. Multivariate analysis found a
relative risk of 1.72 (1.63 1.81) of the PBF benefit in
tuberculosis treatment success when adjusted for the
covariates age group, race, TB type, diabetes mellitus,
alcoholism, directly observed treatment.
Conclusion: The present study using record linkage of TB
and HIV/ aids datasets unprecedented in the country
shows the effectives of ART in TB cure rate in TB-HIV
coinfected cases. The integration between both TB and
HIV programs improve care and support for individuals
affected by both diseases. Although progress is being
made, gaps in the implementation of collaborative
interventions remain far from expected targets.
EP-219-06 Effect of a daily 15 mg/kg dose of

rifapentine on steady state pharmacokinetics
of efavirenz, emtricitabine and tenofovir in
HIV-positive patients
15. From TB diagnosis to outcomes, and

everything in between
C Farenc,1 S Doroumian,1 V Esposito,2 L Perrin,1

C Rauch,1 X Boulenc,1 I Cieren-Puiseux,3 M Maroni3
1
Sanofi, Montpellier, 2Sanofi, Chilly-mazarin, 3Sanofi,
Gentilly, France. e-mail: isabelle.cieren-puiseux@sanofi.com
EP-218-06 TB treatment outcome in TB-HIV coinfected cases stratified by ART, Brazil, 2011
Background: Late 2014, the FDA granted the indication

for the treatment of Latent TB infection for rifapentine
(RPT). This approval was based on the results of the
CDC -T B T C S2 6 w hi c h s h o wed th at we ek l y
RPTisoniazid (900mg each) for 3 months is effective
and well tolerated in active TB prevention. RPT is an
inducer of CYP3A4 and 2B6, isoforms involved in the
metabolism of the anti-retroviral drug, efavirenz (EFV).
A study showed that concomitant use of weekly RPT and
daily EFV is possible with no dose adjustment (EFV
AUC0-24 decrease ,15%). The TBTC, with the aim of
shortening TB treatment duration, is setting up a phase 3
study (S31), including HIV patients, to determine the
efficacy of 17-week regimens containing daily RPT
(1200mg) vs.the standard 26-week regimen. A preliminary study was conducted to assess interactions between
daily RPT and Atriplaw, a combination of EFV 600mg,
emtricitabine (FTC) 200mg, and tenofovir disoproxil
fumarate 300mg. FTC and tenofovir are not metabolized
by CYP450 but are substrates of transporters which may
be up-regulated in presence of RPT.
Design/Methods: 13 HIV/TB-free patients with CD4
Tcell count .350 cells/mm3 and viral load,LOQ,
P Bartholomay Oliveira,1 A Torrens,1 F Dockhorn,1

D Barreira1 1Ministry of Health, Brasilia, Distrito Federal,
Brazil. Fax: (556) 132 138 215. e-mail:
patricia.bartholomay@saude.gov.br
Background: Tuberculosis (TB) and human immunodeficiency virus (HIV) present a huge challenge to health,
social, economic and developmental welfare. In the
context of TB-HIV collaborative activities, World Health
Organization (WHO), recommends that all TB patients
should be screened for HIV and the antiretroviral therapy
(ART) should be offered to HIV-infected TB individuals
irrespective of their CD4 count. This is crucial for
reducing mortality among TB patients living with HIV.
The aim of the current study is to assess TB treatment
successes in TB-HIV co infected cases stratified by ART.
Design/Methods: The study has a retrospective cohort
design using routine program data. It was included in the
study all new TB cases diagnosed and registered in 2011.
Study variables were extracted from a linkage database
of the Aids and TB reporting systems. Percentages were
calculated to report treatment outcomes. Potential
receiving Atriplaw as background therapy, were enrolled

in an open-label, single sequence, 2-period, non-randomized study. PK interactions were assessed after single
and 21days of daily 15mg/kg RPT dosing. Blood samples
were collected over 24 hours post Atriplaw dose, before,
after the first and last dosings of RPT. Plasma concentrations were quantified by LC/MS, PK parameters
calculated by a non-compartmental analysis. Estimates
of Cmin, Cmax, and AUC0-24, geometric mean ratios
with 90% CIs for EFV, FTC and tenofovir co-administered with RPT vs.alone were obtained using linear
mixed effects model.
Results: GEOMETRIC MEAN RATIOS (with/without
repeated doses of RPT) efavirenz (EFV) emtricitabine
(FCT) tenofovir Parameter estimate 90% CI estimate
90% CI estimate 90% CI Cmax 0.83 (0.72-0.94) 1.11
(0.99-1.24) 1.20 (1.02-1.40) Cmin 0.63 (0.52-0.76) 1.14
(0.99-1.31) 1.05 (0.92-1.20) AUC0-24 0.67 (0.59-0.78)
1.01 (0.97-1.06) 1.04 (0.97-1.11). The co-administration of Atriplaw and RPT was well tolerated and no
clinically significant changes of CD4 cell counts or viral
loads were observed.
Conclusion: After 21 daily administrations of 15mg/kg
RPT, FTC and tenofovir steady state exposures were
comparable to Atriplaw administered alone whereas
EFV steady state exposure was decreased by 17% for
Cmax, 37% for Cmin and 33% for AUC0-24. The
magnitude of the EFV-RPT interactions is modest
considering the observed high inter-subject variability
of EFV exposure (AUC0-24 CV%: 77 to 85%). Some
patients had C min and mid dosing interval concentrations of EFV ,1000 ng/ml. Further investigations of EFV
PK and HIV treatment response when RPT is given at
1200mg daily to TB-HIV patients will be done in S31.
EP-220-06 Effect of yoga on lung function and

exercise tolerance in pulmonary impairment
after tuberculosis
R Singla,1 M Mallick,1 S Sharma,1 N Singla,1 S Gupta,1
S Munjal,1 U K Khalid,1 I V Basavaraddi2 1National Institute
of Tuberculosis & Respiratory Diseases, New Delhi, 2Morarji
Desai National Institute of Yoga, New Delhi, India. Fax: (91)
1 126 517 834. e-mail: drrupaksingla@yahoo.com
Background: Yoga includes relaxation, balanced diet,

shatkarma (cleansing practices), yogasana (yogic postures), positive thinking and mediation. Yoga is suited for
promoting relaxation, psycho-emotional stability and
exercise tolerance. After treatment of tuberculosis
significant number of patients are left with sequalae.
Yoga may be helpful in such patients. However, studies
showing the effect of yoga in pulmonary impairment
after tuberculosis are not available. The study was aimed
to see the effects of yoga on lung functions and exercise
tolerance in patients with pulmonary impairment after
tuberculosis.
Design/Methods: A randomized case control study in
which 74 smear positive pulmonary tuberculosis patients, with symptoms or Forced Expiratory Volume in
1st second (FEV1) less than the lower limit of predicted
S431
normal, were enrolled within 3 months after their

treatment completion and randomized in two groups of
37 each: Study group and Control group. Patients in
study group practiced yoga for three months along with
pharmacotherapy whereas patients in control group
continued pharmacotherapy only. Lung function tests
including spirometry, lung volumes and diffusion study
(DLCO), and 6-minute walk test were done at the
baseline and end of 3 months.
Results: 32 patients in study group and 31 patients in
control group completed the study. Patients in study
group showed significant improvement in mean %
improvement in various lung function parameters
compared to control group [FEV1 (13.37% vs.
6.69%, P ,0.001); TLC (8.16% vs. 0.26%, P
0.002); DLCO (16.23% vs.-10.47, P , 0.001); (6minute walk distance (11.95% vs. 1.22%, P , 0.001)].
The study group showed improvement in lung functions
after 3 months of Yoga practice whereas the control
group without addition of Yoga showed deterioration of
lung functions.
Conclusion: Yoga improves pulmonary functions and
exercise tolerance in patients with pulmonary impairment after tuberculosis. It could be a useful adjunct in the
management of pulmonary impairment after tuberculosis.
EP-221-06 Factors associated with high

mortality Among TB-HIV co-infected patients in
Kenya
B Ulo,1 F Ngari,2 D Mobegi,1 J Ongango,3 E Masini,2
M Mungai,1 H K Kipruto,4 S Muhula1 1AMREF Health
Africa Kenya, Nairobi, 2Ministry of Health, Nairobi, 3Kenya
Medical Research Foundation, Nairobi, 4World Health
Organization Kenya Country Office, Nairobi, Kenya. e-mail:
ulobenson@yahoo.co.uk

Kenya, despite decline in TB mortality between 2012
and 2013, mortality among TB-HIV co-infected patients
increased by 23.4 % from 7,700 to 9,500 and accounts
for 60% of TB deaths annually. Kenya has a policy of
immediate start of anti retroviral therapy (ART) to TB
and HIV co-infected patients but ART uptake is 86%. We
sought to understand factors associated with high
mortality among these patients from TB data that is
captured in a case based electronic reporting system
called TIBU.
Intervention or response: Data for all TB and HIV coinfected patients notified in TIBU from January 2013 to
December 2014 from all TB treatment sites countrywide
was extracted. Descriptive statistics used to quantify
proportions of patients who died against various
independent variables. We ran logistic regressions to
determine predictors to mortality using SPSS version 21.
Proportions of key variables were determined against
outcome of died or alive as the dependent variable. Using
Binary logistic regression we determined individually
significant variable that were then analyzed together.
Results and lessons learnt: In total of 179 881 TB
patients were registered in 3,320 facilities between 1st
S432
January 2013 and 31stth December 2014. . Of these,

34.5% [62,023 (30 659 female and 31 364 male)]
patients were co-infected. Mean age in years was 37 and
35 for died and alive respectively. Of all 7886 deaths,
59% (4644) were among TB-HIV co-infected patients.
For those with ART initiation date recorded, 67.5%
(1,129) of deaths occurred within two months before and
after start of TB treatment. Death rate was 10% and 7%
for those within two months before and after TB
treatment start respectively. Only 39.9% of those on
ART were referred from HCC. After adjusting for other
factors, death in TB-HIV co-infected patients were
significantly associated with being male, [OR] .785,
95% confidence interval [CI]: .734-.840), BMI , 18.5,
[OR] .711, 95% [CI]: .868-.737), not receiving ART
[OR] 1.354, 95%CI: 1.237-1.482), and patient types:
Retreatment smear negative [OR] 1.694, 95%CI: 1.044.02), Return after default [OR] 1.926, 95%CI: 1.2243.031 and Retreatment extra pulmonary [OR] 1.751,
95%CI: 1.105-2.776.
Conclusions and key recommendations: TB-HIV coinfected patients had a high case mortality rate during
intensive phase of TB treatment. Close monitoring and
documentation is required especially in the HIV care
clinic.
EP-222-06 Outcomes of HIV-TB co-infected

patients on rifabutin and antiretroviral therapy
in Kenya, 2012
H Weyenga,1 A Katana,1 J Kioko,2 E Masini,2
C Munguti,1 E Ngugi,1 M Schmitz,1 L Nganga1 1Center
for Disease Control, Nairobi, 2Ministry of Hearlth, Nairobi,
Kenya. e-mail: howeyenga@gmail.com
Background: Tuberculosis (TB) is a leading cause of

morbidity and mortality in HIV-infected patients. Optimal TB treatment and HIV viral suppression are critical
predictors of patient outcomes. Unfortunately, drug-drug
interactions between the current first-line (rifampicinbased) TB regimen and commonly used antiretroviral
(ARV) drugs complicate treatment for TB-HIV coinfected patients. Rifampicin significantly reduces the
plasma level of lopinavir/ritonavir (LPV/r). The Kenya
ARV guidelines recommend substitution of rifampicin
with rifabutin for TB-HIV co-infected patients on LPV/r.
Although 83.7% of Pulmonary TB (PTB) and 81.2% of
extra pulmonary TB (EPTB) HIV-infected patients on
rifampicin are successfully treated, outcomes for patients
on rifabutin are unknown. We analyzed national TB data
to describe treatment outcomes for co-infected patients
on rifabutin and LPV/r-based ARV regimen
Design/Methods: We used national TB data to identify
all TB-HIV co-infected patients who had been placed on
rifabutin from January to December 2012. We abstracted
demographic and clinical data and computed point
estimates, for TB treatment outcomes Cured, Failed,
Completed treatment, Died, Defaulted and Successfully
treated which were defined using the WHO definition
and reporting framework
Results: Overall, 57 patients had been placed on

rifabutin and LPV/r regimens; 36 (63.2%) were female.
The median age was 38 years (IQR 30.5-45.0), median
body mass index (BMI) was 18.7Kg/m2 (IQR16.5-20.8)
and median CD4 count was 105 cells/mm3, (IQR 14188). Of these, 42 (73.7%) had PTB, with 14 (33.3%)
having positive sputum smears, whereas 15 had EPTB.
Of the sputum smear positive cases, 9 (64.3%) were
cured, 2 (14.3 %) completed treatment, and 3 (21.4 %)
failed. Among smear negative cases, 23 (82.1%) completed treatment and 5 (17.9 %) died. Of the EPTB cases,
12 (80.0%) completed treatment, 2 (13.3%) died and 1
(6.7%) defaulted. Sex, [OR 0.42; 95%CI (0.09-2.00)],
PTB, [OR 0.63; 95%CI (0.11-2.60)], CD4 count ,50/
mm3, [OR 0.55; 95%CI (0.13-2.39)], and BMI ,16Kg/
m2, [OR 0.55; 95%CI (0.13-2.39)], were not significantly associated with treatment outcomes
Conclusion: TB treatment success for HIV infected
patients on rifabutin was similar to the average treatment
success for all HIV-TB patients on rifampicin. Rifabutin
should be encouraged in programmatic settings in Kenya
to treat TB-HIV co-infected patients on LPV/r based
ART regimens. A larger sample size is required to
determine factors significantly associated with outcomes
EP-223-06 Predictors of tuberculosis treatment

outcomes in a primary care setting in South
Africa
G Louwagie,1 L Ayo-Yusuf,1,2 I Chaparanganda1
University of Pretoria, Pretoria, 2Sefako Makgatho Health
Sciences University, Tshwane, South Africa. e-mail:
goedele.louwagie@up.ac.za
Background: There is limited information on the effects

of socio-economic status, substance use and psychosocial factors on tuberculosis (TB) treatment outcomes in
TB patients with high HIV co-infection rates.
Methods: This was prospective cohort study of 1926
newly diagnosed TB patients at six large clinics in a South
African township. Current tobacco smokers in this
cohort were enrolled in a tobacco cessation trial, but
all participants were followed up for TB treatment
outcomes. We collected information on socio-economic
status, substance use (tobacco smoking, alcohol problem
and illicit drug use) and psycho-social variables (social
support, perceived stress and recent depressive symptoms) using structured interviewer-administered questionnaires. HIV- and TB-related information was obtained from clinical records and TB outcomes from the
TB register. Data analysis entailed univariable and
multivariable multilevel logistic regression with facility
as random effect.
Results: TB treatment was successful (cured or completed) in 77.8% (1258/1618) of the participants. Sixteen
percent of the participants were excluded from the
analysis because of missing outcomes or transfer out. TB
treatment was independently more likely to be successful
in women (Odds Ratio [OR] 1.56, 95% Confidence
Interval [CI] 1.21; 2.02), in married participants (OR
1.52, 95%CI 1.14; 2.03) and in participants who
completed high school versus primary school only (OR

2.14 [95%CI 1.49; 3.09]). HIV-positive participants
(whether on antiretroviral treatment [ART] or not) were
less likely have a successful outcome (OR 0.55 [95%CI
0.34; 0.89], OR 0.50 [95%CI 0.31; 0.80], OR 0.49
[95%CI 0.31; 0.79] for HIV positive on ART, unknown
ART and not on ART respectively, versus HIV negative).
The same variables were predictors for TB treatment
default. Psycho-social variables, tobacco smoking,
alcohol problem or illicit drug use did not influence TB
treatment outcomes.
Conclusions: TB treatment outcomes were worse in
participants who were male, unmarried, uneducated or
HIV-positive. Ever tobacco smoking or social support at
baseline did not predict outcomes. We recommend
tailoring TB adherence programmes to the often neglected high risk groups, in particular unmarried men with
low education. Better monitoring of outcomes is
necessary to differentiate TB treatment interruption
(default) from loss to follow-up due to death.
EP-224-06 Limitations of chest radiography for

TB screening in healthy HIV-infected adults
revealed by comparison with CT
H Esmail,1,2 D Chhiba,3 Q Said-Hartley,3 P Scholtz,3
H Goodman,3 R Wilkinson1,2,4 1University of Cape Town,
Cape Town, South Africa; 2Imperial College London, London,
UK; 3Groote Schuur Hospital, Cape Town, South Africa; 4The
Francis Crick Institute Mill Hill Laboratory, London, UK.
e-mail: hanifesmail@gmail.com
Background: Chest radiographs (CXR) have been widely

used to screen at risk populations for active and inactive
tuberculosis (TB) for over 75 years. Although certain
limitations of TB screening CXR are recognised, it has
never been evaluated in comparison to a gold standard of
computed tomography (CT).
Design/Methods: As part of a study 35 asymptomatic,
HIV-1 infected, antiretroviral therapy nave adults
(median CD4 count 517/mm3), living in Cape Town,
South Africa with evidence of latent TB by Quantiferon
Gold In-tube and no previous history of TB treatment
underwent digital CXR followed by [18F]-fluoro-2deoxy-D-glucose combined positron emission and computed tomography (FDG-PET/CT). Using the CT component (Kv 120, mA 100, slice thickness 3mm),
participants were classified as having changes consistent
with active TB, inactive TB or no evidence of TB, with
allowances made for the lower spatial resolution of
CXR. 7 consultant radiologists, blinded to clinical details
and CT findings, then reported the CXR. A structured
reporting form was used to ensure all sections of the
CXR were reported, with a final classification made as
for CT. All readers reported CXR independently in a
different random order on diagnostic monitors in low
ambient light.
Results: Agreement between the seven readers by kappa
(k) statistic was slight, k 0.10. Agreement of each CXR
reader with CT varied between slight, k 0.05 and fair, k
0.32. Using a binary classification of TB related
changes (active or inactive TB) or no TB related changes,
S433
the sensitivity of CXR readers varied from 33.3%-100%

(median 86.7%) and specificity 25%-94.7% (median
45%). A reciprocal relationship for these parameters was
apparent for each reader, with 5 readers having high
sensitivity and low specificity (over-reading) and 2
readers having low sensitivity and high specificity
(under-reading). The positive and negative predictive
value of CXR readers varied from 50%-83.3% and
66.7%-100% respectively.
Conclusions: This is the first time TB screening CXR
have been evaluated against CT. Agreement between
CXR readers and CT was slight to fair. Some radiologists
had a tendency to over-read and others to under-read
screening CXR. We have provided novel insight into the
limitations of TB screening CXR. CT is not a viable
screening test but is a useful research tool in this setting
and could inform strategies to improve the reporting of
TB screening CXR.
EP-225-06 Genetic polymorphisms of NAcetyltransferase 2 in TB-HIV cotreated

patients in Durban, South Africa
T Mthiyane,1 C Connolly,2 Y Balakrishna,2 T Reddy,2
H Mcilleron,3 A Pym,4 R Rustomjee1 1Medical Reserach
Council, Cape Town, 2Medical Research Council, Durban,
3
University of Cape Town, Cape Town,
4
KwaZuluNatal-Research Institute for Tuberculosis and HIV,
Durban, South Africa. e-mail: thulimthiyane24@gmail.com
Background: Drug induced hepatitis in tuberculosis

patients taking anti-tuberculosis treatment is the most
common adverse event often associated with isoniazid.
Polymorphisms of N-acetyl transferase 2 (NAT2) vary
considerably among different ethnic groups and are said
to affect drug metabolism. We investigated NAT2 single
nucleotide polymorphisms in Durban black South
African population as there is limited information
available.
Design/Methods: One hundred and twenty two HIVinfected participants were enrolled, 102 with cultureconfirmed pulmonary-TB and 20 without TB, whole
blood was drawn and processed genotyped for NAT2
190C.T, 282C.T, 341T.C, 403C.G, 411T.A,
481C.T, 857G.A, 590G.A and 803A.G using Life
Technologies pre-validated Taqman assays (Life Technologies, Paisley, UK). All TB patients received standard
anti-TB treatment. Fishers exact test was used to
compare AE acquisition between the three groups.
Results: 100 out of 102 TB-HIV participants had
genotype results and were evaluated. 61% (61/100) were
found to be slow metabolisers (NAT2*5/*5 and
NAT2*5/*6), 25% (25/100) were intermediate
(NAT2*4/*5 and NAT2*5/12), and 14% (14/100) were
rapid metabolisers (NAT2*4/*11, NAT2*11/12 and
NAT2*12/12). Twenty HIV non-TB participants were
also analysed, 12(60%) were slow 8 (40%) intermediate
metabolisers and none were rapid. NAT2*5 was the most
common haplotype among the TB-HIV and HIV non-TB
groups. The proportion of patients falling into the slow,
intermediate and rapid category did not differ significantly in the TB and non-TB groups (P 0.122). Of the 60
S434
patients who experienced adverse events among the slow,

intermediate and rapid groups, 29(47.5%), 19 (76%)
and 12(86%) respectively, experienced at least one
adverse event, 55% had at least one hepatic adverse
event. There was a significant difference in the prevalence of AE between the three groups (P 0.005). The
proportion who experienced an AE did not significantly
differ between the intermediate and rapid group.
Conclusion: There was a high prevalence of slow
followed by intermediate metabolisers. Results are
consistent with what has been reported in previous
literature among the African population in South Africa,
except for the absence of NAT2*5B and NAT2*6A. In
this study, the number of adverse events, including
hepatic, were highest among the intermediate and rapid
metabolisers. This is contrary to most research results in
this area. We need more trials to validate these results
EP-226-06 Implementation of intensified case

finding and XpertW MTB/RIF as initial TB
diagnostic in HIV-positive individuals with
presumed TB in 21 health facilities in Kampala,
Uganda
J Lemaire,1 R Nyinoburyo,1 M Muttamba,1 D Asiimwe,1
G Lukyamuzi,1 A Nakanwagi-Mukwaya,2 M Joloba,3
F Mugabe4 1Foundation for Innovative New Diagnostics
(FIND), Geneva, Switzerland; 2International Union Against
Tuberculosis and Lung Disease, Kampala, 3Makerere
University, Kampala, 4Ministry of Health, Kampala, Uganda.
e-mail: jeff.lemaire@finddx.org
Background: Kampala district in Uganda accounted for

3808 notified new bacteriologically confirmed (SM/
bac) TB cases from Q3 2012 to Q2 2013 (baseline).
Supported by a TB-REACH Wave 3 grant from the
StopTB Partnership for the period of July 15th 2013 to
June 30th 2015, FIND and partners have rolled-out
Xpert MTB/RIF (Xpert) as an initial test for diagnosis of
TB among HIV positive patients in 21 HIV-TB care
facilities (6 testing and 15 referral sites) in Kampala
coupled with intensified case finding (ICF) activities both
in outpatient departments (OPD) and HIV clinics.
Methods: Four GeneXpert-IV platforms were installed in
Kampala in July 2013 followed by 2 additional
instruments in July 2014. Xpert testing is available to 8
public and 13 private clinics through a motorbike
referral system. Two approaches were newly implemented to increase TB case finding; 1) ICF activities are
carried out in 12 OPD and 16 HIV clinics by actively
screening clinic attendees for TB symptoms through a
small questionnaire. 2) HIV positive () TB suspects are
directly tested by Xpert (compared to following an initial
smear negative (-) result), while HIV- TB suspects
undergo smear microscopy (SM) as per national guidelines.
Results: ICF: From Q3 2013 to Q4 2014, 43 0 094 OPD
and 4 0 774 HIV patients were actively screened for TB
symptoms from which 9 0 465 (22%) and 1 0 732 (36%)
suspects were identified respectively, and referred to
clinicians for further evaluation, after which only 4 0 854
(51%) from OPD (of which 853 HIV) and 894 (52%)
from HIV clinics were referred for laboratory testing. A

total of 727 SM/Xpert were identified through ICF at
OPD (566) and HIV clinics (161). Xpert as initial test:
during the same period, an additional 11 0 077 HIV TB
suspects (passive findings) were tested with Xpert as an
initial diagnostic test. From the resulting 2 0 057 (19%)
confirmed TB cases, 118 (6%) were resistant to
rifampicin. During these 18 months, the NTLP reported
6 0 221 new SM/bac notified cases in Kampala district,
representing 509 additional TB cases (9% relative
increase) as compared to the baseline, to which both
project interventions contributed to.
Conclusion: After 18 months of implementation of this
24 month project, innovative approaches of ICF and
frontloading Xpert for HIV TB suspects have led to the
identification of 2 0 784 new TB cases in 21 health
facilities in Kampala district, and contributing to a 9%
relative increase in case notifications as compared to
baseline.

21. TB epidemiology: post and future
insights into case finding, drug resistance
and mortality
OA-459-06 Tuberculosis case finding through
active screening using mobile phone
application in Jakarta: implementation of a TB
Reach project
Y Permata,1 D Purba,1 R Tjiong,1 A Loebis,1 E Burhan,2
D A Asri,1 I Kurniawati,3 A Shankar4 1Inovasi Sehat
Indonesia, Jakarta, 2Faculty of Medicine Universitas
Indonesia, Jakarta, 3Provincial Health Office, Jakarta,
Indonesia; 4Harvard School of Public Health, Boston, MA,
USA. e-mail: yusie.permata@gmail.com
Background and challenges to implementation: Indonesia is one of the high burden countries for tuberculosis
(TB). The prevalence of TB in Indonesia was 680 000 in
2013, with 64 000 deaths. Worldwide, Indonesia ranks
four for TB incidence number. National TB program in
Indonesia has not included active case finding as one of
the strategies to find TB cases as recommended by WHO.
As part of TB Reach project, we did active screening to
intensify TB case finding in several settings in Jakarta,
Indonesia.
Intervention or response: We have conducted active
screening in out-patient departements (obsetric, diabetes,
and HIV clinic) of hospitals, private clinics, factories,
prisons, and social shelters during the period of
November 2013 to mid April 2015. Trained screeners
interviewed respondents in the study settings using
mobile application developed by Inovasi Sehat Indonesi.
Demographic data, body weight, height, TB-related
symptoms and risk factors were collected. Based on the
questionnaire scoring, respondents were categorized as
high TB suspects, medium TB suspects, or not a supsect.
High and medium TB suspects were requested to give
Oral abstract sessions, Sunday, 6 December
their sputum for GeneXpert MTB/RIF assay in our

diagnostic centers. SPSS program were used for univariate and multivariate data analysis. We excluded
incomplete data from the analysis
Results and lessons learnt: In 15 months period, we have
screened 41 546 people, with 7059 people were
suspected to have TB, but only 5,934 who were willing
to give out their sputum. Of all tested, 714 were
Mycobacterium tuberculosis positive (around 10% of
suspects). Among M. tuberculosis positive, 60 were
rifampicin resistance TB cases (8.4%). The prevalence of
diabetes mellitus (DM) in tested were 13%. Among
tested, 21.7% were manourished, 53.7% normal BMI,
and 24.6% overweight. We found that TB is associated
with malnutrition (RR 2.67; P 0.001) and DM (RR 1.41;
P 0.001). The main challenge we faced was around
15.9% TB suspects screened refused to collect their
sputum. The number of TB cases found could be higher if
all suspects sputum are tested.
Conclusions and key recommendations: The significant
number of TB cases detected in our study shows that
active screening is a crucial strategy to find TB cases.
Active screening should be included as a part of national
TB program in Indonesia, especially in high-risk population such as malnourished people and DM patients, so
that active TB cases can be diagnosed and put in
treatment earlier to halt the transmission.
OA-460-06 Mortality and pre-treatment loss to

follow-up in patients being investigated for TB
in a pilot case manager intervention
N Maraba,1 K Mccarthy,1 G Churchyard,1,2 A Grant,2
V Chihota1 1RESEARCH, The Aurum Institute, Johannesburg,
South Africa; 2London School of Hygiene & Tropical
Medicine, London, UK. e-mail:
nmatabane@auruminstitute.org
Introduction: South Africa has replaced microscopy with

Xpert MTB/RIF, in a centralised laboratory network, as
the first test for TB. Pre-treatment loss to follow-up,
defined as failure to start TB treatment within 28 days
among people with a positive test result, remains high.
We investigated a case manager strategy to improve
linkage to TB and HIV care: we describe mortality and
pre-treatment loss to follow-up.
Methods: Prospective cohort design. Consecutive consenting adults who provided sputum for TB investigations using Xpert MTB/RIF in 5 primary care clinics in
Gauteng and Mpumalanga provinces, South Africa were
enrolled and completed a baseline questionnaire. All
participants were assigned a trained lay case manager
facilitating their linkage into HIV and/or TB care and
were followed up for three months. At 3 months,
mortality was determined by telephonic calls to next of
kin /friends or home visits if the participant could not be
contacted.
Results: We report on preliminary results from the study.
761 persons were screened, 412 (54%) participants were
enrolled with median age 36 years (interquartile range:
30-44), 274 (67%) HIV positive and 228 (55%) being
S435
females. Thirty eight (9%) participants had a positive TB

test result at enrolment. By 28 days after enrollment, 34
(90%) had started TB treatment, 2 (5%) had died and 2
(5%) had not started treatment, giving a total of 4 (11%)
classified as pre-treatment loss to follow-up. At end of 3
months follow-up, 11/412 (3%) participants had died.
Of those who died, 9 (82%) were HIV positive, 1 (9%)
negative and 1 (9%) declined HIV testing.
Conclusion: Despite support from a case manager, early
mortality was high among people being tested for TB.
Additional strategies to reduce mortality among this
group need to be explored.
OA-461-06 Predictors of mortality in patients

diagnosed with preXDR- and XDR-TB: results
from the South African National TB
Programme, 2009-2010
D Evans,1 K Schnippel,2 E Budgell,1 K Kate,1 K Shearer,1
R Berhanu,1,2 L Long,1 S Rosen1,3 1Health Economics and
Epidemiology Research Office, Johannesburg, 2Right to Care,
Johannesburg, South Africa; 3Boston University School of
Public Health, Boston, MA, USA. e-mail: devans@heroza.org
Background: Extensively drug resistant (XDR) TB has

been associated with poor outcomes and high mortality
in multiple cohort studies. South Africa was one of the
first countries to report XDR-TB and annually reports
nearly 1000 XDR-TB diagnoses.
Methods: Retrospective, de-identified descriptive analysis of patients reported in the EDRweb, the electronic
national case register for all DR TB, who initiated
treatment between January 2009 and September 2010.
We used the WHO standard definition of XDR-TB;
preXDR-TB was defined as resistance to either a
fluoroquinolone (FLQ) or a second line injectable (SLI)
in addition to rifampicin and isoniazid (MDR-TB).
Treatment outcomes included success (cured or completed), failed, lost to follow-up, and died. Person-time
accrued from treatment initiation until the earliest of
outcome date recorded or 36 months on treatment. Cox
hazard models were used to identify predictors of allcause mortality.
Results: 10 600 patients initiated DR TB treatment and
were reported in EDRweb. Of these, patients with MDRTB (8611; 81.2%) or rifampicin mono-resistant TB (211,
2.0%) were excluded from the analysis. Of the 1778
(16.8%) analyzed, 936 had XDR TB, 422 had preXDR
FLQ, and 420 had preXDR SLI. 77% (1375/1778) had
an HIV status recorded; of which 62% were co-infected
with HIV, and 76% were receiving ART. Outcomes were
reported for 1231 (69%) patients: 526 (43%) died, 171
(14%) had a successful outcome, 175 (14%) were lost to
follow up and 359 (29%) failed treatment. Median time
to death after starting treatment was 4.4 months.
Mortality was higher among those with XDR-TB
(47%; 1.56/100 pyrs) than those with preXDR FLQ
(38%; 1.11/100 pyrs) or preXDR SLI (38%; 1.12/100
pyrs). Predictors of mortality included history of TB
treatment in patients with preXDR SLI (hazard ratio
2.35 95%CI [1.13-4.87]) and XDR-TB (HR 1.83 [1.07-
S436
3.13]; baseline smear positivity in patients with preXDR

FLQ (HR 2.34 [1.46-3.74]) and XDR-TB (HR 1.84
[1.37-2.47]); and older age (745) (HR 2.18 [1.35-3.52])
in preXDR FLQ patients.
Conclusion: This analysis of South Africas national
database confirms that mortality in preXDR and XDRTB patients remains very high. XDR-TB and preXDR
was defined using laboratory results alone which may
result in exposure misclassification and underestimate
mortality rates. The likelihood of additional high
mortality among those without an outcome reported or
lost to follow up further increases the urgency of
improving XDR-TB treatment.
microbiologically-confirmed TB cases that were RR

increased from 1.8% in 2004 to 5.1% in 2013 (v2 P
0.000). A similar trend was also observed for MDR
proportions. The proportion of RIFR cases that were
confirmed as MDR cases ranged between 90.9% and
63.0% over the study period.
Conclusion: The prevalence of rifampicin resistant TB in
South Africa increased steadily between 2004 and 2012
and has almost tripled from the 2004 baseline. However,
the proportions remain below 10%. The use of RR as
predictor for MDR is not absolute with at least 10%
being mono-resistant rather than MDR. However, the
decline in the predictive value of RR for MDR in later
years may be attributable to the introduction of the Line
Probe Assay that is known to miss a small proportion of
cases with phenotypic isoniazid resistance or to testing
patterns. Our findings of increases in RR prevalence
needs to be further investigated to correct for testing
patterns and further evaluation of this data in the Xpert
MTB/RIF era is also required
Year
OA-462-06 Trends of rifampicin-resistant

pulmonary TB in South Africa: data from the
national laboratory-based TB surveillance
system
1
1 1
A Nanoo, C Ihekweazu, S Candy, N Ismail Centre for

Tuberculosis, National Institute for Communicable Diseases,
Johannesburg, 2Division of Public Health Surveillance and
Outbreak Response, National Institute for Communicable
Diseases, Johannesburg, 3Corporate Data Warehouse,
National Health Laboratory Service, Johannesburg, South
Africa. e-mail: anantan@nicd.ac.za
Background: South Africa has the third highest burden of

pulmonary tuberculosis (TB) globally and while rates of
multidrug resistant TB (MDR) are low, absolute case
numbers are high. Rifampicin resistance (RR) is regarded
as a proxy for MDR and new diagnostics are primarily
targeting this drug for testing. We describe the trends in
RR and MDR-TB over a 9 year period.
Design/Methods: Data on microbiologically-confirmed
TB cases diagnosed between 2004 and 2012 was
extracted from a routine laboratory-based surveillance
system using the National Health Laboratory Services
Corporate Data Warehouse. Basic descriptive statistics
was used to describe the time trends of rifampicin
resistant tuberculosis (RR TB) in South Africa. The
prevalence of MDR-TB relative to RR TB was determined. Data are reported unadjusted for testing rates.
Results: Between 2004 and 2012, 115 129 cases of RR
TB were diagnosed, 93 426 (81.1%) of which were also
confirmed MDR-TB cases (Table 1). The proportion of
2004
2005
2006
2007
2008
2009
2010
2011
2012
TB cases
diagnosed
241
270
312
318
343
339
400
433
405
411
496
507
191
935
956
237
003
836
RR
n (% RR OF TB)
4
3
9
12
12
15
16
19
20
294
443
775
500
793
030
898
744
652
(1.8)
(1.3)
(3.1)
(3.9)
(3.7)
(4.4)
(4.2)
(4.6)
(5.1)
MDR
n (% MDR OF TB;
% MDR of RR TB)
3
3
8
11
11
12
14
14
13
852
164
885
045
201
932
368
974
005
(1.6;
(1.2;
(2.8;
(3.5;
(3.3;
(3.8;
(3.6;
(3.5;
(3.2;
89.7)
91.9)
90.0)
88.4)
87.6)
86.0)
85.0)
75.8)
63.0)
OA-464-06 Rapidly increasing tuberculosis

burden in Brazilian prisons offsets gains in
general population
P Bourdillon,1 J Croda,2,3 A Ko,1,4 J Andrews5 1Yale School
of Medicine, New Haven, CT, USA; 2Federal University of
Grande Dourados, Dourados, Mato Grosso do Sul, 3Oswaldo
Cruz Foundation, Campo Grande, Mato Grosso do Sul,
Brazil; 4Yale School of Public Health, New Haven, CT,
5
Stanford University, Stanford, CA, USA. e-mail:
paul.bourdillon@yale.edu
Background: Brazil has the worlds 4th largest prison

population, which increased from 335 410 to 557 285
from 2007-2013. Although TB rates are high in this risk
group, the population contribution of prison transmission on the Brazilian TB burden has not been evaluated.
Design/Methods: We obtained data on reported tuberculosis cases from 2007 to 2013 in the national
surveillance database (Sistema de Informacao de Agravos
de Notificacao; SINAN). Data was extracted according
to demographics including incarceration status, sex,
race, age, education, and state as well as clinical factors
including disease site, HIV status, and sputum smear. We
used population estimates for Brazil and inmate census
reports to calculate annual TB incidences.
S437
OA-465-06 Under-reporting of sputum smearpositive tuberculosis cases in Kenya

D Tollefson,1 F Ngari,2 M Mwakala,2 D Gethi,3
H K Kipruto,4 K Cain,5 E Bloss1 1U.S. Centers for Disease
Control and Prevention, Atlanta, GA, USA; 2National Leprosy,
Tuberculosis and Lung Disease Unit, Nairobi, 3Kenya Medical
Research Institute (KEMRI), Kisumu, 4World Health
Organization Country Office Kenya, Nairobi, Kenya. 5U.S.
Centers for Disease Control and Prevention (CDC), Kisumu,
Kenya. e-mail: vtu3@cdc.gov
Results: From 2007-2013, the number of cases doubled

from 3511 to 7411 per year, and the proportion of cases
notified from prisons increased from 4.1% to 8.1% of all
cases reported in Brazil. The average incidence among
prisoners was 28.3 times that of the general population,
1216 and 42.9 per 100 000 respectively. The incidence
among prisoners rose by 39% (958 to 1335 per 100 000;
P , 0.0001) compared to a 6% decrease among nonprisoners (44.7 to 41.9 per 100 000; P , 0.0001); the net
effect was just 2% decline in the combined population.
There was significant variability between states, with
average 2007-2012 incidences in prisons ranging from
224 (Amapa State) to 3066 (Rio de Janeiro State) per 100
000. Men comprise 94% of the prison population and
91% prison TB cases, of which 79.3% are aged 20-39.
Although there were fewer female prisoners, they had a
higher average incidence of TB than males (1812 vs
.1179 per 100 000), more extrapulmonary cases (13%
vs. 5.8% of cases), and higher risk of documented HIV
co-infection than males (15.6% vs.10.2% of cases).
Conclusions: The growing TB epidemic in prisons now
accounts for 8% of Brazils disease burden. Given the
high inmate throughput, prisons may serve as reservoirs
for transmission to the general population and present an
increasing obstacle to broader tuberculosis control
efforts. Young men comprise the majority of cases in
prisons, but female prisoners are at higher risk of TB and
HIV co-infection, which may warrant targeted screening.
Effective interventions must be implemented to address
the alarming incidence of tuberculosis in Brazilian
prisons and make an impact on the national disease
burden.

has declared finding missed tuberculosis (TB) cases a top
priority to reach global TB targets, yet the level of underreporting of TB cases is largely unknown in high TB
burden countries. This study was conducted to quantify
the level of under-reporting of sputum smear-positive TB
cases to the national TB surveillance system in Kenya.
Design/Methods: Kenyas National TB Program (NTLDUnit) and the U.S. Centers for Disease Control and
Prevention conducted a national, retrospective TB
inventory study in Kenya, following WHO guidelines.
All sputum acid-fast bacilli smear-positive TB cases
recorded by laboratories (public or private) from April
1 to June 30, 2013 were extracted from paper TB
laboratory registers in 73 sub-counties, which were
selected by stratified random sampling based on TB
burden. These laboratory cases were matched to TB cases
reported within the national, electronic TB surveillance
system (TIBU) in 2013 using name, age, and combinations of sex, geography, and time from diagnosis to
treatment initiation. Lab-diagnosed smear-positive cases
not found in TIBU were defined as unreported to the
NTLD-Unit. A national estimate for under-reporting was
calculated in R software, adjusting for study design.
Results: In sampled sub-counties, 822 of 3,425 (24%) of
smear-positive TB cases found in lab registers during the
study period were not reported in TIBU. The average
proportion of laboratory cases not reported in TIBU was
greatest in high TB burden sub-counties (25%) and
lowest in low TB burden sub-counties (17%). Underreporting in sub-counties ranged from 0% to 44%. The
highest average level of under-reporting occurred in subcounties in Nairobi (34%) and Nyanza region (30%),
while the lowest occurred in sub-counties in Central
(16%) and Eastern (17%) regions. The estimated level of
under-reporting of smear-positive TB cases in Kenya was
23.8% (95%CI: 21.226.4%).
Conclusions: In Kenya, almost one-quarter of smearpositive TB cases diagnosed during the study period were
unreported to the NTLD-Unit. Findings suggest the real
burden of TB is higher than what is currently reported in
Kenya. TB surveillance should be strengthened to ensure
all TB cases are recorded and reported. Additional
research is needed to identify risk factors and reasons
for under-reporting.
S438
OA-466-06 Isoniazid mono-resistance is

associated with increased mortality in a large
cohort of tuberculosis patients in Durban,
South Africa
Y Van Der Heijden,1 F Maruri,1 F Karim,2 G Parker,2
Y Moosa,3 T Sterling,1 T Chinappa,4 P Alexander2
1
Vanderbilt University School of Medicine, Nashville, TN,
USA; 2KwaZulu-Natal Research Institute for TB and HIV,
Durban, 3Nelson R. Mandela School of Medicine, University
of KwaZulu-Natal, Durban, 4Ethekwini Municipality, Durban,
yuri.vanderheijden@vanderbilt.edu
Background: Isoniazid (INH) mono-resistant tuberculosis (TB) is often considered to be of little clinical
significance, though some studies suggest an association
with poor patient outcomes. We hypothesized that
patients with INH mono-resistance had worse clinical
outcomes than patients with drug-susceptible TB.
Design/Methods: We identified patients with INH
mono-resistant TB in a retrospective cohort of TB
patients seen at a single, large, urban TB clinic in
Durban, South Africa (Prince Cyril Zulu Communicable
Disease Centre) from January 2000 through December
2013. For comparison, we identified all patients with
drug-susceptible TB, defined as any patient with Mycobacterium tuberculosis isolates susceptible to INH and
rifampin and all other anti-TB drugs tested. Patients in
both groups received six months of standard four-drug
TB treatment. TB outcomes included all-cause mortality
and poor outcome (defined as death plus treatment
failure).
Results: There were 79 527 TB patients seen at the clinic
during the study period, of whom 19 094 had pulmonary
TB and available drug susceptibility test results. Of these,
488 (2.6%) had INH mono-resistance and 15 374
(80.5%) had drug-susceptible TB. Patients with INH
mono-resistance were more likely to have died (3.3% vs.
1.7%, P 0.01) and treatment failure (3.3% vs. 0.6%, P
,0.001), and less likely to have had treatment success
(42% vs. 50.1%, P ,0.001) compared to patients with
drug-susceptible TB. Patients did not significantly differ
with regard to loss to follow-up or transfer out of care.
Also, patients did not differ with regard to being HIV
seropositive (79% vs. 76%, P 0.34), though only 41%
had known HIV serostatus. In a multivariable logistic
regression model adjusting for age, sex, and race,
patients with INH mono-resistance were more likely to
have died than patients with drug-susceptible TB (odds
ratio [OR] 1.94; 95% confidence interval [CI] 1.16,
3.24; P 0.01; see Table). In a similar model, patients
with INH mono-resistance were more likely to have a
poor outcome than patients with drug-susceptible TB
(OR 3.01; 95%CI 2.07, 4.38; P ,0.001).
Conclusion: Patients with INH mono-resistance had
worse clinical outcomes than patients with drug-susceptible TB. Despite the limitations of a retrospective study,
our findings from this large TB cohort support the need
for further studies investigating appropriate rapid
diagnostic and treatment approaches for patients with
INH mono-resistant TB. Table. Multivariate logistic
regression model of predictors of death. Odds Ratio 95%

Confidence Interval P INH mono-resistant TB 1.94 1.16,
3.24 0.01 Age (per 1-year increase) 1.03 1.02, 1.04
,0.001 Female sex 1.07 0.82, 1.38 0.63 Race Asian vs.
Black 1.15 0.65, 2.03 0.64 Coloured vs. Black 0.96 0.49,
1.88 0.91 White vs. Black 0.41 0.06, 2.94 0.37
22. The nature of the beast: emerging

molecular information about
mycobacteria
OA-467-06 Selective reconstitution of
interferon-gamma gene function in T cells
suffices to restore protection against
mycobacterial lung infection
P Blank,1 K Borst,1 E Grabski,1 T Frenz,1 U Kalinke1
Twincore Institut for Experimental Infection Research,
Hannover, Germany. e-mail: Patrick.blank@twincore.de
Background: In the lung the control of mycobacterial

infection is dependent on an inter-cellular crosstalk of
infected and uninfected cells as well as recruitment of
inflammatory cells. Activation and recruitment of cells is
mediated by chemokines and cytokines. Amongst cytokines, IFN-gamma has previously been shown to be
crucial to mount protective responses against mycobacterial infection in mice and men. IFN-gamma has been
reported to facilitate activation of infected macrophages,
to induce apoptosis as well as to initiate inflammatory
cell recruitment. A combination of these mechanisms is
essential to ensure protection against mycobacterial
infection. Various cells of the innate and adaptive
immune system are able to express IFN-gamma. However, it remained unclear what effect the IFN-gamma
expression of the different cell-types has on antimycobacterial immunity.
Conclusion: To analyze the function of cell-type specific
IFN-gamma we developed a novel mouse model in which
the IFN-gamma gene is conditionally inactivated and can
be reconstituted by Cre-dependent recombination (IFNgammaST/ST). For this study we intercrossed mice
expressing Cre specifically in T cells (CD4-Cre) with
IFN-gammaST/ST. Indeed, CD4-CreIFN-gammaST/ST
mice showed selective reconstitution of IFN-gamma gene
function only in T cells. Intra-tracheal infection of such
mice with Bacillus Calmette-Guerin (BCG) revealed that
unlike IFN-gammaST/ST and conventional IFN-gammako mice, in mice with reconstituted IFN-gamma gene
function in T cells protection was restored. Interestingly,
the cellular infiltration into the lung of BCG infected
CD4-CreIFN-gammaST/ST mice resembled that of
IFN-gammako and IFN-gammaST/ST mice with elevated numbers of eosinophils and neutrophils. These data
indicated that IFN-gammaexpression of Tcells alone was
not sufficient to promote normal inflammatory cell
recruitment. These results have implications for the
development of new tuberculosis vaccines.
22. The nature of the beast: emerging

molecular information about
mycobacteria
OA-468-06 Genomic and proteomic analysis of
Mycobacterium tuberculosis lineage 7 strains
from Ethiopia
S Yimer,1 A Namouchi,1 T Riaz,1 E D Zegeye,2
C Holm-Hansen,3 G Norheim,3 A Aseffa,4,5 T Tnjum1
1
Oslo University Hospital, Oslo, 2University of Bergen, Oslo,
3
Norwegian Institute of Public Health, Oslo, Norway;
4
Armauer Hansen Research Institute, Addis Ababa, Ethiopia;
5
Armauer Hansen Research Institute, Oslo, Norway. e-mail:
yimsolo@yahoo.com
Background: Understanding the various factors driving

the tuberculosis (TB) pandemic is essential to achieve the
Global Plan for TB control and elimination. Recently, a
new phylogenetic M. tuberculosis (Mtb) lineage (lineage
7) has been identified in Ethiopia, which is associated
with prolonged delay among patients and grows slowly
in vitro compared to other lineages. Given the potential
implications for pathogenesis and virulence of the
bacteria, genomic and proteomic studies on M. tuberculosis are warranted. The aim of this study was to
delineate the molecular characteristics of M. tuberculosis
lineage 7 strains by analysing genomic and proteomic
profiles.
Design/Methods: The BACTEC MGIT 960 culture
system was used to cultivate 30 M. tuberculosis lineage
7 isolates. DNA was extracted and whole genome
sequencing was performed using the Illumina Hiseq2000. Overall, after aligning the generated reads for
each strain to the genome sequence of the M. tuberculosis
H37Rv strain, the coverage rate was above 300 times.
High resolution Q-Exactive mass spectrometer was used
to achieve proteome characterization. MaxQuant software was used to define the amounts of the various
proteins present, while post-translational modifications
(PTMs) were identified by Proteome Discoverer and
manual inspection.
Results: More than 800 mutations or single nucleotide
polymorphisms (SNPs) specific to M. tuberculosis
lineage 7 strains were identified. Compared to other
lineages, M. tuberculosis lineage 7 strains exhibited a
high number of mutations in genes involved in carbohydrate transport and metabolism, transcription, energy
production and conversion. Notably, low frequencies of
mutations were observed in genes involved in amino
acid, lipid and coenzyme transport and metabolism, post
translational modification, protein turnover and chaperone. Several interesting proteins, such as transcription
factors, membrane proteins and proteins involved in
pathogenesis of the bacteria were identified in lineage 7
strains. We also identified novel site-specific PTMs.
Conclusion: This study generated novel knowledge
regarding the genomic and proteomic profile of M.
tuberculosis lineage 7 strains. Further research is
essential to understand the consequences of the differential frequencies of mutations observed in the main gene
S439
groups of lineage 7 strains. Taken together, these studies

have generated significant novel knowledge on a new
lineage of M. tuberculosis, which can elucidate how these
strains can have fitness for survival despite their
relatively slow growth.
OA-469-06 Single cell elucidation of

mycobacterial replication dynamics
S Sampson,1,2 J Mouton,1 S Helaine,2 D W Holden2
1
Stellenbosch University, Cape Town, South Africa; 2Imperial
College of London, London, UK. Fax: (27) 21 938 9863.
e-mail: jomien@sun.ac.za
Background: In most Mycobacterium tuberculosis (M.

tuberculosis)-infected individuals, the infection persists
in a latent, asymptomatic state. Therapies that aim to
eliminate tuberculosis (TB) should target these dormant
organisms, since these could resume replication to cause
active disease. It is thought that persistent mycobacteria
exist as a small, viable, but non-replicating (VBNR)
population. Currently the majority of drug therapies
target actively replicating (AR) bacteria, however persister bacteria comprise a subpopulation of bacteria that
is recalcitrant to antibiotic treatment. Little is known
about persistent mycobacteria, since they are very
difficult to isolate. However, recent work has developed
and exploited a technique termed Fluorescence Dilution
(FD), which showed for the first time that the internalization of Salmonella by macrophages induced the
formation of VBNR populations. This approach exploits
2 fluorescent reporters; a constitutive reporter allows the
tracking of viable bacteria, while an inducible reporter
enables the measurement of bacterial replication. Importantly, we have successfully adapted and optimized this
system for use in mycobacteria. The aim of the proposed
work was to exploit the FD technology to assess whether
host exposure induced the formation of VBNR persister bacteria in the model organism M. smegmatis as well
as M. tuberculosis, to broaden the currently limited
knowledge of the biology of these bacteria.
Methods: We applied the dual reporter system to
interrogate replication dynamics of M. smegmatis or
M. tuberculosis during macrophage infection.
RAW264.7 macrophages were infected with M. smegmatis or M. tuberculosis reporter strains and samples
were analysed using flow cytometry and confocal
microscopy.
Results: Characterization of intracellular M. smegmatis
revealed that while the majority of bacteria were non-or
slowly replicating, an unexpected sub-population of
apparently dividing M. smegmatis could be detected.
For M. tuberculosis, FD technology identified a distinct
subpopulation of non-growing mycobacteria. Furthermore, the presence of VBNR and AR mycobacteria
within the same macrophage after 72 hours of infection
were observed.
Conclusion: These results demonstrate the successful
application of the novel dual fluorescent reporter system
both in vitro and in macrophage infection models to
S440
provide a window into mycobacterial population heterogeneity.
which may contribute to increased acquisition of

resistance in patients treated with an MDR regimen.
OA-470-06 Next generation sequencing reveals

hidden ethionamide resistance as a likely driver
towards resistance beyond XDR-TB in a highburden population
OA-471-06 ASAP, an automated sequence

analysis pipeline for whole genome
sequencing data, initial applications in
M Klopper,1 G Hill-Cawthorne,2 R Van Der Merwe,1

S Sampson,1 E Streicher,1 A Abdallah,3 R M Warren,1,3
P Arnab3 1Biomedical Sciences, Stellenbosch University,
Cape Town, South Africa; 2University of Sydney, Sydney,
NSW, Australia; 3King Abdullah University of Science and
Technology, Thuwal, Saudi Arabia. Fax: (27) 21 938 9863.
e-mail: marisat@sun.ac.za
R Van Der Merwe,1 R M Warren,1 A Dippenaar,1

M De Vos,1 P Van Helden,1 A Abdallah,2 P Arnab,2
S Sampson1 1Stellenbosch University, Cape Town, South
Africa; 2King Abdullah University of Science and Technology,
Thuwal-jeddah, Saudi Arabia. e-mail: rvdm@sun.ac.za
Background: In the Eastern Cape Province, South Africa,

90% of XDR-TB strains are of the Atypical Beijing
genotype and are resistant to 10 or more anti-TB drugs.
This wide-spread hyper-resistance prompted next generation sequencing investigation of molecular features
which may predispose these strains to acquire drug
resistance while retaining virulence.
Design/Methods: Atypical Beijing genotype isolates (n
29), ranging from pan-susceptible to beyond-XDR
(known resistance to 11 drugs), were selected for whole
genome sequencing (WGS) on an Illumina platform. In
addition, RNA sequencing was done at mid-log growth
in order to relate chromosomal mutation to gene
expression.
Results: Phylogenetic analyses predicted independent
evolution of Atypical Beijing genotype strains from
MDR-TB to hyper-resistance. Two independent lineages
were identified, each with unique genetic features
including but not limited to drug-resistance causing
mutations. Variant analysis showed an unexpectedly
large number of differences in isolates from the same
patient, vs. instances of low variation between epidemiologically unlinked isolates. RNA-Seq confirmed upregulation of the inhA operon in isolates with an inhA
promoter mutation. Furthermore, we identified upregulation of the toxin-antitoxin (TA) vapBC22 (implicated in persistence), associated with a frameshift
mutation in vapC22. Interestingly, ethA mutations
causing ethionamide (ETH) resistance were found in all
Atypical Beijing genotype isolates, including those
diagnosed as pan-susceptible based on isoniazid (INH)
and rifampicin (RIF) susceptibility.
Conclusion: XDR continues to evolve from an endemic
MDR strain, implying failure to effectively treat MDRTB patients. The deeply rooted mutations conferring
resistance to INH and streptomycin suggest initial
evolution of resistance prior to the introduction of RIF
in the 1970s. Our findings of variable rates of change
challenge the definition of recent transmission by WGS.
Using a combination of WGS and RNA-Seq we
demonstrate the relationship between DNA mutation
and gene expression relating to INH resistance, as well as
in a TA system implicated in persistence. Identification of
an ethA mutation in all Atypical Beijing isolates
irrespective of their 1st line drug susceptibility phenotype
suggest that this lineage is intrinsically resistant to ETH,
Background: The utility of whole genome sequencing

(WGS) in molecular diagnostics is increasing as sequencing costs diminish. Access to a pathogens genome
sequence can help guide clinical intervention through
molecular genotyping and subsequent phenotype prediction bases on existing genotype:phenotype models.
However, processing whole genome sequence data is a
highly complex process with no existing standardized
protocols. As such, there is a need for a simple and
standardized pipeline for WGS data analysis and variant
detection to enable users of all skill levels to reliably
extract pertinent information from WGS data to answer
biologically and clinically relevant questions. Here we
describe ASAP, an automated sequence analysis pipeline
for whole genome sequencing data, and show its
applications toward Mycobacterium tuberculosis WGS
analysis and drug resistance prediction.
Design/Methods: ASAP makes use of open source tools
for automated quality control, reference alignment and
variant calling, followed by novel tools for variant
annotation and phenotype prediction using compiled
databases. Variant detecting accuracy was validated by
comparing it to targeted amplification and Sanger
sequencing. The resistance prediction model of ASAP
for M. tuberculosis was compared to that of culture
based DST as well to other resistance prediction software
based on WGS data.
Results: The accuracy of ASAP matches that of targeted
amplification and Sanger sequencing for regions which
are not exceptionally repetitive. ASAP shows superior
variant detection accuracy and resistance profiling to
existing software. The resistance prediction model used
in ASAP for M. tuberculosis resistance prediction has a
high agreement with DST for first line drugs but with
diminished performance for 2nd line drug resistance
prediction.
Conclusion: ASAP is the first fully automated pipeline for
accurate variant detection in WGS data. Its performance
in M. tuberculosis resistance profiling closely matches
that of DST for resistances with well characterised
molecular markers. The utility of WGS in combination
with tools such as ASAP is set to expand as genotype:phenotype models are continually improved.
S441
OA-472-06 Host targeted therapy for

tuberculosis via aerosol administration of
siRNA targeting IL-10 and STAT3
R Upadhyay,1 A Sanchez-Hidalgo,1 J Arab,1 C Wilusz,1
A Lenaerts,1 M Gonzalez-Juarrero1 1Colorado State
University, Fort Collins, CO, USA. e-mail:
mercedes.gonzalez-juarrero@colostate.edu
Background: Eradication of tuberculosis (TB) depends

on the development of shorter and more effective
treatment regimens, including novel treatment alternatives combined with classical TB therapies. Host targeted
therapies [HDT] in combination with current therapies
for TB has emerged as one approach towards this goal. In
this study we show that it is possible to modulate the lung
immune environment of the Mycobacterium tuberculosis
chronically infected host and increase its own antimicrobial capacity against drug tolerant bacilli. The end
result of this therapy is increased capacity by the host to
eliminate the drug tolerant bacilli remaining in the lungs
during or after standard chemotherapy. Thus, the study
presented below aims at identifying HDT for TB via
aerosol administration of siRNAs.
Design/Methods: Persistent expression of the cytokine
IL-10 and the signal transducer and activator of
transcription (STAT) 3 during a chronic infection with
Mtb impairs the antimicrobial capacity of the host. Mice
were treated during two weeks with isoniazid and
rifampin. Then chemotherpay was stopped and mice
received aerosols of siRNA targeting il-10 and/or stat3.
After siRNA therpay we monitored the pulmonary drug
tolerant bacterial load and expression of antimicrobial
effector products (NO, NADPH, NOS-2 and Arg-1). In
addition, we also monitored changes in the inflammatory
course caused by the therapy by recording the histological outcome after chemo- siRNA therapy.
Results: When chronically infected mice received a
combined therapy regimen consisting of two weeks of
anti-TB chemotherapy followed by intrapulmonary
aerosol delivery of siRNAs targeting il-10 and/or stat3
the pulmonary drug tolerant bacterial load was significantly reduced when compared to control mice receiving
the same chemotherapy but not siRNA/or siRNA sham
treatment (Figure).Furthermore, the pulmonary antimicrobial capacity in these mice was further enhanced as
evidenced by higher expression of antimicrobial effector
molecules and decreased arginase activity when compared to control groups.
Conclusion: As a proof of concept, here, it is shown that
a successful targeting of the host IL-10-STAT3 pathway
via siRNAs aerosol delivery can modulate the lung
immunity to enhance its own antimicrobial capacity and
eliminate more than 90% of the pulmonary drug tolerant
bacterial burden.
OA-473-06 Use of plasma metabolomics at

diagnosis to identify metabolic pathways
associated with pulmonary tuberculosis
clearance: a pilot study
J Frediani,1 E Chong,1 J Alvarez,1 T Yu,1 D Jones,1
N Tukvadze,2 H Blumberg,1 T Ziegler1 1Emory University,
Atlanta, GA, USA; 2National Center for Tuberculosis and
Lung Disease, Tbilisi, Georgia. e-mail: tzieg01@emory.edu
Background: Knowledge of metabolic pathways associated with clearance of Mycobacterium tuberculosis in TB

disease may provide pathophysiologic insight and
identify potential biomarkers. We used plasma metabolomics to identify metabolic pathways associated with
pulmonary TB clearance.
Design/Methods: We studied 61 adults with pulmonary
TB disease (18 with multidrug-resistant TB [MDR-TB])
enrolled in a double blind, controlled, randomized trial
of vitamin D (Vit D) supplementation in Tbilisi, Georgia.
Subjects entered the trial within 7 days of initiation of
conventional anti-TB drug therapy (retreatment cases
were excluded). M. tuberculosis clearance from sputum
cultures (LJ solid media) obtained 8 weeks after entry
was determined. Plasma was obtained at baseline (before
initiation of vitamin D3 or placebo) and metabolomics
analysis performed using high-resolution liquid chromatography mass spectrometry (LC-MS). Individual regression models were analyzed for each detected metabolite
with sputum culture conversion at 8 weeks (wks) as the
independent variable and metabolite intensity as the
S442
dependent variable, adjusted for MDR-TB status,

diabetes status, body mass index, sex, age, treatment
group (vit D vs. placebo), and plasma 25-hydroxyvitamin D level. Statistically significant (P , 0.05) metabolites were subsequently analyzed with a high-throughput metabolomics pathway analysis program
(Mummichog).
Results: After 8 wks of anti-TB drug treatment, 52
subjects (85%) had a negative sputum culture and 9 (15
%) remained culture-positive. Of 5715 metabolites
detected, 251 differed (P , 0.05) between subjects who
had achieved sputum culture conversion to negative
versus those that remained culture positive (130 metabolites were decreased and 121 were increased, respectively). Nine specific metabolic pathways were significantly different between the two groups, including
cytochrome P450 drug metabolism (P 0.013), drug
metabolism related to other enzymes (P 0.001),
glutamate metabolism (P 0.003), aspartate and
asparagine metabolism (P 0.003), amino-sugar metabolism (P 0.021), de novo fatty acid biosynthesis (P
0.02) and leukotriene metabolism (P 0.027).
Conclusion: In patients with pulmonary TB, plasma
metabolomics analysis obtained early after TB diagnosis
identified metabolic pathways related to drug metabolism, amino acid/nutrient metabolism and inflammation
that differentiated those who did or did not become
sputum culture-negative for M. tuberculosis after 8 wks
of anti-TB drug treatment.
OA-474-06 A web-based interface to explore

and analyze Mycobacterium tuberculosis
whole genome sequence data
M Farhat,1 J Honaker,2 C Choirat,2 T Ioerger,3 M Murray4
1
Massachusetts General Hospital, Boston, MA, 2Harvard
University, Boston, MA, 3Texas A & M University, College
Station, TX, 4Harvard Medical School, Boston, MA, USA.
e-mail: marhat@gmail.com
Background: The study of genetic diversity, drug

resistance and other clinical disease phenotypes through
Mycobacterium tuberculosis whole genome sequences
promises to improve our understanding of M. tuberculosis biology and management of TB Disease. Intuitive
and open tools for data analysis can accelerate research
and diagnostic development, and aid clinicians and
laboratory personnel in interpreting these complex data.
We sought to develop a web based interpretation and
analysis tool for M. tuberculosis whole genome sequences.
Methods: We developed a web based tool using a core
dataset of well characterized 123 strains with whole
genome sequences, and 1397 strains with sequences from
28 known or putative drug resistance loci. We also
developed a data pipeline that allow users to contribute
to and expand this dataset. We trained a machine
learning algorithm using the statistical language R that
enables the prediction of drug resistance based on this
data and drug resistance phenotypes, and developed a
mapping tool to display the geographical distribution of
genetic variation, drug resistance, and lineage among

other features.
Results and Conclusion: The interface enables the
prediction of drug resistance phenotypes to 12 drugs
from either whole genome sequences or a minimal set
genetic variants from 21 genes or promoter regions. The
interface also allows the geographic and statistical
exploration of sequence, drug resistance, lineage data
and others that the user may choose to input. To our
knowledge, this tool is the first open access analysis tool
of its type for M. tuberculosis genetic data.
23. TB in children, including MDR

OA-475-06 High frequency of pulmonary nontuberculous mycobacteria in HIV-infected
children with presumed tuberculosis in SouthEast Asia
B Dim,1 S Goyet,1 L Borand,1 T N L Nguyen,2 T H Pham,2
S Godreuil,3,4 V Ung,5 O Marcy1,6 1Institut Pasteur in
Cambodia, Phnom Penh, Cambodia; 2Pham Ngoc Thach
Hospital, Ho Chi Minh City, Viet Nam; 3INSERM U1058,
Montpellier, 4CHU Arnaud de Villeneuve, Montpellier,
France, 5National Pediatric Hospital, Phnom Penh, Cambodia;
6
ISPED - INSERM U897, Bordeaux, France. Fax: (855) 2372
5606. e-mail: dbunnet@pasteur-kh.org
Background: Epidemiology, clinical impact and therapeutic management of Non-Tuberculosis Mycobacteria

(NTM) in human immunodeficiency virus (HIV)-infected children are poorly studied in high tuberculosis (TB)
burden countries. We studied NTM infections in HIV
children suspect of TB in South-East Asia enrolled in the
ANRS 12229 PAANTHER 01 study.
Design/Methods: HIV-infected children aged613 years
with a suspicion of intra-thoracic TB were enrolled after
parental informed consent in pediatric HIV clinics in
Cambodia and Viet Nam, from April 2011 to May 2014.
Smear microscopy, Xpert MTB/RIF, and liquid culture
(MGIT) were performed on respiratory and stool
samples. NTM species identification was performed
using GenoType Mycobacterium CM/ASw (Hain). Children were initiated on TB, NTM, and antiretroviral
(ART) treatments per national recommendations and
followed 6 months. We identified factors associated with
NTM infection by logistic regression.
Results: We enrolled 250 children with 115 (46.0%)
female, 112 (44.8%) on ART at inclusion, median age
6.9 years (IQR: 3.6 9.5), median weight for age Z-score
2.5 (IQR: 3.4 - 1.9), median CD4 15% (IQR: 3 - 25).
Of 245 children with samples performed, 18 (7%) had
culture-confirmed TB and 41 (17%) NTM (4 NTM/TB
co-infections). Factors associated with NTM were CD4
count,5% (P 0.037) and absence of ART (P 0.027)
in univariate analysis, absence of ART only in multivariate analysis. Species identification in 74 of 88 NTM
positive samples found 6 M. simiae, 7 M. scrofulaceum,
12 M. fortuitum, 42 M. avium complex (MAC), 2 M.
gordonae, 2 M. interjectum, 2 M. kansasii, 1 M.
lentiflavum. 5 NTM-positive children had smear-positive

samples, including 1 NTM/TB co-infection; 3 of 4 NTM/
TB co-infections had a positive Xpert. Of 41 NTMpositive children, 28 received TB treatment, 17 received
NTM-specific treatment, initiated 15 days (IQR: 7 53)
after inclusion. At the end of the study (M6), 7 NTM
positive children had died, 1 had withdrawn from the
study, 33 were followed -median weight gain from
inclusion 2.1 (IQR: 1.0 3.0) kgs.
Conclusion: In HIV-infected South-East Asian children
suspect of TB, 17% had culture confirmed NTM on
respiratory and stool samples. Despite important NTM
species diversity, MAC was the most common identified
cause of lung disease in HIV-infected children. In the
absence of rapid diagnostic test, and because of the high
fatality rate, diagnosis and management of NTM are
major issues in these children.
OA-476-06 Tuberculosis among adolescents

and children along the cascade of HIV care in
two regions of Ethiopia
D J Dare,1 W Abebe,2 K T Gari,3 M Aseresa Melese,1
P G Suarez,4 A Ruff5 1Management Sciences for Health,
HEAL TB Project, Addis Ababa, 2Addis Ababa University,
Addis Ababa, 3Hawassa University, Hawassa, Ethiopia;
4
Management Sciences for Health, Center for Health
Services, Arlington, VA, 5Johns Hopkins University, School of
Public Health, Baltimore, MD, USA. Fax: (251) 11 372 4473.
e-mail: degujerene@gmail.com
Background: Understanding the magnitude of TB among

adolescents and children has both public health and
clinical significance. Since adolescents are socially more
active than younger children and adults, their potential
to transmit to their peers is high. Also, adolescents in
general have more tendency to stop or interrupt
treatment, which makes them important targets for
prevention of drug resistant tuberculosis. On the other
hand, younger children are at increased risk acquiring TB
from adults and adolescents. HIV co-infection further
compounds this challenge. However, there is limited data
on the magnitude of tuberculosis in adolescents and
younger children living with HIV. We report results from
a cohort of HIV infected adolescents and children treated
and followed at eight health facilities in two regions of in
Ethiopia. Our objective was to determine the burden of
tuberculosis among adolescents and children living with
HIV along the cascade of chronic HIV care.
Methods: We conducted a retrospective cohort study in
eight health facilities in two regions of Ethiopia. The
study population constituted children (0-9 yr) and
adolescents (age 10-19 years) enrolled in chronic HIV
care between January 2005-December 31 2013. Trained
study nurses assisted by site data clerks retrieved
information from patient charts and registers. Patients
age category was the main predictor variable. Covariates
included baseline WHO clinical stage, CD4, and IPT.
Occurrence of new TB during follow up was the main
outcome variable. We used SPSS for data analysis. We
calculated TB incidence rates as new cases per 100
person-years of observation (PYO). We used Cox
S443
regression analysis to control for confounders. The

national research ethics committee approved the study
protocol.
Results: Of 2058 participants (1072 adolescents and 986
children) included in this analysis, 54% were girls and
90% came from urban areas. Nearly a half (48.5%) were
in advanced WHO stage at presentation. Previous history
of TB was present in 13.9% of the participants. A further
12.1% had history of TB at enrollment. A total of 243
cases (142 in adolescents and 101 in younger children) of
TB were reported during a pre-ART follow up period of
2422 PYO, making the overall incidence rate 10.03
(95%CI; 8.83, 11.35) per 100 PYO for the total cohort.
The unadjusted incidence rate was higher in adolescents
16.32 (95%CI; 13.75, 19.24) than the rate in younger
children (6.51 [95%CI; 5.29, 7.90]; P , 0.0001). After
adjusting for WHO stage and INH use, adolescents were
twice more likely to have TB diagnosis before ART. ART
was associated with 85% reduction in TB incidence in
both groups but the adolescent age group continued to
have significantly higher rates even after ART.
Conclusion: The burden of TB was high both in children
and adolescents but it was even higher in the adolescent
age group, suggesting the need for targeting this age
group for TB prevention. On the other hand, the lower
TB rate in the younger age group could be due to poorer
diagnostic capacity, warranting the need for better
diagnostic tools for TB in the younger children.
Therefore, both adolescents and children living with
HIV should be prioritized for TB prevention interventions.
OA-477-06 Risk factors for active tuberculosis

among child contacts of recently diagnosed
adults with sputum smear-positive
tuberculosis
U Egere,1 T Togun,1 S Abdoulai Daffay,1 J Otu,1
B Kampmann1,2 1Vaccines and Immunity Theme, Medical
Research Council Unit, Fajara, The Gambia; 2Imperial
College, London, UK. e-mail: uegere@mrc.gm
Background: Like in many settings, TB in children is

under-diagnosed and poorly managed in The Gambia.
Although WHO recommended, contact tracing is not
done because of limited resources and lack of expertise.
Better targeted screening might be achievable if particular risk factors are known. We identified risk factors for
TB disease among child contacts of adults recently
diagnosed with TB, which could guide such targeted
strategies for national programs.
Design/Methods: Households of consenting adults diagnosed with sputum smear positive TB in The Gambia
were visited, where a symptom screening questionnaire
and Tuberculin skin testing (TST) were administered to
all children ,15 years. TST response of .10mm was
considered positive. Presence of cough of at least 2 weeks
duration and at least one of other symptoms such as
weight loss, failure to gain weight, fever, night sweats,
was regarded as suggestive of TB. All children with
symptoms suggestive of TB and/or positive TST result
S444
had clinical examination, chest X-ray and sputum

examination at the childhood TB clinic at the MRC.
Children with active TB were commenced on anti-TB
treatment according to national guidelines. Random
effects logistic regression model taking into account
clustering of contacts within households was used to
access the relationship between TB disease and risk
factors.
Results: Of 3565 child contacts screened in the community, 736 (20.6) were symptomatic and/or TST positive of
which 702 had clinical and chest X-ray examination in
the clinic. Sputum examination (smear, GeneXpert and
culture) was then performed for all children. 44 of 702
(6.3%) were diagnosed with active TB. Record of BCG
vaccination was not associated with protection against
active TB. Odds of active TB increased with increasing
gradient of exposure to the index case (OR 4.0 {1.16
13.8} P 0.028 for same room, OR 7.2 {1.83 28.2} P ,
0.0005 for same bed). Positive TST was associated with
increased odds (OR 15.2{5.7 40.8} P ,0 .0005 for TST
10-14mm) but odds did not increase further as TST
increased beyond 14mm.
Conclusion: Child contacts in closest proximity to adult
TB cases have the highest odds to develop active TB.
Though not a new finding, this is significant in The
Gambia where relatively large numbers of contacts live
in separate house to the index case but would still be
eligible for contact tracing. In the face of limited
resources contact tracing could target the closest contacts
only for the highest yield.
Results: 292 children, 144 (49.3%) males, with a median

age of 34 months (IQR 14-84 months) with bacteriologically confirmed TB were recorded; 276 were cultureconfirmed and 16 only confirmed by Xpert. Of cultureconfirmed cases, 241 (87.3%) were pan-susceptible, 20
(7.2%) were MDR, 8 (2.9%) were RIF-resistant/INHsusceptible and 7 (2.5%) were INH-resistant/RIF-susceptible. Of the 16 Xpert-positive only cases (culturenegative), 14 were RIF-susceptible, 2 RIF-resistant and 2
had indeterminate results. Of 290 (99.3%) children
tested, 30 (10.3%) had any RIF-resistance. Of 26 RIFresistant cases tested, ofloxacin and amikacin resistance
were present in 5 (19%) and 4 (15%), respectively. Of
280 (95.9%) children tested, 45 (16.1%) were HIVinfected. Compared to previous surveillance periods, RIF
mono-resistance has increased steadily over each period
from 0 to 2.9%, now surpassing INH mono-resistance.
The prevalence of MDR-TB has stabilised and the HIV
prevalence has decreased significantly from a peak of
29% observed in 2007-2009, to 16% during the last two
surveillance periods.
Conclusion: The steady increase in RIF-resistant/INHsusceptible cases is concerning, while the prevalence of
MDR-TB appears to be stabilising. These data indicate
recent tranmission of DR-TB strains to young children,
reflecting ongoing transmission. The increasing circulation of RIF mono-resistance has important implications
for both treatment and preventive therapy in cases
confirmed to be resistant by Xpert without further
culture and DST.
OA-478-06 Surveillance of childhood

tuberculosis drug resistance in Cape Town,
South Africa: increasing rifampicin monoresistance
OA-479-06 The pharmacokinetics and safety of

ofloxacin for drug-resistant tuberculosis in
children
H S Schaaf,1 E Walters,1 A Hesseling,1,2 C Rautenbach,2,3

C Bosch,1 A Garcia-Prats1 1Stellenbosch University, Cape
Town, 2Tygerberg Hospital, Cape Town, 3National Health
Laboratory Service (NHLS), Cape Town, South Africa. Fax:
(27) 21 938 9138. e-mail: hss@sun.ac.za
Background: Consecutive 2-year tuberculosis (TB) drug

resistance surveys amongst children have been ongoing in
Cape Town, South Africa since March 2003 to evaluate
trends of drug resistance and HIV co-infection amongst
children with culture-confirmed TB. We found a decline
in the prevalence of TB-HIV co-infection and multidrugresistant TB (MDR-TB) in the most recent reporting
period (2011-2013) compared to previous peak occurrences.
Design/Methods: A prospective 2-year surveillance
(March 2013-Feb 2015) of drug resistance in all
bacteriologically confirmed child TB cases (,13 years)
at Tygerberg Hospital, Cape Town, South Africa. Initial
drug susceptibility testing (DST) is done for isoniazid
(INH) and rifampicin (RIF) on a single specimen from
each child with culture-confirmed TB, and RIF DST if
only the Xpert MTB/RIF is positive. Second-line DST
was undertaken if culture-confirmed and resistant to RIF.
HIV status was recorded.
A Garcia-Prats,1 H R Draper,1 K Dooley,2 J Seddon,3

S Thee,4 H Mcilleron,5 H S Schaaf,1 A Hesseling1
1
Desmond Tutu TB Centre, Stellenbosch University,
Tygerberg, South Africa; 2Johns Hopkins University School of
Medicine, Baltimore, MD, USA; 3Imperial College London,
London, UK; 4Charite, Universitatsmedizin Berlin, Berlin,
Germany; 5University of Cape Town, Cape Town, South
Africa. e-mail: garciaprats@sun.ac.za
Background: Ofloxacin is a fluoroquinolone widely used

for the treatment of multidrug-resistant tuberculosis
(MDR-TB). Pharmacokinetic (PK) and safety data in
children are limited, and it is not known whether the
recommended pediatric dose of 15-20mg/kg achieves
exposures reached by adults with TB. After a standard
800mg dose in adults, median PK measures are: area
under the concentration-time curve (AUC)0-24
103lg*h/mL., maximum serum concentration (Cmax)
10.5 lg/mL, and oral clearance (CL/F) 0.12 L/h/kg.
Design/Methods: PK and safety of ofloxacin were
assessed in children ,15 years of age routinely receiving
treatment or preventive therapy for MDR-TB. To assess a
20mg/kg dose, plasma samples were collected pre-dose
and at 1, 2, 4, 8 and either 6 or 11 hours post-dose.
Assays were performed using HPLC MS/MS. PK
parameters were calculated by non-compartmental
analysis. Children with MDR-TB disease were followed
longitudinally in the study for safety, with clinical and

laboratory assessments 1-2 monthly. Adverse events were
graded using standardized DAIDS criteria, and attribution to ofloxacin assessed.
Results: Of 85 children (median age: 3.4 years) 11 (13%)
were HIV-infected and 14 (18%) were underweight.
Mean (range) Cmax and AUC0-8 were 8.97lg/mL (2.4714.40) and 44.2lg*h/mL (12.1-75.8), respectively. In
multivariable linear regression, AUC0-8 was reduced by
1.54lg*h/mL for each additional year of age (P 0.026)
and increased by 0.68lg*h/mL for each additional kg of
body weight (P 0.009). Controlling for age and weight,
no other variables contributed to the model. Because of
, 3 time points in the elimination phase in some patients,
the half-life, CL/F, and AUC0-24 were estimated only in
72 participants. The mean (range) AUC0-24 was
66.7lg*h/mL (18.8-120.7) and half-life was 3.47h
(1.89-6.95). The mean (range) CL/F was 0.33L/h/kg
(0.10-1.04), and more rapid clearance was seen in
younger children (P , 0.001). Among 46 patients with
MDR-TB disease contributing 23.8 years of observation
time, there were no Grade 3 or 4 events potentially
related to ofloxacin.
Conclusion: Ofloxacin was safe and well tolerated in
children with MDR-TB. In these children, Cmax values
approached those in adults after an 800 mg oral dose,
however AUCs were well below reported adult values
suggesting further dose modification in children may be
required to optimize regimens for MDR-TB treatment in
children.
OA-480-06 Developing Swazilands first public

pediatric MDR-TB contact clinic
G Mtetwa,1 P Ustero,1,2 R Golin,1,2 J Glickman,1
P Swamy,2 B Tsabedze,1 M Hlatshwayo,1 A Mandalakas2
1
Baylor College of Medicine Childrens Foundation Swaziland, Mbabane, Swaziland; 2Baylor College of
Medicine and Texas Childrens Hospital, Houston, TX, USA.
Fax: (268) 2404 0214. e-mail: golin@bcm.edu
Background and challenges to implementation: Baylor

College of Medicine Childrens Foundation Swaziland
(BCMCF-SD) provides comprehensive HIV-TB services,
including a standardized approach to pediatric TB
prevention, diagnosis, treatment, and contact tracing.
The high TB incidence (1,382/100 000) together with the
high rates of MDR-TB in Swaziland (7.7% of new cases,
34% of retreatment cases) led BCMCF-SD to develop the
countrys first public pediatric MDR-TB contact clinic.
Intervention: To meet the needs of MDR-TB exposed
children and adolescents, BCMCF-SD established a
monthly MDR-TB Contact Clinic Day. Contacts of
household MDR-TB index cases are identified and
referred to BCMCF-SD. Participating families receive
reminder phone calls prior to their clinic day and receive
transport reimbursement for their first visit. Children
and adolescents are requested to attend clinic with a
caregiver; patients and caregivers attend an education
session on MDR-TB infection. TB screening and risk
assessment is completed by all attendees. Sputum
analysis (GeneXpert [GXP] and culture) and radiological
S445
studies are available at no charge to support childrens

evaluation. The majority of patients will be followed for
1 year, with select patients requiring up to 2 years of
clinical monitoring.
Results and lessons learnt: Since February 2015, a total
of 25 children and adolescents were enrolled. The mean
age was 6.6 years, with 40% (10/25) under 5 years old.
80% (20/25) were HIV negative, 20% (5/25) were HIV
positive. On average, contacts attended clinic 17.7
months after the household index case started MDRTB treatment. TB symptom screening was positive in
32% (8/25) of contacts. Of those, 75% (6/8) completed
GXP testing; GXP testing did not detect MTB in any
specimens. All patients with a positive TB screen had a
normal chest radiograph. Drug sensitivity test results of
the index cases could only be identified for 16% (4/25) of
contacts.
Conclusions: Clinicians and parents welcomed referral to
BCMCF-SDs pediatric MDR-TB contact clinic as they
recognize the importance of monitoring this vulnerable
population. Strategies are needed to reduce the financial
barriers imposed by transport and radiological fees. The
lack of national and international policy concerning
preventive therapy for children and adolescents limits the
full benefit of this program.
OA-481-06 Survival of children with HIV

infection and drug-resistant TB in three
provinces in South Africa, 20052010
S Morris,1 E Hall,1 S Smith,1 N Ismail,2 R Odendaal,3
N Ndjeka,4 H Menzies,1 M Van Der Walt3 1U.S. Centers for
Disease Control, Atlanta, GA, USA; 2National Health
Laboratory Services, Johannesburg, 3South African Medical
Research Council, Cape Town, 4Ministry of Health, Pretoria,
South Africa. e-mail: feu3@cdc.gov
Background: Few studies describe treatment outcomes of

pediatric patients co-infected with HIV and drugresistant tuberculosis (DR-TB). The objective of this
secondary analysis of a pediatric DR-TB study conducted
in three South African provinces was to assess the
survival of these patients based on HIV status and other
risk factors.
Design/Methods: Demographic, clinical, and laboratory
data from pediatric patients treated for DR-TB during
20052010 were retrospectively collected. Multivariable
logistic regression was used to estimate adjusted odds
ratios (aOR) and 95% confidence intervals (CI) comparing the odds of experiencing a poor outcome (death or
treatment failure) to a positive outcome (cure or
treatment completion), while controlling for HIV status,
sex, age, province, year of TB diagnosis, resistance
pattern, weight and 6-month culture growth among
patients diagnosed before 2009. A time-to-event analysis
was conducted using days from start of treatment to
outcome. Bivariate and multivariate stratified Cox
proportional hazard models were used to estimate
hazard ratios.
Results: Of 685 eligible participants, 55% were HIVinfected, 52% were underweight, and 42% were male.
There were 434 participants with documented HIV
S446
status and outcomes; cured or completed treatment

(57%), died (16%), failed (1%), defaulted or lost to
follow-up (9%), and still on treatment, transferred or
unknown (18%). HIV-infected patients had a six fold
increase in odds of poor outcome when compared to
HIV-negative children (aOR6.4; CI 2.119.1). Males,
underweight, and severely underweight patients experienced an increase in odds of death or treatment failure
(aOR2.9; CI 1.18.2, aOR2.6; CI 1.25.6, aOR9.0;
CI 3.4-23.5, respectively) than comparison groups.
Among the HIV-infected patients, those that had never
taken antiretroviral therapy (ART) died or failed
treatment at three times the rate when compared to
those who had taken antiretroviral therapy. The rate of
dying or failing treatment at any point in time during the
follow-up was 2.5 times higher among patients who were
HIV-infected (aHR2.4; CI 1.2-5.0).
Conclusion: HIV-infected, male, and underweight children with DR-TB were more likely to fail treatment or
die when compared to other children with DR TB within
this population. Targeted interventions and early use of
ART are needed to improve clinical outcomes among
HIV-positive and underweight children diagnosed with
and at risk for DR-TB.
OA-482-06 The management of child

multidrug-resistant tuberculosis contacts
across Europe
A Turkova,1,2 M Tebruegge,3,4 F Brinkmann,5 M Tsolia,6
F Mouchet,7 B Kampmann,8,9 J Seddon9 1Medical
Research Council Clinical Trials Unit at University College
London, London, 2Imperial College Healthcare NHS Trust,
London, 3University of Southampton, Southampton,
4
University Hospital Southampton NHS Foundation Trust,
Southampton, UK; 5Childrens Hospital, Ruhr University
Bochum, Bochum, Germany; 6National and Kapodistrian
University of Athens School of Medicine, P. & A. Kyriakou
Childrens Hospital, Athens, Greece; 7Universite Libre de
Bruxelles, Brussels, Belgium; 8Medical Research Council
(MRC) Unit The Gambia, Banjul, Gambia; 9Imperial College

London, London, UK. e-mail: a.turkova@ucl.ac.uk
Background: The burden of multidrug-resistant (MDR)

tuberculosis (TB) is increasing globally, including in the
WHO European region, where an estimated 75,000 cases
occurred in 2013. Consequently, a large number of
children are exposed to MDR-TB, become infected and/
or develop disease. There are no data from clinical trials
to inform the management of child MDR-TB contacts,
and national and international guidance is inconsistent.
However, observational studies suggest that the use of a
multidrug regimen, tailored to the drug susceptibility
results of the source case, can be safe and effective.
Design/Methods: From March to July 2014 we conducted a web-based survey using the Paediatric Tuberculosis
Network European Trials group (pTBNet), the IUATLD
Childhood TB Working Group, and the Childhood
Subgroup of the WHO Stop TB Partnership. The aim
of the study was to capture variations in the clinical
management of children exposed to MDR-TB in the
WHO European region.
Results: Of 176 specialists from 44 countries approached, 72 (41%) responded from 25 countries,
including 28 from 6 countries outside the European
Union/European Economic Area (EU/EEA). Sixty-six
(92%) had .5 years of experience working with TB;
59 (82%) were at senior level and 41 (57%) managed
73 child MDR-TB contacts a year. Forty-two (58%)
stated they were using preventive treatment (PT); 24
(57%) of those reported they were treating all children
with evidence of latent TB infection, 16 (38%) reported
treating exposed children without positive tests of
infection in certain circumstances (young age or immunosuppression). Thirty-one (74%) used tailored regimens with ethambutol, pyrazinamide, high-dose isoniazid, levofloxacin/moxifloxacin and ethionamide being
most commonly used. Twenty-one (50%) stated they
were treating for 6 months, and 14 (33.3%) for .6
months. Most were following cases up for 72 years
regardless of whether PT was given or not (71% & 86%,
respectively). The only factor associated with the
provision of PT was practice within the EU/EEA (vs.
outside the EU/EEA); adjusted odds ratio: 4.07 (95%CI:
1.33-12.5; P 0.014).
Conclusion: We identified a wide spectrum of practice in
Europe regarding MDR-TB-exposed children. Over half
of clinicians were using PT with varying indications, drug
regimens and follow-up periods, reflecting poor evidence
base and variable guidance. The data highlight that there
is an urgent need for prospective studies to inform
management decisions in the future.
24. The wide swath of vulnerable

populations
S447
Conclusions: Today Azerbaijan TB Control in prison

Programme is internationally recognized model for
penitentiaries in and beyond the European Region.
OA-483-06 20 years of TB control in the

penitentiary system of Azerbaijan:
achievements and challenges
E Gurbanova,1 R Mekhdiyev,1 F Huseynov,1 O Baghirov,2
R Tahirli3 1Main Medical Department of the Ministry of
Justice, Baku, 2Specialized Treatment Institution for
Detainees with Tuberculosis, Baku, 3Laboratory of the
Specialized Treatment Institution for Detainees with
Tuberculosis, Baku, Azerbaijan. e-mail:
e.gurbanova@prisonhealth.az
Background: The Main Medical Department (MMD) of

Azerbaijan Ministry of Justice provides healthcare
including TB services to 24 penitentiary institutions with
16 000 population. Design and methods: Following
political commitment, a TB Control Program has been
successfully developing in accordance with WHO
recommendations over 20 years. It includes early TB
case finding, qualitative and rapid diagnostics, standardized TB treatment, strict compliance to DOT and
infection control (IC) protocols as well as uninterrupted
supply with certified drugs. Quality assured TB laboratory is equipped with phenotypic and genotypic diagnostic tools for precise TB diagnostics. Both drugsusceptible (DS) and drug-resistant (DR) TB treatment
is centralized at the Specialized Treatment Institution
(STI) with approximately 900 bed-places. IC include
administrative (patient segregation in accordance with
infectiousness, drug sensitivity profile, treatment phase
and detention regimen) environmental (input and extract
ventilation, UV lamps) and individual. Patients are
offered permanent support of adherence via additional
food and hygienic packages as well as psychological
support to overcome stress related to disease and
imprisonment. Uncompleted TB treatment, initiated
within the PS, proceeds after release. Having ensured
appropriate juridical basis MMD has been collaborating
with MoH and NGO to provide support and follow-up
of TB patients in civilian sector.
Results: Comparing with 1995 in 2014 mortality from
TB among Azerbaijan prison population decreased .65
times (Diagram). The Programme achieved significant
treatment success rate both for new smear positive DSTB and MDR-TB cases equal 91% and 80% respectively.
100% of inmates pass through TB screening in pre-trial
isolators and 95% of the prisoners pass through annual
mass screening for TB including questionnaire, X-ray,
and sputum investigation. Special attention is paid to
identification and prevention of TB among HIV-infected.
Training Center functioning since 2012 at the STI was
declared WHO CC on TB prevention and control in
prisons and trained .700 medical and non-medical staff
of PS. Moreover, delegations from .15 countries visited
the Program to get acquainted with the practice.
Operational researches are performed on the Programmes theoretical and practical basis.
OA-484-06 Deaths due to tuberculosis and

social inequality: a spatial analysis in a priority
municipality for disease control, Brazil, 20062012
M Yamamura,1 M Popolin,1 M Santos Neto,2
L H Arroyo,1 F Chiaravalloti Neto,3 J Crispim,1
A Belchior,1 R Arcencio1 1Nursing School of Ribeirao Preto University of Sao Paulo, Ribeirao Preto, SP, 2Federal University
of Maranhao, Imperatriz, Maranhao, 3Public Health School
University of Sao Paulo, Sao Paulo, SP, Brazil, e-mail:
mellinayamamura@yahoo.com.br
Background: Death due to tuberculosis (TB) is as an

avoidable event there due to the availability of diagnostic
technologies and treatment. The objective was to analyze
the spatial relationship of preventable deaths from TB
with social inequity indicators in Ribeirao Preto - Sao
Paulo, Brazil.
Design/Methods: An ecological study that considered the
primary cause of deaths from TB recorded between 2006
and 2012 in the Mortality Information System. It was
possible to consider literacy variables of household
heads, number of residents per household, and per capita
income obtained through Census of the Brazilian
Institute of Geography and Statistics, in 2010. The
Analysis of Principal Components (PCA) for the construction of social inequity indicators was applied in the
Statistica 10.1 software. The geocoding of deaths was
performed in Terraview 4.2.2. TB mortality rate calculation happened through local and empirical Bayesian
method. It was necessary to define, as geographical units,
the areas covered by Primary Health Care Services.
Multiple linear regression and spatial regression were
used to test the spatial dependence between social
inequity indicators from TB mortality rates, in the R
software.
Results: Death rates, in regions under study, ranged from
0.00 to 797.31 deaths per 100 000 inhabitants/year. In
the APC, three indicators were built with a total variance
S448
of 79.5%. They were named as low, high and intermediate social inequity indicators with a variation of
46.21%, 18.76% and 14.61%, respectively. In multiple
linear regression, the median social inequity indicator or
intermediate was statistically significant (P 0.0005), R2
adjusted by 28.98% and there was spatial dependence
(Moran I 0.21, P 0.0034). Spatial Lag Model to
address the existent dependency resulted in the best one,
and it enabled to verify which areas with median social
inequity or intermediate indicator showed the highest
mortality rates.
Conclusion: Despite evidence of relationship between TB
deaths and extreme social inequality indicators, it was
possible to observe in the study a relation between TB
and areas where families are literate and have incomes
above the minimum wage. Perhaps, the resources are still
insufficient for the social development of communities,
which is an important and protective factor in TB
geography.
OA-485-06 Active case finding: a science of

delivery on the ground
S Pandurangan,1 S Chadha,1 S Mohanty,1
P Banuru Muralidhara,1 B Entoor Ramachandran,1
B Thapa,1 A Das1 1International Union Against Tuberculosis
and Lung Disease South-East Asia Office, New Delhi, India.
Fax: (91) 11 46 05 40 30. e-mail: sripriya14@gmail.com
Background and Challenges: Universal access to tuberculosis (TB) services is critical for timely diagnosis and
treatment. Early case-detection and ensuring complete
treatment of sputum-positive TB patients has always
been a major public health challenge for TB control
programmes. The current programme strategies have
been able to cater to only those patients who visit public
health institutions or those who were identified by
community-based healthcare workers.
Intervention: Project Axshya, supported by the Global
Fund, is being implemented by The Union in 300 districts
of India in partnership with 9 Sub-Recipients and a
network of more than 1200 community based organisations. The primary aim of Project Axshya is to enhance
the access of vulnerable and marginalized populations to
TB services through advocacy, communication and social
mobilisation activities. Under the Project, Active case
finding a strategy being implemented to identify the
missing three million TB cases with a focus on
enhancing accessibility to TB services among the
households. Trained community volunteers visits nearly
1000 households per month per district to sensitize about
TB followed by identification of TB symptomatics
(having cough of 2 weeks) in these households. Identified
symtomatics are either referred to the nearest designated
microscopy centres or if required sputum collection and
transportation is being done. If the symptomatics found
to be positive, they are linked to treatment services.
Results and lessons learnt: A total of 6.2 million
households were visited by the community volunteers
during the period April 2013 March 2015. About
435979 TB symptomatics were identified. Of these 181
107 persons sputum were collected and transported and
the remaining persons were referred to the nearest

microscopy centres. A total of 232 298 (53%) cases
were examined at DMCs and 17 584 (8%) were found to
be positive and 97% of those diagnosed were put on
DOTS.
Conclusion: The Active case finding through community
based organisations successfully enhances access to
diagnosis and treatment of TB for vulnerable and
marginalised populations.
OA-486-06 TB risk perceptions among medical

residents at a tertiary care center in India
G Pardeshi,1 M Parande,1 D Kadam,1 A Chandanwale,1
A Deluca,2 R Bollinger,3 J Farley,4 N Suryavanshi5
1
Byramjee Jeejeebhoy Medical College/Sassoon General
Hospital, Pune, India; 2Johns Hopkins Bloomberg School of
Public Health, Baltimore, MD, 3Johns Hopkins University
School of Medicine, Baltimore, MD, 4Johns Hopkins
University, Baltomore, MD, USA;7Byramjee Jeejeebhoy
Medical College/Johns Hopkins, Pune, India. e-mail:
kanugeet@gmail.com
Background: Risk perception influences an individuals

behavior. Resident doctors in India are at high risk for
tuberculosis (TB). To optimize uptake and effectiveness
of prevention guidelines, we assessed their risk perceptions regarding occupationally acquired TB.
Methods: The study using mixed methodology was
undertaken at BJ Government Medical College, a tertiary
health care center in Pune, India in September 2014.
Consenting post-graduate medical resident doctors
completed a self-administered anonymous structured
Knowledge, Attitude and Practice questionnaire about
factors perceived to increase their risk of TB. Data were
analyzed using v2 /Fishers Exact test. An open-ended
question and two focus group discussions assessed
concerns about their personal risk of acquiring TB.
Responses were analyzed to identify key themes with
exemplary quotes.
Results: Of 305 (94%) residents contacted, 263(86%)
consented and completed the questionnaire. Many
(55%) reported daily exposure to active TB cases.
Thirteen (5%) reported a previous episode of TB and
114(45%) knew of another resident who had been
diagnosed with TB. Almost all medical residents reported
concern with developing TB and drug-resistant TB (98%
& 88%, respectively). Compared to 1st year residents,
more experienced residents (2nd and 3rd year) were
more likely to report specific infection control and
workplace concerns as contributing to their TB risk.
For example, not using Personal Protective Equipment
(PPE) was identified as increasing the risk of TB in 23
(24%) of 97 1st year residents and in 80 (49%) of 163
other residents (P , 0.05). Residents who knew a
colleague with TB were more likely to report risk factors,
compared to those who did not know a colleague with
TB. For example, 71(62%) of those with a TB-infected
colleague reported workload as a TB risk factor,
compared to 59 (43%) of those without a colleague
with TB (P ,0 .05).Key themes identified as increasing
their personal risk, included overexposure to TB patients,
high-risk procedures/sites, patient cough etiquette, poor

use of PPE, poor ventilation, crowded working and living
conditions.
Conclusions: Our study found a high concern for
occupational TB, among medical residents in a large
urban hospital in India. Poor implementation of standard
infection control and workload were frequently identified as contributing to their perceived risk of TB. Less
concern expressed for known TB risk factors among 1st
year residents and among those who did not know a
colleague with TB highlights the need for increased TB
prevention education and better implementation current
TB prevention guidelines.
OA-487-06 Integrating tuberculosis screening

in mother and child health clinics: a policy
change could overcome the challenges of the
second sputum sample
1
2,3
S Maosa, M Kumwenda, A Chilembwe, B Chikuse,

B Chikuse,4 A Dimba,5 C Shaw,6 B-T Nyangwa,7
M Brouwer8 1Sue Ryder Foundation in Malawi, Balaka,
2
Malawi Liverpool Wellcome Trust, Blantyre, 3University of
Malawi, Blantyre, 4District Health Office, Balaka, Malawi,
5
Ministry of Health, Lilongwe, Malawi; 6Target Tuberculosis,
Brighton, 7Independent Consultant, London, UK; 8PHTB
Consult, Tilburg, Netherlands. e-mail:
soniamaosa@yahoo.com
Background: The Malawian National Tuberculosis (TB)

Program (NTP) guidelines require a presumptive TB
patient to submit two sputum samples: one at the time of
consultation and the second one produced early next
morning (the spot-morning or standard approach). A
project integrating active TB screening in mother and
child health clinics found that not all presumptive TB
patients submitted the morning sample. The same day
approach provides the required two samples in a single
visit. In this study we explored reasons why women do
not submit the morning sample. We also compared the
yield between the same day approach and the standard
approach.
Methods: Mixed methods study. Qualitative: focus
group discussions with (presumptive) TB patients and
health care workers. Quantitative: comparison of the
standard with the same day approach through inclusion
of consecutive presumptive TB patients at Balaka District
Hospital. Participants provided a second spot sample one
hour after the first spot sample in addition to the
standard morning second sample. We collected demographic and clinical data. A TB case was any participant
with at least one positive sputum sample by the standard
approach as per NTP guidelines.
Results: Reasons for not submitting a morning sample
included: long distance to the health facility, financial
burden of a repeat visit, fear of TB-HIV diagnosis, long
waiting time at the facility and discourteous attitude of
health care workers. The culturally subordinate position
of women within a household made it difficult for
married women to comply with morning sample
requirements. Being very ill or slackness contributed to
failing to go for a second visit. The same day approach
S449
evaluation has so far included 187 participants; 21

(11%) and 19 (10%) were bacteriological positive
according to the standard and same day approach
respectively. Of the 21 positive patients, 18 had all 3
sputum samples positive. Two of the remaining 3 were
positive on the morning sample only and the third had
both the second spot and the morning sample positive.
Five percent of patients failed to bring the morning
sample versus 1% for the second spot sample.
Conclusion: Women face many barriers in the health
system, at home and in themselves for submitting the
morning sample. A simple policy change from the spotmorning to the same day approach could overcome the
individual and social-economic barriers. Health systems
improvements are necessary to remove some of the other
barriers.
OA-488-06 Tuberculosis-entry screening for

asylum seekers in the Netherlands: challenges
and solutions in a pressurised immigration
context
Y Aartsma,1 J Den Boer,1 M Urban,1 M Vermue,1
Q Waldhober,2 E Kloeze,3 R Van Hest1,4 1Regional
Municipal Health Service Groningen, Groningen,
2
Association of Community Health Services and Regional
Medical Emergency Preparedness and Planning Offices in the
Netherlands, Utrecht, 3University Medical Centre Groningen,
Groningen, 4Regional Municipal Health Service
Rotterdam-Rijnmond, Rotterdam, Netherlands. Fax: (31) 10
4339 950. e-mail: nah.vanhest@rotterdam.nl

Netherlands operates a centralised system of asylum
application. Apart from a small minority at the national
airport Schiphol and unaccompanied minors, the majority of asylum seekers must file their request at the
national reception centre in Ter Apel, almost 22 000
persons in 2014. Within 72 hours their procedure of
registration and identity verification by the Alien Police,
an interview with the Immigration and Nationalisation
Service and mandatory radiographic screening for
intrathoracic tuberculosis (TB), must be completed. We
describe process and yield of daily entry screening,
including a time of a sudden high influx of asylum
seekers, predominantly from Syria and Eritrea. Apart
from the logistics, maintaining good standards of
medical practice under pressure of timely processing of
the asylum application procedure is challenging.
Intervention or response: Chest X-ray (CXR) screening
for TB at the Dutch national reception centre for new
asylum seekers, since April 2012 on a daily basis. CXRs
are digitally transferred for reading (within maximum 24
hours) to the TB Department of the Regional Public
Health Service of Groningen, also allowing teleradiology
in the weekends. Apart from making CXRs, upon
request of the TB Physician, the team of medical
technical assistants at the centre, can also take standardised questionnaires with telephone translators, collect
sputum for bacteriology or perform tuberculin skin
testing for further examination and can isolate individuals.
S450
Results and lessons learnt: The average daily number of

screenings increased from 22 in 2012, 34 in 2013 to 67 in
2014, with outliers above 200 persons per day in 2014.
Of all entrants 1.6% were further examined of which
4.5% were diagnosed with TB. Entry screening during 3
years identified 30 TB cases, of which 8 (27%) were
smear-positive, 26 (87%) culture-positive and 4 (13%)
multidrug resistant. TB was common among Eritreans
but not among Syrians. The annual yield of entry
screening was 114, 57 and 64 cases per 100 000 persons
respectively. A flexible and efficient practice model for
daily radiographic TB screening of a varying influx of
asylum seekers was feasible.
Conclusions and key recommendations: This unique
early diagnosis practice model was effective, with
potential for other European countries. The number of
TB cases found actively and passively after entrance will
be further researched, also as quality assessment of entry
screening.
health care level) were utilized to compare adjusted odds

ratios (aOR) of treatment success between age-groups.
Results: Out of 7539 TB notifications, 21% were among
the elderly. TB notification rates in the HIV negative and
positive population were higher in the elderly (HIV-: 616/
100 000; HIV: 4558/100 000) when compared to
younger adults (HIV-: 283/100 000; HIV: 3216/100
000) (P , 0.01). Among HIVelderly, notifications were
higher among males (5530/100 000) when compared to
females (3437/100 000) (P , 0.01). The overall decline
in notification rates from 2009-2014 was higher in
younger adults (elderly: 1957 to 572/100 000 vs. younger
adults: 2466 to 790/100 000; p,0.01). At TB treatment
commencement, 12% among the 760 had an unknown
HIV status compared to 7 and 6% in the 50-59 yrs and
younger adults (P ,0.01). HIV co-infection was lowest
in the 760 yrs (34%) and higher in the 50-59 yrs (66%)
and younger adults (82%) (P , 0.01). Among new PTB
cases (n 5342), success rate was 68% in 760 yrs
compared to 74% in each of the other age-groups (P ,
0.01). The likelihood of treatment success among all
treatment cases was decreased in the 760 yrs (aOR 0.56,
95%CI 0.47-0.67) and similar in 50-59 yrs old (aOR
1.01; 95%CI 0.83-1.23) in comparison to younger
adults.
Conclusion: The elderly only contributed one fifth to
total adult TB notifications explained by demographic
patterns. However, the 750 yrs old had the highest ageadjusted notifications rates, while HIV testing utilization
and treatment outcomes were inferior in 760 yrs old.
More attention is needed in TB prevention and care
programs for older populations, specifically in the
context of a maturing HIV-TB co-epidemic.
OA-489-06 Rapid expansion of TB prevention

and care programs: are the elderly left behind?
OA-490-06 Effectiveness of routine TB

screening in central Tanzanian prisons using
digital chest X-ray and XpertW MTB/RIF assay
B Kerschberger,1 I Ciglenecki,2 G Mchunu,3

L Rusike-Pasipamire,1 I Zabsonre,1 M M Win,1 B Rusch,2
`
S M Kabore1 1Medecins
Sans Frontieres
(Operational Centre
`
Geneva), Mbabane, Swaziland; 2Medecins
Sans Frontieres
(Operational Centre Geneva), Geneva, Switzerland; 3National
TB Control Programme, Ministry of Health, Manzini,
Swaziland. e-mail: benadi11@yahoo.de
J Van Den Hombergh,1 C Mangu,2 D Kowuor,2

J Malewa,3 E Chilolo,1 M Hoelscher,4,5 P Clowes2,5
1
Stichting PharmAccess International (PAI), Dar Es Salaam,
2
NIMR Mbeya Medical Research Centre, Mbeya, 3Tanzania
Prisons, Dar Es Salaam, Tanzania; 4Medical Center of the
University of Munich, Munich, 5German Centre for Infection
Research, Munich, Germany. e-mail: janushom@gmail.com
Background: Southern African countries scaled-up TB

care with focus on younger adults which are severely
affected by the intertwined HIV-TB epidemic. This
public health approach for care expansion, however,
risks leaving certain populations behind. Little is known
on TB notification rates and treatment outcomes for
elderly populations from high HIV-TB prevalence
settings.
Design/Methods: We compared first line TB drugs agespecific notification rates (2009-2014) and treatment
outcomes (2009-2013) between the elderly (750 yrs)
and younger adults (20-49 yrs) in the rural Shiselweni
region (210,000 population), Swaziland. Annual population figures were obtained from the 2007 census.
Logistic regression models adjusted for baseline factors
(sex, registration year, TB case definition, HIV status,
Background and challenges to implementation: Penitentiary institutions are proven to be a permanent source for
active TB transmission, but data on TB prevalence in
prisons, as well as associated risk factors, are scarce,
especially in Africa where 26% of newly diagnosed cases
occur. Effective and sustainable TB screening followed by
prompt treatment for those found to have active TB is
needed to curb TB transmission in prison settings.
Intervention: As part of a TB REACH grant, all inmates
and new admissions to 3 Tanzanian prisons in Dar es
Salaam are actively screened for TB by using chest X-ray
and the X-pert MTB/RIFw assay. In one of the three
prisons Xpert is only applied for those showing any
abnormality on the chest X-ray. In this prison computeraided diagnosis for TB (CAD4TB) is applied to the X-ray
images. PITC for HIV is offered as an opt-out strategy as
part of the screening protocol. Those who are found to

have active TB and/or HIV infection are registered for
immediate care and treatment following national guidelines.
Results and lessons learnt: From July 2013 to March
2015, a total 6003 inmates and remandees in three
central prisons in Dar es Salaam were screened. Overall
prevalence of bacteriologically proven TB was 2500 per
100 000 (range 1800 to 3100). In the prison with digital
radiography, 2139 prisoners were screened with chest Xray and 688 were found to have abnormalities on the Xray (32%). Among those, 7.8% tested positive on Xpert,
versus 3.6 and 1.8% of all tested with Xpert in prison 2
and 3 respectively. All TB patients detected were treated
promptly, often the same or the next day following
screening. In all 3 prisons TB notifications rose markedly
after introduction of the screening protocol and treatment could be initiated early, thus preventing further
transmission in the prisons and in the community after
release.
Conclusions: Screening prisons populations for TB using
Chest X-ray and Xpert identifies infectious cases of TB
early and facilitates effective treatment, reducing further
transmission. Using a protocol including chest X-ray
reduces the number of Xpert tests required and is
potentially cost-effective. Continuation of TB treatment
after release from the prison remains a major challenge.
Sustained effective TB screening in prisons will eventually bring down the incidence and prevalence of active
TB in prisons and contribute to reduction of transmission
in the community at large. Periodic screening of resident
inmates, followed by ongoing routine screening of new
admissions, should therefore become a core activity of
the prison health services.
S451
Table Activity of enhanced active TB case search and household

contact tracing
Indicators
Enhanced
active TB
case finding
Retrospective
household
contact
tracing
Total
enhanced
household
contact
tracing
Number of
households
visited
682,236
682,236
Number of
index TB
cases
20,805
20,805
Number of
individuals
screened for
TB
2,452,635
77,626
2,530,261
Number (%)
with
presumptive
TB cases
52,262 (2.1%) 9,142 (11.8%)
61,404 (2.4%)
Number (%)
evaluated at
Health
center
18,262 (34.9%) 4,967 (54.3%)
23,229 (37.8%)
All forms of
TB
identified
(% yield)
479 (2.6%)
249 (5.0%)
728 (3.1%)
Number (%)
of smear
positive TB
362 (75.6%) 184 (74.0%)
546 (75%)
cases1
Number (%)
of smear
negative TB
cases
49 (10.2%)
24 (9.6%)
73 (10.0%)
Number (%)
of extrapulmonary
TB cases
68 (14.2%)
41 (16.5%)
109 (15.0%)
1
25. Engagement of CSOs, communities

and patients in TB management and
control
OA-491-06 The yield of TB among contacts of
TB patients treated in the last three years
(retrospective screening) in Ethiopia
Z Dememew,1 M Ensermu,1 D J Dare,1 D Habte,1
M Aseresa Melese,1 K Alemu,2 N Demmelash,1
P G Suarez3 1Management Sciences for Health, Addis
Ababa, 2Management Sciences for Health, Bahir Dar,
Ethiopia; 3Management Sciences for Health, Arlington, VA,
USA. e-mail: zgashu@msh.org
Background and challenges to implementation: As part

of the strategy to strengthen tuberculosis (TB) prevention
and control efforts in Ethiopia, the USAID-funded Help
Ethiopia Address the Low Performance of TB (HEAL
TB) project has collaborated with regional health
bureaus of Oromia and Amhara to implement innovative
enhanced community based TB interventions.
The proportion of SS + is high here because all smear negative and EPTB
suspects are referred to hospitals for diagnosis and our data had not captured
all of them
Intervention or response: We conducted this study in six

selected zones of Oromia and Amhara regions between
June-September 2014. We recruited and trained lay
providers (high school graduates) to do active case
search through two approaches; (a) tracing household
contacts of TB index cases who completed treatment in
the last three years which we labelled as retrospective
contact screening; and (b) screening other households in
the community without index cases. They first visited the
houses of index TB cases and then the remaining
households in the neighboring area using the WHO
symptom screening criteria. They referred the presumptive TB cases to health centers for evaluation.
Results and lessons learnt: The community workers
visited household contacts of 20,805 index TB cases and
screened 77 626 contacts of which 11.8% had presumptive TB. Similarly, of 682,236 other households visited in
the catchment area, 2 452 635 household members were
screened for symptoms of TB of which 2.1% had
presumptive TB which was significantly lower than the
yield among contacts (P , 0.01). Of 23 229 presumptive
TB cases evaluated, 728 (3.1%) all forms of TB were
identified including 546 (75.0%) smear positive (SS),
109 (15.0%) extra pulmonary TB and 73 (10.0%) smear
S452
negative TB cases (SS-). Larger proportion of TB cases

were identified among TB contacts than the comparison
group (5% vs. 2.6%, P , 0.05). Over the same period,
the routine community TB activity by health extension
workers yielded 659 (2.9%) all forms of TB from 22 426
presumptive TB cases evaluated at the health centers and
the yield was less than that of retrospective contact
screening (P , 0.05).
Conclusions and key recommendations: The yield of
retrospective community based contact screening was
significantly higher than the routine community TB
activity and therefore should be considered as additional
case finding approach in the routine health extension
workers community based TB care activities.
OA-492-06 Role of peer educators to increase

HIV testing among TB patients in two ART
hospitals and one TB clinic, Jakarta, 2014-2015
Y Gunawan,1 H Agustine,2 F Susanti,3 M C J Tumbelaka,4
T Handayani,1 W Bruary,1 S Aditya,1 D P Setyawati5 1RED
Institute, Jakarta Timur, 2Persahabatan Hospital, Jakarta
Timur, 3Budhi Asih Hospital, Jakarta Timur, 4Jakarta
Respiratory Center, Jakarta Selatan, 5FHI360 Indonesia,
Jakarta Pusat, Indonesia. e-mail: msamsuri@fhi360.org
Background and challenges to implementation: HIV

testing among TB patients in Indonesia is low due to TB
staff time constraint and PITC skills. Although, several
CSOs are working on TB and HIV, only few of them are
working with both. RED Institute, a CSO with TB-HIV
experience in prisons, commenced its works in 2014 in two
ART hospitals and one TB clinic in Jakarta to promote HIV
testing among TB patients, with TGFs financial support.
Intervention or response: RED staff played the role as peer
educators, conducted TB-HIV IEC for all TB patients at
DOTS clinic in 2 hospitals (Persahabatan Hospital, Budhi
Asih Hospital) and 1 TB clinic (JRC Clinic) to motivate
TB patients to take HIV testing. All of those who agreed
to be tested were referred to HIV laboratory and HIV
treatment, when HIV results were positive.
Results and lessons learnt: Results of increase on HIV
testing among TB patients in 2 hospitals and 1 TB clinic
are as in attached Table.
Conclusions and key recommendations: HIV testing
among TB patients increased in all 3 sites. Major
contributing factor is the role of CSOs peer educators
that provided information and education about TB-HIV
to TB patients. The authors recommend other TB and
HIV CSOs to duplicate in other facilities.
OA-493-06 Effectiveness of sputum collection

and referrals from community screening for
active tuberculosis case finding in Mombasa
County, Kenya
T Kiptai,1 C Mwamsidu,1 A Munene,1 T Abongo,1
M Mungai,1 C Kamau,2 G Moemi,3 B Ulo1 1AMREF Kenya,
Nairobi, 2Christian Health Association of Kenya (CHAK),
Nairobi, 3Kenya Association for the Prevention of
Tuberculosis and Lung Diseases (KAPTLD), Nairobi, Kenya.
e-mail: titus.kiptai@amref.org
Background and challenges to implementation: In Kenya

notified TB cases dropped from 99 159 in 2012 to 89 760
in 2013 (9.48% decline). Mombasa County notified
4726 TB cases in 2013 giving a case notification rate of
469 per 100 000 that is above the national average of 208
per 100 000. In response to the high TB burden, Global
fund approved funds for National TB programs request
to undertake active TB case finding in Mombasa County.
AMREF Kenya is implementing the Grants Civil Service
Organization (CSO) component through two CSOs,
CHAK and KAPTLD. The CSOs work with public and
private facilities that include chemists as well as informal
ones like herbalists respectively to implement activities
through the community strategy.
Intervention or response: Between May 2014 and March
2015, AMREF Kenya through the CSOs trained and
engaged 136 community health volunteers (CHVs) and
76 laboratory technicians. Ten new laboratory technologists were hired and posted to high volume TB
diagnostic facilities to enhance examination of the
additional sputum load. CHVs.implemented active case
finding of TB through door to door screening of
household members. Presumptive TB clients identified
were referred to the nearest diagnostic facilities and
sputum collected for clients who were either not able to
visit the facilities due to lack of transport or unwillingness. CHVs.also followed up the presumptive TB clients
referred by private facilities to diagnostic facilities.
CHVs.were provided with bicycles and monthly stipend
to aid their activities.
Results and lessons learnt: Of the 21 099 people
screened, 9383 (44.5%) presumptive TB cases were
identified. Of these, 3279 (34.9%) were referred for
investigations, out of which 2508 (76.5%) reached the
facilities and 435 (17.3%) diagnosed with TB. For the
6104 (65.1%) who were not able to visit the facilities,
sputum was collected and transported for examination,
out of which sputum for 5492 (90%) were examined and
368 (6.7%) diagnosed with TB. In total 803 TB cases
were identified, increasing notified cases referred by
CHVs.from 4% to 28% during the implementation
period.
Conclusions and key recommendations: Door to door
screening is core in increasing case detection. Sputum
collection and referral reduces diagnostic delays. There
is need to carry out the same activities in areas with high
notifications rates to combat TB infections in the
regions.
OA-494-06 Sputum collection and

transportation in vulnerable and marginalized
communities in India: experiences from
community driven TB project Axshya
1
B Thapa, A Das, S Mohanty, P Banuru Muralidhara,

B Entoor Ramachandran,1 J Tonsing,1 S Chadha1
1
South-East Asia Office, New Delhi, India. e-mail:
bthapa@theunion.org
Background: Community case finding interventions in

vulnerable and marginalized communities are underway
in Global Fund supported civil society driven Tuberculosis (TB) prevention and care project, Axshya in India.
Identified presumptive TB patients (PTBP) are linked
with diagnostic and treatment services through referrals
and sputum collection and transportation (SCT). The
objective is to determine how SCT has contributed to TB
patient diagnosis and treatment in comparison to
referrals.
Design/Methods: Phase II of project Axshya was
implemented in 300 districts in India from April 2013.
Case finding interventions included; active case finding
(ACF), community meetings, mid-media activities, engagement of rural health care providers, facility based
SCT, and engaging groups and networks of people living
with HIV. The intervention wise aggregated data on
number of interventions, PTBP identified, referrals, SCT,
TB diagnosed and initiated on treatment were collected
from the quarterly progress reports from April 2013 to
September 2014 and analysed.
Results: Project identified 665 577 PTBP, 357 972
(54%) were referred and SCT was done from 307 605
(46%). Of those referred, 30% (106 689) could be
examined. PTBP whose SCT was done through all
activities were 3.4 times more likely to get their sputum
examined than referrals. This increased to 4.3 times if
SCT was done following ACF. Altogether, 62% (414
294) PTBP were examined and 34 900 TB patients were
diagnosed and 33 837 (97%) were initiated on
treatment. Highest yield was from was ACF (12 692,
38%). The smear positivity rate was higher in referrals
(8.7%) than SCT (8.3%) (P . 0.05) with overall
positivity rate of 8.4%. TB patients diagnosed through
SCT were 1.03 times more likely to be initiated on
treatment than referrals (P , 0.01). If the project was
able to examine all the referrals that did not reach the
microscopy centre through SCT, an additional 20 857
TB patients would have been diagnosed.
Conclusion: Sputum collection and transportation has
provided accessibility for diagnosis to 46% of those
PTBP identified, diagnosed 25,636 and initiated treatment on 25,062 TB patients. SCT coupled with
community case finding interventions has shown remarkable success for providing access to diagnostic,
diagnosing TB patients and decreasing initial defaults in
marginalized and vulnerable communities in India.
S453
OA-495-06 Empowering faith-based

community health volunteers as partners in TB
control
J Teves,1 F Agbayani,2 T Yu,3 S Dela Cruz,3 S Masulit3
Diocese of Ipil, Zamboanga, 2Integrated Provincial Health
Office, Zamboanga, 3Philippine Business for Social Progress,
Manila, Philippines. e-mail: virgilbelen@gmail.com
1
Background and challenges to implementation: With a

TB cure rate (new smear-positive) of 82% in 2011,
Zamboanga Sibugay province (population: 576 232) in
southern Philippines needed to improve case holding to
meet the national target of 85%. However, distance,
poor road infrastructure, limited availability of transportation, and high transportation fare made it difficult
for public health workers to access the provinces
geographically isolated areas. The Integrated Provincial
Health Office (IPHO) expanded its coverage by working
with the Catholic churchs Community-based Health
Program (CBHP) that had a cadre of over 1000 parish
volunteers distributed across 19 parishes in the 16
municipalities of the province, all committed to serve
the community at practically no cost to the IPHO.
Intervention or response: With USAID/Philippines technical assistance, the IPHO trained more than 600 CBHP
community volunteers to conduct TB education and
counseling sessions and serve as treatment partners to
ensure treatment compliance and completion. They were
also trained to identify individuals with TB symptoms,
refer them to the nearest TB-DOTS center, and regularly
maintain records and submit reports.
Results and lessons learnt: The CBHP volunteers helped
identify 20 new smear-positive (NSP) cases in 2012, 78 in
2013, and 91 in 2014 contributing 4%, 15%, and 16%,
respectively to the provinces CDR (NSP). Correspondingly, the volunteers served as treatment partners to 6,
17, and 30 of the NSP cases who were initiated on
treatment. This resulted in a treatment success rate (TSR)
of 84% in 2012 and 95% in 2013. TSR for the 2014
cohort is unavailable as yet.
Conclusions and key recommendations: Mobilizing
faith-based community ealth volunteers is a low-cost,
viable, and sustainable strategy for TB case finding and
case holding, particulrly in geographically challenged,
access-poor areas.
OA-496-06 Role of community volunteers in

finding missed TB cases: lessons from Three Is
implementation in Zambia
A Nota,1 S Kaminsa,1 M Chizyuka,1 E Chiyeke,1
M Amofa-Sekyi1 1FHI360/TB CARE I, Lusaka, Zambia. Fax:
(260) 211 257 329. e-mail: amosnota@yahoo.com
Background and challenges to implementation: Countries with a high TB-HIV co-infection rate, like Zambia,
are encouraged to focus on Intensified Case Finding
(ICF), Infection Control, and Isoniazid Preventive
Therapy; referred to as the 3 Is to control the TB
epidemic. Tracing household (HH) contacts of smearpositive TB cases is an effective ICF tactic, but not often
employed in Zambia.
S454
Intervention or response: HH contact tracing was

implemented in 15 communities attached to 15 health
facilities in six districts, of Central and Copperbelt
provinces in Zambia. A total of 124 community
volunteers were engaged across these communities from
May 2013 to March 2015. The community volunteers
collected details of all patients diagnosed with TB by
either smear or GeneXpert MTB/RIF from the TB
register with faciltiy staff support. Each volunteer
arranged a convenient meeting time with the HH
members of this diagnosed individual, ensuring that the
majority were present during the visit. The community
volunteer screened all HH members for cough, night
sweats, weight loss and fever. All symptomatic HH
contacts submitted two spot sputum specimens to the
volunteer. The sputum samples were either sent for AFB
smear microscopy (HIV-seronegative) or GeneXpert
MTB/RIF examination (HIV-seropositive). HH contacts
with a sputum or Xpert MTB positive result were
referred or escorted to the health facility for TB
treatment. HH contacts were aslo encouraged to test
for HIV.
Results and lessons learnt: Among the 7455 HH contacts
identified and screened for symptoms of TB, 1607
symptomatic contacts submitted sputum for examination, and 180 (11.2%) were bacteriologically confirmed
with TB. All confirmed contacts were started on TB
treatment. These cases were identified in the community
by community volunteers within a period of 10 months.
Conclusions and key recommendations: We found, like
many others, that HH contact tracing is an extraordinarily effective way to increase case detection. The cases
found over a 10 month period extrapolates to a
population-based case rate of approximately 2,414/100
000 population, 6 times greater than the estimated TB
case in the general Zambian population. Engaging the
community actively in ICF efforts is a very effective
mechanism for finding some of the missing TB cases
within a local population. Involving local communities in
primary health care, may help resource-poor health
systems achieve the goal of universal access for all.
OA-497-06 Active case finding activities

performed by former TB patients: experience
from South-Kivu Province, Democratic Republic
of Congo
E Andre,1 D Bahati Rusumba,2 E Musafiri,2 D Muzigo,2
D Kasereka,2 D Kalumuna,2 O Le Polain De Waroux3
1
Universite de Louvain, Brussels, Belgium; 2Coordination
`
Provinciale Lepre
et Tuberculose du Sud-Kivu, Bukavu,
Democratic Republic of the Congo; 3London School of
olivier.lepolain@lshtm.ac.uk
Background: Late and under-detection of TB is one of the

most important challenges towards TB control in the
Democratic Republic of the Congo. We here report the
results of the first 14 months of implementation of an
Active Case Finding (ACF) programme in the South Kivu
Province (population ~ 6 million), where former TB
patients are being trained to undertake ACF in their

communities. Material/
Methods: ACF activities were performed by screeners in
their close environment, before extending it geographically (stone in the pond approach). Information on
socio-demographic characteristics were recorded for
each individual screened, as well as information on
cough lasting for 715 days. Cases with a cough were
offered a laboratory test (Ziehl Nielsen). We performed
descriptive statistics and explored differences in the
prevalence of cough and TB by socio-demographic
characteristics, through the use of negative binomial
regression.
Results: A total of 14 043 individuals from 2900
households were screened by the 141 screeners in
communities living in 15 of the 34 health districts of
the Province. Overall, 2,103 (15%) individuals reported
a cough for 715 days. This proportion increased linearly
with age from 4.5% in ,5y olds to 28.6% among adults
aged 40 years and above, was higher in males than
females (P , 0.0001), and cases tended to be clustered
within households. About 92% of individuals with a
cough accepted to be tested, of whom 38 (1.8%) were
confirmed with smear positive pulmonary TB (SPP). TB
incidence increased linearly with age and was higher in
males, with a 5% increase in the proportion of diagnosed
cases by 5 year age increase (adjusted incidence rate ratio
for age (aIRR) 1.05 (95%CI 1.04 1.06)) and sex (aIRR
males 1.31 (95%CI 1.01 1.69). The cumulative
incidence of SPP TB in the province, as reported through
the national surveillance system, was 53 per 100 000
(95%CI 50 55). In comparison, the incidence of SPP TB
from the ACF activities over the same period was 231 per
100 000 (95%CI 154 334).
Discussion: Our preliminary results suggest that almost
2% of individuals reporting a cough lasting for more
than 15 days were diagnosed with SPP TB. The higher
incidence among communities targeted by ACF activities
is likely due to a combination of increased detection, as
well as the targeting of more high risk communities, and
suggest ACF activities are important to help curtail the
burden of TB in South Kivu.
OA-498-06 Improving treatment success rate by

use of community mobilizers in Juba, South
Sudan
J Mogga,1 H Lasu,1 S Macharia,2 S Kia,3 B Woldemariam4
Ministry of Health, South Sudan, Juba, 2Management
Sciences for Health, Juba, 3AIDS Resistance Trust, Juba, South
Sudan; 4Management Sciences for Health, Addis Ababa,
Ethiopia. e-mail: smacharia@msh.org
Background and challenges to implementation: Treatment success has stagnated below 80% in the last three
years. The treatment success rate dropped to 72% in
2012 compared to 75% in the 2011 cohort. Juba is
estimated to have a population of over 1 million which is
served by only three TB Management Units (TBMU)
contributing to over 40% of all defaulters in South
Sudan. No community based TB control strategies
applied in Juba city. Directly Observed Treatment (DOT)

is not strictly adhered nor is there a concrete treatment
follow up strategy utilized in Juba to date.
Intervention: To address these challenges, the National
TB, Leprosy and Buruli ulcer Programme (NTLBP), with
support from the Global Fund and USAID-funded TB
CARE I project, implemented a project from October
2013 to 31st March 2014 to retrieve loss follow-up TB
patients in Munuki Primary Health Care Center (PHCC).
A list of these patients was given to a network of 20
community mobilizers who traced the patients to their
respective residence or through telephone contact provided in the TBMU register. The treatment success rate
after the intervention was compared with the reports
previously reported to the NTLBP.
Results and lessons learnt: A list of 539 TB patients had
defaulted from the TBMU register from 1st January 2012
to 31st December 2012 out of which 2 had inadequate
personal information and could were dropped from the
study. A total of 87.2% patients were traced and made to
complete treatment, 4.7% could not be located, 7.8%
had moved to another facility to continue with treatment
and 0.4% died. There were no significant differences in
outcomes according to age and gender. After the
intervention, treatment success rate among new TB cases
increased from 61% to 84%. Challenges encountered
include incomplete recording of patients data, highly
mobile population and crisis on 15th December 2013
that resulted in displacement of TB patients.
community mobilizers in TB control can improve on
adherence to treatment. Close coordination between the
health facility and community mobilizers is required to
provide timely list of loss to follow up. Completeness of
data is of essence to ensure consistence in information in
all patients data sources. This approach can be
replicated in other urban centers where the defaulter
rate is high.
26. Lung health: pneumonia, COPD,

asthma and cystic fibrosis
OA-499-06 A controlled human aerosol model
to study respiratory hygiene interventions to
prevent airborne transmission
K Vanden Driessche,1,2,3 J Zlosnik,2 N Hens,1,4 B Quon,2
B Marais,5 R De Groot,3 D Speert,2 M Cotton6 1Hasselt
University, Center for Statistics (CenStat), Diepenbeek,
Belgium; 2University of British Columbia, Vancouver, BC,
Canada; 3Radboud University, Nijmegen, Netherlands;
4
University of Antwerp, Vaccine and Infectious Disease
Institute, Antwerp, Belgium; 5University of Sydney, Marie
Bashir Institute for Infectious Diseases and Biosecurity and
Childrens Hospital at Westmead, Sydney, NSW, Australia;
6
Stellenbosch University, Cape Town, South Africa. e-mail:
koen.vandendriessche@radboudumc.nl
Background: Mycobacterium tuberculosis is a notorious

airborne transmitted pathogen. Ongoing transmission in
high incidence areas drives the global TB epidemic
S455
including drug resistant disease. There are increasing

concerns about other bacteria being transmitted by air:
Burkholderia cepacia complex and Pseudomonas aeruginosa are associated with increased morbidity and
mortality in cystic fibrosis (CF) patients and both have
been isolated from CF clinic room air. The recently
updated CF Foundation (USA) infection control guidelines now advise all CF patients to wear surgical masks in
healthcare settings. While generally accepted that covering ones cough (with a mask or by practicing cough
etiquette) prevents droplet spread, the effect on airborne
transmission has not been conclusively evaluated. We
have developed a model to measure concentration and
particle size distribution of viable (myco-)bacteria from
human cough aerosols. We set out to improve this model
and validate it as a tool to study interventions against
airborne bacterial spread. Design We recruited CF
patients who were chronically colonized with P. aeruginosa. Participants were asked to cough in two cylindrical
chambers from which the air was extracted through
cascade impactors (Figure). Participants moved between a
chamber where they wore a surgical mask while coughing
and a control chamber where they coughed without a
mask (42 coughs in total equally divided over the two
chambers). We took delayed aerosol samples and excluded microorganisms from stage 1 in the impactors which
captures droplets, as evaluation of airborne transmission
was our primary experimental outcome measure. Generalized linear mixed models were used to measure the effect
of masks on airborne P. aeruginosa load.
Results Eleven CF patients, aged 16-77 years, participated. Viable P. aeruginosa was isolated during 12
sessions from 6/11 (55%) participants. Wearing a
surgical mask reduced the airborne P. aeruginosa load
by an average of 88% (95%CI 81-96%).
Conclusions The controlled human aerosol model used
in our study produced relatively precise results despite a
small sample size. Our data show that surgical masks
reduce the airborne spread of P. aeruginosa in CF,
supporting the new recommendation that CF patients
should wear surgical masks in healthcare settings. Next,
we intend to evaluate the potential impact of different
public health interventions, including cough etiquette,
aimed at reducing M. tuberculosis transmission.
S456
OA-500-06 The correlation between IL-16-295

polymorphism and risk of asthma: evidence
from a meta-analysis
Q Wu,1 Y Ji,1 C Wu,1 Y Wang,1 Q He1 1Sichuan University,
Chengdu, China. e-mail: 76201313@qq.com
Background: IL-16 is one of the immunomodulatory

cytokine whose expression is elevated in the bronchial
epithelium, bronchoalveolar lavage fluid and induced
sputum of asthmatic patients. It promotes the migration
of T cells to target sites, plays a critical role in the
development of coronary artery disease. Several researches reveal that giving medication of antibodies to
IL-16 can decrease development of airway hyperresponsiveness, without having impact on airway eosinophilia in a mouse model of allergic asthma. The aim of
this study is to research the relationship between il-16295 polymorphisms and asthma disease risk by metaanalysis.
Design/Methods: We searched PubMed (http://www.
ncbi.nlm.nih.gov/pubmed/) and China National Knowledge Infrastructure (CNKI) (http://www.cnki.net/) web
databases to cover all studies based on association
between IL-16-295 polymorphisms and risk of asthma
until April 2015. The meta-analysis was performed for
selected case-control studies. At the same time, pooled
odds ratios (ORs) and 95% confidence intervals
(95%CIs) were counted for all genetic models.
Results: A total of 1693 asthma cases and 777 controls
were included from five case-control studies conforming
to our included criteria. Results from overall pooled
analysis suggested no evidence of significant risk between
il-16-295 polymorphism and asthma risk in all genetic
models, such as, allele (C vs. T: OR 1, 95%CI 0.86 to
1.18, P 0.96) , homozygous (CC vs. TT: OR 0.89,
95%CI 0.57 to 1.39, P 0.6) , heterozygous (TC vs.
TT: OR 1.06, 95%CI 0.87 to 1.29, P 0.57) ,
dominant (CC TC vs. TT: OR 1.03, 95%CI 0.86 to
1.25, P 0.73) and recessive (CC vs. TT TC: OR
0.86, 95%CI 0.56 to 1.34, P 0.51) models.
Publication bias was not detected in this meta-analysis.
Conclusion: The results of this meta-analysis indicating
that IL16-295 polymorphism may not have the relationship with asthma. Thus, future large-scale studies with
group of populations are needed to analyze the association between IL-16-295 and asthma.
OA-501-06 Use of Google Earth Pro and mobile

technology in preparation for evaluation of the
Practical Approach to Lung Health in rural
Malawi
B Chisunkha,1,2 H Banda,1,2 K Mortimer,2,3 B Ngwira,1,2
H Mtengo,1,2 J Smedley,3 I Namakhoma,1,2 S B Squire2,3
1
REACH Trust, Lilongwe, Malawi; 2Collaboration for Applied
Health Research & Delivery (CAHRD), Liverpool, 3Liverpool
School of Tropical Medicine, Liverpool, UK. e-mail:
blessingschisunkha@gmail.com
Background: In low and middle income countries

(LMIC), surveys in health research have relied on manual
processes and methodologies like mapping study areas,
household listing and household tracing plus paper-based

questionnaires for data collection. Similarly, there is
physical data transfer of questionnaires to central office
and manual data entry by assigned clerks. To reduce
challenges faced using these traditional approaches we
used Google Earth Pro (GEP) to electronically map a
rural area for population-proportional sampling and for
definition of clusters in preparation for a clusterrandomised controlled trial of the Practical Approach
to Lung Health in rural districts of Dowa and Ntchisi in
Malawi.
Design/Methods: GEP was used to create catchmentarea clusters around Health Facilities (HFs), buffer zones
between clusters and list of every household with
coordinates without requiring field visits. A random
sample of 6,480 households was then obtained from
within each of the clusters. 24 smartphones with installed
GPS Essentials were used by 14 research assistants to
navigate to the selected households. An electronic
questionnaire installed in the smartphones was used to
collect data which was then electronically transferred to
a secure server for storage and management.
Results: We were able to delineate cluster boundaries
with buffer zones around each of the 27 HFs in the two
districts, taking into account geographical features likely
to enhance/hinder patient access, and identify individual
households (figure). This reduced time and costs by
eliminating traveling to conduct household listing and
associated activities. The approach proved accurate,
effective and easy in tracing households. Time and cost
were further saved as both data collection and entry were
done at the point of interviewing. Data were directly
transferred to the server, reducing the risk of data loss.
Conclusion: Using GEP and Mobile technology in health
research has the potential to reduce duration and costs of
field-based projects since electronic mapping performed
on a computer eliminates travel and upkeep costs. This is
especially true in rural and remote regions of LMICs. It
also reduces project costs by removing a separate data
entry exercise. Data cleaning is eased as most errors are
checked at the time of data collection-with automatic
validation checks and skip patterns. Furthermore, it
reduces probability of selection bias by listing and
sampling teams and reduces data loss.
OA-502-06 Short-term health impact of stove

intervention in reducing household air
pollution and illnesses among rural women in
Pakistan
1
T Jamali, A Shahid, A Shahid, Z Fatmi, M M Kadir

Aga Khan University Hospital, Karachi, Pakistan. e-mail:
tanzil.jamali@aku.edu
Background: Improved stoves are considered the mainstay intervention to reduce household air pollution due
to biomass cooking fuel. This paper evaluated stove
intervention programs in two provinces, Sindh and
Punjab, of Pakistan, to determine the short term impact
in reducing household air pollution and illnesses among
rural women, who were the main cook in the household.
Design/Methods: A cross-sectional survey was conducted from June to September 2014 in the villages of Kasur,
Punjab and Dhabeji, Sindh. A total of 83 improved and
209 traditional stoves in Dhabeji and 134 improved and
179 traditional stoves in Kasur were included for
comparison. Interviewers administered the questionnaire
inquiring respiratory (IUTALD), eye and skin symptoms.
Blood pressure was recorded and peak expiratory flow
was measured to determine lung function. 24-hour
particulate matter (,2.5 lm) and carbon monoxide
levels were measured, in a subsample, in intervention and
control kitchens.
Results: Significant reduction for particulate matter and
carbon monoxide were observed among improved stoves
kitchens. Among women in Dhabeji, Sindh program,
significant risk reduction was observed for cough (aRR:
0.27, CI: 0.20, 0.38), phlegm (aRR: 0.27, CI: 0.18, 0.40),
shortness of breath (aRR: 0.16, CI: 0.11, 0.22), chest
tightness (aRR: 0.23, CI: 0.17, 0.31) and attack of
asthma (aRR: 0.33, CI: 0.22, 0.49) (P , 0.001). Risk
reduction were also observed for sandy eyes (aRR: 0.63,
CI: 0.47, 0.97) and itching in eyes (aRR: 0.62, CI: 0.41,
0.95 (P , 0.050). While in Kasur, Punjab program, risk
reduction was observed for phlegm (aRR: 0.60, CI: 0.45,
0.81) and also protection from burns (aRR: 0.56, CI:
0.34, 0.91). Mean peak expiratory flow among improved
stoves users was higher in both Sindh (19.30, CI: 12.90,
25.70) and Punjab (11.70, CI: 6.10, 17.30) programs.
Conclusion: Improved stoves showed substantial emission reduction and improved health indicators over
traditional stoves. However, variation and differential
health gains were noted among the two programs.
Further intervention experience and its evaluation,
standardization of stoves and adaptation according to
the local environmental and social needs is required to
scale up the stove intervention programs.
S457
OA-503-06 Respiratory failure in the medical

intensive care unit at Tikur Anbessa Specialized
Hospital, Addis Ababa, Ethiopia
A Work,1 A Sultan,1 A Shumet,1 K Gsilasie,1
A Binegdie Bekele,1 C Sherman,2 M Parekh,3
N Schluger3,4 1Addis Ababa University, Addis Ababa,
Ethiopia; 2Brown University, Providence, RI, 3Columbia
University, New York, NY, 4World Lung Foundation, New
York, NY, USA. e-mail: madhavi.j.parekh@gmail.com
Background: There is limited data regarding intensive

care unit (ICU) outcomes and specifically for those
respiratory failure in the developing world. In the United
States, ICU mortality is estimated between 10-30%
(SCCM, 2015), but is approximately 25-50% in
developing countries, and higher for those requiring
mechanical ventilation (Anesthesia, 2007). The purpose
of this study is to assess characteristics and outcomes of
patients with respiratory failure at tertiary care center in
Design/Methods: A prospective observational study
included all patients admitted to the MICU from June
2014-April 2015. Demographics, admission diagnosis,
comorbidities, need for mechanical ventilation, etiology
of respiratory failure, and ICU mortality were included in
the analysis. The Mortality Prediction Model (MPM0-II)
was used to estimate severity of illness.
Results: During the study period, 197 patients were
admitted to the MICU. Of these, 34% had respiratory
failure. The mean age of those with respiratory failure
was 35.4 years; 53.8% were female. Hospital acquired
pneumonia was the leading cause of respiratory failure
(22%) followed by pulmonary embolism (17%), aspiration pneumonia (15%), neuromuscular weakness including Guillain-Barre Syndrome and tetanus (13%), community acquired pneumonia (11%), cardiogenic
pulmonary edema (11%), pneumocystis jiroveci pneumonia (9%), severe asthma exacerbation (9%), tuberculosis (6%) and sepsis-related acute respiratory distress
syndrome or ARDS (6%). HIV/AIDS (25.5%) was the
most common comorbid condition, followed by COPD
or asthma (20.9%), and hematologic malignancy
(18.6%). Sixty two percent of those with respiratory
failure required mechanical ventilation, and 34% met
criteria for ARDS. Mortality for those with respiratory
failure was 47.6% as compared to an overall ICU
mortality of 34.8%, despite a similar severity of illness
(mean MPM0-II 18.8 and 17.8, respectively).
Conclusion: This study highlights that respiratory failure
is common in a tertiary care hospital ICU in Ethiopia,
and is associated with a high mortality rate. The majority
of patients with respiratory failure (nearly 50%) were
due to acute bacterial pneumonia (hospital or community acquired, and aspiration). Future study will include
analysis of antibiotic use and ventilator management in
these patients.
S458
OA-504-06 Spectrum and outcomes of patients

admitted with respiratory disease to the
medical wards at Tikur Anbessa Specialized
Hospital
A Shumet,1 K Gsilasie,1 A Work,1 F Oumer,1
Background: Very little published data exists regarding

the characteristics and outcomes of respiratory diseases
requiring inpatient admission in Ethiopia. This study
aims to assess the spectrum of respiratory diseases and
their outcomes among adults admitted to medical wards
at TASH, a tertiary care center in Addis Ababa, Ethiopia.
Design/Methods: A prospective observational study of
patients admitted to the medical wards of TASH between
November 2014 and March 2015 was performed. All
subjects were identified via chart review of inpatient
admissions; their demographic, comorbid conditions,
and clinical data were included in the analysis.
Results: During the study period, 102 patients admitted
to medical wards with a respiratory diagnosis. The mean
age was 44 years, men comprised 55% of patients, and
59% of patients were from rural areas. History of
smoking was present among 25% of patients, and was
exclusively among men. Those with HIV comprised
12%; 17% of patients had a history of tuberculosis
treatment. Active tuberculosis was the primary diagnosis
that required admission in 25% of patients (pulmonary
tuberculosis 15%, disseminated tuberculosis 10%),
followed by lung cancer/mass (17%), bacterial pneumonia (14%), pleural effusion (11%), pulmonary embolism
(7%), and decompensated pulmonary hypertension
(5%). More than one third (35%) of patients had more
than one concurrent respiratory diagnosis; the most
common secondary diagnosis was exudative pleural
effusion (11% of all patients). The most frequently
observed comorbid conditions were hematologic disorders in 21% of patients (most commonly leukemia or
lymphoma), diabetes (17%) and solid malignancies
(11%). The average length of stay was 21.4 days. Thirty
one percent of those studied died during their hospitalization; the majority of these deaths were related to
tuberculosis and lung cancer.
Conclusion: Tuberculosis and lung cancer were the most
frequent causes of admission and death among those
admitted with respiratory disease to the inpatient wards
at TASH. In-hospital mortality was strikingly high in the
study population. Future study will include a larger
sample size and analysis of management strategies.
OA-505-06 COPD among patients with

obstructive airway disease in Ethiopia
K Gsilasie,1 A Work,1 A Shumet,1 F Oumer,1
Background: While chronic obstructive pulmonary

disease (COPD) is considered very common worldwide,
its magnitude and risk factors are not known in Ethiopia.
Physician based clinical diagnosis may be challenging
given the relative low rate of tobacco use overall.
However, COPD may be more common than expected
given significant exposure to biomass fuel smoke in this
population. The aim of this study is to determine the
magnitude of and risk factors for COPD in an outpatient
chest clinic in a tertiary care hospital in Addis Ababa,
Ethiopia.
Methods: A prospective, cross-sectional study of patients
with a diagnosis of obstructive airway disease (asthma,
COPD, bronchiectasis) was performed from December
2014-March 2015. The first phase of the study included
a structured questionnaire of demographic, socioeconomic, and clinical information. The second phase of the
study included pulmonary function testing, specifically
pre-and post-bronchodilator spirometry.
Results: The results of the first phase are reported here.
Fifty-one patients with a physician diagnosis of obstructive airway disease were identified during the study
period. The mean age was 49 and women accounted for
60%. Nearly all (93%) of patients were from an urban
area. Asthma was the most common clinically diagnosed
obstructive lung disease (69%), followed by COPD
(11%), and bronchiectasis (11%); 13% of patients were
given more than one clinical diagnosis. Four patients
shared a diagnosis of COPD with asthma and/or
bronchiectasis. All 8 patients with any tobacco use
history were men, and the majority (63%) were given a
diagnosis of COPD; 80% of patients given a diagnosis of
COPD were men. Only 6% of women were given a
diagnosis of COPD; nearly all were diagnosed with
asthma, despite significant biomass fuel smoke exposure
(83%).
Conclusion: While obstructive lung diseases are common
in Ethiopia, it is unclear if physician-based clinical
diagnoses of specific diseases, notably COPD as opposed
to asthma or bronchiectasis, are correctly identified by
clinical characteristics alone, especially in a population
where tobacco use is uncommon. Perhaps patients who
have COPD are mislabeled as asthma or bronchiectasis,
based on traditional risk factors. Phase one of this study
is reported here, and identifies various diagnoses among
those with obstructive disease, but proper classification
will require spirometry with bronchodilator testing.
OA-506-06 Impact of the 13-valent

pneumococcal conjugate vaccine on clinical
and hypoxemic childhood pneumonia over
three years in central Malawi
E Mccollum,1,2 B Nambiar,2 R Deula,3 B Zadutsa,3
A Bondo,3 C King,2 N Lufesi,4 A Costello,2 T Colbourn2
1
Johns Hopkins School of Medicine, Baltimore, MD, USA;
2
University College London, London, UK; 3Parent and Child
Health Initiative Trust, Lilongwe, 4Ministry of Health,
Lilongwe, Malawi. e-mail: ericdmccollum@gmail.com
Background: The impact of the pneumococcal conjugate

vaccine (PCV) on childhood pneumonia during programmatic conditions in Africa is unknown. Following
PCV13 introduction in Malawi, we evaluated the case
burden and rates of childhood pneumonia.
Design/Methods: Between January 1, 2012-June 30,
2014 we conducted active surveillance for pneumonia in
children ,5 years at seven hospitals, 18 health centres,
and with 38 community health workers in two districts,
central Malawi. Eligible children met non-severe, severe,
or very severe World Health Organization clinical
pneumonia criteria. Oxygenation ,90% defined hypoxemic pneumonia. We quantified the case burden and
rates of clinical pneumonia, hypoxemic pneumonia, and
hospital pneumonia deaths following PCV13 introduction using multivariable time-series regression.
Results: Of 30 630 clinical pneumonia cases, 76,7%
were non-severe or severe, and 19,5% very severe.
Hypoxemic pneumonia occurred in 83% of cases and
32% with hospitalized pneumonia died. At 75% PCV13
coverage, versus 0%, non-severe cases increased 58%
(95%CI, 20%-96%), however, very severe cases decreased 63% (95%CI, 22%-100%) and hypoxemic
pneumonia fell 62% (38%-85%) in children ,5 years.
Among hospitalized cases ,5 years, very severe disease,
hypoxemic pneumonia, and fatality rates declined 65%
(22%-100%), 51% (19%-82%), and 39% (10%-67%).
Very severe and hypoxemia incidence also dropped 46%
(10%-83%) and 76% (39%-100%) at health centres.
Community health worker-diagnosed hypoxemic pneumonia rates declined 85% (17%-100%).
Conclusion: Two and a half years after PCV13 introduction, the health system burden and rates of the most
severe forms of pneumonia substantially decreased in
Malawian children.
S459

50. TB tidbits: diagnosis, surgery and
outcomes
PC-1122-06 Discrepancies between line probe
assays and phenotypical DST for detection of
MDR- and XDR-TB
E Ardizzoni,1,2 S Asnong,1 C Mathieu,2 T Sebastiani,3
I Zabsonre,3 G Torrea,1 F Varaine,2 B De Jong1 1Institute of
Tropical Medicine, Antwerp, Belgium; 2Medecins

Sans
`
Frontieres
(MSF), Paris, France; 3MSF, Geneva, Switzerland.
Fax: (32) 3 247 63 33. e-mail: eardizzoni@itg.be
Background: While WHO recommends to use line probe

assays (LPAs) like the MDR-TBplus and -sl for the rapid
detection of MDR-TB, MDR-TBsl has shown incomplete
sensitivity for fluoroquinolones (FQ) and injectables
agents (IA) resistance detection. In this descriptive study
we compared LPAs and conventional drug susceptibility
testing (DST) during one year of testing, and assessed the
impact of discordant results on treatment decisions.
Design/Methods: On samples sent for routine tuberculosis (TB) diagnosis by Medecins Sans Frontières to the
Institute of Tropical Medicine, Antwerp, LPAs MDRTBplus Vs2 and MDR-TBsl Vs1 were used to detect
multi-drug resistant (MDR) and extensively drug-resistant TB (XDR). Samples were tested by XpertwMTB/
RIF (Xpert) elsewhere. The LPAs were confirmed by DST
on liquid (MGIT) or solid medium. Gold standard for
RMP was rpoB sequence, and DST for the other drugs.
Results were collected from routine databases and
clinical information from patients files.
Results: Samples from 34 patients, of whom 23 (67.6%)
were MDR and 11 (32.4%) were isoniazid (INH) and
RMP susceptible by the gold standard, were also tested
with MDR-TBplus. LPA correctly identified all susceptible cases but missed 5 (21.7%) MDR; 3 MDR that were
false RMP-susceptible in LPA were also missed by Xpert
(mutations Phe251, Phe572), and another 2 were false
INH-susceptible in LPA. On the other hand, MGIT
missed 2 (8.7%) RMP-resistant cases, detected by Xpert
and LPA (Asn526 and Leu526). All cases started MDR
treatment based on Xpert results, except for the 3 false
RMP-susceptible cases in MGIT. In total 20 patients
started XDR empiric treatment by WHO recommendation before their isolates were tested with MDR-TBsl and
DST. Five (25%) were XDR, 3 (15%) were pre-XDR FQ
resistant, and 12 (60%) FQ and IA susceptible by DST.
Four XDR cases were missed by LPA, 1 for both FQ and
IA, and 3 for IA alone. Another pre-XDR was missed for
FQ. All susceptible cases were correctly identified.
Treatment was not modified following testing, whether
correct or not.
Conclusion: MDR-TBplus missed about 13% of RMP
resistant cases due to mutations out of the rpoB core
region, but detected the cases missed by MGIT, related to
disputed mutations(Van Deun 2013), also identified by
Xpert. MDR-TBsl on the other hand missed the majority
S460
Poster discussion sessions, Sunday, 6 December
of XDR cases. In settings where Xpert is used as first test,

LPA tests did not impact on treatment decisions.
PC-1124-06 Comparison of pre- and postXpertW tuberculosis treatment outcomes

among people living with HIV in Botswana
T Agizew,1 S Nyirenda,1 A Mathoma,1 U Mathebula,1
A Date,2 B Kgwaadira,3 P Pono,3 A Finlay1,2 1Centres for
Disease Control and Prevention, Gaborone, Botswana; 2U.S.
USA; 3Ministry of Health, Gaborone, Botswana. Fax: (267)
393 2904. e-mail: agizewt@bw.cdc.gov
Background: Xpert MTB/RIF (Xpert), a rapid molecular

diagnostic test, has a high sensitivity for diagnosing
Mycobacterium tuberculosis among people living with
HIV (PLHIV) compared to smear microscopy and can
potentially reduce time to diagnosis and treatment
leading to improved tuberculosis (TB) treatment outcomes.
Methods: From Aug 2012 - Mar 2014, PLHIV initiating
antiretroviral therapy at 22 HIV care and treatment
centers in Botswana were enrolled and underwent
systematic screening for TB. GeneXpert instruments
were deployed following a step-wedge design in 13 sites
from Nov 2012 - Jun 2013 to serve the centers. National
pre- and post-Xpert intensified case finding algorithms
were followed: initially symptomatic patients were
evaluated by testing sputum samples using smear
microscopy and after GeneXpert instruments were
installed, testing with Xpert. TB treatment outcomes
were compared between patients with drug sensitive TB
diagnosed pre- and post-Xpert, adjusting for intra-site
correlation.
Results: Among 6136 patients, 1878 were enrolled preand 4258 post-Xpert implementation. Upon initial
consultation, 1449 (23.6%) screened positive for one
of four TB symptoms (cough, fever, night sweat, weight
loss) and 244 (4.0%) diagnosed with and treated for
drug-sensitive TB. By Apr 2015, TB treatment outcomes
were available for 219/244 (89.6%); 41 were pre-Xpert
and 178 post-Xpert (Table 1). Among these, 18 patients
were transferred out. Treatment outcomes were unfavorable (died, defaulted or failure) for 8/39 (20.5%) preand 32/162 (19.8%) post-Xpert resulting in a risk ratio
of 1.03 (95% confidence interval [CI]: 0.63-1.71). When
stratified by smear microscopy results, treatment outcomes among patients with a negative smear result (n
149) were similar (risk ratio 1.36, 95%CI: 0.66-2.79).
Half of smear negative post-Xpert patients had a positive
Xpert test result, 64/127 (50.4%). The median number
of days from initial diagnostic test to TB treatment
among patients with smear negative disease was 23
(Interquartile Range [IQR]: 2-34) pre-Xpert compared to
6 (IQR: 3-19) post-Xpert (P 0.16).
Conclusions: Our preliminary findings indicate that
overall treatment outcomes, including those with smear
negative TB disease, were similar among patients treated
for drug-sensitive TB disease, pre- and post-Xpert.
However, Xpert was useful for bacteriologically confirming TB among patients with smear negative disease.
Table. Comparison of TB treatment outcome pre- and postXpert implementation among people living with HIV in
Botswana
PreXpert PostXpert Risk ratio 95% CI*
Treatment outcomes
among all TB
patients
Unfavorable
outcome (death,
default, failure)
Favorable outcome
(cured/completed)
Sub Total
Transferred out or
not evaluated
Treatment
Outcomes among
patients with
smear negative
disease
Unfavorable
outcome (death,
default, failure)
Favorable outcome
(cured/completed)
Sub Total
Transferred out or
not evaluated
n41
8
n178
32
1.03
0.63-1.71
1.36
0.66-2.79
31
130
39 (95%) 162 (91%)
2 (5%)
16 (9%)
n30
n119
17
22
91
28 (93%) 108 (91%)
2 (7%)
11 (9%)
*Adjusted for intra-site correlation
PC-1125-06 Trends in treatment outcome of

new and retreatment tuberculosis cases in two
regions of Ethiopia
K Melkieneh,1 Z Dememew,1 M Nigussie,2
I Jemal Abdulahi,3 Y Anteneh Kassie,1 D J Dare,1
M Aseresa Melese,1 P G Suarez4 1Management Sciences
for Health, Help Ethiopia Address the Low Tuberculosis
Performance (HEAL TB) Project, Addis Ababa, 2Amhara
Regional Health Bureau, Bahar Dar, 3Oromia Regional Health
Bureau, Addis Ababa, Ethiopia; 4Management Sciences for
e-mail: Kmelkieneh@yahoo.com
Background and challenges to implementation: Ethiopia

is one of the 22 high tuberculosis (TB) burden countries.
To address this challenge, the USAID-funded Help
Ethiopia Address the Low TB Performance (HEAL TB)
project in collaboration with the national TB program
has been supporting comprehensive TB prevention and
control interventions in two regions of Ethiopia, covering
50% of Ethiopias 90 million population since 2011.
Here we present trends in treatment outcome in HEAL
TB supported regions.
Intervention or response: The interventions included
capacity building for TB program managers at different
levels, training of health professionals, laboratory
capacity building including quality assurance, drug
supply management, mentoring and supportive supervision. In this report, we analyzed the trends in TB
treatment outcomes for the 10 zones which were
included in the initial phase of project implementation.
Results and lessons learnt: We supported treatment of
34,045 new smear positive pulmonary TB (SS) patients
in the two regions between October 2011 and June 2014.
Treatment success rate (TSR) for the SS improved from
81% at baseline to 95% at the end of the third year, and

cure rate (CR) improved from 52.3% to 88.0% (P ,
0.05). Of 1737 smear positive TB patients on retreatment regimen treated over the same period, TSR
and CR improved from 68.2% to 76.6 % and from
39.5% to 60.7% respectively (P , 0.05) [Figure]. The
annual increment in CR among retreatment cases was
10.6% which is significantly higher than the 6.5%
increment in new SS(P , 0.001). Moreover, the average
TSR and CR for new SS was high: 92% and 88%
respectively.
Conclusions and key recommendations: Treatment
outcomes improved steadily in both new and retreatment cases because of better adherence of patients
to treatment, improved clinical and lab follow ups, and
uninterrupted drug supplies.
up) even after more than a decade into implementation of

the programme. Among NSP the national average lost to
follow-up rate is around 6% of registered NSP cases
compared to 8% in Karnataka. Every fourth treatment
after loss to follow-up patient is again lost to follow-up.
Significant proportions of these lost to follow-up cases
become MDR-TB and come with the challenges of
holding them on a more complex and prolonged
treatment.
Conclusion: The treatment interruption rates are glaringly high among the TB cases (especially among
retreatment TB cases) in Karnataka (population: 63
million) when compared with national averages. Health
systems need to be adequately strengthened to handle
this situation. Focused interventions targeted at the most
at risk groups are needed to address this problem at the
earliest stage of care to avoid complications in the patient
management and reduce burden on the health system.
Loss to
follow-up rates
among notified
new smear
positive
TB cases (%)
PC-1126-06 Treatment interruption: the vicious

cycle
B Naik,1 K G Deepak,1 Y Patel,1 S Shastri,2 A Sreenivas,1
A Singarajipura2 1World Health Organization Country
Office for India, New Delhi, 2Department of Health & Family
Welfare, Bangalore, India. e-mail: naikb@rntcp.org
Background: Treatment interruption is a problem underestimated. The treatment interruption rates among
tuberculosis (TB) patients seem to be within the
acceptable range at country level but masking the interstate/ regional variations which are significant and not
addressed adequately. We attempted to see the trends of
the treatment interruption {programmatically called as
lost to follow-up (erstwhile called as defaulters)}
among new smear positive cases (NSP), relapse cases
and treatment after loss to follow-up cases in the country
vis a vis the state of Karnataka (population: 63 million)
which even after having a programme whose processes
were well appreciated (by national and international
evaluation teams); is having a persistently high rate of
loss to follow-up.
Design/Methods: Secondary data from the national
annual status reports of India and Karnataka state
quarterly reports were analysed and trends were
matched.
Results: As shown in the table; the trend of lost to followup patients in the state of Karnataka has remained higher
than the national average in all the 3 types of cases (new
smear positive, relapse and treatment after loss to follow-
S461
Loss to
follow-up rates
among notified
Relapse
TB cases (%)
Loss to
follow-up rates
among notified
treatment after
loss to
follow-up
TB cases (%)
Year
India
Karnataka
India
Karnataka
India
Karnataka
2005
2006
2007
2008
2009
2010
2011
2012
7
6
6
6
6
6
5
6
9
10
9
8
8
7
7
7
14
14
12
12
12
12
11
11
17
19
16
16
15
15
15
18
20
19
18
17
17
18
17
17
25
26
26
26
27
26
25
26
PC-1127-06 Surgical interventions for drugsusceptible and drug-resistant tuberculosis in

Mumbai
L Anande,1 A Dalal,2 S Shirodkar,2 M Das,3 H Mansoor,3
D Remartinez,3 P Isaakidis3 1Sewri GT Hospital, Mumbai,
2
`
Jupiter Hospital, Mumbai, 3Medecins
Sans Frontieres,
Mumbai, India. e-mail:
msfocb-asia-epidemio@brussels.msf.org
Background: There is a little documented evidence on

surgical interventions carried out in tuberculosis (TB)
patients including drug-resistant TB (DR-TB) patients.
However, there is an urgent need to include surgical
interventions in TB treatment guidelines. We aimed to
describe experiences and programmatic challenges while
providing surgical interventions to TB patients registered
for care in two tertiary-care hospitals in Mumbai, India.
of TB patients receiving surgical interventions in G.T.B
tertiary-care Sewri hospital and Jupiter hospital in
Mumbai, India between March 2012 and March 2015.
Results: A total of 113 TB patients who had undergone
surgical interventions were included in the study. The
cohort consisted of young adults with median (Interquartile range) age of 27 (22-35) years. Majority (74%)
S462
of them were males. Among all TB patients, more than a

quarter (26%) had DR-TB [Multi-drug resistant TB
(MDR-TB): Pre-extensive DR-TB: Extensive DR-TB:
Extremely DR-TB: 16:4:7:2]. Approximately half of the
patients (46%) were operated for pleural thickening and/
or partial/complete lung collapse. About 12% of the
patients were intervened for post-surgical complications
such as broncho-pleural fistula, wound gap or discharging sinus. The most common surgical interventions
carried out were pneumonectomy (32%, 36/113) and
decortications (29%, 33/113) for drug-susceptible TB
patients, while pneumonectomy (45%, 13/29) was the
most common surgical intervention for DR-TB patients.
There were three post-operative deaths recorded in this
cohort.
Conclusion: With increase in complex drug-resistance
forms of tuberculosis and limited number of effective
drugs surgery has become an adjunct to tuberculosis
treatment. Sometimes it remains the only effective option
for XDR-TB and worse form of DR-TB. One-eighth of
patients presenting with post-operative complications
highlight the need for close monitoring. The selection of
cases in need of surgery and appropriate surgical
interventions warrants discussion. Non-resection surgical options for patients with bilateral disease are urgently
needed as long as the current pharmacological treatment
options remain limited.
PC-1128-06 Factors associated with mortality

among TB patients in Kenya
J Njenga,1 A Wairia,1,2 B Mungai,1 F Ngari,2 E Masini,2
P Agunga,1,2 M Maina3 1Centre for Health Solutions Kenya, Nairobi, 2Ministry of Health, Kenya, Nairobi, 3USAID
Kenya, Nairobi, Kenya. e-mail: jnjenga@chskenya.org
Background: Kenya targets to have less than 5% of TB

case fatality by 2018. Available statistics however suggest
that TB case mortality is on the increase from 5.9% in
2012 to 6.6% in 2013. Understanding factors associated
with mortality among TB patients is necessary if the
country is to achieve her set targets.
Design/Methods: We abstracted data from the national
TB surveillance system, TIBU, for all patients started on
TB treatment between January 2012 and December 2013
and who had a treatment outcome assigned by the time
of data abstraction. The outcome of interest was death.
The explanatory variables assessed were whether the
health facility was public, private or prisons, type of
patient, type of TB, DOT by, HIV status, CTX and ART
uptake, nutrition support, sex, age, and patient referral.
Binary logistic regression data analysis methods were
applied using SPSS version 21.
Results: Out of 187 907 patients assessed, 10 973 (5.8%)
died during treatment. Dot by, referred by and CTX
uptake variables were not associated with death outcome. Facility type, sex, age, TB Type, patient type, HIV
status, ART uptake and nutrition support were all found
to be significant variables associated with mortality.
Combined, these factors explained 64% of variation in
the death outcome. Patients in private health facilities
were 19.8% (95%CI 15.7% to 23.8%) less likely to die

compared to patients in public facilities. Females were
16.5% (95%CI 13.0% to 19.8%) less likely to die
compared to males patients while adults were 78.9%
(95%CI 64.8% to 94.3%) more likely to die than
children below 15 years. PTB smear negative, PTB smear
positive and EPTB patients were 24.1% (95%CI 17.9%
to 29.9%), 60.4% (95%CI 56.9% to 63.6%) and 14.8%
(95%CI 7.4% to 21.5%) less likely to die compared to
PTB smear not done patients respectively. Relapse smear
positive and relapse smear negative, return after default
and relapse extra-pulmonary cases were all more likely to
die compared to new patients (P , 0.0001). HIV
negative patients were 73.5% (95%CI 71.5% to
75.3%) less likely to die compared to HIV positive
patients and being on ART reduced the likelihood of
death by 38.5% (95%CI 36.0% to 42.6%) compared to
being not on ART for HIV positive patients. Patients who
received micro nutrients and food support were 8%
(95%CI 2.8% to 13.7%) and 49.9% (95%CI 41.7% to
58.5%) more likely to die compared to those who
received no nutrition support respectively.
Conclusion: TB case mortality in Kenya is associated
with facility type, age, sex, patient type, TB type, HIV
status, ART uptake and nutrition support. There is need
to borrow best practices in private and prison facilities,
come up with innovative interventions targeting men and
ensure all pulmonary TB patients receive smear microscopy in order to reduce TB case fatality. HIV prevention
and ART uptake among HIV positive patients should be
scaled up. Food support should be provided at the right
time. Severely malnourished patients could benefit more
from being admitted and their progress monitored
closely instead of being provided with food support
and left on their own.
PC-1129-06 Osteoplastic thoracoplasty with

minimally invasive access in the complex
treatment of patients with destructive
pulmonary tuberculosis
D Krasnov,1 V Krasnov,1 D Skvortsov,1 I Felker1
Novosibirsk Tuberculosis Research Institute, Novosibirsk,
Russian Federation. e-mail: felkeririna@sibmail.com
1
Background: The collapse surgical operations for patients with pulmonary destructive TB are accompanied
by a significant number of complications, severe pain,
cosmetic defects and poor adherence of patients to this
surgical method. The aim of our study was to develop
and to test a new low-traumatic version of osteoplastic
thoracoplasty (OT) which is performed from minimally
invasive access.
Methods: We conducted a randomized study of 414
patients with complicated destructive TB. OT was
performed for all patients. In the main (I) group (n
191) - from minimally invasive access, in the control
group (II) (n 223) - according to classical method. Lung
resection in all patients was contraindicated. Before
surgery: Radiological signs of TB progression were
identified in 167 (87.4%) persons of I group and in
179 (80.3%) persons of II group (p 0.05, v2). Poly

cavernous lung lesion (2 or more cavities (CV)) was
observed in 118 (61.8 %) and 138 (61.9%) patients (P
0.98, v2). Bilateral subtotal lung damage was in 178
(93.2%) and 213 (95.5%) cases (P 0.30, v2), new
infiltrative foci in the opposite lung in 39 (20.4%) and
in 42 (18.8%) patients (P 0.69, v2) accordingly.
Bilateral CV localization was registered in 43 (22.5%)
patients of I group and in 61 (27.4%) patients of II group
(P 0.26, v2). Smear and culture positive before surgery
were 179 (93.7%) and 207 (92.8%) patients from I and
II accordingly (P 0.72, v2). MDR-TB was registered in
71.2% and 72.2% of cases (P 0.82, v2).
Results: Intraoperative blood loss less than 400 ml in the
main group noted in overwhelming number of patients 187 (95.4%), whereas in the control group - only in 105
(44.1%) (P 0.0001, v2). Various postoperative
complications in I group occurred in 28 (14.7%)
patients, in II group they were registered more frequent
- in 69 (30.9%) persons (p 0.0001, v2). OT application
led to bacteriological conversion in 144 (80.4%) patients
of I group and 143 (69.3%) of II group (P0.01, v2). CV
closure was achieved in 159 (83.2%) and 156 (70.0%)
patients, respectively (P 0.001, v2).
Conclusion: The proposed method is more sparing, the
risk of intraoperative blood loss (RR 10.10; 95%CI
9.20-11.01), and complications in the early postoperative period (RR 1.46; 95%CI 1.38-1.54) for classical
OT is much higher. Application of new techniques from
minimal access allowed to achieve better results than
traditional method, such as bacteriological and cavity
closure (OR 3 .7; 95%CI 3.4-4.01).
PC-1130-06 Persistent sputum smear positivity

in pulmonary tuberculosis patients during
treatment follow-up
C M Muvunyi,1 N Bizimungu,2 J C Ngabonziza Semuto,2
J B Mazarati2 1University of Rwanda, Kigali, 2Rwanda
Biomedical Center, Kigali, Rwanda. e-mail:
clmuvunyi@gmail.com
Background: In resource poor setting, sputum smear

microscopy still remains the method of choice used for
tuberculosis diagnosis and treatment monitoring. At the
same time, identification of acid-fast bacilli (AFB) in the
sputum smear during tuberculosis therapy is considered a
predictor of patient infectivity and treatment failure.
Design/Methods: The study was conducted at national
reference Laboratory in Rwanda to evaluate the magnitude of prolonged positivity of sputum seam during
treatment follow-up in Rwanda. We retrospectively
evaluated a cohort of 205 adult patients (142 men and
63 women) diagnosed with new case pulmonary
tuberculosis.
Results: Of the 205 patients, delayed sputum smear and
culture conversion occurred in 91 (44%) and 47 (23%)
patient beyond 2 months, respectively. Sputum smears
remained positive for AFB beyond 5 months of therapy in
13 patients (6%). Of the 13 patients with positive smears
after 5 months, 10 (77%) converted to negative culture
S463
results before negative smear results, and 3 (23%) had

persistently positive culture results. Thus, in this study
population, there were only three true treatment failures
beyond 5 months.
Conclusion: Persistent sputum smears positive for AFB at
the end of treatment do not necessarily indicate
treatment failure but is more likely to be associated with
negative culture results due to the presence of nonviable
mycobacteria. Based on the data presented here, we
proposed to investigate the pretreatment risk factors
associated delayed sputum smear and culture conversion
in Rwanda where microscopy remains the only tool
widely accessible for TB treatment monitoring.
PC-1131-06 Compliance with tuberculosis

diagnostic guidelines in South African public
health facilities
Z Mclaren,1 A Sharp,1 M Triyana,2 J Zhou,1 A Nanoo3
University of Michigan, Ann Arbor, MI, USA; 2Nanyang
Technological University, Singapore, Singapore; 3National
Health Laboratory Service, Johannesburg, South Africa.
1
Background: The quality of tuberculosis diagnostic care

and treatment varies across health facilities and affects
patient health outcomes. Both health system factors and
patient population factors contribute to non-compliance
with clinical guidelines. National TB Laboratory Registers can be used to generate information about compliance with clinical protocol at the patient and health
facility level. Information on where compliance is poor
enables the targeting of additional resources to locations
where it can save the greatest number of lives. This is the
first study on clinic and treatment quality in South Africa
using national laboratory register data for full sample of
public health facilities.
Laboratory Service database on tuberculosis tests performed on patients in public health facilities for the
period Jan 2004 - Dec 2011, which includes over 30
million test records from over 10 million patients at
health facilities. Culture-confirmed TB-positive cases are
based on the presence of at least one positive result from
TB culture, PCR test or Xpert MTB/RIF. We examine
compliance with clinical guidelines for TB diagnosis and
follow up testing of patients with suspected pulmonary
TB. We perform multivariate regression analysis of the
effect of facility and patient characteristics on the facility
compliance rate using supplemental data from the
District Health Information System (DHIS) and Census.
Results: Overall, 5.2% of patients received two tests
within the first two days, 7.3% of TB patients received
a test at the end of the intensive phase, and 3.6% of TB
patients received a test at the end of treatment. Among
patients who received at least two tests, these values were
35.0%, 40.0%, and 18.4%. Compliance was generally
higher in hospitals than in clinics and varied considerably
by province.
Conclusion: Our results demonstrate that directing
additional health resources to facilities in Limpopo
S464
province, rural areas and in the bottom decile of the

compliance rate distribution could improve compliance
and lead to better patient TB outcomes in these facilities.
The facility TB care indicators we produced are useful as
a standardized measure of quality of care and provides a
source of continuous monitoring.
51. M/XDR treatment: toxicity and stigma

PC-1132-06 Adverse events among MDR-TB
patients receiving second-line treatment in
Cubal, Angola
M L Aznar,1 C Bocanegra,1 E Gabriel,1 A Zacaras,1
I Molina,2 R De Carvalho,1 M Moreno,1 T Lopez1 1Hospital
Nossa Senhora da Paz, Cubal, Angola; 2University Hospital
Vall dHebron, PROSICS Barcelona, Barcelona, Spain. e-mail:
cristinabocanegra@gmail.com
Background: Second-line treatment for multi-drug resistant tuberculosis (MDR-TB) has recently started in
Angola. There is little data about the real burden of
adverse events of this treatment in rural, low-resources
context in wide areas of Sub-Saharan Africa.
Design/Methods: This is a descriptive, prospective study,
performed at Hospital Nossa Senhora da Paz, a 400-bed
rural hospital located in Cubal, in Benguela Province of
Angola. We included all adult patients who started
second-line treatment for MDR-TB between May 2013
and January 2015. Standard treatment in this setting
includes amikacin or kanamycin or capreomycin plus
ethionamide or prothionamide, ofloxacine, cycloserine
and pyridoxine for eight months, followed by twelve
months of treatment with ethionamide or prothionamide, ofloxacine and cycloserine. Prospective data were
analyzed to determine the frequency and characteristics
of adverse effects.
Results: One hundred and twenty-eight patients were
included, with a mean age of 32.3 years (SD 9.4). 56.3%
were male and 8 (6.3%) were co-infected with HIV. The
minimum time of follow-up was three months, with a
mean of 12.1 months (SD 5.9). Overall 90 (70.3%)
patients presented at least one adverse effect; among
these, 29 (22%) were severe. 77 (35%) patients had to
discontinue or lower the dose of at least one drug and 2
(1.5%) abandoned the whole treatment due to adverse
effects. Ototoxic (32%), gastrointestinal (25.8%), peripheral neuropathy (25.8%), together with increased
creatinine (38.3%) were the most commonly adverse
events reported, whereas ototoxic and psychiatric were
the most debilitating. Mean of time for development of
adverse effects was 123 days (SD 107.3); overall, 38.1%
resolved within one month. No differences were observed in risk of adverse effects between infected and
non-infected HIV-patients. Similarly, age was not significantly associated with adverse events.
Conclusion: Patients on second-line treatment experienced a wide range of secondary adverse events, many of
them severe and disabling, at very moment of follow-up.
Appearance of these effects often difficults correct
compliance of the treatment. Management of these

symptoms requires early identification and prompt
management by experienced professionals at all levels
of healthcare. Safer regimens for drug-resistant TB are
advocated.
PC-1133-06 QTc interval in patient treated with

a 9-month regimen for multidrug-resistant
tuberculosis
M-L Mbulula-Mawagali,1 Z Kashongwe Munogolo,2
J Toloko,3 F Kassa,4 M Gninafon,4 W Bekou,4 A Trebucq,5
K G Koura5 1Action Damien, Kinshasa, 2University of
Kinshasa, Kinshasa, 3Provincial Coordination of Tuberculosis
and Leprosy, Kinshasa, Democratic Republic of the Congo;
4
National Tuberculosis Programme of Benin, Cotonou, Benin;
5
Paris, France. e-mail: kgkoura@theunion.org
Background: The QT interval represents the time taken

for ventricular depolarisation and repolarisation. The
corrected QT interval (QTc) estimates the QT interval
according to the heart rate. Moxifloxacin and clofazimine have been shown to cause QTc prolongation with
an increased risk of ventricular arrhythmia. Our goal was
to describe the QTc interval before and one week after
initiation of treatment in multi-drug resistant tuberculosis (MDR-TB) patients treated with a 9-month regimen
including these two drugs at a normal dosage.
Design/Methods: Data on the patients included in the
prospective study in Benin and Democratic Republic of
Congo between September 2013 and January 2015 were
used. Socio-demographic characteristics and human
immunodeficiency virus (HIV) status were collected at
inclusion. An electrocardiogram was performed before
initiation of treatment and one week later. The QT
interval was manually measured and corrected according
to the linear regression Framingham formula: QTc QT
0.154 (1-RR), RR represents the time between two RR
waves. QTc is prolonged if the measure is higher than
440 milliseconds (ms) in men or 460 ms in women.
Students paired t test was used to compare the means of
QTc before and one week after initiation of treatment.
Results: One hundred and twenty-two patients were
included. The median age was 35 years, 40% were
female and 16% of patients were HIV positive. The mean
QTc [95%CI] before initiation of MDR-TB treatment
was 375.0 ms [368.8-381.2] and two patients had a QTc
prolongation: 461 and 522 ms. One week later, the mean
QTc [95%CI] was 388.0 ms [381.0-395.2], which is
significantly higher (P , 10-3) than the mean QTc before
treatment. One patient increased its QTc over the
threshold (from 354 to 603). The two patients with a
QTc prolongation before initiation of the treatment kept
it high: 456 and 543. These three patients, with a QTc
prolongation, are still alive: one has finished his
treatment and has been declared Cured. Treatment is
still ongoing for the other two patients.
Conclusion: The slight but significant increase in mean
QTc after initiation of treatment is consistent with the
cardiac effect of moxifloxacien and clofazimine. However, few patients had a QTc prolongation and the 9-
month Bangladesh regimen was not modified nor

interrupted for these patients.
PC-1134-06 Aminoglycoside ototoxicity in

tuberculosis treatment: a review of the World
Health Organization global pharmacovigilance
database
E Sagwa,1 P C Souverein,1 I Ribeiro,2 H G M Leufkens,1
A Mantel-Teeuwisse1 1Utrecht Institute for Pharmaceutical
Sciences, Utrecht University, Utrecht, Netherlands; 2Centre
for Health Technology Assessment and Drug Research,
Coimbra, Portugal. Fax: (264) 220 361. e-mail:
esagwa@yahoo.com
Background: Ototoxicity due to long-term aminoglycoside use in tuberculosis treatment (TB) is a potentially
debilitating problem. The comparative risk of streptomycin, amikacin, kanamycin and capreomycin in causing
ototoxicity in the context of TB treatment is still unclear.
Our aim was to evaluate the association between
aminoglycosides used for TB and the reporting of
ototoxicity as well as its determinants, using the
pharmacovigilance information that is spontaneously
reported by countries to the Vigibase maintained by the
WHO Collaborating Centre for International Drug
Monitoring at the Uppsala Monitoring Centre (UMC)
in Sweden.
Design/Methods: Case/non-case study among TB patients treated with aminoglycosides, with cases being
reports of ototoxicity, and non-cases being reports of
other adverse drug reactions (ADRs). The influence of
the concomitant use of medications for HIV was also
studied. The strength of the association between amikacin, kanamycin and capreomycin use in TB treatment, in
comparison with streptomycin use, and the risk of
ototoxicity was expressed as reporting odds ratio
(ROR), a measure of disproportionality in pharmacovigilance databases, with 95% confidence intervals (CI).
Results: Up to mid-2014, 3693 individual case safety
reports (ICSRs) were submitted in Vigibase where
streptomycin, amikacin, kanamycin and capreomycin
indicated for TB treatment were suspected of causing the
adverse reaction. Of these ICSRs, ototoxicity was
reported in 602 cases, the rest being other ADRs (noncases). The reported types of ototoxicity included
deafness (n 80), tinnitus (n 95), vertigo (n 404)
and 23 unspecified ototoxicity. In multivariate analysis
(Table1), amikacin was associated most with reported
cases of deafness (adjusted ROR6.5; 95%CI 2.8-15.1),
followed by kanamycin (adjusted ROR4.0; 95%CI 1.213.4). On the other hand, the use of capreomycin
compared to streptomycin use, was inversely associated
with the reporting of vertigo (adjusted ROR0.1; 95%CI
0.01-0.4). None of these three drugs were disproportinately associated with the reporting of tinnitus. Patient age
was an effect-modifier: ,25yrs (reference), 25-49yrs
(aROR1.4; 95%CI 1.0-1.8), 50-74yrs (aROR2.0;
95%CI 1.5-2.7) and 75yrs and older (aROR3.2;
95%CI 2.1-5.0). Sex and concomitant HIV medication
appeared not to influence the risk of reported ototoxicity.
S465
Conclusion: Compared to streptomycin, amikacin use

was associated with the greatest risk of deafness reported
in the VigiBase, followed by kanamycin. We advise
national TB control programs to closely monitor patients
for ototoxicity when using these aminoglycosides in the
long-term treatment of TB. Appropriately designed
studies are needed to confirm the comparative risk of
the various types of ototoxicity occurring in the longterm treatment of drug-resistant tuberculosis using
amikacin, kanamycin and capreomycin.
Table 1: Univariate (crude RORs) and multivariate (adjusted
RORs) analysis of risk of categories of ototoxicity by type of
aminoglycoside
Deafness
(n80)
Exposure
Streptomycin
Kanamycin
Amikacin
Capreomycin
Tinnitus
(n95)
Vertigo
(n404)
Adjusted ROR* Adjusted ROR* Adjusted ROR*

Reference
4.0 (1.2-13.4)
6.5 (2.8-15.1)
1.1 (0.9-2.3)
Reference
1.8 (0.4-7.5)
2.3 (0.7-7.5)
0.5 (0.1-2.1)
Reference
NA
0.3 (0.1-1.4)
0.1 (0.01-0.4)
*Reporting odds ratios (RORs) adjusted for sex and human immunodeficiency virus (HIV) co-treatment. Age not adjusted for because it was an
effect-modifier. NA not possible to calculate.
PC-1135-06 The effect of moxifloxacin and

clofazimine on the corrected QT interval (QTc)
in patients treated with a 9-month regimen for
MDR-TB in Niger
S Hassane Harouna,1 A Piubello,1 B Souleymane,1
S Morou,1 I Boukary,1 A Sala,2 K G Koura3 1Damien
Foundation, Niamey, 2Regional Hospital, Maradi, Niger;
3
Paris, France. e-mail: hassoul20@yahoo.fr
Background: Normal dose of moxifloxacin and clofazimine are currently prescribed as part of a 9-month
treatment regimen for multidrug-resistant tuberculosis
(MDR-TB) in Niger. Both are known to cause a
prolongation of the corrected QT Interval (QTc), with
the resulting risk for fatal arrhythmias. The aims of this
study are to describe the QTc interval at baseline and
during the first month of treatment.
Design/Methods: Between October 2013 and December
2014, a prospective study on a 9-month regimen to treat
MDR-TB patients was conducted in Niger. Breakfasts
were provided 30 min prior to the intake of the drugs. An
electrocardiogram (ECG) was recorded before the start
of the treatment (D0), as well as seven days (D7), and
thirty days (D30) later. The QT interval was manually
measured and corrected according to the linear regression Framingham formula [QTcQT0.154(1-RR)].
QTc was considered prolonged if higher than 450
milliseconds (ms) in men and 470ms in women. All the
ECGs were over-read by a cardiologist blinded, to time,
date, treatment and any data identifying the subject.
Results: Fifty-three patients were included, of whom
twelve (23%) were female. The median age was 36 years
and 11% of patients were HIV positive. At baseline the
mean 6 standard deviation (SD) for the QTc was 370.4
6 24.5 and one patient had QTc prolongation (male
S466
with QTc455ms). The QTc remained within normal

limits on day seven after the start of treatment, with a
mean (6SD) of 374.9 6 24.8. Thirty days after the
treatment initiation, the QTc mean (6SD) was 376.1 6
29.7 which is not significantly higher than QTc before
treatment (P 0.075). The single patient who had QTc
prolongation at D0, had a normal value at D7 (422ms)
and QTc prolongation at D30 (468ms). This patients
treatment was not modified but he was monitored daily
for one week until his QTc prolongation resolved.
Furthermore, no patient manifested a QTc longer than
500ms or a QTc prolongation longer than 60ms from
baseline when we compared the variation between D7
and D30 versus D0. No other significant ECG changes
were observed.
Conclusion: Moxifloxacin and clofazimine administered
in a 9-month regimen for MDR-TB appear to be safe and
well tolerated. An ECG should be performed before the
initiation of treatment to identify patients with QTc
prolongation. Patients with pre-existing QTc prolongation require close monitoring to ensure safety.
PC-1136-06 Hearing thresholds of drugresistant tuberculosis patients: baseline

audiogram configurations and associations
M Fadeyi,1 A Sogebi,2 V Ibeziako,3 A Agbaje,3 E Usoroh,3
S Dutt,3 P Dakum3 1Sacred Heart Hospital DR-TB treatment
centre, Lantoro Abeokuta, Ogun State, 2Olabisi Onabanjo
University, Ago Iwoye, Ogun State, 3Institute of Human
Virology Nigeria, Fct, Abuja, Nigeria. e-mail:
vibeziako@ihvnigeria.org
Background: The treatment of DR-TB involves the use of

aminoglycosides which have been shown to cause
varying degrees of ototoxicity. Majority of DR TB
patients in Nigeria are previous CAT 11 failures most
of which already have varying degrees of hearing loss as a
result of Streptomycin which poses a challenge to clinical
management. This study was aimed at using baseline
audiogram parameters to determine the occurrence of
hearing loss among DR-TB patients and the relationship
between sociodemographic and clinical parameters with
abnormal audiogram configurations.
Design/Methods: One year prospective study of patients
managed at the Ogun state drug resistant-Tuberculosis
center. Patients initial (baseline) audiometry was performed within two weeks of admission to assess hearing
thresholds, using pure tones ranging 0.25-8.0kHz for the
air conduction , and 0.25-4.0 kHz for bone conduction in
each of the ears separately. Patients socio demographic
and clinical data were also obtained. Data was collated
and analyzed
Results: Age range was 8-82 years, Mean6 SD
34.5612.6 years, Male: Female 2:1, 78.0% of the
patients were domiciled outside the home state, 13.6%
were retroviral positive, 94.7% had previous drug
regimen failure. ,10% of the patients had normal
audiograms, 78.8% had sensorineural hearing loss
(SNHL) and two configurations; ascending (40.9%)
and sloping (19.7%) were most common. Pure tone
averages at the low and high frequencies were 33.0 and
40.0 dB HL respectively. 27.3% had normal hearing in

the better ear, 50.8% had mild (25.1-40.0 dB), and others
with higher degrees of hearing loss. Male sex was
associated with deranged audiometric configurations,
while there was no association with age, positive
retroviral status, previous regimen failure and admission
weight ,50.0kg.
Conclusion: Many patients with DR-TB in Nigeria had
mild, suboptimal sensorineural hearing levels bilaterally
with male gender being the only factor associated with
abnormal configurations.
PC-1137-06 Prevalence of depression and

anxiety during treatment of MDR-TB patients
in a tertiary care hospital
Z Barry,1 Sara Khan,1 S Khan,1 O Qureshi,2 A Pasha,2
S Rehman,2 H Hussain,2 A Malik2 1Indus Hospital, Karachi,
2
Interactive Research and Development, Karachi, Pakistan.
e-mail: zainab.barry@industbcontrol.org
Background: Mental health is associated with every

illness. There is scarce literature about mental health
condition of TB patients undergoing prolonged treatment. There are fewer studies on the impact of TB
treatment on a patients mental health, especially MDRTB. The aim of this study is to measure depression and
anxiety in Multi-Drug Resistant-(MDR-TB) patients
during Intensive and continuation phases of TB treatment. Method: We conducted a prospective study to see
the pattern of depression and anxiety at Indus Hospital
from January to December 2014. All patients are
counseled at baseline and during their monthly follow
up visits. A random sample of 171 patients was selected
from both phases of treatment. The Aga Khan University
Anxiety and Depression scale (AKUADS) validated for
Pakistani population was administered. The tool has 25
questions with four responses - never, sometimes, often,
and always. Each question is scored from 0-3. A score
from 0-20 ensures no depression or anxiety; 21-40 is
categorized as mild depression and anxiety; 41-60 is
categorized as moderate depression and above 61 is
labeled as severe depression and anxiety.
Results: Of the 171 patients assessed 48% were females
and 52% were males with the mean age of 28 years. Of
these 58.5% were in intensive phase of treatment and
41.5% were in the continuation phase. Of the patients in
intensive phase 44% had no depression and anxiety,
53% had mild and 3% were severly depressed. Where as,
in continuation phase 53% of the patients had no
depression and anxiety, 16% had mild depression and
2% were still severely depressed.
Conclusion: Given the high rates of depression and
anxiety, counselling should be an essential part of
therapy to ensure adherence.
PC-1138-06 Occurrence of adverse events in

patient receiving community-based therapy for
multidrug-resistant tuberculosis in Pakistan
A Javaid,1 M A Khan,2 S Mehreen,2 A Basit,1
K Afsarafridi,2 U Ullah,2 M Khan3 1Lady Reading Hospital
Peshawar, Peshawar, 2Lady Reading Hospital, Peshawar,
3
Peshawar Medical College, Peshawar, Pakistan. e-mail:
mali_smile2005@yahoo.com
Background: Pakistan is one of the top listed countries

ranking 4th among top 22 multidrug resistant tuberculosis (MDR-TB) countries. The increasing rate of MDRTB in Pakistan underscores the importance of effective
treatment programs of drug-resistant TB. Clinical
management of MDR-TB requires lengthy multidrug
regimens that often cause adverse events (AEs).
Design/Methods: We retrospectively reviewed records of
patients who began MDR-TB treatment during 2012
cohort at a tertiary care hospital in Peshawar, Pakistan.
We sought to ascertain the occurrence of treatment
related adverse events and factors associated with these
events. We also examined the frequency of and reasons
for changing drug regimens. We further sought to
determine whether the occurrence of adverse events
negatively impacts the treatment outcome and management of adverse effects without requiring the discontinuation of MDR-TB therapy. Analysis was performed
using SPSS (SPSS version 16, SPSS Inc., Chicago, IL).
Results: Among 200 enrolled patients 52.5% were
female and most of the patients (81.5%) were from
group age 644 years. Among these patients 155
(78.27%) experienced at least one adverse event during
treatment. The most commonly reported events were
depression (70%), nausea (52.5%) followed by joint
pain (47.5%) and body aches (42.5%). Other events like
hearing loss, gastritis, decrease in sleep, vomiting, tinitis,
anorexia, vertigue and skin rashes occur 22%, 18%,
17%, 11.5%, 10.5%, 9.0%, 8.5% and 7.5% respectively. A change in drug dose due to an adverse events
occurred in 26% cases, while 6% cases had at least one
drug discontinued temporarily. Older age and lung
cavities at baseline were positive association with
occurrence of adverse events. Also association was found
between adverse events and outcomes (OR 0.480, 0.2360.978, P 0.041)
Conclusion: Adverse events were prevalent among
MDR-TB patients treated at PMDT unit Lady Reading
Hospital Peshawar. All patients who were older aged and
cavitatory lungs should be closely monitored for
treatment adverse events.
PC-1139-06 The experiences of adolescents

with multidrug-resistant tuberculosis and
participation in clinical research
K Zimri,1 P Rose,1 A Garcia-Prats,1 A Hesseling,1
H S Schaaf,1 G Hoddinott,1 L Viljoen1 1Desmond Tutu TB
Centre, University of Stellenbosch, Cape Town, South Africa.
e-mail: klassinaz@sun.ac.za
Background: Adolescents experience a large burden of

tuberculosis (TB), including multidrug-resistant (MDR)-
S467
TB. There is limited knowledge of adolescents experiences of MDR-TB, which may pose considerable
emotional and physical burden. The experiences of
adolescents participating in clinical research have been
poorly described to date, particularly in resource-limited
settings, but are important to improve involvement of
this neglected group in future research
Methods: We undertook a qualitative study using indepth interviews. Adolescents aged 10-18 years routinely
diagnosed and treated for MDR-TB in Cape Town, South
Africa, and who participated in an observational
pharmacokinetic study of secondline TB drugs were
included. Individual interviews were conducted with
each participant followed by a body-mapping session
where participants could identify how their bodies were
affected by MDR-TB. Interviews were conducted in
Afrikaans, Xhosa or English and were recorded, transcribed and translated in English. Data were analysed
through a thematic analysis.
Results: Nine adolescents were enrolled and 5 included in
the analysis. During analysis several themes emerged.
Participants had to deal with key challenges including (1)
the burden of the medication, including the pill burden,
medication taste, and pain and fear of injectable drugs;
(2) stigma (including rejection from peers); (3) dealing
with social isolation, including the exclusion from their
family due partly to hospital stay; (4) multiple fears in
terms of health and mortality; and (5) emotional loss
associated with MDR-TB, including many who were
affected by the death of a relative. While facing all of
these challenges, several patients displayed resilience and
good coping mechanisms. Participants found the experience of participation in clinical research empowering,
as they were able to learn about their disease and most of
them expressed interest in sharing their experiences.
Conclusions: Adolescents diagnosed with MDR-TB face
many challenges including stigma, exclusion from their
families due to hospitalisation, and dealing with the
emotional aspects of losing family members. An improved understanding of these challenges in this vulnerable group may help target supportive interventions.
Adolescents overall had a positive view of participation
in research, and were very interested in sharing their
experiences with other adolescents, which is key to
planning research and clinical care.
PC-1140-06 Improving pharmacovigilance of

drug-resistant tuberculosis drugs
N Misra,1 I Master1 1King Dinuzulu Hospital Complex,
Durban, South Africa. Fax: (27) 31 242 6008. e-mail:
nirupa.misra@kznhealth.gov.za
Background and challenges to implementation: King

Dinuzulu Hospital Complex (KDHC) is the centralized
drug resistant tuberculosis (DR-TB) centre in EThekwini,
Kwazulu Natal (KZN), South Africa (SA). KDHC has a
high burden of disease with patient numbers increasing
from 205 in 2000 to 686 cases in 2006 to approximately
2500 patients monthly in 2014. Activation of decentralized sites and satellite sites has increased the need to
S468
strengthen pharmacovigilance activities in EThekwini.

The armamentarium for drug resistant tuberculosis
consists of old drugs, repurposed drugs and new drugs.
Treatment regimen consists of many drugs with overlapping toxicities. This coupled with antiretroviral therapy
in patients co-infected with HIV increases the risk of
adverse drug reactions (ADRs). Despite most patients
presenting with ADRs, the reporting of these at facility
level is very poor as seen by the small number of nonARV ADRs that are reported by facilities to the District
Health Information System (DHIS) compared to ARV
related ADRs. Strengthening ADR reporting in EThekwini is a priority especially with the introduction of new
drugs with limited experience. Using the model of
targeted spontaneous reporting adopted in the ARV
programme, targeted spontaneous reporting of ADRs
related to DR TB medicine is being proposed in
EThekwini.
Intervention or response: Almost every patient initiated
on DR-TB treatment will experience some form of ADR
ranging from mild nausea to severe life threatening
conditions. The completion of ADR reports for every
ADR experienced is not practical and because of the
large number of drugs completion of the standardized
MCC ADR reporting form is too time consuming.
Although a dedicated pre-printed DR-TB ADR reporting
form was introduced by NDOH challenges still existed
with reporting due to the high workload of doctors and
uncertainty as to what to report. The form was reviewed
and changes made to make it more user friendly. A core
list of ADRs that should be reported was developed by
the experts at KDHC and training was conducted on
spontaneous targeted reporting.
Results and lessons learnt: A list of ADRs was developed
to guide clinicians as to what ADR to report. A preprinted DR-TB ADR reporting form was implemented.
The number of ADRs related to DR-TB treatment
reported increased following the intervention.
Conclusions and key recommendations: Targeted spontaneous reporting will improve pharmacovigilance activities in EThekwini.
52. Community-based and selfadministered treatment of drug-resistant

TB
PC-1141-06 A causal loop analysis of the rollout
of community-based MDR-TB care in South
Africa
R Chimatira1 1Right to Care, Johannesburg, South Africa.,
e-mail: raymond.chimatira@gmail.com
Background: South Africa adopted the community-based

MDR-TB care model in 2009. This paper describes an
exploratory operational research project using a qualitative system dynamics approach to identifying the
factors potentially influential on systems performance
for the effective delivery of decentralised DR-TB care at
district level.
Design/Methods: The methods and samples included

review of meeting reports, programme documents and
reports, policy documents, and interviews with provincial, district and sub-district managers and TB Coordinators; healthcare workers providing services in the
facilities. Thematic content analysis was used to minimally organise and describe the data set in rich detail,
and to interpret various aspects of the research topic.
Results: Results from the analysis identified key behavioural, structural, communication and relationship issues
that together determine the quality of services delivered.
The following major recurring systems barriers were
identified: poor communication between levels of care,
poor knowledge of DR-TB programmatic management
(PMDT) at district level and clinical management at
decentralised levels, DR-TB recording and reporting
challenges, lack of financial and other resources (e.g.
availability of vehicles for outreach teams; inadequate
infrastructure and equipment for infection control). A
causal loop diagram (CLD) was developed as an
analytical tool to attempt to understand how the issues
interact through a series of positive and negative
feedback loops to drive quality at a systems level.Using
this framework as a guide, we can develop an understanding of the key issues, opportunities and obstacles
that drive quality in decentralised MDR-TB management
and to systematically explore TB control intervention
options. Solutions include capacitation on PMDT
(including communication systems and referral links
between levels of care), MDR-TB clinical training for
health care workers (including infection control, side
effects monitoring, contact tracing), and regular reviews
of TB and DR-TB programmes, and integration of TB
services into primary health care teams.
Conclusion: This pilot project demonstrates the utility of
the system dynamics method, rather than rigorous policy
analysis based on empirical evidence. The approach
offers a way to make sense of the complex relationships
and information flows between multiple stakeholders.
The CLD framework will help to enhance the quality of
traditional TB control efforts.
PC-1142-06 Self-administered treatment for the

continuation phase of drug-resistant
tuberculosis treatment: feasibility and impact
on patient outcomes
E Mohr,1 H Cox,2 L Wilkinson,1 G Van Cutsem,3 V Cox,1
J Daniels,1 O Muller,1 B Beko,1 J Hughes1 1Medecins

Sans
`
Frontieres
(MSF), Khayelitsha, 2University of Cape Town,
Cape Town, 3MSF, Cape Town, South Africa. e-mail:
mohrek27@gmail.com
Background: Daily directly observed therapy (DOT) is

recommended for drug-resistant tuberculosis (DR-TB)
patients throughout treatment. This potentially poses a
barrier to treatment adherence, particularly during the
continuation phase of treatment when patients have
clinically improved and can return to normal activities. A
pilot program to provide community supported, selfadministered treatment (SAT) for patients in the continuation phase of treatment was initiated progressively
from January 2012 as part of decentralized DR-TB
treatment in Khayelitsha, South Africa. In order to assess
feasibility and impact of SAT, we compared 12 month
outcomes among patients receiving treatment from
primary care clinics with the SAT program to those from
clinics still using DOT.
Methods: All DR-TB patients still receiving treatment at
the end of the intensive phase within four clinics offering
SAT from January 2012 through December 2013 were
compared to those in the remaining five clinics using
DOT. Patients were assessed for eligibility for SAT by
clinicians based on previous adherence record and
clinical condition. SAT patients were assigned a community treatment supporter who maintained contact on a
weekly basis, and patients were seen at their clinic
monthly to monitor treatment progress and collect
medication.
Results: There were 163 patients entering the continuation phase in SAT clinics and 160 patients in DOT
clinics during the study period. Due to phased implementation of the pilot 97/163 (60%) patients were
considered for SAT. Eighty two (50%) were enrolled for
SAT; reasons for exclusion of 15 patients from SAT
included adherence concerns, patient choice, HIV-related
issues, or poor clinical condition/potential treatment
failure. Four of the 82 (5%) patients later returned to
clinic DOT due to adherence concerns. Rates of LFT in
the first 6 months of the continuation phase were 10%
(15/158) compared to 11% (16/153) for all patients in
SAT versus non-SAT clinics, respectively. Similarly, there
was no difference in mortality with 4% and 5% death
rates in SAT and non-SAT clinics, respectively.
Conclusion: While the LFT rate remains high, preliminary outcomes for patients receiving SAT are not worse
than for clinic based DOT. It was feasible to integrate the
pilot into existing primary health care services and
findings suggest that SAT may improve quality of life for
adherent DR-TB patients, while decreasing the burden
placed on clinics by compulsory daily DOT.
S469
PC-1143-06 Decentralization of drug-resistant

tuberculosis: translating policy into practice
N Misra,1 I Master,1 H Sunpath,2 S Fynn2 1King Dinuzulu
Hospital Complex, Durban, 2Ethekwini District Office,
Durban, South Africa. Fax: (27) 31 242 6008. e-mail:
nirupa.misra@kznhealth.gov.za
Background and challenges to implementation: King

Dinuzulu Hospital Complex (KDHC) is the centralized
drug resistant tuberculosis (DR-TB) centre in EThekwini,
Kwazulu Natal (KZN), South Africa (SA). Patients
diagnosed with extremely drug resistant tuberculosis
(XDR TB), pre-XDR-TB and multidrug resistant tuberculosis (MDR-TB) treatment failures are referred to
KDHC from facilities throughout KZN. Patients with
MDR-TB from all facilities in EThekwini and three other
districts in KZN also refer to KDHC. Patient numbers
increased from 205 in 2000 to 686 cases in 2006 to
approximately 2500 patients monthly in 2014. Staff
shortages at KDHC and a shortage of beds have resulted
in a waiting list for patients to be initiated on DR-TB
treatment with the time to treatment initiation in excess
of 3 weeks. Activation of decentralized sites to initiate
patients on MDR-TB treatment and activation of
satellite sites to manage stable MDR-TB patients was
identified as the solution to this challenge.
Intervention or response: Despite policy to guide
decentralization, implementation in KZN was slow.
Barriers included fear of DR-TB, lack of training and
capacity. A steering committee was established to drive
the process. It was decided that implementation will be
done in a collaborative way with stakeholder buy-in
being key to success. A facility assessment tool was
designed and sites assessed for readiness. Workshops
were convened with the multidisciplinary team from
each site. The results of the assessment were presented
and draft standard operating procedures (SOP) to guide
decentralization were presented. The SOPs were finalized
following robust discussions on the practicality of
implementation. Training was also provided on pharmacovigilance and clinical management of TB.
Results and lessons learnt: Criteria for down referral of
patients to satellite sites were finalized. Four CHCs were
identified as ready to manage patients down referred.
SOPs were finalized and signed off by the District
Manager. Three sites in EThekwini and one each in
Amajuba, UThukela and ILembe were activated as
decentralized sites. Stable MDR-TB patients are being
transferred to satellite sites with pre-dispensed medicine
and patients are being initiated at decentralized sites
Conclusions and key recommendations: Discussions of
implementation challenges and barriers prior to finalization of the SOP is key to success. Translating policy into
practice must be done in a collaborative way.
S470
PC-1144-06 HIV and drug-resistant TB comanagement: the importance of strengthening

pharmacovigilance in South Africa
N Ndjeka,1 S Berrada,2 M Dheda,1 J Jooste2 1National
Department of Health, Pretoria, 2SIAPS, Management
Sciences for Health, Hatfield, South Africa. e-mail:
sberrada@msh.org
Background and challenges to implementation: As the

high burden of drug resistant tuberculosis (DR-TB) is
putting increasing pressure on the health system, the
South African National Department of Health (NDoH)
has identified the need to decentralise the management of
these DR-TB patients. In the South African context of
high HIV burden, strengthening clinical and pharmaceutical management is critical to support the effective
implementation of the decentralised model of care. With
new medicines, complex DR-TB regimens and various
co-morbidities, the detection, assessment, understanding
and prevention of adverse drug effects are not sufficient
to improve patients outcomes. Pharmacovigilance needs
to include a management component of the adverse
reaction (ADR) that will take into consideration many
patient, personnel, drug interactions, concomitant conditions and health infrastructure-related confounders.
Intervention or response: NDoH has developed, piloted
and rolled out a decentralised antiretroviral therapy
(ART) pharmacovigilance model. The decentralised
pharmacovigilance model is based on multidisciplinary
clinical clusters where a hospital and its feeder clinics
work together to manage ADRs. The clinical teams meet
once a month to review the ADR reported by the hospital
and the clinics. The ADRs are discussed and recommendations made on the most appropriate clinical management for these patients.
Results and lessons learnt: This patient centered approach to pharmacovigilance appears to be perfectly
suited to the decentralisation of DR-TB patient care.
Building on the lessons learned from the ART programme, the decentralised pharmacovigilance model has
now being tested in three DR-TB hospitals in three
provinces in South Africa. The mixed responses to the
pilot highlighted the need to customise the data
collection forms and training approach to suit the
complexity of the DR-TB programme.
Conclusions and key recommendations: The patient
centered approach for pharmacovigilance has proven
successful for the HIV programme. Although the lessons
learned from the ART programme can be applied to DRTB programme, customisation of the approach is needed
to ensure a successful implementation of the decentralised pharmacovigilance for DR-TB and improved management of the patients.
PC-1145-06 Community-based M/XDR-TB care

in South Africa: feasibility and lessons learnt
`
J Hughes1 1Medecins
Sans Frontieres,
Cape Town, South
Africa. e-mail: msfocb-khayelitsha-tbdoc@brussels.msf.org
Background and challenges to implementation: In the

context of routine hospitalisation and delayed access to
treatment for patients diagnosed with drug-resistant

tuberculosis (DR-TB) in South Africa, a decentralised
model of care was implemented in collaboration with
local stakeholders in Khayelitsha in 2007 to enable
clinically stable DR-TB patients to be diagnosed and
managed by clinicians at a primary health care (PHC)
level.
Intervention or response: Primary care facility staff were
trained, DR-TB monitoring and evaluation systems
established, treatment regimens optimised, patient support staff employed and infection control measures
enforced, while ensuring integration into existing TB
and HIV services.
Results and lessons learnt: Management and responsibility of the decentralised DR-TB programme was
handed over to local authorities at the end of 2013
following notable successes such as increased DR-TB
case detection (~200/year), improved treatment initiation rates (.95% of those diagnosed), reduced time from
diagnosis to treatment initiation (72 days in 2007; 6 days
in 2014) and implementation of a routine standardised
counseling model for all patients starting treatment.
Costing studies indicated that this decentralised model of
care is 42% less costly than a fully centralised hospital
model. While treatment outcomes remained largely the
same as those reported from other areas in South Africa,
a higher proportion of prevalent DR-TB cases in the
community were able to access care due to decentralisation of services. Despite intensive patient support, loss
from treatment rates remain high (~30%), suggesting
that current toxic, lengthy, and poorly efficacious
standardised DR-TB regimens remain a major challenge.
Conclusions and key recommendations: Clinically stable
DR-TB patients can feasibly be managed in PHC and
within their community however treatment outcomes are
still adversely affected by the many challenges related to
provision of adequate adherence support for DR-TB
patients, as well as access to shorter, more tolerable and
more effective treatment regimens. Innovative patient
support strategies along with better access to improved
treatment regimens are required to improve outcomes
and optimise DR-TB care within the community.
PC-1146-06 The impact of decentralization of

Programmatic Management of Drug-resistant
Tuberculosis services in Swaziland
Z Tfwala,1 N Dlamini,1 G Mchunu,1 F Khumalo,1
W Sikhondze1 1Ministry of Health, Mbabane, Swaziland.
e-mail: tmzethy@gmail.com
Background and challenges to implementation: Drug

resistant TB presents a major challenge for TB control
globally. TB patients with drug-resistant strains require
more resources to diagnose and treat successfully than
those without. In 2009, Swaziland conducted a Drug
Resistance Survey (DRS) which revealed a high prevalence strains resistant to both isoniazid and rifampicin
(multidrug-resistant; MDR-TB) among new patients at
7.7% - more than double the 3% average of the African
Region -and a prevalence of 33.7% among previously
treated patients, three times higher than the regional

average of 12 13 %. This implies that about 390 (range:
240-530) MDR-TB cases are expected among new
pulmonary TB patients notified by the programme in
2013 and another 250(210-290) among previously
treated cases. For a country with an estimated population
of 1.2 million this represents a formidable concern.
Intervention or response: In response to MDR-TB
situation, the National TB Control Programme (NTCP)
developed and successfully implemented a plan for the
decentralization of DR-TB services and established
Programmatic Management of DR-TB (PMDT) according to WHO guidelines to provide guidance for DR-TB
management. Presented is the review of PMDT by WHO
as part of an end-evaluation of the National TB Strategic
Plan (2010-2014) in April 2014.
Results and lessons learnt: Since the decentralization of
PMDT services, there has been a progressive improvement in the treatment success rate among MDR-TB
patients from 18% in the 2007 cohort to 58% currently
in the 2011 cohort. PMDT services have been decentralized from one central National TB Hospital in 2010 to 8
more regional DR-TB health facilities by the end of 2014.
These facilities have been reinforced to provide quality
comprehensive treatment care and support services to all
DR-TB patients across the country including community
outreach services and home assessments for Infection
Prevention and Control.
Conclusions and key recommendations: Increasing investment of resources to intensify treatment adherence
through community and family support for MDR-TB
patients, as well as improving access to diagnosis,
treatment and care is key to DR TB control. Swaziland
is now planning to repeat a nationwide, representative
drug resistant survey in 2016 to determine the burden of
drug resistance in the country and possibly any changes
since the last survey in 2009.
PC-1147-06 Community-based management of

MDR-TB patients: experience from Malawi
H S Kanyerere,1 J Mpunga,1 H Banda,2 C Kangombe,1
I Dambe1 1Ministry of Health, National TB Program,
Lilongwe, 2Research for Equity and Community Health Trust,
Lilongwe, Malawi. Fax: (1) 301 941 8427. e-mail:
asmitharthur@urc-chs.com
Background: Malawi started routine surveillance of

MDR-TB among previously treated TB cases in late
1990s. All reported cases were only offered isoniazid and
ethambuttol (as these drugs were believed to reduce
excretion of Mycobacterium tuberculosis). The NTP
planned for an MDR-TB survey 2001 and successfully
applied to the Green Light Committee for access to
concessional quality 2nd line TB drugs for communitybased management of MDR-TB patients.
Intervention: In 2007, NTP adapted WHO MDR-TB
treatment guidelines including recording and reporting
tools. District teams were trained in MDR-TB management. Malawi Government procured the first consignment of 2nd line drugs. Since there are no specialised
S471
facilities in Malawi, NTP has opted for communitybased management of MDR-TB patients.
Results: In September 2007, 8 MDR-TB patients were
initiated on 2nd line TB drugs. By December 2007, 11
MDR-TB patients were on treatment. Since 2007, 186
patients have been diagnosed with MDR-TB and
120(65%) have been started on 2nd line TB treatment.
52 patients (28%) died before treatment initiation.
Majority of patients diagnosed were previously treated
patients with 1st line anti-TB drugs. In 2014 the NTP
diagnosed 17 MDR-TB patients and 13(76%) were
started on 2nd line treatment. On average, it takes 4
weeks from the time of MDR-TB confirmation to start of
treatment. The HIV status varied from 73% in 2007 to
18% in 2014. Almost all MDR-TB patients had their
specimens sent to a supranational laboratory for
confirmation and 2nd line drug susceptibility testing.
No XDR-TB patient has been reported. The treatment
success rates among MDR-TB patients have varied from
62% to 67%. For 11 MDR-TB cases started on
treatment in 2007, only 7 (63%) were successfully
treated. On the other hand, the death rates have been
very high ranging from 20% to 38%. Psychosocial and
nutritional supplementation is the unmet need for the
patients.
Conclusions and recommendations: Community-based
management of MDR-TB patients is feasible in Malawi;
however, more needs to be done. There is need to
expedite diagnosis and treatment initiation to reduce
time leads for the same. Treating these patients at a
specialized MDR-TB facility may improve the treatment
outcomes since most patients may have access to
psychosocial and nutritional support including management of adverse effects. Furthermore, continuous inservice training on MDR-TB patient management for
front-line staff is critical
MDR-TB cases registered and treatment outcomes (2007 to
2011)
MDRTB
cases
Year reg
2007
2008
2009
2010
2011
2012
2013
2014
17
16
15
40
26
27
28
17
No.
started
on
Treatment
11
16
9
26
15
19
11
13
(65%)
(100%)
(60%)
(65%)
(58%)
(70%)
(39%)
(76%)
Treatment
success
7
10
6
18
10
(64%)
(63%)
(67%)
(62%)
(67%)
Died
3
6
3
7
3
Lost to treat- Not

follow- ment evaluup
failed ated
(27%) 1 (9%)
(38%)
0
(33%)
0
(27%) 1 (4%)
(20%) 1 (7%) 1
0
0
0
0
(7%)
0
0
0
0
0
S472
PC-1148-06 MDR-TB treatment outcomes: tasksharing is effective in improving access

without compromising outcomes
J Farley,1 N Ndjeka,2 C Budhathoki,1 E Bergren,1,2
N Mlandu,3 M Van Der Walt4 1Johns Hopkins University,
Baltimore, MD, USA, 2National Department of Health,
Pretoria, 3Jhpiego-South Africa, Pretoria, 4Medical Research
Council, Pretoria, South Africa. e-mail: ebergre1@jhu.edu
Background: Treatment outcomes for multidrug-resistant tuberculosis (MDR-TB) in South Africa remain
poor, particularly in patients with HIV. Interventions that
increase access to quality, timely care are essential. Tasksharing by a clinical nurse practitioner (CNP) and a
medical officer (MO) of MDR-TB-HIV patients is a
potential human resource solution, yet evidence is
required to support this approach.
Design/Methods: Treatment outcomes were evaluated
among MDR-TB patients jointly managed by a CNP and
an MO in an MDR-TB unit in a peri-urban city of
KwaZulu Natal, South Africa. We followed a prospective
cohort who began treatment between January 1 and
December 31, 2012 and followed through January 31,
2015 for treatment outcomes. Patients were assigned a
primary provider (CNP or MO) with task-sharing
throughout patient care. A competing risk analysis with
death, failure or default as competing negative outcomes
is evaluated including primary provider assignment as an
independent variable. A z-test was used to compare the
cohort treatment success rate with the national norm in
2012.
Results: We enrolled 197 patients, 51% females with
75% unemployment with a mean age of 33 years and
mean BMI 19.7. HIV co-infection was 74% (median
CD4 count 250, Q3-Q1 385-141); 75% on ART for
those HIV at baseline. MDR-TB treatment outcome
included 58% success (i.e. cure/completion), 6% failure,
17% death, and 20% default. In comparison, treatment
success for the 2012 cohort across South Africa is 45%;
success rate in our study cohort was significantly higher
than the national norm (z 3.70, P , 0.001). In the
competing risk analysis, there was no difference by HIV
infection status for either failure or default after
adjusting for age, sex and provider type. However,
hazard for death was reduced by 64% in HIV negative
patients (P 0.044).
Conclusion: Overall outcomes in this cohort with tasksharing between CNPs and MOs were better than the
country average for the same year. Despite high ART use,
HIV continues to hasten death among this group.
PC-1149-06 High treatment success for

multidrug-resistant tuberculosis in Africa:
initiation and scale-up of the Ethiopian
countrywide collaborative treatment
D Kokebu,1,2 J Andrews,3 R Hurtado,2,4 E D Ejara,2,5
K Abato,2 T Daniel,2 T Sok,6 A Goldfeld2,7 1St Peters
Tuberculosis Specialized Hospital, Addis Ababa, 2Global
Health Committee, Addis Ababa, Ethiopia; 3Stanford
University School of Medicine, Palo Alto, CA, 4Massachusetts
General Hospital, Boston, MA, USA; 5University of Gondar
Hospital, Gondar, Ethiopia; 6Cambodian Health Committee,
Phnom Penh, Cambodia; 7Childrens Hospital, Boston,
Massachusetts , USA. Fax: (1) 617 726 7653. e-mail:
rh1927@gmail.com
Background: Fewer than half of patients receiving

therapy for multidrug-resistanttuberculosis (MDR-TB)
in Africa are successfully treated, with particularly poor
outcomes reported for HIV co-infected patients. We
report outcomes from the first four years of scale-up of
MDR-TB treatment in Ethiopia.
Methods: A standardized second-line drug (SLD) regimen was used in an NGO-MOH collaborative community- and hospital-based program that provided initial
courses of SLD during program start-up and included
intensive side effect monitoring, home-based adherence
strategies and nutritional supplementation. Clinical
charts and outcomes for patients with at least 24 months
of follow-up were reviewed and predictors of treatment
failure or death were evaluated by Cox proportional
hazards models.
Results: From February 2009 to December 2014, 1044
patients were initiated onSLD. 612 patients had at least
24 months of follow-up, 565 (92.3%) had confirmed
MDR-TB and 44 (7.7%) were suspected MDR-TB cases.
603 (98.5%) had prior TB treatment, 133 (21.7%) were
HIV co-infected, and median BMI was 16.6. Three
hundred and ninety-six (64.7%) were cured, 85 (13.9%)
completed treatment (composite treatment success:
78.6%), 10 (1.6%) failed, 85 (13.9%) died, and 36
(5.9%) were lost to follow-up. Of 479 HIV un-infected
patients, 388 (81.0%) were successfully treated and 58
(12.1%) died and of 133 HIV co-infected patients, 93
(69.9%) were successfully treated and 27 (20.3%) died.
HIV co-infection (AHR 2.48, P , 0.001) and cor
pulmonale (AHR 6.98, P , 0.001) were predictive of
treatment failure or death.
Conclusion: We report here the highest MDR-TB
treatment success outcomes to be achieved to date in
Africa in a population with advanced disease and
profoundly low BMI in a setting with severe resource
constraints, through an NGO-MOH partnership program using a standardized second-line regimen. Intensive
treatment of side effects, nutritional supplementation,
and adherence strategies facilitated by flexible funding,
were key strategies.
53. Infection control

PC-1150-06 Tuberculosis and its presentation in
in-patient units of a university hospital
W De Andrade Pereira De Brito,1 F Dias E Sanches,1,2
1 V Gutterres Silva1 1Pedro Ernesto

C Cangani
De Araujo,
University Hospital / UERJ, Rio De Janeiro, RJ, 2Thorax
Diseases Institute / UFRJ, Rio De Janeiro, RJ, Brazil. e-mail:
enfermeirofernando@hotmail.com
Background and challenges to implementation: Controlling tuberculosis (TB) in Brazil and in the world still
represents a challenge to be conquered, especially in Rio
de Janeiro where it is very commom to find a TB case at
hospital it could represent an extra issue to be solved.
Intervention or response: This descriptive study with a
quantitative approach was developed after another study
at the same hospital with the intention to verify the
impact of the first study. The following objectives:
Identify the inpatient units at risk for exposure to TB;
analyze the number of sputum smears and cultures
required in inpatient units; classify the samples of sputum
smears and positive cultures of the type of TB; correlate
the TB cases reported with positive samples of sputum
smear and cultures. The data was collected during
January-September 2012.
Results and lessons learnt: Clinical units with the highest
rate of cases were found in an internal medicine ward
where there isnt a proper isolating room. Lack of
correlation of institutional profile with the data presented in the municipality of Rio de Janeiro which means
undernotifications. It still remains an urgent need to
continue speaking about TB with patients and theirs
visitors, during shifts with healthcare workers and
residents; as well as classes with students. Those
strategies should be put in practice to avoid the sense
of trivialization and / or no perceived risk of infection
and subsequent illness.
Conclusions and key recommendations: More and more
people are at risk. There isnt enough measures to
evaluate a real impact of TB at hospital. There isnt
enough quantitative of isolating rooms. Need for
development new strategies focused on biosafety. It is
necessary create foruns for discussion under multiprofessionals perspectives.
PC-1151-06 Implementation of infection

control practices to prevent TB transmission
among PLHIV in five healthcare facilities,
Belize, 2014-2015
D Forno,1 F Morey,2 R Lorenzana,3 M Briggs,4 V Lipke4
1
Centers for Disease Control and Prevention, Guatemala City,
Guatemala; 2Belize Ministry of Health, Belmopan, Belize;
3
University Research Company, Guatemala City, Guatemala;
4
GA, USA. e-mail: dforno@cdc.gov
Background: Belize has the highest HIV prevalence

(1.5%) and TB-HIV incidence (8.6 per 100 000) in
Central America. The most common infectious cause of
S473
death among people living with HIV (PLHIV) is TB and

PLHIV are at high risk of contracting all types of TB
disease. Risk for TB transmission is higher in crowded
facilities where staff does not practice TB infection
control (IC) activities. Based on World Health Organization policy, we implemented the CDC TB IC training
package in health facilities to reduce TB transmission.
Design/Methods: The TB IC package was implemented
in the two highest burden districts for TB-HIV in Belize:
three facilities in Belize district and two facilities in Cayo
district. The package included a toolkit to conduct TB IC
risk assessments and to develop site-specific TB IC plans,
training materials, visual aids for clinics, and an
educational video. Facilities were assessed with a 37
item risk assessment tool measuring managerial, administrative, environmental, and personal protective equipment (PPE) categories at baseline.
Results were classified as red if the activity was
implemented incorrectly; yellow if it was in progress;
or green if it was implemented correctly. A total of 25
healthcare workers received a three-day TB IC training
and completed a 13-question pre-/post-test to assess their
knowledge.
Results: Two baseline TB IC activities scored red across
all sites: separation of coughing patients and outdoor
collection of sputum samples (Table 1). Four sites had a
National IC policy, an IC committee, and an IC nurse
assigned to the site, but had no site-specific TB IC plans.
Across sites, an average of 39% of managerial, 35% of
administrative, 29% of environmental, and 66% of PPE
activities were implemented correctly. Average pre and
post-test scores improved from 77% to 92%. A specific
TB IC plan was developed for each site.
Conclusion: The standardized risk assessments showed
poor implementation of several TB IC activities. An
increase in knowledge was shown in most trainees. Sitespecific TB IC plans developed during the training will be
implemented to increase TB IC activities. We will
conduct a six month follow-up evaluation to assess
improvement in TB IC practices following each clinics
plan.
S474
PC-1152-06 Airborne infection control within

drug-resistant tuberculosis centres in Tamil
Nadu, India: does practice follow policy?
PC-1153-06 Natural human-to-guinea pig

model to test TB transmission interventions in
South Africa: ten years later
S R Holur,1 P Rahul,2 D Ranganathan,3 M Lakshmi4

World Health Organization Country Office, Chennai, 2DTO
Office, Vellore, 3Madras Medical College, Chennai, 4State TB
Office, Chennai, India. e-mail: srkholur1979@gmail.com
A Stoltz,1 E De Kock,2 J A Kruger,3 S Keulder,1

A Dharmadhikari,4 P Jensen,5 I Orme,6 E Nardell4
1
University of Pretoria, Pretoria, 2Retrasol, Pretoria, 3Rietvallei
Animal Hospital, Pretoria, South Africa; 4Harvard Medical
School, Boston, MA, 5U.S. Centers for Disease Control and
Prevention, Atlanta, GA, 6Colorado State University, Fort
Collins, CO, USA. e-mail: anton.stoltz@up.ac.za
Background: In 2012, more than 300 000 multidrug

resistant tuberculosis (MDR-TB) cases occurred globally,
of which, approximately 20% were in India. In India, all
MDR-TB cases are required to complete at least 7 days of
inpatient treatment at Drug-resistant Tuberculosis
(DRTB) centres during the initiation of treatment. The
Revised National Control Programme (RNTCP) has
published airborne infection control (AIC) guidelines.
This study aims to assess the AIC knowledge and
practices among health staff and MDR patients within
DR-TB centers.
Design/Methods: We evaluated airborne infection control procedures and practices using a standard checklist
developed by the RNTCP designed to capture the three
major tenants of AIC: administrative controls, environmental controls and availability and use of personal
protective equipment (PPE) in the 6 DR-TB centres in
Tamil Nadu. We assessed the knowledge, attitudes and
practices of all DR-TB centre staff using a standardized
questionnaire. We also selected a convenient sample of
patients undergoing treatment at each DR-TB center to
better assess patient AIC knowledge and practices using a
standardized questionnaire.
Results: dministrative controls: 3(50%) of DR-TB
centres had formal Infection Control Committees in
place and among those, only 1 (17%) had routine and
standing meetings to review performance and practice.
1(17%) had developed a site-specific infection control
plan available. Only 1 center conducted annualized
training and refreshers in AIC. Environmental controls:
air ventilation and exchange (natural or mechanical) was
observed to be adequate at all centers, however, none of
the centers had upper room ultraviolet light for
germicidal irradiation or HEPA-filtration. PPE: Respirators were available in only 2 of 6 centers, however, none
conducted recommended fit testing. A total of 34 staff
members were interviewed. Only 6 (18%) were aware of
AIC, and only 3 (8%) had received formal AIC training.
About a third of staff, (12/34) felt that isolating TB
patients from patient with HIV was not necessary.
Among the 21 patient interviewed, 12 (57%) were not
aware of proper cough etiquette.
Conclusion: This study highlights the urgent need for
immediate standardized AIC training of the administrators and all cadres of staff working in India, especially
those working in high-risk settings like DR-TB centers.
Background: A quarter of the global 9.0 million reported

TB cases are found in the African Region. Hospitals and
other congregate settings are highly efficient sites of TB
transmission, but most recommended interventions have
not been validated in real hospital settings. For the past
10 years we have employed the well-established humanto-guinea pig transmission model of Wells and Riley to
assess a series of conventional and novel infection control
interventions.
Design/Methods: The Sidney Parsons Airborne Infection
Research Facility was established in 2005 to test TB
transmission interventions (UVGI air disinfection, masks
on patients, effective treatment, BCG vaccine, inhaled
drugs, and portable room air cleaners) to prevent TB
transmission. In most experiments, 90 guinea pigs
(intervention chamber) are exposed to exhaust air from
a 6-bed TB ward on days when the intervention being
tested is turned on, and another 90 guinea pigs (control
chamber) are exposed to the exhaust air from the same
ward on alternate days when the intervention is turned
off. The difference between the rates of guinea pig
infections in the two chambers, determined by tuberculin
skin testing, is a precise measure of the efficacy of the
intervention. The alternate day strategy controls for
temporal differences in patient infectiousness. In each
study, up to 6 patients at a time occupy the facility
serving simply to generate aerosolized Mtb. Admission
criteria include subjects recently diagnosed and early in
treatment, preferably sputum smear positive with lung
cavitation. Most subjects have MDR-TB, but some turn
out to have XDR-TB, and are the most likely to transmit
on conventional treatment.
Results: As previously reported, UVGI proved to be
highly effective (80%), masks on patients (56%), but
room air cleaners and inhaled colistin dry powder were
ineffective. As shown by Riley, effective treatment was
shown to be rapidly highly effective. Preliminary results
of BCG vaccination of guinea pigs to prevent infection
demonstrate some efficacy.
Conclusion: As first conceived of by Wells and Riley, the
natural human-to-guinea pig model has for the past 10
years proven to be highly effective in testing conventional
and novel TB transmission interventions and in understanding TB transmission.
PC-1154-06 Results of a baseline infection

control assessment at identified decentralized
MDR-TB facilities in South Africa
A Peters,1 J Farley,1,2 J Pienaar,1 M Mphahlele,1 I Asia1
1
Jhpiego SA, Pretoria, South Africa; 2Johns Hopkins
University, Baltimore, MD, USA. e-mail:
jacqueline.pienaar@jhpiego.co.za

Africas MDR-TB burden ranks third highest in the
world. South Africa also has the highest number of
people living with HIV. Before 2011, national guidelines
mandated that all DR TB patients be initiated on
treatment in the countrys handful of specialized TB
hospitals. New cases outstripped the bed capacity and
the country had to move to the decentralized approach to
the transfer of responsibility for treating MDR-TB
patients to lower levels of care. Infection Control (IC)
is one of the requirements for decentralised MDR-TB
care. The objective of the study is to conduct operational
evaluations of IC in identified MDR-TB decentralized
facilities.
Intervention or response: A cross- sectional descriptive
study of 56 decentralized MDR-TB facilities was
conducted from October 2014 to March 2015 using a
standard assessment instrument comprising of five
categories (IC officer, fit testing, IC Policy, UVGI, N95
Respirators availability).
Results and lessons learnt: Analysis of the 56 clinics
assessed revealed that 14 (25%) have IC officers; 0 (0%)
have done fit testing for N95 Respirators; 2 (3, 57%) do
have IC policies in place, 2 (3, 57%) have working UVGI
lights and only 5 (8, 93%) have N95 respirators
available. One facility implemented 4 of the 5 requirements (1, 79%); 1 implemented 3 of the five requirements (1, 79%); 1 implemented 2 of the five requirements (1, 79%) and 11 implemented 1 requirement (19,
64%) while 42 did not implement any of the requirements (75%). IC practices were poorly implemented in
most of the facilities.
Conclusions and key recommendations: These findings
demonstrate the need to improve infection control
practices and avail resources in identified decentralized
MDR-TB settings before MDR-TB decentralization can
safely and effectively be implemented.
PC-1155-06 A comparative study of the design

of two hospitals on the Cape Flats with regard
to user risk of contracting TB
K Bingham1 1University of Pretoria, Pretoria, South Africa.
e-mail: kgbingham@gmail.com
Background: The incidence levels of Tuberculosis

amongst the population of the Cape Flats, Cape Town,
South Africa, is well documented. This area is served by 2
hospitals, viz. Khayelitsha and Mitchells Plain Hospitals,
both having been designed from a similar architectural
brief, but being markedly different in composition. One
hospital is high-rise with mechanical ventilation, while
the other is low-rise, being predominantly reliant on
S475
natural ventilation, with only selected areas of mechanical ventilation.

Design/Methods: This study compares the architectural
responses of the 2 facilities and the inherent risk of lung
disease through air-borne infection due to these responses. The contextural environment, and its climatic
conditions and prevailing winds are recorded and
compared with the footprints and layouts of the 2
facilities, to assess the impact of external air-flow on the
buildings. Selected floor plan layouts, spatial volumes,
patient flow, and the design of public-staff interfaces are
considered, as well as the intervention of forced draught
ventilation and directional air flow. This is weighed
against the design of room fenestration and the provision
of natural ventilation, and the inherent uncertainty that
this brings.
Results: The mechanically ventilated areas, where
extraction is provided near to areas of infection risk,
allows for the least risk to staff and visitors. Where
provision has been made for natural ventilation, the
effect of the external environment has an adverse effect
of air-flow patterns within the facilities, negating the
design intentions of mechanical systems and producing
the risk of infection to others, particularly in hot climatic
conditions.
Conclusion: Mechanical ventilation and effective differential pressure zones are essential in all health facilities
where there is the risk of infection through air-borne
diseases. Uncontrolled natural ventilation adds to the
risk of infection.
PC-1156-06 Establishment of a TB infection

control demonstration site and training base
for better TB control in Tajikistan
Z Tilloeva,1 O Bobokhojaev,2 A Trusov,1,3 A Hovsepyan,1,4
J Ismoilova,1 T Alieva2 1Project HOPE, Dushanbe,
2
Republican TB Center, Tajikistan, Dushanbe, Tajikistan;
3
Project HOPE, Milwood, VA, USA; 4Project HOPE, Yerevan,
Armenia. Fax: (992) 372 273629. e-mail:
ztilloeva@gmail.com
Background and challenges to implementation: Tajikistan is one of the 27 WHO high multidrug resistant
(MDR) TB burden countries with one of the highest
levels of TB prevalence in WHO Europe region. MDR
cases account for 13% of new and 54% of retreatment
TB cases (2011). Within this poor epidemiological
situation, the lack of proper institutional capacity to
adequately address TB infection control (IC) at health
care facilities (HCFs) is putting health care workers
(HCWs), patients, and caregivers at serious risk of TB
and threatening the sustainability of TB control efforts.
Intervention or response: The Ministry of Health
designated Macheton TB Hospital as the National TB
IC Demonstration Site and Base for Training. A
theoretical and legal assessment was conducted by
reviewing existing guidelines, MoH orders and relevant
documents. Training packages were developed. The
program conducted TB IC risk assessments in pilot
HCFs and results were discussed during the trainings. A
national team of 14 trainers was established who further
S476
rolled out 5 cascade trainings for 61 HCWs from 17 pilot

sites. TB IC plans were developed during the trainings
and further implemented in HCFs. A survey on TB status
of HCWs was conducted by screening 3,853 HCWs for
TB symptoms with further TB testing of those suspected
for TB. The first regional TB IC training was conducted
for 15 NTP representatives from Central Asia Republics
(CAR).
Results and lessons learnt: Monitoring visits to pilot sites
showed considerable improvement in TB IC practices in
all areas. All 17 target HCFs have IC focal person
appointed 15 had established IC committees and started
implementation of TB IC plans 12 are monitoring IC
plan implementation Burden of TB among HCWs was
documented: 85% (3853 out of 4547) of HCWs were
screened for TB 86% (122 out of 142) of HCW from
those screened were identified as TB suspects were
further tested for TB and 11 TB cases were detected
Conclusions and key recommendations: The program
results provide evidence on the successfulness of this
approach in documenting the magnitude of the TB
burden at HCFs, establishing effective practices for TB
IC, and building national and regional interest to
propagate these advances. The program approach should
be further expanded to cover the whole country to
contribute to the establishment of effective TB IC control
practices in Tajikistan, with high potential to be
replicated in CAR.
PC-1157-06 Improving TB infection control in

Odesa, Ukraine
S Yesypenko,1,2 A Aleksandrin,1 M Dolynska1 1USAID
Strengthening Tuberculosis Control in Ukraine Project, Kyiv,
2
Regional Central TB Hospital, Odesa, Ukraine. e-mail:
mdolynska@stbcu.com.ua

Odesa oblast (region) of Ukraine demonstrates the
countrys highest tuberculosis (TB) incidence rate
almost twice as high as the national average (98.2 vs.
59.5 per 100 000). As a result, the regional health care
facilities providing TB care are often overstretched.
Facilities are housed in outdated buildings which make
large-scale renovation to facilitate infection control (IC)
challenging, and a shortfall of health service funding
seriously limits environmental control and individual
protection. This leads to a significant risk of occupational TB in such facilities. Between 2008 and 2012, 23
employees of the local TB services contracted TB.
Intervention or response: Since 2012, the USAID
Strengthening Tuberculosis Control in Ukraine project
has been providing education to health care practitioners
and administrators on cost-effective TB IC interventions.
Following the projects suggestions, the managers of the
Odesa Central TB hospital established proper patients
flow and ensured a strict separation of smear and culturepositive patients, and patients with MDR and XDR-TB.
Layouts of the premises were revised to concentrate highrisk zones and minimize patients interaction. As
mechanical ventilation was not accessible in the facility,
facility specialists, after consultations with the project

experts, provided all wards and areas where aerosolproducing procedures are being performed (bronchoscopy, sputum collection etc.), waiting rooms, and X-ray
departments with shielded UV-radiators. Now doctors
and nurses also insist on proper use of natural ventilation
and keeping windows open, as weather permits.
Results and lessons learnt: In 2014, no new TB cases
were registered among employees of Odesas central TB
hospital.
Conclusions and key recommendations: Inexpensive and
easy-to-implement measures are effective in the prevention of occupational TB, which is a principle indicator of
proper TB IC. A safe medical environment is achievable
even with limited funding.
54. TB and nutrition

PC-1158-06 Role of nutritional support in
improving treatment outcomes among MDRTB patients: experience of the Damien
Foundation
S Mugudalabetta,1 S K Muthusamy,1 L Gujral,1
S Satheesh1 1Damien Foundation India Trust, Chennai, India.
Fax: (91) 44 2836 0496. e-mail:
admin@damienfoundation.in
Background and challenges to implementation: Multi

Drug Resistant Tuberculosis (MDR-TB) is closely linked
with poverty and under nutrition. Many patients cannot
afford to have adequate food due to loss of income and
catastrophic health expenditures. There is very limited
evidence available to suggest that nutritional support in
addition to standard MDR-TB treatment, improves
treatment adherence and outcomes.
Intervention or response: Damien Foundation is providing nutritional support worth 6 USD to poor and needy
MDR-TB patients in the form of rice, wheat, cereals, egg,
cooking oil etc once in two months till completion of
treatment. In 2012, nutritional support was given to 51
MDR-TB patients in south west Delhi and six districts of
South Andhra Pradesh in India. Treatment outcomes for
those who received nutritional support along with those
who didnt was retrieved from MDR-TB register and
analyzed.
Results and lessons learnt: In 2012, 206 MDR-TB
patients were registered for treatment in South West
Delhi and six districts of South Andhra Pradesh. Among
them 51 (25%) patients were identified and received
nutritional support. Analysis of treatment outcomes
showed higher treatment success rate of 74.5% (38/51)
among those who received nutritional support along
with standard MDR-TB treatment compared to 39%
(61/155) among those who received only standard MDRTB treatment. Only five patients (10%) who received
nutritional support were lost to follow up compared to
48 patients (31%) among those received only standard
treatment. Number of MDR-TB patients Treatment
success loss to follow up other outcome Received
nutritonal support with Standard MDR-TB treatment 51

38 (74.5%) 5 (10%) 8 (15.5%) Only standard MDR-TB
treatment 155 61 (39%) 48 (31%) 46 (30%) Total 206
99 (48%) 53 (25.7%) 54 (26%)
Conclusions and key recommendations: Providing nutritional support in addition to standard MDR-TB treatment for patients is likely to motive them to complete
treatment and reduce the default rate.
PC-1159-06 A holistic approach to TB

prevention and care: a study in Jabalpur MP,
India
A Kollamparampil,1 P Varghese,1 M Puthenchirayil,1
G Singh,1 R Mathew,2 B Entoor Ramachandran,3 R Babu3
1
Catholic Bishops Conference of India-Coalition for AIDS
and Related Diseases (CBCI-CARD), Delhi, 2Norbetine Social
Welfare Centre (NSWC), Jabalpur, 3International Union
Office, New Delhi, India. e-mail: abhileshcbci@gmail.com
Background and challenges to implementation: Many

TB patients in India are not able a complete their
treatment because of lack of motivation and support
system. Decentralisation of care along with counseling
helps the patient to complete the treatment. Nutrition is
also a very important component for successful treatment of tuberculosis. Normally TB patients in vulnerable
and marginalised areas, work on daily wages. It becomes
difficult for them to earn their daily living and as a
consequence, having nutritious food along with TB
treatment becomes impossible for them. Catholic Bishops Conference of India Coalition for AIDS and Related
Diseases (CBCI CARD) a sub recipient of Global Fund is
implementing a community based project for TB
prevention & care with support from The Union. The
project is being implemented in 29 districts across 4
states of India. Norbetine social welfare centre (NSWC)
is a partner NGO with CBCI CARD Project Axshya in
Jabalpur in Madhya Pradesh India.
Intervention or response: (NSWC) implementing community based activities in Jabalpur in Madhya Pradesh
India. Jabalpur covering a population of 2 591 552. The
major population consist of bidi workers, slum population or living in remote villages. The main interventions
made by NSWC are involving motivated community
members as DOT providers, nutritional supplementation
to poor TB patients with food from their farms,
collection from schools and other sources. NSWC offers
counselling, treatment, care, and treating patients with
dignity and honour. The centre describes this as a
comprehensive holistic approach to TB prevention and
care.
Results and lessons learnt: NSWC in the last 8 years has
groomed 180 community DOT providers and has
successfully treated 650 patients. Special care and
attention to those families with package of TB services,
minimized the lost to follow up to treatment and death
due to TB
Conclusions and key recommendations: The performance of NSWC shows that when DOTS therapy is
accompanied by professional counselling, nutritional
S477
supplementation, individualized approach and dealing

patients with dignity and honour, success of the
treatment is enhanced significantly.
PC-1160-06 Severe malnutrition impairs

treatment response among Indias initial MDRTB cohorts
M Parmar,1 K S Sachdeva,2 K Rade,1,2 P Darsini,3 R Pant,4
A Sreenivas,1 P Dewan5 1World Health Organization
Country Office for India, New Delhi, 2Central TB Division,
MoH FW, Government of India, New Delhi, 3National
Institute of Research in Tuberculosis, Chennai, 4GD Goenka
World Institute, Gurgaon, 5Bill & Melinda Gates Foundation,
New Delhi, India. e-mail: parmarm@searo.who.int
Background: Globally, India has worlds highest burden

of multidrug-resistant tuberculosis (MDR-TB). As rapid
scale-up of Programmatic Management of Drug resistant
Tuberculosis (PMDT) continues in India, initial cohort
analysis revealed low body mass index and weight loss
during treatment as few factors associated with nonconversion of sputum culture and poor treatment
response. Further analysis was undertaken to evaluate
the effect of severe malnutrition on treatment response in
MDR-TB patients treated under Indias PMDT programme.
Methods: Retrospective cohort analysis of MDR-TB
patients diagnosed among failures of first-line treatment,
registered in the national PMDT programme and treated
with a standard 24-months regimen under DOT from
2007 2010, was done. Data on patients height, weight,
clinical and microbiological response were collected;
sputum conversion and treatment outcome were defined
by standard WHO definition. Logistics regression model
was used to determine the risk of BMI on various
treatment outcomes.
Results: Among 3713 (88%) of MDR-TB patients
registered from Aug2007 Mar2011, 2066 (56%) had
some degree of malnutrition with 1314 (35%) showing
severe acute malnutrition (BMI , 16). Proportion of
patients with sputum culture conversion at 4th month
was higher among those with BMI . 18.5 than those
with BMI , 16 (69% vs.59%, P 0.001) while at 6th
month also this difference was statistically significant
(63% vs. 56%, P 0.001). Patients with BMI , 16 had
two times higher risk of being sputum culture positive at
end of 6 months of treatment (OR 2.21, 95%CI 1.84
2.74, P , 0.0001). Among the 1314 severely malnourished patients, treatment outcome was available for 796
patients; of whom 225 (28%) had treatment success, 239
(30%) had unfavourable response as default or failure
while 332 (42%) died. Among 961 patients with BMI ,
18.5, 393(41%) had treatment success 364 (38%) had
unfavourable response while 204 (21%) died.
Conclusion: Severe malnutrition delays sputum culture
conversion and hence risk of transmission is longer and
higher compared to well-nourished MDR-TB patients.
Treatment success is almost halved and death rate
doubled in severe malnourished MDR-TB patients as
compared to the well-nourished. Nutritional intervention to improve MDR-TB patients response to treatment
S478
will thus have a crucial role achieving the goals of End TB

Strategy.
PC-1162-06 Is food support for malnourished

TB patients associated with treatment
outcomes? TB case mortality in Kenya
PC-1161-06 Role of nutrition status in

multidrug-resistant tuberculosis conversion
J Njenga,1 A Wairia,1,2 B Mungai,1 F Ngari,2 E Masini,2

M Maina,3 P Agunga1 1Centre for Health Solutions - Kenya,
Nairobi, 2Ministry of Health, Kenya, Nairobi, 3USAID Kenya,
Nairobi, Kenya. e-mail: jnjenga@chskenya.org
N Librianty,1 D Soepandi,1 D Handayani,1 A Fuady1

1
Faculty of Medicine University of Indonesia, Jakarta,
Indonesia. e-mail: librianty@gmail.com
Background: Recently, the burden of TB problem were

emerged by MDR-TB, WHO estimate there will be 480
000 new cases of MDR-TB worldwide in 2013 and
Indonesia were be in ranks tenth among high MDR-TB
burden countries. The outcomes of MDR-TB was quite
low with success rate was only 48%. Many factors
associated with culture conversion in MDR-TB treatment including nutrition status. Nutrition was one of
modifiable factor, so it was important to evaluate
nutritional status among MDR-TB patients.
Design/Methods: A retrospective cohort study among
patient treated with standardized MDR-TB regimen
between August 2009 and December 2013 in Persahabatan Hospital. The outcomes of this study are conversion time and association of BMI and HB level with
conversion time and other factors. Data were analyzed
using Kaplan-Meier curves and cox regression analysis.
Ethical clearance was approved by the ethical committee
of Faculty of Medicine Universitas Indonesia.
Results: There are 600 patients between August 2009
and December 2013 and 436 patients included in this
study. Two hundred fifty six patients (58.7%) had
culture conversion at two months. Univariate analysis
showed underweight patients with BMI 6 18.5 kg/m2
[aHR 0.707; 95%CI 0.575-0.870] and anemia [aHR
0.716; 95%CI 0.587-0.847] had associated with longer
culture conversion. Even in multivariate analysis, only
underweight patients [aHR 0.792; 95%CI 0.637-0.983]
that consistently had long time conversion. We also
found obese patients with BMI 7 25 kg/m2 [aHR 1.616;
95%CI 0.882-2.959] had shorter conversion time
compared to patient with normal weight. Other factors
for longer culture conversion were female [aHR 0,808;
95%CI 0,659-0,991], cavity on chest radiograph [aHR
0.781; 95%CI 0.634-0.961], high bacterial load [aHR
0.701; 95%CI 0.501-0.979], previous second line TB
drugs [aHR 0.597; 95%CI 0.415-0.858], resistance with
second line TB drugs [aHR 0.486; 95%CI 0.305-0.776],
and lower lymphocites count [aHR 0.681; 95%CI 0.5240.885].
Conclusion: Nutrition status will affect culture conversion among MDR-TB patients.
Background: Malnutrition is a major cause of morbidity

and mortality in many settings and nutrition therapy
aims at reducing the risk of death (WHO). In Kenya,
national TB program recommends nutrition support to
TB patients who are severely or mildly malnourished
(BMI ,18.5). A number of organizations have been
providing food supplements and food items to TB
patients who meet the set criteria in the course of their
treatment. While modality and frequency of this support
varies, there is little evidence whether it improves
treatment outcomes for malnourished TB patients. The
aim of the study was to assess the impact of nutritional
support on mortality among TB patients in Kenya.
Design/Methods: We compared the risk of death among
eligible TB patients receiving food support with that of
eligible patients receiving no food support. We controlled
for sex, HIV status, ART uptake, TB type, patient type,
facility type and severity of malnutrition (severe malnutrition, BMI ,16 and mild malnutrition, BMI between
16 and 18.5). Data were abstracted from the national
electronic surveillance system, TIBU, for all patients
started on TB treatment between January 2012 and
December 2013 and who had treatment outcome
assigned and their BMI was between 12.0 and 18.5.
Only patients 15 years and above were included in the
analysis.
Results: Of 66 093 patients with BMI between 12.0 and
18.5, 23 816 (36.0%) received food support. Overall
mortality among these patients was 7.0% with 8.7% of
patients who received food support and 6.1% of those
who did not receive food support dying in the course of
TB treatment. Unadjusted odds of dying for patients who
received food were 45% (95%CI 36.5% to 54.0%) more
than those of patients who did not receive food support.
Adjusting for facility type, sex, TB type, patient type,
HIV status, ART status and severity of malnutrition, the
odds of dying for patients on food support were 23%
(95%CI 15.6% to 30.9%) more than those of patients
not on food support. Severity of malnutrition was an
important factor with patients severely malnourished
being 73.1% (95%CI 62.7 to 84.2) more likely to die
than patients mildly malnourished. Further analysis
showed that 36.0% of malnourished (mild and severe)
and 8.5% of patients with normal to obese BMI received
food support and these proportions were statistically
significantly different, P ,0.0001. Mortality among
patients with normal to obese BMI stood at 4.6% and
was statistically significantly lower than that of malnourished patients, P ,0.0001.
Conclusion: While there is evidence that food support is
being provided to deserving patients based on BMI, the
support does not translate into reduction of mortality
among the targeted patients. The national TB program
needs to assess the practices associated with food support

to TB patients in terms of frequency, timing, and quantity
and design other complementary interventions targeted
at the malnourished patients in order to reduce mortality
among this group.
PC-1163-06 Does nutrition support during TB

treatment improve TB treatment outcomes in
Kenya?
1
M Mangut, B Ulo, F Ngari, M Okoth, M Mungai,

R Wambu,2 E Masini,2 T Kiptai1 1AMREF Kenya, Nairobi,
2
Ministry of Health, Nairobi, Kenya. e-mail:
temkomangut@gmail.com
Background and challenges to implementation: Undernutrition is a risk factor for progression from TB
infection to active TB disease and is a predictor of
increased risk of death and TB relapse. In turn TB can
lead to undernutrition. Adequate nutrition interventions
during TB treatment could improve treatment outcomes
in malnourished TB patients. In Kenya, despite TB and
under nutrition affecting more men than women, no
specific facility based nutrition programs focus on men.
Currently, Global fund TB through AMREF Kenya
support 69% of severely malnourished TB patients. We
sought to understand the relationship between food
support and treatment outcomes.
Intervention or response: January to December 2013
data for TB patients notified in TIBU, a case based
electronic system, was extracted. Analysis excluded
patients with invalid Body Mass Index (BMI). Descriptive statistics were used to quantify proportions of
patients by sex, patient type, BMI, TB-HIV co-infection,
Anti- Retroviral Therapy (ART) and duration to TB
treatment outcome against food support received or not.
Logistic regression was done to determine significance of
relationship between food support and death and food
support and out of control.
Results and lessons learnt: Out of 89 760 patients
notified from 3052 TB sites, 69.8% (62 642) had valid
BMI recorded at admission. A total of 9458 (4636
females and 4822 males) patients received food support.
Of the 27 870 (45.5%) with BMI ,18.5, 10 391(37.3%)
had BMI 616.0. Among those who received food,
17.8% (1686) died and 4.5% (424) were out of control.
Among those who did not receive food, 5.4% (3665)
died and 5.5% (3744) were out of control. Treatment
completion was at 50% in each of the two categories.
Higher death rates among those who received food was
observed among TB-HIV co-infected at 70.6% (1217)
compared to 55.2% (2060) among those who did not. Of
the 29 090 with outcome date between 0 and 4 months of
TB treatment 44.9% (4505) died among those who
received food and 39.3% (3584) were out of control
among those who did not receive food. Those who
received food were 1.5 times (P , 0.0001) more likely to
die than those who did not, while those who did not
receive food were 1.2 times (P , 0.0001) more likely to
get out of control.
S479
Conclusions and key recommendations: Death rate was

higher among those who received food and out of control
was higher among those who did not. Late initiation of
nutrition interventions leads to poor treatment outcomes
of out of control and death. Nutrition assessment and
documentation are key for TIBU data to be useful for
decision making.
PC-1164-06 Assessing the impact of nutritional

support on multidrug-resistant tuberculosis
patients in Kenya
R Kiplimo,1 R Wambu,1 H K Kipruto,2 E Masini1 1National
Tuberculosis, Leprosy and Lung Disease Program, Nairobi,
2
World Health Organization, Kenya, Nairobi, Kenya. e-mail:
kiplimorichard@gmail.com
Background: Optimal nutrition is essential for achieving

and preserving health for everyone. Food and nutrition
thus play important roles in care and mitigation of
infections and their effects. The Kenyan DR-TB management model provides both patient and nutritional
supports to its MDR patients. However, treatment
outcomes of Multi-Drug Tuberculosis (MDR-TB) still
remain generally poorer compared to outcomes for drug
sensitive disease patients. Among the risk factors
associated with the poor outcome are malnutrition and
HIV. This paper sought to establish the overall impact of
the nutritional support provided.
Design/Methods: This was a retrospective study of
MDR-TB cases notified to the National program from
2011 to 2013 had been evaluated. Descriptives of the
data were obtained and later univariate analyses (v2) and
multiple logistic regression were performed to determine
the association between the food support and the
outcome and effect of risk factors.
Results: A total of 371 MDR cases were included in the
analysis. A cure rate of 62%, treatment success rate of
81% and death rate of 12% were obtained. While about
60% of the patients are either moderately malnourished
(31%) or severely malnourished(27%), only 44% of the
patients received food support.Among the HIV positives,
30% received food support.Food support is strongly
associated with outcome of MDR patients (v2 29.793,
P 0.000).In reference to out of control outcome,Patients who received food support are 3 times more likely
to have an outcome of cured while HIV positive patients
are 2 times more likely to have an outcome of cured.
Conclusion: Food support and HIV status are indicators
that influence treatment of DR-TB patients both
positively and negatively. The support provides a
platform for continued communication which encourage
patients to move on.
S480
PC-1165-06 Food parcels and adherence to TB

treatment
A Niyazov,1 A Trusov,1 E Kukhranova,1 A Duishekeeva1
Project HOPE, Bishkek, Kyrgyz Republic. e-mail:
aniyazov@projecthope.org

incidence rate of tuberculosis in Kyrgyzstan in 2014
was 109.5 cases per 100 000. Factors contributing to
retaining high rate of TB incidence in Kyrgyzstan are a
high level of poverty, active external and internal
migration of migrants with TB, and restricted access to
health services for high-risk groups (ex-prisoners, drug
and alcohol abusers, homeless people) as well as low
level of government investments in health care. The
majority of TB patients are from high-risk groups, poor
and unemployed. In order to seek a better life, TB
patients move to other countries and often quit TB
treatment. Default rate has varied from 4% to 7% over
the last few years.
Intervention or response: To ensure better adherence to
TB treatment for patients in the continuation phase of
treatment, it was proposed to motivate the patients by
distributing free Food Parcels at places they receive their
medication. The Project Hope team jointly with NTP
medical staff developed the system of distribution of food
parcels. TB patients sign an agreement with the medical
facility to participate in the program with a commitment
to continue TB treatment until their treatment is
completed. Food parcels are distributed once a month.
The content of food parcels include: rice-1kg, pasta- 1
kg, granulated sugar-1 kg, dry milk 1 kg, vegetable oil
1 liter, black leaf tea- 250 gr. Overall in 2014 the absolute
number of patients participating in the food distribution
program was 4916 from the 6095 patients on treatment.
Selection criteria was based on the phase of treatment
(continuation phase), treatment with first line drugs,
New Cases, Relapses, Previously treated (Re-treatment
cases) and patients with extra pulmonary TB. In order to
get food parcels TB patients should not miss more than 5
doses of treatment in a month. Otherwise they will not
receive food parcels.
Results and lessons learnt: The default rate between
patients participating in the program (1.3%) was
significantly lower than among those who didnt
participate in program (26,5%) and the total default
rate of 6,2%. The majority 57% of defaulters are people
who migrate to another country (Russia), 23% alcoholics, 6%- drug addicted, 2% - homeless, 2% due to religious beliefs, 3% - due to side effects, 2% - due
to stigma.
Conclusions and key recommendations: Distribution of
food parcels for TB patients contributed to good
adherence to TB treatment.
55. Molecular epidemiology and

modelling: how can we break the cycle of
transmission?
PC-1166-06 Molecular epidemiology of
tuberculosis in Havana, Cuba, 2009
A Gonzalez Daz,1 T Battaglioli,2 R Daz,3
3 E Gonzalez Ochoa,1 P Van Der Stuyft2,4
R Goza Valdes,
1
Institute of Tropical Medicine Pedro Kour, Havana, Cuba;
2
Institute of Tropical Medicine, Antwerp, Belgium; 3Institute
of Tropical Medicine Pedro Kour, Havana, Cuba; 4Ghent
University, Ghent, Belgium. e-mail: tbattaglioli@itg.be
Background: The objective of the study was to estimate

the proportion of tuberculosis cases attributable to recent
transmission and the risk factors possibly associated with
tuberculosis clustering.
Design/Methods: Population-based study combining
information from epidemiological investigation of tuberculosis cases notified to the national tuberculosis control
program in Havana, Cuba, in 2009 with the results of
genotyping of Mycobacterium tuberculosis isolates with
variable-number tandem-repeat of mycobacterial interspersed repetitive units (MIRU-VNTR) typing.
Results: Out of 186 cases, 61 were genotyped: 33
patterns and 5 clusters with 19, 7, 3, 2 and 2 cases were
found. The index of recent transmission was 45% (95%
confidence interval 33%-58%). Routine contact investigation failed to identify a substantial number of
epidemiological links. A history of living in a closed
setting was strongly associated with clustering.
Conclusion: The index of recent transmission in Havana
in 2009 is high. The existing control measures in closed
settings should be strengthened. A study on a larger
number of cases and for a longer time period should be
carried out to get more precise estimates. Further studies
on the utility and cost-effectiveness of the addition of
molecular epidemiology techniques to support the
progress towards tuberculosis elimination in Cuba, a
low incidence resources-limited setting, are also needed.
PC-1167-06 A bacterial perspective on

tuberculosis in West Africa
F Gehre,1,2 S Kumar,1 L Kendall,2 M E Kitata,3 E Abatih,1
M Antonio,2 D Berkvens,1 B De Jong1 1Institute of Tropical
Medicine, Antwerp, Belgium; 2Medical Research Council
Unit, Fajara, Gambia; 3University of Gondar, Gondar,
Ethiopia. e-mail: fgehre@itg.be
Background: West Africa, with a high burden of

undetected TB, is known to harbour the greatest strain
diversity worldwide, although the mycobacterial population structure has yet to be elucidated on a regional
scale. Understanding TB epidemics from a bacterial
point-of-view, however, is essential for a Systems
Epidemiology approach to TB control, combining
classic epidemiology with host studies and knowledge
on genetic diversity of bacterial populations. Method We
collated novel and published genotyping data and
classified these into mycobacterial lineages and families,
followed by longitudinal and phylogeographic analyses

using statistical tools and GenGIS.
Results Among 3580 isolates from 12 countries, we
identified 6 lineages (L) comprising 32 families, with
91% of infections caused by Euro-American (L4) and M.
africanum families (L5 and L6). We demonstrated
statistically significant phylogeographic separation of
families and that TB due to M. africanum West Africa 2
remains a persistent health concern.
Conclusion The unequal distribution of highly transmissible or resistance-prone mycobacterial families is
important, as various health systems face different public
health challenges. From a study design perspective,
vaccine trials and bacterial geno/phenotype association
studies require high strain diversity and such studies
would preferably be conducted in Western West Africa,
while low strain diversity is advantageous for host
genetics and studies may be better suited for Eastern
West Africa. Our study contributes pathogen information to systems epidemiology-based TB control, and
serves as important resource for the design of future large
scale clinical trials in West Africa.
PC-1168-06 A longitudinal study of the

continued transmission of Mycobacterium
tuberculosis in eastern rural China
B Xu,1 L Wu,1 W Jiang,1 Q Zhao,1 Yi Hu1 1School of Public
Health, Fudan University, Shanghai, China. e-mail:
bxu@shmu.edu.cn
Background: The present study aimed to investigate the

distribution of Mycobacterium tuberculosis which was
persistently transmitted in local population at a period of
8 years, and to explore socio-demographic and clinical
factors affecting the competition and survival of M.
tuberculosis strains.
Design/Methods: The study was carried out in two rural
counties of eastern China during 2005-2012. All
registered TB patients received bacteriologic-based TB
diagnosis. Spoligotyping was used to define the lineage of
Mtb isolates, and 24 loci MIRU-VNTR was applied for
M. tuberculosis genotyping. Based on Middelkoop and
her colleagues study, persistently successful strains were
defined as strains that had been present for at least 2
consecutive years and were still present in 2012
(persistence), and had 72 cases/year; transiently successful strains were strains that did not necessarily persist
to 2012 but occurred for at least 2 consecutive years with
a cluster density of 7 2 cases/year in the consecutive
period; unsuccessful strains were clustered strains did not
fall into these two groups, and singleton strains.
Univariate and multivariate binary logistical regression
analysis were applied to identify factors associated with
strain success.
Results: Of the 982 M. tuberculosis isolates, 793
(80.8%) belonged to the Beijing family. It was found
that 666 isolates (67.8%) showed a unique MIRUVNTR genotype, while 316 isolates (32.2%) shared
genotypes in 104 clusters. The size of cluster ranged from
2 to 15 isolates, and the median of cluster density was 1.4
S481
cases/year. Of the 316 clustered isolates, 36 (11.4%)

were defined as persistently successful strains, 74
(23.4%) were transiently successful and 198 (62.7%)
were unsuccessful strains. Persistently successful strains
were less likely to be resistant to first-line anti-TB drugs
compared with transiently successful strains (27.8%.vs.
50.0%; OR0.179, 95%CI: 0.048-0.565, P 0.002) and
unsuccessful strains (27.8%.vs.82.3%; OR0.068,
95%CI: 0.019-0.122, P 0.001). However, the Beijing
family strains were more likely being persistently
successful strains (83.3%.vs. 60.8%; OR2.328,
95%CI: 1.040-5.040, P , 0.001).
Conclusion: The present study suggests that Beijing
family strains play an important role in the persistent
transmission. Although drug resistant isolates have yet
been dominant in the persistently successful strains, its
adaption with time should be concerned and monitored
longitudinally.
PC-1169-06 Spoligotype profile of

Mycobacterium tuberculosis complex strains
isolated from pulmonary tuberculosis patients
in Khartoum State, Sudan
A Elegail,1 N Y Ibrahim Mohamed,1
E O Mohamed Nour,1 S Hoffner,2 M Haile2 1National
Public Health Laboratory, Khartoum, Sudan; 2Public Health
Agency of Sweden, Solna, Sweden. Fax: (46) 8 301 797.
e-mail: melles.haile@folkhalsomyndigheten.se
Background: Aim of this study was to characterize

specific genotypic lineages of Mycobacterium tuberculosis (MTB) strains circulating among patients with
pulmonary tuberculosis (PTB) in Khartoum, Sudan.Tuberculosis is an infectious disease caused by M. tuberculosis. In this study DNA from a total of 120 M.
tuberculosis strains isolated 2007 - 2009 were sent from
NRL in Khartoum, Sudan for EQA to WHO SRL
Stockholm. Confirmation DST results and mutation
analysis for EQA were performed for all 60 MDR and
60 non-MDR strains.
Design/Methods: Data was collected consecutively from
sputum smear positive patients from selected NTP
diagnostic centers in Khartoum during the period 20072009. Species identification and drug resistance was done
using biochemical tests/DNA methods and conventional
culture technique respectively at NRL. Confirmation
DST was done using LPA and molecular epidemiology by
Spoligotyping to determine genetic diversity and susceptibility profiles of isolates circulating in Sudan.
Results: A total of 120 DNA from M. tuberculosis strains
were studied using Spoligotyping. This revealed a total of
112 patterns from which 4 patterns represented clustered
isolates with a total of 105 (88%) of the strains. In this
study, the CAS1 Delhi/family was the predominant
shared type detected in 61 isolates (51%), where 25 of
the strains were MDR and 36 strains were susceptible. It
was followed by H3/family with 19 (16%) strains, and
11 strains with LAM3/family, 14 strains with T2/T1/
family type and 2 strains each with Beijing type and S
S482
family. One isolate of these where not yet defined in the

SpolDB and therefore regarded as orphan.
Conclusion: The observation obtained using spoligotyping from the Sudanese patients reflects in comparison to
the global distribution highly diverse clades circulating in
the country. The CAS1 Delhi/family was the predominant shared type detected in 61 isolates followed by H3/
family with 19 isolates, and 11isolates with LAM3/
family, 14 isolates with T2/T1/family type and 2 isolates
each with Beijing type and S family. We found 35 of the
isolates were not available in the SpolDB database but
was found as family in our database. These groups
visually reflected some patterns similarities therefore
included in the family group closely related. In general
the study could not see any significant difference between
MDR and non-MDR isolates within the clade types.
Even though study material was limited it indicates that
there exist active transmission of certain spoligotypes.
PC-1170-06 Role of stochasticity in reaching the

2025 global targets
G Huynh,1 M Roh1 1Institute for Disease Modeling, Bellevue,
WA, USA. e-mail: ghuynh@intven.com
Background: Recently, the World Health Organization

developed the End TB Strategy, with a milestone of 75%
reduction in TB deaths and a 50% reduction in TB
incidence from 2015-2025. Stochastic fluctuations in the
intrinsic TB disease dynamics will affect the potential
impact of new interventions and the probability of
reaching the goal by 2025.
Design/Methods: We developed a stochastic model of TB
transmission in China to estimate the probability that
China will achieve the 2025 milestones if new interventions are introduced, including improving treatment
quality and reducing the time to treatment. We consider
how stochasticity affects the probability of maintaining
the TB burden below the goal level beyond 2025, both at
the individual community level and the aggregated
country level. We also explore how the stochastic effects
vary within a plausible range of community sizes.
Results: Stochastic variation increases the time needed
for the incidence to stabilize after the intervention is
implemented; thus, compared to deterministic estimates,
more aggressive TB control is necessary to achieve the
incidence goal with 50% probability. In contrast,
because the mortality goal is small relative to the
population size, stochastic fluctuations increase the
probability of achieving the goal even with less aggressive
TB control. Stochasticity decreases the probability any
single community maintains the TB burden below the
goal levels after 2025, but increases the probability the
entire country stays below the goal.
Conclusion: Stochasticity at the community level contributes significant variability to the observed impact of
new interventions. This can result in dramatic deviations
from the deterministic mean when considering many
communities in aggregate. In China, our model suggests
that stochastic effects will increase the probability of
achieving the mortality goal but decrease the probability
of achieving the incidence goal. These stochastic effects

should be taken into account when estimating how new
interventions might reduce overall TB burden, both in
pilot studies and in the field.
PC-1171-06 A novel regression-based tool to

improve estimates of the proportion of
tuberculosis incidence due to recent
transmission in settings of sparse data
P Kasaie,1 B Mathema,2 W D Kelton,3 A Azman1
J Pennington,1 D Dowdy1 1Johns Hopkins University,
Baltimore, MD, 2Columbia University, New York, NY,
3
University of Cincinnati, Cincinnati, OH, USA. e-mail:
pkasaie@jhu.edu
Background: In developing public health responses to

tuberculosis (TB) epidemics, it is often important to
estimate the proportion of active TB cases that result
from recent infection versus reactivation (recent transmission proportion). This is traditionally done with
molecular epidemiological data (e.g., DNA fingerprinting), but the traditional (n-1) method for converting
these data into estimates of recent transmission carries
substantial bias.
Design/Methods: We developed a stochastic, individualbased simulation model of a TB epidemic that links the
true (simulated) recent transmission proportion to the
molecular clustering pattern that would be observed. We
then carried out a set of controlled simulation-experiments across a variety of epidemiological settings
(derivation set) to develop an improved, user-friendly
tool for estimating the recent transmission proportion as
a function five inputs: underlying TB incidence, sampling
coverage, study duration, clustered proportion of observed cases, and proportion of observed clusters in the
sample. We compared the performance of this tool
(including a simple and a more comprehensive regression
model) against the traditional n-1 0 model, and validated
the findings using an independent set of simulation
experiments (validation set).
Results: In the validation set, the regression tools
reduced the absolute estimation bias (difference between
estimated and true recent transmission proportion) in
the n-1 0 technique by a median [interquartile range] of
60% [9%, 82%] and 69% [30%, 87%]. The bias in the
n-1 0 model was highly sensitive to underlying levels of
study coverage and duration and substantially underestimated the recent transmission proportion in settings of
incomplete data coverage (Figure). By contrast, the
regression models performance was more consistent
across different epidemiological settings and study
characteristics. We provide one of these regression
models as a user-friendly, web-based tool (http://
modeltb.org/recenttrans/).
Conclusion: Novel tools can improve our ability to
estimate the recent TB transmission proportion from
data that are observable (or estimable) by public health
practitioners and guide effective policies and interventions in settings where data are often incomplete and
long studies are impractical.
S483
with random-effects modeling and used stratification to

investigate between-study heterogeneity. Estimating the
population impact of household exposure on tuberculosis was completed by calculating prevalence ratios from
randomly sampled studies and finding a population
attributable fraction of household TB exposure from
these studies. We used the established PAF formula, PAF
P cases[(RR-1) / RR], to calculate the population
impact of household exposure.
Results: We identified 28 studies including 16 732
children exposed to TB in the household and 174 158
unexposed. Overall, children exposed to a contact with
active TB in their households were 3.72 (95%CI 2.94
4.71) times more likely to acquire tuberculosis infection
than their community counterparts. From ten studies
that randomly sampled from the general population, we
estimated that the PAF of transmission due to household
exposure was between 11% and 15%.
Conclusion: A child exposed to tuberculosis in the
household is at almost four times the risk of tuberculosis
transmission compared to a child not exposed. Despite
this, most TB transmission at the population level likely
occurs outside the household.
PC-1172-06 Individual and population-level

effects of household exposure on tuberculosis
transmission: a systematic review and metaanalysis
L Martinez,1 Y Shen,1 E Mupere,2 A Ezeamama,1
S Isaacson,3 A Kizza,1 P Hill,4 C C Whalen1 1University of
Georgia School of Public Health, Athens, GA, USA;
2
Makerere University, Kampala, Uganda; 3Tulane University
School of Public Health and Tropical Medicine, New Orleans,
Louisiana, USA; 4University of Otago School of Medicine,
Centre for International Health and the Otago International
Health Research Network, Dunedin, New Zealand. e-mail:
leomarti@uga.edu
Background: Targeting locations where tuberculosis

(TB) is likely to spread may open up entry points for
interrupting transmission. Whether TB transmission is
more likely to occur inside or outside the household at a
population level is unclear but could assist in allocating
resources in future active case finding efforts.
Design/Methods: We conducted a systematic review and
meta-analysis on studies assessing latent TB infection in
two groups: children exposed to a household member
with TB and children without such exposure. Searches
were conducted in multiple electronic databases. Children were used to represent recent infection and were
defined as below 15 years of age. Household contacts
and community controls had to be matched by age and
location at a minimum. To investigate the individual
impact of household exposure, we pooled odds ratios
PC-1173-06 Mycobacterium tuberculosis

transmission is associated with attendance at
community gathering places in rural Malawi
P Khan,1,2 T Mzembe,1 K Kranzer,2 D Mulawa,1
O Koole,1,2 K Fielding,2 J Glynn,2 A Crampin1,2 1Karonga
Prevention Study/MEIRU, Karonga, Malawi; 2London School
of Hygiene & Tropical Medicine, London, UK. e-mail:
palwasha.khan@lshtm.ac.uk
Background: The majority of Mycobacterium tuberculosis transmission takes places outside the household in
high burden settings. Community gathering places,
including healthcare facilities, churches, funerals, markets and public transport in which crowding occurs, may
be hotspots of M. tuberculosis transmission but their
contribution to transmission is unclear. Identifying
incident M. tuberculosis infection in children , 5 years
can be used as a sentinel of infectious adult tuberculosis
in the community.
S484
Design/Methods: Study hypothesis: Children who attend

gathering places are at increased risk of M. tuberculosis
infection compared to children who do not Setting:
Demographic surveillance site (DSS), northern Malawi
Design and procedures: Eligible children aged , 5 years
resident in the DSS were enrolled in a tuberculin skin test
(TST) cohort. Children with an induration of , 10mm in
the baseline TST survey were retested one year later to
identify incident MTB infection. Guardians of children
were interviewed to assess the childs attendance at
gathering places in the preceding year. Incident infection
was defined as TST conversion with an increase of at
least 10mm from , 10mm to 710mm between rounds.
Results: 5537 children were eligible, of whom 4967
(90%) had a baseline TST result. To date, 2267 children
have had repeat TST, contributing 3334 person-years.
109 children converted from ,10mm to 710mm with
an increment of 710mm between rounds, giving an
estimate of MTB infection incidence of 3.3 per 100
person-years. Data on attendance at gathering places is
available to date for 1242 children (2238 person-years),
of whom 73 TST-converted. After adjusting for age, sex,
household socioeconomic status and population density,
the odds of incident MTB infection were increased in
children who had attended church more than 4 times in
the last year compared to children attending less often
(adjusted OR [aOR] 3.6, 95%CI 1.2-11.1); in those who
visited a healthcare facility at least once in the year (aOR
3.1, 1.0-9.8) and in those travelling by minibus (aOR 1.8,
0.9-3.6); but not with visiting market places (aOR 0.9) or
attending funerals (aOR 0.7).
Conclusion: Community gathering places that involve
contact in confined spaces - church, healthcare facilities,
and minibuses but not outdoor gatherings were
associated with incident MTB infection in children
under-5 years in rural Malawi. These findings support
the call for place-finding, in addition to targeted casefinding in high burden settings.
PC-1327-06 MDR-TB strains have an increased

propensity to acquire PZA resistance in the
context of DOTS-based chemotherapy
F Lanzas,1,2 P Karakousis,3 E Lopez,4 J Jurado,5
J Sacchettini,6 T Ioerger76 1Instituto Conmemorativo
Gorgas de Estudios de la Salud, Panama, 2Universidad de
Panama, Panama, Panama; 3Johns Hopkins University School
of Medicine, Baltimore, MD, USA; 4Hospital Santo Tomas,
Panama, 5Hospital Manuel Amador Guerrero, Colon,

Panama; 6Texas A&M University, College Station, TX, USA.
e-mail: flanzas11@gmail.com
Background: Drug resistance to pyrazinamide (PZA) in

Mycobacterium tuberculosis occurs by chromosomal
mutations, the most of which are SNPs or indels in the
pyrazinamidase gene (pncA). The mutation frequency in
vitro for H37Rv when selected on PZA is very similar to
that in other bacteria, typically around 1X10-5. But,
does this represent the mutation rate in vivo? Previous
studies have shown that mutations in pncA occur with
high frequency (i.e. in a large proportion) in MDR-TB
clinical isolates. This study aims to investigate the
acquistion of pncA mutations at a microscale among a
set of INHRIF resistant strains.
Design/Methods: The genome of 97 clinical isolates (66
MDR-TB and 31 sensitive strains) were sequenced using
Illumina Genome Analyzer IIx. A phylogenetic tree was
constructed base on genome-wide SNP analysis. To
better understand the evolution of PZA resistance in this
population, we decided to develop a crude molecular
clock by quantifying the number of SNPs observed
between contact pairs, which are defined as strains the
came from patients with a known familal relationship
and are adjacent in the phylogenetic tree. We make the
assumption of infection by a common strain in each
contact pair, transmitted from one to the other. Thus any
polymorphism observed between them can be assumed
to have arisen with one host or the other. Analysis of
multiple contact pairs shows an average of 10 SNPs
between them, which were presumably acquired in vivo
over at most only a few years of infection. This
background mutation rate is consistent with what has
been found in other studies (see Timothy Walker et al).
Results: In PncA we observed both a large number and
high diversity of mutations. Among 66 isolates chosen a
priori for resistance to INH and RIF, 36 had mutations in
PncA, and these were explained by 16 distinct mutations.
In contrast, none of the drug sensitive strains displayed
any mutations in pncA. While phylogenetic analysis
suggested that the mutations associated with INH and
RIF resistance where shared among whole clusters, the
PncA mutations were only shared among small subsets of
strains, suggesting they arose recently and independently.
Conclusion: MDR strains that are INH and RIF resistant
are highly susceptible to acquire PZA resistance, and the
probable cause for this effect is not due to increase
mutation rates or reduced fitness costs in vivo, but simply
to the reduction in effective drugs leading to a nearly
monotherapy setting.
56. Tbilisi to Lima: a journey through TB

treatment outcomes
PC-1174-06 Assessment of treatment outcome
and associated factors among retreatment
tuberculosis patients in selected government
health facilities, Eastern Ethiopia
Y Moges,1 K Kibret2 1Haramaya university, Harar, 2Welega
university, Nekemite, Ethiopia. e-mail:
mogesyoni@gmail.com
Background: Tuberculosis is the leading causes of

mortality among infectious diseases worldwide. The
management of previously treated tuberculosis patients
(retreatment cases) is a challenge for National Tuberculosis Programmes in resource-limited settings such as
Ethiopia. So the aim of this study was to determine the
treatment outcomes of the different categories of retreatment tuberculosis patients in Eastern Ethiopia.
Design/Methods: A retrospective cohort study was
conducted from 2002-2006 in selected health facilities
in Dire Dawa town. The study included all TB cases
registered as relapse, failure or default between
September 2002 - March 2006 in Dilchora Referal
Hospital and from September 2004 - March 2006 in
Legehare Health Center. Data were sourced from TB
health unit registers from the two health institutions
using a structured data form. Statistical analysis was
carried out using various statistical tests. Descriptive
statistics were conducted using frequencies and proportions. X2 test was used for testing association of different
characteristics.
Results: A total of 1667 TB patients were treated from
September 2002 - March 2006 in Dilchora Referal
Hospital and in Legehare Health Center. Of these; 103 of
the patients were recorded as TB retreatment cases. The
retreatment rate was 6.24. Among the 103 retreatment
cases, 52 (50.5%) had a treatment outcome registered in
the health unit register; 22 (21.4%) cases were transferred out and 29 (28.2%) had no record of outcome in
the register. Of the 52 category II TB treated patients 29
(55.7%) were cured, 14 (26.9%) died, 5 (9.6%) retreatment failure cases and 4(7.7%) developed multidrug
resistance tuberculosis (MDR-TB). Cure rate was
(40.0%) and (27.4%) in the patients belonging to the
default and relapse category respectively.
Conclusion: The analysis has shown that relapse cases
were the most common retreatment category and
treatment success rates for relapse cases were low
(55%). Persistent monitoring of treatment outcomes of
relapse, failure and default cases is crucial for improving
TB control and to alter the incidence of MDR and XDR
tuberculosis.
S485
PC-1175-06 Treatment outcomes for multidrugresistant tuberculosis patients under DOTSPlus: systematic review and meta-analysis
Y Moges1 1Haramaya university, Harar, Ethiopia. e-mail:
mogesyoni@gmail.com
Background: Anti-tuberculosis (TB) drug resistance is a

major public health problem that threatens progress
made in TB care and control worldwide. The WHO aims
to achieve a 90% treatment success rate for TB patients
by 2015 and current studies on MDR-TB treatment
revealed a huge gap to this target. The aim of this study
was to assess and summarize the available evidence on
MDR-TB treatment outcome under DOTS-Plus.
Design/Methods: Literature based systemic review of
observational studies was conducted. Original studies
providing multi drug resistance tuberculosis under DOTS
plus program according to WHO-defined outcomes at
the end of treatment and follow-up were identified using
data bases such as Medline/PubMed, Google Scholar and
HINARI. The descriptions of original studies were made
using frequency and forest plot. Heterogeneity across
studies was checked using Cochrane Q test statistic and
I2. Pool risk estimates of Multi-drug resistance tuberculosis treatment outcome and sub-grouping analysis were
computed using a Bayesian random effects meta-analysis.
Results: Among articles identified, 13 met our inclusion
criteria from 8 countries. In a pooled analysis, 55.62%
[95%CI 54.3456.89] of patients were cured and
43.48% [58.44 68.51] had successful outcomes, while]
14.21% [11.6016.81] defaulted, 12.6% [8.9816.21]
died, and 7.65% [5.639.68] were transferred out.
Individualised treatment regimens had higher treatment
success than standardized regimens. HIV infection,
alcohol use and previous TB treatment were significantly
associated with poor multi-drug treatment outcome.
Conclusion: In general, the overall performance of
DOTS-Plus was acceptable but still it is far from WHO
treatment success target. In addition, we have identified
high rates of default, which likely contributes to the
development and spread of MDR-TB.
S486
PC-1176-06 Adverse reactions associated with

injectable second-line anti-tuberculosis drugs
among patients with M/XDR-TB in Tbilisi,
Georgia
1,2
M Buziashvili, V Mirtskhulava, M Kipiani,

N Tukvadze,2 J Furin,3 M Magee,4 R Kempker,5
H Blumberg5 1Tbilisi State Medical University, Tbilisi,
2
Georgia; 3Case Western Reserve University, Cleveland, OH,
4
Georgia State University, Atlanta, GA, 5Emory University,
Atlanta, GA, USA. e-mail: mari.buziashvili@yahoo.com
Background: The country of Georgia has high rates of

MDR- and XDR-TB (11% of new and 38% of retreatment cases) which require the use of second-line anti-TB
drugs; these regimens are frequently accompanied with
severe adverse drug reactions (ADRs). The aim of our
study was to determine the rates of and risk factors for
ADRs associated with injectable drug use among M/
XDR-TB patients.
Design/Methods: Consecutive patients, who initiated
treatment for M/XDR-TB at the National Center for
Tuberculosis and Lung Diseases (NCTLD), Tbilisi,
Georgia in 2011, were enrolled in a retrospective cohort
study. All patients received kanamycin or capreomycin.
Data was abstracted from medical and laboratory
records and the national surveillance database using a
standardized form. Ototoxicity was defined as selfreported tinnitus and/or impaired hearing by a patient.
Creatinine clearance (CrCL) was calculated using the
Cockcroft-Gault formula and nephrotoxicity was defined according to the RIFLE (Risk, Injury, Failure, Loss
of kidney function, and End-stage kidney disease)
classification of acute kidney injury (AKI). Bivariate
analysis was performed to assess the association between
patients clinical characteristics and ADRs.
Results: Among 67 patients treated for M/XDR-TB the
median age was 35 years (range 17-73) and 68% were
male. Rates of HIV co-infection were 5%, HCV 15%,
and 9% had diabetes. Median baseline CrCL was 88.9
ml/mi (IQR 68-103). Any AKI (CrCL decrease .25%
compared to baseline) was developed in 36 (54%)
patients (30 received kanamycin and 6 received capreomycin). Renal injury (CrCL decrease .50%) occurred in
8 (12%) patients. Most AKI (76%) occurred in the first 6
months of treatment. Ototoxicity was reported by 36%
including 18 (27%) patients with tinnitus and 22 (33%)
with impaired hearing. Compared to those without AKI,
patients with AKI were significantly more likely to have
an incarceration history (23% vs 3%, P ,0.05). Patients
with AKI had a higher rate of developing ototoxicity
(44% vs. 26%, P 0.11) and higher prevalence of diabetes
at baseline (14% vs. 3%, P 0.13) but the differences did
not reach statistical significance.
Conclusion: AKI and ototoxicity were very common
among patients with M/XDR-TB who received an
injectable drug. A history of prior incarceration was
found to be associated with increased risk for development of AKI. Further defining risk factors for AKI and
ototoxicty will help guide ADR prevention efforts.
PC-1177-06 Poor tuberculosis treatment

outcomes in the district of Manhica,
Mozambique, 2011-2012
A L Garcia-Basteiro,1,2 D Respeito,1 O Joaquim Augusto,1
E Lopez Varela,1,2 C Sacoor,1 I Manhica,3 F Cobelens,4
P Alonso1,2 1Centro de Investigacao
em Saude de Manhica
(CISM), Manhica,
Mozambique; 2Barcelona Institute for
Global Health (ISGlobal), Barcelona, Spain; 3National
Tuberculosis Control Program, Ministry of Health, Maputo,
Mozambique; 4Amsterdam Institute for Global Health and
Development, Academic Medical Centre, Amsterdam,
Netherlands. e-mail: alberto.garcia-basteiro@manhica.net
Background: In Mozambique, there is limited data

regarding the monitoring of tuberculosis (TB) treatment
results and determinants of adverse outcomes under
routine surveillance conditions. The main objective of
this study was to evaluate the treatment outcomes among
TB patients and analyze factors associated with a fatal
outcome. As a secondary objective, we estimated the
proportion of deaths attributable to TB in a very high
HIV- and TB-burden area in southern Mozambique.
Design/Methods: This is a retrospective observational
study based on TB patients diagnosed in the period 20112012 in the district of Manhica, Mozambique. We used
three different data sources: a) TB related variables
collected by the NTP in the district of Manhica for all TB
cases starting treatment in the period 2011-2012. b)
Population estimates for the district were obtained
through the Mozambican National Statistics Institute.
c) Estimates on population growth, number of deaths
and other relevant demographic variables were collected
from the Demographic Surveillance System (DSS) at
CISM. From 2012 onwards, the DSS records the cause of
death in all deaths belonging to the study area. WHO
guidelines were used to define TB cases and treatment
outcomes.
Results: Of the 1957 cases starting TB treatment in the
period 2011-2012, 294 patients (15.1%) died during
anti-tuberculous treatment. Ten per cent of patients
defaulted treatment. The proportion of patients considered to have treatment failure was 1.1%. HIV infection
(OR 2.73; 95%CI: 1.70-4.38), being female (OR: 1.39;
95%CI: 1.31-1.91) and lack of laboratory confirmation
(OR 1.51; 95%CI: 1.10-2.08) were associated with
dying during the course of TB treatment (P , 0.05).
Among HIV positive patients, those who were not on
ART were more likely to die during treatment (uOR:
1.78 95%CI 1.31-2.42). The contribution of TB to the
overall death burden of the district for natural reasons
was 6.5% (95%CI: 5.5-7.6) in the year 2012, higher for
males than for females (7.8%; 95%CI: 6.1-9.5 versus
5.4%; 95%CI: 4.1-6.8 respectively). The age group
within which TB was responsible for the highest
proportion of deaths was 30-34 among males and 2024 among females (20% of all deaths in both cases).
Conclusion: This study shows a very high proportion of
fatal outcomes among TB cases starting treatment. There
is a high contribution of TB to the overall causes of
mortality. These results call for action in order to
improve TB (and TB-HIV) management and thus

treatment outcomes of TB patients.
PC-1178-06 Investigating gender differences in

the profile and treatment outcomes of
tuberculosis in Ebonyi State, Nigeria:
implications for tuberculosis control
1
1 1
S Oshi, I Alobu, K Ukwaja, D Oshi Centre for

Development and Reproductive Health, Enugu, Enugu State,
2
National Tuberculosis and Leprosy Control Programme,
Ministry of Health, Abakaliki, Ebonyi State, 3Department of
Ebonyi State, Nigeria. e-mail: ukwajakingsley@yahoo.co.uk
Background: Globally, twice as many men as women are

being diagnosed with tuberculosis (TB) annually. Little is
known about gender differentials in TB in Africa. The
aim of this study was to describe gender differences in the
epidemiological characteristics and treatment outcomes
of TB in Nigeria.
Design/Methods: A retrospective cohort analysis of
routine data was carried out among adult TB patients
treated between 2011 and 2012 in two large health
facilities in Nigeria. Gender differences in their demographic characteristics and treatment outcomes were
analyzed accordingly.
Results: Of 1668 TB patients enrolled, the male to female
ratio was 1.4:1. The mean ages of males and females
were 40.2 6 14.7 and 36.1 6 14.6 years respectively (ttest 6.62, P , 0.001). Male gender was associated with a
higher failure to smear convert after 2 (21.8% vs. 17.5%;
P 0.06) and 5 months (4.3% vs. 1.5%; P 0.02) of
treatment for smear-positive TB patients. Also, men were
more likely than women to fail treatment (2.2% vs.
0.7%; P 0.01). No significant differences exist in the
treatment success rates between women and men (78.2%
vs. 74.5%; P 0.08). Adjusted analyses showed
significant association between being an urban-male
and a HIV-infected female with unsuccessful outcome
adjusted by socio-demographic and clinical factors.
Conclusion: We found that gender disparities exist in TB
profile and treatment outcome in Nigeria and genderspecific strategies are needed to optimize TB management. We recommend that: (1) urgent strategies to reduce
treatment default and mortality in men and women with
TB need to be implemented; (2) TB case finding in older
women should be improved; (3) specific measures to
reduce unsuccessful outcomes for patients in the highrisk groups should be implemented; (4) socioeconomic
interventions to improve treatment success for rural TB
patients, especially women, should be implemented; and
(5) further studies on the impact of biological factors on
TB treatment outcomes in men and women need to be
carried out and the influence of HIV co-infection, sex
hormones and poverty assessed.
S487
PC-1179-06 Agbami tuberculosis control and

management initiative in Nigeria
S Kashim,1 H Deji,1 O Sam,1 A Thomas,1 O Pitan,1
C Onwurah,1 O Sunday,1 E Egbochukwu1 1Star Deep - a
Chevron Company, Llagos, Nigeria. e-mail:
gidadomansu@yahoo.com.au
Background: For over 50 years Chevron has operated in

areas of Nigeria where weakened health systems have
contributed to the prevalence of the triad of HIV, Malaria
and TB. These diseases pose challenges which are faced
by our employees, their families, the communities in
which we operate, and ultimately, the business. We
recognize that supporting public health is core to the
success of our business in the society. Methodology: The
strategy is to achieve early detection and proper
management of tuberculosis, chest and lung diseases in
Nigeria which aligns with the national strategy. The
objectives are to provide infrastructural support; medical
equipment and supplies; capacity building targeting
health workers; and advocacy on TB disease in 25 States
over 8 years. Activities already undertaken include the
conduct of a needs assessment, building and equipping of
23 chest clinics in state-owned health institutions across
the country, provision of medical supplies and training of
health workers. The clinics have consulting rooms,
wards, and fully-equipped laboratories with mobile Xray units and molecular diagnostic (GeneXpert) machines. In addition to delivering health benefits, the
program has generated more than 1000 jobs and
opportunities for people and local firms.
Results and lessons learnt: A practical training on
molecular-genetic diagnostics of tuberculosis was organized for over 100 laboratory attendants and supervisors
in collaboration with the National TB Program as at
2014. The training focused on providing participants
with basic computer skills, equipment operation, turnaround maintenance competency and record keeping;
this training targeted health workers from the chest
clinics and neighboring hospitals. The program has
provided 18 GeneXpert machines representing 19 % of
the National donor contributions. From which a total of
31 571 sputum were tested. In 2014 a total of 14 352
presumptive TB cases were registered in the chest clinics;
with 1124 smear positive cases examined. 701 smear
positive TB cases were traced and identified while 321
TB contacts were confirmed positive. A total of 506
MDR suspects were traced and identified with 92 drug
resistant cases identified and 53 enrolled for care.
Conclusion: The above best practice summarises the
contribution of the Agbami Coventurers towards control
and management of TB in Nigeria.
S488
PC-1180-06 Characteristics of patients with

active TB and their relation to treatment
outcomes in patients previously treated in two
DISAS of Metropolitan Lima
PC-1181-06 Prevalence, risk factors, and

treatment outcomes of isoniazid- and
rifampicin-monoresistant pulmonary
tuberculosis in Lima, Peru
C C Contreras,1 M Lindeborg,2 D Chen,1 M F Franke,3

S Shin3 1Socios En Salud Sucursal Peru, Lima, Peru; 2Harvard
College, Boston, MA, 3Harvard Medical School, Boston, MA,
USA. e-mail: ccontreras_ses@pih.org
L Villegas,1 L Otero,2 T Sterling,1 E Gotuzzo,2 C Seas2

Vanderbilt University School of Medicine, Nashville, TN,
USA; 2Universidad Peruana Cayetano Heredia, Lima, Peru.
e-mail: leonela.a.villegas@vanderbilt.edu
Background: Tuberculosis (TB) is a product of socioeconomic, political, and medical factors. This study identifies characteristics, in patients who have already been
treated, that influence the risk of having an unfavorable
treatment outcome in Lima, Peru. It focuses on differences between patients from the public sector (PS) and
patients who auto-administered treatment (AA), or
obtained treatment from outside of the PS.
Design/Methods: 1545 patients older than 18 years of
age, with a past treatment and current diagnosis of active
TB, of 145 establishments in Lima Metropolitana, were
included between January 2005 and May 2008. Information was taken from clinical histories, treatment
cards, and laboratory registries. A statistical analysis
was carried out through Epi Info 3.4.3.
Results: Within the cohort of 1545, 1404 were from the
public sector (PS) and 141 were auto-administered (AA).
Following univariate analysis, AA patients had better
outcomes than SP patients (OR 1.85 [1.25, 2.74], P
0.002). For example, non-adherence to TB treatment
(more than 20% of days without treatment, n 1530)
was greater in the SP group than in the AA group (26.0%
vs. 16.7%, P 0.02). However, the correlation
descreased upon multivariate analysis (OR 1.46, [0.94
2.26] P 0.10). With respect to demographics, multivariate analysis revealed that finishing secondary school
(OR 1.85 [1.45, 2.36], P ,0.0001) and having a higher
body mass (OR 1.12 [1.08, 1.16], P , 0.0001) increase
the probability of favorable results. On the other hand,
the consumption of alcohol (OR 0.59 [0.44, 0.81], P
0.001) and drugs (OR 0.49 [0.35, 0.70], P 0.0001);
coinfection with HIV (OR 0.67 [0.45, 0.98], P 0.04);
and having resistance to H or R (OR 0.45 [0.30, 0.69], P
0.0002) and to H and R (OR 0.29 [0.21, 0.40], P
,0.0001), are all negatively associated with treatment
outcomes.
Conclusion: The recurrence of TB should be attended by
state health services and the public sector, taking into
account the diverse social factors that influence clinical
results of patients. There should be greater communication between the public and private sectors, in order to
acheive effective management, monitoring, and reporting of cases following national guidelines.
Background: Isoniazid (H) and rifampicin (R) are the

two most efficacious first-line agents for tuberculosis
(TB) treatment. Peru has had stable rates of multidrug
resistant (MDR) TB, accounting for 3.9% of new cases
and 35% of retreatment cases in 2013; however, there
are limited data on H and R monoresistance. We assessed
the prevalence of H and R monoresistance, associated
factors, and the effect on treatment outcomes.
Methods: A prospective observational cohort study
actively enrolled adults with a first episode of smearpositive pulmonary TB from 34 health facilities in North
Lima, Peru, from March 2010 through December 2011.
A structured questionnaire was administered at enrollment. A sputum sample at diagnosis was cultured on
Lowenstein-Jensen
(LJ) media and drug susceptibility
testing (DST) was performed using the LJ proportion

method at the following concentrations: H at 0.2 ug/ml
and 1 ug/ml and R at 40 ug/ml. Patients were followed up
at the end of each treatment. We determined the patients
characteristics associated with H and R monoresistance
and evaluated the effect of resistance on treatment
outcomes (lost to follow up, death, transfer, and failure
were considered poor outcomes).
Results: Of 1292 patients enrolled, 1039 (80%) were
culture-confirmed. H monoresistance was present in 85
(8%) patients, of whom 45 (53%) were resistant to 1 ug/
ml of H. R monoresistance was present in 24 (2%)
patients. The best prediction models for H monoresistance (compared to H susceptible) and for R monoresistance (compared to R susceptible) are shown in the Table.
Variables that were not predictors of monoresistance
included prior H prophylaxis, history of tobacco use, and
age. H monoresistant patients had a higher risk of poor
treatment outcomes (OR 2.46; 95%CI: 1.40-4.30)
after adjusting for prior imprisonment (OR .25;
95%CI: .03-1.88) and alcohol abuse (OR .57;
95%CI: .29-1.11). In addition, R monoresistant patients
had a higher risk of poor treatment outcomes (OR
2.53; 95%CI: 1.02-6.27) with no measured confounders.
Conclusion: H and R monoresistance were associated
with a significantly increased risk of poor outcomes, after
controlling for potentially confounding variables. This
underscores the importance of early case detection and
initiating appropriate TB treatment regimens in these
patients.
57. When the tail wags the dog: drug

resistance in TB programmes
PC-1182-06 Sputum bacterial load predicts
multidrug-resistant tuberculosis in retreatment
patients: a case-control study
M Sander,1 C Vuchas,1 H Numfor,1 A Nsimen,1
J-L Abena Foe,2 J Noeske,3 A Van Deun,4,5 K Morgan6
1
Tuberculosis Reference Laboratory Bamenda, Bamenda,
2
National TB Program, Yaounde, 3Independent Consultant,
Yaounde, Cameroon; 4Institute of Tropical Medicine,
Antwerp, Belgium; 5International Union Against Tuberculosis
and Lung Disease, Paris, France; 6University of Liverpool,
Leahurst, UK. e-mail: melissa.sander@gmail.com
Background: Rapid and effective diagnosis of multidrugresistant tuberculosis (MDR-TB) is an essential component of global tuberculosis control, but most MDR-TB
cases are still not diagnosed. Our objective was to assess
if patient sputum bacterial load can be used to identify
patients that are at higher risk of multidrug resistance.
Design/Methods: We used a case-control study and
multivariable logistic regression models to investigate
associations between MDR-TB and sputum bacterial
load among retreatment TB patients at a reference
laboratory in Cameroon.
Results: MDR-TB was associated with a smear microscopy grade of 3 (odds ratio, 21.9; 95% confidence
interval [CI] 6.2-76.8) or 2 (odds ratio, 10.8; 95%CI
2.9-40.7), as compared to a result of 1, scanty or smearnegative, among 80 MDR-TB cases and 521 controls.
MDR-TB was associated with automated time to
detection of 6160 hours (odds ratio, 2.2; 95%CI, 1.14.7) as compared to .160 hours among a subpopulation of 47 cases and 350 controls.
Conclusion: A higher sputum bacterial load is associated
with MDR-TB in retreatment patients in Cameroon. If
this association proves generalizable to other populations, prioritizing more highly smear positive patients for
laboratory screening could provide a simple and effective
approach to improve the detection of MDR-TB in
PC-1183-06 Is initial loss to follow-up

overestimated in Indias TB programme?
S Shastri,1 S Burugina Nagaraja,2 S Katakol1 1Lady
Willingdon State TB Centre, Bangalore, 2ESIC Medical
College and PGIMSR, Bangalore, India. e-mail:
susha007@gmail.com
Background: In India, despite the efforts from national

TB control programme, tuberculosis remains the major
public health problem. As a policy in the programme, all
those patients diagnosed as smear positive tuberculosis
are initiated on treatment within seven days of diagnosis.
Those smear-positive pulmonary tuberculosis (PTB)
patients not confirmed as starting treatment are reported
as initial loss to follow-up; irrespective of whether the
patients die, migrate or access to private treatment. We
S489
conducted the study to determine the reasons for initial

loss to follow-up under programmatic settings.
Methods: The study was conducted in Karnataka, a
southern state in India with a population of 63 million. A
validated data collection format was used to gather the
information from the available programmatic records
and reports. The data was collected from 31 districts for
the period October to December 2014. The field level
program managers were trained to collect the information by making home visits; if the patients were not
available during the visits information was gathered from
the patients relatives or neighbours. The data collection
forms were entered in microsoft excel formats and were
compiled at state level for analysis.
Results: A total of 12 159 new sputum smear positive
cases were diagnosed during the study period and 8318
(68%) patients were registered for TB treatment. Out of
the 3841 that were initial loss to follow up as per the
definition, 264(7%) patients died before start of the
treatment, 204(5%) patients migrated/provided wrong
address, 149(5%) patients referred outside the state but
no feedback was received about starting the treatment
and 547(14%) patients registered for the treatment in the
subsequent 3 months.
Conclusion: The study findings suggest that the current
recording and reporting system in the programme is
inflating the initial loss to follow-up with inclusion of
deaths. Policy change in the recording and reporting has
to be made to ascertain the cause of initial loss to
follow-up routinely. Provision of electronic technologies
at the microscopy centre level to track the patients will
strengthen the timely initiation of treatment.
PC-1184-06 Why the current TB performance in

Nigeria? Obvious but hidden deterrent factors
M Gidado,1 S Useni,1 J Onazi,1 N Peter,1 E Mitchell,1
C Fischer,2 G Akang,3 R Eneogu3 1KNCV TB Foundation,
Abuja, 2USAID, Abuja, 3Federal Ministry of Health, Abuja,
Nigeria. e-mail: gidadomansu@yahoo.com.au
Introduction: Nigeria is ranked 3rd among high TB

burden countries in the world. Despite its existing service
coverage for microscopy labs of 1:109 000, and 79%
DOTS coverage by population; the case notification rate
for 2013 was 57.3/100 000 population which is 17% of
the estimated incidence (338/100 000 population). The
aim of the assessment was to unveil the hidden deterrents
on case notification to guideline specific interventions.
Methods: An assessment tool was developed which
includes TB disease burden; HIV burden; service
coverage and distribution by population, LGA, and
among all existing health care facilities in the states;
functionality of existing DOTS and laboratory facilities;
and health system related factors. Teams of assessors
were trained on the tool and subsequently the tool was
pre-tested and adjusted. The assessment was conducted
within two weeks in 12 selected states representing each
of the six zones in Nigeria. This involved desk review of
2014 program data at state & LGA levels, in-depth
interview/focused group discussions and health facility
S490
visits. Data was analyzed using simple descriptive

statistics of all the indicators disaggregated by LGAs.
Among all the states there was very low TB CNR an
average of 52/100 000 population (range 31-77/100 000
popn); proportion of childhood TB among notified cases
varied by states from 1.5% to 13%; HIV prevalence
varied significantly by state (range 1% to 15.2%); DOTS
and AFB microscopy coverage by population was 0.8/
25,000 popn and 0.5/50 000 popn, with huge variation
by state and LGAs respectively; proportion of existing
DOTS clinics not reporting a case within a period of two
quarters was 29.5% (range 4.3% to 54%); proportion of
microscopy centers not reporting within a quarter was
24.5% (range 2.6% to 64.%); the coverage of TB
services within total health care facilities was 20% (range
7.6% to 39%).
Conclusion and recommendations: TB program interventions should be state and LGA specific; emphasis
should be on fixing not reporting sites rather than
expansion; focus on community and health care staff
awareness on TB and TB service delivery points; and
develop a mechanism for partners coordination at state
level using state specific operational plans.
PC-1185-06 Prevalence of pre-XDR-TB and XDRTB among MDR-TB patients registered at the
Ojha Institute of Chest Diseases, Karachi
N Rao,1 S Baig,1 N Ahmed,1 D Rao2 1Ojha Institute of Chest
Diseases, Karachi, Hamdard College of Medicine & Dentistry,
Karachi, Pakistan. Fax: (92) 219 926 1470. e-mail:
nisar.rao@aku.edu
Background: Drug resistant tuberculosis is a major

public health problem. Serveillance of sensitivity pattern
of MTB against anti-tuberculosis drugs is crucial for the
sucessful management of tuberculosis. Pre-XDR and
XDR-TB are threats for the sucessful control of
tuberculosis. Data on this issue will help us to determine
sucess of our already running TB control program.
Design/Methods: We reviewed the record of all patients
registered in MDR clinic of Ojha Institute of Chest
Diseases between January 2010 and December 2014.
Data was collected about clinical details and resistance
pattern.
Results: A total of 919 patients were enrolled during the
period. 465(51%) were male. Most of the patients (718/
919) were young i.e. from 15-44 age group. The preXDR-TB and XDR-TB prevalence rate among MDR-TB
patients were 363/919 (39.5%) and 18/919 (02%)
respectively. Out of 18 XDR-TB patients 05 were
resistant to Km (Kanamycin) Am (Amikacin) Cm
(Capreomycin) FQ (Fluoquinolone), 04 were resistant
to Km FQ, 03 against AmFQ and 06 against CmFQ.
Out of 363 Pre-XDR-TB patients 332 were resistant to
FQ, 04 to Km, 07 to Cm, 02 were resistant to Km Am
and 18 were resistant to Km Am Cm.
Conclusion: There is high prevalence of pre-XDR
tuberculosis in patients registered at Ojha Institute of
Chest Diseases, Karachi, Pakistan. The XDR- TB cases
are also reported and registered. There is a need to closely
follow the new TB patients registered in chest clinics for

proper treatment and compliance to prevent drug
resistant tuberculosis. There is also a need to ban over
the counter sale of ATT and specifically quinolones.
PC-1186-06 Description of new pulmonary

tuberculosis cases in southern India
G Roy,1 R Kubiak,2 A Sivaprakasam,1 N Hochberg,2
S Govindarajan,3 P Salgame,4 J Ellner,2 S Sarkar1
1
Jawaharlal Institute of Postgraduate Medical Education and
Research, Pondicherry, India; 2Boston University Medical
Center, Boston, MA, USA; 4Government of Puducherry,
Puducherry, India; 5Rutgers New Jersey Medical School,
Newark, NJ, USA. e-mail: rachkubiak@gmail.com
Background: India has the largest tuberculosis (TB)

burden with an estimated 2 million cases annually
accounting for over 20% of the worlds cases. Recent
data are lacking about the demographic characteristics of
TB cases in southern India. Understanding the epidemiology of TB in India is critical for the development of
effective TB control measures.
Design/Methods: Data were collected through the
Regional Prospective Observational Research for TB
(RePORT)-India collaboration. New sputum smearpositive pulmonary TB cases diagnosed by the Revised
National TB Control Programme (RNTCP) in Puducherry and the Cuddalore district of Tamil Nadu, India
were enrolled within one week of treatment initiation
starting in May 2014. Standardized questionnaires
including the Alcohol Use Disorder Identification Test
(AUDIT-C) were used to collect a medical history and
socio-demographic data. Sputum was collected to
document culture positivity. We compared our findings
to 2011 Indian census data and the existing literature and
used the binomial test for significance testing.
Results: Of the 232 smear-positive patients diagnosed
through RNTCP, 210 were culture-confirmed and
included in this analysis. A greater proportion of cases
were male than in the general populations in both
Puducherry (72.6% vs. 49.1%, P , 0.0001) and
Cuddalore (74.3% vs. 50.1%, P 0.0058) and 15-59
years old; 140 (80%) in Puducherry and 30 (85.7%) in
Cuddalore, compared to 66.0% (P , 0.0001, 0.0160,
respectively) in the general populations. Among cases,
the mean reported household size in Pondicherry was 3.8
people (range 2-7) and 3.1 in Cuddalore (range 2-4)
compared to 4.1 in both states. Most cases were lower
castes (206, 98.1%) and reported low monthly household incomes of 3000-5000 rupees (48-80 USD; 102,
45%). The World Health Organization estimates 27.1%
of south Indians currently or formerly smoke compared
with 102 (48.6%) of cases. Harmful or hazardous
alcohol use was reported by 76 (36.2%) cases, which is
markedly higher than the prevalence rates of 3.7% and
14.9% found in recent studies in rural Tamil Nadu.
Previous studies found diabetes prevalence rates in
southern Indian populations in the range of 6-14%,
while among TB cases in this study, 62 (29.7%) reported
a previous diagnosis of diabetes. Only one (0.05%) case
S491
was HIV positive suggesting a prevalence similar to the

national rate of 0.27%.
Conclusion: New sputum-positive pulmonary TB cases
were more commonly male, of lower socioeconomic
status, and had a higher prevalence of diabetes and
alcoholism (but not HIV) compared to the general
population.
PC-1188-06 Risk factors associated with

tuberculosis in Kazakhstan: implications for
disease prevention and control
PC-1187-06 Role of exogenous reinfection in

patients with recurrent tuberculosis in the
United States
A Davis,1 A Aifah,1 A Terlikbayeva,2 Z Zhumadilov,2

T Abildaev,2 M Darisheva,2 N Schluger,1 N El-Bassel1
1
Columbia University, New York, NY, USA; 2Columbia
University Global Health Research Center of Central Asia,
Almaty, Kazakhstan. e-mail: alissad22@gmail.com
J Interrante,1,2 M Haddad,1,2 L Kim,1 N Gandhi2 1U.S.

2
Emory University, Atlanta, GA, USA. e-mail:
jdi5w@virginia.edu
Background: Although substantial gains in reducing

tuberculosis incidence in the United States have occurred
over the past 2 decades, progress toward the national
goal of tuberculosis elimination is slowing. One area to
target interventions in order to achieve elimination is in
reducing the proportion of tuberculosis cases that recur.
The etiology of recurrent tuberculosis is typically thought
to be reactivation of residual Mycobacterium tuberculosis infection; however, recent studies suggest that
reinfection with new strains may account for a greater
proportion of recurrent tuberculosis in the United States
than previously recognized. The objective of this study
was to characterize the etiology of recurrent tuberculosis
in the United States using genotyping of initial and
recurrent M. tuberculosis isolates.
Design/Methods: The study population for this retrospective national population-based cohort in the United
States was drawn from all persons reported to the
National Tuberculosis Surveillance System as having 2
episodes of tuberculosis during 19932011. Genotyping
results were available from both tuberculosis episodes for
194 cases. We analyzed the proportion of recurrent
tuberculosis attributable to reinfection and the risk
factors associated with reinfection.
Results: Among 136 recurrences meeting the inclusion
criteria for this study, 20 (15%) had differing genotypes
between the initial and recurrent M. tuberculosis isolates,
and thus, were likely attributable to reinfection; genotypes were the same for 116 (85%) cases and were
categorized as reactivation. Using exact logistic regression, factors associated with reinfection included Mexican birth with immigration to the United States within
12 years of the second episode (adjusted odds ratio, 10.7;
95% confidence interval, 1.786.3), and treatment of the
first episode exclusively by directly observed therapy
(adjusted odds ratio, 4.5; 95% confidence interval, 1.0
29.2).
Conclusion: Reinfection with a new M. tuberculosis
strain was the cause for approximately 15% of recurrent
tuberculosis cases in this US cohort. The proportion
caused by reinfection increased to as high as 60% in
certain subpopulations, such as recent immigrants from
Mexico. Public health initiatives should focus on these
populations and their areas of residence, work, and
recreation to stop further transmission of M. tuberculosis.
Background: Kazakhstan has experienced an increase in

tuberculosis (TB) infections. Identifying key populations
and the social determinants of TB is essential for
controlling the spread of the disease. Studies in other
regions have indicated that smoking, alcohol and
substance use are risk factors for TB and decrease the
effectiveness of TB control efforts. Furthermore, other
studies have found that HIV positive individuals and
individuals with a history of incarceration have an
elevated risk for TB. This study examines the relationship
between TB, criminal history, tobacco, alcohol and
substance use, and HIV in Kazakhstan.
Design/Methods: A case-control study included 1590
participants recruited from four regions in Kazakhstan.
We defined index cases as those who had a positive
culture, confirmation by positive smear and nucleic acid
testing or X-ray confirmation and response to antituberculosis drugs. Of the total sample, 598 (37.6%) met
the criteria as index cases (e.g. diagnosed with TB).
Participants completed a survey containing items on
sociodemographics, history with the criminal justice
system, tobacco, alcohol and drug use, and HIV
diagnosis. Multivariate regression was used to examine
associations between a TB diagnosis and a number of risk
factors: criminal history, tobacco, alcohol and drug use
and injection. Fishers exact tests were used to calculate
significant differences in HIV positivity between individuals with TB and without TB.
Results: After controlling for demographic variables
(sex, age, ethnicity and education), participants who
had ever been charged with a criminal offense (adjusted
odds ratio (AOR) 3.51, 95%CI [1.93-6.37]), spent time
in jail (AOR 6.26, 95%CI [2.66-14.70]), ever used nonsmoking tobacco (AOR 1.99, 95%CI [1.26-3.14]), ever
smoked tobacco (AOR 1.59, 95%CI [1.21-2.08]), ever
used alcohol (AOR 1.54, 95%CI [1.22-1.95]), and ever
used marijuana (AOR 1.99 [1.04-3.82]) were more likely
to have a TB diagnosis. No significant associations were
found between TB and injecting drug use. Individuals
with TB had higher rates of HIV positivity than
individuals without TB (1.2% vs. 0.1%, P 6 0.05).
Conclusion: The findings about risk of TB are consistent
with previous studies. The association between TB and
criminal history and tobacco, alcohol and drug use in
Kazakhstan suggests a need for targeted intervention and
prevention approaches for individuals in the criminal
justice system and for those who use substances.
S492
PC-1189-06 Projecting the impact of alternative

drug susceptibility testing strategies on future
TB drug resistance
M Fofana,1 S Shrestha,1 G Knight,2 T Cohen,3 R White,2
F Cobelens,4 D Dowdy1 1Johns Hopkins Bloomberg School
of Public Health, Baltimore, MD, USA; 2London School of
Hygiene & Tropical Medicine, London, UK; 3Yale School of
Public Health, New Haven, Connecticut, USA; 4KNCV
Tuberculosis Foundation, The Hague, Netherlands. e-mail:
mariam.fofana@gmail.com
Background: Drug resistance presents a major obstacle

to tuberculosis (TB) control efforts. Although the
incidence of TB is decreasing in most settings, many
areas have seen an increase in multidrug-resistant (MDR)
and extensively drug-resistant forms of TB. Understanding the potential future trajectories of drug resistance and
the impact of interventions such as drug susceptibility
testing (DST) can help improve TB control policies.
Methods: We used a multi-strain dynamic transmission
model of TB calibrated to the Southeast Asia region to
project trajectories of resistance to rifampin, moxifloxacin, and pyrazinamide from 2015 to 2035. We
randomly sampled values of key parameters for which
data are unavailable or of poor quality, including
probabilities of acquiring drug resistance during treatment, relative transmission fitness of drug-resistant
strains, and treatment outcomes for different drug
resistance profiles. We then selected only those parameter
sets that were consistent with current data on TB
incidence, prevalence and drug resistance in Southeast
Asia (n 1190), and used them to evaluate a range of
possible future epidemics of MDR-TB under different
algorithms for the implementation of DST.
Results: Initial MDR prevalence in 2015 was calibrated
to a median of 3.1% (interquartile range [IQR] 2.53.8%). Under the assumption of no DST, the proportion
of MDR among all TB cases by 2035 varied widely
(median 7.5%, IQR 5.2-12.4%), and was consistently
higher among previously treated (median 30%) than
treatment-nave cases (median 5%). A focused strategy,
in which TB patients receive DST for rifampin only if
they experience treatment failure, substantially reduced
projected MDR prevalence among all TB cases in 2035
(median 5.0%, IQR 3.7-7.4%). Testing for rifampin
resistance among all patients who were previously
treated for TB provided an additional but small
incremental benefit (median proportion MDR in 2035
4.6%, IQR 3.4-6.6%).
Conclusions: Given current epidemiologic data, the
future trajectory of TB drug resistance in Southeast Asia
remains highly uncertain; future trends depend on the
transmission fitness of resistant strains which may vary
over time, adding further uncertainty. Nevertheless,
implementation of DST can have an important population-level impact, by reducing the probability of epidemiologic scenarios in which a large proportion of TB
cases are MDR. Scale-up of DST should prioritize
patients with a previous history of treatment failure.
58. Advancing the END TB strategy and

other policy issues
PC-1191-06 Impact of social support in all forms
of TB case detection and in MDR-TB treatment
outcome
S Islam,1 M Rana,1 M Siddiqui,1 M Akramul Islam,1
S Reza,1 S Alam,1 M Haque2 1BRAC, Dhaka, 2NTP, Dhaka,
Background and challenges to implementation: BRAC

has been implementing TB control programme in
Bangladesh in collaboration with the National TB
control Programme for about 20 years. Since its
inception, the programme has evolved to its current
stage through identifying implementation gaps and
adapting the best solution in country settings. The social
support is one of those initiatives that were adopted
recently to overcome the economic, geographical, sociocultural and health system barriers for TB patients,
especially for the poor and protect them from increasing
socio-economic vulnerability due to more out of pocket
expenditure for TB diagnostic purpose.
Intervention or response: In a country like Bangladesh
the diagnosis of smear negative and extra pulmonary TB
is expensive. To mitigate this issue, BRAC began
providing financial support for the poor presumptive
TB patients from January 2013 as a social protection
initiative. Poor TB symptomatic with smear negative,
receive investigation and transportation costs to attend
diagnostic facilities, and consultation fees in case they
visited private sector and diagnosed as TB (USD 13).
Poor people with extra pulmonary TB symptoms and
child presumptive cases receive same supports (USD 33)
for diagnosis. In addition, drug resistant (DR) TB
patients are getting nutritional support during treatment
period for treatment adherence.
Results and lessons learnt: Among 1 73 568 presumptive
TB patients who received social support in 2013, the
number of TB patients identification was 48 322 (28%).
Similarly of 1 52 063 presumptive TB patients who
received social support in 2014, the number of TB case
identified 49,888 (33%). Considering total patient
identified in 2012, 2013 and 2014 was respectively 1
02 716, 1 21 239 and 1 24 286 which showed an

increasing pattern. Treatment outcome of DR TB cases
was 71% in 2012 cohort.
Conclusions and key recommendations: Introducing
social support effectively by providing diagnostic support to poor TB presumptive increased all types of case
notification and improved DR TB treatment success rate
which finally facilitate in reducing TB burden in the
community.
PC-1192-06 School-based TB prevention

activities in Georgia
G Bakhturidze,1,2 Nana Peikrishvili1,2 1Georgian Health
Promotion and Education Foundation, Tbilisi, 2Georgian
Health Promotion and Education Foundation, Tbilisi, Georgia.
e-mail: iayd@yahoo.com
Background and challenges to implementation: While

adults are more locked into existing stereotypes regarding TB, youth are more open to the new information and
have joined information activities with great enthusiasm.
This experience we earned in 2013, when implemented
project related information and communication campaigns in Universities and in some public schools
supported by USAID.
Intervention or response: Considering above needs, we
organized school based TB prevention actions targeted
youth from all regions of Georgia during 2014. We
worked very actively with the teachers of public schools
in target towns. The special trainings of teachers were
organized by participation of 130 teachers. Also indirectly we approached about 400 other teachers and
school administration members, who were involved in
information meetings and other events in target schools.
The second main targets for our activities were pupils of
public schools. We selected 130 schools throughout
Georgia and totally directly involve about 3000 pupils in
the age of 13-17 years through information meetings and
presentations. Indirectly, we approached about 20 000
pupils and 10 000 adults through different activities
organized by schools and distribution of information
materials.
Results and lessons learnt: Methodical guideline for
teachers regarding TB created. 130 teachers have
received ToT regarding TB issues and raised their
capacity. Information flier for youth regarding TB issues
distributed (20 000 copies). T-shirts with slogans and
logos (500 pieces) were used during school activities.
Information meetings in 120 schools organized and
about 3000 pupils attended. At least 20.000 pupils and
10.000 adults (parents, teachers, etc.) reached through
teachers trainings, pupils activities, distribution of
printed materials and received information about TB.
At least 80 different activities/events organized by
schools. 3 articles appeared in regional newspapers and
2 reportages in regional TV/radio reaching totally about
300 000 population. 5 talk-shows organized in internet
TV and radio covering all issues of TB reaching about
500 000 people.
S493
Conclusions and key recommendations: School based TB

prevention programs are essential to raise awareness
among new generation. Appropriate curriculas and
trained teachers will sustain the process and prevent TB
among youth and their rounding communities.
PC-1193-06 Understanding the effectiveness of

engaging pharmacists in TB care and control
from the TB patients perspective
V Panibatla,1 S Prasad2 1TB Alert India, Hyderabad, 2Eli Lilly
and Company (India) Ltd., New Delhi, India. e-mail:
vikas@tbalertindia.org
Background: Pharmacists are important community

stakeholders in India. They also are the first contact
point for consultation for any sort of illnesses. In a
project, with aim of engaging pharmacists in National
TB Program (NTP), supported by UWW &Lilly MDRTB Partnership and implemented by TB Alert India
results are encouraging. Pharmacists from Feb 13 to Feb
15 referred 2175 people with TB symptoms for testing.
Around 92% (1989/2175) got tested and 11% (225/
1989) are diagnosed with TB. To understand how much
time people diagnosed with TB would have taken for TB
testing if pharmacist would have not referred, a study
was taken up. Study also focused on how satisfied are
they with pharmacists service.
Intervention: An independent external consultant interviewed 19% (42/225) of people who were diagnosed
with TB due referral by pharmacists in two districts
among four implementing districts in Telangana state,
India. Line listing of all such in two districts was
prepared. People diagnosed with TB listed at even serial
number in the line listing were contacted for willingness
to participate in the study. Based on the response TB
patients were either interviewed over phone or through
personal interview.
Results: Around 81% (34/42) male and 19% (8/42)
female were interviewed. Majority (32 /42) were in age
group of 19-59 years. About 52% (22/42) reported that
they reached pharmacy shop within 7 days of onset of
symptoms. Around 26 out of 42 interviewed (62%)
reported that it was their first visit to pharmacy shop
after onset of symptoms. About 61% (26/42) of referred
people reached for TB testing next day of the referral and
with no out of pocket expenditure. Near about 3/4th (34
/42) were referred for testing in the first visit to
pharmacy. About 25% (11) shifted to NTP treatment
from private hospitals after visiting Pharmacist. Major
chunk (70%) reported that they would have waited for
another 7-10 days before going to any doctor if
pharmacist did not refer them. About 86% (36/42)
expressed satisfaction with information given and way
pharmacist behaved with them. Referrals made by
pharmacists reached 10-15 days earlier for testing than
other TB patients at NTP clinic.
Conclusion: Engaging pharmacists proved very useful in
early identification and early treatment initiation.
Pharmacists were able to give correct information and
convince people to go for testing. Effective training,
S494
frequent follow up and reiteration of motivation are key

factors for success.
PC-1195-06 Measuring the effect of two

community-based intervention strategies on
diagnostic delay in Southern Ethiopia
PC-1194-06 Increasing TB case detection in

small-scale private facilities through sputum
sample networking in Nyanza, Kenya
L Blok,1 G Asnake,2 D Gemechu,2,3 M Bakker4 1Royal

Tropical Institute (KIT), Amsterdam, Netherlands;
2
REACH-Ethiopia, Hawasa, Ethiopia; 3London School of
Tropical Medicine, Liverpool, UK; 4Royal Tropical Institute
(KIT), Amsterdam, Netherlands. Fax: (31) 20 568 8444.
e-mail: l.blok@kit.nl
P Izulla,1 J H Ochieng,1 G Cheruiyot,1 J Marwanga1

Population Services Kenya, Nairobi, Kenya. e-mail:
PIzulla@gmail.com
Background and challenges to implementation: Kenya is

among the 22 countries that account for 80% of TB cases
in the world. Close to 20% of TB cases in Kenya are in
Nyanza region in part due to the higher HIV prevalence.
Despite close to 50% of Kenyans seeking health services
in private facilities, their involvement in TB detection
and management has been limited. With funding from
TB Reach, Population Services Kenya (PS Kenya) in
partnership with Kisumu Medical and Education Trust
(K-MET) are carrying out a one year pilot project whose
aim is to increase case detection and treatment of TB in
48 franchised private facilities in Nyanza. At baseline
only 12 facilities could offer sputum microscopy and 12
were offering treatment. Service providers were trained
in TB case management and sputum microscopy where
applicable. 92 community health volunteers (CHVs)
were trained to conduct community screening and
mapped to the facilities for referral of TB suspects.
Screening only sites were linked to facilities with
microscopy and treatment services for referral of TB
suspects. Despite increase of sputum microscopy services
from 12 to 24 facilities and treatment services to 18
facilities in first 2 quarters, many TB suspects were lost to
follow up through double referrals from community to
screening only sites and then to microscopy sites.
Intervention or response: Sputum sample networking
system for non-diagnostic facilities was introduced in
Q3. Screening only facilities were provided with sputum
containers for collection of sputum and cooler boxes for
transportation to the nearest diagnostic center.
CHVs.transported the cooler boxes from one site to the
next and followed up on the results. Sputum was
collected on designated days and transport costs to the
referral sites catered for by the program to minimize
delays in specimen movement. Data was captured on a
paper based data collection tool, verified and subsequently fed into the program HMIS. Service quality was
assured by field based demand creation and clinical
coordinators that work closely with the providers and
CHVs.
Results and lessons learnt: There was a four-fold increase
in the numbers who tested sputum positive by microscopy from 79 in Q2 to 328 in Q3.
Conclusions and key recommendations: Sputum sample
networking improved TB case detection among patients
seeking TB services in small scale private facilities with
inadequate infrastructure by minimizing movement of
patients from one facility to another.
Background: In Ethiopia health extension workers

(HEW) as first line cadres have been involved in
identifying and referring people at risk of tuberculosis.
REACH Ethiopia has implemented an innovative package of community based enhanced TB case finding
expanding the role of HEW from referral of symptomatic
people to include sputum collection, smear preparation,
and treatment follow up once diagnosis is confirmed.
This approach led to a marked increase in case
notification in Sidama zone. Since 2013 the program
successfully extended into 4 additional zones trying
slightly less intensive packages of support in three of
these zones. This study aims at measuring the effect of the
full package (FP) and the targeted package (TP) on
diagnostic delay.
Design/Methods: Diagnostic delay was defined as time
between onset of symptoms and date of diagnosis, and
further broken down in health seeking delay (time from
symptoms till first visit to a health provider) and health
system (HS) delay (from first visit till diagnosis). Data
were collected through systematic interviews with SS
patients and validated against suspect and treatment
registers during baseline and intervention. Seventy seven
patients were included, 40 baseline and 37 intervention
(21 FP and 16 TP). Median number of days were
calculated and compared using a non-parametric test
(Kruskal Wallis).
Results: The overall diagnostic delay reduced from a
median of 77.0 days (IQR 31.5-122.5) at baseline to 31
days (IQR 28.0-62), P 0.0795, due to a significant
reduction in HS delay (from median 27.5 days (IQR 2.563.5) to 7 days (IQR 0-31), P 0.0366). The median HS
delay under the FP was slightly shorter than under TP (4
and 7.5 days respectively), but this difference was not
significant P 0.893. No change was observed in the
delay in seeking care.
Conclusion: Our study shows that community based case
finding efforts can accomplish significant reductions in
time till diagnosis in particular reducing health systems
delay and referral time. The seemingly little influence on
delay in seeking care is however worrisome given its
relative importance in total delay. Recall bias may have
affected the accuracy of time estimates, but was partially
compensated by triangulation of data for the health
system period. The difference in diagnostic delay
between the two intervention packages was non-significant and will need to be further evaluated against cost
implications.
PC-1196-06 Engaging commercial transporters

(Keke NAPEP) in mobilizing community
members for TB services: experiences from Imo
state, Nigeria
.A F Omoniyi, A Awe, G Akang, G Madu,
1
E Ubochioma,2 M Gidado,4 C Orji,3 R Eneogu2 1World

Health Organization, Abuja, 2NTBLCP, Abuja, 3Imo State TBL
Control Programme, Owerri, 4KNCV, Abuja, Nigeria. e-mail:
omoniyifadare@yahoo.com
Background and challenges to implementation: Tuberculosis is major public health problem in Imo state, a
state with a population of about 3.9million inhabitants.
The National KAP survey conducted in 2012 revealed
that only 27% of the populace have correct knowledge of
the cause of TB, early symptoms and signs of TB. This
low knowledge is one of the major factors responsible for
low TB case finding in all the states in Nigeria, Imo state
not exempted. Only 8% of the estimated TB cases in Imo
state were notified in 2013. The programme in order to
address this conducted an outreach campaign in Imo
state during which commercial transporters popularly
referred to as Keke NAPEP were used to mobilized
members of the community for TB services for one week.
The aim of this study is to assess what worked and the
lessons learnt in the use of commercial transporters
(Keke NAPEP) in mobilizing community members for
TB services.
Intervention or response: Amakohia/Akauma community in Owerri North LGA of Imo state was selected for the
outreach after analysis of state TB programme data.
Advocacy visits was conducted to the community leaders
during which the use of commercial transporters was
agreed upon for community mobilization during the
outreach. A total of 20 commercial transporters were
selected, the transporters were sensitized for about 1
hour on the content of information to be passed to the
community. They were also given IEC materials in local
languages. Health workers were also trained and
necessary commodities provided.
Results and lessons learnt: An increased average daily out
patient load was observed in the PHC in the community
from an average of about 5 patients per day to about 50
patients daily. A total of 308 presumptive TB cases (125
males & 183 females) were sent for TB diagnosis during
the one week outreach campaign, 5% (15) of the
presumptive TB cases were diagnosed with TB and were
treated.
commercial transporters for active community mobilization resulted in massive turnout of TB suspects for TB
screening during the outreach. These transporters are
well known in the community and therefore present an
opportunity for continuous community mobilization and
raising awareness about TB in the community.
S495
PC-1197-06 Engaging commercial motorbike

riders in creating TB awareness in the
communities in Oyo State, Nigeria
O Lawal,1 O Oladimeji,2,3 O Eltayeb,4 O Daniel,5 A Awe,5
M Gidado6 1Oyo state Ministry of Health, Ibadan, 2Center
for Community Health Care, Research and Development,
Abuja, Nigeria; 3Liverpool School of Tropical Medicine,
Liverpool, UK; 4Damien Foundation, Ibadan, 5World Health
Organization, Abuja, 6KNCV Nigeria, Abuja, Nigeria. e-mail:
droladfb@gmail.com
Background: Poor awareness of free screening and

treatment of TB is one of the major factors responsible
for missed TB cases and delayed diagnosis in Oyo state,
which is one of the thirteen high burden states in Nigeria.
It is therefore necessary to evolve ways of improving the
situation of TB screening and diagnosis. Motorbikes are
major means of transportation in most streets in cities,
towns and villages either commercially or privately
owned. The objective is to highlight the contribution of
this innovative approach to improve TB case notification
in the communities.
Design/Methods: Advocacy meetings with members of
the association of commercial motorcyclists in four
LGAs to facilitate awareness building on TB and contract
engagement to carry the TB banners. Capacity building
on TB and referral linkages involving the 2 DOTS
officers per LGA on data collection tools and 80
members of the association selected on the basis of 20
motorbikes per LGA from 4 LGAS.
Results: A total of 24 DOTS officers and 80 motorcyclists were involved in the project resulting in 3945
presumptive TB referral and 2562 phone calls and text
messages received in the 4th Q 2014 compared to 2341
screened in 3rd Q 2014. The case notification also
increased by 12% from 1623 in Q4 2013. HIV positivity
rate was 7%. Eight percent of those referred were new
smear ve TB cases.
Conclusion: The results indicated that engaging the
motor bikers to carry the banners in their rears have
positive and cost effective impacts for creating awareness
to also compliment the efforts of CBO and CV. We
suggest a bigger try of this strategy.
PC-1198-06 Tuberculosis contact investigation:

perspectives of TB contacts in Kampala,
Uganda
I Ayakaka,1 S Ackerman,2 L Davis,3 A Katamba,4 E Fair,5
A Cattamanchi,3 M Handley,5 J Ggita4 1Infectious Diseases
Research Collaboration, Kampala, Uganda; 2University
California San Francisco, San Francisco, CA, 3Curry
International TB Center, San Francisco, CA, USA; 4Makerere
University, College of Health Sciences, Kampala, Uganda;
5
San Francisco Genaral Hospital, San Francisco, CA, USA.
e-mail: ayakaka@gmail.com

WHO recommended TB contact investigation (CI) be
routinely implemented in high-burden countries like
Uganda. Many national guidelines recommend CI, but
data on how to implement this in routine practice is
S496
limited. We conducted focus group discussions (FGDs)

and in-depth interviews (IDIs) with household TB
contacts to describe their knowledge and attitudes
towards contact investigation and identify barriers to
and enablers of CI in Kampala, Uganda.
Intervention or response: We developed a theoryinformed discussion guide based on the Behavior Change
Wheel, a synthesis of 19 behavior-change frameworks to
assess how contacts described their capabilities, opportunities and motivations (COM) in regard to CI. We
explored acceptability of potential interventions, like
home sputum collection and HIV testing and use of
technology to collect clinical information and communicate results, to address identified barriers. A Ugandan
social scientist and TB epidemiologist led the discussions.Thematic analysis of de-identified FGDs and
interview transcripts was performed to identify barriers
related to COM.
Results and lessons learnt: We held 1 FGD and 13 IDIs
with 36 TB household contacts from 14 households. Of
the 14, 7 had been visited by a community health worker
(CHW) to monitor treatment adherence and the other 7
had not. The main barrier related to capability was
limited TB knowledge especially in households not
visited by a CHW. Barriers to physical opportunity
included clinic travel costs and inconvenient clinic hours.
Barriers to motivation included fear of community
stigma from identification of CHWs in uniform. Most
prevalent in households not yet visited by a CHW,
contacts described lack of confidence in health centers
due to long queues, inattentive medical workers and drug
stock outs in government facilities. Main enablers of CI
cited were CHWs, valued for the ability to provide much
needed information about TB and direct linkage to
providers, reducing clinic waiting times. Communication
with CHWs through home visits or phone contact were
both acceptable and desirable.
Conclusions and key recommendations: Cited barriers to
CI were lack of knowledge, belief that health units
provide poor, challenging to access services. Our analysis
suggests improved education, convenient access to TB
diagnostics, streamlined clinic visits and improved
communication may overcome identified barriers.
Homebased sputum is acceptable.
59. Basic science: host

PC-1199-06 Quantitative membrane
proteomics reveals new functions of CD13 in
Mycobacterium tuberculosis infection
C-P Kuo,1 R Lin1 1MacKay Memorial Hospital, Taipei, Taiwan.
e-mail: pongow@gmail.com
Background: The cellular immune response for Mycobacterium tuberculosis infection remains incompletely
understood. To uncover membrane proteins involved in
this infection mechanism, an integrated approach consisting of an organic solvent-assisted membrane protein
digestion, stable-isotope dimethyl labeling and LC-MS/
MS (chromatographymass spectrometry/mass spectrometry) analysis was used to comparatively profile

the membrane protein expression of human dendritic
cells upon heat-killed M. tuberculosis treatment.
Design/Methods: Organic solvent-assisted trypsin digestion coupled with stable-isotope labeling and LC-MS/MS
analysis was applied to quantitatively analyze the
membrane protein expression of THP-1 derived dendritic
cells. We evaluated proteins that were up-regulated in
response to heat-killed M. tuberculosis treatment, and
applied STRING (Search Tool for the Retrieval of
Interacting Genes/Proteins) website database to analyze
the correlations between these proteins.
Results: By using confocal microscopy and flow cytometry, we found that membranous CD13 expression was
upregulated and was capable of binding to mycobacteria.
Treatment dendritic cell with anti-CD13 antibody during
M. tuberculosis infection enhanced the ability of T cell
activation.
Conclusion: Increasing expression of CD13 on dendritic
cells while M. tuberculosis infection and enhancement of
T cell activation after CD13 treated with anti-CD13
antibody indicates CD13 positively involved in the
pathogenesis of M. tuberculosis.
PC-1200-06 T-lymphocyte dysfunction is

responsible for immunosuppression in active
tuberculosis
C-C Shu,1 J-Y Wang,1 C-J Yu1 1National Taiwan University
Hospital, Taipei, Taiwan. e-mail: stree139@yahoo.com.tw
Background: Tuberculosis (TB) remains one of the most

common infectious diseases and leading causes for
mortality. During active tuberculosis, immunosuppression happens but the nature is unclear. Design: We
enrolled 133 patients with TB and 193 IGRA-screening
TB contacts (95 of them IGRA positive) to compare the
cytotoxic ability by serum level of Fas ligand (FasL) and
tumor necrosis factor receptor 1 (TNFR1). T-SPOT was
used as IGRA method to screen latent TB infection.
Furthermore, we enrolled 3 healthy subjects and 3
patients with TB to check the expression of programmed
cell death-1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) over CD4 lymphocyte.
Results: Serum FasL level was higher in IGRA-positive
contacts than IGRA-negative contacts and TB patients
whereas TNFR1 was higher in TB patients than IGRApositive group (Figure 1). Further check in 3 healthy
subjects and TB patients, PD-1 and CTLA-4 were higher
in TB patients CD4 lymphocyte than control group
(Figure 2). Th1 lymphocyte by double positive stain of
CD4 and interferon-gamma was lower in TB group but
the PD-1 expression was higher. By contrast, Foxp3
positive CD4 lymphocyte was higher in TB patients,
whos PD-1 was also higher than healthy control.
Conclusion: In TB infection, extrinsic pathway of
apoptosis was down-regulated in active infection than
latent status. In cellular pathway, PD-1 and CTLA-4 may
play the important role in the down-regulation mechanism.
S497
entiated T-cells (P 0.016) than blood. Comparing

patients with tuberculosis and LTBI, the proportion of
CD4 CD45RO CD27 CCR7- effector-memory Tcells in blood trended to be reduced. After antigenspecific stimulation of in BAL cells, interleukin (IL)-2
and IFNc were mainly produced from a CD4
CD45RO CD27- CCR7- subpopulation.
Conclusion: Effector T-cells in BAL of patients with
pulmonary tuberculosis are specifically recruited into the
lungs and produce antigen-specific cytokines. Due to the
loss of the CCR7 homing receptor, the intermediate
differentiated T-cells will persists at the focus of infection
and not move back to the lymphnodes. Results from this
study may have implications for vaccine development or
therapeutic interventions.
PC-1202-06 Smoking and interleukin-4 gene C589T polymorphism in patients with

chemoresistant pulmonary tuberculosis
PC-1201-06 Recruitment of antigen specific Tcells in pulmonary tuberculosis
B Kalsdorf,1,2 T Schmidt,3 E Petruccioli,4 J Hofmeister,1
M Sester,3 D Goletti,4 C Lange1,2 1Research Center Borstel,
Borstel, 2German Center for Infection Research, Borstel,
3
University of the Saarland, Homburg, Germany; 4National
Institute for Infectious Diseases, Rome, Italy. Fax: (49)
4537188 3130. e-mail: kalsdorfb@yahoo.de
Background: T-cell mediated immunity is critical for

immune control of tuberculosis. To investigate how
antigen-specific T-cells are attained to the lungs, cells
from peripheral blood and bronchoalveolar lavage (BAL)
of individuals with pulmonary tuberculosis and LTBI
were compared.
Design/Methods: In three centers in Europe, peripheral
blood was drawn and a bronchoscopy was performed on
patients who were suspected of having pulmonary
tuberculosis with negative acid-fast bacilli sputum
smears, who underwent regular bronchoscopy as part
of the diagnostic workflow. Healthy persons with LTBI
(defined by the positive interferon gamma (IFNc)
immune response to M. tuberculosis-specific antigens in
the blood but not in the BAL) served as a control group.
Cells were left unstimulated or stimulated with aCD3,
PPD, ESAT-6 / CFP-10 and CMV for 18 hours.
Immunophenotype was charaterized by effector/ memory (CD3, CD4, CD8, CD27, CD45RO), homing receptor
expression (CCR7) and functionality by antigen-specific
IFNc and IL-2 production. Statistical analysis for paired
data was performed by Wilcoxon Signed Rank test and
comparative analysis between groups by the MannWhitney U-test for nonparametric data.
Results: Fourteen out of 69 patients, in whom the
diagnosis of tuberculosis was confirmed by culture
positivity, were compared with ten healthy participants.
Irrespective of tuberculosis status, BAL T-cells had
significant higher percentages of CD4 CD27- CCR7intermediate differentiated effector T-cells (P , 0.01)
and CD4 CD45RO- CD27- CCR7- terminally differ-
D Butov,1 M Kuzhko,2 N I Makeeva,1 T Butova,1

A Stepanenko,1 N N Pitenko1 1Kharkiv National Medical
University, Kharkiv, 2National Institute on Phtysiatry &
Pulmonology, F.G. Yanovsky NAMS of Ukraine, Kiev, Ukraine.
e-mail: dddimad@yandex.ru
Background and objective: Determine IL-4 gene C-589T

polymorphism in smokers with multi-drug-resistant
tuberculosis (MDR-TB) in lungs on the background of
the respective cytokine production of blood.
Methods: The study included 170 people in Kharkiv
region of Ukraine including 60 patients with MDR-TB
and smoking (1st group), 36 - without MDR-TB and
smoking (2nd group), 14-with MDR-TB and nonsmoking (3rd group ) 30-without MDR-TB and nonsmoking (4th group) and 30 healthy donors (5th group).
Serum levels of IL-4 were evaluated by ELISA. Investigations of gene polymorphisms of this cytokine were
performed using restriction analysis of the amplification
products of specific regions of the genome, i.e., promoter
region C-589T of IL-4 gene. Statistical evaluation: the
obtained data were evaluated by standard Student t-test
and probability relative frequency deviation from a
constant probability of independent trials. The difference
was considered to be significant at P , 0.05.
Results: In the 1st group the levels of IL-4 were
(7.3560.27) pg/L, 2nd (10.1960.42) pg/L, 3rd
(10.5860.66) pg/L, 4th(12.6160.50) and 5th
(29.9961.27) pg/L. Differences between the 1st and
3rd, 2nd and 4th, TB patients and 5th groups were
significant (P , 0.001). Normal homozygote genotype
IL-4 predominated in healthy donors was 56.6769.05%
(n 17) compared to patients with TB: 1st group
11.6764.14% (n 7), 2nd-8.3364.61% (n 3), 3rd
21.43611.38% (n 3) and 4th26.6768.07% (n 9).
Heterozygous genotype IL-4 was observed in
83.3364.81% (n 50) in 1st, 22.2266.93% (n 8) in
2nd, 21.43611.38% (n 3) in the 3rd, 3.3363.28% (n
1) in 4th and 23.3367.72% (n 7) in 5th groups.
Mutation homozygous genotype IL-4 was 5.0062.81%
(n 3) in the 1st group, in 2nd69.4467.68% (n 25),
S498
3rd57.14613.73% (n 8), 4th66.6768.61% (n 20)

and 5th20.0067.30% (n 6). Differences between the
1stand 3rd, 2nd and 4th, TB patients and 5th groups
were significant (P , 0.05).
Conclusion. Smoking contributes to significant activation polymorphism C-589T gene IL-4 heterozygous type
that may lead to changes in the immune system making
susceptible to MDR-TB. Compared to healthy controls
patients with TB had significantly decreases levels of
serum IL-4. This coincided with greater frequency of
heterozygous and heterozygous genotype polymorphism
C-589T of IL-4 gene. In addition, these studies revealed a
significant influence of the polymorphism C-589T gene
IL-4 (with polymorphism IL-2 and IL-10 genes, which
we also studied) on the changes in the population of Thlymphocytes, clinical symptoms, relapse of tuberculosis,
formation destructions in the lung, which may treatment
outcomes in patients with MDR-TB.
PC-1204-06 Epigenomic effects of hookah

tobacco smoke in normal human respiratory
epithelia and lung cancer cells
Similar to CSC, WPC decreased H4K16Ac and

H4K20Me3 levels in SAEC and HBEC. In addition, like
CSC, WPC increased miR-31 and decreased miR-487b
expression in SAEC, HBEC and lung cancer cells.
Furthermore, in a dose-dependent manner, WPC increased expression of Cyp1a1, Cyp1b1, ABCG2,
Lin28B, Myc, SALL4, CTCF, and JARID2 in SAEC
and/or HBEC cells, and up-regulated Cyp1a1, Cyp1b1,
and ABCG2 in lung cancer cells. WPC (2mg/ml)
exposure significantly decreased expression of Dkk-1 in
HBEC and Calu-6 cells. Cigarette smoke induced similar
changes in these cells.
Conclusion: These preliminary findings support more
comprehensive array-based analysis of this in-vitro
model system to further delineate epigenomic mechanisms by which hookah smoke may contribute to
initiation and progression of human lung cancers.
60. Imagining and prediction rules for TB

diagnosis
D Schrump,1 Y Xiong,1 S Xi,1 S Azoury,1 M Zhang,1

J Shan,2 J Hong1 1National Cancer Institute, Bethesda, MD,
2
National Cancer Institute, Frederick, MD USA. e-mail:
David_Schrump@nih.gov
PC-1206-06 An improved hardware neutral

computer aided detection platform for
tuberculosis interpretation from chest X-rays
with advanced false-positive reduction
Background: Although perceived by many to be a safe

alternative to cigarettes, hookah smoke contains a
number of carcinogens present in cigarette smoke. To
date, epidemiologic associations between hookah tobacco exposure and lung cancer risk have not been firmly
established. In the present study, a well-established invitro model system was used to directly compare the
epigenomic effects of hookah smoke relative to cigarette
smoke in normal human respiratory epithelia and lung
cancer cells.
Design/Methods: Waterpipe condensates (WPC) and
cigarette smoke condensates (CSC) were prepared from
commercial Al Fakher tobacco and 3R4F cigarettes (UKentucky Research Labs), respectively, using Borgwaldt
precision smoking machines. Normal human small
airway epithelial cells (SAEC), and cdk-4/hTERTimmortalized human bronchial epithelial cells (HBEC),
as well as Calu-6 (p53 -/-) and A549 ( p53 wt) NSCLC
lines were cultured in the presence or absence of varying
concentrations of WPC or CSC. MTT assays were used
to evaluate effects of condensates on cell proliferation;
qRT-PCR, and immunoblot techniques were used to
examine mRNA, microRNA and histone alterations
mediated by WPC relative to CSC.
Results: Five day WPC exposures mediated variable
effects in cultured respiratory epithelia and lung cancer
cells. In SAEC and HBEC, WPC exposure mediated dosedependent growth inhibition. At lower doses (0.1mg/ml
and 0.5mg/ml), WPC enhanced, whereas at higher
concentrations (1.0mg/ml and 2.0mg/ml), WPC inhibited
proliferation of Calu-6 cells. WPC had minimal effects
on growth of A549 cells except at high concentrations
(2.0mg/ml), which inhibited proliferation of these cells.
M Acharyya1 1Advenio TecnoSys, Chandigarh, India. e-mail:

mausumi.acharyya@adveniotecnosys.com
Background: An automated CADx tool for consistent

high performance interpretation of pulmonary tuberculosis (PTB) in dCXRs with a focus on reducing falsepositives (FPs) is proposed.
Design/Methods: A two-layered coarse-to-fine cascade
framework under supervised learning is proposed. The
first step is candidate generation module on a training
data set. A true candidate is a localized region within the
lung which exhibits TB patterns. Candidate generation
step involves extracting several descriptive features
within annotated localized patches corresponding to
normal and TB infected and subsequently training the
first level classifier (random tree) with these features. At
this stage we achieve high sensitivity but also with a high
false positive level. The candidates classified as TB
infected are positive examples are then described by
regions of interest (ROI) and are used as input for the
second level classifier. In the second level each candidate
is resampled via scaling, translations, and rotations to
increase the variation of the training data and to avoid
overfitting. These multiple views of candidates are used
to train a deep Convolutional Neural Network (CNN)
classifier. The CNN provides probability for each ROI
view which are averaged and a single probability
measure is obtained for each ROI or candidate. For an
unknown test dCXR image, candidates are similarly
generated and the CNN classifier is applied to produce
the probability measures for each ROI averaging which
gives the single probability of each candidate and
classifies it to be normal or abnormal based on a
threshold value of the probability measure. This second
level classifier behaves as a highly discriminating process

to discard challenging false positives while still achieving
high sensitivity. The training dataset is representative of
clinical and community settings. Around 1000 digital Xray of chest dCXR (posterior-anterior view) would be
obtained in partnership with The National Institute for
Research in Tuberculosis (WHO Supranational Reference Laboratory), Chennai & Asian Institute of Public
Health (grantee StopTB REACH programs in East
India), Bhubaneswar. The data represents the spectrum
of PTB radiological markers with confirmatory clinical
annotations (microbiology culture as gold standard).
Results: ROC and AUC are calculated. 90% sensitivity
with 80% specificity is acheived
Conclusion: Proposed tool consistently interprets dCXRs
with high sensitivity and specificity.
PC-1207-06 Imaging characterization of

cavitatory lesion of lungs and
histopathological correlation: a hospital-based
prospective observational study
S Dahal1 1National Academy of Medical Sciences,
Kathmandu, Nepal. e-mail: shishir15@mail.ru
Background: Pulmonary tuberculosis is an important

cause of lung cavity along with bronchogenic carcinoma.
Imaging characterization of cavitatory lesion in chest
radiograph and/or in computated tomography has great
importance in identifying benign versus malignant lesion
where suspicious malignant lesion requires further
invasive investigation. Differentiating benign cavity from
malignant one is crucial in further investigations and
management.
Design/Methods: This cross sectional, hospital based,
observational study was conducted at National Academy
of MedicalSciences .A total of 54 patients with cavitatory
lung lesion were evaluated in contrast enhanced computed tomography and computed tomography guided
biopsy were taken from the cavitatory wall for histopathological diagnosis. In computed tomography pulmonary cavities were characterized by size, wall thickness, and smoothness of inner wall, content of cavity and
enhancement of cavitatory wall. Each of these parameters are correlated with the Fine Needle Aspiration
Cytology findings of malignant and inflammatory
(tubercular and non tubercular) cavity. Univariate
analysis is presented as Spearman correlation coefficient.
Results: Malignant cavity showed the strongest correlation with larger cavity with mean diameter more than 20
mm (r0.6) and irregularity of inner cavitatory wall
(r0.7).While significance enhancement of cavitatory
wall has weakly correlated with malignancy (r0.35).
Correlation between malignant cavity and content is of
no statistical significance. All non malignant cavities are
tubercular in histopathology. Fluid content in the cavity
has stronger negative correlation with non malignant
cavity (r2).There is no statistical significant correlation
between non malignant cavities rest of other characteristics of cavity.
S499
Conclusion: Irregularity of inner wall of pulmonary

cavity and larger size cavities correlates best with
malignancy while fluid content in the cavities show
strong negative correlation with non malignant cavities.
PC-1208-06 Objective computerized chest

radiography screening to detect tuberculosis in
the Philippines
R Philipsen,1 C Sanchez,1 P Maduskar,1 J Melendez,1
B Van Ginneken,1 W J Lew2 1Radboud University Medical
Center, Nijmegen, Netherlands; 2World Health Organization,
Philippine Country Office, Manila, Philippines. e-mail:
rphilipsen@gmail.com
Background: In the Palawan provincial areas of the

Philippines, screening for pulmonary tuberculosis (TB) is
performed in a WHO project by symptom screening
combined with chest radiography; if at least one of these
is positive, subjects undergo molecular Xpert testing.
However, interpreting chest radiographs (CXRs) requires
skilled personnel whose availability is limited and
training is expensive and logistically challenging. Computerized reading is a potential solution: CXRs are
objectively interpreted by software for the presence of TB
related abnormalities. Operational costs are minimal and
a result is available within one minute. We retrospectively engaged computerized reading in the DetecTB
(Diagnostic Enhanced Tools for Extra Cases of TB)
project, an intensified case finding program targeting
prison settings and high-risk communities on Palawan
Island in the Philippines.
Methods: A retrospective study was carried out on all
cases in the DetecTB program available mid 2014. We
compared the performance of CAD4TB v4.10, a
commercially available software package, against the
performance of the physicians readings. CAD4TB scores
a posterior-anterior CXR for TB related abnormalities on
a scale from 0 (normal) to 100 (abnormal). We
performed Receiver Operator Characteristic (ROC)
curve analysis to compare the computerized and human
reading and visually inspected CXRs marked as normal
by the physician but as highly suspicious by the software.
Results: In total, 12 256 radiographs were available for
analysis. The number of cases positively read by the
physicians was 3068 (25.0%). 379 subjects (3.1%) tested
positive on Xpert. Defining these cases as positive, and
the remaining cases as negative, the area under the ROC
curve (Az) achieved by CAD4TB v4.10 was 0.911 [CI:
0.895-0.926]. The software achieved 90.0% sensitivity
at 80.0% specificity. The physician had a sensitivity of
97.6% and a specificity of 77.3%, which is slightly better
than the software. However, this reading is biased as
cases that were considered normal by the physician were
considered negative by definition in our study design. An
independent human observer considered 78 out of the
100 cases read as normal by the physician with the
highest CAD4TB scores to be abnormal.
Conclusion: Computerized reading provides high sensitivity and specificity and may be used to assist or possibly
S500
replace a human reader in active case finding programs,

and thus improve screening throughput.
PC-1209-06 Computerized reading of chest

radiographs in The Gambia National
Tuberculosis Prevalence Survey: retrospective
comparison with human experts
CXR: Normal, Active TB, Abnormal, healed TB,

or Other pathology. The software (CAD4TB, Radboud
University Medical Center, Nijmegen, The Netherlands)
computed a TB score between 0-100 (0-Normal, 100Abnormal) for each CXR. The area under receiver
operating characteristic curve (AUC) was calculated for
CAD4TB and various cut-off points were chosen on the
TB score to compare specificities at the sensitivities of the
field and central readings.
Results: The field reading achieved a sensitivity of 86.3%
at 72.8% specificity. The central reading had a sensitivity
of 67.1% at 93.3% specificity, and when the healed TB
category was considered abnormal, sensitivity increased
to 87.7% with a decrease in specificity to 65.2%.
CAD4TB attained an AUC of 0.90 and at all three
sensitivity levels of human readers, specificity was nearly
identical and not significantly different (P . 0.05): at
field sensitivity, CAD4TB had a specificity of 72.2%, and
for central sensitivity levels, specificities of 93.7% and
66.0% (healed TB as abnormal) were obtained.
Conclusions: When selecting subjects in a prevalence
survey on the basis of chest radiography to undergo
confirmatory testing for TB, the performance of computerized reading is not significantly different from field
and central readings by human experts. The software has
potential to improve the efficiency of TB prevalence
surveys.
P Maduskar,1 I Adetifa,2,3 E Leroy Terquem,4

J Van Den Hombergh,5 A Fasan-Odunsi,6 C Sanchez,1
U Dalessandro,2 B Van Ginneken1 1Radboud University
Medical Center, Nijmegen, Netherlands; 2Medical Research
Council Unit, Fajara, Gambia; 3London School of Hygiene
&Tropical Medicine, London, UK; 4Centre Hospitalier de
Meulan les Mureaux, Meulan-en-yvelines, France;
5
PharmAccess International, Dar Es Salaam, Tanzania; 6Lagos
State University Teaching Hospital, Ikeja, Lagos, Nigeria.
e-mail: pragnya.m@gmail.com
Rationale: Tuberculosis (TB) prevalence surveys require

recruitment of trained personnel for reading chest
radiographs (CXRs), which is scarce in high TB burden
countries. Computerized CXR reading could replace
human readers in such a scenario and we retrospectively
investigate this possibility on data from the 2011-2013
TB prevalence survey in The Gambia.
Methods: Computerized readings were compared with
field and central readings on 4552 CXRs. The survey
participants were screened based on symptoms and CXR
findings. Field readers judged CXRs on radiological
findings as: Normal, Abnormal, suggestive of TB or
Other abnormalities; the latter two were considered
abnormal. In case of symptoms and/or abnormal CXR,
sputum samples were collected and bacteriological tests
(fluorescence microscopy and BACTEC MGITe culture)
were performed for confirmatory TB diagnosis. Following the analysis, 73 subjects with proven TB were
considered Abnormal and the remainder as Normal.
The CXRs were centrally audited by an expert reader
and one of the following categories was assigned to each
PC-1210-06 Development of a clinical

prediction rule for the diagnosis of pleural
tuberculosis in Peru
A Soto,1,2 L Solari,1 J Agapito,3 P Van Der Stuyft1,4
1
Institute of Tropical Medicine of Antwerp, Antwerp,
Belgium; 2Hospital Nacional Hipolito Unanue, Lima,
3
Universidad Peruana Cayetano Heredia, Lima, Peru; 4Ghent
University, Ghent, Belgium. e-mail: lelysol@hotmail.com
Background: Pleural tuberculosis is one of the most

common forms of extra-pulmonary tuberculosis and is
still a diagnostic challenge. Some Clinical Prediction
Rules (CPRs) are available for pleural tuberculosis, but
their external validity is uncertain. The aim of this study
was to develop a CPR for diagnosis of pleural tuberculosis in Peru.
Design/Methods: Adult patients with pleural exudates of

unknown origin attending 2 reference hospitals in Lima,
Peru, from 2009 to 2011 were evaluated through basic
laboratory tests on pleural fluid and microbiologic tests
(smears for Acid-fast bacilli (AFB), cultures and Real
Time PCR). Closed pleural biopsy was performed for
diagnosis of tuberculosis, which was defined as a positive
smear for AFB, a positive culture or positive PCR in
pleural fluid or biopsy OR presence of caseating
granulomas in histopathology OR suggestive histopathological results with resolution of clinical symptoms at
the third month of treatment initiation. Logistic regression defined the parameters significantly associated with
pleural tuberculosis and a score was developed on this
basis. The required sample size was 259.
Table Results of the logistic regression model for pleural
tuberculosis and number of points assigned to each predictive
finding in the developed Clinical Prediction Rule.
Predictive findings
Clinical predictors
Gender (male)
Contact with a
tuberculosis case
Disease duration in
weeks
Haemoptysis
Lymphadenopathy
Age
Pleural fluid analysis
Proteins in gr/dL
Adenosine deaminase
(ADA) in U/L
Odds Ratio
(95% confidence
interval)
p value
2.85 (1.23-6.56)
3.03 (1.26-7.24)
0.01
0.01
0.93 (0.87-1.00)
0.05
Number
of points
assigned
11
14
-(duration of
disease/2)
0.15 (0.04-0.66)
0.01 -15
0.21 (0.08-0.57) ,0.001 -17
0.97 (0.94-0.99) ,0.001 -(Age in
years/2)
1.75 (1.18-2.62) ,0.001 8*(proteins )
1.04 (1.03-1.06) ,0.001 (ADA)/2
Results: We included 268 patients with complete data

available. 181 (67%) had pleural tuberculosis according
to our criteria. The parameters associated to tuberculosis
were age, gender, contact with a case of tuberculosis,
duration of disease, haemoptysis, lymphadenopathy,
pleural fluid proteins and ADA. Night sweats, fever,
blood cell count, lymphocyte predominance in pleural
fluid and pleural LDH were not associated to tuberculosis. According to the results of the logistic regression
model, a CPR was proposed. Table 1 shows these results
and the CPR. Its area under the ROC curve was 0.94. At
a cutoff of 38 points, the CPR attained a sensitivity of
90%, a specificity of 91%, a positive likelihood ratio
(PLR) of 9.72 and a negative likelihood ratio (NLR) of
0.11. If 2 cutoffs instead of one were used at 30 and 50
points, where patients with ,30 points were taken as
low probability and patients with .50 as high probability of pleural tuberculosis, the PLR would be .10 and
the NLR ,0.10 and decisions could be taken on 92% of
patients.
Conclusion: The developed CPR can be used as a tool for
diagnostic evaluation of patients with clinical suspicion
of pleural tuberculosis. Its diagnostic performance in this
population was high. Prospective validation is needed
before widespread implementation in Peruvian hospitals.
S501
PC-1211-06 A systematic review of prediction

models for estimating the probability of
pulmonary tuberculosis in adults
S Van Wyk,1 H-H Lin,2 M Claassens1 1Desmond Tutu TB
Centre, Department of Paediatrics and Child Health,
Stellenbosch University, Cape Town, South Africa; 2Institute
of Epidemiology and Preventive Medicine, National Taiwan
University, Taipei, Taiwan. e-mail: susansvanwyk@gmail.com
Background: Current symptom screening for tuberculosis (TB) has sub-optimal sensitivity and TB cases are
missed. A prediction model combining socio-demographic characteristics and medical history to estimate
an adults absolute probability of having active pulmonary TB (PTB) could be useful in identifying presumptive
TB cases at an early stage, leading to early diagnosis and
treatment. A systematic review was conducted to
describe the characteristics of prediction models for
estimating the probability of having PTB in adults.
Design/Methods: A prediction model was defined as the
combination of two or more clinical predictors designed
to estimate the probability of having TB. Studies using
culture confirmed PTB as reference standard were
included. Models for inpatients, children or specific
patient populations (e.g. dialysis patients) were excluded.
PubMed, Scopus and the Cochrane Library and abstracts
from the Union, American Thoracic Society and European Respiratory Society conferences were searched.
Reference lists from identified studies were checked. The
TRIPOD statement was used to assess completeness of
reports and model classification. The CHARMS checklist guided risk of bias assessment.
Results: From 13 671 identified records, 10 were
included for data extraction; 9 were from high burden
countries; 5 assessed only smear negative PTB as
outcome and 7 focused on human immunodeficiency
virus (HIV) infected individuals only. Full appraisal of
models was difficult due to lack of clarity and incomplete
reports. Eight studies were type 1a prediction model
studies where predictive performance was evaluated
using exactly the same data set as was used for
development (apparent performance); one study was a
type 1b model where internal validation was done using
bootstrapping techniques and one was a type 4 study,
where predictive performance of a published prediction
model in inpatients was assessed on outpatient data. The
type 4 model included hemoptysis, age .45, weight loss,
expectoration and apical infiltrates on chest radiography
as predictors and might be useful for ruling out PTB in
smear negative presumptive TB cases without previous
history of TB in hospital outpatient settings (score,0:
NPV 94.9% [95%CI: 86-99]). Results could not be
pooled due to heterogeneity.
Conclusion: Existing prediction models for estimating
the probability of having PTB in adults at primary care
level are poorly reported and validated and therefore not
useful for TB screening. WHO symptom screening is still
recommended.
S502
61. Alternative diagnostics for drug

resistance
PC-1214-06 Multicenter evaluation of
high resolution melting curve for
detection of extensively drug-resistant
Y Pang,1 X Hui,1 H Dong,2 Z Zhang,2 S Huan,3 J
Zhang,2
D Chin,3 Y Zhao1
1
Chinese Center for Disease Control and Prevention, Beijing,

PATH, Beijing, 3Bill and Melinda Gates Foundation, Beijing,
China. e-mail: pangyu@chinatb.org
2
Drug-resistant tuberculosis (TB), especially multidrugresistant TB (MDR-TB) and extensively drug-resistant

TB (XDR-TB), is still one of the most serious threats to
TB control worldwide. Early diagnosis of drug-resistant
is essential for generating the effective chemotherapy
regimen. High resolution melting curve (HRM) is a rapid
method for detecting XDR-TB. Several preliminary data
have demonstrated that HRM shows excellent performance to detect the rifampcin- and isonizid-resistant TB,
while it has not been effectively evaluated among large
clinical samples. In this study, we evaluated the
performance of HRM for XDR-TB in 772 TB smearpositive patients in two TB laboratory of China. When
setting the MGIT 960 drug susceptibility testing (DST) as
golden standard, our results revealed that the sensitivity
and specificity of HRM were 90.1% and 96.3% for
rifampin (RIF) resistance, and 81.5% and 98.3% for
isoniazid (INH) resistance, respectively. For second-line
anti-TB drugs, the sensitivity and specificity of HRM
were 76.3% and 98.3% for ofloxcin (OFLX) resistance,
and 50.0% and 98.4% for kanamycin (KAN) resistance,
respectively. In conclusion, our research demonstrates
that HRM can be used as an alternative for detecting RIF,
INH and OFLX resistance. The low cost of HRM makes
it available to be broadened in limited-resource laboratories in China. The unsatisfactory performance of HRM
to detect KAN resistance highlight the urgent need for
explore the potential KAN resistant mechanism of
Mycobacterium tuberculosis.
PC-1215-06 Revisiting the microbial diagnosis

of drug-resistant tuberculosis
E Cambau1 1Assistance Publique Hotpitaux

de Paris, Paris,
France. Fax: (33) 1 4995 8537. e-mail:
emmanuelle.cambau@lrb.aphp.fr
Background: Detection of drug-resistance and reliable

detection of susceptibility are both required in order to
design an efficient individualized treatment regimen for
multidrug resistant tuberculosis. While genotypic detection of resistance can be done for some drugs only,
phenotypic drug susceptibility testing is difficult to
standardize for second line antituberculous drugs.
Design/Methods: We aim to combine phenotypic drug

susceptibility testing and genotypic detection of resistance to first line and second line drugs for the microbial
diagnosis of drug-resistant Mycobacterium tuberculosis
strains. We performed first a new protocol of DST, using
the TBeXiST-Epicenter software in MGIT 960. This
protocol includes a first phase of resistance screening
with one critical concentration. In case of resistance,
TBeXiST DST has a second phase of quantitative
determination of resistance by testing growth at increasing drug concentrations, complemented by genetic
assessment of resistant mutations.
Results: Combining phenotypic and genotypic assessment of resistance provided highly reliable results for
resistance. Screening at critical concentration closed to
epidemiological cut-off (ECOFF) provided highly reliable results for susceptibility. In case of MDR or XDRTB, the determination of the resistance level helps in
setting a 4 to 5 drugs regimen affordable for these cases.
The final results are obtained rapidly (median 10 days)
since the growth is observed in a real-time manner and
not endpoint of 28-42 days. Resistance can be observed
even shorter, after 4-5 days. Finally, heterogeneous
resistance is easily detected since 10-fold diluted growth
controls are run in parallel
Conclusion: Microbial diagnosis of drug resistance in TB
was revisited by combining a new phenotypic DST
protocol and extensive genotypic characterization providing a reliable and rapid protocol for assessment of
susceptibility and resistance.
PC-1216-06 Mutations associated with secondline tuberculosis drug resistance in Georgia

N Bablishvili,1 N Tukvadze,1 E Shashkina,2 B Mathema,3
Z Avaliani,1 N Gandhi,4 H Blumberg,4 R Kempker4
1
Georgia; 2Public Health Research Institute TB Center in
Newark, Newark, NJ, 3Department of Epidemiology,
Columbia University, New York, NY, 4Division of Infectious
Diseases Department of Medicine, Emory University School
of Medicine, Atlanta, GA, USA. e-mail:
nino_bablishvili@yahoo.com
Background: Rapid detection of multidrug and extensively drug-resistant tuberculosis (M/XDR-TB) is an

essential component of TB-control. In prior work, we
found the MTBDRsl assay, a commercially available
molecular test, to have suboptimal performance in the
detection of fluoroquinolone and injectable drug resistance. To better understand the genetic determinants of
resistance we conducted targeted genetic sequencing of
Mycobacterium tuberculosis (MTB).
Design/Methods: We performed a cross sectional study
at the National Reference Laboratory (NRL) of the
National Center for Tuberculosis and Lung Diseases
(NCTLD) in Tbilisi, Georgia and the Public Health
Research Institute (PHRI) TB Center in Newark, New
Jersey. MTB isolates were stored samples obtained from
a previous prospective study evaluating the performance
of the MTBDRsl assay. DNA extraction was performed
using the QIAamp DNA-kit for all positive subcultures
of MTB. Targeted DNA sequencing was performed for

gyrA, gyrB, rrs and eisgenesand compared to phenotypic
drug susceptibility results.
Results: Among 112 MDR-TB isolates examined by
phenotypic testing, 16(14%) were resistant to ofloxacin
(OFX), 17(15%) to capreomycin (CAP), and 63 (56%)
to kanamycin (KAN). Sequencing was successfully
performed for each gene as follows: 111(99%) gyrA,
107(96%) gyrB, 104(93%) rrs, and 103(92%) eis. DNA
sequencing identified mutations in gyrA and gyrB genes
in 9(56%) of 16 and 5(31%) of 16 MTB isolates
phenotypically resistant to OFX, respectively. The
addition of gyrB mutations increased the sensitivity of
OFX resistance detection by 13%. No gyrA/B mutations
were found in 95 susceptible isolates. The A1401G rrs
mutation was detected among 7 of 15(47%) and 7(12%)
of 60 available MTB isolates resistant to CAP and KAN,
respectively. No rrs mutations were found in susceptible
isolates. Any eismutations were found in 14(24%) of 59
and 7(14%) of 49 available isolates resistant and
susceptible to KAN, respectively. In contrast, the C14T mutation was found in 6(10%) of 59 KAN resistant
isolates and in none of the KAM susceptible isolates. The
addition of the C-14T mutations increased the sensitivity
of KAN resistance detection by 10%.
Conclusion: The inclusion of C-14T eis and gyrB gene
mutations would improve the sensitivity of rapid
detection assays for KAN and fluoroquinolone resistance, respectively. In many MTB isolates, CAP and
KAN resistance is not associated with known drug
resistance mutations in rrs and eis genes.
PC-1217-06 Changes in mutations: a closer look

at GenoType MTBDRplus assay results for the
molecular diagnosis of drug-resistant M.
tuberculosis in Lima, Peru
R Calderon,1 C Pantoja,1 K Tintaya,1 L Lecca,1 J Coit,2
D Coleman,3 G R Davies,4 C Mitnick2 1Socios En Salud
Sucursal Peru, Lima, Peru; 2Harvard Medical School, Boston,
MA, USA; 3Saint Georges University of London, London,
4
University of Liverpool, Liverpool, UK. Fax: (51) 1 547 2121.
e-mail: rcalderon_ses@pih.org
Background: Multi-drug resistant tuberculosis (MDR)

seriously threatens TB control and prevention programs.
Effective control requires rapid diagnosis to facilitate
prompt implementation of adequate therapy. The GenoType MTBDR assay is an effective diagnostic tool for
MDR but it also provides key information on gene
mutations involved in drug resistance. Here, we report
resistance prevalence and mutation frequency in the
context of a clinical trial in Lima, Peru
Design/Methods: Drug susceptibility testing (DST) for
first-line anti-TB drugs was performed on sputum
samples from patients newly diagnosed with TB and
who consented to participate in a trial between 20132015 in Lima. DST for isoniazid (INH) and rifampin
(RIF) was performed using BD BACTEC MGIT 960
system and GenoType MTBDR assay at the Socios En
Salud Laboratory (SES-Lab).
S503
Results: Among the 351 patiens who consented, 328

were tested with the GenoType. Of those, 60 (20.1%)
were drug-resistant; 28 (8.5%) were INH monoresistant,
32 (9.8%) were MDR and 5 were RIF monoresistant.
GenoType and DST results are currently available for
197 patients; of these, 5 of were MDR on both methods
(100% concordance) and mutations involved were rpoB
S531L & D516V, katG S315T and inhA C15T. In all
resistant specimens rpoB S531L, codon 533 & D516V,
katG S315T and inhA C15T & T8C were involved. . In
the group resistant detected by GenoType, the frequency
of mutations for rpoB D516V: 11 (28.9%); rpoB 526
codon: 0 (0%) and rpoB S531L: 27 (71.1%); for katG
S315T: 36 (59.0%) and; for inhA C15T: 14 (23.0%).
Moreover, mixed infectious were detected in 15 (22.4%)
of the patients. No discordant results between the two
methods were found for RIF susceptibility, but discordance with INH was detected in 5; 3 without mutations
and 1 with inhA T8C and 1 with inhA C15T
Conclusion: In Peru, TB control requires rapid DSTs for
RIF as well as INH due to the high prevalence of INH
monoresistance. This work also shows a possible change
in the frequency of mutations. Therefore, given the high
prevalence of MDR found in this group of newly
diagnosed TB patients, continuous monitoring and deep
evaluation on existing molecular DST will become
important in the coming years in order to retain accuracy
in results reporting. Genomic analysis may be necessary
in order to further dissect these findings. Lastly, the
application of the Genotype MTBDR assay is important
and shows excellent concordance with standard drug
susceptibility testing done in Peru.
PC-1218-06 Predictive value of the new

Genotype MTBDRsl V2 for the baseline
detection of drug resistance mutations among
multidrug resistant tuberculosis patients
A W Dreyer,1,2 S Omar,1 N Ismail1,3 1National Institute for
Communicable Diseases, Johannesburg, 2University of
Witwatersrand, Johannesburg, 3University of Pretoria,
Pretoria, South Africa. e-mail: aw.dreyer@gmail.com
Background: The management of multidrug resistant

tuberculosis (MDR-TB) patients in most cases consists of
empiric second line drug regimens with subsequent long
turnaround times for phenotypic results and potentially
suboptimal therapy and associated drug resistance with
poor outcomes. Molecular diagnostics assays have
decreased turnaround times for the rapid detection of
both Mycobacterium tuberculosis (MTB) and associated
drug resistance significantly. The Gentype MTBDRsl
V2(Hain Lifesciences, Germany) is a commercially
available assay for the detection of secondline drug
resistance and utilize hybridization technology to assess
both intact wildtype and associated mutations for
fluoroquinolones(FQ) and second line injectables(SLI).
The improved version includes detection associated with
mutations in both gyrA, gyrB, rrs and eis.
Design/Methods: Baseline sputum samples were collected from MDR-TB patients initiating treatment. MTB
S504
was isolated using liquid culture (Bactec MGIT960

system) and sequencing performed for secondline drug
resistance mutations targeting gyrA, gyrB (FQ) and rrs
(SLI) regions. The frequency of the resistance mutations
were determined and expressed as a percentage. The
performance of the assay to predict secondline resistance
was assessed.
Results: Of the 254 MTB isolates available for sequencing a total of 49(19.29%) showed mutations in the gyrA
region and 11(4.33%) mutations in the gyrB region. For
the fluoroquinolones A90V (42.86%) and D49G
(46.94%) were the most common mutations detected
within gyrA. Among the 11 mutations within gyrB and
K537Q (36.36) and L566F (18.18) were more frequent.
Mutations within the rrs region was detected among
65(25.6%) of isolates and showed great variability with
11 different mutations detected of which A514C
(17.65%) and C517T (23.53%) were the most frequent.
The predictive value of the Genotype MTBDRsl assay
was calculated at 97.96% for gyrA when considering
probing for both absence of wildtype and mutations and
45.45% of gyrB mutations. The mutation missed in gyrA
was a double mutation, L105R_G112A, with corresponding phenotypic resistance. Only 14.71% of the rrs
mutataions with would have been detected with the
assay.
Conclusion: The Genotype MTBDRsl assay V2 is a
useful tool for the rapid baseline assessment for second
line drug resistance pending phenotypic drug susceptibility results and could potentially improve patient
outcome by early initiation of appropriate therapy.
using the H37Rv strain. Mutant discrimination was

assessed using five clinical isolates. Mutant genotypes
were determined by Sanger sequencing and drugresistant phenotypes were determined by BACTEC
MGIT 960 culture. Analytical sensitivity for the identification of a RIF mono-resist strain was evaluated using a
clinical isolate with a His526Tyr substitution in rpoB,
while analytical sensitivity for the identification of an
INH mono-resistant strain was evaluated using a clinical
isolate with a Ser315Thr substitution in katG. Analytical
sensitivity for detection of multi-drug resistant strains
was evaluated using three isolates with mutations in both
rpoB and katG genes; namely, Ser531Leu/Ser315Thr,
His526Asp/Ser315Thr and Gln513Lys/Ser315 rpoB/
katG mutants.
Results: Analytical inclusivity, specificity, and crossreactivity were verified for five M. tuberculosis complex
organisms and five nontuberculous mycobacteria
(NTMs). Analytical sensitivity of the assay was found
to be less than 50 cfu/mL-sputum for M. tuberculosis
detection and less than 250 cfu/mL-sputum for rifampicin and isoniazid resistance detection. The assay successfully detected all five M. tuberculosis complex organisms
tested. There was no detectable cross-reactivity with any
of the five NTMs. The BD MAXTM MDR-TB assay is
currently under development and not available for sale or
use.
PC-1219-06 Preliminary analytical performance

of the BD MAX MDR-TB assay for the detection
of M. tuberculosis complex and rifampicin and
isoniazid resistance
PC-1221-06 Discordant rifampicin resistance

results on MGIT960 and XpertW MTB/RIF in the
TB Control in Prisons Programme of Azerbaijan
1 1
I Ashman, M Porter Becton Dickinson, Sparks, MD, USA.

e-mail: iramarc_ashman@bd.com
Background: Tuberculosis is a deadly infectious disease

and continues to remain one of the worlds top health
challenges. Rapid and accurate detection of Mycobacterium tuberculosis and drug resistance is of paramount
importance to appropriately identify and treat patients
suffering with the disease, reduce the death rate and stop
the spread of TB. The BD MAXTM MDR-TB assay is an
automated, qualitative test that allows for the direct
detection of M. tuberculosis complex and identification
of genetic mutations correlated with rifampicin (RIF)
and isoniazid (INH) resistance from raw (unprocessed)
and NALC/NaOH-treated (processed) clinical sputum
samples. The assay is designed for use on the BD
MAXTM System, which offers true walk-away automation by incorporating sample lysis, extraction, amplification and detection all in one system.
Design/Methods: Analytical sensitivity of the assay for
both M. tuberculosis complex detection and mutant
discrimination was determined using spiked negative
sputum. BD MAXTM MDR-TB analytical sensitivity for
the detection of M. tuberculosis complex was evaluated
62. Insights into drug-resistant

tuberculosis from Ghana to Peru
E Gurbanova,1 K Blondal,2 F Mirzayev,3 R Tahirli,4

V Popova,4 R Mekhdiyev,1 A Ismayilov,5 A Altraja6 1Main
Medical Department of the Ministry of Justice, Baku,
Azerbaijan; 2Department of Communicable Disease
Prevention and Control, Reykjavik Health Care Services,
Reykjavik, Iceland; 3Laboratories, Diagnostics & Drug
Resistance Unit, World Health Organization, Geneva,
Switzerland; 4Reference Laboratory, Specialized Treatment
Institution for Detainees with Tuberculosis, Baku, Azerbaijan;
5
Project HOPE, Dushanbe, Tajikistan; 6University of Tartu,
Tartu, Estonia. e-mail: e.gurbanova@prisonhealth.az
Background: The Main Medical Department of Azerbaijan Ministry of Justice implements a comprehensive TB
Control programme in the penitentiary system (PS). This
programme practices active tuberculosis (TB) case
finding among inmates and provides effective treatment
for all identified cases of both susceptible and drugresistant TB in the Special Treatment Institution (STI).
The STI also houses a modern reference laboratory that is
quality-assured by the Supra-national Reference Laboratory in Borstel, Germany and performs all WHOrecommended diagnostics including Xpert MTB/RIF and
culture with following drug-susceptibility testing (DST)
on MGIT960. As soon as susceptibility result on R is
available from either Xpert MTB/RIF or MGIT960, the
patient starts adequate treatment with regimen adjustment after full DST is available. Since 2010, frozen
cultures of all culture-positive samples processed in this
laboratory are maintained.
Objectives: To assess the proportion of discordant R
resistance results on Xpert MTB/RIF and MGIT960
among patients identified in Azerbaijan PS during 20112015.
Design and methods: All pulmonary TB cases undergoing
testing for TB in STI were included into this retrospective
study. TB cases under treatment monitoring were
excluded. Sputum samples from admitted patients were
investigated both with Xpert MTB/RIF and MGIT960
culture and DST.To increase the number of strains
evaluated, additional strains were revived from culture
archive.
Results: Overall 532 strains were included, 286 consecutive sputum samples from admitted patients and 246
frozen strains. Out of the 532 strains, 33 (6.2%) had
discordant results, where 21 (3.9%) were R-resistant
according to Xpert MTB/RIF, but R-sensitive according
to MGIT960 and 12 (2.3%) were R-sensitive on Xpert
MTB/RIF but R-resistant according to MGIT960 (Table
1). Both tests, however, were in a good agreement with
each other (Cohens kappa 0.825, P , 0.001). If MGIT
results were taken as reference, then sensitivity, specificity, positive predictive value (PPV), negative predictive
value (NPV) and test accuracy of Xpert MTB/RIF were
90.8%, 95.2%, 84.9%, 97.2% and 94.2% respectively.
Table 1. Rifampicin susceptibility results by Xpert MTB/
RIF and MGIT960. Xpert MTB/RIF Sensitive Resistant
Total MGIT 960 Sensitive 393 21 414 Resistant 12 106
118 Total 405 127 532
Conclusions: Despite the 6.2% discordance between the
methods, the results by Xpert MTB/RIF had acceptable
sensitivity, specificity, PPV, NPV and accuracy and were
in a good agreement with those obtained with MGIT960.
Further research and sequencing will be necessary to
better understand the genetic profile of strains that
exhibit discordance in Rif susceptibility results between
these two diagnostic techniques.97.2% and 94.2%
respectively.
PC-1222-06 Feasibility of multidrug-resistant

tuberculosis treatment with improved
regimens based on drug resistance profiles
M H Wu,1 R Jou1,2 1Centers for Disease Control, Taipei,
2
Background: Since multidrug-resistant tuberculosis

(MDR-TB) cases with additional resistance to a fluoroquinolone or a second-line injectable drugs might be
qualified for treatment with bedaquiline and cases with
no resistance to fluoroquinolones and second-line
injectable drugs could be treated with a short-course
regimen, we analyzed drug-resistant profiles of MDR
Mycobacterium tuberculosis to assess possible use of
improved regimens for MDR-TB treatment.
S505
Methods: This is a prospective laboratory-based study.

M. tuberculosis isolates were subsequently collected
from 196 MDR-TB cases during 2013-2014. Drug
susceptibility testing was performed using the Middlebrook 7H11 agar proportion method. In addition, the
resazurin microtiter assay was used to determine the
minimum inhibitory concentrations (MICs) of clofazimine and linezolid. The susceptibility breakpoints for
clofazimine and linezolid are suggested to be 1.0 lg/mL
and 1.0 lg/mL, respectively. Extensively drug-resistant
(XDR) TB was defined as MDR-TB plus resistance to
any fluoroquinolone and at least one second-line
injectable drugs.
Results: Of the 196 isolates, ratios of resistance were
34.2%(67/196) to pyrazinamide, 15.8%(31/196) to
ofloxacin, 16.3%(32/196) to moxifloxacin, 12.8%(25/
196) to levofloxacin, 5.6%(11/196) to gatifloxacin,
3.6%(7/196) to kanamycin, 2.6%(5/196) to amikacin,
1.5%(3/196) to capreomycin, 24.0%(47/196) to ethionamide, 8.7%(17/196) to para-aminosalicylate,
89.8%(176/196) to rifabutin, 0% to cycloserine. No
XDR-TB case was found. The range of MIC against
tested clinical strains was 0.06 to 4 lg/mL for clofazimine and 0.06 to 1 lg/mL for linezolid. Furthermore,
only 5.6% (11/196) of MDR-TB cases were resistant to
clofazimine (MIC. 1lg/mL) and no isolates resistance to
linezolid (MIC. 1lg/mL). We found that MDR M.
tuberculosis isolates resistant to clofazimine were not
associated with ofloxacin and moxifloxacin resistance
with P 0.1344 and 0.1525, respectively.
Conclusion: Of 196 MDR-TB cases, 157 (80.1%) cases
susceptible to fluoroquinolones and second-line injectable drugs might be qualified to be treated with a shortcourse regimen. Whereas, of 185 clofazimine susceptible
cases, 28 (15.1%) cases with resistance to any fluoroquinolone or 7 (3.8%) cases with resistance to any
second-line injectable drug might be qualified to be
treated with bedaquiline-containing regimen.
PC-1223-06 First- and second-line

Mycobacterium tuberculosis drug resistance at
a tertiary hospital chest clinic in Ghana
S Kudzawu,1 M Omari,1 A Forson,1 A Badoo,1 H Ganu,1
B C De Jong,2 J Otu,3 A Kwara,4 M Antonio3 1Korle-Bu
Teaching Hospital, Accra, Ghana; 2International Tropical
Institute, Antwerp, Belgium; 3Medical Research Council Unit,
Fajara, Gambia; 4Brown University, Providence, RI, USA.
e-mail: skudzay@yahoo.com
Background: Mycobacterium tuberculosis drug resistance is a major challenge to the use of standardized
regimens for tuberculosis (TB) therapy, especially among
previously treated patients. To inform drug selection for
previously treated TB patients at Korle-Bu Teaching
Hospital, we investigated the frequency and pattern
ofdrug resistance among predominantly previously
treated patients with smear-positive pulmonary tuberculosis.
Design/Methods: A cross-sectional survey was carried
out from 2010-2013 of patients with AFB smear positive
pulmonary TB who were previously treated or were
S506
smear positive at month 2 or 5 of first-line therapy. M.

tuberculosis cultured isolates were shipped to the TB
diagnostics Laboratory at The MRC Unit, The Gambia
for confirmatory testing and phenotypic Drug Susceptibility Testing (DST), for isoniazid, rifampin, streptomycin and ethambutol. Multi-drug resistant (MDR) M.
tuberculosis isolates were additionally subjected to
second-line DST including ofloxacin, capreomycin,
kanamycin and ethionamide.
Results: 170 M. tuberculosis isolates were submitted for
DST, of which 29 were not viable and 21 were considered
contaminants. Of the 120 evaluable samples, 107
(89.2%) were from previously treated TB cases and 13
from newly-diagnosed TB patients. Ninety patients
(75%) were males, and median age (IQR) was 37 (28.0
48.0) years. Overall 62 isolates (51.6%) were
susceptible to the first-line drugs tested, 28 (23.3%)
were MDR, 42 (35.0%) were resistant to streptomycin
and 22 (18.3%) were mono resistant. Of the MDR
isolates, 26 (92.8%) were from previously treated
patients, 20 (71.4%) were also resistant to streptomycin
and 6 (21.4%) were pre-XDR. Among the 13 newly
diagnosed cases, 2 (15.4%) had MDR-TB and 3 (23.1%)
had streptomycin resistance. Both cases of MDR-TB had
a history of contact with a previously treated case that
had MDR-TB.
Conclusion: There is an urgent need for routine DST and
the early use of individualized regimens in previously
treated TB patients in Ghana given the high prevalence of
MDR-TB and streptomycin resistance among these
patients. The high prevalence of pre-XDR among the
MDR-TB isolates requires that new cases of TB with
known MDR-TB exposure should be prioritized for
culture and DST for both first and second-line drugs.
This is necessary to optimize outcomes among patients
with high risk for MDR-TB, and minimize the risk of
transmission of MDR/XDR-TB to their contacts.
PC-1224-06 Stable anti-tuberculosis drug

resistance prevalence in Bamako, Mali, 20062014
B Diarra,1,2 M Sanogo,2 B Baya,2 A Togo,2 S Tounkara,2
S Dao,2 B De Jong,1 S Diallo2 1Biomedical Sciences, Institute
of Tropical Medicine, Antwerp, Belgium; 2SEREFO, University
of Sciences, Techniques and Technologies of Bamako,
Bamako, Mali. Fax: (223) 202 27 513. e-mail:
bdiarra@icermali.org
Background: Drug resistant tuberculosis (DR-TB) continues to threaten global tuberculosis control, resulting
from either primary infection with resistant bacteria or
from acquired resistance. Mali, a large country with
limited resources, reported 9,000 new TB cases, including 12 cases of MDR-TB in 2013. In Mali, individual
patient drug susceptibility testing (DST) is not universally accessible, however, at SEREFO, DST has been
conducted since 2006 for each patient enrolled in
research protocols. Given the absence of a nationwide
drug resistance survey, we present here the evolution of
TB drug resistance among new and retreatment cases
over the 9-year period.
Design/Methods: Between 2006 and 2014, we conducted a descriptive cross-sectional study by enrolling
pulmonary TB cases from local reference centers and
the University Teaching Hospital at Point G, in Bamako,
Mali. The study protocols were approved by the Ethics
Committee of the FMPOS-Mali and IRB of the NIAID/
NIH, USA. Confirmed Mycobacterium tuberculosis
complex (MTBc) isolates underwent indirect first line
DST using BD-MGIT AST/SIRE System, and in the
context of a regional collaboration, subcultures had DST
repeated for first line drugs, plus second line DST by agar
proportion method for MDR isolates, at the Medical
Research Council (MRC), the Gambia.
Results: A total of 1186 mycobacterial cultures were
performed on samples from 522 patients, including 1105
sputum and 81 blood samples. Of these patients, 120
(23%) had a negative culture, 43 (8.2%) grew NTM, and
343 (65.7%) grew MTBc. Phenotypic DST was performed on 337 (98.3%) of the MTBc isolates. Among the
337 patients, 127 (37.7%) had resistance to at least one
drug, including 75 (22.3%) multidrug resistant (MDR)
cases. The overall prevalence of MDR-TB was 3.4%
among new cases and 66.3% among retreatment cases.
The rates of resistance, both MDR and other resistances,
remained stable during the study period, with a small
increase observed in 2009, and 2010. Second line DST
was available for 35 (46.7%) of MDR cases, and no
XDR-TB was identified.
Conclusion: The drug resistance levels, including MDR,
found in this study are relatively high, likely related to the
selected referral population. While worrisome, the
numbers remained stable over the study period. As
culture and DST are only available in Bamako, the
nationwide surveillance of retreatment cases will provide
more accurate results on countrywide drug resistance
rates.
PC-1225-06 From facility to finding drug

resistance mutations: drug resistance survey in
Nigeria
V Sebastian,1,2 I Uko,2 G Akang,3 T Abiola,3 S Ogiri,4
Denn Onotu,2 N. Nwokoye,5 M Zignol6 1NTBLCP, Nigeria
National DR-TB Survey Consultant, Abuja, 2US-Centers for
Disease Control and Prevention (US-CDC) Nigeria, Abuja,
3
Federal Ministry of Health Nigeria (FMOH), Abuja, 4World
Health Organization Country Office Nigeria, Abuja, 5Nigeria
Institute of Medical Research (NIMR), Lagos, Nigeria; 6World
Health Organization, Geneva, Switzerland. e-mail:
vsebastian@cdc.gov
Background: Nigeria notifies about 100 000 sputum

smear positive (SS) tuberculosis (TB) cases annually and
is listed among the 22 high TB Burden (HTB) countries.
We determined the prevalence and pattern of resistance
to anti-tuberculosis drugs: rifampicin (RIF) and isoniazid
(INH) using Line Probe Assay (LPA) and compared with
culture drug sensitivity testing (DST).
Design/Methods: By cluster sampling we systematically
randomly sampled 30 clusters of DOTS facilities from a
national sampling frame of all DOTS facilities. Each
cluster was proportionate to size of new sputum smear
positive (SS) TB cases reported. In each survey cluster,

consenting SS TB patients (.15yrs old) were enrolled
consecutively as patients visit these facilities till a
required national total sample size of 1,723 respondents
new and re-treatment cases - were enrolled. Two most
positive sputum samples per respondent were transported under cold-chain in triple layered-packaging to incountry TB reference laboratories for LPA (using
Genotype MTBDR plus) testing and cultured (LJ slopes)
to obtain isolates. Isolates of all resistant and 10% of
sensitive samples were transferred under cold chain to
Supra-national Reference Laboratory at CDC Atlanta
and retested by LPA, solid and liquid culture DST to
determine (and compare) resistance to RIF and INH and
other anti-TB drugs.
Results: Of 1723 patients enrolled, 1449 had samples
successfully subjected to LPA, 70 were found to be
MDR-TB strains, giving a general prevalence of 4.8%
(95%CI: 3.8 6.1 %), comprising a prevalence of 2.9%
(weighted) among new TB cases (95%CI: 2.1 4.0%)
and 14.3% (95%CI: 10.219.3%) among retreatment
TB cases. We found resistance to other anti-TB drugs and
the strongest predictor of MDR-TB was previous anti-TB
treatment (P , 0.000). The results of LPA done directly
on smear-positive sputum concurred highly with culture
DST performed on isolates using MGIT 960 0 . Sensitivity, specificity, positive and negative predictive values of
Genotype MTBDR plus relative to conventional DST
using MGIT 960 and agar proportion method for
detection of resistance were 96.8, 92.1, 81.1 and 98.8
percent respectively for RIF and 73.8; 98.7; 96.9 and
87.5 percent respectively for INH. No XDR-TB strain
was found.
Conclusion: Line Probe assay (Genotype MTBDR plus
assay) was found to be feasible, rapid, sensitive and
specific test for routine DST for diagnosis of RIF
resistance. There is a notable prevalence MDR-TB in
Nigeria and potential for XDR-TB, history of prior antiTB treatment a very strong predictor of MDR-TB.
PC-1226-06 Scaling up of a commercial line

probe assay for diagnosis of drug resistance
and evaluation against an agar proportion
method in Peru
Z Puyen,1 J R Acosta Barriga,1 L Solari1 1Instituto Nacional
de Salud, Lima, Peru. e-mail: zpuyeng@gmail.com
Background: In Peru, a commercial line probe assay

(LPA), Genotype MTBDRplusw, was introduced by the
National Tuberculosis Control Program (NTCP) to
detect resistance to Isoniazid (H) and/or Rifampin (R).
Sputum samples or cultured strains in Lowenstein-Jensen
media, (L-J) from patients entering treatment were sent
to the National Institute of Health (NIH) to be tested.
Genotype MTBDRplus is used as a screening test: if
patients are found to have resistance to H or R, it must be
checked using the agar proportion (AP) method, which
includes other first and second line drugs. These results
define the final regimen to be given. The aim of this study
is to describe the process of scaling up of the LPA test and
S507
to compare its results with those obtained through AP for

evaluation of resistance to H and R.
Design/Methods: This is a descriptive cross-sectional
study. All results from sputum samples/ cultured strains
of patients with smear-positive pulmonary tuberculosis
in Peru that were sent to the NIH for LPA during 20112014 were included. The number of tests done was
compared with the number of cases according to the
NTCP for each year to calculate the coverage. Comparison between LPA and APP for resistance to H and R was
calculated using the positive predictive value, the
agreement and Cohens j coefficient for resistance to
H, R, and MDR.
Results: From 2011 to 2014, the total number of
Genotype MTBDRplus tests performed was 1075,
5472, 9621, and 11 954. This represents a coverage of
6% 31%, 57% and 71% respectively. 15%-43% of these
patients had previous episodes of tuberculosis. The
Figure shows the proportion of patients showing
resistance to H and R in each of the studied groups.
The positive predictive values of Genotype MTBDRplus
were 98%, 92% and 94% for resistance to H, R and
MDR respectively. The agreement and j coefficients for
resistance to isoniazid were 94.9% and 0.85, for
resistance to rifampin were 91.9% and 0.84 and for
MDR were 91.9% and 0.84.
Conclusion: The LPA Genotype MTBDRplus is a
suitable screening test for resistance to H, R and MDR,
allowing early diagnosis of resistance. Since 2011, it has
been scaled up in Peru by 15-25% per year, increasing its
coverage to 71% of cases with positive sputum smears in
2014. The concordance found between LPA and AP is
very good for resistance to H, R and MDR. The
universalization of this test in all the country is strongly
recommended.
S508
PC-1227-06 Evaluation of Mycobacterium

tuberculosis resistance levels and crossresistance to isoniazid and rifampin with their
respective structural analogues
1,2
1,2
I Rukasha, H Said, S Omar, A W Dreyer, A Hoosen,

N Ismail1 1National Institute for Communicable Diseases,
Johannesburg, 2University of Free State, Bloemfontein, South
Africa. Fax: (27) 71 477 8978. e-mail:
irukasha@yahoo.co.uk
Background: Studies have shown that some second-line

structural analogues of first-line drugs may have
equivalent or greater potency for MDR treatment. In
developing countries, second-line analogues are initiated
on failure of first line drugs without confirming their
efficacy. Structural analogue drugs such as Rifampicin
(RIF) and rifabutin (RFB) resistance to Mycobcaterium
tuberculosis is mutually associated with mutations
within the rifampicin resistance determining region
(RRDR) of the rpoB gene. Similarly, Isoniazid (INH)
and ethionamide (ETH) resistance is mediated by the
shared target InhA. Thus cross-resistance between
structural analogues is common. Particular genes in
rpoB mutations and in promoter of the inhA gene have
been shown to be likely associated with high levels of
resistance while, others may be found in phenotypically
susceptible isolates. Correlating specific mutation to
MICs allow for prompt detection of resistance.
Design/Methods: The aims of the present study were to
analyze the MICs data to determine whether crossresistance occurs between the structural analogues and to
correlate the MICs of RIF and RFB to the rpoB mutation
as well as MICs of ETH and INH to inhA and katG
mutations in order to determine levels of MICs in
relation to each mutation.The Sensititre MYCOTB assay
was used to determine the MIC levels. Sanger sequencing
was used to determine mutation. Kruskall-Wallis was
used to determine correlation between the MICs and
different mutations.
Results: Cross-resistance between the analogues RIF and
RFB was 89% while for INH and ETH 75%. Isolates
with mutations S531L, H526Y, H526R, S531Q in the
RRDR were associated with RIF/RFB resistance while,
mutations D516V and L533P were RIF resistance/RFB
susceptibility. All mutations in the inhA promoter and
katG were associated with high-level of resistance to
both INH and ETH.
Conclusion: There was a high level of cross resistance
between the analogues. Isolates with high INH and RIF
resistance levels also displayed high MIC values for their
structural analogs. These findings stress the need to test
the effectiveness of second-line structural analogues
before their incorporation in the therapeutic schemes.
The study has shown that the usefulness correlating
MICs to specific mutations which can allow rapid
detection of cross-resistance.
PC-1228-06 Routine surveillance for antituberculosis drug resistance in two districts of

Botswana, 20122014
A Finlay,1 J Basotli,1 C Modongo,2 E Click,3 J Oeltmann,3
R Boyd,1 N Zetola,2 P Moonan3 1Centers for Disease
Control and Prevention, Gaborone, 2Botswana University of
Pennsylvania Partnership, Gaborone, Botswana; 3U.S.
USA. e-mail: afinlay@cdc.gov
Background: Previous nationwide antituberculosis (TB)

drug resistance (DR) surveys demonstrated an increasing
trend of multidrug-resistant (MDR) TB from 1995
2008 in Botswana, a country with the second highest
prevalence of HIV in the world. The World Health
Organization recommends routine DR surveillance
through universal drug susceptibility testing (DST) to
monitor control efforts. We describe the performance of
routine DR surveillance among patients enrolled in a TB
transmission study in Gaborone an urban setting with
moderate TB incidence and high HIV prevalence and
Ghanzi a rural setting with high TB incidence and
moderate HIV prevalence.
Methods: From Aug 2012 Dec 2014, extra sputum
samples were collected from patients within two weeks
of initiating TB treatment in Gaborone and Ghanzi
districts. All Mycobacterium tuberculosis isolates underwent liquid or solid based DST for isoniazid, rifampin,
ethambutol and streptomycin. HIV status was collected
through patient history or rapid testing. District-wide
resistance prevalence estimates and 95% confidence
intervals (95%CI) were calculated. Differences in proportions for each anti-TB drug were compared to
previous survey data.
Results: During the study period, 3159 patients were
registered for TB treatment, among whom 2717 (86%)
were enrolled. Of these, 2499 (92%) submitted 7 1
sample for testing, 1423 (52%) had a culture positive for
M. tuberculosis and 1225 had complete DST results.
Resistance to isoniazid and rifampin (i.e., MDR-TB) was
detected among 37 (3.0 %) patients: 1.9% (n 15)
among those with new smear positive disease; 1.4% (n
3) among those with new smear negative disease; 9.5%
(n 19) among previously treated (Table). The prevalence of MDR-TB was 2.7% (95%CI 1.73.7) in
Gaborone and 4.3% (95%CI 1.76.9) in Ghanzi.
MDR-TB prevalence did not differ when stratified by
HIV status. The proportion of resistance for each anti-TB
drug was similar to previous surveys, except for any
ethambutol resistance, which was slightly higher 4.4%
(95%CI 3.0 5.9) vs. 1.8% (95%CI 1.12.9) in 2008.
Conclusions: Results of routine DR surveillance from
two districts show anti-TB drug resistance patterns
consistent with the 2008 national survey suggesting
current control efforts have not reduced the prevalence of
drug-resistant TB. Routine anti-TB DR surveillance is
feasible if efforts to collect initial diagnostic specimens
are improved and a modest increase in laboratory
capacity can be supported.
Table: Comparison of Antituberculosis Drug Desistance

Between the 2008 National Antituberculosis Drug Resistance
Survey and Routine surveillance in Two Districts in Botswana,
20142015
Resistance Pattern*
Newly diagnosed,
smear positive
Any drug
resistance
Any INH, RIF,
EMB
Any isoniazid
Any rifampicin
Any ethambutol
Any streptomycin
Any monodrug
resistance
Isoniazid
monodrug
resistance
Rifampicin
monodrug
resistance
Multidrug
resistance
Previously treated
patients
Any drug
resistance
Any INH, RIF,
EMB
Any isoniazid
Any rifampicin
Any ethambutol
Any streptomycin
Any monodrug
resistance
Isoniazid
monodrug
resistance
Rifampicin
monodrug
resistance
Multidrug
resistance
National Survey,
2008
Prevalence
(95% CI, n)
Routine
Surveillance
2014 2015
Prevalence
(95% CI, n)
n 924
n 810
16.9% (14.619.4,
20.3% (17.523.0,
156)
164)
8.9% (7.210.8, 82) 10.4% (0.812.5, 84)
7.6% (6.09.4, 70)
7.2% (5.48.9, 58)
3.6% (2.54.9, 33)
3.2% (2.04.4, 26)
1.8% (1.12.9, 17)
4.4% (3.05.9, 36)
10.6% (8.712.8, 96) 13.2% (10.815.6,
107)
13.0% (10.915.3,
15.7% (13.218.2,
120)
127)
3.8% (2.75.2, 35)
2.7% (1.63.8, 22)
1.1% (0.61.9, 10)
0.7% (0.21.9, 3)
2.5% (1.63.7, 23)
1.9% (0.92.8, 15)
n 137
n 200
23.4% (16.831.0, 32) 30.5% (24.036.9, 61)

23 (16.8% (11.223.8,
23)
10.2% (5.916.2, 14)
13.1% (8.219.6, 18)
6.6% (3.311.7, 9)
11.8% (7.118.0, 16)
14.6% (9.421.3, 20)
18.5% (13.123.9,37)
14.5%) (9.619.4, 29)
13.5% (8.718.3, 27)
5.5% (2.38.7, 11)
23.5% (17.629.4, 47)
17% (11.722.3, 34)
2.2% (0.65.8, 3)
2.0% (0.04.0, 4)
5.8% (2.710.8, 8)
3.0% (0.65.4, 6)
6.6% (3.311.7, 9)
9.5% (5.413.6, 19)
63. MDR-TB: detection, prevalence and

drivers
PC-1229-06 Why MDR-TB patients do not
accept treatment
Z Xu,1 L Bai1 1TB Control Institution of Hunan province,
Changsha, China. e-mail: evan.xu21@gmail.com
Background: There were 1091 MDR-TB patients detected through drug sensitive test (DST) in designated MDRTB hospital of Hunan province, China, from January 1,
2012 to December 31, 2013. According to the provincial
MDR-TB control guideline, all patients have to accept a
24 months treatment regime including first 4-6 weeks
hospitalization. However, only 419 cases started the
treatment to the end of March 2014, the other 672
S509
(61.6%) patients did not receive any proper treatment

although following up and tracing efforts were made by
all level TB control personnel. This study intends to
identify the reasons of patients who did not receive
MDR-TB treatment .
Design/Methods: The designated MDR-TB hospital
would inform the county CDC staff about MDR-TB
patients information once the diagnosis was confirmed.
CDCs personnel and/or the village doctors were due to
trace patients to the designated hospital for treatment,
and feedback the tracing results to hospital; then
hospitals would record patients treatment information
and feedback to the county CDC every month. The
patients not receiving treatment would be traced by
county CDC as well as newly diagnosed MDR-TB
patients, until all MDR-TB patients started treatment.
The information feedback was done mainly through email, QQ, MSN and telephone and the tracing procedure
and results were recorded, collected and analyzed.
Results: Among the 672 cases who did not started MDRTB treatment, the identified causes were the economic
hardship (386, 57.4%), moving to other city for work
(126, 18.8%), lacking of detailed address and telephone
number (47, 7%), refusing treatment (32, 4.8%), with
other severe illness or too weak to be hospitalized (28,
4.2%), and death (14, 2.1%).
Conclusion: The leading causing of patients not accepting MDR-TB treatment was economic hardship in
Hunan province. In addition, moving to other city for
work, unknown contact information, refusing treatment,
too weak or with other severe diseases and death were
other important reasons. Therefore, reducing treatment
expense and increasing the coverage of health insurance,
improving the MDR-TB treatment affordability and
availability need to be addressed to enhance the
proportion of patients of accepting MDR-TB treatment.
PC-1230-06 RNTCP experience in rolling out

base-line second-line drug susceptibility
testing among MDR-TB cases
M Ghedia,1 K S Sachdeva,1,2 A Amar Shah,1 S Anand,1
P Parmar,1 R Ramachandran,1 A Sreenivas,1
F Sayyed Imran1 1World Health Organization Country
Office for India, Delhi, 2Central TB Division, Delhi, India. Fax:
(91) 112 306 3226. e-mail: drmayank381981@gmail.com
Background and challenges to implementation: India is

highest MDR-TB burden country with approximately 61
000 MDR-TB cases among notified pulmonary TB cases.
The precise prevalence of XDR-TB among MDR-TB in
not known with only few survey have captured data that
is also very limited. RNTCP has well-articulated PMDT
services available across country with rapid diagnosis of
MDR-TB with CB-NAAT or LPA. The second line Drug
Susceptibility testing for MDR-TB was available to very
selected patients whose follow-up culture positive at fifth
month. The performing base line Second line DST in
MDR-TB was limited because of lack of adequate
number of SLD certified laboratories and no guideline
on SLD certification for laboratories.
S510
Intervention or response: To address these challenges the

Central Tuberculosis Division in coordination with
National Referance laboratory developed guidelines for
certifying the laboratory on second line DST in January
2013 with participation of all NRLs, finalized of
Standard Operative Procedure for Second Line Drug
susceptibility testing in liquid Culture (MGIT). Training
schedule for Laboratory Personal and proficiency testing
methods conducted the training of identified laboratory,
expedited the proficiency testing among identified
laboratories. All the document was duly approved by
technical committee and ministry of health. Training of
laboratory personal was conducted in NRL and on-site
in liquid culture (MGIT).
Results and lessons learnt: RTNCP has only three
laboratory (National Referance Laboratory) certified
for Second Line DST in 2013 and in March 2015 has 16
laboratory certified for SLD ( 2013-1, 2014- 6 2015- 6).
Development of guidance document for certification,
involvement of all NRLs for formulating guidelines and
training method will help to achieve rapid sacle-up of
SLD certified laboratories.
Conclusions and key recommendations: Based upon our
experience well-coordinated and step-wise approach
involving all stake holders e.g. National Programme,
Technical Committee, National Referance Laboratory
and Intermediate Laboratory generated notable result
and with a short time. With one year RNTCP reached 16
SLD certified laboratories which is currently providing
services for base line Second line DST among MDR-TB
patient in selected states.
PC-1231-06 Genetic characterization of

multidrug-resistant Mycobacterium
tuberculosis strains from various regions of
Delhi
J Arora,1 Z Sidiq,1 M Bhalla,1 P Visalakshi,1 V Myneedu,1
U Singh,2 D Behera,2,3 R Sarin1 1National Institute of
Tuberculosis and Respiratory Diseases, New Delhi, 2AIIMS,
New Delhi, 3PGIMER, Chandigarh, India. e-mail:
arojyoti@gmail.com
Background: Multi drug-resistant (MDR) tuberculosis is

a potential threat to tuberculosis (TB) control programs
throughout the world. The present study was aimed to
evaluate the population structure among multidrug
resistant Mycobacterium tuberculosis (M. tuberculosis)
strains isolated from MDR suspects registered at 6 DOTS
centers located in South, West and South-West zones
using spoligotyping.
Design/Methods: Sputum samples obtained were processed and inoculated on to Lowenstein Jensen slants.
Positive cultures were subjected to drug susceptibility
testing against four first line drugs- streptomycin (4lg/
ml), isoniazid, (0.1 lg/ml), rifampicin (40 lg/ml) and
ethambutol (1 lg/ml). Spoligotyping was performed
according to the standard protocol on 200 MDR isolates.
Results: Spoligotyping yielded 81 different patterns, 67
of these patterns were unique (1 isolate only) whereas 14
patterns containing 133 isolates were clustered (2 or
more isolates) with a clustering rate of 66.5%. SIT 26

predominated in this study with 49/200 (24.5%) isolates,
followed by SIT 1 with 45/200 (22.5%) isolates. The
predominant spoligotyping families in this study were
found to be Central Asian (CAS) family (43.0%),
followed by Beijing Family (23.0%), the East AfricanIndian family (8.0%) and the ill defined T family (5.0%).
The LAM, X and MANU were present as minor families.
The study showed an intriguing, marked distribution of
two strain families (CAS and Beijing) in all the regions.
Conclusion: The data obtained shows an almost similar
population structure of MDR strains in different regions
of Delhi. SIT 26 and SIT 1 were responsible for
approximately 50% of burden of MDR-TB in Delhi,
suggesting an enhanced ability of these strains to
transmit.
PC-1232-06 Multidrug-resistant tuberculosis

among foreign-born persons in Japan
N Kobayashi,1 T Hattori,2 T Kirikae,3 S Kato,4 J Suzuki,1
K Ohta1 1NHO Tokyo National Hospital, Kiyose, 2Tohoku
University, Sendai, 3National Center for Global Health and
Medicine, Tokyo, 4The Research Institute of Tuberculosis,
Kiyose, Japan. e-mail: kobayashin@tokyo-hosp.jp
Background: Although the rates of tuberculosis (TB) in

Japan have been decreasing, the proportion of TB cases
in the foreign-born has been steadily increasing from
2.1% to 5.2% of the national total during 1998-2013.
Recently, we demonstrated that multidrug-resistant TB
(MDR-TB) in Tokyo was more prevalent in foreign-born
patients compared with Japanese patients, and indicated,
by genotypic analysis of M. tuberculosis isolates, that
foreign-born patients most coming from high/relatively
high TB burden countries would have the potential risk
of spreading to other residents in Japan. The purpose of
this study is to clarify the current epidemiological status
of MDR-TB and contribution of foreign-born cases to
MDR-TB in Japan.
Design/Methods: We conducted a questionnaire survey
among MDR-TB patients registered from January 2011
to December 2013 in Japan. We analyzed the clinical and
microbiological differences between foreign- and Japanese-born MDR-TB patients.
Results: In the three-year period under study, a total of
184 TB cases were reported as MDR: mean age was 52.7
years, 65.2% were male. Ninety nine (53.8%) patients
were new cases and 84 (45.6%) were previously treated
cases. Among the MDR-TB, 20 (10.9%) cases showed
resistance to LVFX and KM. Japanese patients were 132
(71.7%) and foreign-born were 52 (28.3%). The foreignborn MDR-TB patients were female dominant and
markedly younger than native Japanese patients. Of the
52 foreign-born patients, 23 (44.2%) originated from
China. Other countries of origin were the Philippines (n
9), Myanmar (n 5), Mongolia (n 2), other Asian
countries (n 12) and Ukraine (n 1). Eight (20%) of 52
foreign-born patients with MDR-TB had organisms
resistant to LVFX, 3 (8%) to KM and 1 (3%) to EVM.
There was a higher prevalence of extensively drugresistant tuberculosis (XDR-TB) among Japanese MDR-
TB than among foreigner MDR-TB (14% vs.4%).

Treatment outcomes of foreign-born MDR-TB: 33
(65.3%) patients underwent successful treatment, 11
(23.1%) patients returned to their home countries during
treatment, 2 (3.8%) patients failed and 1 (1.9%) patient
died (3 patients unknown).
Conclusion: Among recently registered MDR-TB in
Japan, foreign-born patients accounted for about 30%,
who were mainly from Asian high burden TB countries.
Treatment outcome of foreigner MDR-TB was more
successful than Japanese MDR-TB. The return to their
home countries before completion of treatment might be
a problem.
PC-1233-06 Drug-resistant tuberculosis and risk

factors of multidrug resistance in Ukraine:
results of the first nationwide survey
E Pavlenko,1 A Barbova,2 A Hovhannesyan,3
Z Tsenilova,4 A Slavuckij,4 A Dean,5 M Dara,6
A Dadu7 1Temporary adviser of the Ministry of Health of
Ukraine, Kyiv, 2Central Reference Laboratory on TB
Microbiological Diagnostics of the Ministry of Health, Kyiv,
Ukraine; 3World Health Organization Temporary Adviser,
Yerevan, Armenia; 4World Health Organization, Country
Office, Kyiv, Ukraine; 5World Health Organization, Geneva,
Switzerland; 6World Health Organization Office at the
European Union, Brussels, Belgium; 7World Health
Organization, Regional Office for Europe, Copenhagen,
Denmark. e-mail: dad@euro.who.int
Background: Ukraine belongs to the worlds 27 countries

with high burden of multidrug resistant tuberculosis
(MDR-TB) was one of the few without sufficiently
representative data on level of drug-resistance.
Design/Methods: To investigate the levels and patterns of
resistance to first-line anti-TB drugs among new and
previously treated pulmonary TB cases and explore the
risk factors for multidrug-resistant (MDR) TB. Weighted
cluster sampling approach was used to select the study
participants. Patients enrolled between November 2013
and May 2014. Microscopy, culture and isolation of M.
tuberculosis were performed at regional level, while
drug-sensitivity tests (DST) were performed at five
selected inter-regional TB laboratories. Internal and
external quality assurance was performed correspondingly at National Reference Laboratory in Kiev and
Supranational Reference Laboratory in Riga.
Results: Of the 1590 patients enrolled into study, 1266
were newly-diagnosed and 324 previously treated TB
patients. MDR- TB was detected in 22.7% (95%
Confidence Interval (CI): 19.6-25.9) among the new
patients and 58.3% (95%CI: 52.1-64.6) among the
previously treated patients. Human immunodeficiency
virus (HIV) was found in 249 (15.7%) patients.
Independent risk factors for MDR-TB among new
patients were HIV (Odd Ratio (OR): 1.44; 95%CI:
1.02-2.03), poor socio-economic status (OR: 1.55;
95%CI: 1.16-2.05), and living in South-East oblasts
(OR: 2.37; 95%CI: 1.51-3.72). Age was inversely
associated with MDR-TB in linear pattern: every 10
years increase in age reduced Odd of MDR in new cases
S511
by 19% (OR: 0.81; 95%CI 0.72-0.91). Among the

previously treated patients the risk factors associated
with MDR-TB were the setting and outcome of previous
treatment: previously treated patients living in large cities
had 72% lower Odd of MDR-TB compared to these that
live in towns (OR: 0.24; 95%CI: 0.11-0.51), while
patients that were lost to follow-up had lower Odd of
MDR-TB compared to those relapsed (OR: 0.42;
95%CI: 0.21-0.83).
Conclusion: The prevalence of MDR among new TB
patients in Ukraine is higher compared to WHO
estimates and routine drug resistance surveillance data.
Findings suggest that there is urgent need to for better
HIV-TB collaborative activities, addressing social determinant of TB and accelerating the MDR-TB detection,
treatment and improve treatment adherence. Further
operational research are needed for full understanding of
DR TB in Ukraine.
PC-1234-06 Prevalence of MDR-TB and drug

resistance patterns in retreatment cases in
Punjab, Pakistan
A Majeed,1 M Awais Arif,1 H Javed,1 S Kanwal1 1Provincial
TB Control Program, Punjab, Lahore, Pakistan. e-mail:
hasnain_javed@hotmail.com
Background: Tuberculosis has emerged a serious public

health problem in recent years. In 2013, 9 million people
fell ill with TB and 1.5 million died from the disease
while an estimated 480 000 people developed multidrugresistant TB (MDR-TB). Pakistan ranks 4th in 22 high
burden TB countries and ranks 4th in 27 high burden
MDR-TB countries. However little or no data exists on
prevalence and/or drug resistance pattern amongst high
risk MDR-TB groups from Punjab, the most populous
province of the country. Previously treated cases are
considered at high priority for drug susceptibility testing
to identify cases with multi-drug resistance (MDR-TB).
Design/Methods: A total of 1250 retreatment cases were
included in this study from January 2010 to December
2014. This cross sectional study was carried out in 9
tertiary care hospitals implementing Programatic Management of Drug Resistance TB (PMDT) in Punjab. TB
culture and drug resistance pattern study was performed
at National TB Reference Lab Islamabad, Agha Khan
diagnostic and research Laboratories and Provincial TB
Reference Lab, Lahore. Data was collected from the
Electronic Nominal Review System (ENRS) and analysis
was done with SPSS version 17.
Results: Among 1250 retreatment cases, 861 (69%) were
MDR-TB while 31% of the remaining showed mono
resistance. The mean age was 32 years; amongst them
53% were males and 97% resided urban area of Punjab.
Our study showed that MDR-TB is more prevalent in the
most productive age group 15-45 years (57%; P
0.000), in the urban area of the Punjab (97%, P 0.040)
and in the pulmonary site (68%; P 0.041). Overall
41% of retreatment cases showed resistance to all first
line anti-TB drugs (RHZES) while 28% showed resistance to oral first line anti-TB drugs (RHZE). Five
S512
hundred and twenty six retreatment cases also have the

resistance to FQs and SLD, amongst them 420 (81%, P
0.000) patients showed highly significant resistance to
(FQs) and 5% had Ethionamide resistance.
Conclusion: Prevalence of MDR-TB among enrolled
retreatment cases is alarmingly high. Drug-resistant
tuberculosis including MDR-TB has emerged as a serious
public health problem in Pakistan and will greatly impact
TB control strategies. This study may have important
implications for TB control as it highlights the need to
fully implement National Tuberculosis guidelines with
main focus on expansion and effective implementation of
DOTS in order to prevent further emergence of drug
resistance.
PC-1235-06 Estimating true MDR prevalence

rates using routinely collected data
Z Mclaren,1 R Burger2 1University of Michigan, Ann Arbor,
MI, USA; 2Stellenbosch University, Stellenbosch, South Africa.
Background: In the case of the tuberculosis epidemic in

South Africa, where TB has been the leading cause of
death for over a decade, the lack of reliable estimates of
local tuberculosis prevalence leads to inefficient allocation of government resources. Surveillance studies can
provide an unbiased estimate of prevalence, but are
costly and therefore infrequent. In the absence of
inexpensive, accurate diagnostic tests accessible to the
majority of the patient population, statistical methods
can produce an unbiased estimate of prevalence from
routinely collected data.
Design/Methods: In this context, we use high-quality
tuberculosis test result data from South Africas National
Health Laboratory Service (NHLS) to estimate the
prevalence of MDR-TB. The data set spans 2004-2011
and includes 32 058 877 test records from over 12 445
000 patients at health facilities around the country. We
develop a theoretical model of the clinicians decision to
perform drug susceptibility testing on a patient with
suspected TB based on the patients risk factor profile
and subject to clinical guidelines and limited resources.
Our identification strategy employs a simulated maximum likelihood estimation (SMLE) method and a
generalized method of moments (GMM) to generate
estimates. We rely on exogenous variation in drug
resistance testing rates induced by time period, national
health policy changes, and local resource allocation to
estimate our parameters.
Results: We estimate that the MDR rate in South Africa
could be as high as 3.63%. This is likely to be an upper
bound because there is information that the clinicians
have access to that we, as statisticians, do not. The
subgroup analysis demonstrates that the estimated
MDR-TB rates are higher in health clinics and health
centers, for women, those between ages 30-60, and for
Northern Cape, Mpumalanga, North West and Eastern
Cape (KZN omitted from analysis due to data limitations).
Conclusion: Our results suggest that resource allocations

for tuberculosis diagnosis and treatment should be
revised upwards. Resources should be targeted towards
provinces that have the lowest index of suspicion for DR
or that have the highest MDR rates. More frequent
surveillance of MDR-TB is needed to track the evolution
of MDR-TB prevalence over time. Between surveillance
studies, modeling of disease prevalence rates are an
inexpensive and efficient way to guide the allocation of
TB resources.
PC-1236-06 Drug resistance patterns in MDR-TB

patients: challenges for future regimen design
H Stagg,1 J Brown,1,2 E Ibraim,3 V Riekstina,4,5
P Viiklepp,6 G Dravniece,7,8 C Jackson,1 P White9,10
1
University College London, London, 2Royal Free London NHS
Foundation Trust, London, UK; 3Marius Nasta Institute of
Pulmonology, Bucharest, Romania; 4Riga East University
Hospital, Riga, 5Centre of TB and Lung Diseases, Riga East
University Hospital, Riga, Latvia; 6Estonian Tuberculosis
Registry, National Institute for Health Development, Tallinn,
Estonia; 7KNCV Tuberculosis Foundation, The Hague,
Netherlands; 8Otsuka Pharmaceutical, Geneva, Switzerland;
9
Imperial College School of Public Health, London, 10Public
Health England, London, UK. e-mail: h.stagg@ucl.ac.uk
Background: Globally, numbers of multi-drug resistant

tuberculosis (MDR-TB) cases are growing, with many
concerns about the challenges of treating what may be
increasingly complex drug resistance patterns. We sought
to inform future regimen selection by examining
additional drug resistance in MDR-TB cases over time
in three epidemiologically differing nations of the WHO
European region: Estonia, Latvia and Romania.
Methods: Analysis of demographic (vulnerable groups)
and microbiological (drug sensitivity testing (DST))
surveillance data for Estonian, Latvian and Romanian
MDR-TB cases, 2004-13.
Results: In 2013 Estonia, Latvia and Romania had 50, 76
and 575 MDR-TB cases, respectively. The percentage of
cases undergoing DST to second line drugs was high (98100%) for Latvian and Estonian cases, but only 67% in
Romania. The percentage of MDR-TB cases resistant to
fluoroquinolones and/or injectables was concerning in all
three countries (Estonia 50%, Latvia 47%, Romania
25%). In Romania more MDR-TB cases are resistant to
second-line injectables than fluoroquinolones; the reverse is true in Estonia. Since 2004 absolute numbers of
MDR-TB cases have decreased in all countries (27-59%
decrease). Although the proportion of MDR-TB cases
with resistance to injectables has decreased during the
study period, there has been an increase in cases with
additional resistance to fluoroquinolones. The proportion of MDR-TB cases with XDR-TB was 1.7-fold higher
in Latvia in 2013 (17%) vs. 2004 (10%). The percentage
of cases resistant to pyrazinamide in 2013, where tested,
was high for both retreatment and new cases (69-75%).
The percentage of MDR-TB cases resistant to ofloxacin
was high in Estonia and for retreatment cases in Latvia
and Romania. Prothionamide resistance was common
(43-73%) in Estonian retreatment cases and both new
and retreatment cases in Latvia. Latvia had the highest

percentage of MDR-TB cases known to be co-infected
with HIV (24%).
Conclusions: High levels of additional resistance to
currently recommended second line drugs was seen in
all countries, with important variations between countries in drug sensitivity profiles. Accurate DST and DST
data for second line drugs from a variety of countries is
vital for comparison to WHO treatment recommendations in order to inform the development of future MDRTB treatment regimens that meet the needs of all regions
and patient groups.
64. Patient-centered care and methods to

improve treatment adherence
PC-1237-06 Health-related quality of life
assessment using EQ-5D in newly diagnosed
pulmonary tuberculosis patients
S Mcallister,1 S Imaculata,2 Y Laurence,3 S Kerry,4
J Critchley,4 P Hill,1 R Ruslami,2 B Alisjahbana2 1University
of Otago, Dunedin, New Zealand; 2Padjadjaran University,
Bandung, Indonesia; 3London School of Hygiene & Tropical
Medicine, London, 4St Georges University of London,
London, UK. e-mail: sue.mcallister@otago.ac.nz
Background: Tuberculosis (TB) remains a major public

health problem worldwide. TB studies have focussed
mainly on severity and clinical outcomes but TB can have
an impact on patient health-related quality of life
(HRQOL). However, there is limited information available on this, particularly in TB endemic countries. Our
study evaluated HRQOL and associated characteristics
in pulmonary TB patients in Bandung, Indonesia.
Design/Methods: As part of the TANDEM program on
TB and diabetes in different countries, we conducted a
cross-sectional study of newly diagnosed pulmonary TB
patients aged .18 years recruited from Community
Health Clinics and DOTS clinic. In-person interviews
were conducted to ask about socio-demographic characteristics and HRQOL using the EuroQol Group EQ5D-5L Indonesian validated measure (dichotomised to 2levels for analysis No problems; Any problems), and
Visual Analogue Scale (VAS: a scale from 0 to 100). The
Karnofsky score was clinician-assessed. Odds ratios
(OR), 95% confidence intervals (CI), coefficients and P
s were generated using logistic regression and ordered
Results: A total of 275 TB patients were recruited from
February 2014 to January 2015. The proportion
reporting Any problems was highest for EQ-5D Pain/
discomfort (75%), followed by Anxiety/depression
(59%), Usual activities (45%), Mobility (28%), and
Self-care (13%). The mean VAS score was 70.4 (SD
17.8). People with a bank account had significantly
reduced risk of reporting Any problems on four EQ-5D
variables: Mobility (OR 0.41, 95%CI 0.19-0.83), Usual
activities (OR 0.44, 95%CI 0.25-0.78), Pain/discomfort
(OR 0.39, 95%CI 0.22-0.73) and Anxiety/depression
S513
(OR 0.52, 95%CI 0.29-0.90). People with higher

education had a substantially reduced risk of Any
problems for two EQ-5D variables: Usual activities
(OR 0.13, 95%CI 0.03-0.52) and Pain/discomfort (OR
0.19, 95%CI 0.05-0.70). Those aged 760 years were
significantly more likely to report mobility problems
(OR 3.29, 95%CI 1.34-8.08). All the EQ-5D variables
correlated with the Karnofsky score, but the correlation
was relatively weak for anxiety/depression.
Conclusion: The EQ-5D provides important information
on HRQOL in TB patients, particularly in understanding
pain/discomfort and anxiety/depression. Poorer TB
patients are likely to have other concerns that affect
their HRQOL. The EQ-5D can provide clinicians with
insights into patients HRQOL that may be amenable to
intervention.
PC-1238-06 Audit and improvement of the

clinical pathway for adult pulmonary
tuberculosis in Bethesda Hospital, Yogyakarta,
Indonesia
T Lestari,1 A Probandari,1,2 H A Sanjoto,1 H Djasri,1
Iswanto,3 Y Mahendradhata,1 A Utarini1 1Center for
Health Policy and Management, Yogyakarta, 2Sebelas Maret
University, Surakarta, 3Bethesda Hospital, Yogyakarta,
Indonesia. Fax: (62) 27454 9425. e-mail:
trisasilestari@gmail.com
Background and challenges to implementation: Clinical

pathway (CP) is a structured multidisciplinary care plans
which detail essential steps in the care of patients with
specific clinical problem. Bethesda hospital has been
using a CP for adult pulmonary tuberculosis (TB) since
2008 and it has been proven effective to improve
compliance of a clinical care team to a standard clinical
guideline for TB. However, the CP has not been reviewed
since its development.
Intervention or response: This was a one cycle action
research to improve the quality of CP for TB in Bethesda
hospital as part of a project to develop a sustainable
hospital disease management system. In April 2015, a
clinical care team audited the CP using the Integrated
Clinical Pathway Appraisal Tool (ICPAT) to helps the
team identified the weakness of the CP and make
improvements. There are 6 review dimensions, i.e. the
structure, the documentation, the development, the
implementation, the maintenance and the role of
organisation. Each dimension contains a list of review
items to the content and the quality of CP. The CP was
revised according to the audits finding and tested for two
months before it is implemented.
Results and lessons learnt: The CP for tuberculosis in
Bethesda hospital has fulfilled 42 out of 58 (72%)
criterias for good content of CP and 35 out of 49 (71%)
criterias for good quality of CP. Appraisal scores for each
review dimensions were as followed: 75% in the
structure, 56% in the documentation, 73% in the
development, 83% in the implementation, 65% in the
maintenance, and 100% in the role of organisation.
Majority of unmet criterias were related to patients
involvement in the treatment (43%). Other unmet
S514
criterias in the content of CP were the function of CP as a

reminder for possible variations in the treatment,
structure of the document, legality, and storage of CP.
Unmet criterias in the quality of CP were identification of
responsible person for each activities, identification of
priority activities, schedule for review, and specific code
for existing variations. The CP is currently under revision
and will be tested in May to June 2015.
Conclusions and key recommendations: The clinical care
team has not been considered patients involvement in
making clinical decision in their course of treatment for
TB which is important in the era of patient-centeredness.
Audit of CP using the ICPAT can improve the quality of
care for TB in hospital.
PC-1239-06 Exito del TAES en personas con TB,

centrada
afromestizos, mediante la atencion
por enfermera, en Cuajinicuilapa, Costa Chica,
Guerrero, Mexico
R E Huicochea,1 M Clemente,1 V Leyva1 1Secretara de
Salud, Chilapancingo, Guerrero, Mexico. Fax: (52) 55 261
46 438. e-mail: arceliaavena@yahoo.com.mx
Justificacion
: El municipio de Cuajinicuilapa registra
cada ano
18 casos nuevos de TB, aportando 8.5% de la
morbilidad de la region.
El 16.3%, de los enfermos son
afromestizos, que de acuerdo a su cosmovision
(creencias
de hechicera, brujera),de caracter fuerte, representan un
reto para los servicios de salud para llevarlos al e xito del
Tratamiento, aunado a factores socioeconomicos,
que lo
condicionan, lo cual hace necesario brindar la atencion
centrada en estas personas.

Metodologa : Se desarrollo la estrategia Atencion
integral, en cinco fases implementadas por la enfermera

de Red TAES; la primera empoderamiento al enfermo
sobre la TB, haciendole sentir respeto, confianza y
reduccion
del tiempo en sala de espera, llamandolo por
su nombre, monitoreando reacciones adversas; la segunda: visitas domiciliarias mensuales, la tercera: atencion
psicologica,
la cuarta: colectas alimenticias en ferias TB,
rifas y boteo economico,

quinta: premiacion
al esfuerzo,
otorgando estmulos, resultado de BK de control
negativo, ganancia de peso e invitandole un almuerzo;
se analizo historial de la TB en este grupo e tnico.
Resultados: de 1996 a 2015, se diagnosticaron 59 casos
nuevos TBTF, 98.3% TBP y 1.7%miliar, 72.9% fueron
detectados en consulta externa, 27.9% en pesquisa casa a
casa, 33.9% en contactos. La positividad fue de 28.1%
(),13.6% (), 32.2%(),25.4% por Rx; comorbilidades: desnutricion
(69.5%), Diabetes Mellitus 23.7% y
VIH 5.0%, rango edad 21 a 79 anos.
66.1% fueron
hombres, 33.6%mujeres, beneficiandose 59 personas

afromestizas, dotacion
alimentaria quincenal, consultas
y suplementos alimenticios mensual, pasajes; gestion
de
actas y constancias de identidad, polizas
de seguro
popular, laboratorios, Rx particulares, curo 94.9% y

5.1% continua
en tratamiento.
Conclusiones: La estrategia de atencion
integral personalizada por enfermera en personas afromestizas afectadas por TB, no modifico la cosmovision
con relacion
a la
enfermedad, solo
garantizo el e xito del TAES, demostrando la importancia de la deteccion
en los contactos,
para incidir en el control de la TB, en el Municipio de
Cuajinicuilapa, Guerrero.
PC-1240-06 Patient-centred behaviour change

communication materials can help to reduce
social stigma and improve treatment
adherence
G Mishra,1 L Lochting,2 H Amdal,2 B Lamichhane,3
A Thapa3 1LHL International Tuberculosis Foundation,
Norway, Kathmandu, Nepal; 2LHL International Tuberculosis
Foundation, Oslo, Norway; 3National Tuberculosis Centre,
Bhaktapur, Nepal. e-mail: gokulmishra@gmail.com
Background and challenges to implementation: There

was a serious lack of patient centred behaviour change
communication materials in the Nepal National Tuberculosis Programme (NTP). Considering this fact, the
LHL International Tuberculosis Foundation, Norway
and NTP jointly organised three patient centred needs
assessment workshops and two focus group discussions
with the health workers and patients together in different
parts of Nepal- to assess the real information needs of TB
patients on tuberculosis in 2010. This first edition of
patient centred TB booklet (first published in 2011)
contains a description of how you can use this booklet,
basic facts about tuberculosis, facts about DR-TB and
HIV/AIDS, how to cope with tuberculosis and the side
effects of treatment and keeping body and soul well when
you have TB.
Intervention or response: Patient Centred Tuberculosis
booklet was developed with patients and health workers
together. Health communication trainings were organised for health workers and volunteers and TB booklets
were distributed to health workers, volunteers and TB
patients at Kathmandu Metropolitan City, Nepal.
Results and lessons learnt: 100% of TB patients shared
their knowledge that TB patients should cover their
mouth and nose while coughing and sneezing. The
majority of the TB patients used a mask for covering their
mouth and nose followed by handkerchief (32.8%),
blocking with palms (13.8%). 63.9 % TB patients
received advice about good nutrition during TB treatment from health workers while 36.1 % of TB patients
received the same information from a TB booklet. All
respondents reported experiencing less social stigmatism
as a result of contracting TB after sharing the TB booklet
with their family. Most of the TB patients also reported
that the booklet made it easier for them to speak to their
family and friends about TB.
Conclusions and key recommendations: The booklet was
found to be an effective tool in improving relationships
with TB patients, bringing good practices, motivating
health workers while dealing with TB patients and
reducing discrimination and stigma.
PC-1241-06 Unveiling treatment supporter

roles behind completion of treatment by TB
patients: implications for a new agenda setting
for lung health
1
R Olukolade, A Hassan, R Kusimo, L Okwuonye,

J Okechukwu,2 J Osho,1 S T Abdurrahman,2 C Ogbuji1
1
Association for Reproductive and Family Health, Abuja,
2
Federal Capital Territory Tuberculosis and Leprosy Control
Programme, Abuja, Nigeria. e-mail: richkolade@gmail.com
Background: Tuberculosis control programme in Nigeria

allows for the involvement of Treatment Supporters (TS)
basically in monitoring patients daily drug intake.
However, the identification of the support needs of TB
patients is more appropriate to inform the engagement
and the expected roles of TS. This paper unveils what the
roles of TS should be, from the perspectives of TB
patients.
Methods: Qualitative data was collected from three
categories of participants; TB patients, TS and DOTS
Staff, selected from six TB clinics representing each of the
six Area Councils of the Federal Capital Territory (FCT)
of Abuja, Nigeria. Eleven Focus Group
Discussions; (5 male and 6 female groups) was conducted
among TB patients currently on treatment. Fifteen (15)
TS identified from the facilities TB registers were
interviewed as key informants. In-depth interview was
conducted with one (1) TB clinic staff, in each of the six
facilities where the FGDs were conducted. Data generated was content analyzed and interpreted in relation to
the objectives of the study and themes in the Discussion
Guide.
Results: Majority of TB patients expressed their challenges while on treatment. Four key areas of TB patients
support needs were identified; (1) monitoring and
supervision of daily drug in-take, (2) motivational
support to take the drugs as expected, (3) provision of
support in feeding (when there is no food or means of
eating), and (4) provision of support in supplying
transportation cost to visit TB clinic if there is no means
of doing so. Left on their own, TB patients may not be
committed to taking their drugs and complete their
treatment regimen, if any of these supports is lacking
when needed. Patients with TS who offer these supports
tend to complete their treatment regimen and not likely
to default.
Recommendations: Although TB drugs are free in
Nigeria and TS are being engaged to support Treatment,
TB patients support needs should inform the design of
Community TB care.
S515
PC-1242-06 A patient-centered TB counseling

strategy impacts out-patient follow-up in
Yunnan, China
M Li,1 A Innes,2,4 X Zhao,3 l Ling,3 M Ma,1 M Li,1 C Du,1
X Xu3 1Kunming No. 3 Hospital, Kunming, Yunnan Province,
China; 2FHI 360, Bangkok, Thailand; 3FHI 360 China,
Kunming, Yunnan Province, China; 4FHI 360 Asia Pacific
Regional Office, Bangkok, Thailand. e-mail:
ainnes@fhi360.org
Background: Adherence counseling on TB and multidrug resistant tuberculosis (MDR-TB) using a patientcentered strategy is unique in China, where communication is normally approached from a providers perspective. The USAID Control and Prevention-Tuberculosis
(CAP-TB) China program, managed by FHI 360,
partnered with Kunming No.3 Hospital to use a
patient-centered counseling strategy that was integrated
into daily education. The goal was to prepare inpatients
for outpatient treatment and adherence following
hospital discharge.
Intervention: A total of 150 newly registered TB
inpatients at Kunming No. 3 Hospital participated over
four months in 2014. During inpatient hospitalization,
nurses or TB peer educators provided each patient with
at least three in-depth, one-on-one counseling sessions:
upon admission, during inpatient hospitalization, and
prior to discharge. On the day of TB treatment initiation
and on the day before hospital discharge, patients were
tested on their basic TB knowledge and awareness. Preand postcounseling TB knowledge was compared using
Z-test for proportions.
Results: Among the 150 patients, 133 (89%) were newly
diagnosed TB cases, and 17 (11%) were previously
treated. Seventy-five of the patients were from outside
Kunming City; 58% had lower than secondary education; and 63% self-reported as farmers, migrant or
mobile workers, or unemployed. Twenty (13%) reported
previous history of TB among family members. TB
knowledge increased significantly after inpatient counseling when compared to pre-counseling (all P , 0.005).
The post-counseling test showed that 98% knew their TB
status correctly, compared to 49% on the pre-test.
Knowledge of TB transmission increased (from 16.7%
to 74%); as well as knowledge on drug-susceptible
treatment course (from 50.7% to 92.7%) and MDR-TB
treatment course (41% to 85%). Patients also improved
their knowledge on timing for follow-up sputum testing
(60% increase). After this inpatient counseling strategy
was implemented, the number of TB outpatients at
Kunming No.3 Hospital nearly doubled and was
maintained at increased levels compared to pre-counseling intervention.
Conclusions: Using a patient-centered TB counseling
strategy increased inpatient TB knowledge and was
associated with improved attendance for outpatient
follow-up visits. Longer-term outcomes, such as impact
on treatment completion and cure, are still to be
determined and will be followed for this cohort.
S516
Table 1: TB knowledge and awareness in pre- and postcounseling tests among TB inpatients at Kunming No. 3
Hospital
TB knowledge
and awareness
Total number
of patients

Know their TB
category (TB,MDRTB, newly treated or
retreatment)
Know TB
transmission route
Know TB symptoms
Know TB and MDRTB are curable
Know treatment
course for TB
Know treatment
course for MDR-TB
Know TB medication
has side effects
Know TB treatment
principles
Know when a newly
treated patient should
return for sputum test
Know when a
retreatment patient
should return for
sputum test
Know what causes
MDR-TB
Pre-counseling Post-counseling
N
150
100
150
100
p-value
74
49.3
147
98.0
,0.001
25
95
16.7
63.3
111
131
74.0
87.3
,0.001
,0.001
61
40.7
111
74.0
,0.001
76
50.7
139
92.7
,0.001
62
41.3
128
85.3
,0.001
113
75.3
132
88.0
,0.005
96
64.0
137
91.3
,0.001
20
13.3
119
79.3
,0.001
15
10.0
105
70.0
,0.001
27
18.0
59
40.0
,0.001
PC-1243-06 Barriers and facilitators to

treatment adherence among MDR-TB patients
in Yunnan, China
X Lin,1 l Ling,2 X Xu,2 C Wong,3 Y Guo,1 A Innes4 1Yunnan
Center for Disease Control and Prevention, Kunming,
Yunnan Province, 2FHI 360 China, Kunming, Yunnan
Province, China; 3FHI 360, Durham, NC, USA; 4FHI 360 Asia
Pacific Regional Office, Bangkok, Thailand. e-mail:
ainnes@fhi360.org
Background: China is a high-burden TB country with the

second highest number of multi-drug resistant (MDR)
TB cases worldwide. Although TB prevalence in China
declined from 2000-2010, the MDR-TB burden is high,
underscoring challenges to treatment adherence. We
explored the barriers and facilitators to treatment
adherence in order to develop communication and
support strategies for MDR-TB patients, with the goal
to improve treatment outcomes.
Intervention: Semi-structured interviews were conducted
with the first 24 MDR-TB patients enrolled in the China
Global Fund TB program in Yunnan, China in 2013.
Patients were asked how they were diagnosed with TB or
MDR-TB; about the barriers and facilitators for treatment adherence; and the impact of TB and MDR-TB on
their lives, with a specific focus on issues impacting
treatment adherence. Interviews were transcribed or
written up as expanded notes and analyzed using applied
thematic analysis.
Results: We identified four overarching barriers to

treatment adherence: economic and geographic factors
that influence access to care; experience-based knowledge or beliefs about healthcare choices; perceived and
actual fears of treatment side effects; and complexity of
information on TB treatment. Economic hardship was
the most frequently reported barrier, when patients could
not pay for drugs, hospitalization, and transportation.
Some patients from rural areas discontinued follow-up
visits and then bought drugs from private pharmacies
when they felt sick, or they took Chinese traditional
medicine. Some also stopped treatment when their cough
disappeared, thinking that they had fully recovered.
Other barriers included self-stigma around TB and the
perceived fear of how the side effects would disrupt their
lives, as well as the actual side effects experienced, which
led to reduction or discontinuation of treatment. In
contrast, facilitators of adherence included physician
support and encouragement during clinic visits and
phone calls. Communication among patients was also
important for facilitating adherence, enabling them to
learn from their peers about TB and how to overcome
challenges.
Conclusions: There are many barriers that impact MDRTB treatment adherence. To improve treatment outcomes, a patient-centered strategy should be used that
addresses stigma, economic, and other barriers; that
simplifies TB education to improve comprehension; and
that teaches about side effects and their treatment.
PC-1244-06 Patient satisfaction and adherence

to medication among adult patients with
pulmonary tuberculosis in Uganda
H Babikako,1 D Neuhauser,2 A Katamba,3 E Mupere,3
C C Whalen4 1Makerere University, College of Health
Sciences, School of Medicine, Kampala, 2Case Western
Reserve University, Kampala, 3College of Health Sciences,
Makerere University, Kampala, Uganda; 4College of Public
Health, University of Georgia, Athens, GA, USA. Fax: (256)
414 533 531. e-mail: mupez@yahoo.com
Background: Non-adherence to tuberculosis (TB) treatment arises from the interaction of multiple factors
affecting the quality of TB care. Current TB program
services and clinical research have focused largely on
outcomes of mortality and microbiologic cure, and have
not considered patients preferences; yet patients beliefs
and attitudes may influence how patients take drugs. We
therefore established whether patient satisfaction was
associated with non-adherence to prescribed TB treatment in urban Uganda.
Methods: In a cross-sectional study of 469 smear positive
TB patients attending five program clinics in Kampala
city, we measured patient satisfaction with service care
(PS) and satisfaction with TB medication information
received (SIMS) using a PS 13-item and SIMS 17-item
questionnaires, respectively, and non-adherence using
Morisky scale among patients completing 2, 5 and 8
months of treatment. We used our previously published
subscales for PS 13-item of clinic organization set-up,
technical quality, and hospital care; and for SIMS 17-
item of information about action and about effects of

medication. Sub-scales were scored as summated scales
on a 0 100 level where higher scores indicated better
satisfaction.
Results: Scores for satisfaction with information about
action and about effects of medication were all below
50% at 49.8 6 14.9 and 28.2 6 28.2, respectively. Of the
three subscales for PS 13-item tool, clinic organization
set-up had the lowest score at 63.3 6 7.8. Scores for
clinic organization set-up, hospital care and information
about effects of the medication were significantly higher
in HIV positive compared to negative patients. For
example, among 135 HIV positive patients, score for
clinic organization set-up was 64.5 6 6.1 versus 62.6 6
8.6 in 314 HIV negative patients, respectively (P
0.016). Among 135 non-adherent patients, clinic organization satisfaction score was significantly lower 62.1 6
8.8 versus 63.8 6 7.3 in 334 adherent patients,
respectively (P 0.014). A unit increase in clinic
organization set-up satisfaction score was associated
with a 4% less likelihood for adherence (OR 0.964;
(95%CI: 0.933, 0.995)) among HIV negative patients.
Conclusion: We found sub-optimal patient satisfaction
with information about action and about effects of
medication, and HIV positive status was associated with
improved satisfaction. Low satisfaction with clinic
organization was associated with non-adherent in HIV
negative patients.
PC-1245-06 The effects of peer education on

treatment adherence of MDR-TB patients in
Kunming, China
G Nie,1 W Zhang1 1Institute of Tuberculosis Prevention and
Control, Yunnan Center for Disease Control and Prevention,
Kunming, China. e-mail: tudor@icn.ch
Background: The incidence of MDR-TB in China is

8.3%, and 7.0% in Kunming, Yunnan Province. Primary
cause of MDR-TB is irregular treatment of drugsusceptible TB. This study sought to explore the
influence of peer education on treatment adherence for
patients with multi-drug resistant tuberculosis (MDRTB).
Design/Methods: 157 patients with MDR-TB were
randomly divided into an observation group (77 patients) and control group (80 patients). The patients in
the observation group received peer education and
routine health education. The patients in the control
group received routine health education. We compared
patient knowledge about MDR-TB, changes in mood
and anxiety and treatment adherence between the two
groups after a six-month intervention.
Results: Patients in the peer education group had
significantly better knowledge about MDR-TB (P ,
0.01), reduced anxiety (P 0.05), and improved
treatment adherence than the control group (P 0.01).
Conclusion: Peer education can significantly improve
patient knowledge about MDR-TB, reduce anxiety and
improve treatment adherence in patients with MDR-TB.
S517
PC-1246-06 Where there is hope: a qualitative

study examining adherence to MDR-TB
treatment in Karakalpakstan, Uzbekistan
S Horter,1 B Stringer,1 J Greig,1 P Du Cros,1 S Dietrich,2
M Tillyashaykhov,3 N N Parpieva,3 Z Tigay4 1Medecins

`
Sans Frontieres
(MSF), London, UK; 2 MSF, Berlin, Germany;
3
Ministry of Health of the Republic of Uzbekistan, Tashkent,
4
TB II Hospital, Nukus, Uzbekistan. e-mail:
shona.horter@london.msf.org
Background: Multi-drug resistant tuberculosis (MDRTB) treatment is complex and lengthy, with a cocktail of
toxic drugs that have limited efficacy. Adherence to
treatment is challenging, and non-completion of treatment has significant medical and public health implications. A qualitative study was conducted to understand
patients experiences with adherence to MDR-TB treatment in Karakalpakstan, Uzbekistan, to inform future
strategies of adherence support.
Methods: Participants were recruited purposively to
ensure a variety of treatment experiences, patient
characteristics, and health practitioner perspectives. 52
in-depth interviews were conducted with MDR-TB
patients (n 35) and health practitioners (n 12),
including 5 follow-up interviews to probe emerging
themes. Interview transcripts were analysed thematically
using coding. Analysis drew upon principles of grounded
theory, with constant comparison of codes and categories
within and between cases to actively seek discrepancies
and generate concepts from participant accounts.
Results: Several factors affected patients adherence to
MDR-TB treatment. Many patients had limited TB
knowledge, influencing their belief in the need for and
value of treatment and their hope for cure, and
undermining their motivation for treatment. Patients
ability to choose treatment and prioritise it over
conflicting demands within their lives relied on their
sense of autonomy and control, as well as their
ownership and self-responsibility for treatment. Patients
authority over their treatment-taking and their engagement with it was influenced by common beliefs, myths
and stigma, as well as by hierarchical practitioner-patient
relationships. Womens societal role as daughter-in-law
could undermine their authority over treatment. Patients
tolerance of treatment was linked to their views about
drug effects, with a positive mindset seen as improving
side effects experienced.
Conclusion: This qualitative study reveals three main
themes influencing adherence to MDR-TB treatment:
disease knowledge and belief, patient autonomy and
control, and views on TB drugs. It reinforces the policy
for an individualised, holistic and patient-centred approach to treatment support for effective TB control and
treatment. Patient knowledge about their disease and
treatment, and engagement of patients as active participants in their care was endorsed as a key enabler for
adherence, strategies that acknowledge this should be
included in programming.
S518
65. Why are you late? Delay in treatment

access
PC-1247-06 Delays in TB diagnosis at Parakou,
Benin
S Amidou,1 R Baguidi,2 A Wachinou3 1RESEARCH,
National Programme against HIV, Parakou, 2National
Programme against TB, Parakou, 3RESEARCH, National
Programme against TB, Parakou, Benin. e-mail:
salmaneamidou@gmail.com
Background and challenges to implementation: Delays in

Tuberculosis disease (TB) diagnosis may lead to death
even with appropriate TB treatment or significant decline
in lung function after recovery. The main objective of this
study was to determine the factors related to delay before
TB care.
Intervention or response: A cross sectional study was
conducted among TB patient enrolled in 2011 at the
center of detection and treatment of TB at Parakou. Two
types of delay were defined. Patient delay was related to
the length of time between the onset of the symptoms and
the first contact with a health facility and System delay
related to the period between the first contact with a
health centre and the beginning of the TB treatment. A
standardized questionnaire was use to collect data.
Results and lessons learnt: Sixty six patients were
interviewed and 72.7% had delayed over 21 days. The
burden of delay was due mostly to patients delay which
represented 84.4% of Total delay. The means of Total
delay per patient was 90 169 days, Patient delay mean
was 76 157 days and System delay was 14 28 days.
The 3rd quartiles for System, Patient and Total delays
were respectively 12, 67 and 82 days. Access to health
centre was good (82.5% lived near a health centre), but
TB facilities were not available in the nearest centre from
patients house (90.5%). Most of patient (93.6%) usually
go to health centre first when sick, but also use
alternative options like self-medication, praise or plants
(52.4%). Highest delays were not related to income,
knowledge, access to health centre, nor treatment first
choice when sick but Systems delay above 3rd quartile
(12 days) were more observed among patient who firstly
sought care in public centre than those who chose private
clinic (OR6.4; C.I.5%:[1.5-27.2]).
Conclusions and key recommendations: The main delay
to TB care occurs before contact with the health system.
System delay was mainly due to unavailability of TB
facilities. To reduce delay, the National TB Programme
must improve policies, give more training to health
workers, and expand TB care availability. Populations
education should continue to improve global use of
modern health care facilities as first option.
PC-1248-06 Patient and provider delay in

pulmonary tuberculosis patients: a crosssectional study in Addis Ababa city, Ethiopia
D Assefa Lemma,1 N Kamp,2 E Klinkenberg3 1KNCV
Tuberculosis Foundation, Addis Ababa, Ethiopia; 22. KNCV
Tuberculosis Foundation, The Hague, The Hague,
Netherlands;3RESEARCH, KNCV Tuberculosis Foundation,
The Hague, Netherlands. e-mail: dawita8246@gmail.com
Background: Ethiopia is one of the high TB burden

countries, and is implementing TB control strategies
which are fully aligned with the globally recommended
Stop TB Strategy. High level of public awareness is
believed to influence health seeking behavior positively
and this may contribute to early TB case identification
and hence limit transmission of the disease in the
community. This study aims to assess the extent of
pulmonary TB patient delay and associated factors in
Addis Ababa.
Design/Methods: Cross sectional survey was employed
to interview pulmonary TB patients during their intensive phase of daily supervised treatment. Study conducted in Addis Ababa, in four purposely selected high TB
burden sub-cities from Aug 1- Oct 31, 2014.
Results: A total of 535 pulmonary TB patients (59.5%
being male) were enrolled. Average household size was
3.8 (SD 1.81) and average rooms per house were 2.12
(SD 1.47). The median patient delay was 30 days (IQR
30 - 60) while provide delay was 5 days (IQR 3 15).
Only 2 (0.4%) of PTB patients had correct knowledge on
the mode of TB transmission that said through breathing
when someone cough or sneeze but 109 (20.5%) PTB
patients said by sharing household utensils and 422
(79.2%) eating together from the same plate. When
asked about the practice at the start of TB related
symptoms, 82 (15.5%) reported they tried home
remedies, 49 (9.2%) reported to have visited a pharmacy
while 18 (3.4%) sought care from traditional healer.
When PTB patients requested what made them to seek
health care at health facility, 435 (81.6%) reported self
and advise from family members while only 17 (3%) said
the community Health Extension Workers (HEW).
Conclusion: Patient delay is a problem for TB control
and majority of PTB patients sought health care by
themselves or family advice. PTB patients knowledge on
the mode of TB transmission in an urban setting like
Addis Ababa was found to be very limited. Community
TB care implementation through HEWs in order to
improve public awareness and health seeking behavior
need to be strengthened.
Table 1: Who informed PTB patients to seek care at health
facility
Who made you seek care?
Myself and family member
Community HEWs advice / refer
Civil society organization members
Other, e.g. someone I know, etc
Total
Number / percentage
435
17
4
77
533
(81.6%)
(3.2%)
(0.8%)
(14.5%)
(100%)
PC-1249-06 TB treatment delays in Odisha,

India: is it expected even after so many years of
RNTCP implementation?
K Ilangovan,1 S Burugina Nagaraja,2
R Ananthakrishnan,3 A Jacob,4 J Tripathy,5 D Tamang4
1
Health Systems Research India Initiative,
Thirvananthapuram, 2ESIC Medical College and PGIMSR,
Bengaluru, 3Resource Group for Education and Advocacy for
Community Health, Chennai, 4International Union Against
Delhi, 5School of Public Health, PGIMER, Chandigarh, India.
Fax: (914) 428 173 947. e-mail:
kumaravelilangovan@gmail.com
Background: In India, the Revised National TB Control

Programme (RNTCP) envisages initiation of TB treatment within seven days of diagnosis among smear
positive patients. After nearly two decades of RNTCP
implementation treatment dealys are usually not expected. We conducted this study to determine the proportion
of sputum smear positive TB patients who were initiated
on treatment after seven days and their associated risk
factors.
Methods: The study was conducted in Cuttack and
Rayagada districts of Odisha. It was a retrospective
cohort study that involves reviwes of TB treatment
registers and laboratory registers for 2013.
Results: Among 1800 pulmonary TB (PTB) patients,
1074 (60%) had been initiated on treatment within seven
days of diagnosis, 721 (40%) initiated on treatment more
than seven days and 354 (20%) had delays of more than
15 days. The mean duration between diagnosis and
treatment initiation was 21 days with range of 8-207
days (Median 14days). Odds of treatment delay of
more than seven days were 4.9 times (95% confidence
interval [CI} 3.3-6.6) among those who had been
previously treated, 6.2 times (95%CI 1.3-29.7) among
those infected with HIV and 1.8 times (95%CI 1.1-2.9)
among those diagnosed outside district designated
microscopy centre (DMC).
Conclusion: Delay in initiation of TB treatment occured
in majority of the smear positive patients. The RNTCP
should focus on the core areas of providing quality TB
services with time tested srategies. To have real time
monitoring mechanism for diagnosed smear positive TB
patients is expected to be the way forward.
PC-1250-06 Health-seeking behaviour and

delay in diagnosis of tuberculosis in India
S Nayak,1 P Sharma1 1International Union Against
Delhi, India. Fax: (91) 4605 4429. e-mail:
SNayak.HR@theunion.org

health seeking behaviour of those with TB determines
different degrees of engagement with the public health
system due to many reasons. Thus utilization of health
services can be influenced by availability, distance factor,
economic loss, perceived monetary value of the public
health systems quality services. There is limited infor-
S519
mation in India about the TB patients perspective in

explaining health-seeking behaviour to address the issue
of delay / discontinuity in managing people with TB. To
identify and measure the health seeking behaviour,
challenges faced by presumptiveTB patients in accessing
public health facilities and delay in diagnosis of TB by
using data from Project Axshyas intervention in 30
districts of India from November 2012 and April 2013.
Intervention or response: A cross sectional communitybased survey in 30 districts of India conducted by the
trained field investigators appointed by The Union. The
respondents (4804)including general population, TB
patients(496),health service providers, opinion leaders,
representatives of NGOs identified and interviewed
through a door-to-door survey by using semi-structured
questionnaire.
Results and lessons learnt: Onset of symptoms 27%
preferred qualified private doctor/clinic their first point
of contact. While 73% with symptoms actively sought
health care between 1 week to 1 month after diagnosis.(71%) aware of diagnosis / symptoms, consider TB
a serious disease, heard of free diagnosis, (83%) as
curable. (21%) had cough for , 3 weeks and 9% by 2
weeks. The current strategy of identifying TB suspect,
delays the diagnosis of TB cases (21-9) by 15%. Hospital/
doctors/TV are major source of information related to TB
and treatment is free of cost under DOTS known by
59%. 76% take treatment regularly and 55% about
correct duration of treatment (6-8 months). 81%
respondents visited 2 or more providers prior diagnosis.
39% sought health care from private health care sector
due to a perception that government centres are distantly
located and long waiting hours. The results show that not
only in terms of delays in diagnosis (and finally the
treatment) is attributable to the long distance of
government facility, socioeconomic condition of TB
patient, which force them to continue for daily wage
for family sustenance instead to hospitals.
Conclusions and key recommendations: Increased access
to health care by increasing the frequency of mobile clinic
services and strengthening the community health worker
strategy are possible options.
PC-1251-06 Clinical characteristics and

diagnostic delay in spinal tuberculosis patients
in the Netherlands
D Ijdema,1 C Magis-Escurra,1 P Horsting,2 C Erkens,3
R Aarnoutse,1 M Boeree1 1Radboud University Medical
Centre, Nijmegen, 2Maartenskliniek, Nijmegen, 3KNCV, The
Hague, Netherlands. e-mail:
cecile.magis-escurra@radboudumc.nl
Background: With declining TB incidences diagnostic

delays, especially in extra-pulmonary TB, may increase.
We describe the patient and clinical characteristics of
patients with spinal TB and assessed the course of
diagnostic delays of spinal TB from 2000-2011 in The
Netherlands.
S520
Table Determinants of delay
Deter
minant
Sex*
value
M
F
Age
0-34 ys
Patients
delay
Median in
months
(min-max)
Origin*
NL
Fb
Risk
Yes
factors*
No
Trias of
symptoms*
Yes
No
Neurol
Yes
sympt*
No
Diagnostic
delay
Median in
months
(min-max)
0.75 P 2.25 P
(0- 0.139 (0-24) 0.052
37.5)
1.00
4.00
(0-72)
(0-52)
0.75 P 2.00 P
(0-9) 0.235 (0-52) 0.022
35-64 ys 1.00
(0-72)
. 65 ys
Doctors
delay
Median in
months
(min-max)
1.50
(0-18)
1.5
(0-12)
1.0
(0-72)
1.0
(0-72)
4.5
(0.2538.5)
5.5
(0-76)
4.5
(0.2552)
4.00
5.75
(0.25(0.7527)
76)
3.25
5.00
(0-24)
(0-25)
P
2.5
P
5.0
0.883 (0-24) 0.370 (0-30)
3.0
5.0
(0-52)
(0.2576)
P
3.0
P 5.0 0.750.967 (0- 0.581
46)
12.5)
3.0
5.25
(0-52)
(0-52)
1.0
(037.5)
0.5
P
4.0
P
(0-35) 0.483 (0-27) 0.050
1.0
2.0
(0(0-24)
72)
0.88 P 3.75 P
(0-18) 0.049 (0.25- 0.403
24)
1.0
2.25
(0-72)
(0-13)
P
0.004
P
0.006
P
0.665
P
0.889
5.50
(0.546)
5.0
(0-76)
P
0.582
4.25
(0.2524)
5.5
(0-76)
P
0.070
*Mann-Whitney test, #Kruskal Wallis test, Fb foreign born, Trias of

symptoms night sweats, weight loss and back pain, Neurol sympt
neurological symptoms at presentation.
Design/Methods: Data from the Netherlands Tuberculosis Registry were studied, completed with basic
demographic data and data considering patients-, doctors- and total diagnostic delay retrieved from the patient
records at the public municipal health services.
Results: A total of 274 cases were studied. Median
diagnostic delay was five months and stable during this
period. Sex and age groups were associated with
significant differences in diagnostic delay (male 4.5 vs.f
emale 5.5 months), and 4.5-5 months in the youngest age
group and persons .65 years but 5.75 months in patients
aged 35-64 years . No difference was observed between
origin of patients, patients presenting with TB risk
factors or with neurological symptoms. Typical TB
symptoms at presentation lead, surprisingly, to significantly increased doctors delay (typical symptoms 4.0 vs.
no typical symptoms 2.0 months, P 0.05).
Conclusion: Considering spinal TB diagnosis and act
expeditious is necessary to limit the time to diagnosis in
spinal TB. Refresher courses should be offered both to
family physicians and clinical specialists in The Netherlands.
PC-1252-06 Barriers to accessing rapid secondline treatment initiation for drug-resistant TB

in South Africa
L Dickson-Hall,1 C Tomlinson,1 H Cox,1 J Furin,2
A Vant Hoog,3 A Grant,4 C Oosthuizen,1 M Nicol1,5
1
University of Cape Town, Cape Town, South Africa; 2Harvard
Medical School, Boston, MA, USA; 3University of Amsterdam,
Amsterdam, Netherlands; 4London School of Hygiene &
Tropical Medicine, London, UK; 5National Health Laboratory
Services, Cape Town, South Africa. e-mail:
lindydickson4@gmail.com
Background: Following the roll-out of the Xpert MTB/

RIF rapid test in South Africa, early reports suggested
that the number of patients subsequently initiating
treatment and the time taken for this to be achieved fell
short of what was expected. We aimed to assess health
system and patient level barriers that may impact on
treatment access and time to RR-TB treatment initiation.
Methods: A mixed methods approach was used to
identify factors associated with time to RR-TB treatment
initiation in South Africa. 40 successive cases with RR TB
were identified within National Health Laboratory
records in early 2013 from selected urban and rural
areas in each of 3 provinces. Cases were classified into
those that initiated treatment rapidly (within 0-30 days),
delayed treatment (31-180 days) and did not initiate
treatment (.180 days, if at all). Semi-structured interviews were recorded and transcribed from patients,
family members or friends and health care workers, who
were presumed to have insight into the patients care
initiation. The data was manually analysed for themes
denoting barriers, enablers and missed opportunities for
care linkage and more rapid treatment initiation.
Results: Among the 91 patients included, 78 cases
initiated treatment at a median of 29.5 days after the
diagnostic specimen was taken; 24 patients initiated
treatment within 10 days, 45 between 11-60 days, 9
between 2-6 months. 13 patients had not initiated
treatment by 6 months. Factors most commonly mentioned as barriers to care from a health care perspective
included: Difficulties in accessing and actioning laboratory results, challenges tracing patients for result
provision due to administrative deficiencies, complexities in referring patients to centralised facilities and
patient socio-economic situations. Factors most mentioned as barriers to care from a patient perspective
included: Low levels of family or staff support, socioeconomic challenges, and having to leave the household
or work place to initiate treatment in a hospital or at a
distant site. Enablers and missed opportunities most
mentioned related to communication between different
health care providers and efforts made advising and
encouraging patients.
Conclusions: A laboratory record showing RR-TB does
not guarantee access to treatment. Exploring barriers to
care, enablers and missed opportunities yields important
insights which can be used to direct improvement
initiatives involving treatment access and time to
treatment.
PC-1253-06 Why they said they got tested for

TB
D Skinner,1 M Claassens2 1Research on Heatlh and Society,
Cape Town, 2Desmond Tutu TB Treatment Centre, Cape
Town, South Africa. e-mail: dskinner@sun.ac.za
Background: Patients may be infectious with TB for an

extended period for they begin treatment. Early testing
for TB is crucial for controlling the epidemic, especially
in communities where there are high existing levels of
disease such as in South Africa, which has one of the
highest rates of TB internationally. It is particularly
important now with the overlapping HIV epidemic and
the increasing prevalence of drug resistant TB. The
objective of the study was to understand why people who
had presented to be tested for TB initially opted to do this
at the time that they were tested and the context in which
this occurred.
Design/Methods: This study was nested in a larger study
looking at patients who do not present for treatment.
Depth interviews were done with 41 patients across five
provinces in South Africa. Generally only one patient
was interviewed from a clinic and the sample covered
patients who had taken their treatment regularly and
others who did not. All interviews were recorded and
then transcribed and translated into English. A contextualised interpretative content analysis method was
used.
Results: The reason why people present themselves for
testing included an awareness of having symptoms of TB
and wanting to protect family. Knowledge of TB was
high in the community. Unfortunately many wait until
severe symptoms start and thus remain infectious for
longer. There were concerns about stigma and the
problems of being on treatment especially having to visit
the clinic daily for DOTS. Many are tested as part of
screening programmes run in the community or when
they come to the clinic for a different purpose. While this
is an important intervention in the epidemic, some
respondents reported being unclear of the nature of the
tests being done and the implications of the test. A
number claimed that they were not told what the tests
were for or felt forced into being tested. Some expressed
shock when they were given the results of the test.
Conclusion: Patients do present for testing, but despite
high levels of knowledge of the symptoms, do not take
fully into account that they may be infectious in the early
part of their illness. The screening was accepted as an
important intervention, but a number did feel that they
were not respected in the process. More education is
needed on the need for testing and to encourage testing
before severe symptoms develop.
S521
PC-1254-06 Diagnostic delay and associated

factors among patients with pulmonary
tuberculosis in Dar es Salaam, Tanzania
K Ibrahim,1,2,3 J Hella,1,2,3 F Mhimbira,1,2,3 L Fenner,1,2
T Maroa,3 M Chiryamkubi,4 T Maroa3 1Swiss Tropical and
Public Health Institute, Basel, 2University of Basel, Basel,
Switzerland; 3Ifakara Health Institute, Dar Es Salaam,
4
National TB and Leprosy Control Programme, Dar Es Salaam,
Tanzania. e-mail: ksaid@ihi.or.tz
Background: Tuberculosis (TB) has a long infectious

period, and therefore early diagnosis and reduction of
undiagnosed TB is crucial to minimize the spread of
disease in the community. We aimed to assess the
diagnostic delay of infectious TB and to identify
associated risk factors.
Methods: We analysed new adult smear-positive TB
patients enrolled in an on-going TB Cohort Study in Dar
es Salaam, Tanzania between 2013 and 2015. TB
patients were interviewed to collect information on their
health seeking behaviours. Diagnostic delay was defined
as the interval from onset of TB symptoms until the time
of TB diagnosis (63 or . 3 weeks). All patients were put
on treatment within the first two days of TB diagnosis.
We used logistic regression models to identify risk factors
for diagnostic delay.
Results: We included data from 452 new adult TB
patients. The median age was 34 years (interquartile
range [IQR] 28-42), 310 (69%) were males, and 148
(33%) were HIV-positive, 64/148 (43%) on ART. More
than half of the patients (251, 56%) lived in rented
houses, median household size was 3 persons (IQR 2-4),
and 354 (78%) of the patients earned ,200 USD per
month. Almost all patients (449/452) reported coughing
at time of diagnosis. Median delay time was 3 weeks
(IQR 3.0-3.0), 24 (5%) had a delay of more than four
weeks. Majority of patients reported taking self-treatment prior to TB diagnosis (403, 89%), amoxicillin was
the first drug of choice (377, 93%) and ampicillin was
taken by 12 (3%). Prior to TB diagnosis, 389 (86%)
reported seeking care from formal heath providers, of
those, 207 (53%) sought care 3 times or more. Only 5
(1%) visited traditional healers. In univariate analysis,
diagnostic delay was significantly associated with higher
income level, smaller household size, no previous visit to
a heath facility, non-use of self-treatment and chest pain
(Table). Age, sex, education level and HIV status did not
show significant association. In multivariate analysis,
higher income level, smaller household size, and no
previous visit to a health facility were predictors of
treatment delay of more than 3 weeks.
Conclusion: In urban Tanzania, diagnostic delay of TB
was more common among patients with higher income
and smaller household size, but less likely in patients
seeking health care multiple times. Consequent systematic screening for TB at facilities and community-level
could further reduce the delay to interrupt transmission
and control of TB in the communities.
S522
Table. Association of socio-demographic and health related

factors with patient delay (defined as three or more weeks)
among 452 pulmonary tuberculosis patients in Dar es Salaam,
Tanzania.
Characteristic
Age
18-24
25-44
.44
Cases
n (%)
OR
(95%CI)
p
value
aOR
(95% CI)
0.4
57 (13)
302 (67)
1
0.42
(0.14-1.21)
93 (93)
0.5
(0.15-1.64)
Sex
Female
Male
142 (31)
310 (69)
1
0.97
(0.60-1.58)
HIV status
Positive
Negative
148 (33)
304 (67)
1
1.11
(0.69-1.81)
0.7
1
0.65
(0.29-1.43)
0.99
(0.39-2.50)
0.9
0.5
1
0.81
(0.46-1.44)
0.66
Occupation
Unemployed 89 (20)
1
Unskilled
60 (13)
1.79
labour
(0.79-4.04)
Semiskilled 303 (67)
1.55
labour
(0.83-2.93)
p
value
0.8
1
1.02
(0.58-1.78)
0.2
0.5
1
1.17
(0.45-3.05)
0.86
(0.39-1.89)
Income class
,0.001
Earns less
177 (39)
1
1
than $120
Earns $120 275 (61)
3.31
3.00
or more
(1.92-5.70)
(1.59-5.66)
0.001
Household size
,0.001
,0.001
1 in the
41 (9)
1
1
household
2-3 in the
292 (65)
0.45
0.37
household
(0.23-0.88)
(0.18-0.78)
74 in the
119 (26)
0.13
0.14
household
(0.05-0.32)
(0.05-0.36)
Visiting health
facility
No visit
58 (13)
1-2 visits 182 (40)
1
0.66
(0.35-1.23)
3 or more 207 (47)
0.20
visit
(0.10-0.39)
Self-treatment
No
49 (11)
Yes
403 (89)
Chest pain
No chest
pain
Chest pain
0.003
,0.001
1
0.95
(0.31-2.94)
0.35
(0.08-1.19)
0.002
1
0.38
(0.02-0.71)
0.6
1
0.77
(0.25-2.40)
0.002
0.4
102 (23)
350 (77)
0.46
(0.28-0.75)
1.39
(0.70-2.74)
OR, Odds Ratio; aOR adjusted Odds Ratio; 95% CI, 95% Confidence
Interval.
All variables in the table were included in the multivariate model.
PC-1255-06 Patients suspecting TB are

diagnosed earlier
S Mutembo,1,2 J M Mutanga,1,2 J Sekandi Nabbuye,1,3
B Mwangelwa,2 C Kanene,4 K Musokotwane,2
J Chinyonga,2 C C Whalen1 1University of Georgia, Athens,
GA, USA; 2Ministry of Health, Zambia, Choma, Zambia;
3
University of Makerere, Kampala, Uganda; 4Center for
Disease Control and Prevention Zambia Country Office,
Choma, Zambia. Fax: (26) 213 221 446. e-mail:
simon.mutembo@gmail.com
Background: Early diagnosis of Tuberculosis (TB) and

treatment are essential for an effective TB control
program. Delay in diagnosis affects both disease prognosis at the individual level and transmission within the
community. This study aims to determine the magnitude
of delay to diagnosis and the associated risk factors in a
rural setting where HIV prevalence is high.
Methods: We conducted a retrospective cohort study of
patients aged 716 years who were treated for TB at
Mazabuka District Hospital in Zambia between March
and July 2014.We administered a questionnaire to TB
patients who were on treatment for less than 6 months.
Descriptive analysis was performed followed by survival
analysis to estimate the time to final diagnosis from onset
of symptoms. We performed a cox proportion hazard
analysis to evaluate patient and health provider factors
associated with time to final diagnosis.
Results: We enrolled 171 participants with a median age
of 34 years (IQR 27, 42), of whom 63% were male. One
hundred and twenty patients (71%) were HIV positive
and 94% of these were on combination antiretroviral
treatment. The overall median time from the onset of
symptoms to diagnosis was 101 days (IQR; 68, 152). The
median time from first contact with health care provider
to diagnosis was 28 days (IQR 14, 56). All patients
received TB drugs the day that the diagnosis of TB was
made. The first point of contact for 84% of the patients
was a family member and 26% first visited either a drug
store or a traditional healer. None of the participants
consulted a health care provider in the initial step.
Patients who suspected that they had TB were 61% more
likely to achieve an earlier diagnosis (HR 1.6; 95%CI:
1.1, 2.3). Time to diagnosis was similar between HIV
negative and positive patients.
Conclusion: The largest delay in diagnosis occurred
before patients reached the health care system but in
patients who suspected that they had TB, the diagnosis
was often made earlier. We recommend community
programs to raise awareness on TB and influence change
of health seeking behavior.
PC-1256-06 Specimen transportation system in

Zimbabwe: an innovative way to overcome
barriers to accessing early TB diagnosis in a
resource limited settings
C Zishiri,1 R Ncube,1 M Nqobile,1 K Charambira,1
C Sandy,2 A Kanjanga,3 M Jonasi,1 P Shiri2 1International
Union Against Tuberculosis and Lung Disease, Harare,
2
Ministry of Health and Child Care, Harare, 3Riders for
Health, Harare, Zimbabwe. e-mail: czishiri@theunion.org
Background and challenges to implementation: Tuberculosis (TB) case detection rate in Zimbabwe has fallen
from 56% in 2000 to 42% in 2013 according to the
WHO 2014 Global TB report. This decline calls for a real
need to strengthen efforts for TB case finding at all levels.
The country is one of the top 22 high burdened countries
which account for more than 80% of the TB cases
globally. Sputum smear microscopy remains the first line
laboratory diagnostic technique for pulmonary tuberculosis (TB) in Zimbabwe. Despite significant investment in
the infrastructure of TB laboratories in recent years, most
rural communities remain with sputum smear microscopy access challenges.
Intervention or response: To overcome barriers to
accessing sputum smear microscopy, TB CARE I in
partnership with Riders for Health launched a specimen
transport (ST) system in June 2010.The system transports sputum and other samples using motorcycles to
bring specimens to the nearest diagnostic centre. This is
currently a daily routine and where large distances are
covered this becomes a weekly routine.
Results and lessons learnt: The system now covers up to
24 districts with 42 motorcycles which serve 649 health
facilities (40%) of all health facilities in Zimbabwe.
There was a fourfold increase in specimens transported
from 38,663 in 2010 to 182,265 in 2014. The proportion
of other non sputum specimens increased from 56% in
2010 to 70% in 2014. This increase represents a notable
overall strengthening of the health systems. The turnaround time for sputum specimen results decreased from
a range of two to three weeks (pre ST period) to a range
of one to seven days with ST. The percentage of TB
patients with sputum not done plummeted from 19% in
2010 to 7% in 2014 and through improved access to
follow up specimens, the cure rate increased from 71% in
2010 to 75% in 2013.
Conclusions and key recommendations: The ST system is
a reliable and efficient system that minimises barriers to
accessing TB diagnostic services, delays in diagnosis and
barriers to follow up sputum specimens for monitoring
treatment response.
S523
66. Gender issues in TB and tobacco

PC-1257-06 Involvement of female community
health workers in BRAC-supported areas for TB
control
S Alam,1 S Reza,1 M Basher,1 M Rana,1 M Akramul Islam,1
M Chowdhury,1 M Haque2 1BRAC, Dhaka, 2NTP, Dhaka,
Background and challenges to implementation: Community based tuberculosis control program in Bangladesh has been implementing by BRAC since 1994 under
the stewardship of National TB Control Program (NTP).
Recently BRAC is working following this model in twothirds of the country covering approximately 93 million
population in both rural and urban areas. Female
community health workers (Shasthya Shebikas) are the
nucleus in TB control activities at community level.
Increase accessibility of TB services and ensuring daily
intake of drugs (DOT) are disputes for TB control.
Intervention or response: Shasthya Shebikas are usually
being selected from BRACs village microcredit organizations. Each Shebika provides essential health care
services to about 250 households. They receive basic
training before starting work and a one-day refresher
training in every month. During the routine household
visits, they disseminate TB information, identify TB
presumptive and refer them for sputum test, provide
home-based DOT and refer TB patients for any sideeffects. Patients are encouraged to deposit taka 200
(US$2.5) as bond money for treatment completions
which is totally refundable. Respective Shebikas receive
taka 500 (US$6) for successful DOT treatment completion of each patient from program as an incentive.
Results and lessons learnt: Currently about 63 810
Shasthya Shebikas are actively involved in TB control
activities in BRAC supported areas and more than 80%
DOT is ensured by these Shasthya Shebikas. In 2014, a
total of 124 286 patients diagnosed in BRAC supported
areas. Of them, 72,231 were new smear positive and case
notification rate was 78 and 134 per 100 000 population
for new smear positive and all cases respectively.
Treatment success rate in 2013 was 95% among new
smear positive patients.
Conclusions and key recommendations: As Shastha
Shebikas are the residences of community, they can
enhance early detection of TB cases and better treatment
compliance. To sustain a favorable treatment outcome
more than 90% is also possible through implementation
of daily DOT by them.
S524
PC-1258-06 Smoking-attributed mortality in

women in Tianjin, China: case-control study of
all adult deaths from 2010 to 2013
PC-1259-06 Gender differences in the

presentation of tuberculosis, Medellin,
Colombia, 2010
W Li,1 G H Jiang,1,2 Z L Xu,1 D Z Wang,1 W Zheng1 1Tianjin

Centers for Disease Control and Prevention, Tianjin, 2Tianjin
Medical University, Tianjin, China. e-mail: liweicdc@126.com
L Villa,1 M P Arbelaez1 1Universidad de Antioquia, Medellin,

Colombia. e-mail: lvillavelez@gmail.com
Objectives: In 2009, Tianjin became the first city which

added questions on smoking of the deceased in the death
certificate in China and Asia. We used the information
from the death certificate to analyze all cause of deaths
attributable to smoking in women age 35-79 years in
Tianjin. To assess smoking induced mortality in women
in Tianjin.
Methods: Death data were recorded from death certificates collected through the Cause of Death Registration
System (CDRS) maintained by the Tianjin Centers for
Disease Control and Prevention (TJCDC), which monitors the entire household population in Tianjin. The
information about smoking was obtained by doctors
through inquiring the relative living together with the
deceased. From 2010 to 2013, 50 412 (current smokers
& never smokers) were included in this analysis (44 145
cases & 98 67 controls). Unmatched multiple logistic
regression was used to calculate odds ratios (OR) of
mortality for current smokers vs. never smokers adjusted
by five year age group, education, year of death and
marital status.
Results: 6.72% of the women death cases were caused by
smoking in Tianjin and smokers lost 5 years of life due to
the habit of smoking. Compared to the non-smokers, the
most extreme OR was in lung cancer group, and it was
highest in 65-79 years group, the OR was 5.22 (4.735.76). The ORs were 3.27 (2.70, 3.96) and 2.13 (1.203.78) in 50-64 and 35-49 years group respectively.
Factors as longer smoking years and plenty daily
cigarette consumption were both associated with risk
of high mortality. The all-cause OR was highest in
smokers aged 50-64 years who consumed 715 cigarettes
per day (2.11, 1.69-2.63) when compared with never
smokers, than those who consumed 1-14 cigarettes
(1.40, 1.13-1.73). Smokers who smoking more than 15
years showed a greater OR than those who smoking less
or on 15 years (OR: 1.63 vs. 1.28).
Conclusions: Results from our study addressed the
problem that smoking had been a major risk factor for
mortality and productivity loss in female adults in
Tianjin and also highlight the early effect the tobacco
control work in Tianjin and the continued need to take
urgent programs to reduce the tobacco use in Tianjin,
and even China, because comprehensive tobacco control
programs and policies have been proven effective for
controlling tobacco use.
Background: In Colombia, Tuberculosis (TB) is still a

public health problem. TB is a relevant cause of mortality
in low-income populations and women (most during
reproductive ages), which generates a bigger social
problem given that many of them are head of household
mothers. The objective of this study is to determine
gender differences in the presentation of tuberculosis
(TB) in a group of patients from Colombia
Design/Methods: The target population was identified
from the databases of a cohort study of cohabitants
conducted in the cities of Medelln, Popayan and Cali
(Colombia) as part of the activities of the Colombian TB
Research Center (CCITB) between 2005 and 2008. The
sample included 492 pulmonary TB index cases and
2816 household contacts
Results: 55.3% of the index cases were men who
correspond to a male/female ratio of 1.2. Most cases of
TB occured in men over 50 years old and women at
younger ages (15-49 year old). More cases of coinfection
with HIV were found in men 13 (4.9%) compared to
women 5 (2.3%). At follow-up 36 incident cases were
found, of which 22 (61.1%) corresponded to women,
with a slightly increased risk among those between 15-24
years of age and a higher incidence of disease (RR 1.55
95%CI 0.5-5.1). Women showed less evidence of BCG
vaccination; and higher incidence of TB compared to
men when they were malnourished, living in overcrowded conditions, had fewer open areas in the household or
perceived to have poor ventilation in it. There was delay
in the diagnosis (56 days) and treatment (12 days) in
men, while for women was 38 and 2 days respectively.
Symptoms like cough and expectoration were found to
be more common among women. More cases of mixed
TB in men and extra-pulmonary TB were found in
women; additionally, women possessed higher bacterial
load and higher co-morbidity with diabetes mellitus.
Conclusion: This study is relevant for TB control
programs in order seek, treat and follow the cases from
a patient-centered care perspective, because it is focused
on vulnerable and special populations. In the local
context must be considered the fact that TB incidence
is higher in younger women and older men, and the risk is
associated with gender and age. Additionally, in Colombia the burden of TB significantly affects women in their
reproductive age. Moreover comorbidity of diabetes
must be addressed because their risk of developing TB
becomes higher and requires active search strategies
especially in their family environment.
PC-1260-06 TB control amongst brothel- and

street-based female sex workers: findings from
the Western sex corridor, Maharashtra
O Bera,1 M Chandge,2 S Kamble,3 S Nayak,1 S Mohanty,1
R Deshmukh,4 S Chadha,1 P Banuru Muralidhara1
1
South-East Asia Office, New Delhi, 2District TB Office,
Mumbai, 3State TB Office, Pune, 4World Health Organization
South East Asia Office, New Delhi, India. e-mail:
Background: Female sex workers (FSW) are usually

transient, hard to reach group, often new migrants living
in crowded conditions coming in contact with number of
clients and associated TB-HIV co-infection amplifies the
risk of spread of tuberculosis (TB). FSWs risk of
acquiring HIV is reduced in recent years; however, the
risk of getting TB infection has remained the same. The
intervention study was designed with the objective of
assessing awareness level of tuberculosis and impact of
communication intervention among brothel based and
street based female sex workers of Pune and Sangli,
Maharashtra.
Intervention: At the onset, brothel based and street based
FSWs were mapped and baseline knowledge, attitude
and practice (KAP) survey was done using questionnaire
prepared in local language among randomly selected 200
brothel based and 200 street based FSW in Jan 2014. It
was followed by interventions which included intensive
three months door to door brothel campaign and
personal sensitization of street based FSW by trained
NGO health staff using pictorial tool kits developed by
project Axshya (Global fund round 9 Project) imparting
knowledge on TB disease, diagnosis, treatment to 3000
FSW in Pune and 2000 FSW in Sangli and was overall
monitored and supervised by corporation health officials. An observatory period of nine months was
maintained before the end line knowledge and practice
assessment of new batch of randomly selected 200
brothel based and 200 street sex workers was conducted
in Dec 2014.
Results and lessons learnt: Amongst brothel based FSWs
of the 200 interviews in base line survey only 105
(52.5%) heard of TB as compared to 146 (73%) in end
line survey. Significant (P , 0.05) increase of knowledge
was observed in end line as compared to base line in
variables such as cough of two week as major symptom
(71%), sputum testing as method of diagnosis (78.7%),
high chance of acquiring TB in HIV patients (72%), free
treatment at all government health facilities (47.3%).
Significant positive change of attitude regarding decreased stigma associated with TB (12.3%), increased
sharing of TB Status (46%) and increased acceptance of
TB treatment at Government and NGO health facility
(57.5%). Practice of covering face while coughing or
sneezing (21.2%) and HIV infected FSW to visit health
facility even if cough of one day (48.6%) was significantly higher as compared to baseline study. Street based
FSW were not having any knowledge of cough of two
weeks as main symptom of TB, contact examination,
DOT as treatment of TB during baseline study. After the
S525
intervention, there was significant change in knowledge

of cough of two week as major symptom, sputum testing,
increased chances of having TB in HIV infected, DOTS as
preferred treatment and treatment is free in government
health facilities. It was disappointing that the attitude of
status sharing and stigma attached didnt change much.
Table: Comparative analysis of baseline and end line KAP
survey among Female Sex Workers (Brothel based and Street
Based)
Brothel Based FSW
Variables
Street based FSW
Baseline
End line
Baseline
End line
KAP Study KAP Study KAP Study KAP Study
(Total
(Total
(Total
(Total
200)
200)
200)
200)
N (%)
N (%)
N (%)
N (%)
Knowledge
Heard of TB
105(52.5)
**Cough of two
weeks as
symptom of
TB
38(36.2)
**TB spreads
when a person
having TB
cough or
sneezes
25(23.8)
**Sputum testing
as method of
diagnosis
32(30.4)
**HIV increases
the chances of
having TB
18(17)
** Knowledge of
TB Contact
examination
26(24.7)
**DOTS as TB
treatment
11(10.4)
**TB treatment
better in
private health
facility
78(74.2)
**TB Treatment
free in nearest
government
health facility.
32(30.4)
** TB is curable
45(43)
Attitude
**TB creates
stigma
42(40)
**TB Status
sharing
11(10.5)
**TB treatment
in Govt./ NGO
Health facility
30(28.5)
Practice
**Cover your
face with cloth
while coughing 12(11.4)
**Cough of 2
weeks is a
must to attend
health clinics
22(21)
**HIV patient to
attend health
facility even if
cough of one
or more day.
13(12.4)
146(73)
76(38)
104(71.2)
12(15.7)
48(47.1)
111(76)
10(13.2)
52(51)
115(78.7)
16(21)
47(46.1)
2(2.6)
27(26.4)
90(61.6)
21(20.5)
76(52.1)
34(33.3)
105(72)
102(51)
75(51.2)
52(68.4)
54(53)
69(47.3)
108(74)
23(30.2)
43(56.5)
85(83.3)
88(86.3)
18(12.3)
65(85.5)
83(81.3)
68(46.6)
18(17.6)
84(57.5)
57(75)
82(80.4)
31(21.2)
5(7)
43(42)
120(82.2)
37(36.3)
71(48.6)
27(26.5)
** This questions are only for those who have heard of TB
Conclusions and key recommendations: This study is the

first of its kind in bringing the impact of health
S526
communication intervention amongst brothel based and

street based FSWs. The above study demonstrates the
importance of customized active awareness campaign for
high risk populations. Even though there is low
awareness and least health seeking behaviour amongst
FSWs, there is no established health promotion strategy
for this marginalised and vulnerable community. FSWs
have the potential to be moving time bombs of TB-HIV
co-infection and coupled with low socio-economic
conditions and the huge clientele they serve, it is the
right time to create high impact large scale awareness
campaigns coupled with active screening in high risk
populations.
PC-1261-06 TB screening and testing at

antenatal care clinics in Burkina Faso
A Roggi,1 Y Moyenga,2 G Sulis,1 S Gnanou,2 I Zombra,2
2 F Castelli,1 A Matteelli1 1Clinic of Infectious and
S Diande,
Tropical Diseases, University of Brescia (UNIBS), WHO
Collaborating Centre for TB-HIV and TB Elimination, Brescia,
Italy, 2National Tuberculosis Programme (NTP), Ministry of
Health, Ouagadougou, Burkina Faso, e-mail:
al.roggi@gmail.com

Burkina Faso 5520 case of tuberculosis (TB) all forms
were reported in 2013, including 3616 case of smear
positive pulmonary TB (941 women and 2676 men).
Figures on notification rates among pregnant women are
not routinely available in the country. In 2014 the NTP
initiated a project to introduce TB screening and testing
among pregnant women attending antenatal care clinics
(ANC) in the Central region of Burkina Faso. Technical
assistance was provided by the University of Brescia and
financial support by WHO.
Intervention or response: Women attending 55 ANC in 4
Districts were screened for signs or symptoms of active
TB, using a standardized WHO questionnaire. Women
with presumptive TB were expected to give sputum
samples to be analyzed at the National Reference
Laboratory (NRL). At NRL samples were tested for TB
and rifampicin resistance using the Xpert MTB/RIF
platform. A two-day training course for health workers
was performed in all selected ANC prior to the
intervention. Data were recorded on ANC registers and
collected during supervisory visits by NTP staff.
Results and lessons learnt: Between August 2014 and
March 2015, 73 289 pregnant women were seen at
intervention sites and 2035 (2.8%) reported TB signs or
symptoms. However, only 31 samples were collected and
referred for Xpert MTB/RIF analysis. All tests were
negatives for M. tuberculosis. The average time between
sputum collection and analysis was 3.3 days. During the
same period of time no case of TB was notified among
pregnant women in the intervention area, though 33 had
been estimated to occur.
Conclusions and key recommendations: TB is likely
severely underdiagnosed among ANC clinic attendants
in Burkina Faso. Questionnaire screening was feasible
and identified a significant number of women with
presumptive TB, however, only a small fraction of
women with presumptive TB were tested for TB. To

promote appropriate diagnostic procedures in these
cases, additional targeted interventions are warranted
including training, supervision, and staff retention
measures. Regional Health Offices, rather than PNT
should take primary responsibility to integrate TB into
the ANC prevention and care package.
PC-1262-06 Frequency and predictors of oral

lesions in women with smokeless tobacco use:
a cross-sectional study in squatter settlements
of Karachi, Pakistan
M Irfan,1 F Idrees,1 J Khan1 1Aga Khan University, Karachi,
Pakistan. Fax: (92) 213 493 4294. e-mail:
muhammad.irfan@aku.edu
Background: Incidence of oral cancers is rising in

Pakistan, particularly in women who use smokeless
tobacco (SLT). Oral malignancy is usually preceded by a
pre-malignant lesion such as leukoplakia, erythroplakia
etc. Increased incidence of such lesion in women would
indicate a further rise in oral malignancy in future.
Objective: To determine the frequency and predictors of
oral lesions in women with habitual use of SLT.
Methods: A cross-sectional survey was done in January
2015, in two squatter settlements of Karachi on women
aged 15-75 years. A questionnaire was filled after
consent, icluding details of SLT use. The oral cavity of
each subject was examined after rinsing the mouth with
water and pre-malignant lesions, if any, were noted.
Results: Of the 385 women, 277 (72%) were using SLT,
however only 102 consented to take part in the study.
The mean age was 36.6612.4 years. 20 (19.6 %) were
working as housemaids. The starting age was 20.5610.9
years. About 48 (47%) were found to have oral lesions, 8
had 3, 22 had 2 and 48 had at least 1 lesions. The most
common lesions were oral ulcers 42 (41%), leukoplakia
22 (21%), melanotic macules 19 (18%), submucous
fibrosis 5 (4%) and erythroplakia 3(2%). Two patients
were found to have large irregular oral lesions suspicious
of malignancy. 42% (n 43) women had gingivitis and
loosening of teeth. The highest frequency of oral lesions
were found in ghutka users 57.4% (n 27/47), followed
by Naswaar 53.3%(n 8/15), Paan25.9% (7/27) and
Mawa25%(3/12) . 98 (97%) reported at least one of the
family members using some forms of tobacco. 43 women
reported at least 1 child below 10 using SLT. The
frequency of use greater than 3 times /day (OR 2.62,
95%CI (1.17-5.86), the age of initiation below 20(OR
4.173, 95%CI (1.74-10.01) and duration of use greater
than 10 years (OR 4.917, 95%CI (2.07-11.67) were
independent predictors associated with increased incidence of oral lesions.
Conclusion: A high incidence of oral lesions including the
pre-malignant ones among women with habitual use of
smokeless tobacco is a cause of concern as it is an
indicator of increased morbidity and mortality in
Pakistan in the future due to oral cancer. Appropriate
measures are required to address this issue.
PC-1263-06 Gender profile of TB screening and

diagnosis in Ugu District, South Africa
L Oldja,1 R Forse,1 A Mohiuddin,1 S Khowaja,2 A Khan,1
L Page-Shipp3 1IRD SA, Johannesburg, South Africa; 2IRD,
Karachi, Pakistan; 3The Aurum Institute, Johannesburg,
South Africa. e-mail: loldja@gmail.com
Background: The primary driver of the TB epidemic in

South Africa is HIV/AIDS, with over 60% of TB patients
being co-infected. In this setting, young women face the
highest risk of HIV, which consequently feminizes the TB
epidemic compared to other settings. The female-to-male
TB incidence ratio in South Africa (0.83) is among the
highest in the world, yet males still bear the majority of
the disease burden.
Design/Methods: Between February and mid-April
2015, systematic TB symptom screening was implemented at 14 high-volume Department of Health facilities in
Ugu District, KwaZulu-Natal, South Africa. Screeners
used a 4-symptom verbal questionnaire to identify people
to send for testing among patients and attendants over 5
years of age. All samples were transported to the
National Health Laboratory Service (NHLS) for testing
with the Xpert MTB/RIF assay. Each step of the TB care
pathway was analysed, disaggregated by age and gender.
Results: Twice as many females (6335) as males (3108)
were screened, yet males were more likely to be
symptomatic (41% of males vs.32% of females, P ,
0.0001). Adolescent (5-15 years) females were the only
age group in which more females among those screened
reported symptoms than males (60% of females vs.48%
of males, P 0.0343). There were no significant
differences in sputum submission rates, transport to
NHLS or lab-reported errors (e.g. not leaked, insufficient, etc) between genders. Among valid, returned test
results, MTB-positivity in males was two-times higher
than that of females (14% vs. 7%, P , 0.0001). Even
though twice as many females were screened, 48% more
bacteriologically confirmed pulmonary TB patients were
detected among males. Of these, 17% of male cases and
9% of female cases were MTB/Rif, although these
findings do not show a significant difference between the
groups. Among patients without prior self-reported
history of TB diagnosis and treament, 60 males (14.6%
of Xpert MTB/- results returned) and 44 females (6.9%
of Xpert MTB/- results returned, P , 0.0001) were
found to be positive for TB. Male Female Indicator N
(%) N (%) P Number Screened 3108 (33%) 6335 (67%)
Number Presumptive TB 1267 (40.8%) 2038 (32.2%)
,0.0001 Number able to produce sputum for Xpert
testing 847 (66.9%) 1261 (61.9%) 0.0041 Number of
samples collected & sent for testing on Xpert 842
(99.4%) 1254 (99.4%) 0.9999 Number of Xpert results
obtained (any) 611 (72.6%) 877 (69.9%) 0.2020
Number of Xpert results (MTB /-) 501 (82.0%) 714
(81.4%) 0.7858 Number of cases detected with TB SS/B
70 (14.0%) 47 (6.6%) ,0.0001 Number of Rif TB SS/
B cases detected 12 (17.1%) 4 (8.5%) 0.2728.
Conclusion: Early and frequent engagement with the
health system is critical to combat TB-HIV. The higher
numbers of women screened indicates gender-based
S527
differences in accessing care, due in part to utilizing

maternal and child health services and experiencing less
stigma when accessing care. Despite higher rates of HIV
in young women, men continue to present higher TB
positivity rates, which may reflect delays in patient careseeking. This pattern can be generalized across both new
and retreatment cases. The gendered barriers to care in
South Africa need further exploration to improve early
TB case detection.
PC-1264-06 Gender differences in treatment

completion of tuberculosis patients in Kampala
City, Uganda
A Etwom,1 D Mumpe Mwanja,2 M Kenneth,1 D Sama,1
D Lukoye,1 D Kimuli,1 W Amutuhaire,3 M Ruhweza,1
P G Suarez1 1Management Sciences for Health, Kampala,
2
Ministry of Health, Kampala, 3Kampala Capital City
Authorities, Kampala, Uganda. e-mail: etwoma@yahoo.com
Background and challenges to implementation: Gender

disparities in tuberculosis (TB) treatment outcomes have
been reported worldwide. Though socio-cultural factors
have been implicated elsewhere, the disparities in
treatment outcomes by gender for patients in Kampala
had not been documented before. An analysis of TB
treatment outcomes was conducted to compare treatment completion by gender among patients treated for
tuberculosis in Kampala.
Intervention or response: KCCA health team with
support from the USAID funded TRACK TB Project
conducted a retrospective study to determine gender
associated factors influencing treatment outcomes of
newly diagnosed TB patients. Records of patients started
on anti TB treatment from July to December 2012 in 15
public (PHF) and 35 private (PFP) TB Diagnostic and
Treatment Units (DTUs) in Kampala were extracted from
the DTU databases. Gender based comparative bivariate
and multivariate logistic regression analysis by Age, HIV
Status, Ownership of treating facility and DOT status
was done.
Results and lessons learnt: A total of 4142 records were
extracted, of which 1609 (38.8%) belonged to female
patients. Overall treatment completion for the cohort
was 75% and the variation by gender and other
parameters is illustrated in Figure 1. A comparison of
treatment completion by gender showed that females
from PFP had better odds ratios (OR) of successfully
completing treatment compared to females in the PHF
(OR 1.43, P , 0.001). Also males from PFP were more
likely to successfully complete treatment compared to
males in the PHF (OR 1.39, P 0.019). Pulmonary
Bacteriologically Confirmed (PBC) females and males
had better chances of successfully completing treatment
compared to those in other diagnostic categories.
Females on DOT had better treatment completion (OR
1.82, P , 0.001) compared to males (OR 1.60, P ,
0.001) under same conditions. Furthermore, HIV negative females had better odds (OR 1.63, P , 0.001, OR
1.57, P , 0.001) of completing treatment compared to
HIV negative males.
S528
Conclusion: Females diagnosed with TB have better odds

of completing treatment compared to their male counterparts despite case notification being higher among
males.TB patients accessing treatment from PFPs are
more likely to complete treatment compared to those
receiving care from PFPs.We recommend further qualitative research to establish the determinants of these
outcomes and to inform interventions to improve
treatment completion among males and patients attending PFPs in urban settings.
PC-1265-06 Capacity building among women in

tobacco control
B I Chitindi,1 B Chipopo2 1Advocacy, Tobacco Free
Association of Zambia, Lusaka, 2Advocacy, Zambia Heart and
Stroke Foundation, Lusaka, Zambia. e-mail:
bichitindi@yahoo.com
Background and challenges to implementation: Due to

lack of policy and inadequate enforcing mechanism for
existing laws, control of tobacco use is nonexistence in
Zambia. Zambian population can substantively reduce
and not initiate tobacco use with a comprehensive
community prevention programs. Communities influence social norms and reach people where they live,
work, and play, learn, associate and worship. In line with
Zambias traditional values, women in the community
face higher health risk factors and costs than men, due to
their greater use of health care, yet they are more likely
than their male counterparts to be poor and unemployed.
Women are major contributors to health systems through
their roles as primary care givers in the family and are a
backbone and are exposed to greater occupational health
risks in their roles as informal health care providers in the
communities. They are unremunerated in their voluntary
work, but have a voice that can make change.
Interventions or response: Women health and the
participation of women in health systems has been a
concern for World Health Organization. The life-course
approach reveals the importance of womens multiple
contributions to society, in both their productive and
reproductive roles. In 2012 women in four communities,
three in urban and one in rural areas of Zambia, a total
number 320 women were mobilized, trained and
empowered with the knowledge in tobacco control to
be peer educators in communities. Women were eager to
take responsibilities to form support groups to advance
tobacco control in the nation of Zambia.
Results and lessons learnt: After acquiring knowledge,
women formed support groups voluntary, created
tobacco control corners in health centers. Sensitized on
the harmful use of tobacco. The program has voluntarily
spread to other parts of the country. The International
Tobacco Control (ITC) Zambia Wave 1 Survey report
indicates that 82% of tobacco users in Zambia are
supportive to stronger tobacco control policies and they
are ready to support the policies if implemented.
Conclusion and key recommendations: From both
women training and ITC Zambia Wave 1 Survey report,
the Zambian population are aware of the harms of
tobacco use and that government should play a stronger

role in tobacco control. Policy makers should create a
stronger and more comprehensive tobacco control
programs and ensure that they are supported, implemented and enforced.
67. The spectrum of life: miners, drug

users, the elderly
PC-1266-06 Active case finding for tuberculosis
among elderly individuals (55 years and older)
in Tboung Khmum Operational District,
Cambodia
S Scholtissen,1 O Camelique,1 M Bonnet,2 M Bastard,2
C Hewison,1 A Palomares,1 D Dowdy,3 I Chikwanha1
1
`
Medecins
Sans Frontieres
France, Paris, 2Epicentre, Paris,
France; 3Johns Hopkins Bloomberg School of Public Health,
Baltimore, MD, USA. e-mail: sofschol@yahoo.com
Background: If declines in tuberculosis (TB) prevalence

in Cambodia are to be accelerated, novel approaches are
required to identify populations with high prevalence of
undiagnosed TB, such as the growing number of elderly.
As shown in the national TB prevalence survey, older
patients are also under-diagnosed as they do not present
themselves to health facilities. We therefore sought to
screen a population of age 55 years or more using chest
X-ray (CXR) and TB symptoms, with an ultimate goal to
develop an appropriate Active Case Finding model for
this population.
Methods: We targeted a total of 17 health centers (HC) in
two phases. The first phase screened the elderly
population of 6 HCs at a single site, whereas the second
phase used decentralized screening (8 sites for 11 HCs).
Participants meeting any of the following definitions had
a sputum specimen collected for GeneXpert MTB/RIF
(Xpert) and/or culture testing: (a) cough 7 2 weeks; (b)
cough , 2 weeks with any TB symptoms (sputum/
hemoptysis, fever, weight or appetite loss, night sweats,
lymph nodes); (c) CXR abnormal possibly or likely active
TB (grade 2 or 3); or (d) for phase 2 only, sputum
production/hemoptysis with no cough. All TB cases were
referred for treatment.
Results: In the first phase (centralized), we screened 4098
individuals from a target population of 9517 (43%
participation). Among them, 1352 (33%) had a positive
screening, of whom 831 (62%) received Xpert testing. A
total of 125 TB cases were identified, which is 3.1% of
the screened population: 52 bacteriologically confirmed
(1.3%, 47 by Xpert, 5 by culture), 68 bacteriologically
negative, and 5 Extra-pulmonary TB (EPTB). In the
second phase (decentralized), 8237 participants were
screened from a target population of 13 994 (59%
participation). Among them, 3091 (37.5%) had a
positive screening, of whom 3048 (99%) were sent for
testing. So far, 158 TB cases were identified, which is
1.9% of the screened population: 92 bacteriologically
confirmed (1.1%, 79 by Xpert, 13 by culture), 65
bacteriologically negative (6 Xpert negative with cultures

ongoing), and 1 EPTB.
Conclusions: Active Case Finding among elderly in
Cambodia identified a large proportion of individuals
with active TB, who may be sources of ongoing
transmission in the community. A decentralized approach improved participation. The large-scale feasibility and cost-effectiveness of screening older population in
Southeast Asian settings merit further investigation.
Table 1 - Proportion of Pulmonary TB cases bacteriologically
confirmed among those tested by GeneXpert and distribution of
screening criteria. ACF phase 1 & 2.
X-RAY SCREENING*
SYMPTOMS
SCREENING
PHASE 1 (n47)
Cough 7 2 weeks
Cough , 2 weeks
Any TB
symptoms**
No cough
Sputum/
hemoptysis
Negative
Total
PHASE 2 (n79)
Cough 7 2 weeks
Cough , 2 weeks
Any TB
symptoms**
No cough
Sputum/
hemoptysis
Negative
Total
Abnormal
unlikely
active TB
(Grade 1)
Abnormal Abnormal
possibly
likely
active TB active TB
(Grade 2) (Grade 3)
Total
2/157
(1.3%)
1/120
(0.8%)
5/109
(4.6%)
1/69
(1.4%)
13/36
(36.1%)
3/11
(27.3%)
20/302
(6.6%)
5/200
(2.5%)
1/84
(1.2%)
2/43
(4.7%)
4/12
(33.3%)
7/139
(5.0%)
0/50
4/411
(1.0%)
4/150
(2.7%)
12/371
(3.2%)
11/35
(31.4%)
31/94
(33.0%)
15/235
(6.4%)
47/876
(5.4%)
2/258
(0.8%)
2/353
(0.6%)
7/218
(3.2%)
5/198
(2.5%)
18/42
(42.9%)
15/38
(39.5%)
27/518
(5.2%)
22/589
(3.7%)
0/78
3/33
(9.1%)
2/5
(40.0%)
5/116
(4.3%)
0/7
16/688
(2.3%)
31/1137
(2.7%)
9/75
(12.0%)
44/160
(27.5%)
25/770
(3.3%)
79/1993
(4.0%)
4/696
(0.6%)
* Normal X-rays were not integrated in this table as no TB case was found in
this group of participants.
** Sputum/hemoptysis, fever, weight loss, loss of appetite, night sweats,
lymph nodes.
PC-1267-06 Identification of rifampicin

resistance among injecting drug users at entry
screening in Cipinang DC: where will it lead?
Y N Sumarli,1 E Nurhidayat,1 Y I Hiola,2 U Salamah,3
N Tandirerung,4 M Samsuri,5 M R Christian6 1Cipinang
Detention Center, Jakarta, 2Provincial Office of Law and
Human Right, Jakarta, 3Ministry of Law and Human Rights,
Jakarta, 4Ministry of Health, Jakarta, 5Family Health
International 360, Jakarta, 6World Health Organization,
Jakarta, Indonesia. Fax: (62) 218 591 1415. e-mail:
dr.yulius.st@gmail.com
Background and challenges to implementation: A study

in Indonesia in 2011 has found that people with HIV
who injected drugs were 85% more likely to have HIVassociated tuberculosis than those who were not injecting
drug users. Drug use has been associated with higher
prevalence of latent TB infection (LTBI) and incidence of
S529
TB disease. Over capacity is also a risk to accelerate TB

transmission. Cipinang detention center has capacity of
1136 inmates but was occupied by 3487 inmates. In
2014 there were 27 TB Patients and 117 HIV patients.The high turnover of the prison population, between
prisons and to the wider community, is a major
challenge. This facilitates transmission and consequently
the spread of both drug-sensitive and drug-resistant
forms of TB.
Intervention or response: To have an overview of
injecting drug users burden among new inmetes and
enhance the process of early detection of drug resistant
TB, a structured questionnaire of Injecting drug profile
was integrated in the entry screening form. All new
inmates will be screened upon entry for TB and HIV
using this questionnaire. Inmates with smear negative
and HIV positive will be further screened by geneXpert.
Results and lessons learnt: In 2014, from 4432 new
inmates,we identified 106 PWID, of whom all were
tested for HIV, and 80 (75.6%) were HIV positive and 26
(24.4%) were HIV negative We performed sputum
examination to those with HIV positive, and we found
6 were positive for TB. From those with negative result,
we performed 28 GeneXpert testing and found out that
24 were negative for MTB, and 2 were MTB Positive
Rifampicin susceptible and 2 were MTB Positive
Rifampicin Resistant. In total, out of 106 PWID there
were 8 (7.5%) TB cases and 2 (1.9%) were RR TB
patients.
Conclusions and key recommendations: Through a
comprehensive entry screening, Cipinang DC managed
to identify PWID with Rifampicin resistant early. The
barriers to access a comprehensive care for Rifampicin
resistant Tuberculosis may lead to the high prevalence of
MDR-TB among drug users in Cipinang DC as well as
the physiological effects of drug use, along with the
environment and risk behaviors of drug users. Therefore
a specific intervention and in Cipinang Detention should
be addressed.
PC-1268-06 Characteristics of tuberculosis

among elderly TB patients in Kenya
R Muthee,1 D Nyangahu,1 B Langat,2 R Kiplimo,1
M Githiomi1 1National Tuberculosis, Leprosy and Lung
Disease Program, Nairobi, 2National Treasury, Nairobi, Kenya.
e-mail: mutheewanjiru@gmail.com
Background: Aging is an emerging risk factor for

Tuberculosis (TB) disease. The great longevity of tubercle
bacilli permits them to remain viable within a healthy
host for many years. Majority of TB in the elderly is
secondary to reactivation of latent TB infection to active
TB. Factors contributing to this include age-associated
diseases and poor nutrition. The elderly account for a
large proportion of TB cases discovered at autopsy
illustrating the difficulty of clinical diagnosis in this age
group. This paper identified the characteristics of elderly
TB patients (55 years and over) notified to the TB
program and compared their treatment outcomes to
S530
assess whether there are any peculiar factors fueling the

TB burden in this cohort.
Design/Methods: A retrospective study based on the
Kenya National TB Surveillance data for 2013 and 2014
(Q1-Q3). The study population was all elderly TB
patients registered during the period. Descriptive statistics were generated and v2 tests were applied to establish
if there was any association among the paired variables
using SPSS.
Results: In 2013, 11.2% of all TB cases were the elderly
while in 2014, this increased to 14%. Male preponderance was noted which accounted for 63.5 % of all cases.
81 % of the cases notified from the whole country had
pulmonary TB. In 2013, Uasin Gishu County recorded
the highest (38%) number of Extra Pulmonary TB cases
while in 2014 Nairobi County recorded the highest
(37%). Overally, 21% of all the elderly cases received
food support. This was more than half of the moderately
and severely malnourished. 19.5% of all the elderly cases
were HIV positive with Homabay and Kisumu Counties
displaying the greatest burden in both years. 30% of all
the elderly TB cases were smear positive. 19.5% of the
cases were from the private sector and faith-based
organizations with 1% from the prisons. Further analysis
on revealed that there was no statistical significance
between treatment outcomes and HIV status (P 0.18).
Also, there was a positive correlation in the treatment
outcome between the elderly and the young.
Conclusion: There was an uneven distribution of elderly
in different counties which elicits the need for targeted
county-specific fact-finding missions to elucidate these
reasons and enact plausible interventions. In addition, an
interplay of factors such as quality of care, health-seeking
patterns, nutritional support and treatment adherence
behavior among the elderly should be explored.
PC-1269-06 Situational assessment of

tuberculosis in mines in Myanmar
S T Aung,1 O Myint,2 T Kyaw,3 E Cooreman3 1National TB
Programme, Nay Pyi Taw, 2Ayeyarwaddy Regional TB Center,
National TB Programme, Pathein, 3TB Unit, Yangon,
Myanmar. Fax: (95) 6742 1201. e-mail:
dr.sta.ntp@gmail.com
Background and challenges to implementation: Tuberculosis is one of the major public health problems in
Myanmar. Estimated TB prevalence and incidence is 473
and 373 per 100 000 population, respectively. Countrywide case notification rate (all forms) was 297/100 000
in 2013. One of the case finding strategies is screening
among high risk groups. Miners were assumed to have
higher TB incidence rates. Mobile teams actively screen
for TB among the mining population. Though this
strategy is not really new to Myanmar, knowledge on
TB prevalence in miners was very limited. The situational
assessment was conducted for successful implementation
and scaling up of the activity. The objectives of the
assessment are; to review the geographic, demographic
and available health services in mining areas, and to
assess TB prevalence in miners, their families and
communities.
Intervention or response: Assessment teams used a

checklist. The teams visited mines in Shan State and
Tanintharyi Region between 2014 and early 2015.
Symptoms screening was done with proforma. People
with broad symptoms compatible with TB were subjected to digital chest X-ray. Sputum was also checked for
those with X-ray compatible with TB.
Results and lessons learnt: Mining areas were hard to
reach and accessibility to quality health services was
challenging. Total attendees to mobile clinics were 2742
including 779 miners. Nearly 90% of miners did not
know their HIV status. Thirty seven miners (4.7%) had
history of TB treatment. Cough was the most common
(though not universal) symptom among symptomatic
people (41.8%). X-ray was taken from 2355 individuals
(including 760 miners). Of them, 1900 (81%) were
considered as normal, 52 (2.2%) active TB, 91 (3.9%)
healed lesions, 127 (5.4%) presumed TB and 185 (7.9%)
other pulmonary disease. Sputum was examined from
active and presumed TB (179 persons). Sputum was
positive in 24 cases (13.4%). In total, 73 people
(including 13 miners) were diagnosed with TB. This
means an overall prevalence of 2.7% and a prevalence of
1.7% among miners.
Conclusions and key recommendations: Multi-sectorial
coordination and collaboration was very important for
undertaking this assessment. TB prevalence in mine
communities was significantly higher than that of the
country. Moreover, mobile team could identify the
hidden TB cases in this risk group residing in hard-toreach areas. Thus, TB screening in mining areas should
target the entire mining population and not merely
miners.
PC-1270-06 TB case finding among the elderly

in a congregate setting
R Fabella,1 C Malibiran,2 R Bunao2 1Impact Project,
Philippine Tuberculosis Society, Inc., Quezon
City2Department of Social Welfare and Development,
Muntinlupa City, Philippines. e-mail: allanfabella@yahoo.com

elderly, especially those in congregate settings, are
vulnerable to tuberculosis. The Philippines has scant
TB data and no institutional policy on TB control for this
population. The setting of this project is a government
facility that provides residential care to abandoned
seniors 60 years old and above. From October 1, 2013
to December 5, 2014, the facility conducted TB case
finding among its 251 residents using developed protocols.
Interventions or response: Supported by USADI/Philippines, protocols were developed to evaluate clients with
TB symptoms and for routine chest X-ray screening. Two
rounds of mass screening, using symptomatology with
DSSM and chest Xray, respectively, were conducted 9
months apart.
Results and lessons learnt: In the initial screening, a total
of 247 residents were screened for TB symptoms, of
whom 12 (5%) had cough of more than two weeks. The
number doubles if cough of any duration was used. From

the same group 136 clients underwent direct sputum
smear microscopy (DSSM); of these, five patients had
positive DSSM and were subsequently enrolled for
treatment. In the second screening, 193 patients underwent chest X-ray and 41 had positive TB findings.
Thirty-five underwent DSSM and one had a positive
result. A physician diagnosed TB in all 41 patients, and
31 cases were registered for treatment. The rest were not
treated because of non-cooperation of psychiatric patients (8), and no consent for treatment (2). An additional
nine patients were subsequently detected and treated
after referral for symptoms or on routine chest X-ray
upon entry. With a total of 53 new TB patients diagnosed
out of 251 clients, computed incidence rate is 20 000 per
100 000 which is 65 times more than community rates.
Rate of bacteriologically confirmed TB was 4% (5/136)
while rate of clinically diagnosed TB through chest X-ray
was 25% (50/205). Even among those with no cough,
rate of TB diagnosis was 15% (33/219).
Conclusion: A mix of approaches to TB case finding in
congregate settings (e.g., use of chest X-ray as a more
sensitive screening tool and lowering symptom threshold
from cough of two weeks to cough of any duration)
could result in more TB cases detected and treated
compared with traditional approaches. This experience
provides evidence for the development of national
policies and guidelines for TB control among the elderly
in congregate setting.
PC-1271-06 Psychological distress and its

relationship with non-adherence to TB
treatment: findings from a multicentre clinical
trial
G Theron,1 J Peter,1 L Zijenah,2 D M Chanda,3 C Mangu,4
P Clowes,4 D Stein,1 K Dheda1 1University of Cape Town,
Cape Town, South Africa; 2University of Zimbabwe, Harare,
Zimbabwe; 3University Teaching Hospital, Lusaka, Zambia;
4
Mbeya Medical Research Centre, Mbeya, Tanzania. e-mail:
grant.theron@gmail.com
Background: The successful cure of tuberculosis (TB) is

dependent on adherence to treatment. Various factors
influence adherence, however, few are easily modifiable.
There are limited data regarding correlates of psychological distress and their association with treatment
adherence.
Design/Methods: In a trial of a new TB test, we measured
psychological distress (K-10 score), TB-related health
literacy, and morbidity (TBscore) prior to diagnosis in
1502 patients with symptoms of pulmonary TB recruited
from clinics in Cape Town (n 419), Harare (n 400),
Lusaka (n 400), Durban (n 200), and Mbeya (n 83).
Socioeconomic, demographic, and alcohol usage-related
data were captured. Patients initiated on treatment had
their DOTS cards reviewed at two- and six-months.
Results: 22% (95%CI: 20%, 25%) of patients had severe
psychological distress (K-10730). In a multivariable
linear regression model, increased K-10 was independently associated with previous TB [estimate (95%CI)
0.98(0.09-1.87); P 0.0304], increased TB score
S531
[1(0.80, 1.20); P , 0.0001], and heavy alcohol use

[3.08(1.26, 4.91); P 0.0010], whereas male gender was
protective [1.47(2.28, 0.62); P 0.0007]. 26%
(95%CI: 21%, 32%) of 261 patients with cultureconfirmed TB were non-adherent. In a multivariable
logistic regression model, reduced TB score [OR
(95%CI) 0.639(0.497, 0.797); P 0.0001], health
literacy score [0.798(0.696, 0.906); P 0.0008], and
increased K-10 [1.082(1.033, 1.137); P 0.0012], and
heavy alcohol usage [14.83(2.083, 122.9); P 0.0002],
were independently associated with non-adherence.
Culture-positive patients with K-10730 were morelikely to be non-adherent (OR2.290(1.033-5.126); P
0.0416].
Conclusion: Severe psychological distress is frequent
amongst TB patients in Southern Africa. Targeted
interventions to alleviate psychological distress, alcohol
use, and improve health literacy in newly-diagnosed TB
patients could reduce treatment non-adherence.
PC-1272-06 Implementing a high volume TB

screening programme in peri-mining
communities in South Africa
F Popane,1 M Radile,1 C Rithuri,1 D Ratema,1 M Dube,1
L Page-Shipp1 1The Aurum Institute, Johannesburg, South
Africa. e-mail: fpopane@auruminstitute.org

collaboration with the National, Provincial and District
Departments of Health, we are implementing a high
volume TB screening programme between April 2014
and March 2016 in peri-mining communities in 6
districts in South Africa; West Rand, Lejweleputswa,
Dr Kenneth Kaunda, Bojanala, Waterberg and Sekhukhune. Peri-mining communities are considered to be
vulnerable to TB. Community members are reached in
their homes, workplaces, at public events, public places
and schools.
Intervention or response: Mobile teams perform a TB
symptom screen using the National TB screening tool
(cough, weight loss, night sweats and fever .24 hours).
Spot specimens are collected where possible or clients are
referred to clinics for sputum collection. Xpert MTB/RIF
testing is done through the NHLS. Screening includes an
HCT offer. Where logistically possible, HCT testing is
done in the field; otherwise clients are referred to clinics
or other partners. Those diagnosed with TB and those
who are HIV positive are referred to clinics for care.
Results and lessons learnt: To date, 253 241 people have
been screened for TB: 17 856 (7.1%) were TB
symptomatic, of which 10 163 (57%) were able to
produce spot sputum, results were received for 9249
(91%); 122 (1.3%) were positive. HCT was offered to
222 113 (88%), 11 296 (5.1%) were performed in the
field; others were referred to the nearest facility. Of HCT
performed in the field; 785 (7%) tested positive.
Strengthening of public- private mix, community engagement and the use of flexible, skilled teams; are
critical programme components.
S532

screening was widely accepted. Challenges include
difficulty collecting quality spot sputum specimens in
the field and operational challenges with in-field HIV
testing. Investigation is required to further understand
the epidermiology of TB in peri-mining communities.
PC-1273-06 Using TB refill sites to determine

the burden of TB amongst ex-miners in
Swaziland to improve community contact
tracin
V Masuku,1 G Mchunu2 1University Research South Africa,
Mbabane, 2National TB Control Program, Manzini,
Swaziland. Fax: (268) 2505 3248. e-mail:
victoriam@urc-sa.com
Background and challenges to implementation: Swaziland has highest TB burden, incidence 1349/100 000
(National Tuberculosis Strategic Plan 2015-2019). Current TB strategy prioritizes mining populations high risk
however, of the over 22 000 Swaziland ex miners, the
number with TB is unknown. Routine TB data collection
tools have not, tracked TB patients by occupational
exposure. There is therefore limited data on the burden
of TB amongst ex miners making it difficult to design
targeted interventions for this population group. Furthermore, contact tracing strategies are not systematic
increasing the risk of TB transmission at household level.
Intervention or response: In March 2015, University
Research South Africa (URSA) worked with the national
TB program (NTCP) to profile TB patients exposure to
mining. A total of 8 high volume TB facilities were
identified and an interview guide was developed. All TB
patients coming to refill medicines over 4 weeks were
interviewed. Data was collected using face to face
interviews during treatment refill to determine number
of TB patients who are miners, ex miners or relatives of
ex miners.
Results and lessons learnt: from March to April 2015, of
1218 TB patients who came for treatment refills, 81
(7%) were ex miners, 8 (1%) current miners and 178
(15%) were relatives of ex miners. TB contacts were
mapped to guide follow up contact tracing at household
level. On the relatives of ex miners with TB, a process is
ongoing to confirm ex miner relative having confirmed
TB or undergone TB treatment. The NTCP has since
revised the TB data collection tools to include occupation.
Conclusions and key recommendations: The NTCP
wishes to continue tracing ex miners with TB and linking
them with community treatment adherence officers to
intensify active case finding amongst high risk populations. With the revision of TB data collection tools in
Swaziland, it will now be possible to track and attain
routine TB data to facilitate this.
PC-1274-06 Addressing TB among people who

inject drugs through the community initiated
treatment intervention in Ukraine
O Denisiuk,1 A Tyshkevych,1 P Smyrnov1 1ICF International
HIV/AIDS Alliance in Ukraine, Kyiv, Ukraine. Fax: (380) 444
905 489. e-mail: o.denisiuk@gmail.com
Background and challenges to implementation: Access to

HIV and TB care for People Who Inject Drugs (PWID) in
Ukraine is limited due to stigma and low awareness. Data
from Ukrainian National AIDS Center suggest that
PWID register for medical care on average 2 years after
positive HIV test. Delay in HIV treatment initiation leads
to immunosuppression and increase of opportunistic
infections such as tuberculosis (TB). In April 2013, the
International HIV/AIDS Alliance in Ukraine (Alliance
Ukraine) started the Community Initiated Treatment.
Intervention (CITI) to improve linkage to HIV care and
treatment for PWID through peer case-management
approach. CITI provides 6 months support to start
PWID on ART. During this time PWID are also provided
with support in regards of TB diagnostics and treatment.
Intervention or response: CITI combines peer-navigation, outreach case-management and community support
to achieve better treatment uptake and outcomes for
HIV-positive PWID. The approach delivers individual
care to a client for the purpose of registering at the
regional AIDS Centre and getting ART prescribed as well
as a comprehensive set of related medical services such as
TB diagnosis and treatment. CITI clients are recently
tested positive or might know their status for a long time
but are not accessing medical services. Most case
managers are former outreach workers from harm
reduction projects.
Results and lessons learnt: 26 NGOs started CITI in
March, 2013 across 11 regions of Ukraine. Since 2013 3073 PWID were registered in AIDS clinics, 2778 clients
started ART with the support of 60 case-managers. The
average time to link PWID to HIV care and treatment is
22 days. Individual data from 3754 clients who finished
CITI support program during 2013-2014s shows next
results: 503 clients have been diagnosed with TB and
started TB treatment, among them 283 started ARV
treatment. 485 clients were provided with TB prophylactics after registration in AIDS clinic.
Conclusions and key recommendations: PWID are at
increased risk of HIV-TB co-infection and are hard to
reach population with medical services. The casemanagement approach, which is the core component of
the CITI, along with HIV services improves PWID
engagement in TB care. Peer-navigation, outreach casemanagement and community support are crucially
important interventions to provide timely TB diagnostics
and linkage to care for HIV positive PWID.
68. Paediatric TB: IPT, contact tracing and

child perspectives
PC-1275-06 Contact screening and expanding
childhood TB diagnosis and treatment of TB in
Bangladesh
S Islam,1 S Reza,1 M Akramul Islam,1 V Begum,2
K Khatun,1 M Haque3 1BRAC, Dhaka, 2World Health
Organization, Dhaka, 3NTP, Dhaka, Bangladesh. Fax: (88)
2988 8026. e-mail: shayla.dhaka@gmail.com
Background and challenges to implementation: Bangladesh is one of the high TB burden countries in the world.
BRAC, a non government organization is jointly working
with national TB control programme to combat against
tuberculosis since 1994. Due to unavailability of
diagnostic facilities and scarcity of pediatricians, diagnosis and management of child tuberculosis is still a
major challenge for the country especially in rural
setting. It is necessary to raise awareness in the
community and increase screening through contact
tracing of household cases within the existing programme for identifying more child TB cases.
Intervention or response: Considering the different
challenges, BRAC started conducting orientation to
pediatrician at tertiary hospitals and district levels. A
basic training is given to health care providers on sign
symptoms of child TB to increase awareness in the
community during mobilization. A contact tracing
register was introduced in the existing programme of
BRAC in late 2013 and contact tracing initiated among
the family members who are in close contact with
pulmonary TB cases and were sent for screening. During
their households visit, BRACs frontline health workers
disseminate basic messages, identify presumptive cases
and refer them to the specific centres. Poor child TB
symptomatic gets financial support from BRAC for
investigation and transportation purpose.
Results and lessons learnt: In 2014, a total of 929
pediatricians were reached through 64 networking
meetings on child TB diagnosis and management at
tertiary and district level hospitals. In addition, 1029
health care providers were trained on child tuberculosis
including basic messages, identification and referral.
There was an increase in child case identification from
2865 in 2013 to 3554 in 2014. In 2014, 9.7% of the total
child TB cases were identified through contact tracing.
Conclusions and key recommendations: Evaluation of
symptomatic contacts may increase identification of
child cases. An orientation session could be planned for
health care providers to improve the contact tracing
skills. The children who are in close contact of
pulmonary tuberculosis patient if even asymptomatic
should be referred for screening and should do follow-up.
S533
PC-1276-06 INH preventive therapy for underfive children in two regions of Ethiopia
Y Tadessse,1 M Nigussie,2 I Jemal Abdulahi,3 D Habte,1
S Negash,1 D J Dare,1 J Degu,1 Y Haile,4 M Aseresa
Melese,1 P G Suarez5 1Management Sciences for Health,
Addis Ababa, 2Amhara Regional Health Bureau, Bahir Dar,
3
Oromia Regional Health Bureau, Addis Ababa, 4United
States Agency for International Development (USAID), Addis
Ababa, Ethiopia; 5Management Sciences for Health,
Arlington, Virginia, USA. e-mail: ytadesse@msh.org
Background and challenges to implementation: A sixmonth course of Isoniazid preventive therapy (IPT) is
among the recommended strategies for prevention of
tuberculosis (TB) in under-five (U5) children who are in
close contact with active pulmonary TB cases. There is
limited experience with implementation of this strategy
in high TB burden settings. The objective of this analysis
was to share the experience of routine implementation of
IPT for U5-children under routine program condition in
two regions of Ethiopia.
Intervention or response: The USAID funded HEAL TB
project in collaboration with the Amhara and Oromia
RHBs introduced contact screening of smear positive
pulmonary TB index cases. Using contact screening as
entry point, we identified and enrolled eligible U5
contacts to IPT. We oriented the health workers on the
importance of IPT, provided job aids, and supplied
recording and reporting formats for routine use. We also
provided direct site level support through regular
mentoring visits and continuing medical education
specifically designed for mid-and low-level health workers. In this analysis we included all children who initiated
IPT during October 2013 June 2014 in 28 health
facilities (7 hospitals and 21 health centers).
Results and lessons learnt: A total of 504 index smear
positive pulmonary TB cases were reported during the
implementation period. There were 282 U5- children
registered as household contacts of the smear positive
pulmonary TB index cases accounting for 17.9% of all
household contacts. Of these, 237 (84%) were evaluated
for active TB and presumptive TB was identified in16
(6.8%) children. TB diagnosis was confirmed in five
children making the overall yield 1.77% (95% Confidence interval, 0.76-4.08%). Of 232 children, 142
(61.2%) received IPT. The proportion who completed
the 6 months INH preventive therapy was 114 (80.3%)
while 28 (19.7%) interrupted treatment, and no child
developed active TB disease while on IPT. The major
factors for interruption of IPT were care taker refusal (n
12), INH drug stock out (n 9) and being lost to followup (n 4).
Conclusions and key recommendations: Tracing infants
and young children who are contacts of infectious TB
cases and offering them chemoprophylaxis was feasible
in the Ethiopian setting. Contact screening among adult
index cases served as a key entry point for identification
childhood TB cases. The IPT completion rate was good
but the remaining gap should be addressed.
S534
Results and lessons learnt: During the study period, there

were 131 contacts (85 household 46 close) of the index
cases, interviewed during internal evaluation exercise. Of
the household contacts, 48 were children, of which 83%
(39) received isoniazid chemoprophylaxis and 2% (1)
was detected with TB, while 65% (24/37 adult household contacts) were evaluated for TB.
Conclusions and key recommendations: None of the 46
close contacts, living in adjoining slum household, could
be evaluated due to unclear operational guidelines to
screen them and social causes. RNTCP could consider in
close contact tracing and evaluation of TB with chest
symptoms of all members living in adjoining houeholds
in urban slums as a policy.
PC-1278-06 TB disease prevalence among

children exposed to drug-susceptible TB
patients: the Indus Hospital TB program
F Amanullah,1 A Malik,1,2 K Asif,1 M Jaswal,1 R Jafri,1
H Hussain1,2 1Indus Hospital, Karachi, 2Interactive Research
and Development, Karachi, Pakistan. e-mail:
amyn.malik@irdresearch.org
PC-1277-06 Close contact tracing of children in

urban slums of Delhi, India
D Kundu,1 S Chandra,1 K Chopra,2 A Khanna3 1World
Health Organization Country Office for India, Delhi, 2New
Delhi Tuberculosis Centre, Delhi, 3Directorate of Health
Services, Delhi, India. e-mail: debashishkundu@yahoo.com
Background and challenges to implementation: As per

the WHO recommendations on contact investigation, all
household contacts and other persons who are in close
contact with TB patients are to be screened for TB. In
India, under the Revised National Tuberculosis Control
Programme (RNTCP) all children less than 6 years of
age, who are household contacts of the family member
suffering with active TB are screened for TB and
provided with Isoniazid chemoprophylaxis once active
TB is ruled out. However, there are constraints observed
in the state of Delhi while screening household contacts,
especially close contact for TB in the urban community
due to social, cultural and behavioural factors.
Intervention or response: We reviewed contact tracing
efforts in the state of Delhi from interviews of sputum
positive patients living in urban slums during various
internal evaluations conducted between the periods from
October 2012 to November 2013. The proportion of
contacts screened for TB above and below 6 years of age;
and proportion of eligible children contacts (below six
years of age) prophylactically provided with Isoniazid
preventive therapy were found.
Background: Children in contact with patients with

pulmonary tuberculosis (TB) are a known high risk
group for TB disease. Investigation of child TB contacts
allows for early TB case detection and effective treatment
including isoniazid preventive therapy (IPT). Children
particularly 0-4 year olds are more likely to acquire TB
infection and progress to TB disease compared to adults.
The objective of this study was to assess the prevalence of
TB disease among 0-4 year old children exposed to
infectious pulmonary TB patients.
Design/Methods: We retrospectively analyzed the findings and outcomes of our TB screening and IPT program
for 0-4 year old children exposed at home to newly
diagnosed bacteriologic positive TB patients from 2011
to 2014. All 0-4 year old household contacts who
presented for evaluation, underwent a history, clinical
examination, CXR and TST to rule out TB disease prior
to initiating Isoniazid preventive therapy.
Results: A total of 256, 0-4 year olds were evaluated for
TB disease, 27% of whom were ,1 year olds. We found
that 90/256 (35%) children evaluated were severely
underweight. We found TB infection (TST/.10 mm) in
5% children at initial evaluation. TB disease was
diagnosed in 18/256 (7%) children, 10/18 (55.5%) of
the children who developed disease were either never
initiated on IPT or did not complete IPT. The remainder
16% (3/18) developed disease during IPT treatment and
28% (5/18) were diagnosed at initial evaluation. None of
the children who completed IPT developed TB disease
during a one year follow-up period.
Conclusion: Our findings highlight the need for an
effective IPT program. Children 0-4 year old who
completed 6 months of IPT did not develop disease,
whereas children in whom Isoniazid preventive therapy
was either not started or not completed were more likely
to develop disease. IPT programs need strengthening to
ensure treatment adherence and completion.
S535
PC-1279-06 Lessons learned from conducting

clinical trials for new TB drugs in children
PC-1280-06 What can visual methods contribute

to paediatric TB treatment experience?
M T Gler,1 M Frias,2 R Reyes,1 J Hafkin1 1Otsuka Manila

Research Center, Makati, 2De La Salle Health Sciences
Institute, Dasmarinas, Philippines. Fax: (632) 754 1450.
e-mail: maricelle-tc-gler@otsuka.us.com
K Abney1 1University of Cape Town, Cape Town, South

Africa. e-mail: kate.abney@uct.ac.za
Background: Delamanid is a new anti-tuberculosis agent

recently endorsed by the World Health Organization for
the treatment of multi-drug resistant tuberculosis (MDRTB) in adults. In order to accurately characterize the
safety and define the dose of delamanid in children,
Otsuka pharmaceuticals launched a phase 1 PK study
(Trial 232) in the Philippines at De La Salle Health
Sciences Institute (DLSHSI) in Cavite Province.Study
Design: Trial 232 is an 18-day phase 1 open-label trial in
which children with MDR-TB received delamanid for 10
days combined with an optimized background regimen
(OBR) followed by OBR alone prescribed according to
the national guidelines of the Philippines. This trial
requires a total enrollment of at least 36 children from
ages 0-17. Older children were enrolled first and
progressively younger children will subsequently be
enrolled as safety and pharmacokinetics are established
among the older children.
Challenges: A number of challenges were encountered
both before and during the implementation of the trial.
The key challenges prior to the start of the study revolved
around recruitment, regulatory approval, and logistical
issues. Although the Philippines has a high prevalence of
TB disease (12.8%) and TB infection (65.6%) among
household contacts of MDR-TB patients, the program
historically has only focused on treating bacteriologically
confirmed disease. Indeed, since the treatment programs
inception, there have been only a handful of children
treated for MDR-TB out of the total population of
children thought to actually have the disease. In order to
implement this trial, DLSHI had to undertake significant
capacity building which included educating study staff
and regional clinicians, networking with the NTP for
identification of suspected cases, and performing recruitment activities in MDR-TB centers in the Metro Manila
and the Luzon areas. Active case finding via contact
tracing of the adult MDR-TB patients in the MDR-TB
unit was also done. Other challenges encountered were
ensuring the acceptability of frequent venipuncture for
PK determination, creating retention materials for each
cohort to maintain interest in study participation,
confirming that schooling would not be affected during
the course of the trial, and reducing the daily burdens on
the broader family caring for the child. Lessons learned:
With our experience for the first 2 cohorts, we have
learned that clinical trials for novel TB drugs in children
are feasible.
Background: There is a move to document and convey

the experiences of people being treated for severe forms of
TB within inpatient treatment facilities. However, childrens experiences do not feature as prominently as those
of adults within clinical settings. One reason is that
children are minors and have very little authority or say in
their individual treatment and living situations. A second
reason is that research methods employed are not always
appropriate when working with children. The study was
premised on trying to apprehend and understand a
childs eye view of inpatient treatment for TB.
Design/Methods: The larger project comprised of one
year of qualitative research facilitated within a childrens
ward in a sub-acute long term stay TB hospital. Data was
collected alongside child patients living and being treated
at Brooklyn Chest Hospital, located in Cape Town,
South Africa. The data presented here was based on
qualitative methods used in research alongside 50 child
patients. Child-centred visual methods included; drawing, photography and sculpting clay. Visual methods
were used to engage child patients in discussion of their
experiences of hospital life, individual treatment and
their subjective understanding of their TB.
Results: Visual methods were a productive means to
generate data in four areas: 1) individual conceptions of
TB, 2) treatment challenges, 3) hospital life and 4)
educational challenges during duration of treatment.
Conclusion: Visual methods are beneficial to research
with paediatric TB patients, and provide a more nuanced
perspective on the challenges children face during
inpatient treatment. Findings are useful and allow for
the development of evidence-based child-centred hospitalisation policy and practice in sub-acute long term stay
hospitals.
PC-1281-06 Outcomes of IPT usage among

children , 5 years old in a TB Reach contact
tracing project in Swaziland
R Golin,1,2 P Ustero,1,2 P Swamy,2 J Glickman,1 K Ngo,2
S Dlamini,1 M Hlatshwayo,1 A Mandalakas2 1Baylor
College of Medicine Childrens Foundation - Swaziland,
Mbabane, Swaziland; 2Baylor College of Medicine and Texas
Childrens Hospital, Houston, TX, USA. Fax: (268) 2404
0214. e-mail: golin@bcm.edu
Background: Swaziland is the epicenter of the HIV and

TB pandemics. Despite national policy recommending
Isoniazid Preventive Therapy (IPT) for all pulmonary
tuberculosis contacts who are under age 5 years (U5) or
HIV-infected there is no data regarding IPT uptake in
Swazi U5 contacts and IPT usage is recognized as suboptimal. Perceived barriers to IPT delivery include
transport fees to healthcare facilities, facility and
radiology fees, frequent drug shortages, and a broad lack
of recognition regarding the importance of IPT. In concert
with our WHO sponsored TB REACH contact tracing
S536
program, Baylor College of Medicine Childrens Foundation Swaziland (BCMCF-SD) promoted and monitored IPT uptake among contacts of index cases starting
anti TB treatment (ATT) at 7 local TB clinics (TBC).
Intervention/Response: BCMCF-SDs contact tracing
enabled cough monitors to identify all contacts eligible
for IPT and make referrals for evaluation and management at the TBC. Barriers to healthcare access were
addressed through home visits, transport reimbursement,
and vouchers for chest X-rays. Analysis of IPT uptake
was limited to U5 contacts as the validity of reported
HIV-status was unknown.
Results and lessons learnt: Of the 4171 index cases
enrolled, 2589 (62%) underwent contact tracing, yielding 9803 reported contacts. 16.4% (1608) of contacts
were U5, including 3.4% (55) HIV-infected, 51.7%
(828) HIV-uninfected, 44.8% (718) HIV unknown.
Despite systematic referral of all eligible U5 contacts,
IPT uptake was low (16% for all 7 TBC; range 1% 36%). Contacts were more likely to initiate IPT if they
were younger (P 0.0012), HIV-negative (P , 0.001),
slept in the same room as the index case (P , 0.001), and
were the child of the index case (P , 0.001). IPT uptake
was not associated with contacts gender, recent household death, completion of a home visit, or presence of a
miner in the household.
Table 1: Comparison of U5 contacts who did and did not
initiate IPT
CHARACTERISTIC
CONTACTS
ON IPT
CONTACTS
NOT
ON IPT
OVERALL [n (%)]
162 (10%)
1446 (90%)
MEAN AGE
[years 6 SD]
SEX
Male [n (%)]
Female [n (%)]
HIV STATUS
Positive [n (%)]
Negative [n (%)]
Unknown [n (%)]
DEATH IN HOUSE
IN LAST 2 YEARS
Yes [n (%)]
No [n (%)]
SLEEP LOCATIONS
Different House
[n (%)]
Same House [n (%)]
Same Room [n (%)]
Same Bed [n (%)]
RELATIONSHIP TO
INDEX CASE
Child [n (%)]
Other [n (%)]
HOME VISIT DONE
Yes [n (%)]
No [n (%)]
FAMILY MEMBER
EMPLOYED AS
MINER
Yes [n (%)]
No [n (%)]
PVALUE*
N/A
2.2661.49
2.6561.44
0.0012
69 (42.6%)
93 (57.4%)
707 (48.9%)
740 (51.1%)
0.13
7 (4.2%)
106 (65.4%)
48 (29.6%)
48 (3.3%) , 0.001
722 (49.9%)
670 (46.3%)
39 (24.1%)
118 (72.8%)
417 (28.8%)
993 (68.7%)
0.22
35 (21.6%)
565 (39.1%) , 0.001
38 (23.5%)
67 (41.4%)
22 (13.6%)
400 (27.7%)
355 (24.6%)
125 (8.64%)
94 (58%)
68 (42%)
480 (33.2%) , 0.001

966 (66.8%)
73 (45.1%)
89 (54.9%)
630 (43.6%)
816 (56.4%)
0.72
48 (29.6%)
404 (27.9%)
114 (70.4%) 1042 (72.1%)
0.65
* Test of significance comparing U5 contacts who did and did not initiate IPT
Conclusions and key recommendations: Uptake of IPT

among U5 contacts is critically low in Swaziland.
Although we identified an association between IPT
uptake and some important factors such as HIV status
and sleep location, there remains much uncertainty
regarding factors limiting IPT uptake that might inform
focused interventions to increase IPT uptake and
adherence. BCMCF-SDs experience emphasizes the
ongoing need to prioritize a national approach for
evaluation and implementation of IPT for this vulnerable
population.
PC-1282-06 Feasibility and yield of hospitalbased screening of household child contacts of

newly diagnosed adult tuberculosis patients
J Kiwanuka,1,2 C Kyakwera,2 N Kyomugasho,3
D Atwine,3 M Nansumba,3 Y Boum Ii,1,3
J Mwanga-Amumpaire,1,3 M Bonnet3,4 1Mbarara
University of Science and Technology, Mbarara, 2Mbarara
Regional Reference Hospital, Mbarara, 3Epicentre, Mbarara,
Uganda ; 4Institut de Recherche pour le Developpement

(IRD), Montpellier, France. e-mail:
Background: Contact investigations of infectious tuberculosis (TB) patients can be an efficient method of
finding new child TB cases and identifying at-risk
children for preventive therapy. However, despite recommendations this is rarely practiced in most lowresource countries. We assessed the feasibility and yield
of hospital-based contact investigation and treatment of
children in household contact with newly diagnosed
adult TB case in south-western Uganda.
Methods: Over a 24-month period, adults newly
diagnosed with smear- and/or culture-positive pulmonary TB were invited to bring children ,5 years in their
households to hospital for assessment. Children were
assessed for active TB or latent TB infection (LTBI) using
a standardised symptom screen, tuberculin skin test
(TST) and chest radiography. Children with suspected
active TB were referred for further evaluation and
treatment. The rest, classified as LTBI or uninfected
were started on a 6-month regimen of isoniazid
preventive therapy (IPT) and followed for 9 months,
with repeat TST at 3 months for initially TST-negative
contacts.
Results: A total of 162 adult index cases were diagnosed
and 133 enrolled. A total of 281 out of 339 (82.4%)
eligible children (median of 2 child contacts per index
case) identified were assessed; 53% female, median age
33 months, 4.5% HIV infected. The median interval
between diagnosis of index cases and screening of
contacts was 17 days (IQR 6, 53 days). Forty three
children (15.3%) were referred for TB treatment. Three
children did not return for TST reading, 102 (36.3%)
were classified as LTBI and 133 (47.3%) as uninfected.
Of 235 children classified as LTBI or uninfected 8 did not
receive IPT; 227 started and 185 (81.5 %) completed 6
months of IPT. Children with low level exposure (a few
hours during the day) to the index case were less likely to
complete IPT (aOR 0.4, 95%CI 0.2-0.9). Two children
on IPT were reclassified as active TB during follow-up,

and 10 out of 133 (7.5%) non-infected children
converted to TST positive on IPT.
Conclusio:n Hospital-based tuberculosis contact investigation was feasible, and yielded a high number of new
cases of active TB and LTBI among children in household
contact with adult TB cases. IPT was effective and well
accepted by the majority of contacts. Repeat TST during
follow-up is important to identify additional LTBI cases.
Home-based follow-up could help to optimize uptake
and completion of investigation and IPT.
PC-1283-06 Symptom-based screening of child

contacts of tuberculosis cases: a prospective
cohort study in Uganda
M Bonnet,1,2 C Kyakwera,3 N Kyomugasho,2 D Atwine,2
M Nansumba,2 Y Boum Ii,2,4 J Mwanga-Amumpaire,2,4 J
Kiwanuka3,4 1Institut de Recherche pour le Developpement

(IRD), Montpellier, France; 2Epicentre, Mbarara, 3Mbarara
Regional Reference Hospital, Mbarara, 4Mbarara University
of Science and Technology, Mbarara, Uganda. e-mail:
Background: Tuberculosis (TB) screening and isoniazid

preventive therapy (IPT) of young child contacts of adult
TB cases can reduce the risk of active TB in children.
However, this is poorly implemented in high burden
countries partially due to difficulties in screening
children for active TB at lower level health facilities.
We assessed the performance of symptom-based screening to facilitate the routine implementation of TB
screening for paediatric contacts.
Methods: Children younger than 5 years, in contact with
TB confirmed adult index cases were assessed with
standardized symptom screening, chest X-ray and
tuberculin skin test (TST). Children diagnosed as active
TB disease based on clinical, radiological and bacteriological investigations were referred for treatment; the rest
who were either uninfected or with latent TB infection
(LTBI) were systematically treated with 6 months IPT
and followed for 9 months. The accuracy of symptombased screening for the diagnosis of active TB was
estimated against a reference definition of active TB
based on chest X-ray findings independent of clinical
findings.
Results: Of 281 child contacts (median age 33 months;
53% female and 4.5% HIV infected), 43 (15.3%) were
initially classified as active TB, 102 (36.3%) as LTBI and
136 (48.4%) as non-infected cases. Two additional
patients were diagnosed as active TB during follow-up
on IPT. Active TB was initially diagnosed in 33/69
(47.8%) of symptomatic children vs. 10/212 (4.7%) of
non-symptomatic children, P , 0.001. Chest X-ray was
highly suggestive of active TB in 35/280 (12.5%)
patients. After adjustment on age, HIV and sex, cough
(aOR 3.2, 95%CI 1.4-7.6), reduced playfulness (aOR
9.7, 95%CI 2.5-37.7) and moderate to severe malnutrition based on weight-for-height z-score (aOR 13.0,
95%CI 1.8-91.9) were associated with radiological
active TB. Screening based on the presence of one of
these 3 signs showed 71.4% (25/35, 95%CI 53.7-85.4)
S537
sensitivity, 63.1% (154/244, 95%CI 56.7-69.2) specificity and 93.9% (154/164, 95%CI 89.1-97.1) negative
predictive value. Of 10 patients with negative screening
and radiological active TB, 9 were started on TB
treatment and 1 on IPT.
Conclusion Symptom-based screening could be an
effective and easy tool to identify paediatric TB contacts
requiring further evaluation for active TB. However, our
symptom-based screening needs to be optimised to
reduce the risk of initiating IPT in children with
potentially active TB. Further evaluation using a more
accurate reference standard for active TB is required.
69. Paediatric TB: epidemiology - II

PC-1284-06 Systematic review of the
epidemiology of and programmatic response
to TB in children in South Africa
A Garcia-Prats,1 F Adeniyi,1 S Nicol,1 K Du Preez,1
H Rabie1 1Stellenbosch University, Cape Town, South Africa.
e-mail: liesl.nicol@gmail.com
Background: Tuberculosis (TB) causes substantial morbidity and mortality in children. Challenges in diagnosis,
a focus on smear-positive disease which is uncommon in
children, and an underestimate of the morbidity and
mortality from childhood TB, has resulted in a lack of
attention to children (i.e. ,15 years of age) in TB
programmes. Age-related differences in TB epidemiology, pathophysiology, disease manifestations, diagnosis,
prevention and treatment require children be considered
specifically in TB programmes. With the aim of
relevant research on the epidemiology of and programmatic response to TB in children in South Africa.
Design/Methods: A comprehensive search for published
and unpublished research was conducted. Based on
specific inclusion criteria research investigating TB
incidence, prevalence, risk factors, prevention, treatment, clinical outcomes and care in children in South
Africa were included in this systematic review.
Results: 190 studies were included in the review and
reported on one or more of the review outcomes. Table 1
indicates the number of studies reporting on each of the
main outcomes as well as information on where the
studies were conducted.
Conclusion: This review demonstrated a high burden of
childhood TB, suggesting a failure of existing TB control.
There was little health systems and services research and
the available studies reported serious problems with
implementation of key programs, such as delivery of
isoniazid preventive therapy. Other important findings
were a paucity of community-based research and a lack
of geographic distribution, with most studies being from
the Western Cape. Renewed efforts to address key
research and implementation gaps will be needed to
improve the programmatic response to childhood TB in
South Africa.
S538
with confirmed TB died before treatment initiation. Of

the remaining 126 children started on treatment, 103
(81.7%) had a chest X-ray suggestive of TB.
Conclusions: The prevalence of HIV infection and
malnutrition was high in this cohort. Despite high
sensitivity of the Xpert MTB/RIF as compared to culture,
very few patients started on treatment had confirmed TB
and majority were treated based on chest X-ray. There is
need to routinely and systematically evaluate malnourished and HIV infected children for TB. More robust
diagnostic tests are required to increase the proportion of
children treated on the basis of confirmed TB.
PC-1285-06 Childhood tuberculosis in Mbarara

National Referral Hospital, Uganda: a
description of patients and diagnosis
1,2
1,2
E Kumbakumba, M Nansumba., J Kiwanuka,

M Bastard,4 P Orikiriza,3 D Nansera,1,2 Y Boum Ii,1,3
M Bonnet3,5 1Mbarara University of Science and Technology,
Mbarara, 2Mbarara National Referral Hospital, Mbarara,
Uganda, 3Epicentre, Mbarara, Uganda ; 4Epicentre, Paris,
5
Institut de Recherche pour le Developpement

(IRD),
Montpellier, France. e-mail:
Background: Tuberculosis (TB) is common in young

children with HIV infection and malnutrition. We
investigated a cohort of children for active tuberculosis
by age, HIV co-infection and nutrition status. Methodology: Children , 14 years old with at least one of the
following symptoms (unexplained weight loss, persistent
respiratory symptoms, fever, fatigue or reduced playfulness) were investigated with chest X-ray, sputum
microscopy, XpertMTB/RIF and culture and stool
XpertMTB/RIF and culture on a subset of patients
started on treatment. Decision to initiate treatment was
based on clinical, mycobacterial and radiological findings.
Results: Of the 396 children enrolled, 217 (54.8%) were
male and 123 (31.1%) were HIV positive. The median
age was 4 years (IQR 1.4-7.5) and 79 (20.3%) children
had moderate or severe acute malnutrition (weight for
height Z- score). Seventy-six (19.2%) children had a TB
contact history and 144 (36.4%) were started on TB
treatment. These children were significantly younger
than those who were classified as not TB, with a median
age of 2.5 (IQR 1.1-5.4) years versus 5 (IQR 2.1-8.3)
years (P , 0.001). A significantly higher proportion of
these children had moderate or severe acute malnutrition
(33.8% versus 12.5%, P ,0 .001) but there were no
differences in HIV co-infection rates. Overall, TB was
confirmed in 20 (5.0%) children: 13 had a positive Xpert
MTB/RIF on sputum and 7 on stool and 13 had positive
culture. Of 351 children tested with both culture and
Xpert MTB/RIF on sputum, sensitivity of Xpert MTB/
RIF was 90.9% (10/11) and specificity was 98.8% (336/
340). Sixty-four children started on TB treatment had
XpertMTB/RIF done on both sputum and stool: 8
(12.5%) were positive; 4 (6.3%) on both, 3 (4.7%) on
sputum only and 1 on stool only (1.6%). Two children
PC-1286-06 High burden of tuberculosis and

delayed initiation of treatment among children
in a resource restricted setting
S Kizito,1 A Katamba,1 C Atuhaire,1 G Amanya,1
E Obuku1 1Makerere University, Kampala, Uganda. e-mail:
somekizito@yahoo.com
Background: In Uganda, childhood TB is neglected.

Contact tracing lags, no disaggregation of reported data
for children, limited diagnostic capacity. With high HIV
and adult TB burden, and 51.4% of population being
children, child TB burden is expected to be high. Use of
Desk Guide is limited. TB diagnosis is mainly based on
smear results. There is paucity of data and uncertainty in
prevalence of childhood TB in Uganda. We aimed at
determining prevalence; factors associated with PTB in
children attending PHC facilities in Kampala and missed
opportunity to initiate treatment within a month of
diagnosis.
Design/Methods: Cross-sectional study among 255
children 5 15 years, in 6 health facilities in Kampala,
March 2015. Socio-demographic, clinical, and laboratory data were collected using questionnaire. TB diagnosed
using Desk Guide algorithms. Sputum based on ZN/
FM and/or Gene-xpert. Logistic regression used to assess
associations with outcome.
Results: Prevalence of all-forms, smear positive, smear
negative TB were respectively: 13.7% (2.6 24.8), 6.3%
(0.9 11.6) and 7.5% (2.8 14.6). Among HIV positive:
41.7%, 33.3%, and 8.3%. Among malnourished;
21.7%, 8.3%, and 5.3% respectively. Among contacts,
all-form 89.3% and smear positive 78.6%. Factors
associated with smear positive PTB; windows/room
(OR 0.1, 95%CI: 0.01 0.48), contact (OR 8.3,
95%CI: 2.02 33.70), malnutrition (OR 3.6, 95%CI:
1.28 9.99). Factors for smear negative PTB: tobacco
smoker at home (OR 8.1, 95 % CI: 2.72 24.14), 5 Km
or more from hospital (OR 3.1, 95 % CI: 1.02 9.66),
use of firewood as fuel (OR 15.7, 95 % CI: 2.30
106.56). Factors for all-form PTB: tobacco smoker at
home (OR 1.6, 95 % CI: 1.07 6.86), stunting (OR
2.2, 95 % CI: 1.01 4.15). Only 5.3% of smear-negative,
81.3% of smear-positive, overall 40% of diagnosed
children were initiated on treatment within a month of
diagnosis.
Conclusion: We found high pediatric TB burden more-so
in HIV positive, malnourished, contacts. Delayed initi-
ation of TB treatment. There is need to intensify

screening HIV positive and malnourished for TB,
intensify contact tracing, and need to follow up
diagnosed children to ensure timely treatment initiation.
Variable
Overall
Categories
255
All form
TB
Smear
positive
TB
Smear
negative
TB
Prevalence
Prevalence
Prevalence
13.7 (2.6
24.8)
HIV
Negative (236)
12.7 (0.6
24.8)
Positive (12)
41.7 (12.0
71.3)
Nutrition Good (194)
11.3 (4.5
18.2)
Poor( 60)
21.7 (0.1
53.1)
Sex
Female (145)
9.7 (1.2
18.1)
Male (110)
19.1 (0.7
34.5)
Contact
No contact
10.7 (1.0
22.5)
Had contact
89.3 (77.5
99.9)
Education Not enrolled (17) 17.6 (4.1
31.2)
Pre-primary (26) 11.5 (4.2
27.3)
Primary
12.8 (1.5
24.1)
Secondary (48)
16.7 (0.1
38.1)
6.3 (0.9
7.5 (2.8
11.6)
14.6)
5.1 (0.9
7.6 (0.4
10.3)
15.7)
33.3 (4.0
8.3 (0.1
62.6)
31.0)
4.1 (0.1
7.2 (1.2
13.2)
13.2)
13.3 (0.5
5.3 (0.5
22.0)
13.0)
3.4 (1.1
6.2 (1.0
5.8)
13.5)
10.0 (0.8
9.1 (0.3
20.8)
18.6)
21.4 (1.4
43.0)
78.6 (57.0
99.9)
5.8 (0.2 11.8 (0.4
9.5)
28.5)
11.5 (0.4 No obser 27.3)
vation
3.7 (0.1
9.1 (0.8
8.5)
19.1)
12.5 (3.7
4.2 (0.2
28.7)
9.9)
S539
district hospitals (14.8% of 4428), referral hospitals

(12.2% of 9315) compared to HCIII (7.8% of 743) and
HCIV (6.2% of 2427). Treatment success was better in
the lower level facilities (referral hospitals (46%), district
hospitals (56%), HCIV (80%), HCIII (78%)). Deaths
were higher in HC III (13%) and referral hospitals (8%)
while LTFU was higher in referral (20%) and district
hospitals (16%). While 48% of the facilities had the
national TB guidelines available, only 24% of them were
conducting contact tracing. Screening for TB in children
using the recommended tool was evident in 80% of the
facilities. There was limited involvement of specialists in
childhood TB management as shown below. Persistent
cough (30%), weight loss (29%), and persistent fevers
(16%) were the most identified childhood TB symptoms.
Microscopy (37%), CXR (29%), and Xpert (14%) were
the most listed appropriate childhood TB diagnostic
tests.Clinical symptoms (26%) and smear microscopy
(25%) were the most commonly used diagnostics.
Sputum (27%) and gastric aspirate s(21%) were the
most collected samples. Only 21% of the facilities had
access to onsite Xpert test. 91% of the facilities were
providing anti-TB medicines. Most of the facilities
manage TB-HIV co-infected children in the TB (43%)
and HIV (25%) clinics.
Conclusions and key recommendations: Findings indicate minimal involvement of specialists in child TB
management, limited access to recommended diagnostics
hence an urgent need to streamline and strengthen
national capacity for the diagnosis and management of
TB in children.
Table showing the prevalence of TB among children 5 to 15 years with

features suggestive of tuberculosis, Kampala Uganda 2015
PC-1287-06 Capacity of health facilities to

diagnose and manage childhood TB in Uganda
M Sekadde,1 R Asiimwe,1 B Ngwatu,2P Bakkabulindi,2
E Namusoke-Magongo,1 E Quinto,1 S Okokwu,3
F Mugabe1 1Ministry of Health, Kampala, 2Clinton Health
Access Initiative, Kampala, 3UNICEF, Kampala, Uganda.
e-mail: moorine.sekadde@gmail.com
Background and challenges to implementation: With a

childhood TB case notification of 3% among bacteriologically confirmed incident TB cases, the TB Program in
Uganda conducted an assessment to determine the
capacity for childhood TB diagnosis and management.
Intervention or response: A cross sectional study. 112
public and private health facilities were randomly
selected from the 12 regions. Capacity was defined as
practices and resources for childhood TB screening,
diagnosis, treatment, prevention, and recording and
reporting. A pretested tool was used to collect data from
key facility personnel. Univariate and bivariate analysis
was done using SPSS version 16.
Results and lessons learnt: Of the facilities assessed 29
(26%), 38 (33%), 31 (28%), and 14 (12%) were Health
Center (HC) III, IV, district and referral hospitals
respectively and 87% were public sector facilities.
Childhood TB case notification was highest in the
PC-1288-06 DETECT Child TB: increasing access

to TB services for children in Uganda
S Muyanja,1 A Detjen,2 F Mugabe,3 M Sekadde,3,4
J Nakaweesi,5 M Murungi,4 S M Graham,2
A Nakanwagi-Mukwaya1 1International Union Against
Tuberculosis and Lung Disease (The Union), Kampala,
Uganda; 2The Union, Paris, France; 3The Uganda National TB
and Leprosy Program, Kampala, 4Baylor College of Medicine
Childrens Foundation Uganda, Kampala, 5The Mild May
Hospital Uganda, Kampala, Uganda. e-mail:
szmuyanja@yahoo.com
Background: The Decentralize TB services and Engage

Communities to Transform lives of children with TB
(DETECT Child TB) project aims to develop a districtwide health systems model for routine childhood TB care
S540
in two districts in Uganda. The Project will train and

mentor faciltity-based health care workers in prevention,
diagnosis and management of childhood TB and engage
village health teams (VHTs) to perform contact screening
and treatment support.
Objective: To determine baseline data on child TB case
finding, treatment and prevention services in Wakiso and
Kabarole districts.
Methods: A cross sectional survey was conducted in all
hospitals, level IV health facilities and a selection of level
III health facilities.Information was obtained on current
practices concerning the prevention, diagnosis and
treatment of TB in children. In addition, information
on VHTs and their existing linkages to facilities and roles
in TB care was collected.Document reviews of NTLP
Facility Registers were done to verify the number and
disease category of children notified to the NTLP.
Results: A total of 58 health facilities were included in
the survey, covering 60% of all facilities down to level III
in the two participating districts. From January 2013 to
December 2014, 404 children 0-14 years were notified
with TB from these health facilities, representing 9% of
all TB cases notified during this time period. Of these
only 4% (18) were under 5 years and 10% (43) were
bacteriologically confirmed. The majority of children
(60%, n 241) were diagnosed at hospitals, which made
up only 14% (8) of all health facilities sampled. ZN
Microscopy and CXR were the main diagnostic tools to
aid TB diagnosis in children. 42% of health facilities
reported doing routine contact tracing although only 10
% n 6 had a record of this. IPT was routinely provided
at only 18% (10) health facilities while only 19% (11)
facilities utilized VHTs for TB Services.
Conclusion: Provision of child TB services in Uganda is
still highly centralized and should be decentralized to
lower health facilities. At these health facilities, contact
screening and IPT should be integrated into routine care
and healthcare workers should be trained to confidently
make a clinical diagnosis of TB in children.
PC-1289-06 Looking where the light is bad:

engagement in care among individuals with
LTBI in the United States
J Mancuso1 1Uniformed Services University of the Health
Sciences, Bethesda, MD, USA. e-mail:
james.mancuso@usuhs.edu
Background: An estimated 12 million people (4.2%) in

the U.S. had LTBI in 2000, and prevalence was higher
among the foreign born and certain ethnic groups and
races. However, the proportion of these infected persons
who have been diagnosed or appropriately treated is
unknown. The objective of this study was to estimate
engagement in LTBI care in the United States and
examine risk factors for lack of engagement.
Design/Methods: We used the demographic and TB
components of the National Health and Nutrition
Examination Survey (NHANES) from 2011-2012. The
population at risk was all those with LTBI, as determined
by a TB skin test induration of 10 mm or larger. Those
with a history of active TB were excluded from this study.

Engagement in care was measured by self-reported
history of testing, diagnosis, treatment, and completion
of therapy. Prevalence estimates for the U.S. population
were calculated using NHANES 2011-2012 Medical
Examination Center 2-year weights to adjust for
probability of selection and nonresponse to the survey,
with additional reweighting to account for nonresponse
to TB testing.
Results: There were a total of 11.9 million people
infected with LTBI in the United States in 2011-2012. Of
these, 71% had been tested for TB, but only 23%
reported a history of a diagnosis of TB (Figure). Only
12% had a history of treatment and 11% a history of
completion of therapy. Thus, 10.6 million people (89%)
infected with LTBI never completed a course of therapy
and remain at higher risk for developing TB disease.
Infected people who were older or were contacts of a case
of TB disease were more likely to be aware of the
diagnosis of LTBI. Foreign born and members of higher
risk racial and ethnic groups were not more likely to be
aware of their LTBI than people who were US-born or
Caucasian.
Conclusion: As seen in the cascade of LTBI care shown in
the Figure, most persons in the USA with LTBI had been
tested previously, but few recalled a history of diagnosis
and almost 90% never completed a course of treatment.
Although those who start therapy for LTBI in the USA
are likely to complete it, this is a poor measure of TB
control program effectiveness. To contribute meaningfully towards TB elimination, efforts should be directed
towards looking where the light is badthat is,
identifying those with LTBI and ensuring that they are
diagnosed and engaged in treatment.
PC-1290-06 Association between latent

tuberculous infection and indoor air pollution
among household contacts of pulmonary
tuberculosis cases
1
S K Sahu, D Reddy, A Mcintosh, R Kubiak, G Roy,

J Ellner,2 S Sarkar,1 N Hochberg,2 1Jawaharlal Institute of
Postgraduate Medical Education and Research, Pondicherry,
India; 2Boston University, School of Medicine, Boston, MA,
3
Boston University, School of Public Health, Boston, MA,
4
Albert Einstein College of Medicine, Bronx, NY, USA. Fax:
(1) 718 678 1020. e-mail: divya.reddy@einstein.yu.edu
Background: One component of the effort to control

tuberculosis disease (TB) is to identify modifiable factors
associated with latent tuberculosis infection (LTBI).
Approximately half the worlds population uses unprocessed solid fuels such as coal and biomass (e.g. wood,
dung, crop residues) for household cooking that cause
indoor air pollution (IAP). Limited data suggest an
increased risk for TB due to IAP, but its association with
LTBI has not been evaluated. This study aims to assess
the association between use of biomass cooking fuels and
LTBI.
Methods: We analyzed data from culture-confirmed
pulmonary TB cases and their household contacts (.6
years) recruited into the Regional Prospective Observational Research for TB (RePORT) cohort in Pondicherry
and Tamil Nadu, India. Demographic, clinical, and
household data were collected and tuberculin skin tests
(TST) administered for household contacts. LTBI was
defined as a TST induration of 7 10 mm. Univariate
associations were calculated using logistic generalized
estimating equations to account for household-level
clustering effects.
Results: Of the 242 household contacts in the 74
households with environmental data, 144 (60%) were
female, and the median age was 22 years (IQR 24).
Among household contacts, 223 (92%) had never
smoked tobacco, but 109 (45%) lived in households
where smoking was allowed, including 58 (53%) on a
daily basis. 75 (31%) household contacts had LTBI.
Household contacts with LTBI were more likely to live in
households where tobacco smoking was allowed (OR
1.9; 95%CI 0.9-3.6; P 0.07). The primary cooking
fuels were liquid petroleum gas (52; 70%), wood (18;
24.3%) and kerosene (2; 2.7%); 1 (1.3%) household
used electricity. Among those using wood as the primary
fuel, 11 (61%) had an outdoor kitchen. Compared to
households using electricity for cooking, those using
wood were more likely to have household contacts with
LTBI (OR 2.9) but the difference was not statistically
significant (95%CI 0.7-12.5; P 0.16).
Conclusion: About one-quarter of the studied households use wood for cooking, which is a known
contributor of IAP, however the majority use an outdoor
kitchen. While our data do not show a statistically
significant association between biomass fuel use and
LTBI, the effect estimate is substantially larger than 1,
despite the small sample size. Additional analyses are
ongoing to assess the role of potential socioeconomic
confounders including tobacco use and ventilation.
S541
PC-1291-06 Clinical and immunologic profile of

TB-HIV co-infected children at a paediatric HIV
clinic in Uganda
H Lukolyo,1,2 M Murungi,1 A Mandalakas,2,3 D Damba,1
K Ngo,3 A Kekitiinwa1,2 1Baylor College of Medicine-Bristol
Myers Squibb Childrens Clinical Center of Excellence at
Mulago Hospital, Kampala, Uganda; 2Baylor College of
Medicine, Houston, TX, 3The Global TB Program, Texas
Childrens Hospital, Houston, Texas, USA. e-mail:
heather.lukolyo@bcm.edu
Background: Tuberculosis (TB) is one of the most

common and deadliest opportunistic infections among
people living with human immunodeficiency virus (HIV)
in resource-limited settings. However studies describing
the clinical characteristics and outcomes of children with
TB-HIV co-infection are lacking.
Design/Methods: This is a retrospective cohort analysis
of all new cases of active TB disease among children 0-15
years diagnosed in 2013 at the Baylor College of
Medicine Childrens Foundation of Uganda. Data were
extracted from the electronic medical record. Descriptive
statistical analysis was performed to describe clinical
characteristics of cases.
Results: There were 4036 HIV-positive and HIV-exposed
children ages 0-15 years active in care at the end of 2013.
During 2013, 90 children were diagnosed with TB and
started on anti-TB treatment (ATT); three were found to
be HIV-negative and excluded from analysis. Of the 87
remaining cases, 43 (49.4%) were male with median age
at TB diagnosis of 4.3 years (interquartile range (IQR)
6.5 years). Forty-six (52.9%) cases were diagnosed
among children already on antiretroviral therapy
(ART) while 41 (47.1%) cases were in ART-naive
patients. There were 82 (94.3%) pulmonary TB (PTB)
and five (5.7%) extrapulmonary TB cases. The majority
of PTB cases were diagnosed clinically. For children 0-5
years old, the mean CD4 percent was 15.1% (IQR 13%)
at TB diagnosis and increased to 28.0% (IQR 15%) after
completion of ATT. For children 5-15 years old, the
median CD4 was 293/mm3 (IQR 694/mm3) at TB
diagnosis and increased to 776/mm3 (IQR 565/mm3) by
completion of ATT. The majority (60.8%) of children
were malnourished at the time of TB diagnosis; by ATT
completion 88.7% had adequate nutrition, demonstrating a decreased odds of malnutrition (OR 0.057, 95%CI
0.023-0.14). Sixty-two (71.3%) patients completed ATT,
while 21 (24.1%) died, one (1.1%) was lost to follow up
and three (3.4%) transferred out prior to ATT completion.
Conclusion: In this population of HIV-positive and HIVexposed children, the incidence of new TB cases was 22
per 1,000 children. This study demonstrated a high
mortality rate of 24.1% among TB-HIV co-infected
children. Most children were immunosuppressed and
malnourished at time of TB diagnosis; those completing
ATT demonstrated improvements in both immunologic
and nutritional status. The few patients with bacteriologically confirmed TB highlights the challenge of
accurate TB diagnosis in children.
S542
also less likely to have a documented HIV result. There is

a missed opportunity for adolescent HIV testing at the
time of TB diagnosis in this setting. Future studies of
larger cohorts should explore default and failure rates
among 10-24 year-olds treated for TB, particularly
among 15-19 year-olds. Poor outcomes should be further
characterized, and treatment challenges should be
explored, in order to optimize TB treatment in this age
group.
PC-1292-06 Adolescent TB treatment outcomes

in Gaborone, Botswana
L Enane,1 E Lowenthal,1,2 T Arscott-Mills,1,2,3
B Kgwaadira,4 S Coffin,1,2 A Steenhoff1,2,3 1Childrens
Hospital of Philadelphia, Philadelphia, PA, 2University of
Pennsylvania, Philadelphia, PA, USA; 3Botswana-UPenn
Partnership, Gaborone, 4Botswana National TB Programme,
Gaborone, Botswana. e-mail: leslie.enane@gmail.com
Background: Adolescents are at high risk for TB,

particularly if they are HIV-infected. HIV is the leading
cause of death among adolescents in Africa; the impact of
TB in hastening these deaths is unknown. Furthermore,
there has been very little research on adolescent TB
treatment outcomes. The neglect of adolescent TB
patients is an impediment to their care, and to global
TB control. We sought to describe TB treatment
outcomes among adolescents and young adults (10-24
years) in Gaborone, Botswana, and to compare the rates
of poor outcomes with those of older adults (.25 years).
of standardized TB patient register data at three clinics in
Gaborone, Botswana, from January 2012-January 2014.
Using a paper data abstraction form, we sampled every
adolescent and young adult case, and every third older
adult as a comparison group. Rates of poor treatment
outcomes were calculated for 5-year age groups: 10-14,
15-19, 20-24, and for all 10-24 year-olds, and compared
with those of older adults (.25).
Results: We enrolled 99 adolescent and young adult
patients and 157 older adults. Median ages (interquartile
range) were 21 (18; 23) and 37 (31; 45), respectively. The
adolescent and young adult group, when compared to
older adults, was predominantly female (56% vs. 36%, P
, 0.01); less likely to be HIV-positive (19% vs. 75%, P ,
0.001); and less likely to have a documented HIV test
result (8% vs. 2%, P , 0.04). Adolescents and young
adults had higher default rates (12% vs. 9%), though this
was not statistically significant. The 15-19 year old
group had the highest default rate (16%) (Figure). All of
the treatment failures were in the adolescent and youngadult age groups (5%). The 15-19 year-old group had the
highest rate of treatment failure (8%). Seventeen patients
(7%) did not have a treatment outcome recorded as they
transferred out to receive care at another site.
Conclusion: While adolescent and young adult TB
patients were less likely to be HIV-positive, they were
70. The two Ts: linking TB and tobacco

PC-1293-06 Tobacco smoking and recurrent
pulmonary tuberculosis in Armenia
A Harutyunyan,1 V Petrosyan,1 N Truzyan1 1American
University of Armenia, Yerevan, Armenia. e-mail:
aharutyunyan@aua.am
Background: Tuberculosis (TB) is second only to HIV/

AIDS as the greatest killer worldwide due to a single
infectious agent. In 2013, an estimated 9.0 million people
developed TB and 1.5 million died from the disease.
Poverty, HIV infection and tobacco smoking are the main
determinants for TB. This cross-sectional study evaluated smoking behavior of TB patients and the role of
smoking for the recurrence of TB.
Design/Methods: All the patients with drug sensitive
pulmonary TB that started continuation phase of TB
treatment in March-December 2014 were recruited
through 52 outpatient TB offices in Armenia. Data were
collected through interviewer administered surveys using
a structured questionnaire. The wealth score was
calculated based on the weighted values of self-reported
families general standard of living, household monthly
expenditure, employment and migrant statuses of the TB
patients. The level of depression was assessed using the
Center for Epidemiological Studies Depression (CES-D)
scale. The study used descriptive statistics and logistic
regression for data analysis.
Results: Overall, 395 TB patients participated in the
study. The mean age of participants was 46.7 years
(SD15.4), 78.0% were male. Smoking prevalence was
significantly higher among males compared to females
(67.5% vs. 5.8%; P , 0.001). The mean number of

cigarettes smoked per day was 19.5 (SD12.5). About
80% of current non-smoking males reported to be past
smokers which resulted in 75.4% lifetime smoking
prevalence with average smoking history of 28.0 years
(SD14.7). One-fifth (19.8%) of TB patients were
recurrent cases. The odds of having recurrent TB was
higher among current smokers (OR1.93, P 0.014),
males (OR1.72, P 0.117), older (OR1.03, P ,
0.001) and depressed (OR1.78, P 0.046) patients and
those who had lower wealth score (OR0.82, P 0.006).
Multiple logistic regression revealed that after adjusting
for age, wealth score and depression status, the odds of
having recurrent TB was 1.90 times statistically significantly higher among current smokers compared to nonsmokers (P 0.032).
Conclusion: Smoking is very prevalent among drug
sensitive pulmonary TB patients in Armenia and it is a
statistically significant risk factor for recurrence of
pulmonary TB. Joint efforts should be made to coordinate national tobacco and TB control programs to raise
awareness about the TB and smoking association and to
provide assistance for quitting to smoking TB patients.
PC-1294-06 Incorporating tobacco control with

community-based TB services in improving TB
cure rates and tobacco quit rates in urban
centers in Bangladesh
M Basher,1 S Islam,1 S Reza,1 S Alam,1 M Siddiqui,1
M Akramul Islam,1 M Haque2 1BRAC, Dhaka, 2NTP, Dhaka,
Bangladesh. e-mail: basher.ak@brac.net
Background and challenges to implementation: Last few

decades, tobacco consumption through inhalation has
increased substantially especially in developing countries
like Bangladesh. BRAC started integration of smoking
cessation intervention into Directly Observed Therapy
(DOT) programs in 17 DOTS centers of Dhaka since
2011 as tobacco smoking is associated with tuberculosis
(TB) and decrease the effectiveness of TB treatment.
Main aim of the initiative is to explore current smokers
among TB patients and see the effectiveness of smoking
cessation and prevention on TB treatment outcomes.
Intervention or response: BRAC TB programme staff was
trained on tobacco control with precise focus on harmful
effect of smoking and its effect on treatment outcome. All
tools were developed in Bengali based on the guideline
Smoking Cessation and Smoke-free environment for TB
patients of the Union. Counseling is given to patients for
smoking cessation during initiation of treatment and
subsequent visits to TB centre.
Results and lessons learnt: From 2011 to 2014, a total of
14 202 TB patients were enrolled in the intervention
areas. Among them, 23% patients were smokers of
whom 98% were male and 2% female. The level of
addiction was high in 31% cases and low in 69% cases.
Among registered patients between 2011 and 2013, 73%
people quitted smoking among smokers (22%). Treatment outcome 91% which is much better among the TB
S543
patients and quit rate was 80% among the successfully

treated TB patients
Conclusions and key recommendations: Tobacco cessation counselling within TB Patients is found more
effective which may replicate within community based
DOTS programme.
PC-1295-06 Does smoking during tuberculosis

treatment affect outcomes? A cross-sectional
study in Chandigarh, India
S Goel,1 J Kathiresan,1 A Garg,2 R J Singh3 1Post Graduate
Institute of Medical Education and Research, Chandigarh,
2
Health & Family Welfare, U T Chandigarh, Chandigarh,
3
4993. e-mail: sonugoel007@yahoo.co.in
Background: World Health Organization (2014) reported that Tuberculosis (TB) is one of the worlds deadliest
communicable diseases, with an estimated 9 million
developing T.B and 1.5 million succumb from illness. In
year 2013, a total of 240 000 deaths occurred due to TB
in India, with a prevalence rate of 19/100 000
population. A WHO/The Union monograph on TB and
tobacco control has confirmed a strong link between
active and passive tobacco smoking and a range of TB
outcomes including infection, response to treatment,
relapse rates and mortality and had strongly recommended co-ordination between National TB and tobacco
control programmes. There is a lack of data from India
on the direct effects of smoking on TB treatment
outcomes. The study aim to analyze smoking pattern
among T.B patients and the effect of smoking on
treatment outcome of TB patients.
Methods: The study used cross-sectional design among
pulmonary tuberculosis patients over 15 years of age in
Union Territory (UT) of Chandigarh, located in north
India. All the TB patients who enrolled themselves for
DOTS in two quarters (January till June 2013) in total 17
Designated Microscopy Centres (DMCs) of Chandigarh
were target population in the study and were followed up
till their final outcomes. The data was analyzed using
SPSS version 16.0. The study was ethically approved by
Institute Ethics Committee, PGIMER, Chandigarh.
Results: Out of 686 T.B patients enrolled, 78.4% fall in
category 1 of treatment category. Around 12.6% (n 87)
were daily smokers, 10% (n 69) occasional smokers
and almost similar number (n 72, 10.4%) were
quitters. 46.3% (n 318) were smokeless tobacco users.
Over two third (68.6%) initiated their smoking habit
before 20 years of age and almost all smokers take their
first puff within 30 minutes of their wake-up in morning.
Half (57.5%) of the smokers made at least one quit
attempt in their lifetime. The cure rate was significant
higher in non-smokers as compared to smokers. Transfer
out rate and death cases was also found less in nonsmokers, however not statistically significant. The
defaulters and failure cases were significantly more in
smokers as compared to non-smokers.
Conclusions: The TB patients who smoke have higher
chance of poor treatment outcomes as compared to non-
S544
smokers. The tobacco control interventions should

especially focus on tuberculosis patients as they are the
captive and receptive audience during their treatment.
PC-1296-06 Assessment of pulmonary

tuberculosis and its association with smoking
habits in Paniya tribes of India
S Palliyal1 1DM Wayanad Institute of Medical Sciences,
Kalpetta, India. e-mail: shanupalliyal@gmail.com

tribal populations throughout India have remained
socially and culturally alienated from mainstream Indian
society until developmental and conservation activities in
tribal areas forced interactions between the M. tuberculosis (TB) is a major public health problem among Paniya
tribes, a marginalized tribal group in Kerala state, South
India. Previous studies have documented a high prevalence of tobacco use among Paniya Tribals in Wayanad.
However, little is known about their correlation exists
between tobacco habit and pulmonary tuberculosis.The
aim of this study was to evaluate the pulmonary
tuberculosis among nonsmoking and smoking Paniya
tribes.
Intervention or response: A cross sectional survey was
done among 688 nonsmoking and 483 smoking Paniya
tribal populations of Wayanad District, India from
January 2013 to June 2013 after approval from the
Institutional ethical committee. Information on smoking
status, type of tobacco smoked, quantity of tobacco
smoked, and duration of tobacco smoking was collected
from cases and controls using a questionnaire.
Results and lessons learnt: In this study pulmonary
tuberculosis was found to be far more prevalent among
smoking Paniya tribes than among the non smoking
Paniya tribes. (P , 0.0001). The prevalence of severe
tuberculosis was found to be 32% amongst smokers.
This was much higher than the 2% found among the non
smokers.Among the Smokers a statistically significant
relationship was observed between pulmonary tuberculosis and poor access to health care (P , 0.001).
Conclusions and key recommendations: There is a
positive association between tobacco smoking and
pulmonary tuberculosis. The present study demonstrates
gross disparities in pulmonary tuberculosis among
smoking and nonsmoking Paniya tribes. There is an
urgent need to develop and implement culturally
appropriate awareness raising activities to target smoking to support the efforts to control TB in this
community.
to smoking. A study at IGMC, Shimla revealed weak

immune system in people of HP, leading to diseases like
TB.The presence of heavy metals like chromium in soil &
water in industrial area leading to lung ailments.
Education on TB prevention, cure and treatment,
sustainability of smoke free public places and more
initiatives by Govt. is required. Non participation of
community in TB Control, unavailability of treatment up
to remote areas, discontinuation of treatment by the
patients, tobacco usage, increased pollution are a few
identified threats and challenges.
Intervention or response: According to health officials
there is a gap in the access to the DOTS centre and
treatment of patients and patients suffer from malnutrition. ASHA workers are helping detect patients and
monitor the treatment for six to eight months. Migrant
laborers and workers exposed to hazards of pollution
and dust from mines are being focused. HP is a smoke
free state and progressing towards Tobacco free state.
Appointment of nodal officers, strong advocacy, frequent
raids, seizure of Gutka, liaising with administration ,
discussions in steering and monitoring committee meetings, monthly meetings, Gram Sabhas, involvement of
media , serving of notices to public place managers for
violations , removal of promotional boards under sec 5 of
COTPA are few activities done to enforce law. More than
20 000 challans were done during the year 2012-2013
with a fine collection of more than Rs. 20 lakh. Similar
aggressive drives are required by each state for preventable epidemics like TB, HIV and for TAT Free India.
PC-1297-06 Tobacco and tuberculosis free India

S Gupta1 1Individual, Shimla, India. e-mail:
smshellyhp@yahoo.co.in

Health Departments target is to make the state TB free
by 2035. This target seems to be a tough task. About 14
000 new TB patients have been detected last year. The
state continues to report increased lung cancer cases due
Results and lessons learnt: Tobacco related Lung diseases

can be reduced by formulation , implementation &
Evaluation of TB Control Strategies ; Ensuring Smoke
Free Public Places , reaching the unreached & Pollution

free environment.
Conclusions and key recommendations: Sustainability
Monitoring , Evaluation Public Private Partnership
Prevention , Education & In time Quality Treatment to
reduce morbidity & mortality . Research & Development
Effective Enforcement of smoke free laws.
PC-1298-06 Smoking among tuberculosis

patients in India: proactive cessation efforts are
urgently needed
S Nayak,1 P Sharma1 1International Union Against
Delhi, India. Fax: (91) 11 4605 4429. e-mail:
SNayak.HR@theunion.org
Background and challenges to implementation: The new

Stop TB strategy recognizes prevention of most frequent
risk factors as an important contributor to the sustainability of a case management intervention for TB control.
Available evidences show that tobacco smoking is one of
the key risk factors that favour the progression from
latent TB infection to pulmonary disease. The comorbidity of TB and Tobacco has been well established.
However, the effort to help TB patients, quitting tobacco
by explaining the harmful effects/ dangers, especially
during the treatment should be an absolute must.
Simultaneously, required essential support system like
accessibility of cessation centres to the patients either in
their communities or in the hospitals must be established.
As Tobacco creates addiction, hence, cessation can occur
only at moments when an individual finds both a reason
and desire to stop smoking. Smokingsubstantially
increases the risk of TB and death from TB.
Intervention or response: A cross sectional communitybased knowledge, attitude and practices (KAP) survey
was conducted in 30 districts of India by The Union. The
respondents, including general population, TB patients,
health service providers, opinion leaders and representatives of NGOs interviewed (November 2012 to April
2013) through door-to-door survey conducted by using a
semi-structured questionnaire on tobacco use/control.
Results and lessons learnt: All the respondents were
asked about using tobacco in any form, to which 175
(35%) responded affirmatively. Among those respondents who used tobacco, 24% smoke, 36% smokeless
tobacco and 5% were dual users. Noticeably, 49%
tobacco users have been using tobacco for , 10 years. A
majority of the smokers use beedi and among those using
smokeless tobacco, 55% chew tobacco.Though, there
was fair knowledge about TB and tobacco linkages in
general, 73% were advised to quit using tobacco during
their hospital visits. Over and above first hand smokers,
32% are further exposed to secondhand smoke, 38%
knew the consequences, 35% use tobacco despite being a
TB patient, 38% knew the association between TB and
tobacco, 56% knew smoking can increase the risk of
developing TB disease and 48% knew the consequences
of smoking during TB treatment.
S545
Conclusions and key recommendations: Not to smoke

warning messages often go unheeded among TB patients.
Proactive efforts are needed to encourage health staff and
DOTS providers to give strong cessation messages.
PC-1299-06 Integrating smoking management

and TB control program: a way forward
l Pangaribuan,1 B Dwihardiani,2 D Perwitasari,1
T Tejayanti,1 J Irianto,1 M Farid,3 A V Siagian4 1Ministry of
Health Republic Indonesia, Jakarta Pusat, 2World Health
Organization, Indonesia Country Office, Jakarta Pusat,
3
Tuberculosis Operational Research Group, Jakarta Pusat,
4
Ministry of Health Republic Indonesia, Jakarta Pusat,
Indonesia. e-mail: bintari_dwihardiani@yahoo.com
Background: The relation of smoking and TB has been

proven in many literatures. Because both cause lung
problem, the collaboration of TB control program and
smoking cessation project may help control the two
problems. Based on Indonesian national TB prevalence
survey results, we propose findings to support the
integration of the two control programs.
Design/Methods: National TB prevalence survey (NPS)
which was conducted in 2013-2014 applied smoking
history question in its questionnaire. NPS was conducted
as a cross sectional survey with multi staged cluster
sampling among population aged 15 years old and
above. Eligible participants were invited to attend
interview and chest X-ray screening. Those with positive
screening symptoms or chest X-ray were requested to
submit sputum specimen, in which LJ culture and smear
microscopy were conducted. Bacteriologically confirmed
TB case was defined by panel team based on laboratory
results supported by chest X-ray and clinical information. Positively screened individuals and 10% of negatively screened individuals were interviewed for TB
knowledge, attitude, and practice.
Results: Among NPS participants, 33.9% admitted that
they were smoking at the time of interview. Smoker is
significantly higher in the male population. The proportion of TB cases is higher among smoker even though the
association is not significant. As TB prevalence is higher
among male, male sex can be a confounder of the
association of smoking and TB cases. Among TB cases,
smokers had slightly higher proportion of infiltrates and
cavities shown in chest X-ray. Among the participants of
KAP interview, there is no difference in knowledge about
TB symptoms, transmission, and the access of TB
treatment between smoker and non smoker.
Conclusion: Considering high smoker prevalence in
Indonesia, focus on smoker to detect TB may open the
opportunity to manage smoking problem. Moreover,
most of the smoker are men and TB prevalence is higher
among men, therefore integration of TB and smoking
management which focus on male population may work.
Further research is needed to evaluate the yield, cost
effectiveness on TB control and smoking control, and
feasibility of integrating the two programs.
S546
PC-1300-06 Deaths due to respiratory

neoplasms and the trend of tobacco use in Sri
Lanka, 19602010
L Athauda1 1Faculty of Medicine, University of Kelaniya,
Ragama, Sri Lanka. e-mail: lathika@kln.ac.lk
Background: The prevalence of current tobacco smoking

among adults in Sri Lanka in 2012 was estimated as
15.0%, the prevalence being 29.9% for males and 0.4%
for females. Deaths due to neoplasms of the respiratory
tract have been on the rise over time. This study aimed at
describing the tobacco use trends from 19602010 and
its influence on deaths due to respiratory disease
neoplasms.
Intervention or response: Tobacco consumption data was
obtained from the Food and Agricultural Organization
database, population data from the UN database and
deaths data on respiratory disease from the department
of census and statistics Sri Lanka. Deaths due to
respiratory neoplasms included, neoplasms of the larynx,
trachea bronchus and lungs, as well as neoplasms of intra
thoracic organs. Correlation and regression analysis were
carried out.
Results and lessons learnt: The average tobacco use from
1960-2010 was 560.87223.43kg/capita/year. The trend
of tobacco use displayed and increment from 1960 1978, and a declining trend thereafter. The average death
rate due to neoplasms of the respiratory tract from 1960 2009 were 1.47 1.04. There was a decrease in mortality
due to respiratory neoplasms beginning in 1995; this
decrease corresponds to the decrease in smoking since
1978 having a lag period of about 15 years (r 0.002 P
0.021, R20.151). Tobacco control policy was implemented at a national legislature ony in the year 2006,
though execution has been problematic.
Conclusions and key recommendations: The rise in
deaths due to neoplasms of the respiratory tract in Sri
Lanka shows a clear association with tobacco use of a lag
of 15 years. The trends of respiratory neoplasm deaths
are yet on the rise. The future may be predicted to be a
downward trend, as Sri Lanka follows a global trend of
redcution in tobacco use. However, the reduction of
deaths can be influenced by the improving health care
and techonology of the country.
in China. When a survey was conducted in 14 Chinese

cities, questions related to E-cigarette were also included.
Design/Methods: An urban representative population of
individuals age 15 or above was selected with PPS
method in 14 cities. A target sample size of 2000
respondents was required for each city. The adjusted
sample size was 2500 after taking into account the
potential loss due to ineligibility and non-response
(assuming an 80% final response rate). In Beijing, the
adjusted sample size was 4,800 in the urban areas. Inhouse interview was conducted using handhold devices
from October 2013 to August 2014. Questions of Have
you ever heard of e-cigarette? Do you currently use ecigarette? were asked to the subjects. SPSS was used for
data cleaning and data management; SAS-callable
SUDAAN was used in data analysis. Sample weights
were calculated for each respondent, and were also poststratified to the total adult population of each city
respectively, by gender and age groups, using population
counts from the 2010 China National Population
Census.
Results: 31 151 of 37 300 respondents completed the
survey, representing a population of 52.8 million. In the
current survey, more than 40.0% of adults reported
having heard of e-cigarettes in each of the 14 cities.
However, in all 14 cities, the percentage of adults who
currently used e-cigarettes was low, with the highest
found in Luoyang at 1.7%. The survey showed that the
majority of the current e-cigarette users were also current
cigarette smokers (80.0% or more in all cities). However,
some e-cigarette users reported that they were not
current or occasional tobacco smokers, including some
who never smoked tobacco before
Conclusion: E-cigarettes attained a high degree of
awareness in the 14 participating cities, while a small
percentage of adults reported using the products. It is
worth noting that current e-cigarette users consisted of
current tobacco smokers and adults who were not
current tobacco smokers, some of whom had never
smoked tobacco before. Additional research and regulation are needed to be conducted on multiple areas of ecigarette use.
PC-1302-06 End of ENDS (Electronic Nicotine

Delivery System) in state of Punjab, India
R Gupta,1 V Mahajan,1 K Singh,1 A Singla,1 S Goel2
Tobacco Control Cell Punjab, Chandigarh, 2Post Graduate
Institute of Medical Education and Research, Chandigarg,
India. Fax: (91) 1725 012 356. e-mail:
rakesh60.mahajan@gmail.com
71. From old to new: chicha to ecigarettes

PC-1301-06 The emerging E-cigarette in China
C Wang,1 J Wang,1 L Zhao,1,2 Y Yang,1 Y Jiang1 1Chinese
Center for Disease Control and Prevention, Beijing, China;
2
GA, USA. e-mail: wangcongxiao@163.com
Background: E-cigarette was first invented and sold in

China in the 2000s. However, it was not widely used in
China but well recognized in the western. As it is more
and more popular worldwide, the situation also changes
Background and challenges to implementation: Electronic Nicotine Delivery System (ENDS) popularly
known as e-cigarettes are highly addicting and potentially lethal products. It is mostly being used by children
and youth because these are glamorised by the tobacco
industry. Though not generally available in stores, they
are widely promoted through social media, email
marketing with discount offers. Sales are increasing
sharply all over the world. Currently, these are not
regulated by any national authority in India. Punjab
became the first state in India to declare Electronic

Nicotine Delivery System (ENDS) as illegal in 2013.
Intervention or response: Various steps were undertaken
during the process to institutionalise the monitoring
mechanism for abuse of ENDS. These included sensitization of political leadership and bureaucracy at state
level; advocacy meetings of all Deputy Commissioners
(administrative head of district) and Civil Surgeons
(administrative head of health department of district);
meetings with tobacco wholesalers to make them aware
about the violation of Drugs and Cosmetics Act,which
allows nicotine (2 mg and 4 mg) only as gum and
lozenges to be manufactured and sold and sensitization
of media about the importance of making e-cigarettes
unavailable for sale.
Results and lessons learnt: A circular was issued by
Government of Punjab regarding declaration of manufacture and sale of ENDS as illegal. District level Task
Force effectively implements the ban through conducting
raids at suspected point of sale, which is being monitored
monthly by Deputy Commissioners. At state level,
Health Minister and Principal Secretary regularly monitor the implementation with Civil Surgeons. The media
sensitises the public through regular news-articles/stories/talk-shows etc. about the importance of making ecigarettes unavailable for sale. Violators are being
booked under Drugs and Cosmetics Act. Court cases
have been filed against violators. In a special 3 day
campaign in March 2015, 1050 Points of sale were
checked all over Punjab and no sale of ENDS/ECigarettes was found.
Conclusions and key recommendations: The declaration
by the Government of Punjab opens the way for other
states to follow suit and prevent ENDS becoming an
additional marketing strategy for tobacco companies.
PC-1303-06 E-cigarettes: prevalence of use and

perceptions among adolescent college
students of Mangalore city
E Aluckal1 1A J Institute of Dental Sciences, Mangalore,
India. Fax: (91) 824 222 4968. e-mail: draluckal@gmail.com
Background: Tobacco usage is the largest, preventable

cause of premature mortality in the world. E-Cigarettes
are battery-powered devices that deliver vaporized
nicotine without the tobacco combustion of regular
cigarettes. Tobacco control advocates and researchers are
concerned that e-cigarettes could act as gateway
devices, getting novice users, particularly the young
people, addicted to nicotine and encouraging future
tobacco use. The adolescent college students are a group
that is at risk for smoking initiation and may use ecigarettes as an experimentation to start smoking. These
products are gaining popularity in many countries but
little is known about their use among college students in
India.
Design/Methods: A descriptive cross sectional study was
conducted on 1074 adolescent college going students of
different pre-university colleges of Mangalore city using
cluster sampling. Information was elicited on their socio-
S547
demographic characteristics, usage of e-cigarettes, perceptions and practices related to that. Statistical analysis
was done using SPSS version 21.
Results: The mean age of study population was 16.4
years with 59% males. Around half of respondents
(46.4%) had seen e-cigarettes advertised. A total of
16.1% reported trying an e-cigarette (5.1% nonsmokers,
18.4% former smokers, and 34.5% current smokers),
and 5.9% reported use in the past 30 days. Compared to
non-smokers, former smokers and current smokers were
more likely to have tried e-cigarettes (OR 4.15 and OR
9.84 respectively), and current smokers were more likely
to have tried e-cigarettes than former smokers (OR 2.37).
Current smokers were also more likely to be current users
of e-cigarettes than both former smokers (OR 15.24) and
non-smokers (OR 4.83). Smokers were interested in
trying e-cigarettes to help them quit smoking (80.4%), as
a long-term replacement for cigarettes (74.8%), or to use
in places where they cannot smoke (82.1%).
Conclusion: Most adolescent college students were
aware of e-cigarettes, and a substantial minority reported
trying them, with evidence of use among nonsmokers.
Given that even experimentation with e-cigarettes could
lead to nicotine dependence and subsequent tobacco
usage, regulatory and behavioral interventions are
needed to prevent gateway use by adolescent college
students. Education about e-cigarettes could help providers deliver comprehensive preventive services to
adolescents at risk of tobacco use.
PC-1304-06 Hookah smoking and lung cancer in

Kashmir valley, Indian subcontinent
M Dar1 1Directorate of Health Services Kashmir, Srinagar,
India. e-mail: uberaltaf@yahoo.co.in
Background: The literature about the causal relationship

between lung cancer and tobacco smoking mostly
concerns cigarettes. Hookah smoking is popular in the
Kashmir valley of the Indian subcontinent, and is
generally believed to be innocuous because of the passage
of the smoke through water before inhalation. Objective:
To determine the relationship of hookah smoking to lung
cancer in Kashmir.Hookah smoking is widely practiced
in Kashmir and was found to be the commonest form of
smoking amongst the patients with lung cancer.
Design/Methods: The study was conducted in the
SheriKashmir Institute of Medical Sciences, Srinagar,
Kashmir (India), a 650 bedded tertiary care university
hospital that serves as the main referral center for the
Pulmonary and Oncology cases of the Kashmir valley of
the Indian subcontinent. Hookah (locally called Jajeer)
in Kashmir is traditionally made of an earthern-ware
water bowl or base (made in current times of tinned
copper or brass with exquisite surface carvings) that is
half flled with water and connects to a separable
earthenware head (or Chillam) by a cane wood conduit
or body, the lower end of which stays submerged in the
water of the base. Another inverted
Results: The cases included 209 males and 42 females
whereas the controls included 328 male and 72 female
S548
subjects. The mean age in males was 58.4 years and in

females 56.51 years. Seventy seven percent of the cases
and 44.6% of the controls were ever smokers. Table 1
shows distribution of the smoking patterns in the cases
and the controls. Smoking was strongly associated with
the presence of lung cancer, with 194 cases being ever
smokers (178 current smokers, 16 ex smokers) as
compared to 173 controls hookah smoking was the
commonest pattern of smoking either exclusively or with
cigarettes.
Conclusion: Our hospital based study provides evidence
that hookah smoking is associated with an increased risk
of lung cancer in ethnic Kashmiri population with the
risk being 6 times more as compared to non smokers. The
study reaffrms the previous report by by Nafae et al in the
sixties who found hookah smoking as the commonest
form of smoking in a cohort of 25 patients of lung cancer,
seen in 20 of the 25, being exclusive in 17.
PC-1306-06 Water-pipe tobacco (shisha)

use among undergraduate health
professional students, College of Health
Sciences, Nairobi University, Kenya, 2014
A Maina,1,2 D Kiptui,2 W Arvelo1,3
1
Field Epidemiology and Laboratory Training Programme,

Nairobi, 2Division of Non-Communicable Diseases, Ministry
of Health, Nairobi, 3Centres for Disease Control and
Prevention, Nairobi, Kenya. e-mail: akmaina@yahoo.com
Background: Tobacco use contributes to more than 6

million annual deaths globally. A growing body of
knowledge shows a rise in the use of water pipe tobacco
(shisha), spreading from the traditional Eastern Mediterranean and Northern African regions to other parts of
the world. In Kenya, anecdotal reports from the media
and observations made in social places have shown a rise
in shisha use among the youth who perceive it as a safer
alternative to cigarette use. We assessed the prevalence of
and factors associated with shisha smoking among
undergraduate health profession students in the College
of Health Sciences, University of Nairobi, Kenya.
among final year undergraduate students enrolled in the
medicine, nursing, pharmacy and dentistry programmes
in August 2014. Students completed a questionnaire
adapted from the Global Tobacco Surveillance System.
Information on ever and current use of shisha, social
demographic variables, alcohol and cigarette use was
obtained. Data was analyzed using Epi Info 3.5.1.
Results: A total of 246 students were interviewed with a
mean age of 23 years. Majority were female 145 (59%)
and in the Medicine program 128 (52%). Fifty three
(21.5%) of the respondents were current shisha users, 84
(34.1%) had ever used shisha in their lifetime and 11
(4.5%) were current cigarette smokers. Current shisha
use was more commonly reported among males 24
(23.8%), medical students 37 (28.9%), self-sponsored
students 33 (25%) and students residing in a rented
house 14 (37.8%). Majority 36 (69%) smoked weekly;

mainly at entertainment spots 47 (89%). Concurrent
alcohol and cigarette use among current shisha users was
47 (90%) and 8 (15.4%) respectively. Factors associated
with current shisha use were catholic faith (OR 2.03;
95%CI 1.04, 3.96), residence in a rented house
(OR2.65; 95%CI 1.25, 5.61), alcohol use (OR13.46;
95%CI 5.47-33.06) and family member who smokes
shisha (OR 6.43 95%CI 3.32-12.43).
Conclusion: The use of shisha as an alternative form of
tobacco is high among university students undertaking
health professional courses. There is need for initiatives
geared towards behavior change among these students to
boost tobacco control efforts among the youth in Kenya.
PC-1307-06 An evidence-based public

campaign against waterpipe / hookah smoking
in Turkey
T Durgut1 1Turkish Green Crescent, Istanbul, Turkey. e-mail:
erdebirtuba@hotmail.com
Background: Hookah use has been spreading rapidly in

epidemic proportions in Turkey, especially among
teenagers and youth. For this reason the Turkish Green
Crescent Society has developed a hookah awareness
campaign. Both qualitative and quantitative pre-campaign research has been conducted in order to define the
perception, attitudes and behaviors of the target population regarding hookah. In this stage we conducted
focus group studies and a baseline survey encompassing
interviews with 1288 people.As a result of the precampaign analyses, we observed that the level of
awareness on the dangers of hookah was very low
compared to that of cigarettes, that hookah was not
perceived as a tobacco product, that it was preferred due
its scents and flavors and that many false beliefs were
propagated such as its smoke being less dangerous due to
passing from water.
Intervention: Measurable objectives of our campaign,
campaign strategies and concepts were defined according
to the data collected through the pre-campaign analyses.We applied our campaign in four phases. In each
phase we delivered different messages and the public was
gradually informed of hookah facts through the use of
TV and radio ads, outdoor materials, brochures, PR
studies as well as the use of internet and social media
Results and lessons learnt: Following the campaign we
surveyed 1266 people in order to analyze the effectiveness of the campaign. The recall rate of the campaign is
high by 72,7%, in total group.TV ads recalled are
evaluated believable (83%) and effective since it
would discourage people from smoking hookah (77%)
and also taught persons something new (75%.) The
awareness level about the negative health effects of
smoking hookah significantly increased in post-campaign period. More than half of the people, who recall
the campaign say that their interest to try smoking
hookah diminished and also ads made them to advise
people nearby to quit smoking hookah.
Conclusions and key recommendations: Legal and

environmental regulations and restrictions are necessary
in order to prevent the hookah epidemic, in addition to
the social marketing and awareness campaigns. Restrictions on the production of flavored hookah tobacco
would positively impact our struggle against prevalent
hookah use, since the analyses of target population
showed that 88% of users preferred flavored hookah.
Legal restrictions on the fast spreading hookah cafes
would also reduce the prevalence of hookah use.
S549
health and government responses range from cautious

acknowledgement of potential cessation benefits to calling
for product bans. Current literature shows that key areas
of governments regulation include sale, use, advertising,
promotion, taxation and product classification. Many
countries have bans or laws that prohibit or restrict the
sale of e-cigarettes. Some regulate e-cigarettes as medicinal
products. Others regulate e-cigarettes as tobacco products
(or tobacco-related products). A very small number
regulate nicotine-containing e-cigarettes as poisons.
Results and lessons learnt: One size does not fit all.
However, there are common objectives which should
guide governments policy on e-cigarettes, e.g. Minimize
health risk to consumers; Ensure control over nicotine/
other substances; Prohibit unproven health claims and
marketing that prevents smokers from quitting, promotes uptake by non-smokers, youth and pregant
women;Protect tobacco control policies from interference. They can be adapted to specific national situation
and modified as the science on e-cigarettes evolves.
Conclusions: Public health and governments must
continue to conduct/use credible scientific evidence with
the objective of protecting lung health. This session
provides an overview of current government regulatory
responses and related issues.
72. Holding hands: high-level

cooperation on tobacco control
PC-1308-06 Lung health and e-cigarettes: a

review of recent government actions on ecigarettes and electronic nicotine delivery
devices
PC-1309-06 SmokeHaz: an example of a

scientific society-led advocacy tool for tobacco
control
M Wisotzky1 1International Union Against Tuberculosis and

Lung Disease, Edinburgh, UK. e-mail:
mwisotzky@theunion.org
C Vardavas,1 C Jimenez-Ruiz,2 J I De Granda ,3

C Gratziou4 1University of Crete, Rethymnon, Greece;
2
de Neumologa y Ciruga Toracica,
Sociedad Madrilena
Madrid, 3Hospital 12 de Octubre, Madrid, Spain; 4Athens
University, Athens, Greece. e-mail: maeve.barry@ersnet.org
Background and challenges to implementation: Research

on e-cigarettes safety and efficiacy as a cessation aid
continues to emerge. A single new study shows that ecigarettes compromise the immune system in the lungs of
mice and make their lungs more vulnerable to infection.
Human population studies show that, while the chemicals in e-cigarettes are at much lower levels than
traditional cigarettes, the long-term effects are still
unknown. Studies indicate that youth are taking up ecigarettes with some evidence that they may be more
likely to begin using traditional cigarettes. The huge
surge in marketing/promotion, investment by transnational tobacco companies, increased awareness and
use of ENDs, lack of long-term data on safety or
effectiveness for cessation creates challenges for governments making policy decisions to protect lung health.
Intervention or response: In response to the quickly
growing e-cigarette market, national and international
public health organizations, including The Union, have
assessed existing literature and research and developed
specific policy recommendations for governments. Public
Background and challenges to implementation: Smoking

is the leading preventable cause of morbidity and
mortality worldwide. At European level, approximately
700 000 EU citizens die prematurely every year because
of tobacco consumption and 13 millions are affected by
smoking related diseases. A large number of non smokers
are victims of passive smoke exposure.
Intervention or response: As a medical society, the
European Respiratory Society (ERS) is highly involved
in activities for Tobacco Control since 1990. Most
recently its Tobacco Control Committee created a new
website the SMOKEHAZ.eu to provide a one-stop web
platform that intended as an easily accessible and usable
source of information regarding the impact of active and
passive smoking on lung health in adults and children.
Results and lessons learnt: SMOKEHAZ.eu presents the
results of systematic reviews and meta-analysis of a large
number of studies selected by strict criteria, to assess the
effects of active and passive smoking on adults and
children respiratory health. This is a collaboration
project between ERS the University of Nottingham, the
S550
UK Centre of Excellence for Tobacco and Alcohol

Studies and the European Lung Foundation (ELF). The
new ERS site contains health messages based on scientific
evidence with specific figures and sources of the
information, so that all points can be easily traced to
the source and examined in detail.
Conclusions and key recommendations: The SMOKEHAZ website is a valuable and unique advocacy tool for
tobacco control targeted at policymakers, but also serves
as an educational tool for health professionals and other
researchers including relevant recommendation based on
scientific evidence.
PC-1310-06 Partner network consolidation for

national policy on tobacco control in Brazil
E Rangel,1 T Cavalcante,1 A P Teixeira1 1Ministry of Health,
Rio De Janeiro, RJ, Brazil. e-mail: ecavalcanti@inca.gov.br

network construction process was initiated in 1989 by
the National Programme for Tobacco Control articulated by the National Cancer Institute with the state and
municipal health departments. In 2001, this process has
become systematic and strengthened by the need to ratify
the WHO Framework Convention for Tobacco Control
and to combat the tobacco industry strategies trying to
delay the accession of Brazil. The Executive Secretariat of
the National Commission for Implementation of the
FCTC was created to support and expand the construction of the network partnership between Ministries,
Universities, Medical Societies, Congress and NGOs.
Intervention or response: The articulation of the network
is to monitor and establish contacts with organizations
that converge to strengthen the National Tobacco Control
Policy. The interaction is established by telephone,
narrowed by technical visits, technical training and
seminars for the empowerment of organizations referenced to become spokespersons of common interests.
Results and lessons learnt: The partnership allowed the
Brazilian State to ratify the WHO FCTC and implement
several measures of this treaty. As a result of the
partnership, in 2013 Brazil, as a key facilitator of the
Working Group of Articles 17 & 18 (Provision of
support for economically viable alternative activities &
Protection of the environment and the health of persons)
established by the Conference of the Parties to the FCTC,
received in the city of Pelotas, Rio Grande do Sul (federal
state with predominance of tobacco growing), 18
countries to demonstrate the implementation methodology of National Programme of Production Diversifica
tion on Tabacco Growing Areas
coordinated by the
Ministry of Agrarian Development. There was also the
commitment agreement signed between the Brazilian
Institute of Environment and Renewable Natural Resources (IBAMA), through Attorney Generals Office
(AGU), and the Interstate Association of Tobacco
Industry to establish ways of combating deforestation
in the Atlantic Forest for tobacco plantation.
Conclusions and key recommendations: The Brazilian
experience has enabled, empowered and legitimized
several organizations to position themselves in favor of

Tobacco Control Policies and provided great social
adherence to the advancement and strengthening actions.
The partnership has been a powerful strategy for
strengthening the National Policy on Tobacco Control.
PC-1311-06 Tobacco control in social networks:

the Brazilian experience with the virtual health
library prevention and cancer control
L Casado Costa,1 R Feijo,1 L Labrego1 1Instituto Nacional
de Cancer
Jose Alencar Gomes da Silva, Rio De Janeiro, RJ,
Brazil. Fax: (55) 21 3207 6068. e-mail: leticiac@inca.gov.br
Introduction: Social networking is an important media

strategy for the dissemination of information in virtual
environments with interactivity and sharing of information in real time and dynamically. Lately it has been used
by large companies and institutions. One way of
inserting tobacco control issues on social network in
Brazil was through the development of a fan page of the
Virtual Health Library Prevention and Cancer Control
on Facebook. The focus of this fan page is to promote,
through posts, issues related to cancer control. Smoking,
known as a risk factor for several types of cancer,
occupies a prominent place on this page which allows the
interested public to access quality information in a
friendly language.
Methodology: Posts are selected through a search on
Facebook and other websites, of issues which could be
interesting for the fan page public. Videos, papers,
information on legislation and other publications are
prioritized.
Results: The Facebook page became public in July 2013
and until the end of April, 1,925 people had liked it and
consequently became fans. Of this total, 81 % are
women and 19 % men. By sex and age group, 25 % are
women between 25 and 34 years old, 21 % between 35
and 44 years and 13 % are between 45 and 54 years.
Among males, 6 % of the fans of the page are between 25
and 34 years and 3 % between 45 and 54 years. Most
fans are Brazilians (1,841). Actually, there are fans in 24
countries around the world. Among Brazilians, 381 are
from Rio de Janeiro and 143 from Sao Paulo. This
distribution is related to the fact that INCA, leader of
VHL, is located in Rio de Janeiro. The most visualized
posts were, in first place, a video of a choir of 12 people
with laryngectomy, victims of tobacco and patients of
AC Camargo Cancer Center, which occupied a museum
in Sao Paulo performing Beatles songs. The post was
visualized by 3410 people and shared 41 times. In second
place, a news article regarding the World Conference on
Tobacco or Health on which the World Health Organization highlights the harms of waterpipes. The post was
visualized by 516 people and shared 20 times.
Conclusion: The Library page on Facebook has become
an important means of mobilizing people interested in the
topic. The disclosure of matters related to smoking on the
page represents a strong strategy for publicizing the issue,
in addition to attracting users to the library where they
have access to the publications available in its collection.
PC-1312-06 The role of Bloomberg Initiative

grants in tobacco control in India and
Bangladesh
B Gopalan,1 R J Singh,1 R Thakur,2 S S Mahbubul Alam,3,4
I Chowdhury4 1International International Union Against
Tuberculosis and Lung Disease (The Union), South-East Asia
Office, New Delhi, New Delhi, 2Consultant, Shimla, India;
3
The Union, Dhaka, Bangladesh. Fax: (91)11 4605 4430.
e-mail: bgopalan@theunion.org
Background: Studies have proved that the health outcomes of India and Bangladesh were highly affected due to
tobacco use. Both Countries have legislations in place but
implementations of activities are sub optimal at grass root
level. To support tobacco control policy development &
implementation, total grant of US$12.987 million (59
grants in 16 rounds) have been received. These grants are
managed by The Union with population coverage of 652
million in both these countries since 2007.
Intervention: The project team designed and implemented customized interventions as per the policy changes of
India and Bangladesh at par with MPOWER strategy.
The interventions focused on administrative framework
and infrastructure, Capacity Building, Enforcement,
coalition building and network, Public Education, Stop
Tobacco Industry Interferences and Policy focused
research, monitoring and evaluation. Grants were analyzed in terms of population coverage, funding support,
capacity building, smoke free jurisdictions, building
coalition and partnerships, inter-sectoral involvement.
Results: The activities under MPOWER implementation
are progressing towards achieving the objectives in both
countries. Grants have built the capacity of National and
sub national Governments and local public Health institutions, civil societies and law makers who directly and
indirectly associated with tobacco control in the region.
Description
India
Population coverage
under BI
Legislation
Amendment/
strengthening of
Legislation
Population protected
from secondhand
smoke
No. of BI Grants
Human capital
under existing BI
project
BI grants obligated
Mean utilization
Outcomes: Building
of stake holders
No. of persons
trained
No. of network
institutions
Rate of compliance
(SF) in project
area
No. of smokefree
jurisdictions
Bangladesh
539 million
2003
113 million
2005
20082014
2013
196 million (36%) 12 million (11%)

41
18
93
US$9.617 million
78%
71
US$3.370 million
85%
0.250 million
0.100 million
35
)6
90%
85%
90 (districts/cities)
16 854 (Public Places)
S551
Conclusion: The BI Grants contributed significantly in

prioritising tobacco control in India and Bangladesh.
National and sub national governments have benefited
from the BI resulting in overall implementation of the
MPOWER strategy in India and Bangladesh. BIs
continued support is the need of the hour till national/
sub national Governments takes complete lead in
Tobacco Control Activities.
PC-1313-06 WHO engagement with Ministry of

Finance in South-East Region
N N Kyaing,1 M Goodchild,2 A M Perucic,2 M Huq,3
V Munish,4 F Qureshi,5 D Kania5 1World Health
Organization South-East Asia Regional Office, New Delhi,
India; 2World Health Organization, Geneva, Switzerland; 3
World Health Organization Bangladesh, Dhaka, Bangladesh;
4
World Health Organization India, New Delhi, India; 5 World
Health Organization Indonesia, Jakarta, Indonesia. Fax: (91)
112 337 0197. e-mail: kyaingn@who.int
Background and challenges to implementation: Bangladesh, Sri Lanka and Thailand were the only countries in
the Region where the share exceeded 70% of the retail
price in 2012. While some countries in the Region
employ simple structures of tobacco taxation, others
have very complicated structures combining import
tariff, excise tax (specific and/or ad valorem), surcharge,
local tax and value-added tax. In most countries, the
policies have been targeted at cigarettes and have made
them more expensive; whereas other tobacco products
are subject to lower tax rates, and are allowed to be sold
at lower prices. These structures encourage smokers,
especially the poor, to switch from the more expensive
cigarettes to cheaper products such as hand-rolled
cigarettes, bidis and smokeless tobacco. Consequently,
tobacco control policies have become less effective.
Intervention or response: World Health Organization
had engaged with the Ministry of Finance (MOF) in
Member States of WHO South-East Asia Region
including Bangladesh, India, Indonesia, Myanmar, Nepal, and Thailand since 2009. MOF personnel get engaged
with WHO through courtesy calls, briefings, advocacy
workshops, symposia, trainings, study tours and secondment of MOF personnel at WHO, Geneva. WHO tax
simulation was applied in the trainings and symposia
using country data to estimate increase in tax revenues
after raising taxes and simplifying tax systems.
Results and lessons learnt: Bangladesh, India and
Indonesia had removed some tiers from the tax system
and simplify it. After series of meetings with MOF in
India, Ministry of Finance raised the taxes significantly in
2014. Bangladesh and India raised taxes on bidis too and
Myanmar has increased the taxes on non-cigarette
tobacco products and chewing tobacco twice. Thailand
introduced minimum specific floor and increased taxes
on roll your own products.
Conclusions and key recommendations: Engaging with
Ministry of Finance has brought significant increases in
tobacco taxes increasing government revenues and
reducing tobacco consumption. The engagement should
S552
be further enhancement and the same measures should be

applied in all Member States of WHO.
PC-1314-06 Government and community

established model of high compliance of
smoke-free provisions of legislation in Tehri, a
difficult hilly jurisdiction of India
G Tripathi,1 R J Singh,1 Y Pant,2 P Tripathi3 1International
Office, Delhi, 2Government of Uttrakhand, Tehri, 3University
of Delhi, Delhi, India. e-mail: tripathi.govind@gmail.com
Background: Indian tobacco control legislation (Cigarette and Other Tobacco Product Act-2003) prohibits
smoking in public places. The law mandates that a
specific signage informs people about smoke free status
of a public place must be displayed at prominent places.
Under the law, violators of the smoke free provisions will
be fined up to INR 200/- . Tehri (population 6 16 409 ),a
hilly district of Indian state of Uttarakhand, implemented
various steps for enforcing smoke free legislation through
massive awareness activities, series of capacity building
programmes followed by effective law enforcement .This
study conducted with an objective to assess the current
level of compliance to the smoke free provisions of the
law.
Design/Methods: Smokefree compliance surveys are
important tools to validate levels of compliance. An
unobtrusive cross sectional survey of randomly selected
405 public places in nine administrative blocks of Tehri
district was done in the month of January, 2015 by
trained investigators using pretested checklist. The five
core parameters of evaluation were: Presence of signage,
absence of active smoking, absence of smoking aids,
absence of tobacco litter and absence of tobacco smell.
Results: The No smoking signage informing general
public and tourists about smoke free provisions were
observed at 86 % of public places; While 95% of public
places were found without active smoking. 96.21%
public places were observed free from Smoking aids like
ashtrays, match boxes & lighters. More than 91% of
sampled public places didnt have any tobacco litter
(cigarette butts and bidi ends). Over 95% of public
places dint have evidence of recent smoking as evident of
absence of tobacco smell.
Conclusion: Tehri district has achieved high level of
compliance to smoke free provisions of the legislation as
a result of increased awareness among general public and
custodians of public places. Robust enforcement mechanism established. Pro-active district administration
involved all important stakeholders led to this historic
achievement. The administration has declared Tehri as
first smoke free district of the state. This model has
motivated and speeds up the Implementation mechanism
of tobacco control in entire hilly and difficult state.
PC-1315-06 Governments pro-public health

policies and effective coordination is key for
success in tobacco control: a case study from
Himachal Pradesh, India
K Thakur,1 R Thakur2 1Government of Himachal Pradesh,
Shimla, 2World Health Organization-Randstad, Shimla, India.
e-mail: healthmin-hp@nic.in
Background and challenges of implementation: Himachal Pradesh (Pop: 7 million, Area: 55 673 km2) is a state
in northern part of India. Despite the Indian tobacco
control legislation (COTPA), the state was been facing
huge tobacco burden. More than one-fifth of adult
population in the State is tobacco user. State is not
covered under Indias National Tobacco Control Programme; as a result the state had major resource
constraints in terms of manpower, finance and technical
expertise. However, an effective collaboration between
the state government, network of local NGOs (HPVHA)
and The Union- an International organization worked
very well and state became a model for tobacco control in
India.
Intervention or response: HPVHA generated massive
awareness among general public about the issue,
strategically carried out political advocacy with policy
makers. The Union provided funding and technical
assistance. Government issued relevant circulars, notified
squads and simplified the enforcement procedures.
Results and lessons learnt: Collaborated efforts resulted
in setting up of an institutional framework for implementation of tobacco control policies. Stringent enforcement was carried out across the state. Till May 2014,
more than 80 000 violations has been reported and near
Rs 9 million has been collected as fine amount which is
further utilized for tobacco control activities. TAPS
violations at points of sale are nearly rooted out from
the state. First conviction in the country under for TAPS
violations, pictorial health warnings on tobacco packs
and gutkha ban was carried out in the State. Based upon
the findings of compliance survey on an International
protocol, the state of Himachal Pradesh was declared
smoke free in July 2013 by Health Minister. WHO
awarded HPVHA and State government in the year 2011
and 2012 respectively for implementing tobacco control
policies effectively.
Conclusion and key recommendations: State Governments pro-public health policies formulation and implementationand strategic collaboration between Government and NGOs are always vital for successful
implementation of any public health initiative as
demonstrated in State of Himachal Pradesh. This model
can be replicated in other states in India or other
developing countries with similar settings. State Government is currently working on the plan to regulate the to
ban the single cigarette sale and mandatory licensing of
tobacco vendors.
S553
PC-1316-06 How a standardized capacity

building program scaled up National Tobacco
Control Program implementation in India
PC-1317-06 Potential of Institute for Global

Tobacco Controls online course for health
professionals capacity building: an evaluation
G Tripathi,1 R J Singh1 1International Union Against

Tuberculosis and Lung Disease, South-East Asia, New Delhi,
India. e-mail: tripathi.govind@gmail.com
T Ahmad1 1Doctors Network for the Eradication of Tobacco

(DOCNET), Moradabad, India. e-mail:
combatobacco@gmail.com
Background: India launched its National Tobacco

Control Programme (NTCP) in 2007-08 in 9 states (18
districts) and 12 states (24 districts) in 2008-09.Currently in 53 districts of 29 states. Under the BI initiative The
Union provide technical and financial support in 40
districts of 20 states with the objective to build
supportive political and administrative environment,
establish and strengthen the system, institutionalise
enforcement of provisions of National legislation. An
analysis conducted to understand that how effectively
standardised Capacity building program enhances the
pace of Tobacco control implementation.
Intervention: Analysis conducted of 8 state level, 20
district level, 11 law enforcement, 5 municipal corporations, 38 piggy back, 18 WNTD, 68 state and district
level training and review programs. In all 168 training/
reviews conducted and trained 2550 officials and
stakeholders under 8 types of different training programs
during Oct-2013 to Mar 2015. Efforts shown good
results and capacitated various state and district level
authorities, administrative heads, department heads,
local bodies, PRI members, media personals, civil society
members, Tax authorities, tourism authorities and
others. The core parameters of impact evaluation were:
Establishment of system, availability of IEC material,
establishment of enforcement mechanism, convergence
with other stakeholders, Active involvement for issuance
of directives/policy notifications, increased level of
engagement of Media and civil society organizations
Results: Tobacco control infrastructure established at
85% jurisdictions, availability of awareness related
material found at 90% places. Enforcement mechanism
is established at 60% places. Multi-stakeholders engagement initiated at 50% places of intervention areas.
Issuance of directives/policy notifications increased at
85% places. Increased level of engagement of Media and
civil society organizations at 95% places of the all
intervention areas.
Conclusions and recommendations: Despite many constraints, a standardised capacity building program under
NTCP project has resulted considerable rise in terms of
enhancing the capacity of stakeholders. The analysis
shows that standardised capacity building programme
can have substantial positive effects to implement the
NTCP programme in country. While the other tools such
as policy level intervention and related to the same, can
be readdressed to analyse further strategy for getting
good results.
Background: Capacity building of health professionals is

one of priority areas in the global fight against tobacco.
Institute for Global TobaccoControl (IGTC) at Johns
Hopkins Bloomberg School of Public Health, which is a
global leader in tobacco-control education andtraining,
has recently launched a short online course aimed at
health professionals (http://hp.globaltobaccocontrol.org/
online_training). This course entitled, Learning from the
Experts: A Course for Healthcare Professionals is
available internationally free of cost. Keeping in mind
the broader picture of capacity building in low-resource
settings, we have undertaken an evaluation of this online
course based on the structured feedback received from
MBBS students and interns enrolled in a medical college
in India.
Design/Methods: Kirkpatricks four-level evaluation
model for training programs was utilised for formulating
the design of this evaluation. This pre-, post-, follow-up
study was administered to a cohort of 121 final year
medical students and interns of Teerthanker Mahaveer
Medical College and Research Center, Moradabad. After
the baseline survey and briefing about the online course,
the participants were required to complete the online
course and report to the investigators. A pre-structured
questionnaire was used to collect the data in the baseline
as well as the post-course survey. Parameters assessed
included the feedback of study subjects on various
aspects of the course module design and perceived utility.
By comparing with the baseline response a change in
knowledge and attitude in relation to tobacco-control
was also assessed.
Results: Of the 121 students included in the study 97
students completed the post-course survey. 91% of these
rated the course as Excellent, while 9% of the subjects
rated it as Good. 93% believed that the course would be
Very much useful in their career as doctors, while 7%
believed it to be Somewhat useful. All the subjects
(100%) responded that they would Recommend this
course to their peers. In the baseline survey only 18%
expressed an interest in participating in a research related
to tobacco-control while in the post-course survey this
rose to 64% (P , 0.001). 98% of the students found the
course Easy to follow. 27% of the subjects in the
baseline survey responded with Strongly Agree to the
statement that tobacco-control ought to be taught in
greater detail in the medicalcurriculum, while in the postsurvey this rose to 72% (P , 0.001).
Conclusion: The online course Learning from the
Experts: A Course for Healthcare Professionals can help
fill the tobacco control curriculargap in medical schools
in a cost-effective way. The slide lecture with audio
format employed in the course delivery can work
intechnologically compromised settings also. Innovative
marketing strategies may be applied with designated
S554
funding to promotethis course amongst healthcare

professionals worldwide.
can significantly predict active TB and may suggest

preventive therapy if no contraindication.
73. IPT: attitudes, application and

adherence
PC-1318-06 Predicting development of active
tuberculosis in dialysis patients by serial
follow-up of latent tuberculous infection
C-C Shu,1,2 J-Y Wang,1,2 V-C Wu,1,2 F-J Yang,3 C-J Yu,1,2
L-N Lee,1,2 1National Taiwan University, Taipei, 2National
Taiwan University Hospital, Taipei, 3National Taiwan
University Hospital, Yun-Lin Branch, Yun-lin County, Taiwan.
e-mail: stree139@yahoo.com.tw
Background: Tuberculosis (TB) is still one of the most

common infectious diseases. In preventing from transmission of TB, early screening and treating patients with
latent TB infection (LTBI) will be a milestone to eliminate
TB, especially in TB closely contacts, HIV-infected
patients, and patients using immunosuppressants. However, the efficacy for predicting active TB by LTBI is
unclear in patients requiring dialysis, who actually have
10~25 fold risk of active TB. Therefore, we conducted
the present study following dialysis patients LTBI status
and analyze the performance predicting active TB.
Design: Patients received long-term dialysis (73 months)
in multiple medical centers in Taiwan were screened from
April 2011 to March 2013 under approval of institutional review board. We excluded patients with past
history of active tuberculosis, recent exposure to
pulmonary TB, and HIV infection or recent use of
immunosuppressant. Finally, 876 were enrolled and
received Quantiferron-TB Gold In-tube (QFT-GIT) test,
used for defining the status of latent TB infection. Among
them, 589 patients received QFT-GIT follow up after six
months. Primary outcome was active TB, defined by
mycobacterial culture positive for Mycobacterium tuberculosis or typical histology finding plus treatment
response.
Results: The enrolled 876 dialysis patients were at
average age of 62.3 years (SD: 13.2) and had male
proportion of 53% and hemodialysis at 84.2%. The first
time of QFT-GIT was positive in 192 patients (21.9%),
negative in 656 (74.9%) and indeterminate in 28 (3.2%).
The QFT-GIT six months later reported 98 (16.7%)
positive and 474 (80.6%) negative. Among them, there
was 69 patients with persistent positive QFT-GIT. In the
following up, six (0.68% of 876) patients had been
diagnosed as active TB including 3 plueral TB, 1
pericarditis, 1 disseminated type and 1 pulmonary TB.
Only 3 (0.51% of 589) of them was diagnosed after 6
months. Two of them had initial positive QFT-GIT (1%
of 192) and persistent positivity of QFT-GIT (2.9% of
69). In Kaplan Meier curve, the persistent positive QFTGIT results was significant associated with active TB
development (P 0.004, log rank test).
Conclusion: In dialysis patients, LTBI defined by QFTGIT test, especially persistent positive in six months later,
PC-1319-06 Perceptions and knowledge of

health care workers plays a role in increasing
isoniazid preventive therapy uptake amongst
people living with HIV
M Calnan,1 M Pasipamire,2 S Mazibuko,2 S Haumba,1
N Shongwe1 1University Research Co. Mbabane, 2Swaziland
National AIDS Programme, Mbabane, Swaziland. Fax: (268)
2404 7199. e-mail: maricalnan@yahoo.co.uk
Background and challenges to implementation: In 2007,

Swaziland implemented the TB-HIV policy guidelines
but only in 2011 were the 3 0 Is guidelines produced
emphasizing provision of Isoniazid Preventative Therapy
(IPT) to reduce the risk of active tuberculosis among
people living with HIV. Health care workers (HCWs)
received orientation to the 3 0 Is guidelines and facilities
supported to initiate IPT. However, the uptake for IPT
has remained low (9% nationally). This study aimed to
determine HCWs perceived role in IPT uptake.
Intervention or response: Mixed methods approach
including focus group discussion (FGD) for HCWs
stationed in HIV clinics and a collection of retrospective
data on IPT initiations, adherence and outcomes from
electronic medical record and pharmacy databases,
patient files and IPT log books in 4 high volume HIV
clinics. Data was from patients initiated on isoniazid
from January 2012 through to December 2013 and
analysed using StatsDirect
Results and lessons learnt: The FGD revealed that 50%
(8/16) of HCWs were aware of IPT but are reluctant to
provide it due to fears of creating isoniazid resistance or
Isoniazid side effects. The other 50% were not knowledgeable about IPT and thus could not talk to patients
about it. Documentation by the HCWs in patient files
and IPT log book was poor with 45% of patients not
being assigned an outcome. Outcomes were retrieved
from the electronic database and pharmacy database.
Between January 2012 and December 2013, only 1.1%
(445/38 756) of eligible patients were initiated on IPT
across the four facilities. Of the 350 records reviewed,
67% were female and 95% were aged 20 to 60 years.
Only 32% (112) completed treatment, while 56% (196)
defaulted their IPT, 5% were LTFU and 2% (7)
transferred out. Of patients who had a documented
default date, 30% adhered through 4 months of IPT and
28% defaulted after 1 month. Among the defaulters,

88% adhered to ART treatment for 22 months after the
default.
Conclusions and key recommendations: The data suggests that patient adherence or clinic attendance and
accessibility are not the major problems; rather Healthcare worker perceptions and knowledge affect IPT
service delivery negatively and need to be addressed as
a priority through training and ongoing mentoring.
PC-1320-06 Predictors of isoniazid preventive

therapy completion among adults newly
diagnosed with HIV in rural Malawi
K Little,1 G Barnes,2 M Khundi,3 L G Ngwira,3
R E Chaisson,2 A Nkhoma,3 L Corbett,3,4 D Dowdy1 1Johns
Hopkins Bloomberg School of Public Health, Baltimore, MD,
2
MD, USA; 3Malawi-Liverpool Wellcome Research Trust,
Blantyre, Malawi; 4London School of Hygiene & Tropical
Medicine, London, UK. e-mail: klittle@jhsph.edu
Background: As a part of comprehensive HIV care, the

WHO recommends that people living with HIV (PLHIV)
receive at least 6 months of isoniazid preventive therapy
(IPT) once tuberculosis (TB) disease has been excluded.
IPT reduces the risk of TB disease in PLHIV, both before
and during antiretroviral therapy (ART). However,
completion of IPT remains a major challenge in
Design/Methods: We evaluated predictors of IPT completion (defined as receipt of 7150 doses of isoniazid) in
newly diagnosed PLHIV recruited from 8 rural primary
clinics in Malawi, as part of a cluster-randomized trial of
TB screening. Participants who presented for antenatal
care or symptomatic diagnosis were screened for TB and
offered IPT on the same day as their HIV test. Individuals
reporting one or more symptoms of TB (cough, fever,
night sweats, and/or weight loss) received further
microbiological evaluation via sputum smear microscopy
or Xpert MTB/RIF before initiating IPT in order to rule
out active TB disease. ART initiation criteria included
WHO Stage 3 or 4, CD4 cell count ,350, or pregnancy/
breastfeeding. Our analysis includes all participants who
started IPT by June 1, 2014. Predictors of IPT
completion were evaluated using a multilevel logistic
regression model with patient- and clinic-level characteristics.
Results: Of 828 trial participants screened negative for
active TB and started on IPT, median (interquartile
range, IQR) age was 34 years (26-40), and 67% were
women (551/828, 66.6%) (Table 1). Overall, 612
(73.9%) completed 7150 doses of IPT, although 8%
of these individuals reported treatment interruptions of
72 months before completion. The mean duration of
IPT taken was 190.3 days (IQR: 167-202) in completers
and 90.3 days (IQR: 28-139) in non-completers (P ,
0.001). On univariate analysis in a multi-level model,
older age (OR: 1.21, 95%CI: 1.10-1.32), and receipt of
ART (OR: 3.26, 95%CI: 2.00-5.31) were significantly
associated with IPT completion. After controlling for
potential confounders, concomitant ART (aOR: 3.85,
S555
95%CI: 2.27-6.52) and older age (aOR: 1.16, 95%CI:

1.03-1.30) significantly increased the odds of IPT
completion, while WHO stage III/IV at baseline was
significantly associated with decreased odds of completion (aOR: 0.51, 95%CI: 0.31-0.85).
Conclusion: In this population of rural Malawian adults
recently diagnosed with HIV, completion of IPT was
associated with concomitant receipt of ART, older age,
and WHO stage. Our findings highlight the importance
of additional efforts to ensure IPT completion among
younger HIV patients, those with clinically advanced
HIV, and those not immediately eligible for ART.
PC-1321-06 Population-level impact of different

durations of isoniazid preventive therapy
among people living with HIV
A Kunkel,1,2 F Crawford,2 T Cohen2 1Harvard School of
Public Health, Boston, MA, 2Yale School of Public Health,
New Haven, CT, USA. e-mail: agkunkel@gmail.com
Background: The World Health Organization recommends the use of at least 6-9 months of isoniazid
preventive therapy (IPT) among people living with HIV
(PLHIV) who are deemed unlikely to have active TB on
the basis of symptom screening. Longer durations of IPT
among PLHIV could further reduce risks of TB at both
the individual and community levels. However, extending the duration of IPT could also exacerbate concerns
about the risk of side effects and the potential for
increased resistance to isoniazid.
Design/Methods: We have created a mathematical model
to assess the potential health costs and benefits of varying
IPT durations in terms of TB incidence, drug resistance,
and mortality. Our model incorporates TB transmission
and treatment dynamics as well as HIV incidence and
antiretroviral uptake. Unlike previous IPT models, this
model simulates IPT on the backdrop of existing
isoniazid, rifampicin, and multi-drug resistant TB and
includes mixed infections with up to four distinct strain
types (pan-sensitive, isoniazid mono-resistant, rifampicin
mono-resistant, and multi-drug resistant). We estimated
model parameters using data from the joint TB and HIV
epidemics in Botswana.
Results: Using this model, we evaluated a range of IPT
durations in Botswana, using mortality as the primary
outcome measure to evaluate the composite effects of IPT
on isoniazid-sensitive and isoniazid-resistant TB. Our
results show that changing the duration of IPT would
affect the prevalence of both isoniazid-sensitive and
isoniazid-resistant TB. The optimal duration of IPT
varies depending on the time horizon of interest, with
more aggressive IPT strategies supported in some
circumstances despite increases in the prevalence of drug
resistance.
Conclusion: These results suggest that concerns about
increases in isoniazid resistance that could result from
longer IPT durations should be balanced against the
potential benefit of reductions in drug sensitive TB. The
development of novel TB drugs and regimens will play a
crucial role in mitigating concerns about future increases
S556
in isoniazid resistance resulting from extended IPT

durations among PLHIV.
PC-1323-06 Improving implementation of

isoniazid preventive therapy through
community-based referral in rural South Africa
PC-1322-06 Biologic anti-rheumatoid agents

alter interferon-gamma release assay results
for detection of latent tuberculous infection
S Shenoi,1 A P Moll,2 T Van Der Merwe,2 P Vranken,3

L Andrews,1 R P Brooks,1 E K Dokubo,4 G Friedland1 1Yale
University School of Medicine, New Haven, CT, USA; 2Church
of Scotland Hospital, Tugela Ferry, 3Centers for Disease
Control and Prevention, Pretoria, South Africa; 4U.S. Centers
for Disease Control and Prevention, Atlanta, GA, USA. Fax:
(1) 203 737 4051. e-mail: sheela.shenoi@yale.edu
W Thanassi,1,2 J Chang,1 J Sokolove1,2 1Department of

Veterans Affairs, Palo Alto, CA, 2Stanford Medical Center,
Palo Alto, CA, USA. e-mail: wendy.thanassi@gmail.com
Background: Since the risk for development of active

tuberculosis disease is several-fold greater in individuals
taking immunosuppressive medication for rheumatoid
arthritis and other autoimmune conditions, screening for
tuberculosis (TB) infection is routine practice. Testing for
latent TB using interferon-gamma release assays (IGRA)
is commonplace. Our objective was to determine the
influence of two common biologic anti-rheumatoid
agents on the results of the two available interferongamma release assays in a known positive test result
cohort.
Methods: Whole blood from ten adults in California,
United States who had childhood TB exposure risk
factors and who had tested positive by IGRA twice
previously was exposed in-vitro to a 1:120 dilution
(twice therapeutic) of etanercept (Enbrele) and abatacept (Orenciae). The IGRAs elispot (TB.TSPOTe) and
the enzyme-linked immuno-assay (ELISA)
(QuantiFERONe) (QFT) were run on the drug-exposed
and unexposed samples by Oxford Laboratories, Tennessee and in the Palo Alto Veterans Affairs Hospital,
California clinical laboratory, respectively. Control tubes
had an equal volume of polyclonal immunoglobulin
without biologic agent added. Results of the assays were
compared against their nils by paired t-test.
Results: One individual was excluded for excessive
reactivity at baseline. All nine remaining participants
had declines in their QFT result with abatacept (P ,
0.001) and etanercept (P , 0.01). Individual declines
varied from 7-60% with abatacept and from 34-81%
with etanercept (see Figure ). Neither the ESAT-6 nor
CFP-10 TB antigen panels on the elispot showed a
significant change in response to either drug.
Conclusions: While anti-tumor necrosis factor-a chemotherapy has been associated with decreased IGRA
response to tuberculin antigens, such response has not
been shown with the CTLA-4 Ig agents abatacept or
belatacept. We theorize that the elispot did not show the
same response decline as the elisa IGRA because the cells
are washed prior to exposure to TB antigen, which may
result in a more accurate immunologic response. Such
findings have implications for the selection of IGRA used
in TB screening of adults being treated by the chemotherapeutic agents and drug classes.
Background: Isoniazid preventive therapy (IPT) is an

effective World Health Organization-endorsed strategy
to reduce TB incidence among HIV-positive persons.
South Africa has among the highest TB incidence rates
globally (860/100 000) and more than 5 million people
living with HIV. IPT has been slow to be effected and faces
obstacles to widespread implementation. New strategies
to improve its successful uptake are urgently needed.
Design/Methods: Through a community-based intensive
case finding program in rural South Africa, we identified
HIV-positive community members without fever, cough,
night sweats, or weight loss and referred them to the
government health care clinics for IPT. We measured IPT
uptake, provided adherence training, and measured
monthly monitoring, and six-month treatment completion. Adherence was measured by self report and
isoniazid metabolite testing in urine.
Results: Among the first 184 community members
enrolled, the median age was 35 (IQR 29-45), 152
(82.6%) were female, and 92 (50%) had CD4 .350
mm3. Subjects made 640 (85%) of 752 scheduled visits;
to date, 58 have completed the IPT course, 3 were lost to
follow up, and 1 patient died of unrelated causes. Among
640 urine tests conducted, 603 (94%) demonstrated INH
metabolites. Self report overwhelmingly (93%) concurred with urine testing results.
Conclusion: Preliminary data suggests that community
members in TB and HIV endemic areas are interested in
TB prevention and willing to participate in IPT.
Community members who are HIV-positive but have
not entered treatment can be engaged in care through
IPT, and even if not yet qualifying for ART by CD4
criteria can successfully adhere to IPT alone. Self report
may be a successful predictor of adherence. Offering IPT
to those not yet engaged in HIV care may be a viable
method of expanding IPT and reducing TB incidence.
PC-1324-06 Impact of pregnancy on latent

tuberculous infection diagnostics in HIVinfected pregnant women in western Kenya
S Lacourse,1 L Cranmer,1,2 D Matemo,1,3 J Kinuthia,1,3
D Horne,1 G John-Stewart1 1University of Washington,
Seattle, WA, 2Emory University School of Medicine and
Childrens Healthcare of Atlanta, Atlanta, GA, USA;
3
Kenyatta National Hospital, Nairobi, Kenya. e-mail:
smlacourse@gmail.com
Background: Maternal TB-HIV co-infection is associated

with poor maternal and infant outcomes. Pregnancy
provides a unique opportunity for TB prevention efforts.
Latent TB infection (LTBI) treatment benefit is most

clearly demonstrated in those with evidence of M.
tuberculosis infection by tuberculin skin tests (TST) or
interferon-gamma release assays (IGRA). The impact of
pregnancy on LTBI diagnostics in HIV-infected women
in sub-Saharan Africa is unknown.
Design/Methods: In this ongoing longitudinal study, LTBI
testing was performed on HIV-infected pregnant women
in two antenatal clinics in western Kenya. TST and IGRA
[QuantiFERON (QFT)] were performed at enrollment
during pregnancy. Sputum culture was performed irrespective of symptoms. Women with negative or discordant QFT/TST were retested 6 weeks postpartum.
Results: Of the 100 women enrolled between August
2014-January 2015 with negative sputum culture, 65
have completed post-partum follow-up with median age
of 27 years (IQR 22-32), CD4 of 563 cells/ll (IQR 315780), and gestational age of 30 weeks (IQR 24-32). The
majority (95%) were taking combination antiretroviral
therapy. More women had positive QFT vs. TST at
baseline (21/65 [32.3%] vs. 6/65 [9.2%], P0.06); 6.2%
were concordant positive (QFT/TST) and 50.8% were
concordant negative. Ten (15.4%) women had indeterminate QFT (9 TST-, 1 TST). IGRA/TST agreement in
pregnancy was 56.9% (j 0.13, 95%CI 0.02-0.21).
Between pregnancy and 6 weeks postpartum, 9 (16.1%)
had TST conversion; 1 had QFT conversion. LTBI test
reversion occurred in 3 women (1 TST, 2 QFT). The
majority of women with indeterminate pregnancy QFTs
(9/10) were negative at postpartum. Postpartum IGRA/
TST had higher agreement (86.2%, j0.61 95%CI 0.360.85), with 15/58 (25.9%) and 11/58 (19.0%) positive
QFT and TST, respectively (P 0.40). Overall mean QFT
IFN-T was higher in postpartum (.96 vs. 2.43 IU/ml, P
0.04). Baseline report of TB contact was associated with
postpartum positive QFT but not TST (QFT OR 3.8,
95% Cl 1.1-13.7 vs. TST OR 1.7 95%CI 0.4-6.7).
Conclusion: LTBI testing performance varies by pregnancy stage among HIV-infected pregnant women in western
Kenya, with more QFT positive women compared to TST
during pregnancy. TST conversion in the early postpartum period in the setting of relatively consistent IGRA
positivity may reflect pregnancy associated TST anergy. A
reliance on TST could miss a considerable number of
women potentially benefiting from LTBI treatment.
S557
PC-1325-06 Tuberculosis screening and

isoniazid Preventive therapy provision among
patients initiating antiretroviral therapy,
Nigeria, 20042012
I Pathmanathan,1 E K Dokubo,1 S Ray,1 D Onotu,2
S Odafe,2 I Dalhatu,2 H Debem,2 S Agolory1 1U.S. Centers
for Disease Control and Prevention, Atlanta, GA, USA;
2
Centers for Disease Control and Prevention, Abuja, Nigeria.
e-mail: ydi6@cdc.gov
Background: Tuberculosis (TB) is the leading cause of

mortality among people living with HIV (PLHIV), and
often goes undiagnosed. Nigeria had the most AIDSrelated deaths worldwide in 2013 (210 000), of which
39% were associated with TB. Since 2004, the World
Health Organization (WHO) has recommended collaborative TB and HIV activities, including routine screening of all adult PLHIV for TB using a four-symptom
screen, and the offer of isoniazid preventative therapy
(IPT) to those unlikely to have active TB. We assessed TB
screening and IPT provision among adults initiating ART
in Nigeria between 2004 and 2012 to identify missed
opportunities for TB diagnosis and prevention.
Design/Methods: We analyzed retrospective cohort data
from a nationally representative sample of ART patients
715 years old who were enrolled in U.S. Presidents
Emergency Plan for AIDS Relief (PEPFAR)-supported
care and treatment programs from 2004 to 2012. Data
were abstracted from 3,496 patient records at 35 sites
selected using probability-proportional-to-size (PPS)
sampling. Analyses were weighted and controlled for
the complex survey design. Logistic regression was used
to assess if ART start year was associated with TB
screening among PLHIV initiating ART.
Results: Ninety-two percent (95%CI 91.1-92.9%) of
patients initiating ART were screened for TB. Of those,
4.2% (95%CI 3.2-5.4%) were already on TB treatment,
1.2% (95%CI 0.6-2.3%) reported prior TB, and 6.7%
(95%CI 4.7-9.6%) were documented as having suspected TB, indicating a positive TB symptom screen.
There was no statistically significant association between
year of ART enrollment and TB screening, with those
enrolled in 2010-2012 having an odds of being screened
only 1.1 times greater than those enrolled in 2004-2006
(OR: 1.1; 95%CI 0.5-2.7). Only 0.5% (95%CI 0.21.1%) of patients eligible for IPT (without suspected or
prior TB and not on TB treatment) were on IPT at ART
initiation.
Conclusion: A high proportion of PLHIV initiating ART
are screened for TB in Nigeria; HIV facilities should
continue to screen these patients at every clinical
encounter to facilitate timely TB diagnosis and treatment
for co-infected patients. IPT uptake in Nigeria for eligible
PLHIV is poor. An effective implementation system is
required to increase provision of IPT for PLHIV with a
negative TB symptom screen, in accordance with WHO
and national guidelines, in order to reduce their risk of
developing active TB.
S558
PC-1326-06 Modelling the post-treatment

effect of isoniazid plus antiretroviral therapy
on tuberculosis incidence
T Sumner,1 R Houben,1 G Maartens,2 M Rangaka,3
R White1 1London School of Hygiene & Tropical Medicine,
London, UK; 2University of Cape Town, Cape Town, South
Africa; 3University College London, London, UK. e-mail:
tom.sumner@lshtm.ac.uk
Background: Trials of isoniazid preventive therapy (IPT)

for people living with HIV in high tuberculosis (TB)
incidence settings in the absence of antiretroviral therapy
(ART) have shown limited duration of protection against
TB disease following completion of therapy. This may be
due to reinfection or reactivation of existing infections
which were insufficiently treated. Recently published
data from a pragmatic trial of IPT among people on ART
in Khayelitsha, Cape Town, South Africa (Rangaka et al,
2014) suggested that the effect of IPT persisted in the first
year after completion of IPT, longer than observed in
trials in ART-naive individuals. We used mathematical
modelling to explore if this increased duration of
protection may be due to an increased curative ability
of IPT when given in combination with ART.
Methods: A mathematical model describing Mycobacterium tuberculosis transmission in a cohort of HIVinfected individuals after completion of IPT was used to
analyse data from the Khayelitsha trial to estimate the
annual risk of infection (ARI) and the proportion of
individuals whose latent TB infection (LTBI) was cured
by IPT.
Results: The model captures the trends in TB incidence in
the placebo and interventions arms observed during the
trial. The estimated ARI was 4.0% (2.6-5.8) and the
estimated proportion of individuals whose LTBI was
cured following IPT was 35.4% (2.4-76.4%). The
proportion cured was sensitive to assumptions about
isoniazid resistance; if all infections were assumed to be
susceptible the proportion cured was lower at 23.2%
(0.1-61.5%). Immediately following completion of IPT
the majority of disease in the model was due to
reactivation of uncured infection, however over time
the proportion of disease due to recent infection
increased. The cumulative proportion of disease that
occurred in individuals cured by IPT was estimated to be
7.8% (0.4-32.1).
Conclusions: Our results suggest that IPT can cure LTBI
in approximately 35% of HIV positive individuals also
taking ART, and therefore provide protection beyond the
period of IPT. In low incidence settings, 12 months of IPT
in combination with ART may provide additional long
term benefit to HIV-positive individuals. However the
durability of this protection will be limited in high
incidence settings where there is a continued risk of
exposure to re-infection and continuous preventive
therapy together with improved infection control efforts
would be required to provide long-term protection
against TB.
SHORT ORAL ABSTRACT SESSIONS

01. Across the spectrum of TB
management
SOA-600-06 New strategy for the treatment of
tuberculosis with hepatotoxicity free antituberculosis drug
C-Y Tsai,1 T-Y Shih,1 H-T Ho,1 W-J Su,2 Y-W Huang,3
W-C Perng,4 F-Y Chang,4 O Y-P Hu1 1National Defense
Medical Center, Taipei, 2Taipei Veterans General Hospital,
Taipei, 3Chang Hua Hospital, Chang Hua, 4Tri-Service General
Hospital, Taipei, Taiwan. e-mail: chanyentw@gmail.com
Background: Tuberculosis (TB) is a major cause of

morbidity and mortality worldwide. The World Health
Organization (WHO) reported that one-third of the
worlds population has latent TB, and roughly 9 million
cases of active TB emerge annually, resulting in 2-3
million deaths. These data exemplify the importance of
the problem, especially as most new cases occur in
populated nations, including India and China. Currently
available anti-TB drugs (isoniazid [INH], rifampin [RIF],
and pyrazinamide [PZA]) cause serious hepatic side
effects, especially in patients with liver dysfunction.
Cytochrome P450 2E1 (CYP2E1) and amidase synergistically contribute to RIF- and INH-induced hepatotoxicity. To prevent anti-TB drug-induced hepatotoxicity, a
new, hepatotoxicity-free formulation of traditional antiTB drugs was developed by inhibiting CYP2E1 and
amidase activity.
Methods: Numerous known pharmaceutical excipients
were screened as CYP2E1 or amidase inhibitors in vitro
and in vivo. The most active inhibitor was used to
develop a hepatotoxicity-free formulation with INH,
RIF, and PZA.
Results: An FDA-approved, classified as Generally
Recognized As Safe (GRAS), excipient was shown to
be a CYP2E1 inhibitor. This excipient can also inhibit
amidase activity at similar concentrations. In both
animal and human studies, the selected agent significantly inhibited the activity of CYP2E1 (up to 58%) and
amidase (up to 70%) in vivo. Pathology data revealed
that drug-induced hepatotoxicity was nearly eradicated.
Furthermore, the inhibitor did not affect the area under
the curve (AUC),maximum concentration (Cmax), and
anti-TB effects of active ingredients (INH, RIF and PZA).
A randomized, double-blind, active drug controlled,
multi-center phase III clinical study is currently underway in 15 medical centers to assess the efficacy of the
new anti-TB formulation to reduce anti-TB drug-induced
liver injury.
Conclusion: We have developed a new, hepatotoxicityfree formulation of traditional anti-TB drugs, which has
successfully entered phase III study. We expect this
hepatotoxicity-free anti-TB drug at high doses will both
shorten TB treatment time and improve compliance in
TB patients. Consequently, this may decrease the
incidence of multi-drug resistant (MDR)-TB. Thus, this
hepatotoxicity-free anti-TB drug will dramatically

change TB prevention and treatment.
SOA-601-06 High transfer-out affecting

hospitals performance in tuberculosis
treatment outcomes
1
S Gebreyes, D Habte, Z Dememew, I Jemal Abdulahi,

D J Dare,1 M Nigussie,3 M Aseresa Melese,1 P G Suarez4
1
Low TB Performance (HEAL TB), Addis Ababa, 2Oromia
Regional Health Bureau, Addis Ababa, 3Amhara Regional
Health Bureau, Bahar Dar, Ethiopia; 4Management Sciences
for Health, Center for Health Services, Arlington, VA, USA.
Fax: (25) 111 661 6824. e-mail: solomon_ngsh@yahoo.com
Background and challenges to implementation: Directly

Observed Treatment for TB patients is implemented at
hospital and primary health care unit levels. The USAIDfunded Help Ethiopia Address the Low Performance of
TB (HEAL TB) project in collaboration with the national
TB program (NTP) has been providing technical and
material support to all the Hospitals and Health Centers
(HCs) in Amhara and Oromia Regions since 2011.
Intervention or response: NTP with support from HEAL
TB provided capacity building for TB program managers
at different levels, training of health professionals at
hospital and HC levels, laboratory capacity building
including quality assurance, mentoring and supportive
supervision. In this report, we compared the TB
treatment outcomes between Hospital and HC registered
TB patients in the 10 zones of the initial phase of project
implementation.
Results and lessons learnt: The project supported
treatment of 18 874 new smear positive pulmonary TB
(SS) patients in the hospitals (1208) and HCs (17 666)
between October 2011 and June 2014. Treatment
success rate (TSR) among hospital enrolled patients
improved from 59.6% at baseline to 80.2% at the end of
the third year and from 85.7% to 90.9% in the HCs. The
improvement in cure rate (CR) over three years in the
hospitals was from 51.6% to 74.2% whereas the
corresponding progress in the HCs was from 69.1% to
85.3% (Figure 1). The average rate of annual decline in
unfavorable outcomes (transfer out (TO), default, death
S559
and treatment failure) was higher in the hospital group as

compared to the HC counterpart (19.1% vs.4.0%, P ,0
.05). The baseline and latest TO in the hospital were
28.7% and 10.3 % in contrast with the 6.4% and 2.5%
at the HCs (P , 0.05). Default and death showed
decreasing trend in both hospitals and HCs with the
latest rate for the hospital and HCs being 2.8% and
1.5% default (P . 0.05) respectively; and 2.8% and
2.6% deaths (P . 0.05) respectively. Treatment failure at
the end of the third year was 4% and 2.4% (P . 0.05) at
hospitals and HCs respectively.
Conclusions and key recommendations: The relatively
higher TO in the hospitals and failure of the hospitals to
trace the final outcome underestimated their TSR and
CR as compared to the HCs. In the hospitals, the TO rate
has declined significantly in three years because of the
referral of hospital diagnosed TB patients to the patients
nearest health center for DOT. Further effort is required
to reduce the TO in hospitals and trace the outcome from
the receiving health facilities.
SOA-602-06 Thrice-weekly dose of antituberculosis drugs: efficacy under Indias

program conditions
R Kakar1 1RNTCP, District TB Center, Faridabad, State
Haryana, India. e-mail: raman24march@yahoo.com
Background: When discovered, anti tuberculosis drugs

were always administered in Daily-doses. That worked
wonders! Later, some studies claimed that drugs were
equally effective even if given as thrice-weekly doses
(intermittently). Not everyone believed that. Three
Reviews (Cochrane, Thorax, Azhar) appear unconvinced! Most nations worldwide (including 20 High TB
Burden Countries) remain skeptic and continue to
practice time-tested Daily dosing! However, India is
exception. In 2000, India adopted a unique thrice-weekly
regimen. During 14 years, under national program,
Thrice-weekly doses have been administered to 14.2
million TB patients, self-claiming 85% success. Objective: To revisit long-term fate of all registered patients;
identify the ones, who clearly didnt achieve lasting cure
with a single course of thrice-weekly treatment, were
later re-registered for a second innings; thus, their names
figure twice in national registry (nicknamed herein as
Repeaters). Setting: Author heads District Tuberculosis
Center, Faridabad, India; his job has been to monitor
thousands of current / ex-patients on Thrice-weekly
regimen, by providence, offering him unique access to
patients, data and entire network of TB workers of
district.
Design/Methods: Two methods of data collection:
Retrospective Record Review: Author motivated district
work-force; jointly conducted systematic review of handwritten registers of past 14 years, one at a time thus
identified 1575 Repeaters. Clinical Audit: Author
conducted exhaustive clinical audit in his busy TB
OPD, cherry-picked (on an average) one repeater per
day; thus tracked 1600 Repeaters in 42 months (Of these,
800 were referred in for Re-treatment, duly diagnosed at
S560
Delhis premier TB institutes and needed local registration).

Results: Long-term fate of 27 120 registered TB patients
was analyzed. 3175 patients noticed to have returned
sick, meriting Re-treatment; thus over 11.7 % identified
as Repeaters - registered twice/thrice (names, unique
govt. IDs, irrefutable evidence enclosed) revealing an
unfavorable outcome, unreported anywhere thus far - a
bubble within. Besides, study found 1282 (5.4%) Deaths;
2271 (9.6%) Defaulters recorded similarly with
clinching details.
Conclusion: Under Indias routine program conditions,
thrice- weekly regimen is ineffective. Too many (11.7%)
patients come back sick, which may promote drugresistance on industrial scale!
SOA-603-06 Assessing the quality of diagnosis,

treatment service and adherence monitoring
by private practioners to TB patients in India: a
case study from Himachal
A Singh1 1Office of the CMO, Solan, India. Fax: (91) 179
222 4151. e-mail: dpo3solan@gmail.com
Background: Private Practioners in India treats more

than half of tuberculosis patients; infact are the first point
of contact. Private sector often has been blamed for nonstandardized diagnosis, irrational treatment and biggest
contributor to spreading drug resistance. Present study
was conducted with an objective to assess quality of
diagnosis, treatment service and adherence monitoring
provided by private practioners in Himachal Pradesh in
India.
Design/Methods: In April 2015, a list of TB patients
diagnosed, initiated on treatment and notified by private
sector from third quarter 2012 to 1st quarter 2015 in
Solan district in Himachal Pradesh was extracted from
the web based system called Nikshay. After obtaining
consent from patients, each one of them were contacted
and interviewed; responses were recorded in a case
record form. Medical records were also studied in details.
Results: Out of total 126 patients notified by the private
sector, 90 patients and attendants of four deceased
patients were interviewed. Cough (85%) was the most
common symptoms reported by the patients. Near onethird patients also reported fever along with cough. Few
patients (16%) also reported loss of weight. X-ray
examination (79%) was the most common investigation
followed by blood investigation (74%), sputum smear
examination (55%), pleural fluid cytology (22%), and
FNAC (1%). Patient also reported a variable treatment
history; 11% patients were on single drug, 28% patients
were on two drugs, whereas near half of the patients were
on three drugs regimen. Near one-fourth of patients were
also given injections. No INH prophylaxis was given to
children lesser than six years of age who were in contact
of 13% of cases. Treatment duration ranged from three
months in 13% patients to .1 year in 8%. All doses were
self-administered. 12 cases opted for taking DOTS in the
public hospital. None of the patients were monitored by
the PP. Total 22 (23%) cases were cured, 46 (48%)
relapsed, 17 (18%) defaulted, 5 (5%) were failure and 4

(4%) died.
Conclusion: There are major limitations in the quality of
diagnostic and treatment services by the private practioners, which is reflected as a poor treatment outcome.
Non-standardized diagnostics and inappropriate treatment regimen by PP are a major challenge for tuberculosis control and a cause for impending drug resistance.
Each private practioner must be provided training on
Standard for TB care in India.
SOA-604-06 Factors for diagnostic delay in

smear-positive tuberculosis patients: Patna,
India
S Mannan,1 S Dhawan,2 S Upadhyay,3 K Sahai,3 K Rade,1
A Sreenivas,1 N Kulshreshta4 1World Health Organization
Country Office, New Delhi, 2International Union Against
Delhi, 3State Health Services, Bihar, Patna, 4Central TB
Division, New Delhi, India. e-mail: alitangni@gmail.com
Setting: East Indian District of Patna with a population

of 5.6 million (urban 43% and rural 57%).
Objective: To estimate the duration between onset of
symptoms and diagnosis of smear positive TB patient in
the Microscopy Centres of Patna and assess whether age,
gender, literacy, knowledge on TB and the source of first
health care provider are associated with delay from onset
of symptoms to diagnosis of TB.
Method: A community based cross-sectional study was
conducted with structured questionnaire, which included
data from the laboratory register for smear positive
diagnosed.
Results *: During the study period (December 2012February 2013), 311 Smear-positive tuberculosis patients
were diagnosed, of whom 164 were enrolled, 140 (85%)
were successfully interviewed, 18 (11%) could not be
traced, 4 (2%) refused and 2 (1%) died. The median age
was 27 years (range: 14-70), 51% lived in rural and 49%
in urban areas. The reason for approaching government
health provider were Proximity (40/61, 66%) and for
private provider were Faith in the provider (27/38,
71%). The delay was longer if the patient, was a male (75
days vs. 30 days), unemployed (135 vs. 60) and contacted
providers other than the government (150 days vs. 60
days P , 0.001). Awareness about DOTS had shorter
delays and the odds were 0.9 (0.4-1.8) compared to
participants not aware. Participants having knowledge
about tuberculosis before being diagnosed had an odd of
1.8 (.9-3.6) to delay to those who had no knowledge of
tuberculosis. The median delay for patients with cough
74 (53%) was 60 days while it was 30 days for those with
fever 40 (29%).
Conclusions and key recommendations: After 14 years of
NTP in the district, there is median delay of 53 days in
diagnosing smear positive TB. Dissemination of information regarding the signs/symptoms of TB and the
availability of free services under NTP is to be reexamined considering the delay in patients with cough
more than that with fever. Though the patients are
reaching the diagnostic facility, ensuring laboratory
Short oral abstract sessions, Sunday, 6 December
examination is done on the same day will reduce delay

within the system. Also delay is less if the first care
provider is government though 56% first sought medical
services from non-government providers. The policy of
the state with regards to the rural health providers need
to be reassessed considering the proportion of patients
for whom they are the first health care provider.
SOA-605-06 Major delays in the diagnosis and

management of tuberculosis patients in Nepal
R Mahato,1 W Laohasiriwong,1 R Koju,2 K Biraj,2
K Vaeteewootacharn1 1Khon Kaen University, Khon Kaen,
Thailand; 2Dhulikhel Hospital, Kathmandu University
Hospital, Dhulikhel, Nepal. Fax: (977) 011 490 707. e-mail:
mahatoroshank@gmail.com
Background: Early diagnosis is determining factor for

spread of this lethal disease as delay in diagnosis and
treatment of tuberculosis geometrically increases spread
and infectivity of the disease and is associated with
higher risk of mortality. The present study aimed to
investigate the delays associated with diagnosis and
treatment among new pulmonary tuberculosis patient in
central development region of Nepal.
Design/Methods: An analytical cross-sectional study was
conducted by administration structured questionnaire
interview and reviewing the medical record of the new
sputum smear positive pulmonary tuberculosis cases
during January April 2015. Simple random sampling
was applied to select 5 districts form 19 districts of all 3
ecological regions of Nepal. Delay in diagnosis is
collective term for patient delay and health system delay.
The patient delay refers to the time interval between
onset of disease till patients approaches the formal health
facilities. Similarly, health system delay is the time for
complete diagnosis of the disease after patient approaches for formal health care provider. An acceptable patient
and health system delay was 30days and 7 days
respectively.
Results: A total of 374 new sputum smear positive
pulmonary tuberculosis cases were included in the study.
The mean age of participants was 39.67 years 618.56
years. The majority (62.57%) of the patients were male
and married (66.58%). A total of 33.16% were from the
rural part of Nepal and 30.75 % were illiterate. The
median patient delay, health system delay, and total delay
was 32days, 3 days and 39.5days respectively. An
unacceptably high patient and health system delays were
observed in 52.83% and 23.89% patients respectively.
Conclusion: Nepal has still a significant problem of TB
diagnosis and treatment. Most of the delay was caused by
patients because of the accessibility. Identifying major
delays and developing evidence-based approaches to
address those delays will help in advancing tuberculosis
prevention and management in low-income settings. This
will also be the key to success for reducing TB burden.
S561
SOA-606-06 The impact of the duration of

clofazimine administration with first-line drugs
in a mouse model of tuberculosis
chemotherapy
N Ammerman,1,2 R Swanson,1,3 B Ngcobo,1 C Moodley,1
A Dorasamy,1 S Modley,1 S Singh,4 J Grosset1,2
1
Durban, South Africa; 2Johns Hopkins University School of
Medicine, Baltimore, MD, USA; 3University of KwaZulu-Natal,
Durban, 4University of KwaZulu-Natal, Westville, South
Africa. e-mail: nicole.ammerman@gmail.com
Background: The addition of the anti-leprosy drug

clofazimine to the first-line drug regimen for tuberculosis
(TB) decreased the duration of treatment (from 6 months
to 3 months) necessary to achieve relapse-free cure in the
mouse model of TB chemotherapy. Because clofazimine
accumulates to extremely high levels in the tissues and
has a very long half-life (weeks to months), we
hypothesized that clofazimine may not need to be
administered with the first-line regimen for the entire
duration of TB treatment to exert its potent bactericidal
and treatment-shortening activity.
Design/Methods: To address our hypothesis, we used the
mouse model of TB chemotherapy to evaluate the
activity of first-line regimens in which clofazimine has
been incorporated in different administration schemes.
Mycobacterium tuberculosis-infected BALB/c mice were
treated for 4 months with one of the following regimens:
(i) no drug control; (ii) standard regimen control,
2RHZE/2RH (2 months of rifampicin [R], isoniazid
[H], pyrazinamide [Z] and ethambutol [E] followed by 2
months of R and H); (iii) 2RHZC/2RHC (previously
tested clofazimine [C]-containing first-line regimen);
2RHZC/2RH (clofazimine only administered during
the first 2 months); and (iv) 2RHZC/2C (only clofazimine administered after the first 2 months). Treatment
outcomes were bacterial counts in the lungs during
treatment and the proportion of culture-positive relapses
6 months after stopping treatment.
Results: After 3 months of treatment, mice receiving any
clofazimine-containing regimen had significantly lower
bacterial burdens in the lungs compared to mice receiving
the standard regimen; however, after 4 months of
treatment, only mice receiving the 2RHZC/2C remained
culture-positive. Six months after stopping treatment,
mice that received the clofazimine-containing regimens
had significantly less culture-positive relapse compared
to mice receiving the standard regimen: 2RHZE/2RH,
2RHZC/2RHC, 2RHZC/2RH, and 2RHZC/2C had
relapse-free cure rates of 53%, 87%, 93% and 80%,
respectively.
Conclusion: As part of a first-line regimen, clofazimine
might equally contribute to the overall antimicrobial
activity when included only in the intensive phase (first 2
months) of TB treatment. Further studies are needed to
optimize the dosing of clofazimine, especially considering the role of the sustained release of the drug from the
tissues after stopping treatment, to maximize its therapeutic potential for TB.
S562
SOA-607-06 Hepatotoxicity of a high-dose

rifampicin treatment regimen among HIVinfected tuberculosis patients in West Africa:
findings from the RAFA trial
C Merle,1 S Floyd,2 D Affolabi,3 B Bah,4 A Ndiaye,5
I Mbaye,5 M Sarr,5 O Bah-Sow4 1World Health
Organization, Geneva, Switzerland; 2 London School of
Hygiene & Tropical Medicine, London, UK; 3Centre National
Hospitalier de Pneumo-Phtisiologie, Cotonou, Benin; 4Ignace
Deen Hospital, Conakry, Guinea; 5Programme National de
Lutte Contre le Paludisme, Dakar, Senegal. e-mail:
merlec@who.int
Background: Around 30% of HIV-TB patients die by 12

months after starting TB treatment and the early
mortality appears to be TB-related, rather than HIVrelated. More aggressive TB treatment might improve
treatment outcomes and reduce mortality.
Design/Methods: The RAFA trial is an open label, 3
parallel arms, randomised controlled trial, and includes a
nested pharmacokinetic study. TB-HIV patients, who are
ARV-naive and with a CD4 cell-count .50 cells/mm3,
are recruited in Benin, Guinea and Senegal. The 3 trial
arms are 1) ARV initiation at week 2 combined with
standard TB treatment (Arm A); 2) ARV initiation at
week 8 combined with standard TB treatment (Arm B);
3) ARV initiation at week 8 with high-dose of rifampicin
(15mg/kg) during the intensive treatment phase and
standard dose during the continuation phase (Arm C).
The primary endpoint is 12-month mortality. Treatment
hepatoxicity is monitored at weeks 2, 4, 6, 8 and 24 of
treatment, with ALAT, ASAT, phospathase alcaline and
bilirubin measurements.
Results: 599 patients with pulmonary TB were included
in the analysis, with follow-up to at least the end of the
intensive phase. Baseline CD4 count was .200 cellcount/mm3 for 46%. Among patients with the high-dose
rifampicin regimen, 20% experienced liver injury based
on ALAT measurements (40/201); 1.0% (2/201) with
grade 3 or 4. Corresponding figures from Arms A and B
combined were 23% (92/398) and 0.3% (1/398) with
grade 3 or 4, with no statistical evidence of a difference
among the 3 treatment arms for the risk of a grade 3 or 4
measurement (P 0.26).
Conclusion: The risk of severe treatment hepatotoxicity
is low with a TB treatment regimen using a 15mg/kg dose
of rifampicin in TB-HIV patients.
SOA-608-06 Diagnosed but not treated: the

problem of untreated diagnosed TB in Uganda
P Kabagambe,1 J Muhangi,1 S Orach,1 J Emuto,1
C Edemaga1 1Uganda Episcopal Conference, Kampala,
Uganda. e-mail: pattkab@gmail.com
Background and challenges to implementation: Coughing patients who have been diagnosed with bacteriologically positive TB detected through Sputum Smear
microscopy or GeneXpert testing of their sputum
samples are particularly at a high risk of TB transmission
to other persons especially if they are not started on TB
treatment immediately. Untreated, diagnosed TB is also
hazardous to the patient, leading to poor outcomes like

complications of TB as well as high mortality. Without
an intervention to ensure that diagnosed bacteriologically positive TB patients are treated immediately, TB
control efforts will be futile. Possible causes of delayed or
no treatment of diagnosed TB are; drug shortages, delay
in receiving TB results hence them returning late when
the patient has left the facility, poor documentation, lack
of knowledge among laboratory staff and health workers
about the urgency of TB treatment, patient factors like
transport and long distance problems, poor facilitycommunity linkages and death before initiation of
treatment.
Intervention or response: We performed an audit of
bacteriologically positive TB patients identified by
Sputum-Smear Microscopy and GeneXpert testing in
the period from 1st March 2013 to 23rd April 2015 in 11
health facilities in Mpigi district in Central Uganda.
Laboratory registers were reviewed to list all B-positive
patients in the review period. Names were then cross
referenced from the Unit TB register (Treatment register)
for documented evidence of treatment. Documented
evidence of transfer out or referral was also sought.
Results and lessons learnt: A total of 481 patients were
diagnosed with bacteriologically positive TB (172 by
GeneXpert (35.8%) and 309 by SS (64.2%)) in the
review period. However, 185 (38.5%) patients TB
treatment was unaccounted for.
Conclusions and key recommendations: We propose to
conduct focus group discussions with health workers and
Village Health Teams in Mpigi to find out why the
treatment status of these patients is unaccounted for, on a
case by case basis. Follow-up in form of phone calls and
visits to the patients will then be instituted. We
recommend such audits periodically, not only for TB
but also for other infectious diseases like HIV so as to
strengthen timely linkage to care and treatment. We
recommend a treatment-linkage focal person or team to
ensure that all positive results from the laboratory are
returned to requesting clinicians and followed by timely
initiation of treatment
SOA-609-06 Using electronic data management

system leads to improved TB patient
monitoring in Kampala, Uganda
A Etwom,1 D Lukoye,1 D Kimuli,1 M Kenneth,1
M Ruhweza,1 P G Suarez2 1Management Sciences for
Health, Kampala, Uganda; 2Management Sciences for
Health, Medford, MA, USA. e-mail: etwoma@yahoo.com
Background and challenges to implementation: Kampala

capital city (KCCA) has an annual TB notification of
about 9,000 cases. However, the 2013 treatment
completion was about 54.3% and over 20% getting lost
to follow up mainly because there was no proper way of
identifying treatment interrupters and defaulters. Paper
based registers could not support easy retrieval of records
during drug refills and tracking of missed appointments.
Also at the same time, KCCA had low (20%) timely
submission rate of reports to Ministry of Health (MOH)
and affecting timely decision making. We explored

benefits of using a computerized data management
system for following up patients in Kampala.
Intervention or response: A USAID and PEPFAR funded
TRACK TB Project in collaboration with KCCA,
developed a database based on TB registers. The project
trained 14 health workers on use of the system.
Computers (PC) and Internet were given to routinely
document patient data that is, physical and phone
addresses, diagnosis, refill schedules and HIV services
for patients that were started on TB treatment between
April 2013 and March 2014 from 60 health facilities.
The information was then routed to the central server PC
managed by data officers. Data officers routinely
generated lists of patients who either missed appointments or interrupted treatment and linked them to
Village Health Teams (VHT) who in turn traced these
patients from the communities, counseled and linked
them back to treatment.
Results and lessons learnt: A total of 9,447 were patient
records managed till end of treatment period. As a result
of using computerized system to timely inform health
workers about patients who are interrupting treatment
and using CLFs, quarterly treatment completion improved from 54.3% to 85% over a period of 1 year. The
number of treatment interrupters traced and linked back
to treatment increased from about 150 to 1594 and
hence lost to follow up reduced from 2160 (24%) to 566
(6%). In addition KCCA achieved 100% timeliness and
accuracy in reports submitted to MOH because the
computerized system generated reports automatically.
Also time spent by 12 health workers in compiling
quarterly reports was cut from about 24 hours to 2
hours. The data also facilitated numerous decisions.
Conclusions and key recommendations: Results show
that use of a computerized system to manage TB patients
will improves treatment completion and facilitates quick
decision- making.
SOA-610-06 How do patients access the private

sector in Chennai, India? An evaluation of
delays in tuberculosis diagnosis
L Bronner,1,2 R Ananthakrishnan,3 A Swaminathan,3
N Krishnan,3 M Pai,4 D Dowdy1,2 1Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD, 2Johns
Hopkins School of Medicine, Baltimore, MD, USA; 3REACH,
Chennai, Tamil Nadu, India; 4McGill University, Montreal, QC,
Canada. e-mail: lizabronner@gmail.com
Background: Diagnosis and treatment of tuberculosis

(TB) in India are characterized by heavy private-sector
involvement. Prolonged delays from symptom onset to
treatment are important contributors to TB transmission,
but these delays remain poorly characterized among
patients who seek care in the Indian private sector. We
assessed delays in TB diagnosis and treatment initiation
among patients in the private sector. Design Crosssectional survey of 289 consecutive patients who were
eventually diagnosed with TB in the private sector and
referred to a public-private mix program for TB
treatment in Chennai city from January 2014 to February
S563
2015. We conducted interviews using standardized

questionnaires to collect data on demographics, TB
disease, healthcare providers consulted, and intervals
from symptom onset to treatment initiation. We defined
formal providers as allopathic and non-allopathic
practitioners with formal training in the public or private
sector. Informal private providers included traditional
healers, unlicensed practitioners, and pharmacists. Total
delay was analyzed using multivariable generalized linear
and Poisson regression models.
Results: Among 212 patients diagnosed with pulmonary
TB in this public-private mix program, 90% first
contacted a formal private provider, and 78% were
diagnosed by the first or second provider seen. Patients
who first consulted the formal private sector experienced
longer mean patient delay, but health system delay was
longer among patients who first consulted the public and
informal sectors (Figure). For patients first seeking care
in the informal private sector, mean total delay was
longer versus the formal private sector (mean difference:
28 days, 95%CI: 12.5-44.3, P , 0.001) or public sector
(mean difference: 27 days, 95%CI: 0.6-53.3, P 0.05).
After adjusting for potential confounders, consulting an
informal provider as the first point of care, compared to a
provider with formal training, resulted in a significant
increase in total delay (absolute increase: 23 days,
95%CI: 6.2-39.5), and risk of having prolonged delay
.90 days (aRR 2.4, 95%CI: 1.3-4.4).
Conclusion: Even among patients seeking care in the
private sector in Chennai, delays in TB diagnosis are
substantial, often two months or more. Seeking care in
the informal private sector is associated with increased
diagnostic delay of about one additional month. Novel
TB diagnostic strategies are required across all health
sectors to reduce the time patients spend seeking
diagnosis after onset of infectiousness.
S564
02. How to reach the END of TB

SOA-611-06 Feasibility of a zero medical
burden for pulmonary tuberculosis outpatients through Chinas new rural cooperative
medical system
Z Xu,1 L Bai1 1TB Control Institution of Hunan Province,
Changsha, China. e-mail: evan.xu21@gmail.com
Background: Although China has free TB diagnosis and

treatment policy, most patients have to pay additional
diagnosis and treatment expenses (such as liver and
kidney function test, liver protect medicine), which may
lead to treatment interruption due to patients low
affordability. Therefore, Hunan province launched a
so-called case-based payment pilot project in fourteen
counties successively to cover all patient expenses
through Chinas new rural cooperative medical system.
Under this pilot project, patients would not pay any
medical fee in whole treatment course. This research is to
explore the feasibility of case-based payment (zero
medical expense) for pulmonary tuberculosis (PTB)
outpatients through new rural cooperative medical
system (NRCMS).
Methods: Cross section and qualitative study were
applied. All 131 counties of Hunan province were
investigated to identify the current situation of medical
security in tuberculosis patients by questionnaire and
telephone interview. Thirty-six staff of nine pilot counties
which already implemented case-based payment were
interviewed on their opinions of case-based payment,
interviewees were from health bureau, NRCMS sectors
and centers for disease control and prevention (CDC) of
each county.
Results: Outpatient PTB treatment for rural patients was
covered by NRCMS in 118 of all 131 counties.
Reimbursement models consist of proportional reimbursement, fixed amount reimbursement and case-based
payment. However, the proportion of proportional
reimbursement has a range of 30%~80%, the payment
of amount reimbursement varies from 200 to 1500 RMB
Yuan per case. Patients covered by proportional reimbursement and fixed amount reimbursement need to pay
medical expense of 588yuan/case and 463yuan/case
respectively, comparing zero expenditure for patients of
case-based payment. NRCMS need input about 17.8
million Yuan more per year to reimburse all rural PTB
outpatients medical expenses, which only accounted for
0.12% of total NRCMS fund. In those counties which
already implemented case-based payment policy, the
number of patients visited CDCs has increased dramatically, follow-up and treatment compliance improved
obviously as well.
Conclusion: Zero outpatient medical burden for rural
tuberculosis patients is feasible and realizable through
case-based payment under Chinas NRCMS.
SOA-612-06 Innovative social protection

mechanism for alleviating catastrophic
expenses for multidrug-resistant tuberculosis
patients
D Kundu,1 V Katre,2 M Deshpande,3 A Sreenivas,1
P Parmar1 1World Health Organization Country Office for
India, Delhi, 2RSBY and MSBY State Nodal Agency,
Directorate of Health Services, Raipur, 3Directorate of Health
Services, Raipur, India. e-mail: debashishkundu@yahoo.com
Background and challenges to implementation: MultiDrug Resistant Tuberculosis (MDR-TB) patients incur
huge expenditure for its diagnosis and treatment, which
can be reduced through a well designed and implemented
social health insurance mechanism. Post-2015 WHO
Global TB Strategy Framework, one of the key targets set
for 2035 is no affected families face catastrophic costs
due to tuberculosis.
Intervention or response: The state of Chhattisgarh in
India successfully established partnership between Revised National TB Control Programme (RNTCP) and
Health Insurance Programme, a Universal Health Insurance Scheme (UHIS) for all, by establishing RSBY and
MSBY MDR-TB packages [Table] to absorb catastrophic expenses incurred by MDR- TB patients from
diagnosis to treatment completion in public and private
sector. Data on uptake of insurance claims through
innovative MDR-TB packages from December 2013 to
April 2014 was collected to discuss initial experience of
linking health insurance programme tailor made to cover
catastrophic health expenditure for MDR-TB patients.
Results and lessons learnt: 207 insurance claims of
RSBY and MSBY MDR-TB Packages were processed,
of which 20 claims were from the private and 187 from
the public health establishments, empanelled under the
health insurance programme, free of cost. Catastrophic
expenditure, ~20000 USD, was saved through RSBY and
MSBY insurance mechanism. For the first time in India,
national health insurance was successfully linked with
the national TB control programme through creation of
special packages for the MDR-TB patients. Early finding
of implementation of RSBY and MSBY MDR-TB
Packages shows that this innovative mechanism is
working.
Conclusions and key recommendations: Innovative
RSBY and MSBY MDR-TB health insurance package
is a step towards reducing catastrophic expenses.
Considering that both drug resistant and drug sensitive
TB treatment is ambulatory and takes longer time to
complete, out-patient care needs to be included into the
mainstream health insurance. In order to address the
social protection component of post 2015 End TB
Strategy, mechanisms emphasizing collaboration with
existing social health insurance schemes needs attention
in the national policy framework for TB control.
Table Rashtriya Swasthya Bima Yojna (RSBY), Mukhyamantri

Swasthya Bima Yojna (MSBY) and Innovative RSBY & MSBY
MDR-TB Package Details
RSBY and MSBY
What is the scheme?
Coverage
Amount
Financing
RSBY is a cashless, paperless Total

Beneficiaries need to
and portable social health
insurance
pay only 0.42 USD
insurance scheme through
coverage
as registration fee
smart cards. RSBY and
amount is
while Central and
MSBY, The Chief Ministers upto 30000 State Government
Health Insurance Scheme in INR (500
pays the premium
Chhattisgarh, also managed USD*) per to the insurer
by RSBY, provides health
annum
selected by the
insurance coverage and
17 USD
State Government
protection to people below
(~2 USD
on the basis of a
and above the poverty line
Per Visit)
competitive
(APL and BPL) to fund
bidding. Every
medical treatment for
beneficiary family
packages covered in Round
is issued a
IV of RSBY and MSBY
biometric enabled
disease list in Chhattisgarh
smart card
1. On admission basis
containing their
2. For five enrolled
fingerprints and
members in a family;
photographs.
3. Covers transportation
expenses in a year
across all RSBY
empanelled network
hospitals (both private
and public) in the state.
1
USD* 60 INR
Innovative RSBY and MSBY MDR-TB Package Details in Scheme

Year 4
Medical/
Surgical
Medical
Conditions
Medical
Conditions
MDR
TB Package
Name
Package Details
No. of
Times/ Days
claims
Package
can be
Cost
processed
PreChest X-Ray,
4000
One Time
Treatrelevant
INR
ment
haematological
(67 USD)
Evaluand biochemical
ations
tests: Complete
Blood Count
(CBC), Liver
Function Test
(LFT), Thyroid
Function Tests
(TFT), Blood
Urea Nitrogen
(BUN),
Creatinine,
Urine (Routine
& Microscopic),
Urinary
Pregnancy Tests
(UPT)
Follow- Chest X-Ray,
3300 INR Maximum
up
relevant
(55 USD) Five
Evaluhaematological
Times
ations
and biochemical
for Creattests: Complete
inine and
Blood Count
rest tests
(CBC), Liver
for
Function Test
maximum
(LFT), Blood
two times
Urea Nitrogen
(BUN),
Creatinine,
Urine (Routine
& Microscopic)
Medical
Conditions
S565
Hospital Bed Charges,

5600 INR Maximum 7
Stay
Doctors
(93 USD days stay
Consultation
@13
on proFees and any
USD/
rota basis
other additional Day)
/ ancillary drugs
SOA-613-06 Identification of solutions and

their levels of acceptance by private health care
providers for adoption of key tuberculosis care
standards in India
O George,1 R Ganesan,1 V Sharma,1 R Swamickan,2
M Benezet3 1Abt Associates Inc., New Delhi, 2United States
Agency for International Development, New Delhi, India;
3
Abt Associates Inc., Bethesda, MD, USA. Fax: (91) 112 653
5172. e-mail: oommen.geo@gmail.com
Background: Combating tuberculosis (TB) in India and

mitigating the growing threat of drug-resistant TB
requires adoption of TB care standards by private health
care providers (pHCP). Experience suggests that few
pHCP utilize sputum tests for diagnosis of pulmonary
TB, notify patients or accept intermittent treatment
regimens. Treatment monitoring is largely dependent on
patient revisits. Engagement of pHCP has hitherto been
on the terms of Indias National TB Control Program.
Solutions influencing pHCP adoption of TB care
standards need to be identified and tested.
Design/Methods: A 2-stage study was conducted in
2013-14 with qualified pHCP in Karnataka, India. The
objective of stage 1 was to identify possible solutions for
pHCP adoption of TB care standards through iterative
exploratory interviews using semi-structured questionnaires with 56 doctors. In stage 2, structured interviews
with 170 doctors explored potential pHCP behavior
change if packages of these solutions were applied.
Results: In stage 1, pHCP identified 8 solutions: (1)

continuing medical education, (2) laboratory access
through sputum transportation, (3) assured patient
retention or feedback on those referred, (4) support for
additional documentation and patient monitoring, (5)
flexibility to adopt daily drug regimens, (6) non-financial
rewards, (7) non-disclosure of personal identifiers (PII) of
S566
patients, and (8) financial incentives. Stage 2 showed that

solutions (1), (2), (4) and (5) had strong and wide appeal
while solutions (3), (6), (7) and (8) appealed to niche
groups. Half the pHCP interviewed reported no concerns
with notification, but expressed the lack of means for
this. Financial incentives, however, evoked strong negative reactions from some providers. Provider segments
and potential value perceived by pHCP to interventions
suggest that inputs that create and support a provider
network, especially those impacting patient satisfaction
of provider services, are valued by a large group of pHCP.
A base package offering solutions (1) to (6) had strong
effects on utilization of sputum tests (from 27% to 65%),
and notification (from 46% to 72%). Incremental effect
of adding solutions (7) or (8) or both to the base package
were similar, impacting different sub-groups of pHCP.
Conclusion: The study recommends action to support
pHCP adoption of key standards for TB care. The base
solution package has the potential to substantially
improve provider compliance to sputum tests and
notification.
company representatives and the profit margins associated with drugs. All respondents said that, no patients
buy the entire course of drugs at a time TB patients are
mostly daily wage workers. They get money in the
evening and buy medicines. Doctors though did not
explicitly ask patients to go to a particular RPP, but
ensured that they go to the pharmacy attached to their
clinic, by writing medicines that were available only
there. Pharmacists reported doctors receiving commission upto 40% from pharma companies.
Recommendations: Potential of RPPs has remained
untapped in NTP. A policy to systematically involve
PRPs should be considered. The PRPs when trained, can
create community awareness about TB, help in early case
detection/referral and act as DOTS provider and counsel
patients. Enforcing regulatory monitoring of sale TB
drugs under schedule H1, to check the indiscriminate use
is urgent to prevent multi drug resistant TB.
SOA-614-06 Retail private pharmacists

knowledge and perceptions about
collaborating with the National Tuberculosis
Programme: a qualitative study, South India
S Vijayan,1 P Mhatre,2 R Gandhi,1 R Chopra,1 J Thakker,1

P Kandasamay,1 V Jondhale,1 Mohu Datta1 1PATH,
Mumbai, 2Mumbai Mission for TB Control, Mumbai, India.
Fax: (91) 114 605 4430. e-mail: shibuvij@gmail.com
Y Vijayashree,1,2 H Bindu,2 N Rao,2 N Devadasan2

Institute of Tropical Medicine, Antwerp, Belgium; 2Institute
of Public Health, Bangalore, India. e-mail:
vijayashree@iphindia.org
Background: There is documented evidence of vouchers

for offsetting patient costs in Maternal and Child health
(MCH) services thus reducing financial constraints
amongst the lower socio economic segment. Taking cues
from this Private Provider Interface Agency (PPIA) in
Mumbai designed and deployed a voucher system for TB
suspects and patients who approach private health care
providers.
Intervention: The PPIA voucher scheme uses 3 differently
colored voucher booklets(20 vouchers) for each of its 3
major services :(i) Free Chest X-ray - availed at
networked quality controlled facilities, (ii) Highly
subsidized CBNAAT (XpertMTB/RIF) amongst the
networked laboratories and (iii) free treatment for the
first line drugs availed at the networked chemists. Each
voucher has 3 copies (with unique voucher number), for
patient, physician and NGO field officer. Two subcontracted NGOs are responsible for voucher stock management and placing as well as replenishment of
vouchers at each of the networked facilities. The NGO
coordinator forwards the collated details of reimbursable
amounts from the service providers with required
documents to PPIA - Grant management team. The
reimbursement is made against the invoice generated on
monthly basis from the NGOs. The reimbursement of
every credit note is completed within 7 days of the
subsequent month.
Results: In 7 months, 1300 private medical providers,
~100 health facilities, 5 laboratories, 80 X-ray laboratories and ~100 chemists were empanelled with NGO
field representatives placing ~25 000 X-ray vouchers at
the networked health facilities in Mumbai. ~2000 X-ray
vouchers have been issued by Less Than Fully Qualified
Background: In India, RPPs are often first and repeated

point of contact for patients. NTP is involving RPPs
through Indian Pharmaceutical Association by training
them to identify and refer chest symptomatics to NTP for
TB diagnosis. We conducted this study to assess RPPs (i)
knowledge and referral practices (ii) stocking and
dispensing of TB drugs (iii) kickbacks to providers.
Methods: Semi-structured interviews were conducted
with 40 RPPs in Bangalore (urban 19) and Tumkur
district (rural 21) during 2013 from Karnataka, India.
RPPs were randomly selected from the register maintained with district drug controller.
Results: None of the respondents had received any TBrelated training, but majority of them knew about mode
of spread of TB. RPPs described the typical profile of TB
patients as old and financially weak. Mix of providers,
such as allopathic, traditional medicine and quacks
referred TB patients to RPPs. Self-referrals were common
among the economically poorer sections as they did not
see value in paying an additional cost to the doctor as
consultation fee and purchased medicines based on the
RPPs recommendation. One respondents believed that
pharmacists have experience and that they know what is
best for patients. They reported seeing 10-15 chest
symptomatic per day and dispensing cough syrups, and
antibiotics to such patients. RPPs in rural tend to refer
cases to NTP as compared to urban RPPs. They tend to
stock TB drugs and their choice was mainly determined
by the doctors prescription patterns, opinions of drug
SOA-615-06 Use of voucher schemes:

promising higher quality service delivery for
early diagnosis and treatment of TB in Mumbai
(LTFQ ) (Utilization 70%) providers for the suspect

patients and~1400 X-ray vouchers have been issued
from networked health facilities (Utilization 80%)
implying a cost of about INR 6 lacs (9500 USD) A total
of 1750 gene expert testing vouchers (Utilization 90%)
incurring a cost of ~INR 22.5 lacs (36 000 USD) and
about 9000 treatment vouchers (for first line drugs)
incurring ~INR 34.5 lacs (55 000 USD) were issued
Conclusions: The voucher scheme offsets patient costs
leading to higher utilization of services and provides
private providers with free diagnostic and treatment
packages to be directly offered to the patients. An Evoucher mechanism is foreseen for scaling up voucher
scheme operations to widen the reach of the program.
The future of E vouchers is promising for higher cost
effectiveness of the program.
SOA-616-06 Disparities in treatment delay

among patients with pulmonary TB in South
West Nigeria
M Alagi,1 O Popoola,2 O Uchendu,2 E Owoaje2 1National
Agency for Control of AIDS, Abuja, 2University and University
College Hospital, Ibadan, Nigeria. e-mail:
drpopee@gmail.com
Background: Nigeria has the highest burden of tuberculosis PTB in Africa. Delays in initiation of DOTS therapy
is known to increase the probability of PTB transmission
from infected patients to uninfected contacts. This study
therefore aimed to estimate the time lapse in initiation of
treatment for PTB and explore factors associated with
delay including disparities in delay between socioeconomic groups.
Design/Methods: An analytical cross sectional study
patients newly initiated on DOTS treatment across was
conducted. A multistage sampling technique was used to
sample from rural and urban DOTS centres. The WHO
EMRO tuberculosis delay survey questionnaire was
adopted. Treatment delay was defined as the period
between onset of PTB symptoms and initiation of antiKochs treatment. The Kaplan-Meier survival analysis
technique was used as well as multivariate logistic
regression. Level of statistical significance was set at a
0.05.
Results: A total of 403 PTB patients newly initiated on
DOTS therapy were studied. 204 respondents were
sampled from urban 4 urban DOTS centres while, 199
respondents were sampled from 8 rural DOTS centres.
Persistent cough 59.5% and fever 12.7% were the most
frequently reported symptoms which prompted care
seeking. The median treatment delay was significantly
longer among urban PTB patients 49 (range 2 - 363) days
compared to rural 40 (range 6 - 344) days. Factors
associated with treatment delay in excess of 4 weeks
included attempts at concealing ill health OR 3.0 95%CI
(1.5 5.2); out of pocket expenditure OR 2.5 95%CI (1.1
- 5.5); utilizing more than 2 health care providers OR 2.2
95%CI (1.2 - 4.0; technical unskilled workers (artisans,
drivers & traders) OR 2.3 95%CI (1.2 4.5 and males
OR 1.8 P 0.047 95%CI (1.0 3.3).
S567
Conclusion: The study showed unacceptably long delays

in initiating PTB in south west Nigeria. Efforts to address
disparities in treatment delay should identify that males,
unskilled manual workers, individuals without health
insurance and with poor health access as being more
vulnerable to treatment delay.
SOA-617-06 EXPAND-TB: contribution of an

innovative project in increasing global MDR-TB
detection
D Orozco,1 K Kao,1 J Lemaire,1 C N Paramasivan,1
R Mazitov,1 T Verges,2 Y Mundade,3 F Mirzayev4 1FIND,
Geneva, 2STOP TB PARTNERSHIP, Geneva, 3UNITAID, Geneva,
4
World Health Organization, Geneva, Switzerland. e-mail:
kekeletso.kao@finddx.org

EXPAND-TB project is a 7 year (2009-2015) initiative
to help countries turn the tide of the TB epidemic by
introducing new and faster TB diagnostics in 27
countries, 15 of which are high-burden TB or high
MDR-TB burden countries. The EXPAND-TB project
aims to address the obstacles to correct diagnosis and
treatment of MDR-TB in low-resource settings by
improving laboratory capacity to diagnose drug-resistant
and HIV-associated TB using WHO endorsed diagnostic
tools. The project is focused on establishing or upgrading
103 central and regional reference TB laboratories and
47 Xpert MTB/RIF testing sites.
Intervention or response: 101 central and regional TB
reference laboratories and 47 Xpert MTB/RIF testing
sites have been established; commodities for MGIT
liquid culture and drug susceptibility testing, line probe
assay, and rapid speciation are regularly delivered to the
countries to meet their testing requirements; comprehensive training packages and SOPs have been developed
and training was delivered building the capacity laboratory staff, clinicians and policy makers. Analysis of the
EXPAND-TB projects contribution to global MDR-TB
notifications was made using data from the latest WHO
TB report (2013). The MDR-TB estimates from the 27
EXPAND-TB supported countries were pooled and
compared to global estimates.
Results and lessons learnt: The project was implemented
in 3 phases: phase 1-laboratory preparedness; phase 2technology transfer and validation; phase 3- routine
testing and monitoring. As the project progressed
through the implementation phases its contribution to
the global MDR-TB notifications steadily increased. The
increase was noticeable in 2012 and 2013 as the majority
of countries entered phase 3 (Figure 1). In 2013, from the
299 800 MDR-TB cases estimated globally, only 136 412
(45%) were notified and 63 761 (47%) were notified
within the 27 project countries. Amongst those notified
within the 27 countries 35 736 (56%) cases were
detected directly through EXPAND-TB support and the
remainder by other means or in other labs not supported
by the project. By the 31st December 2014 a total of 106
983 MDR-TB cases had been diagnosed with EXPANDTB support since project initiation in 2009.
S568
Conclusions and key recommendations: The EXPANDTB project has successfully contributed towards the
global increase in MDR-TB notifications
SOA-618-06 Customizable economic analysis

tools to support evidence-based tuberculosis
control decision-making
E Carlson,1 T Miller,1 B Naik,2 A Vemulapalli1 1University
of North Texas Health Science Center, Fort Worth, TX, USA;
2
Revised National Tuberculosis Control Programme,
Bangalore, India. Fax: (1) 817 735 0446. e-mail:
erin.carlson@unthsc.edu
Background: Tuberculosis (TB) control requires substantial public resources. Effective, sustainable public strategies consider diagnostic and treatment options, as well
as associated efficiencies and costs. Economic analysis
can bolster effective advocacy for TB programs, identify
efficiencies and optimize program efforts. However, such
work may be beyond technical capacity of local health
authorities. An economic evaluation tool customizable to
local, regional, or national circumstances may be a
valuable means to provide economic information to
facilitate evidence-based decision making by health
authorities.
Intervention: We created customizable economic analysis
tools to evaluate costs and efficiencies of regional TB
programs in the United States and India. These tools
combine standard techniques for health care economic
analyses with local cost, outcome, and prevalence data in
a Microsoft Excel platform. The tools allow health
authorities to model potential health and economic
effects of proposed programmatic changes and consider
how their program may be affected by changes in
diagnostics, medications, personnel, food packages, local
epidemiology, etc. Outcome measures are customizable
and may include case rates, mortality, and money costs or
savings in local currency or over time. Two such
modeling tools were developed and customized to
support TB control activities by the Texas Department
of State Health Services (TDSHS) in the United States
and the Government of Indias Revised National
Tuberculosis Control Programme (RNTCP).
Results: The model estimates net monetary costs
associated with incidence and other inputs, the number
of TB and MDR cases to result from secondary
transmission, and the number of deaths caused by index
cases during a given time. This baseline data is compared
to hypothetical conditions for a newly proposed or
modified DOTS or other TB program, assuming conservative observed, published, and estimated cost and
outcome differences. Under these assumptions, the
model provides information to determine how many
dollars, lives, and new cases may potentially be saved
under the proposed program compared to the existing
program.
Conclusions: A customizable economic analysis tool may
support TB program policy through rapid comparisons
of costs and outcomes for proposed program changes.
SOA-619-06 Modelling the cost-effectiveness

for the health system of XpertW MTB/RIF scaleup by district profile in Tanzania
I Langley,1 B Doulla2 1Liverpool School of Tropical Medicine,
Liverpool, UK; 2National TB and Leprosy Programme, Dar Es
Salaam, Tanzania. e-mail: ivor.langley@lstmed.ac.uk
Background: The computer modelling of patient pathways (known as Virtual Implementation) has been used
to assist national policy makers to assess options for the
scale-up of new diagnostics (including Xpert MTB/RIF)
for pulmonary tuberculosis in Tanzania. This presentation will demonstrate how a new version of the model
designed for local staff from the National Tuberculosis
Programme to use themselves has been employed to
model diagnostic algorithms at district level.
Design/Methods: A model using the Virtual Implementation approach has been developed which is easy to use
and contains an Excel user interface for data input and
reviewing outputs in a TB diagnostics dashboard.
Individuals in Tanzania have been trained and software
installed to enable staff to model for themselves the
impact of alternative diagnostic algorithms. The model
has been used to assess eight different diagnostic
algorithms including those using Xpert MTB/RIF in 10
different diagnostic districts and 5 diagnostic district
types. Incremental cost effectiveness ratios have been
calculated.
Results: Implementation of Xpert MTB/RIF in all
diagnostic districts modelled would be cost effective
and have the greatest benefit, measured in DALYs
averted and cures, but with highest additional health
system costs. Depending on the objective of implementation the priority sequence for implementation of new
diagnostic algorithms across the country varies. The
projections show that same day LED is particularly
effective when funds are tight and reduction in DALYs
and increasing cures is top priority. If more funds are
available full roll-out of Xpert is the preferred strategy
starting with the largest centres and those with high HIV.
If MDR-TB detection is the top priority then LED same
day will not assist greatly and early roll-out of Xpert is
preferred. Similarly if time to start treatment is top
priority then reducing high levels of empirically diagnosis
becomes the priority so implementing Xpert in districts
with a high proportion of smear negative TB should be
considered first.
Conclusion: The analysis suggests that an effective policy

for pulmonary TB diagnostics in Tanzania would be full
roll-out of Xpert MTB/RIF in larger districts with high

HIV first and same day LED microscopy elsewhere. The
Virtual Implementation approach is a useful tool that
should be retained in Tanzania to model new implementations of Xpert at district level and other new tools as
they become available.
S569
than vaccinating adolescents. The most effective vaccine

explored may reduce TB incidence and mortality rates by
almost a third by 2050, averting up to 3.7million cases
over this time horizon.
SOA-620-06 Future trends in TB epidemiology

in China and the potential impact of novel agetargeted TB vaccines: a modelling study
R Harris,1 T Sumner,1 G Knight,2 R White1 1London School
of Hygiene & Tropical Medicine, London, 2Imperial College
London, London, UK. e-mail: rebeccaclaireharris@gmail.com
Background: Ageing is an important risk factor for TB in

China, with prevalence rates in the elderly 3-times higher
than all-age estimates. The impact of population ageing
on Mycobacterium tuberculosis transmission and the
future burden of tuberculosis (TB) disease in China is
poorly understood. Furthermore, with a pipeline of new
vaccines, understanding the currently unknown benefits
of targeting older populations is important for planning
development and implementation. These new vaccines
are likely to be required for China to reach WHO 2050
elimination targets.
Design/Methods: An M. tuberculosis transmission model
incorporating Chinese social mixing patterns was calibrated to age-stratified estimates of population size and
TB prevalence, notification and mortality rates. The
model assessed the elderly (65years) contribution to M.
tuberculosis transmission and new cases, and the impact
of vaccinating older adults (55-64years) or adolescents
(15-24years). Novel vaccines with pre-infection, postinfection and multi-stage mechanisms of action were
explored (efficacious in individuals uninfected, latently
infected, and regardless of infection status, respectively),
at a range of efficacies (40%-80%) and coverages (30%,
70%).
Results: Preliminary results suggest the elderly contribution to M. tuberculosis transmission will increase from
18% to 53%, and to new TB cases from 13% to 71%,
between 1990-2050. Compared to baseline, vaccination
of older adults or adolescents reduced 2050 TB incidence
rates by 2%-31% or ,1%-7%, respectively, and
mortality by 2%-28% or ,1%-5%, respectively; thus
averting 199 000-3 676 000 or 10 000-963 000 cases,
and 7900-152 700 or 300-26 700 deaths 2025-50. Postinfection vaccines provided substantially greater impact
than pre-infection vaccines when delivered to older
adults, whereas if targeted to adolescents pre-infection
vaccines achieved greatest impact. The minimum achievable incidence by 2050 was 22/100 000pop/yr, with 70%
coverage of older adults with an 80% efficacious multistage vaccine.
Conclusion: Our preliminary results suggest that by 2050
the elderly will be the main contributor to TB burden in
China. Strengthening of TB control efforts for the elderly
is likely required to achieve WHO 2050 TB elimination
goals. Novel TB vaccines targeted to older adults
provided consistently higher population level impact
SOA-621-06 Comparing the impact of the

Global Fund on health systems in Papua New
Guinea and Nepal
I Harper,1 A Street,1 K Dahal2 1University of Edinburgh,
Edinburgh, UK; 2HERD, Kathmandu, Nepal. e-mail:
ian.harper@ed.ac.uk
Background: Both Nepal and Papua New Guinea with

high burdens of tuberculosis - are dependent on GFATM
grants for the financing of their TB programmes. Each
country also has amongst the remotest, poorest, hardestto-reach populations in the world, coupled with weak
primary health care infrastructure.
Design/Methods: The research questions focused on the
impact of GFATM funding structures and evaluation
metrics on relationships between health institutions, the
distribution of financial and infrastructural resources,
and programmatic performance. Independent ethnographic research in both countries was conducted,
supplemented with in depth open-ended interviews with
government officials and health workers, NGOs project
staff, and other stakeholders in GFATM funded projects.
Inductive analysis compared data from both countries, to
draw up generalizable themes and cross cutting issues.
Results: Both countries have significant resource and
programmatic issues facing their TB control linked to
geography, weak communication and transportation
infrastructure, and a growing number of actors involved
in health governance and delivery. The current drive of
developing M & E metrics in line with GFATM policy
and practice, while generating important data for
monitoring performance, has important unintended
consequences. A tendency to focus on achieving targets
and an increase in bureaucratic procedure has channeled
institutional resources into generating these metrics,
informants argue, at the expense of broader health
service delivery issues. Marginal populations in the hard
to reach areas are the most likely to suffer from these
unintended consequences.
Conclusion: The current recording and reporting metrics
developed for accountability purposes for GFATM
funded programmes have unintended effects. While the
S570
development of M & E is a crucial component of health

system development, more attention should be focused
on the impact these metrics themselves have on the
systems. In particular, we show they can reinforce
inequalities between infrastructure-dense and infrastructure-sparse areas of the health system. The impact of M
& E on access to services for the most marginal and the
hardest to reach populations should be a routine part of
health system evaluation.
03. HIV and TB: snapshots

SOA-622-06 Blended learning as capacity
building option for TB-HIV services: results of a
comparative study in Ethiopia
G Tsegaye,1 T Anteneh,2 K Merga,1 M Abebe,1
M Behute,1 M Aseresa Melese,2 D J Dare,2 P G Suarez3
1
All Africa Leprosy, Tuberculosis, Rehabilitation and Research
Training Centre, Addis Ababa, 2Management Sciences for
Health, HEAL TB Project, Addis Ababa, Ethiopia;
3
Management Sciences for Health, Center for Health
Services, Arlington, VA, USA. Fax: (251) 11 372 4473.
e-mail: degujerene@gmail.com
Background: Shortage of trained health workers is a

critical barrier to strengthening tuberculosis prevention
and control programs. Health workers frequently stay
away from routine duty and family for a number of days
for off-site workshop trainings. Our objective was to
assess if blended learning (BL) approaches which
combine distance-based learning with off-site trainings
can be implemented without compromising the quality
of services.
Methods: We carried out a comparative study using the
Kirkpatrick training evaluation model. The BL group
received self-reading materials for 3 weeks followed by a
3-day face-to-face off-site training while the traditional
group (TD) received a 6-day off-site workshop-style
trainings. Training materials were tailored to the specific
modality of training but nearly the same content
developed and endorsed by the Ministry of Health were
used. Health workers with no recent training in the
updated national TB-HIV guidelines from two selected
zones participated in the study. Health workers with no
recent training in the updated national TB-HIV guidelines from two selected zones participated in the study.
The evaluation criteria included participant satisfaction,
learning, behavioral changes, impact according to the 4level Kirkpatrick training evaluation model. We compared the level of satisfaction using a 32-item questionnaire between the two groups, and computed pre-test
and post-test results between and within the two groups
using one way ANOVA test. We also estimated the unit
direct cost per trainee for each training modality. This
paper presents preliminary results for the first two levels,
not inclusive of the behavior change and outcome.
Results: At the time of this analysis, complete data was
available for 209 trainees (63 from the TD group and
146 from the BL group). The two groups did not differ in
terms of education and work experience. The two groups

had similar level of satisfaction [composite score; 57 in
TD vs.55 in BL]. The mean pre-test result was higher for
the TD group (56.2 vs. 51.2) while their post test result
was similar (66.7 vs. 63.5; F1.73; P . 0.1). Also, the BL
group had higher mean (SD) improvement in post-test
performance (15.6 vs. 10.1; F7.52; P , 0.01). The unit
cost per trainee was lower in the BL group [116.8 USD
versus 235 dollars].
Conclusion: Both groups had similar level of satisfaction
with the training approach and improvement in knowledge level was better in the BL group. The cost and time
taken to deliver the BL approach was lower by about a
half. This approach could be a more efficient way of
delivering in-service trainings but further analysis is
needed to compare behavioral and service delivery
outcomes between the two groups.
SOA-623-06 Incremental diagnostic yield of

incorporating TB-LAM test and XpertW MTB/
RIF into a diagnostic algorithm in HIV-positive
patients
H Huerga,1 G Ferlazzo Natoli,2 P Bevilacqua,1 B Kirubi,3
E Ardizzoni,2,4 J K Sitienei,5 M Bonnet1 1Epicentre, Paris,
2
`
Medecins
Sans Frontieres
(MSF), Paris, France; 3MSF, Nairobi,
4
Kenya ; Institute of Tropical Medicine, Antwerp, Belgium;
5
National Tuberculosis Program, Nairobi, Kenya. e-mail:
helena.huerga@epicentre.msf.org
Background: The value of LAM test when is incorporated into a diagnostic algorithm has not been elucidated.
We aimed to assess the diagnostic yield of an algorithm
including LAM test and LAM plus Xpert MTB/RIF in
HIV positive patients.
Design/Methods: Prospective observational study conducted in Homa Bay (HIV prevalent area of Kenya). HIV
positive patients with suspicion of pulmonary tuberculosis (TB) whether hospitalized, with CD4,200 cell/ll,
with BMI,17Kg/m2 or severely ill were eligible. Patients
received clinical exam, chest X-ray, LAM test (positive
cut-off grade 2), Xpert MTB/RIF test and MTB culture
(Thin-Layer-Agar and Lowenstein-Jensen).
We used TB
treatment initiation as primary end point and MTB

culture as reference (positive culture as TB confirmed).
Results: In total 397 patients were included: 201 (50.8%)
women; median age 35 years (IQR: 29-44); 308 (77.6%)
hospitalized; 270 (69.1%) with CD4,200cells/ll; 335
(84.4%) severely ill; 169 with BMI,17 (43.4%). A forth
of the patients could not produce sputum. Of the 300
patients with laboratory results: 49 (16.3%) were smear
positive, 73 (24.3%) Xpert positive, 104 (34.7%) LAM
positive and 83 (27.7%) culture positive. LAM test
sensitivity and specificity was 54.9% (95%CI: 43.565.9) and 72.8% (95%CI: 66.4-78.6). In total, 273/397
(68.8%) patients started TB treatment. The reasons to
start TB treatment among the 83 culture positive patients
were: 47.0% clinical situation, 16.9% suggestive X-ray,
21.7% positive LAM test, 7.2% positive Xpert, 7.2%
positive culture. The addition of the LAM test increased
the proportion of confirmed TB cases started on TB
treatment from 63.8% to 85.5% compared to the
clinical-radiological algorithm and to 92.8% when Xpert

was also included (Table 1).
Conclusion: Despite its low sensitivity, the addition of
the urine LAM to the clinical and radiological algorithm
allowed TB treatment initiation in more than 80% of
confirmed TB patients. This finding is particularly
important for patients in severe clinical situation who
usually have more difficulties to produce sputum.
Although, the low specificity of the LAM test in this
study is a concern, it might be partially due to the
imperfect reference standard in severely ill patients who
are more likely to present negative culture disseminated
TB. TB-LAM should be added to the clinical-radiological
algorithm in HIV positive patients with TB suspicion not
only when CD4 is less than 200 cells/ll but also if
patients are severely ill or hospitalized.
Overall*
(N83)
Hospitalized
(N58)
CD4,200
cells/ll
(N66)
Severely ill**
(N75)
BMI,17
(N41)
Clinicalradiological
algorithm
Additional
LAM
Diagnostic
yield, %
(95%CI)
Diagnostic
Diagnostic
yield, % Incremental yield, %
(95%CI)
yield, % (95%CI)
63.8
(52.8-73.6)
56.9
(43.6-69.3)
66.7
(54.2-77.2)
85.5
(76.0-91.7)
82.8
(70.4-90.6)
90.1
(80.9-95.9)
21.7
65.3
(53.7-75.4)
65.9
(49.6-79.1)
86.7
(76.7-92.8)
82.9
(67.5-91.9)
21.4
Additional
Xpert
25.9
23.4
17.0
92.8
(84.6-96.8)
89.7
(78.4-95.4)
95.5
(86.5-98.6)
92.0
(83.0- 96.4)
92.7
(78.8-97.7)
*Diagnostic yield: proportion of patients started on TB treatment using a

diagnostic algorithm among culture positive patients; Incremental yield:
increased proportion of patients started on TB treatment among culture
positive patients when an additional test is included in the previous
algorithm.
**Severely ill: respiratory rate higher than 30 respirations per minute,
temperature higher than 398C, cardiac rate higher than 120 beats/ minute, or
unable to walk without help.
SOA-624-06 Pulmonary tuberculosis in HIVinfected patients presenting with normal chest

radiograph and negative sputum smear
M Jojula,1 C Chatla,2 S P Chakramahanty,3
A Raghuramreddy,1 A Mounika,1 P Mamatha1 1Sri Shivani
College of Pharmacy, Warangal; 2World Health Organization,
New Delhi, 3State TB Cell, Hyderabad, India. e-mail:
chatlac@rntcp.org
Background: Deaths due to tuberculosis (TB) are high

among the TB-HIV co-infected patients. Among PLHIV
most of the instances the sputum smear is found to be
negative for MTb. Chest X-rays also dont yield any
diagnostic value mainly due to the advanced immunosuppressed condition. This study makes an attempt to
find out the utility of alternate staining methods such as
Light Emitting Diode (LED), Fluorescent Microscopy
(FM) and solid culture.
Intervention: Two sputum samples were collected from
100 successive presumptive TB cases among PLHIVs
S571
visiting ICTC at MGM hospital, Warangal, Telangana

State, India whose smear microscopy and X-ray were
negative for M. tuberculosis. All the 100 samples were
repeated with ZN microscopy and X-ray. Additionally
LED and FM microscopy procedures were conducted
with the NALC-NAOH concentration method. All the
samples were inoculated on LJ medium for solid culture
and all the positive cultures were subjected for biochemical test to identify phenotypic appearance, nitrate
reduction, niacin and PNB susceptibility test for all the
first line anti TB drugs as per standard guidelines.
Samples found positive on microscopy were cross
checked with Line probe assay (LPA).
Results: All the 100 samples collected showed negative
for M. tuberculosis on ZN technique and negative for
pulmonary TB on chest X-rays. 15 samples were positive
for MTb both on LED and FM. LPA reconfirmed the M.
tuberculosis in all the 15 samples tested with 11 sensitive
for both INH and Rif, 3 INH mono resistant and 1 Rif
mono resistance patterns. Of the 100 inoculations in LJ
medium, 25 culture inoculations were positive for M.
tuberculosis growth and also were confirmed as M.
tuberculosis strains based on morphological, biochemical test and growth was seen after 4th week of
inoculation. Of the 25 culture positives 20 were sensitive
for both INH and Rif, 4 INH mono resistant and 1
resistant to all first line anti TB drugs.
Conclusions: In smear negative and chest X-ray negative
presumptive TB cases especially in immune compromised
groups such as PLHIV, it is found to be useful to subject
the sputum samples to LED and FM methods and solid
culture wherever available. These methods clearly
demonstrated additional yield over conventional ZN
staining which can be recommended especially in the
settings where high throughput equipment such as Xpert
MTB/RIF is not available. These proactive measures can
help in early diagnosis of TB which in turn can reduce
mortality due to TB among PLHIV and break the chain of
transmission of TB.
S572
SOA-625-06 TB mortality measurement:

comparing verbal autopsy methods to
necropsy in a setting of high HIV prevalence in
Siaya County, Kenya
S Murithi,1,2 J Sitienei,1 E Mitchell,3,4 H Kipruto,1,5
E Amukoye,6 C Muturi,7 Z Nganga,2 K Cain8 1National
Tuberculosis and Lung Disease Program, Nairobi, 2Jomo
Kenyatta University of Agriculture and Technology, Nairobi,
Kenya; 3KNCV Tuberculosis Foundation, Netherlands, The
Hague, 4Amsterdam Institute for Global Health &
Development, Amsterdam, Netherlands; 5World Health
Organization, Nairobi, 6Kenya Medical Research Institute,
Nairobi, 7Jaramogi Oginga Odinga Training and Referral
Hospital, Kisumu, 8Center for Global Health, Centers for
Disease Control, Kisumu, Kenya. e-mail:
smgacherism@gmail.com
Background and challenges to implementation: Most

African countries lack complete vital registration systems. In such settings, verbal autopsy (VA) is recommended to estimate the mortality burden of Tuberculosis
(TB).It involves assigning causes of death from interviews with next-of-kin and health record review. Coding
is either by trained human coders or computer algorithms. Although widely endorsed as a legitimate
method, the performance of VA to detect TB related
deaths has never been validated against gold standard.
This study sought to assess the accuracy of VA in a
sample of deceased individuals above age one presenting
with respiratory-related symptoms in a rural Kenyan
setting with high TB-HIV co-infection.
Intervention or response: The study was conducted
between 2012- 2013. Post-mortems (PM) were done by
a pathologist. Tissue and fluid samples from major
organs were tested with liquid culture and XpertMTB/
RIF. VA interviews were conducted after a median of 50
days mourning period. InterVA4 algorithm was applied
to interview data and the probable causes of death were
ascertained. If TB appeared as among highest, second or
third highest likelihood causes on InterVA4, this was
classified as a TB death.VA questionnaires were independently reviewed by two clinicians certified to
determine causes of death from VA.
Results and lessons learnt: A total of 156 deaths were
notified. Of the eligible cases (n 98), 80% of the
families consented to autopsy. PM detected 11 deaths, all
bacteriologically-confirmed. All but one was co-morbid
with HIV. The clinicians diagnosed 20 and 29 deaths
respectively as TB-related and upon harmonization they
settled on 21 deaths. This yielded a combined overdiagnosis of 91%. Among the 11 TB deaths; clinicians
VA correctly detected 3 (Sensitivity: 27% [95%CI 1057]) and InterVA4 algorithm detected 2 (Sensitivity: 18%
[95%CI 5-48]). Among the 66 patients with non-TB as
cause of death, InterVA4 over-diagnosed TB in 14
(specificity: 78%, [95%CI 66-86]) and clinicians overdiagnosed TB in 18 (specificity: 72% [95%CI 65-86]).
Agreement between the clinicians and InterVA4 was at
(0.51 [95%CI 0.29-0.73]).
Conclusions and key recomendations: Current VA
methods overestimate TB mortality burden in a setting
with high HIV prevalence. Low sensitivity and specificity
imply they cannot function as interim proxy measures of

TB mortality.Vital registration systems and minimally
invasive autopsy are options for policy makers to
consider.
SOA-626-06 A TB-HIV integration intervention

and morbidity, mortality and retention in care
among individuals with TB and HIV: MERGE
cluster randomised trial
Kufa),1,2 K Fielding,3 P Hippner,1
T Chakezha (nee
4
K Kielmann, A Vassall,3 G Churchyard,1,2 A Grant,2,3
S Charalambous1,2 1RESEARCH, The Aurum Institute,
Johannesburg, 2University of the Witwatersrand,
Johannesburg, South Africa; 3London School of Hygiene &
Tropical Medicine, London, 4Queen Margaret University,
Edinburgh, UK. e-mail: tendesayik@yahoo.com
Background: There are limited data on effectiveness of

interventions to improve TB and HIV integration and
their effect on patient relevant outcomes. We describe
results from a cluster-randomised trial to evaluate the
effect of a TB and HIV integration intervention on
morbidity and/ or mortality and retention in care, in
South Africa.
Design/Methods: Eighteen primary care clinics were
randomised to intervention or control. The intervention
was clinic-specific activities supported by intervention
staff (TB-HIV integration and TB screening officers) that
helped address gaps in integration and improve TB
screening for HIV-positive people. The control arm was
current standard of care. The composite outcome of
death or hospitalisation during 12 months of follow up
was measured in two cohorts of newly diagnosed (i) TB
and (ii) HIV patients. Retention in care was measured as
the proportion of HIV-positive TB patients who
remained in HIV care 12 months after enrolment.
Results: Of 4003 individuals screened, 3024 and 731
individuals newly diagnosed with HIV and TB, respectively, were enrolled. In the HIV and TB cohorts 62.3%
and 45.5% were female, median ages 34 and 36 years. In
the TB cohort 79.6% (582) were HIV-positive. The
incidence of death/hospitalisation was similar between
the arms: i) HIV cohort: adjusted rate ratio (aRR) 1.09
(95% confidence interval [CI] 0.85- 1.40) and ii) TB
cohort: aRR 1.16 (95%CI 0.65- 2.07). The proportion of
HIV-positive individuals newly diagnosed with TB and
retained in care at 12 months was also similar between
the arms: aRR 1.08 (95%CI 0.91- 1.28)(see Table).
Although there was evidence that the intervention was
understood and well received by participating clinic staff,
intervention staff spent most of their time providing
direct patient care as opposed to the more strategic roles
that were intended. At the end of the intervention period,
there was no evidence of increased co-location of services
or higher coverage of integrated TB-HIV services at
intervention vs. control clinics.
Conclusion: The intervention as implemented did not
have an effect on morbidity, mortality or retention in care
among individuals newly diagnosed with TB and with
HIV, likely because the intervention did not improve
integration. Other interventions to improve integration

need to be evaluated.
Table: Rates, adjusted and unadjusted risk ratios for the
outcome comparing intervention with control arms
Outcome
Intervention Control
Rate /
Rate /
100 pyrs
100 pyrs Unadjusted
(# cases / (# cases / Rate Ratio1
(95% CI)
pyrs)
pyrs)
Adjusted
Rate
Ratio1,2
(95% CI)
Morbidity and/
mortality
within 12
months
among
individuals :
Newly
13
10.8
1.18
1.09
diagnosed
(182/1395 ) (146/1352) (0.93- 1.50) (0.85- 1.40)
with HIV
Newly
diagnosed
with TB
Retention in
care among
HIV positive
individuals
newly
diagnosed
with TB
17.9
(67/375)
11.4
(32/280)
1.37
1.16
(0.77- 2.43) (0.65- 2.07)
% (n/N)
63.4%
231/336
adjusted
Risk
unadjusted
Ratio1,2
Risk Ratio1
(95% CI)
(95% CI)
% (n/N)
56.0%
1.11
1.08
141/252 (0.89 - 1.37) (0.91- 1.28)
Comparing intervention vs control; 2adjusting for age, sex and baseline

imbalances
SOA-627-06 An integrated community based

TB-HIV adherence support model provides an
alternative to DOT for TB patients in Cape
Town, South Africa
R Kaplan,1 J Caldwell,2 S Adriaanse,2 S Hermans,1,3
L Mtwisha,1 L-G Bekker,1 K Jennings,2 R Wood1
1
University of Cape Town, Cape Town, 2City Health, Cape
Town, South Africa; 3University of Amsterdam, Amsterdam,
Netherlands. e-mail: Richard.Kaplan@hiv-research.org.za
Background: Directly observed therapy (DOT) is advocated by the WHO to support adherence to TB
treatment. In Cape Town, with TB-HIV co-infection
rates of 50% and patients on antiretroviral treatment
(ART) self-supervising treatment with the support of
community care workers (CCWs), the need was perceived for a single adherence support programme for TB
and HIV-positive patients. In 2011, the City of Cape
Town introduced an integrated adherence support model
as an alternative to DOT at primary care TB and HIV
clinics. The model consisted of educational sessions on
TB and HIV treatment by facility based counselors; home
assessments by CCWs and a multidisciplinary meeting
after 2 weeks of in-clinic DOT to decide whether patients
were eligible for community support with weekly home
visits by CCWs or whether they should continue with
daily supervised treatment as either in-clinic, field-based
S573
or workplace DOT. We evaluated the impact of this

model on TB default and death rates.
Methods: We retrospectively analysed the electronic TB
register data of patients treated during a 6 month period
before and after the introduction of the new adherence
model. Multivariable random effects Cox regression
analyses were used to compare the hazard of default and
death before and after the intervention adjusting for age,
gender, HIV status, calendar year, type of TB and for
clustering by clinic.
Results: The pre-intervention population consisted of
8761 TB patients and the post-intervention population of
7433 treated in 46 primary care TB clinics. Baseline
characteristics were similar; 50% were HIV positive,
median age was 32 years (IQR 24-41) and 55% were
male. More HIV-positive patients were on ART at the
start of TB treatment after the intervention (31% vs.
24%) and more patients who were not on ART, started
antiretroviral treatment after the intervention (76% vs.
60%). After the intervention, 79% of patients were
placed out for community adherence support compared
to 51% who received field-based DOT before the
intervention. The adjusted hazard ratio (aHR) for death
associated with the intervention was 1.03 (95%CI 0.781.36) and the aHR for default was 1.01 (95%CI 0.821.25).
Conclusion: The integrated adherence support model did
not result in higher default or death rates when compared
to DOT and was associated with an increase in ART
uptake. It could provide a viable alternative adherence
support system for an integrated TB and HIV treatment
programme.
SOA-628-06 CD4 count distribution, ART uptake

and TB treatment outcomes among HIVpositive TB patients in Cape Town, 2009-2013
R Kaplan,1 S Hermans,1,2 J Caldwell,3 K Jennings,3
L-G Bekker,1 R Wood1 1University of Cape Town, Cape
Town, South Africa; 2University of Amsterdam, Amsterdam,
Netherlands; 3City Health, Cape Town, South Africa. e-mail:
Richard.Kaplan@hiv-research.org.za
Background: In Cape Town, the roll-out of antiretroviral

treatment (ART) in primary health care clinics has
increased steadily over the last decade with an estimated
coverage of 63% of HIV positive patients in 2013. How
this has influenced the characteristics of the population
of patients presenting to the primary care TB programme
is unknown. This study aimed to examine the HIV
prevalence and ART usage among TB patients in the
primary care TB programme in Cape Town from 2009 to
2013 and to evaluate the degree of immunosuppression
through an analysis of the CD4 count distribution.
Methods: Data from the electronic TB register (20092013) on all newly registered TB patients 715 years
were analyzed retrospectively. Descriptive statistics were
used to quantify the annual changes in CD4 counts, ART
usage at start of TB treatment, ART uptake during TB
treatment and TB treatment outcomes (available for
years 2009-2012). Differences in HIV prevalence and
S574
median CD4 counts were tested using the Chi2 test for
trend and the Cuzick nonparametric test for trend
respectively.
Results: 122 762 patients were treated in 101 primary
care TB clinics over the 5 year period. The absolute
number of notified HIV-positive TB cases decreased by
11% over the time period, from 12 567 in 2009 to 11
162 in 2013. The HIV prevalence among TB patients
decreased from 50% in 2009 to 48% in 2013 (P ,
0.001). The median CD4 count at the start of TB
treatment increased from 152 cells/mm3 [interquartile
range (IQR) 69-278] in 2009 to 176 cells/mm3 [IQR 74324] in 2013 (P , 0.001). The CD4 count distribution
showed a shift to CD4 categories .200 cells/mm3 in
2013 with the greatest decrease occurring in the
categories between 100 200 cells/mm3 (Figure). More
patients entered the TB programme on ART in 2013
compared to 2009 (34% vs. 10%) and ART uptake
among those not on ART increased substantially from
37% to 76%. TB treatment outcomes showed an 82%
cure or completion in 2012 compared to a 79% cure or
completion rate in 2009.
Conclusion: The increase in median CD4 counts, the
high treatment cure or completion rate and the increase
in ART uptake indicate well-functioning TB and ART
treatment programmes. However the absolute decrease
in HIV-positive TB cases, while promising, was modest
and the proportion of highly immune compromised
patients was still high, indicating that the increased ART
coverage has not lead to a substantial decline in TB cases.
SOA-629-06 TB as a cause of hospitalization

and death among people living with HIV
worldwide: a systematic review and metaanalysis
N Ford,1 Z Shubber,2 A Matteelli,1 G A Meintjes,3
K Kranzer,4 B Grinsztejn,5 M Doherty,1 H Getahun1
1
World Health Organization, Geneva, Switzerland; 2Imperial
College, London, UK; 3University of Cape Town, Cape Town,
South Africa; 4London School of Hygiene & Tropical
Medicine, London, UK; 5Ministry of Health, Brazil, Rio de
Janeiro, Brazil. e-mail: fordn@who.int
Background: Despite significant progress in improving

access to antiretroviral therapy (ART) over the last
decade, substantial numbers of people living with HIV/

AIDS (PLHIV) in all regions continue to experience
severe illness and require hospitalization. We undertook
a global review of the proportion of hospitalizations and
deaths due to tuberculosis (TB) in PLHIV.
Design/Methods: Seven databases were searched using a
pre-defined protocol for studies reporting causes of
hospitalizations among PLHIV since 01 January 2007
irrespective of age or geographical region and without
language restriction (earlier cohorts were included
provided they also reported data from 2007 onwards).
The proportion of hospitalizations and in-hospital
mortality attributed to TB and other illnesses was
estimated using random effects meta-analysis and the
influence of antiretroviral therapy and calendar year
(study midpoint) was assessed by meta-regression.
Results: From an initial screen of 9049 records, 68

cohorts were identified, providing data on 30 981 adults
(32 832 hospitalizations) and 2526 children (2792
hospitalizations) across 39 countries. 70.7% (95%CI
49.0-92.4%) of patients were already known to be HIV
positive at the time of admission but only 47.7% (31.264.1%) were on antiretroviral therapy. Overall, 18.0%
(15.5-21.6%) of all adult hospitalizations were due to
TB, making it the leading cause of hospitalization
overall; the proportion of adult hospitalizations due to
TB exceeded 10% in all regions except the European
region (Figure). 10.8% (7.6-13.9%) of all paediatric
hospitalizations were due to TB and TB was the 4th
leading cause of hospitalization (after bacterial pneumonia, diarrhoea, and severe anaemia). Among patients
who died following hospitalization, TB accounted for
26.1% (19.4-32.8%) of all adult deaths and 30.1%
(11.2-48.9%) of all paediatric deaths. In meta-regression, an increase in the proportion of patients on ART at
admission was associated with a decrease in the
proportion of hospitalizations attributable to TB (P ,
0.01). There was no change in TB as a cause of
hospitalization over time.
Conclusion: TB continues to be a leading cause of

hospitalization and death among adults and children
living with HIV worldwide. Increased coverage of
antiretroviral therapy reduces the risk of hospitalization;
however, the relative importance of TB as a cause of
severe illness has not decreased in recent years. The
contribution of TB to hospitalization and death may be
underestimated due to limited availability and sensitivity
of diagnostics in resource-limited settings. This study
underscores the continued need to scale up TB prevention and expedite TB diagnosis among PLHIV.
SOA-630-06 HIV and alcohol intake are

important determinants of non-adherence to
tuberculosis treatment in pragmatic settings in
urban Uganda
H Babikako,1 A Okwera,2 C C Whalen,3 E Mupere1
College of Health Sciences, Makerere University, Kampala,
2
Mulago National Referral Teaching Hospital, Kampala,
Uganda; 3College of Public Health, University, Athens, GA,
USA. e-mail: babikako@yahoo.com
Background: Control of tuberculosis (TB) epidemic in

most sub-Saharan Africa countries is challenged by
human immunodeficiency virus (HIV) infection, and
yet the systematic understanding of the impact of HIV
and its treatment on adherence to TB treatment is not
well characterized.We established non-adherence to TB
treatment among HIV positive and negative adult
patients attending programme clinics in Kampala, the
capital of Uganda in Africa.
Design/Methods: Using a cross-sectional study design,
469 smear positive TB patients attending five programme
clinics in Kampala were studied. Adherence was measured using a Morisky scale among patients completing
2, 5 and 8 months of treatment.
Results: The study found 29% (135/469) of the study
participants to be non-adherent. Of the 135 participants
that were non-adherent, 33% were HIV positive. Nonadherence increased with the duration of TB treatment.The proportion for non-adherence was 21% (40/
190) at 2 months, 30% (41/135) at 5 months, and 38%
(54/144) at 8 months. HIV positive patients who had
completed medication at 5 and 8 months had an
increased likelihood for non-adherence, respectively.
For example, HIV positive patients who had completed
medication at 5 months were three times more likely to
be non-adherent (OR 3.21 (95%CI: 1.354, 7.647))
compared to patients that had completed treatment at
month two, adjusting for age, sex, patient being new or
transfer in, alcohol intake status, duration of treatment.
Among 314 HIV negative participants, patients with
history of alcohol intake were two times (OR 2.211
(95%CI: 1.229, 3.980)) more likely to be non-adherence,
adjusting for age, sex, patient being new or transfer in,
duration of treatment.
Conclusion: The study found a high burden of nonadherence among patients receiving care in national
programme TB clinics in urban Uganda; HIV positive
patients and HIV negative with history of alcohol intake
S575
were more likely to be non-adherent. Tuberculosis

programmes should enhance pre- and in-treatment
adherence counseling and education throughout the
duration of therapy, and HIV positive and patients with
history of alcohol intake should be targeted.
SOA-631-06 Point-of-care C-reactive protein

based TB screening may improve the efficiency
of ICF and increase uptake of IPT among
patients new to HIV/AIDS care
C Yoon,1 F Semitala,2,3,4 J Katende,2 P Byanyima,4
A Andama,3,4 I Ayakaka,4 M Kamya,2,3,4 A Cattamanchi1
1
University of California, San Francisco, San Francisco, CA,
USA; 2Makerere University Joint AIDS Program, Kampala,
3
4
Makerere University-University of California, San Francisco
Research Collaboration, Kampala, Uganda. Fax: (1 )415 695
1551. e-mail: christina.yoon@ucsf.edu
Background: Tuberculosis (TB) is the leading killer of

PLHIV. Early initiation of antiretroviral therapy (ART),
early diagnosis of TB, and early provision of isoniazid
preventive therapy (IPT) can substantially reduce the
number of HIV-TB deaths. However, many patients with
HIV present late to care leading to delayed ART
initiation and increased risk of TB. We compared the
efficiency of intensified case finding (ICF) and opportunities to provide IPT when patients new to HIV/AIDS
care were screened for TB using point-of-care (POCCRP) levels vs. symptom-based assessment.
Methods: We performed symptom-screen assessment,
POC-CRP testing (normal 610 mg/l), and sputum-based
TB testing (Xpert x1, culture x2) on consecutive ARTnave adults with CD4 6350 cells/ll entering care at an
HIV/AIDS clinic in Uganda from July 2014 to April
2015. We compared the proportion of TB cases detected
by Xpert and culture among patients who screenedpositive, and the proportion of patients eligible for
immediate IPT (specifity), by either screening strategy.
Results: Of 261 patients new to care (median CD4 144
cells/ul), 64 (25%) had culture-positive pulmonary TB.
Nearly all patients (92%, 241/261) screened-positive for
TB by symptoms including 63/64 patients with TB
(sensitivity 98%, 95%CI: 92-100). Xpert testing of the
241 symptomatic patients identified 35 (57%) culturepositive TB cases. Of the 197 patients without TB, only
19 patients screened-negative by symptoms (specificity
10%, 95%CI: 6-15) and would have been immediately
eligible for IPT. In contrast, 50% (131/261) of all
patients screened-positive for TB by POC-CRP, including
62/64 patients with TB (sensitivity 97%, 95%CI: 89100). Xpert testing of the 131 patients with elevated
POC-CRP also identified 35 (57%) culture-positive TB
cases but required only half as many Xpert assays. Of the
197 patients without TB, 128 patients screened-negative
by POC-CRP levels (specificity 65%, 95%CI: 58-72), a
greater than 6-fold increase in the proportion of patients
immediately eligible for IPT without the need for
confirmatory testing, compared to the symptom screen.
Conclusion: 1 in 4 patients new to HIV/AIDS care have
culture-positive TB; this population is an extremely high-
S576
risk group for whom TB screening should be prioritized.

Targeting POC-CRP based- rather than symptom-based
screening to patients new to care can improve the
efficiency of ICF and increase uptake of IPT while
maintaining a similar TB case detection yield.
SOA-632-06 Chronic respiratory disease in

older children and adolescents with vertically
acquired HIV infection
G Mchugh,1 J Rylance,2,3 E Majonga,1,4 J Metcalfe,5
T Bandason,1 H Mujuru,6 K Kranzer,4 R Ferrand1,4
1
Biomedical Research and Training Institute, Harare,
Zimbabwe; 2University of Liverpool, Liverpool, 3Liverpool
School of Tropical Medicine, Liverpool, 4London School of
Hygiene & Tropical Medicine, London, UK; 5University of
California San Francisco, San Francisco, CA, USA; 6University
of Zimbabwe, Harare, Zimbabwe. e-mail:
graceandandre@gmail.com
Background: Chronic lung disease is a major cause of

morbidity in vertically HIV-infected children and adolescents. While we have previously shown that the
burden in these children is extremely high, the association between chronic lung disease and HIV has not been
established due to the lack of a comparison group. We
compared symptoms of lung disease and spirometry in
HIV-infected and HIV-uninfected older children and
adolescents.
Design/Methods: HIV infected older children and
adolescents (age 6-16 years) with vertically-acquired
HIV on antiretroviral therapy for at least 6 months
attending an HIV clinic in Harare, Zimbabwe were
consecutively recruited. Assessment included a symptom
questionnaire, MRC dyspnoea scale, clinical examination, spirometry and shuttle walk testing. Frequency agematched HIV-negative children were recruited from 7
primary health care clinics in Harare.
Results: 200 HIV-infected and 69 HIV-uninfected children were recruited. Age (median 11 years) and sex of the
two groups was comparable (see table).In the HIV
infected group, median CD4 count was 729cells/ll (iqr
479-935), mean age at diagnosis was 5 years old and 157
(84%) had an undetectable HIV viral load (,1000cp/
ml). One third of the HIV-infected children (n 76, 38%)
had been treated for tuberculosis compared to only one

child among the HIV-uninfected children (1%) The
proportion of children previously diagnosed with asthma
was similar in both groups (3% vs. 3.5%). A quarter of
the HIV-infected children (n 41, 25%) who had
spirometry performed showed either obstruction or
reduced FVC with poor reversibility, compared to only
4% of HIV-uninfected participants with spirometry data
HIV N 200 HIV- N 69 P Age,y,median (iqr) 11(912) 11(9-13) 0.72 Sex-Male,% 52 39 0.06 MRC
Dyspnoea scale .1,% 14 0 ,0.01 Height for age, z
score ,-2,% 34 10 ,0.01 Weight for age, z score,-2,%
28 9 ,0.01 Oxygen saturation ,88% at rest, % 1 0 0.39
Desaturation 75% on exercise,% 12 9 0.39 Duration of
shuttle walk, mins, mean 10m 23s 11m28s 0.03 FEV
zscore,mean(sd) 0.73(1.20) 0.29(0.82) ,0.01 FVC
zscore, mean(sd) 0.79(1.24) 0.25(0.90) ,0.01 FEF
25%-75% zscore mean(sd) 0.082(1.21) 0.08(1.02)
0.49
Conclusion: There is a substantial burden of chronic lung
disease among HIV-infected older children despite
treatment with antiretroviral therapy, in comparison to
HIV negative counterparts. The aetiology of such chronic
lung disease needs investigation and interventions to
treat lung disease are urgently required.
04. TB epidemiology: from Malawi to

Taiwan
SOA-633-06 Risk stratification of adult
tuberculosis contacts: a population-based
study in Taiwan
J-Y Feng,1 S Chen,1 W Fan,1 S Wei-Juin,1 C Liu1 1Taipei
Veterans General Hospital, Taipei, Taiwan. e-mail:
peterofeng@vghtpe.gov.tw
Background and challenges to implementation: Contact

investigation is an important and effective active-case
finding strategy for TB control programs, but also
requires abundant resources in public health. The current
World Health Organization (WHO) recommendation
for contact investigation suggests follow-up screening in
the first year after exposure. However, most of the TB
contacts studies have included both adults and children,
or enrolled children only. The adequacy of one-year
follow-up screening for all TB contacts remains uncertain. Understanding the dynamic pattern of active TB
occurrence among TB contact and identifying contacts
with the highest risk are of paramount clinical importance.
Intervention or response: We retrieved the data from the
National Health Insurance Research Database in Taiwan
from 2000 to 2011, which covered more than 99% of the
23 million people of Taiwan. TB contacts were identified
by diagnostic codes (V01.1 in the International Classification of Diseases, 9th revision, and the clinical
modification ICD-9-CM) in conjunction with chest plain
film and/or sputum acid-fast smear examination and/or
tuberculin skin test (TST). We matched each TB contact
with four non-exposed subjects by age and sex, and

presence of comorbidities on the same index date of
diagnosis. The occurrence of TB, the major outcome of
the present study, was determined based on the
compatible ICD-9-CM code (010-018 in ICD-9-CM).
The diagnosis of TB was validated by the prescription of
at least two anti-TB medications for more than 28 days.
We compared the difference in the occurrence of active
TB between TB contact and matched cohort. The clinical
risk factors associated with TB occurrence were also
investigated.
Results and lessons learnt: A total of 8659 adult TB
contacts and 35 636 matched cohorts were included for
analysis. Overall, 72 TB contacts and 67 matched
cohorts were documented with active TB during the
follow-up period, and the mean annual incidence of TB
were 312.8 and 72.2 cases per 100 000 person-years,
respectively (adjusted HR 4.39, 95%CI 3.156.12, P ,
0.001). The incidence rate of active TB in the first,
second, and third years were 567.1 cases, 106.4 cases,
and 229.3 cases per 100 000 person-years, respectively,
for TB contacts. The adjusted hazard ratio of the first,
second, and third years were 12.30 (95%CI 6.7322.48),
1.64 (95%CI 0.644.23) and 4.10 (95%CI 1.6310.33),
respectively. In the multivariate analysis, the independent
risk factors associated with active TB development
among adult TB contacts were being 7 65 years old
(HR 2.15, 95%CI 1.223.81), male (HR 1.74, 95%CI
1.072.81), having autoimmune diseases (HR 2.60,
95%CI 1.404.81), COPD (HR 1.95, 95%CI 1.16
3.27), and liver cirrhosis (HR 2.61, 95%CI 1.036.65).
When categorize the TB contacts into high-risk population (two or more risk factors) and low-risk population
(less than two risk factors), Kaplan-Meier analysis
demonstrated that high-risk TB contacts had a significantly higher cumulative incidence of active TB and the
curves separated immediately after exposure.
Conclusions and key recommendations: This nationwide, population-based cohort study from a TB endemic
area indicated that although the annual incidence of TB
is highest during the first year, the annual incidence
remains significantly higher in TB contacts during the
third year. The extension of the follow-up period to three
years should be considered for adult TB contacts,
especially those with more risk factors.
S577
SOA-634-06 Deciphering tuberculosis dynamics

in East Greenland: genome based molecular
epidemiology of Mycobacterium tuberculosis
in a high-incidence area
K Bjorn-Mortensen,1 A Bengaard Andersen,2 B Soborg,1
T Lillebaek,1 A Koch,1 K Ladefoged,3 T Kohl,4
S Niemann4,5 1Statens Serum Institut, Copenhagen,
2
University Hospital of Copenhagen, Copenhagen, Denmark;
3
Queen Ingrid Hospital, Nuuk, Greenland;
4
Forschungszentrum Borstel, Leibniz-Zentrum fur
Medizin
und Biowissenschaften, Borstel, 5German Center for
Infection Research (DZIF), Borstel, Germany. e-mail:
kabm@ssi.dk
Background: Tuberculosis (TB) is again a major health

problem in Greenland. TB outbreaks often occur in
isolated settings with complex social networks revealed
by contact tracing. This complexity makes it difficult to
identify cases of active Mycobacterium tuberculosis
transmission in order to confine the outbreaks. TB
incidence in the isolated East Greenland has increased
dramatically in recent years to approximately 1000
notified TB cases per 100 000 inhabitants and approximately 40% of the East Greenlandic population is
infected with M. tuberculosis. Since 2004, all culturepositive TB samples from East Greenland have been
genotyped with mycobacterial interspersed repetitive
unit-variable number of tandem repeats (MIRU-VNTR).
Available data showed that almost all strains were from
two major clusters with very similar MIRU-VNTR
genotypes, making it difficult to distinguish new from
old cases.
Objective: The study aimed to characterize transmission
patterns in a population-wide whole genome sequencing
(WGS) study including all culture-positive cases within a
21-year period. The main objective was to determine to
what extend WGS contributes to the understanding of an
outbreak in a high TB incidence, low diversity setting.
Methods: Using Illumina Nextera XT library preparation kits and the NextSeq500 next generation sequencing
platform, we performed WGS of 182 isolates (98% of
culture-positive cases). After reference mapping to the
H37Rv genome, detected variants were combined to
yield a set of phylogenetically informative single nucleotide polymorphism (SNP) positions.
Results: A maximum parsimony tree based on these
positions revealed the same pattern as seen with MIRUVNTR, but with a much higher resolution. Employing a
maximum distance of 12 SNPs between isolates, four
groups were created, which overall correlated well with
the MIRU-VNTR clustering. Both time and place of
notification correlated well with WGS data, and
transmission links suggested by WGS fit well with
expansion over time. Also, epidemiological data and
WGS data together suggest a very clear geographical
distinction of clones.
Conclusion: WGS based genotyping describes in detail
the longitudinal transmission pattern of tuberculosis in
an isolated, high incidence, low variance setting.
Therefore, even among isolates with the same, or almost
identical, MIRU-VNTR type in a high incidence setting,
S578
WGS performs well and offers a much higher resolution

than previously used genotyping methods.
SOA-636-06 Diversity of MIRU-VNTR among

Mycobacterium tuberculosis Central Asian
family isolates from Nepalese patients
SOA-635-06 Increasing burden of tuberculosis

among foreign-born populations in Japan:
trend and issues
Y Shah,1 A Poudel,1 H Mahmoud Diab,1 J Thapa,1

C Menda,1 B Maharjan,1,2 C Nakajima,1 Y Suzuki1
1
Hokkaido University, Sapporo, Japan; 2German Nepal TB
project (GENETUP), Kathmandu, Nepal. e-mail:
yogendra.90@gmail.com
L Kawatsu,1 A Ohkado,1,2 K Uchimura,1 S Yoshimatsu,1

K Izumi,1,2 K Ito1 1Research Institute of Tuberculosis, Japan
Anti-Tuberculosis Association, Tokyo, 2Nagasaki University
Graduate School of Biomedical Sciences, Nagasaki, Japan.
e-mail: kawatsu@jata.or.jp
Background: Compared to other industrialized countries, the burden of tuberculosis (TB) among the foreignborn populations has been relatively low in Japan however, with social and economic factors, including the
shortage of workers in several industries such as
construction and care for the elderly, pushing Japan for
a more open immigration policy, more foreigners are
expected to enter Japan. TB control among foreign-born
persons has thus become an urgent issue.
Objective: To identify the trend of TB among foreignborn populations in Japan and issues for TB control.
Method: We analyzed the publicly available data from
the Tuberculosis Surveillance Center to identify the trend
of TB by age-groups, nationality and time of entry to
Japan for the period 2003-2013. We also compared the
treatment results of foreign-born to Japanese patients,
and in terms of occupational category. v2 test was
performed to compare proportions and calculate odds
ratios.
Results: The number of foreign-borns among newly
notified TB patients had increased from 906 in 2003 to
1064 in 2013. The proportion of foreign-borns among
the newly notified TB patients was largest among those
aged 20 to 29 years in 2013, 42.7% of patients aged 2029 were born outside Japan. Of the 1064 foreign-born
patients, 432, of whom 34% were from China, had
entered Japan within the last five years. Of the 184
foreign-born patients who had been staying in Japan for
more than five years, 29.1% were from the Philippines.
Compared to the Japanese patients, lost to follow-up and
transferred-out was significantly higher among the
foreign-borns both in the age groups 20-59 (odds ratio
(OR) 3.54, 95% Confidence Interval (CI) 2.47-5.05, P
0.000) and aged above 60 (OR5.44, 95%CI 2.1112.50, P 0.000). However, death was significantly
higher among the Japanese patients than foreign-borns in
the age group 20-59 (OR7.16, 95%CI 1.22-290.00, P
0.016). Multi-drug resistance was significantly higher
among foreign-born patients (v252.27, P 0.000).
Treatment results did not differ significantly in terms of
occupational category among the foreign-born patients.
Conclusion: Lost to follow-up and transferred-out was
significantly higher in foreign-born patients than the
Japanese counterparts in both age groups 20-29 and 60
and above. Public health effort must be strengthened to
deliver effective patient education and adherence support, as we continue to expect increasing number of
foreigners arriving to Japan.
Background: Tuberculosis (TB) is caused by Mycobacterium tuberculosis and poses major challenges and
public health problems in Nepal. Central Asian Strain
(CAS) family has been reported as one of the most
predominant genotypes of M. tuberculosis in South
Asian countries including Nepal. Mycobacterial interspersed repetitive units-variable number of tandem
repeats (MIRU-VNTR) is a reliable and reproducible
genotyping method for the differentiation of M. tuberculosis isolates. The main aim of this study was to
elucidate the epidemiological link among M. tuberculosis
CAS family strains within Nepal and those in surrounding countries, as well as providing insight into the
transmission of TB for better monitoring of disease and
strengthening of control measures.
Design/Methods :A total of 145 M. tuberculosis CAS
isolates from Nepalese patients were analyzed by
spoligotyping and 24 loci MIRU-VNTR. The individual
and cumulative Hunter Gaston Discriminatory index
(HGDI) were calculated and result categorized as poor,
moderate and highly discriminatory. Clustering analysis
was analyzed using www.miru-vntrplus.org.
Results: Based on spoligotyping results, SIT26 was the
most prevalent (47%, n 63) shared type followed by
SIT599 (10.34%, n 15). The cumulative HGDI of 24locus VNTR were equal to the 15-loci MIRU-VNTR
(HGDI, 0.9942; clustering rate, 26.2%). Combining
both MIRU-VNTR and spoligotyping identified 47
isolates exhibiting 19 clusters (RTI-20%), which suggested that those isolates were from patients with
ongoing transmission between them. On other hand,
the majority of (80%) isolates did not show clustering
but had unique patterns suggesting a long history of M.
tuberculosis CAS strain in Nepal.
Conclusion: The proposed 15 loci MIRU-VNTR typing
scheme are well suited to assess the M. tuberculosis
population structure and diversity, trace back the
transmission dynamic and epidemiological link among
M. tuberculosis CAS family strain within Nepal and
surrounding countries; this will enhance the TB national
epidemiological surveillance and management program
for control of TB transmission in Nepal.
SOA-637-06 High mortality and prevalence of

undiagnosed HIV and tuberculosis in adults
with chronic cough in Malawi: a prospective
cohort study
M Nliwasa,1 P Macpherson,2 M Mukaka,3 A Choko,3
A Mdolo,3 G Chipungu,1 H Mwandumba,3 L Corbett1,3,4
1
Helse Nord Tuberculosis Initiative, Blantyre, Malawi;
2
University of Liverpool, Liverpool, UK; 3College of Medicine,
University of Malawi, Blantyre, Malawi; 4London School of
mnliwasa@gmail.com
Background: Chronic cough is associated with high

mortality in health facility attenders and hospitalized
adults in Africa. However, the fate of adults with chronic
cough at community level has not previously been
investigated.
Methods: Adults reporting chronic cough (72 weeks)
during a household census of 19 936 adults in urban poor
setting in Blantyre, Malawi were randomly sampled and
age-frequency matched with adults without chronic
cough (controls). At 12 months, participants were traced
to establish vital status, offered HIV testing and
counselling and investigated for tuberculosis (sputum
for Xpertw MTB/RIF, smear microscopy and culture,
followed by hospital clinician review) if symptomatic.
Verbal autopsies to establish cause of death were
performed.
Results: Seventy one per cent (245/345) of participants
with baseline cough and 70.4% (243/345) of controls
were traced. A significantly higher proportion of
participants with baseline cough (16/178, 8.9%) compared to controls (7/187, 3.7%; P 0.039) were
diagnosed with tuberculosis. Over one-third (56/162,
34.6%) of participants with baseline cough were HIVpositive, substantially higher than among controls (32/
170, 18.8%; P 0.005), with 26.8% and 18.8%
respectively being previously undiagnosed. The 12month risk of death was high in both participants with
baseline cough (10/245, 4.1%) and controls (6/243,
2.5%, P 0.317).
Conclusion: In high HIV and tuberculosis burden
settings, undiagnosed HIV and tuberculosis are worryingly common among adult community members with
chronic cough, and the subsequent risk of death is high.
Upscaling of community active case finding and interventions to promote timely seeking of HIV and TB care
are urgently needed.
SOA-638-06 Will the European Union reach the

United Nations Millennium Declaration target
of a 50% reduction of tuberculosis mortality
between 1990 and 2015?
M Van Der Werf,1 F Hruba,1 S Bonfigli1 1European Centre
for Disease Prevention and Control, Stockholm, Sweden. Fax:
(46) 8 5860 1296. e-mail:
marieke.vanderwerf@ecdc.europa.eu
Background: At the end of 2015, the world will assess

achievements on the Millennium Development Goals
(MDG), including MDG 6 targets to reduce tuberculosis
S579
(TB) death rate by 50% compared with 1990. In the

European Union (EU), high quality vital registration data
with detailed information on cause of death allows
assessing whether the target set for TB mortality has been
reached or is within reach by 2015.
Design/Methods: We obtained information from the
statistical office of the European Union (Eurostat)
database on population number, total number of deaths,
and causes of death for 28 EU countries for years 19942011. Using the number of TB deaths reported in the
causes of death data registration and the population data
on the 1st of January in the same year, we calculated the
TB death rate and estimated its reduction. We also
calculated the adjusted TB death rate by taking into
account incomplete coverage (deaths with no cause
documented) and ill-defined causes of death (International Classification of Diseases 10th Revision codes
R00R99) as described in the technical appendix of the
Global Tuberculosis Report 2014. Finally, we interpolated missing data between existing data points and
predicted the death rate for the years up to 2015 using
exponential smoothing state space models for time series.
Due to data completeness, data from years 1999-2011
were used for EU estimates.
Results: The TB death rates in EU countries ranged from
0.27 to 11.3 per 100 000 in 1994 and from 0.22 to 7.04
per 100 000 in 2011. Between 1999 and 2011 the TB
death rate decreased from 1.81 to 1.04 per 100 000, i.e.,
by 42% for the 28 EU countries. Adjusting the TB death
rate for incomplete coverage and ill-defined causes of
death provided a 49% reduction, from 2.09 in 1999 to
1.08 per 100 000 in 2011 in the EU. The projected
reduction in adjusted TB death rate for 2015 was 64%.
Conclusion: We predict that the EU Member States will
reach the MDG target for reduction of TB death rate by
50% by 2015. The new End TB Strategy requires a
further reduction of the number of TB deaths of 35% by
2020 compared to 2015, which will challenge TB
prevention and care in the EU.
SOA-639-06 A historic view on understanding

causes of the tuberculosis epidemic: mortality
and living conditions in Bern, Switzerland,
between 1890 and 1950
1
K Zurcher,
M Ballif,1 M Zwahlen,1 M Egger,1 L Fenner1,2
Institute of Social and Preventive Medicine, Bern, 2Swiss

Tropical and Public Health Institute, Basel, Switzerland.
e-mail:
1
Background: Poor living conditions are the breeding

ground for poverty-related diseases such as tuberculosis
(TB). In low-income countries with a high TB incidence,
such living conditions are common, but not in highincome settings. We studied TB mortality and living
conditions in the capital city of Bern, Switzerland, a
hundred years ago.
Methods: We screened mortality registries (certified by
official autopsies) and public health reports from 1890 to
1950 which were available at the city council of Bern.
S580
Causes of death were grouped in three categories:

injuries, communicable and non-communicable diseases.
Results: Overall mortality steadily decreased from 225 to
86 deaths per 10 000 population between 1890 and 1950
(Figure), but peaked in 1918 due to the influenza
epidemic. During the same time period, the population
increased from 46 000 to 146 700. The absolute number
of deaths due to communicable diseases declined from
467 in 1890 to 48 in 1950, while deaths due to
noncommunicable diseases increased from 644 to
1099. During this period, TB mortality decreased 14fold, from 47.4 to 3.3/10 000 (Figure). Construction of
new residential areas offered better living conditions in
the city outskirts from 1898 onwards, but improvements
in the city centre started only later in 1911. TB mortality
differed substantially within the city, with higher
mortality in areas of higher population density (correlation coefficient r 0.98, P 0.003), areas with a lower
number of rooms per apartment (r 0.79, P 0.007),
and areas with a lower number of windows per
apartment ( r 0.72, P 0.02, based on data from
1911). The densely populated city centre showed a 2-fold
higher TB mortality compared to the outskirts: 36.6
versus 15.1/10 000 in 1911-1915, and 20.3 versus 13.7/
10 000 in 1921-1925. According to a survey from 1896,
the city centre had a higher dwelling density compared to
the outskirts (2.2 versus 1.4 persons per room), and
considerably lower indoor airspace (18 versus 60 m3 /
person).
Conclusion: Poor living conditions characterized by
overcrowding and poor indoor ventilation fuelled the
TB epidemic in the Swiss capital city Bern in the last
century. Similar conditions can be found today in large
cities of low-income countries. Improved living conditions and public control measures likely led to the decline
in TB mortality in Bern before effective treatment options
became widely available.
SOA-640-06 Tuberculosis control strategies to

reach the 2035 global targets in China: the role
of changing demographics and reactivation
disease
G Huynh,1 D Klein,1 D Chin,2 B Wagner,1 P Eckhoff,1
R Li,3 L-X Wang3 1Institute for Disease Modeling, Bellevue,
Washington, 2The Bill & Melinda Gates Foundation, Seattle,
Washington, USA; 3Chinese Center for Disease Control and
Prevention, Beijing, China. e-mail: ghuynh@intven.com
Background: In the last 20 years, China ramped up a

DOTS (directly observed treatment, short-course)-based
tuberculosis (TB) control program with 80% population
coverage, achieving the 2015 Millennium Development
Goal of a 50% reduction in TB prevalence and mortality.
Recently, the World Health Organization developed the
End TB Strategy, with an overall goal of a 90% reduction
in TB incidence and a 95% reduction in TB deaths from
2015-2035. As the TB burden shifts to older individuals
and Chinas overall population ages, it is unclear if
maintaining the current DOTS strategy will be sufficient
for China to reach the global targets.
Design/Methods: We developed an individual-based
computational model of TB transmission, implementing
realistic age demographics and fitting to country-level
data of age-dependent prevalence over time. We explored
the trajectory of TB burden if the DOTS strategy is
maintained or if new interventions are introduced using
currently available and soon-to-be-available tools. These
interventions include increasing population coverage of
DOTS, reducing time to treatment, increasing treatment
success, and active case finding among elders . 65 years
old. We also considered preventative therapy in latently
infected elders, a strategy limited by resource constraints
and the risk of adverse events.
Results: Maintenance of the DOTS strategy reduces TB

incidence and mortality by 42% (95% credible interval,
27-59%) and 41% (5-64%), respectively, between 2015
and 2035. A combination of all feasible interventions
nears the 2035 mortality target, reducing TB incidence
and mortality by 59% (50-76%) and 83% (73-94%).
S581
Addition of preventative therapy for elders would enable

China to nearly reach both the incidence and mortality
targets, reducing incidence and mortality by 84% (7893%) and 92% (86-98%).
Conclusion: The current decline in incidence is driven by
two factors: maintaining a low level of new infections in
young individuals and the aging out of older latently
infected individuals who contribute incidence due to
reactivation disease. While further reducing the level of
new infections has a modest effect on burden, interventions that limit reactivation have a greater impact on TB
burden. Tools that make preventative therapy more
feasible on a large scale and in elders will help China
achieve the global targets.
Conclusion: We found a high TB burden among acute

emergency-displaced populations and countries affected
by conflicts had the highest estimated TB burden based
on national TB prevalence compared with countries
affected by other types of emergencies. This raises
concern about disrupted treatments, including for TB
control. Implementing TB prevention and control
programs as soon as possible in the acute phase of a
humanitarian emergency will minimize the risk of
transmission and maximize chances of completing
therapy.
SOA-641-06 The burden of tuberculosis among

acute emergency-affected displaced
populations
Y Xie,1 W Cronin,2 R Oatis,2 S Cohn,3 L Paulos,2 J Razeq,2

R Chen,1 S Dorman3 1Tuberculosis Research Section,
National Institute of Allergy and Infectious Diseases (NIAID),
National Institutes of Health (NIH), Bethesda, MD, 2Maryland
Department of Health and Mental Hygiene, Baltimore, MD,
3
MD, USA. Fax: (1) 301 480 5705. e-mail:
yingda.xie@nih.gov
A Tate,1 S Mcnabb,1 S Cookson2 1Rollins School of Public

Health, Atlanta, GA, 2U.S. Centers for Disease Control and
Prevention, Atlanta, GA, USA. e-mail: sgc0@cdc.gov
Background: A substantial proportion of the worlds

population is affected by conflict-generated (humanitarian) or natural emergencies. By the end of 2013, 51.2
million people had been forcibly displaced, many acutely.
Unfortunately, little is known about the tuberculosis (TB)
burden among displaced populations. Yet TB is prevalent, and these settings both increase the risk of new
infection and decrease the chances of completing TB
therapy. So quantifying the TB burden in these settings
will help guide the need for prevention and control
efforts.
Design/Methods: Countries in 2009, 2011, and 2013
affected by acute emergencies with 350 000 new
internally displaced persons (IDPs) were identified from
several datasets. Estimated TB burden (number and
proportion) among newly displaced IDPs, and newly
returning IDPs and refugees was determined using WHO
national TB prevalence rates. Weighted estimates were
aggregated by WHO Regions, emergency type, and
World Bank country economic classification.
Results: Forty-five countries were identified; the newly
displaced population ranged from 17 814 666 in 2009 to
29 906 491 in 2013. Most countries experienced both
humanitarian (conflict-generated) and natural emergencies; the WHO African Region had the most countries
with both (n 18). Natural disasters produced the
greatest number of IDPs (.35 million) and the highest
estimate of displaced populations with TB (55 846).
Although countries experiencing only conflict did not
have the highest number of IDPs (.19 million), the
weighted TB burden was highest in these countries, with
IDPs accounting for 1.83% of the estimated TB burden
compared with 0.44% for natural disasters and 1.61%
for both types (Table). According to the World Banks
economic classification, the TB burden among IDPs
ranged from 0.03% for high-income to 1.32% for lowincome countries.
SOA-642-06 Use of nucleic acid amplification

tests for determining risk of tuberculosis
transmission
Background: Sputum smear microscopy-negative pulmonary tuberculosis (PTB) patients typically are less
infectious than smear (SM)-positive PTB patients.
Nevertheless, molecular epidemiology studies have
shown that SM-negative PTB patients contribute significantly to transmission. Compared to smear microscopy,
nucleic acid amplification tests (NAAT) have a lower
(diagnostically more favorable) limit of bacillary detection. We hypothesized that the risk of transmission from
sputum NAAT-negative PTB patients is exceedingly low.
Objective: To estimate the risk of M. tuberculosis
transmission from sputum NAAT-negative vs. sputum
NAAT-positive PTB patients.
Design/Methods: Retrospective study was performed
using existing TB program data in MD, USA, from 2004
to 2009, during which M. tuberculosis fingerprinting and
contact investigations were performed routinely for all
culture-positive PTB patients and NAAT (Amplified
Mycobacterium tuberculosis Direct Test [Gen-Probe,
San Diego, CA]) was performed routinely for PTB
suspects. Patients with M. tuberculosis isolates having
the same fingerprint were assigned to clusters; transmission events were approximated by collection order of
individuals first culture-positive specimens. Transmission risk was assessed in 2 ways. First, we assessed the
risk of being identified as a genotypic cluster index case.
Second, we assessed the number of consecutive transmission events among secondary cases who immediately
followed and had the same sputum NAAT result as the
index case. As a point-of-reference, we also estimated
risk of transmission from sputum SM-positive vs. sputum
SM-negative PTB patients.
Results: Among 822 culture-positive PTB individuals,
697 (85%) had complete laboratory and genotyping data
and were analyzed. Compared with NAAT-positive PTB
patients, NAAT-negative PTB patients were 81% less
likely to be classified as index cases (RR0.19, 95%CI
S582
0.01-1.17) and 91% less likely to transmit TB (RR0.09,

95%CI 0.01 to 0.52). Compared with SM-positive PTB
patients, SM-negative PTB patients were 34% less likely
to be classified as index cases (RR0.66, 95%CI 0.361.18) and 61% less likely to transmit TB (RR0.39,
95%CI 0.23-0.64).
Conclusion: Risk of transmission from NAAT-negative
PTB patients is substantially less than that from NAATpositive PTB patients. NAATs may have a role in
infection control and decisions about respiratory isolation and contact investigations, and therefore may be
important public health tools.
showed 92% of laboratories improved their biosafety

measures, 80% improved in documentation in the TB
laboratory register, all facilities had the required level of
reagents and consumables and almost all (70%) facilities
performed IQC
Conclusions and key recommendations: Onsite coaching
is an effective and innovative method to improve
laboratories performance in EQA by ensuring every staff
is trained in their own work environment. We recommend onsite coaching for all laboratories performing
below the national EQA performance target of 85%.
05. Training for the fight against TB
SOA-644-06 How medical colleges in India are

contributing to ensuring universal access to TB
care in India: a case study from Himachal
Pradesh State
SOA-643-06 Onsite coaching: an effective way

to improve laboratory performance in external
quality assurance, Ghana
F Dzata,1 E Mensah,2 F Sorvor,3 F Bonsu1 1National TB
Control Program, Accra, 2Public Health Laboratory, Effia
Nkwanta, 3Konongo Hospital, Konongo, Ghana. e-mail:
francaus@ymail.com

Laboratory plays a critical role in TB diagnosis because
early detection will lead to early treatment and reducing
the risk of spread of the infection. External Quality
Assurance (EQA) for TB microscopy is an integral part of
the TB program and it is done to ensure laboratories
produce accurate and reliable results for patient management. A review of Laboratories performance in EQA
showed the Ashanti region which has the second highest
number of diagnostic centers (laboratories) had only
57.6% of these facilities attaining the national performance target of 85%.
Intervention or response: The national TB laboratory
team organized onsite coaching on TB microscopy at the
different diagnostic centers in the region in the first
quarter of 2014. The objective was to train laboratory
staff on TB microscopy and all components of EQA. The
onsite coaching was to ensure that all staff receive the
training in their work setting using their own equipment
and not only selected staff being trained at a central
point. This was also to enable the national team assess
the functionality of the microscopes and other equipment
and reagents on site. The team visited the laboratories in
the region and at each site staff were given hands on
training on documentation, logistics management, smear
preparation, biosafety measures, microscope use/maintenance and Internal Quality Control (IQC). Training
was in form of power point presentations and practical
sessions in the laboratory.
Results and lessons learnt: After one week onsite
coaching of laboratories in the region, EQA assessment
for all quarters in 2014 showed 92.5% of laboratories in
the region attained the national performance target of
85%, an improvement of 35.4% over the 2013
performance. Monitoring and support visits paid to
laboratories in the region after the onsite coaching also
A Bhardwaj1 1Dr. Rajendra Prasad Government Medical

College, Tanda, Kangra, India. e-mail:
ashoknonu@gmail.com
Background Revised National Tuberculosis Program

(RNTCP) in India in 1997 recognized the role of Medical
College in providing comprehensive TB care services and
now is the largest national health program globally to
have documented contribution of Medical Colleges in
program implementation. Himachal Pradesh (pop; 7
million) is a hilly state in India with difficult geographical
terrains and sparse population. There are two medical
colleges (IGMC Shimla and RPGMC Tanda) in the state.
During initial years of RNTCP launch in Himachal
between 1997 and 2002, the program faced several
challenges in service delivery, lack of infrastructure and
skilled human resource. Medical Colleges involvement
provided solutions to many of the problems; as a result,
program has been achieving its objective of case
detection and cure rate for last several years. This study
documents the functioning of medical colleges for
implementing RNTCP and contributions made by the
medical colleges in state of Himachal Pradesh.
Intervention or response: A State Task Force (STF) at
state level and core committees in both Medical Colleges
were formed to implement and review the program.
DMC cum DOTS centers were opened in both colleges
with an objective to provide the quality assured diagnosis
and standard treatment. Series of trainings were conducted both at national and state level. Diagnostic and
treatment services were established at both medical
colleges for MDR-TB cases.
Results and lesson learnt: Medical colleges in Himachal
Pradesh contributed in a major way in finding more TB
cases. In 2013 and 2014, medical colleges contributed 15
percent of total cases, 10% of smear negative and 40% of
extra-pulmonary TB cases in the state. However, there is
some gap in service delivery. All referred for treatment
cases are not being followed up in the districts; as a
result, there is a high initial default rate. Medical
Colleges faculties are earnestly following RNTCP
guidelines. State Task Force in 2014 supported two PG
dissertations, 14 major operational research projects and
ten Short Term operational research projects under

RNTCP.
Conclusion: There is a systematic involvement of medical
colleges in RNTCP implementation in Himachal Pradesh. However, in view of providing Universal access to
quality TB care and meeting new targets under National
Strategic Plan 2012-17, medical colleges has to play a
bigger role in capacity building, service delivery by
improving coordination and operational research.
S583
level, only 32 were found from 450 villages/wards

covering about 90,000 households.
Conclusion: The knowledge about TB among health
service providers in India whether government or private
providers is inadequate. Non-availability of qualified
government doctors at village/ward level is an issue
requiring intervention at the earliest. This situation if not
intervened immediately will have serious implications in
TB prevention and care, posing a threat to the End TB
strategy.
SOA-645-06 Health service providers

availability and knowledge about TB in 45
districts from India: implications for TB
prevention and care
K Sagili,1 A Das,1 B Entoor Ramachandran,1 B Thapa,1
P Banuru Muralidhara,1 S Pandurangan,1 S Mohanty,1
S Chadha1 1International Union Against Tuberculosis and
Lung Disease, South East Asia Office, New Delhi, India. Fax:
(91) 11 4605 4430. e-mail: ksagili@theunion.org
Background: Availability of trained Health Service

Provider (HSP) and their correct knowledge about health
conditions like Tuberculosis (TB) is essential for early
identification and appropriate disease management. This
reduces the delay in diagnosis and treatment initiation
and duration of transmission which is essential in
reducing the burden of TB.
Design/Methods: A KAP survey was conducted to
capture the knowledge, attitude and practice about TB
among HSPs in 45 districts from 17 states in 2013.
Stratified cluster sampling was carried for districts
selection. From each district, 10 primary sampling units
(PSU) were selected by population proportionate to size
method, with each unit having an approximate 200
households. Hence a total of 450 PSUs covering about 90
000 households were surveyed. The HSPs were identified
with the reference from general population and interviewed using a semi-structured questionnaire. For the
study purpose, the HSPs were defined as those to whom
general population would approach for their day-to-day
medical needs at village or ward level. These included
qualified government doctors, nurses, auxiliary nurse
midwives (ANM), anganwadi workers (AWW) and
primary health centre (PHC) staff from public health
system and chemists and qualified/non-qualified practitioners from private sector.
Results: A total of 767 health service providers were linelisted and interviewed during the survey. Of these
majority identified based on their availability were
AWWs (27%) followed by private practitioners (26%),
ANMs (16%) and chemists (16%) (Figure). The
availability of qualified government doctors was minimal
(4%). Among the providers from public health system,
69% doctors, 38% nurses, 50% ANMs and 35% PHC
workers and AWWs had correct knowledge about TB
(symptomdiagnosistreatment). And among private
sector 24% practitioners and 15% chemists had correct
knowledge. Important issue to be noted is the nonavailability of government doctors at village or ward
SOA-646-06 Effectiveness of a training course

on quality assurance of chest radiography in
Laos
A Ohkado,1 P Vangvichit,2 T Sorsavanh,2 O Rasphone,3
M Mercader,4 K Osuga,5 T Date6 1Research Institute of
Tuberculosis / Japan Anti-Tuberculosis Association, Kiyose,
Japan; 2National Tuberculosis Control Programme, Ministry
of Health, Vientianne, 3X-ray Unit of Health Science
University of Lao PDR, Vientianne, Laos; 4Radiology
Department, Gat Andres Bonifacio Memorial Hospital,
Manila, Philippines; 5Stop TB, Office of the World Health
Organization Representative in Viet Nam, Western Pacific
Regional Office, Hanoi, Viet Nam; 6College of Healthcare
Management, Miyama, Japan. Fax: (81) 424 93 5529.
e-mail: ohkadoa@jata.or.jp
Background: Chest radiography (CXR) plays a major

part of X-ray facility functions in the world. It is of
critical importance to improve and maintain the quality
of CXR to avoid faulty diagnosis of smear-negative/CXR
suggestive pulmonary tuberculosis (PTB) patients. The
present study aims to determine the effectiveness of
training in improving the quality of CXR among
radiological technologists (RTs) in Laos.
Design/Methods: This was a cross-sectional study
through on-site investigation of X-ray facilities, interviewing RTs, and reviewing CXR films before and after
the training in 2013 in Laos. In all, 19 government X-ray
facilities with RTs selected by Laos National Tuberculosis Control Program were eligible for inclusion. The
profiles of the X-ray facilities were collected using a semistructured questionnaire form. Three recently taken
CXR films of men and another three of women aged
over 15 years were collected both before and after the
S584
training course from all of the RTs. The quality of the

CXR films was assessed using Assessment sheet for
Imaging Quality of Chest Radiography developed by
the Tuberculosis Coalition for Technical Assistance. The
scores per every six assessment factor were summed.
Wilcoxon matched pairs test was applied to compare the
scores on CXR films taken before and after the RTs
attended the training course.
Results: In all, 19 RTs from 19 facilities at 16 provinces
across Laos participated in the training course. Among
them, 17 RTs submitted the required set of CXR films,
i.e. 204 in total. A wide range of X-ray machine settings
became apparent, e.g. tube voltage ranged from 40 to
120 kV and tube current from 50 to 500 mA. The
assessment of CXR films indicated that the training was
effective in improving the CXR quality with respect to
contrast (P 0.005), sharpness (P 0.004) (Figure), and
total sum score of the six assessment factors (P 0.009).
While the training did not seem effective in improving
other assessment factors such as identification mark,
patient positioning, density, artefacts, and total assessment score sum, but the sum scores of both the
identification mark (P 0.07) and total assessment score
(P 0.06) tended to show an improvement.
Conclusion: The significant improvement of the total
sum score of the six assessment factors in addition to
contrast and sharpness strongly suggests the positive
impact of the training course to improve the quality of
CXR in a sample-population of RTs in Laos.
SOA-647-06 Distance-learning pilot course for

front-line staff of the Peruvian National
Tuberculosis Program guidelines
1
J Lulli,1 A Mendoza-Ticona,2 L Otero,1 G Cordova,

K Castillo,3 C Montero,2 A Valentina Antonieta,2
E Gotuzzo1 1Instituto de Medicina Tropical Alexander von
Humboldt, Universidad Peruana Cayetano Heredia, Lima,
2
y Control de
Estrategia Sanitaria Nacional para la Prevencion
la Tuberculosis, Ministerio de Salud, Peru, 3Facultad de
Medicina Alberto Hurtado, Universidad Peruana Cayetano
Heredia, Lima, Peru. e-mail: jose.lulli@upch.pe
Background: The Peruvian National Tuberculosis (TB)

Program (NTP) issued a new version of the TB guidelines
in November 2013. A distance education program to
train front line TB workers was implemented in a joint

collaboration with the NTP and the Instituto de
Medicina Tropical Alexander von Humboldt and the
Telemedicine Unit of the School of Medicine, at
Universidad Peruana Cayetano Heredia. We report the
pilot implementation.
Methods: A well organized and good performing district
health division in Lima was selected for the pilot. A single
face-to-face training session introduced students to the
online platform. Upon completion of a baseline survey, a
pretest evaluation, participants started the 12 week, 6
hours-per-week course which included: 20-minute lectures recorded by country experts, scientific and programmatic literature, videos, brief post-tests and forum
interaction. On a weekly basis, two district coordinators
provided standard answers for pooled questions from the
forum. Personalized SMS delivered course and TBrelated news to participants as well as single reminders
to those about to miss a deadline. Technical support was
provided by phone and mail. Participants missing at least
7 post-tests were considered dropouts.
Results: We registered 108 students (30 medical doctors
(MD), 39 registered nurses (RN) and 39 nurse aids (NA).
Among the 92 who completed the survey: 86% (79) were
women, median age was 39 years (IQR: 32-47). Median
time from graduation was 8 years (IQR: 6-17) and they
had been in their post for 3 years (IQR: 1-6). Prior
participation in a distance education program was
reported by 20% (18), 84% (77) owned a home
computer and 27% (25), a smartphone. Internet use
was reported daily by 30% (28), several times per week
by 27% (25) and less than once per week by 37% (34).
The pretest evaluation median was 16 (IQR: 14-17) with
no significant differences by staff profession. Mid-term
scores were 11.8 for RN, 13.0 for MD and 10.9 for NT.
Dropouts began on week 4 and increased on weeks 5 to
10 irregularly. In total, 17 (15.7%) students dropped out
with no difference by staff profession. Among the
dropouts, 70% cited time constraints as the main reason
for dropout and 30% cited constraints in accessing a
computer.
Conclusions: Overall course retention was high, most
NTP staff were active in most course activities. Course
support to students may increase course participation. A
distance-learning course is a good alternative for NTP
staff training.
SOA-648-06 Fortalecimiento de la
e implementacion
de nuevas
comunicacion
tecnicas para la coleccion de muestras de

esputo para incrementar la carga bacilar
K Tintaya,1 M Mendoza,1 J Jajaycucho,1 C Pinedo,1
L Lecca,1 C C Contreras,2 J Coit,2 C Mitnick2 1Socios En
Lima, Peru ; 2Harvard Medical School,
Salud Sucursal Peru,
Boston, MA, USA. e-mail: ktintaya_ses@pih.org
Background: Se identifico la necesidad de establecer

mejoras en la educacion
del paciente y adicionar nuevas
tecnicas en la coleccion
de muestras de esputo. Esto
debido a que aun cuando existen guas proporcionadas
por el Ministerio de Salud (MINSA), existe sobrecarga de

trabajo y falta de tiempo en el personal de salud, para
brindar educacion
al paciente y obtener una muestra
adecuada que permitan obtener resultados y tratamientos oportunos. En un estudio previo, de 107 encuestados,
el 80% obtuvo una buena calidad de sus muestras y
fueron aquellos que recibieron una educacion
previa por
parte del personal (Armas, Y. 2010).
Design/Methods: Se busco a pacientes con baciloscopa
() o () para incluirlos a un ensayo clnico. Se busco
en 61 establecimientos de salud de Lima Metropolitana
durante 13 meses (febrero 2014 a febrero 2015). Se
incluyo a todos los pacientes que tuvieron un resultado
de baciloscopa positiva () a traves de un personal del
estudio. Se entrevisto a cada paciente y se brindo
materiales para coleccion
de muestra que deba de usar
durante la obtencion
de su segunda muestra (frasco de
coleccion,
cepillo dental y agua gasificada). Ademas se
educo al paciente acerca de los criterios que se debe
considerar para mejorar la calidad de la muestra. Al da
siguiente, el paciente entregaba su segunda muestra al
personal de salud, quienes enviaban para su procesamiento, al mismo laboratorio que brindo el primer
resultado.
Results: Se identificaron 196 pacientes que tuvieron una
baciloscopa () en su primera muestra. Luego de
implementar las nuevas tecnicas y brindar educacion
para mejora de la calidad de la muestra, se obtuvo que el

22% mejoro la carga bacilar a () o (); el 30%
mantuvo su misma carga bacilar; el 13% obtuvo una
segunda muestra negativa; y en 35% no se pudo obtener
informacion
de la segunda muestra. Nuestro estudio
intenta demostrar que la educacion
que se brinda al
paciente y la implementacion
de las tecnicas incrementa
la calidad de la muestra.
Conclusion: Intervenciones en educacion
e incorporacion
de nuevas tecnicas que mejoren la coleccion

de esputo,
ayuda a que el paciente incremente sus posibilidades de
un diagnostico
y tratamiento oportuno.Se recomienda
que los programas locales de TB consideren que

intervenciones sencillas, puedes contribuir a mejorar la
carga bacilar y a su vez, fortalecer la relacion
entre los
pacientes y el personal de salud.
SOA-649-06 Interactions between nurses and

doctors in caring for TB patients
T Fedotkina,1 V Sarkisova,2 O Komissarova,2 P Volkova,3
N Serebrennikova,2 M Driever,4 B Mandleco5 1Tomsk
Oblast TB Center, Tomsk, 2Russian Nurses Association, St
Petersburg, 3Marii El TB Department, Joshkar Ola, Russian
Federation; 4Self-Employed, Seattle, WA, 5Brigham Young
University, Provo, UT, USA. e-mail: fedot-tanya@yandex.ru
Background: It is obvious that caring for TB patients

requires constant interaction between doctors and
nurses, the quality of which plays an important role in
providing quality patient-centred TB care. Therefore,
this study aimed to identify key issues discussed by
doctors and nurses in the process of their interaction and
describe the kind of interaction happening between
S585
doctors and nurses while caring for TB patients and

their families.
Design/Methods: A qualitative descriptive study was
conducted to look at the interaction between doctors and
nurses. Respondents completed a questionnaire consisting of demographics, five open-ended questions and an
assessment tool to measure the benefits of joint doctornurse discussions about patient care on a scale of one to
10.
Results: Eighty-three TB doctors and nurses (43 doctors
and 40 nurses) from 16 Russian regions participated.
Twenty-one doctors (49%) and 20 nurses (50%)
responded that joint discussions of patient care issues
was of maximal benefit. However, respondents raised a
number of obstacles to such productive professional
dialogue. High workload and lack of time to jointly
discuss patient care issues was raised as a barrier. Both
doctors and nurses reported that the clear distinction
between the roles of doctors and nurses presented a
challenge for teamwork and joint patient care management. For example, doctors tend to regard a nurses job
as unimportant and nurses lack confidence in their own
professional value for doctors and patients. Nurses
reported that they sometimes doubt doctors professional
competence (appropriate treatment regimen, prescriptions not ordered on time, unwillingness to accept and
use the information about the patient, etc.) and doctors
may doubt nurses qualifications which hindered useful
interaction. Nurses felt doctors underestimate the nurses
influence on patients treatment outcomes.
Conclusion: About half of the doctors and nurses
included in the study were satisfied with their professional relationship and describe their interaction as
mutually beneficial. To improve the quality of interprofessional cooperation joint conferences should be
organized for nurses and doctors, providing a platform to
discuss individualized patient-centred care, teamwork
and mutual trust in colleagues professionalism.
SOA-650-06 Evaluation of different models of

training to improve health care worker
knowledge of childhood TB at primary health
care level in South Africa
L Du Plessis,1 K Du Preez,1 M Palmer,1 V Azevedo,2
E Mhlope,3 A Hesseling1 1University of Stellenbosch, Cape
Town, 2City Health, Cape Town, 3USAID TB Program,
Johannesburg, South Africa. e-mail:
lienkiduplessis@gmail.com
Background: There is limited evidence regarding appropriate childhood TB training strategies in high-burden
low-resource settings. There is currenlty a shift in
education models, moving away from didactic approaches, towards decentralised interactive self-study models.
Guidance regarding feasible and effective training
strategies is needed.
Design/Methods: As part of a paediatric TB health
system strengthening project (Kid-care) implemented in
Khayelitsha, Cape Town (2013/4), we conducted 15 four
week training sessions on childhood TB at primary care
S586
level. Groups were randomized to 1 of 3 models of

support (lectures, peer-groups or self-learning). Impact of
the training on knowledge was measured with a 55question test at three time points: pre-training, posttraining and 10 weeks thereafter.
Results: 131 nurses (median age 41 years (SD 10.2),
122(95%) female) were enrolled from 9 clinics; 100
(76%) completed post-test and 79 (60%) retention
testing, while 27(21%) had provided TB care during
the preceding 2 years. The mean baseline knowledge was
66% (SD4.6), and was not age-associated. Overall prepost knowledge gained was 11%, with the lecture model
gaining 14.8% and peer-groups 8.4% compared to selflearning (Table). No significant difference between
knowledge retention was observed between training
models (Table). Externally facilitated lectures were better
attended than participant facilitated sessions (attending
.2/4 sessions: 94% lecture learners, 26% self-learners
and 30% peer-group learners). Despite higher baseline
knowledge amongst TB nurses [72% (95%CI 68-75] vs.
64% (95%CI 62-67), P 0.001), there was a comparable increase in knowledge [8% (95%CI 1-16) vs.1 2%
(95%CI 7-17), P 0.3210.
Conclusion: Focussed training increased knowledge of
childhood TB amongst all participants irrespective of
work area, highlighting the need for training initiatives to
target HCWs across services. This is of particular
importance as children may access care at multiple
health services entry points and staff turn-over is often
high. Although the lecture model achieved the highest
mean increase, the results may reflect dedicated time
spent on course content rather than the effects of a
didactic model. Training initiatives need to consider
multiple setting specific factors to ensure participant
engagement. Longer term retention assessment should be
considered and the impact of training on clinical practice
and patient outcomes evaluated.
SOA-651-06 Targeted, repetitive intervention

to improve MDR-TB knowledge retention in
Rayong Province, Thailand
D Punpiputt,1 S Chareonsiri,1 J Indrasap,2 C Thibbadee,3
A Innes1 1FHI 360 Asia Pacific Regional Office, Bangkok,
2
Rayong Provincial Health Office, Rayong, 3Rayong Provincial
Hospital, Rayong, Thailand. e-mail: ainnes@fhi360.org
Background: Effective pre-service education and inservice training are critically important for the health
care work force to maintain and update knowledge.
Multi-drug resistant tuberculosis (MDR-TB) is a complex disease, and effective patient management requires
knowledge and application of clinical and public health
guidelines. Repetitive, interval reinforcement creates a
spacing effect that has been shown to improve
knowledge retention. Presenting information and reinforcing it over intervals of time is the principle behind
QStream, an internet-based application that has been
shown in clinical trials to improve knowledge retention,
change behavior to improve adherence to guidelines, and
increase learner engagement.
Intervention: In Rayong, Thailand, the USAID Control
and Prevention-Tuberculosis (CAP-TB) project pilottested QStream to teach clinical concepts for TB,
MDR-TB, and TB-HIV. Regular teaching conferences
with the TB team identified key learning points that
formed the basis for quiz questions (multiple choice or
true/false). Following each teaching conference, these
quiz questions were sent to the team in Rayong, a multidisciplinary group of physicians, registered nurses,
pharmacists, laboratory technicians, and coordinators.
The QStream platform determined the frequency between each question based on the number of questions
(4-5) in each quiz and the duration of the quiz (1-2
months). Each question had to be answered twice
correctly before it could be retired. Questions were
designed to take less than 5 minutes to complete,
maximizing efficient learning while on the job.
Results: Over one year, a total of 6 quizzes were sent to

the Rayong TB team. At the end of the year, a final test
was administered that comprised all questions from the
year. We found that TB knowledge was positively
correlated with participation in the teaching conferences
and QStream quizzes throughout the year (R 0.59,

Figure). This result shows that people who did well on
the quizzes throughout the year also did well on the final
test.
Conclusion: Although the sample size is small, the results
suggest that a targeted, repetitive method of teaching is
effective to acquire and retain TB knowledge. Given the
high turnover among TB staff in Thailand, it is important
to identify an effective method to certify and maintain
basic TB knowledge among all staff, both new and old.
Using QStream to reinforce and solidify key concepts
may be one solution.
S587

Abstract Book 2015-Web

Uploaded by

Copyright:

Available Formats

Abstract Book 2015-Web

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Abstract Book 2015-Web

Uploaded by

Copyright:

Available Formats

VOLUME 19

ISSN 1027 3719

46th World Conference

CAPE TOWN SOUTH AFRICA

24. Why and how to use tobacco taxes as funding

SATURDAY 5 DECEMBER 2015

S16 52. Developing a rights-based approach to prevention

Poster discussion sessions

S295 17. TB & NCD co-morbidity

Poster discussion sessions

SUNDAY 6 DECEMBER 2015

Oral abstract sessions

S438 22. The nature of the beast: emerging molecular

Guy Marks, Woolcock Institute of Medical Research, Sydney, NSW, Australia

ROGELIO HERNANDEZ PANDO

ELLEN MITCHELL (The Netherlands)

TIM STERLING (USA)

Expert statistical review panel

INT J TUBERC LUNG DIS 19(12s2):S1S587

46th World Conference on Lung

02. Best practice in management of side

Background and Challenges: Drug-resistant TB has

country in 2012, the gap between diagnosed and enrolled

03. Pre-approval of access to new TB

Symposium abstracts, Friday, 4 December

Drug, and Cosmetic Act sets out the legal framework

04. Outstanding issues in HIV/AIDS

Background: Diagnosis and confirmation of tuberculosis

implications for TB programmes. TB programmes in

Pros and cons: empiric TB treatment is critical to

Strategies for reducing tuberculosis (TB) mortality

Symposium abstracts, Friday, 4 December

TB before or without microbiologic evidence can result

05. Zoonotic TB session I - Trends,

A quarter of the global burden of human tuberculosis

currently used BCG vaccine has not been successful in

Rapid bTB: a novel lateral flow test able to

Symposium abstracts, Friday, 4 December

ylococcus). In the Cameroon laboratory direct detection

focus on detecting active disease and then culling infected

How zoonotic tuberculosis is laying new

Ferrets: a small animal model for zoonotic

Tuberculosis epidemics develop when one case is

Background: During the early 1990s, there were limited

Symposium abstracts, Friday, 4 December

scientific contributions have been published. Studies on

abattoirs had positive culture. Furthermore, 10.4% (48/

Prevalence, molecular epidemiology and risk

06. One pilot after another: how can we

S Cadmus,1 V Akinseye,1 O Adedipe,1 AKwaghe,2

Viet Nam: TB CARE I project leading to the

Background: Tuberculosis (TB) and bovine TB (bTB)

Background: Viet Nam is a high TB burden country but

T H Nguyen 1KNCV Tuberculosis Foundation, Hanoi, Viet

Symposium abstracts, Friday, 4 December

the wider health care sector by the NTP strengthening

07. Role of surgery in management of TB:

for complications and sequel, 1316 patients were

Experience from the Russian Federation: