Nephrology 17 (2012) 681–688

Original Article

Occurrence of infection among children with nephrotic syndrome

during hospitalizations nep_1650 681..688

CHANG-CHING WEI, I-WEN YU, HSIANG-WEN LIN2,3 and ALAN C TSAI4,5

1 2

1
Division of Pediatric Nephrology, Department of Pediatrics, 2Department of Pharmacy, China Medical University Hospital, 3Graduate Institute and School of
Pharmacy, College of Pharmacy, 4Department of Health Services, School of Public Health, China Medical University, and 5Department of Healthcare
Administration, Asia University, Wufeng, Taichung, Taiwan

KEY WORDS: ABSTRACT:

hospitalization, infection, nephrotic syndrome.
Aim: The present study was conducted to investigate the trends of child-
Correspondence: hood nephrotic syndrome (NS) admissions and factors associated with
Dr Hsiang-Wen Lin, School of Pharmacy and childhood NS admissions with major infections in Taiwan.
Graduate Institute, College of Pharmacy, China Methods: A retrospective analysis was performed using Taiwan National
Medical University, Pharmacy Department,
Health Research Insurance Database (NHIRD) to explore the associated
China Medical University Hospital, No. 91
Hsueh-Shih road, Taichung, 40402 Taiwan.
factors and health care burden for childhood NS admissions with major
Email: hsiangwl@mail.cmu.edu.tw, infections in 1997 to 2007.
hsiangwl@yahoo.com Results: Of 133 927 children, a total of 176 children had NS, which incurred
Dr Alan C Tsai, Department of Healthcare 508 hospital admissions. Nineteen percent of admissions were associated
Administration, Asia University, Wufeng, with major infections. Pneumonia was the most common infection (49%),
Taichung 413, Taiwan and Department of followed by urinary tract infection (UTI), bacteraemia/sepsis, peritonitis and
Health Services, School of Public Health, China
cellulitis. Pneumonia was the most common infection among children age
Medical University, Taichung 404, Taiwan.
Email: atsai@umich.edu
younger than 10 years, whereas UTI was more common among children aged
greater than 10 years. NS admission with infections had longer periods of
Accepted for publication 28 July 2012. hospital length of stay and higher hospital total costs compared to those
Accepted manuscript online 7 August 2012. without infections. Regression analysis reveals that younger age, regional
hospitals, admission hospital located in middle and south areas and admis-
doi:10.1111/j.1440-1797.2012.01650.x sion made in spring were associated with increased risk for developing
Disclosures: None. major infections.
Conclusions: While 19% of childhood NS admissions were associated with
major infections, young age, admissions made in spring, located in middle
SUMMARY AT A GLANCE and south Taiwan and in regional hospitals were the major associated
The authors report the incidence of factors for infection. Age plays an important role in risk and types of
infection and its associated factor in infection.
children with nephrotic syndrome (NS)
from the viewpoint of public health in
Taiwan, using the Taiwan National Health
Insurance Database.

INTRODUCTION other immunosuppressants.2 Immune-compromised status

of NS can result in increased susceptibility to various infec-
Nephrotic syndrome (NS) is among the most common tions, which are potentially serious and life-threatening
chronic kidney diseases occurring in childhood. The inci- without prompt diagnosis and treatment.3 Major infections
dence of idiopathic NS is reported at two to three cases per of NS include spontaneous bacterial peritonitis, bacter-
100 000 children.1,2 Defence mechanisms are impaired in aemia, sepsis, cellulitis, pneumonia and urinary tract
patients with NS due to its consequences of tissue oedema, infection (UTI).4–8 Episodes of NS complicated with
urinary loss of immunoglobulin and complement and major infections usually require admission for aggressive
secondary effects to its treatment, including steroids and treatment. NS complicated with infection can affect not

© 2012 The Authors

Nephrology © 2012 Asian Pacific Society of Nephrology 681
C-C Wei et al.

only patient outcomes but also increase the health care Table 1 World Health Organization International Classification of Diseases,
burden.3 Ninth Revision, Clinical Modification (ICD-9-CM) for index disease
Although advances in knowledge of care and medication Disease category ICD-9-CM code
have improved outcomes of NS and reduced comorbidity and
Nephrotic syndrome 581.0–581.3, 581.8, 581.9
mortality of NS in recent years, infection remains one of the Nephrotic syndrome histologic
most common complications and a significant cause of mor- subtype
tality in childhood NS.3,9–13 Infection rate varies among Diffuse mesangial proliferation 581.0
patients, nephrotic state, different renal histology, therapeu- Focal and segmental 581.1
tic regimens, geographic region and various health care glomerulosclerosis
facilities.8,14–18 There is limited population-based study in the Mesangioproliferative 581.2
glomerulonephritis
literature addressing the trends of hospital admissions and
Minimal change disease 581.3
potential associated factors about the infections among chil- Lupus nephritis 581.8 and 710.0
dren with NS. The present study was conducted to investi- Unspecified pathological lesion 581.9
gate the trends of childhood NS admissions and factors Major infection
associated with childhood NS admissions with major infec- Sepsis/bacteraemia 995.91–995.92, 777.81, 790.7,
tions in Taiwan. 038.0–038.4 and 038.8–038.9
Cellulitis 528.3, 681.0–681.1, 681.9 and
682.0–682.9
Pneumonia 480.0–480.3, 480.8–480.9, 481,
METHODS
482.0–482.4, 482.8–482.9,
483.0–483.1, 483.8,484.1, 484.3,
Data sources 484.5–484.8, 485, 486 and 487.0
Peritonitis 567.0, 567.1, 567.2, 567.8, 567.9,
The implementation of Taiwan’s National Health Insurance (NHI), as
728.89
a single-payer, social insurance plan, has covered almost all citizens
Urinary tract infection 590, 599.0
with modest cost sharing.19 The Bureau of NHI (BNHI) has contracted
Thromboembolism
with 97% of hospitals since 1996 to ensure sufficient access in Taiwan.
Acute pulmonary heart disease 415.XX
At present, the coverage of population and hospitals are both as high
Occlusion of cerebral arteries 434.XX
as 99%. NHI also provides the datasets for corresponding research on
Arterial embolism and thrombosis 444.XX
issues related to cost, quality of health services, medical practice Portal vein thrombosis 452.XX
patterns, accessibility to health care programs and treatment out- Other venous embolism and 453.XX
comes at national or local levels. This study used the one-million thrombosis
sampling claimed data of the National Health Research Insurance Other respiratory infection (e.g. upper 4660, 462, 4779, 490, 3829, 4659,
Database (NHIRD), which represents the entire insured Taiwanese respiratory infection) and 493XX, 518XX
population (i.e. 23 million). More specifically, we used the Longitu- respiratory diseases (e.g.
dinal Health Insurance Database 2005 (LHID 2005) for the analyses. asthma)
Further, this study was exempt from the Institutional Review Board Acute gastroenteritis 558.9
because the NHIRD database contains the de-identified person iden- Systemic lupus erythematosus 710
tifiers and is publicly available through the proper application process. Acute renal failure 584.XX, 586
Chronic kidney diseases 585.XX

Study design
Those childhood nephrotic syndrome (NS) patients (i.e. <19 year-old tion period, they were categorized as infection admissions and oth-
and with discharge diagnosis of ICD-9: 581.0–581.3, 581.8, 581.9) erwise. Specifically, the occurrence of infection diseases, focused on
admitted to hospitals during 1997–2007 were identified as our study major infections (i.e. pneumonia, bacteraemia/sepsis, cellulitis, peri-
cohort. The trends of childhood NS admission and sick children tonitis and UTI defined on the basis of ICD-9-CM) (Table 1), during
admitted to hospitals (i.e. with or without major and specific infec- hospitalization were the primary interest of outcome. The length of
tions) and the number of admissions per person among those child- stay (LOS) and total hospital cost consumed by those patients with
hood NS patients who were ever infected or not during various types of infection were compared with those without infec-
hospitalization across years were examined. For all childhood NS tion. The trends of various types of infection were observed toward
patient admissions within 11 years, the patients’ hospital admissions childhood NS patients at different age groups (i.e. <1, 1–4, 5–9,
were compared by their corresponding age, gender, disease statuses 10–14, 15–18). The occurrence of other complications and
(e.g. causes of hospitalizations (thromboembolism, other respiratory co-morbidities for NS, i.e. acute renal failure (ARF) and renal his-
infection, acute gastroenteritis, Table 1)), admission characteristics tology for NS (i.e. lupus nephritis (LN), focal and segmental glom-
(i.e. number of admissions per year, season, hospital level, hospital erulosclerosis (FSGS), mesangioproliferative glomerulonephritis
location) between the admissions with and without infection. Once (MPGN), minimal change disease (MCD) and unspecified histology),
childhood NS patients ever admitted to hospital due to infection (i.e. thromboembolism, other respiratory diseases, which were defined
major infections in the corresponding hospital admission records on the basis of ICD-9-CM (Table 1), were examined. In Taiwan,
(one primary and up to four secondary diagnoses)) during observa- there are 22.6 million people living in the land area 36 188 km2, so

© 2012 The Authors

682 Nephrology © 2012 Asian Pacific Society of Nephrology
 Childhood NS with major infection

the population density is 625/km2. Most (95.6%) of the population NS-related admissions, during 1997 to 2007 in Taiwan. The
live in the western part of Taiwan. Only 5.4% live in eastern Taiwan, occurrence of childhood NS hospitalization in Taiwan was
where medical care and socioeconomic status are underprivileged. 1.31 per 1000 children. Of childhood NS admissions, the
Thus, their enrolments in the corresponding divisions were divided occurrence of major infections and pneumonia had similar
into the northern region (including Taipei Division and Northern
patterns across years, whereas the highest was observed in
Division), central region, southern region (including Southern Divi-
2005 and the lowest was observed in 1998 and then 2000.
sion and Kaoping Division) and eastern region upon the definitions
proposed by BNHI and its NHIRD coding book. Further, NHI con-
The number of admissions per person among those non-
tracted hospitals were classified based upon the hospital size, care infection childhood NS patients was higher than those
and teaching capacity. The hospital areas and their levels were com- infected NS patients, expect in 2005 (Fig. 1). While the 95%
pared as well. To explore the associations between potential factors confidence intervals for the number of admissions per person
and experience of infection admissions during hospitalization, the among two groups were almost overlapping across years, it
admissions were classified into two types (‘ever infection’ vs ‘never represents no statistically significant difference among the
infection’) among childhood NS admissions. ‘Ever infection’ two groups of childhood NS patients during the observation
described those admissions in which there was one or more than one period.
NS admission associated with infection and; ‘never infection’
described those NS admissions without any of major infection diag-
noses during the observation year. Demographic and medical characteristics of
NS admissions
Statistical analyses
Ninety-seven of the 508 admissions (19.1%) were associated
The corresponding findings were presented as mean 1 standard
with defined major infections. The NS-related admissions,
deviation, or frequency (relative frequency, %). The differences
including NS with unspecified pathological lesions in
among continuous variables were analyzed using independent
t-tests and the differences among discrete variables were analyzed
kidney, hypoalbuminaemia, proteinuria and hypovolaemia,
using Pearson’s c2 test. The overlapping of 95% confidence intervals accounted for the major causes of hospitalization (Table 2).
for the number of admissions per person between different groups Only 13.2% of admissions were coded with histological
across years was examined. Associated factors for infections among diagnosis (8.2% vs 14.4% among NS with and without infec-
childhood NS admissions were examined using simple logistic tions).The other admissions were then mostly driven by
regression, where it was assumed that each admission from an major infections (i.e. pneumonia, UTI, peritonitis), other res-
individual child was independent. Statistical significance was piratory diseases (e.g. asthma) and then acute gastroenteritis.
accepted at a = 0.05. SAS version 9.1 (SAS Institute, Cary, NC, USA) Acute renal failure, chronic renal diseases (i.e. chronic glom-
was used to manage and analyze the data. erulonephritis (CGN) with lesion of membranous glomeru-
lonephritis) and arterial embolism/thrombosis accounted for
RESULTS
very small amount of childhood NS admissions (Table 3).
There were statistically significant, different patterns of hos-
Hospital admission trends of childhood NS and
pitalization causes among those infection and non-infection
sick children
admissions (except for the other respiratory diseases). As a
Of 133 927 children, a total of 176 children were admitted result, there were 61 (34.7%) childhood NS patients ever
to hospitals with NS diagnosis, which contributed to 508 infected during hospitalization. The mean age of children

3.8
3.6
3.4
3.2
3
Number of admission/person

2.8
2.6
2.4
2.2
2
1.8
1.6
1.4
1.2
1
0.8
Fig. 1 Admission percentage for different 0.6
types of infection stratified by age groups 0.4
among patients with childhood nephrotic 0.2
syndrome (NS). Infection ( ); non-infection 0
( ). 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

© 2012 The Authors

Nephrology © 2012 Asian Pacific Society of Nephrology 683
C-C Wei et al.

Table 2 Demographic characteristics of childhood nephrotic syndrome admissions with or without infections

Features Overall Infection Non-infection P*

Sample size of admission (n) 508 97 411

Age (years) in admission (mean 1 SD) 8.0 1 4.8 6.5 1 4.3 8.4 1 4.9 <0.001
1–4 34.2% 45.4% 31.4% 0.003
5–9 32.9% 33.0% 32.9%
10–18 32.3% 18.6% 35.5%
Male 73.10% 74.20% 72.80% 0.778
Histology
LN 1.60% 2.00% 1.50% 0.0393
FSGS 4.30% 3.10% 4.60%
MPGN 2.40% 3.10% 2.20%
MCD 4.90% 0.00% 6.10%
Unspecified 86.80% 91.80% 85.60%
MCD + Unspecified 91.70% 91.80% 91.70% 0.979
Admission characteristics
Admission season
Spring 26.60% 38.10% 23.80% 0.021
Summer 28.50% 20.60% 30.20%
Fall 23.20% 18.60% 24.30%
Winter 21.90% 22.70% 21.70%
Admission hospital level
Medical centre 71.10% 57.70% 74.20% 0.006
Regional hospital 25.20% 37.10% 22.40%
District hospital 3.70% 5.20% 3.40%
Location of admission hospital
North 34.60% 26.80% 36.50% 0.296
Middle 28.30% 29.90% 28.00%
South 34.80% 40.20% 33.60%
East 1.80% 3.10% 1.90%

Infection included bacteremia/sepsis, pneumonia, cellulitis, peritonitis and urinary tract infection). *Using independent t-tests for continuous variables and
Pearson’s c2 tests for discrete variables to compared the differences. FSGS, focal and segmental glomerulosclerosis; LN, lupus nephritis; MCD, minimal change
disease; MPGN, membranoproliferative glomerulonephritis.

with NS admission was 8.0 1 4.8 years, where the mean age was in the middle of the observation period. The LOS and
of NS with infections was younger (6.5 1 4.3) than NS total hospital cost among those who were NS with
without infections (8.4 1 4.9) (Table 2). Seventy-three bacteraemia/sepsis were statistically significant longer and
per cent of the NS patients were boys and there was no higher, respectively (P = 0.0008, P = 0.0014, respectively)
significant gender difference between NS with and without than non-infected NS admissions (Table 4).
infections. The distributions of age group and histology types
were statistically different between those with and without
Infection patterns in different age groups among
infections (P = 0.003, 0.0393, respectively) (Table 2). For
NS admissions
admission seasons, 38% of admissions were made in spring
among infection NS children but 30% in summer among Among infection admissions, pneumonia was the most
those without infection. Most admissions were made toward common infection (49%), followed by UTI (30%),
medical centres (71%), whereas 37% of infection NS and bacteraemia/sepsis (11%), peritonitis (11%) and cellulitis
22% of non-infection NS admissions were made toward (5%). Pneumonia was the most common infection among
regional hospitals. For admission hospital locations, the NS children younger than 10 years of age, whereas UTI was
majority of NS admissions were made in western Taiwan and more common among NS children aged greater than 10
only a few were in Eastern Taiwan (98.2% vs 1.8%). years (Fig. 2).
The average total hospitalization cost for all childhood NS
was $US 993 1 1401, in which the infected NS admissions
Associations of childhood NS admissions and
consumed more than non-infected NS admissions ($US
major infections
1278 1 1693 vs $US 87011034, P = 0.0024) (Table 3). Specifi-
cally, we used the exchange rate of $US to New Taiwan Compared to the corresponding reference groups of child-
Dollars (NTD) as 30.219 on 30 June 2002 because that date hood NS patients (age 10–18, female, admitted hospital

© 2012 The Authors

684 Nephrology © 2012 Asian Pacific Society of Nephrology
 Childhood NS with major infection

Table 3 Medical characteristics of childhood nephrotic syndrome admissions with or without infections

Features Overall Infection Non-infection P*

Sample size of admission (n) 508 97 411

Mean admission/person for 10 years 2.9 1 4.3 1.6 1 1.5 2.1 1 2.6 0.1682
Admission/person for 10 years
1 117 (66.5%) 46 (75.4%) 71 (61.7%) 0.439
2 24 (13.6%) 8 (13.1%) 16 (13.9%)
3 15 (8.5%) 3 (4.9%) 12 (10.4%)
4 7 (4.0%) 1 (1.6%) 6 (5.2%)
5 6 (3.4%) 1 (1.6%) 5 (4.3%)
ⱖ6 7 (4.0%) 2 (3.2%) 5 (4.4%)
Total discharge diagnoses 1164 245 506
Major causes of hospitalizations (%)†
NS 199 (28.47) 63 (25.71) 153 (30.24) 0.199
Pneumonia 36 (5.15) 26 (10.61) 0 (0.00) <0.001
UTI 29 (4.14) 20 (8.16) 0 (0.00) <0.001
Peritonitis 5 (0.72) 5 (2.04) 0 (0.00) 0.001
Arterial embolism/thrombosis 1 (0.14) 1 (0.41) 0 (0.00) 0.15
Other respiratory diseases‡ 44 (6.29) 23 (9.39) 75 (14.82) 0.038
Acute gastroenteritis 19 (2.72) 7 (2.86) 14 (2.77) 0.944
Systemic lupus erythematosus 8 (1.14) 1 (0.41) 7 (1.38) 0.222
Acute renal failure 7 (1.00) 2 (0.82) 6 (1.19) 0.644
Chronic kidney disease 4 (0.57) 1 (0.41) 3 (0.59) 0.744

Infection included bacteremia/sepsis, pneumonia, cellulitis, peritonitis and urinary tract infection). *Using independent t-tests for continuous variables and
Pearson’s c2 tests for discrete variables to compared the differences. †Refers % of all discharge diagnoses among corresponding group. ‡Other respiratory
diseases included acute pharyngitis, bronchitis, unspecified otitis media, upper respiratory infection, acute bronchitis, allergic rhinitis, asthma, other diseases of
lung, not elsewhere classified. NS, nephrotic syndrome; UTI, urinary tract infection.

Table 4 Length of stay and hospitalization total cost among patients with childhood nephrotic syndrome (NS), classified based on various types of infection

n Length of stay (days) Hospital total costs ($US)

Mean 1 SD P Mean 1 SD P

NS without infection 411 5.9 1 5.9 Reference 870 1 1034 Reference

NS with infection 97 7.3 1 6.2 0.0007 1278 1 1693 0.0024
NS with bacteraemia/sepsis 11 11.4 1 6.4 0.0008 1898 1 1481 0.0014
NS with cellulitis 5 6.6 1 1.1 0.7682 1150 1 658 0.5461
NS with peritonitis 11 4.9 1 3.0 0.5339 1488 1 1932 0.0581
NS with pneumonia 50 5.8 1 2.7 0.8956 929 1 744 0.6972
NS with UTI 31 5.4 1 2.9 0.6040 648 1 436 0.2357

*Using independent t-tests for continuous variables to compare the differences. NS, nephrotic syndrome; SD, standard deviation; UTI, urinary tract infection.

located in north, in medical centre, in summer of admission Among childhood NS patients, younger age was more likely
season), the odds ratios (ORs) of infections among the other to contract infection during hospitalizations. Pneumonia was
two age groups, admission hospital located in middle and the most common infection when children were equal to
south area, regional hospitals and admissions made in spring and younger than 10 years but more UTI among children
were statistically significant higher (Table 5). In particular, were older than 10 years. Given that infection is one of the
the ORs of infection among NS children aged 1–4 years old most common complications and still remains a significant
was 5.312 (95%CI; 2.8–10.1). cause of morbidity and occasionally mortality in childhood
NS, our results suggest that clinicians could be more aggres-
sive on infection prevention in children ⱕ4 years and on
DISCUSSION
infection intervention to prevent younger childhood NS
In the current study, we did not find the resembling patterns patients from pneumonia and older childhood NS patients
of infection occurrence among those sick children admissions from UTI.
and those children admissions with NS. We have noted that Instead, the current study observed that the number of
age played an important role on infection in childhood NS. admissions per person among the non-infection group was

© 2012 The Authors

Nephrology © 2012 Asian Pacific Society of Nephrology 685
C-C Wei et al.

80
72
NS was secondary NS and LN was the most common cause of
70 secondary NS in children. We also noted that renal histology
60 patterns among infection and non-infection NS children
Percent (%)

50
50
were different. Further studies might be needed to explore
43
36
such difference and its contributing factors.
40 34
27 29 In fact, circannual variations of NS, which presented as
30
22
18 18
initial episodes with an autumn peak and relapses with a
20 14 14
14
11
9
spring peak, have been reported in some other literature.21–24
10 6 6
5
0 0
Allergens triggers and infection triggers were found to be
0 associated with initial episodes and relapses of NS. Our
1–4 5–9 10–14 15–18
results also revealed a seasonal variation, that is, a spring
Fig. 2 Trends of nephrotic syndrome (NS) admission and sick child hospital- peak occurred. This seasonal variation might be explained in
ization among national representative sample in Taiwan during 1997 to 2007. that the infection could be one of the triggers or the conse-
*refers to sick children admissions without NS diagnoses during hospitaliza- quence of NS. Therefore, the NS admissions associated with
tions. All sick children admissions = Sick children admissions plus All childhood major infections increased the health care burden. The NS
NS admissions. Bacteraemia ( ); cellulitis ( ); peritonitis ( ); pneumonia ( ); admissions with bacteraemia and sepsis contributed almost
urinary tract infection (UTI) ( ).
twofold increments in LOS and total hospitalization costs in
this study. That is because bacteraemia and sepsis are the
most severe, life-threatening systemic infections, which need
higher than those among the ever infected group, except in more health care resources and aggressive treatment during
2005, although there was no statistical significance. hospitalization. As such, it would be beneficial to come up
There might be several reasons for this. First, we only con- with efficient strategies to prevent NS children from infection
sidered sepsis, pneumonia, UTI, cellulitis and peritonitis, in and facilitate their appropriate self-care and management.
terms of the major infections. Some other rare or mild Nevertheless, this study has some limitations. First, the
infections were not included in this infection group. As a findings were exploratory in nature to focus on the popu-
result, the infection rate (19%) in our population was not lation of interest ‘children with nephrotic syndrome during
as high as previous studies.20 Second, most of our NS chil- hospitalizations.’ Therefore, the interpretation of findings
dren experienced single time hospital admission (75.4% in (except in Fig. 3) should be cautious because it cannot
the infection group vs 61.7% in the non-infection group) directly apply to all sick children ever hospitalized during
during the observation period. Therefore, more other the study period. Although we are actually conducting a
potential factors should be addressed and examined in study to examine the infection patterns and its associated
future studies. factors among children with or without NS, further studies
The likelihood of infection and types of infection in child- would be needed to verify our findings using different data-
hood NS also varied in different geographic regions around bases or using different study designs. Second, there is a
the world. A retrospective study conducted in India observed lack of detailed information about medication use history
that 38% of NS children had at least one infection and UTI and clinical outcomes (e.g. laboratory data and clinical
was the commonest infection (13.7%) followed by pulmo- course), based upon the inpatient data of NHIRD. Although
nary tuberculosis (10.4%), peritonitis (9.1%), skin infections the accuracy of the NHIRD has been validated in some
(5.2%), upper respiratory infections (5.2%) and lower res- other diseases,25 there is no similar study available on the
piratory tract infections (3.9%).18 Another study from Paris diseases that occurred among the children (including neph-
found that 8% of children admitted for NS children had rotic syndrome). However, the NHIRD provides a robust
bacterial infections and half of the infections were peritonitis population-based data for patient demographics, disease,
and 50% of the identified germs were Streptococcus pneumo- drug and procedures and health care expenses. In fact, the
nia.6 In Karachi, acute respiratory infection, cellulitis and NHIRD databases were very useful for health care providers
enteritis were most common infections followed by UTI and and policy makers to examine issues on the epidemiology
peritonitis.17 In contrast, we found in Taiwan, 19% of NS of the disease and its associated complications, cost-
admissions were associated with major infections. And pneu- effectiveness, health care quality and utilization among
monia was the most common infection, followed by UTI, hospitals and geographic regions to improve quality of care
bacteraemia/sepsis, peritonitis and cellulitis. and reduce disease-economic burden. Third, this study did
Indeed, idiopathic NS in children is a clinical syndrome not include the outpatient information and non-hospital
associated with a variety of glomerular lesions. MCD is the costs, in terms of indirect cost. It is important to include the
most common cause of idiopathic NS. Kidney biopsy usually outpatient data to estimate the prevalence and/or incidence
is not performed in childhood NS because most of them are of childhood NS, its contributing factors on infection and
MCD that respond to corticosteroids therapy. In this study, burden of diseases. Forth, the detailed data of medication
we found the kidney biopsy rate was 13.2%. Only 1.6% of use was not included as well to explore the use of steroids,

© 2012 The Authors

686 Nephrology © 2012 Asian Pacific Society of Nephrology
 Childhood NS with major infection

Table 5 Associations between the patients’ characteristics and their experience of infection admission during hospitalization among children admitted for
nephrotic syndrome (n = 176)

Number of admissions Ever infective Never infective Odds ratio P Value*

97 243 (95% CI)

Age (n, %)
1–4 44 (45.4%) 52 (21.4%) 5.312 (2.8–10.1) <.0001
5–9 32 (33.0%) 78 (32.1%) 2.575 (1.4–4.9) 0.0041
10–18 18 (18.6%) 113 (46.5%) Referent
Gender (n, %)
Female 25 (25.8%) 87 (35.8%) Referent
Male 72 (74.2%) 156 (64.2%) 1.606 (0.95–2.715) 0.0771
Location of admitted hospital (n, %)
North 26 (26.8%) 113 (46.5%) Referent
Middle 29 (29.9%) 68 (28.0%) 1.854 (1.008–3.407) 0.047
South 39 (40.2%) 56 (23.1%) 3.027 (1.677–5.464) 0.0002
East 3 (3.1%) 6 (2.5%) 2.173 (0.510–9.264) 0.2941
Hospital level (n, %)
Medical centre 56 (57.7%) 181 (74.5%) Referent
Regional hospital 36 (37.1%) 56 (23.1%) 2.078 (1.242–3.478) 0.0054
District hospital 5 (5.2%) 6 (2.5%) 2.694 (0.792–9.162) 0.1125
Admission season (n, %)
Spring 37 (38.1%) 58 (23.9%) 2.296 (1.205–4.375) 0.0115
Summer 20 (20.6%) 72 (29.6%) Referent
Fall 18 (18.6%) 55 (22.6%) 1.178 (0.569–2.437) 0.6589
Winter 22 (22.7%) 58 (23.9%) 1.365 (0.68–2.742) 0.3816

*Through the simple logistic regression analysis.

160 80
150 75
140 70
130 65

Number of admissions (NS only)

120 60
Number of admissions (×100)

110 55
100 50
90 45
80 40
70 35
60 30
50 25
40 20
30 15
20 10
10 5
0 0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Fig. 3 The number of admissions/person among those childhood nephrotic syndrome (NS) patients who ever infected or not during hospitalization during 1997
to 2007. Bar line represents the 95% confidence interval. All sick children admissions ( ); Sick children admissions* ( ); Sick children admissions* with major
infections ( ); Sick children admissions* with pneumonia ( ); Sick children admissions* with urinary tract infection (UTI) ( ); All childhood NS admissions
( ); Childhood NS admissions with major infections ( ); Childhood NS admissions with pneumonia ( ); Childhood NS admissions with UTI ( ).

© 2012 The Authors

Nephrology © 2012 Asian Pacific Society of Nephrology 687
C-C Wei et al.

other immunosuppressants and its impact on childhood NS 6. Liponski I, Cochat P, Gagnadoux MF et al. Bacterial complications
infection. Last, the simple logistic regression analysis was of nephrotic syndrome in children. Presse Med. 1995; 24: 19–22.
7. Adeleke SI, Asani MO. Urinary tract infection in children with
performed, rather than the multiple variable logistic regres-
nephrotic syndrome in Kano, Nigeria. Ann. Afr. Med. 2009; 8:
sion, due to the limited sample size for childhood NS. As
38–41.
such, further studies focusing on all childhood NS patients 8. Adedoyin OT, Ojuawo IA, Odimayo MS, Anigalaje EA. Urinary
might be needed to confirm the findings obtained from this tract infections in children with primary nephrotic syndrome and
study and its generalizability. acute glomerulonephritis. West Afr. J. Med. 2010; 29: 235–8.
Our study is one of few population-based studies to 9. Gulati A, Sinha A, Sreenivas V, Math A, Hari P, Bagga A. Daily
analyze childhood NS admissions associated with major corticosteroids reduce infection-associated relapses in frequently
relapsing nephrotic syndrome: A randomized controlled trial. Clin.
infections in Taiwan. Nineteen percent of childhood NS
J. Am. Soc. Nephrol. 2010; 6: 63–9.
admissions were associated with major infections. Young
10. Mekahli D, Liutkus A, Ranchin B et al. Long-term outcome of
age, admissions made in spring, located in central and south- idiopathic steroid-resistant nephrotic syndrome: A multicenter
ern Taiwan and in regional hospitals were the major associ- study. Pediatr. Nephrol. 2009; 24: 1525–32.
ated factors for infection. Further, age plays an important 11. Hodson EM, Willis NS, Craig JC. Non-corticosteroid treatment for
role in risk and types of infection. Given that the infection is nephrotic syndrome in children. Cochrane Database Syst. Rev. 2008;
associated with a great increase in LOS and hospital costs, (1): CD002290.
12. Gipson DS, Massengill SF, Yao L et al. Management of childhood
efficient strategies to prevent such complications could
onset nephrotic syndrome. Pediatrics 2009; 124: 747–57.
have a major impact on both the quality and cost of medical
13. Anonymous Minimal change nephrotic syndrome in children:
care. Further analyses of the clinical effectiveness and Deaths during the first 5 to 15 years’ observation. Report of the
cost-effectiveness of NS are warranted. International Study of Kidney Disease in Children. Pediatrics 1984;
73: 497–501.
14. Soeiro EM, Koch VH, Fujimura MD, Okay Y. Influence of
ACKNOWLEDGEMENTS nephrotic state on the infectious profile in childhood idiopathic
nephrotic syndrome. Rev. Hosp. Clin. Fac. Med. Sao Paulo 2004; 59:
This study is based in part on data from the National Health
273–8.
Insurance Research Database provided by the Bureau of
15. Hodson EM, Willis NS, Craig JC. Corticosteroid therapy for
National Health Insurance, Department of Health and nephrotic syndrome in children. Cochrane Database Syst. Rev. 2007;
managed by National Health Research Institutes. The inter- (4): CD001533.
pretation and conclusions contained herein do not represent 16. Chen YL, Chen JH. Approach of influence factors on infectious
those of the Bureau of National Health Insurance, Depart- complications in patients with primary nephrotic syndrome.
ment of Health or National Health Research Institutes. This Zhejiang Da Xue Xue Bao Yi Xue Ban 2003; 32: 145–8.
17. Moorani KN, Khan KM, Ramzan A. Infections in children with
work was partly funded by the China Medical University
nephrotic syndrome. J. Coll. Physicians Surg. Pak. 2003; 13: 337–9.
Hospital Research Grants DMR-97-010 and DMR-99-140.
18. Gulati S, Kher V, Gupta A, Arora P, Rai PK, Sharma RK. Spectrum
The funding body had no role in study design; in the collec- of infections in Indian children with nephrotic syndrome. Pediatr.
tion, analysis and interpretation of data; in the writing of the Nephrol. 1995; 9: 431–4.
manuscript; and in the decision to submit the manuscript for 19. Davis K, Huang AT. Learning from Taiwan: Experience with
publication. Further, the authors express their gratitude to universal health insurance. Ann. Intern. Med. 2008; 148: 313–14.
Dr Chia-Hung Chou for providing his insight and comment 20. Alwadhi RK, Mathew JL, Rath B. Clinical profile of children with
nephrotic syndrome not on glucorticoid therapy, but presenting
for the manuscript.
with infection. J. Paediatr. Child Health 2004; 40: 28–32.
21. Bekajlo D, Czabak R, Lech M, Molga B, Stefaniak J. Effect of
REFERENCES various environmental factors on the incidence of nephrotic
syndrome in children. Pol. Tyg. Lek. 1984; 39: 241–4.
1. Nephrotic syndrome in children: Prediction of histopathology from 22. Reeves WG, Cameron JS, Johansson SG, Ogg CS, Peters DK,
clinical and laboratory characteristics at time of diagnosis. A report Weller RO. Seasonal nephrotic syndrome. Description and
of the International Study of Kidney Disease in Children. Kidney immunological findings. Clin. Allergy 1975; 5: 121–37.
Int. 1978; 13: 159–65. 23. Toyabe S, Nakamizo M, Uchiyama M, Akazawa K. Circannual
2. Eddy AA, Symons JM. Nephrotic syndrome in childhood. Lancet variation in the onset and relapse of steroid-sensitive nephrotic
2003; 362: 629–39. syndrome. Pediatr. Nephrol. 2005; 20: 470–73.
3. Chen LZ, Jiang XY, Lu HY, Zhang QL, Mo Y. Efficacy and safety of 24. MacDonald NE, Wolfish N, McLaine P, Phipps P, Rossier E. Role of
cyclosporine A in treatment of refractory nephrotic syndrome in respiratory viruses in exacerbations of primary nephrotic
children: A systematic review of randomized controlled trials. syndrome. J. Pediatr. 1986; 108: 378–82.
Zhonghua Er Ke Za Zhi 2009; 47: 898–903. 25. Lien HM, Chou SY, Liu JT. Hospital ownership and performance:
4. Wilfert CM, Katz SL. Etiology of bacterial sepsis in nephrotic Evidence from stroke and cardiac treatment in Taiwan. J. Health
children 1963–1967. Pediatrics 1968; 42: 840–43. Econ. 2008; 27: 1208–23.
5. Krensky AM, Ingelfinger JR, Grupe WE. Peritonitis in childhood
nephrotic syndrome: 1970–1980. Am. J. Dis. Child. 1982; 136:
732–6.

© 2012 The Authors

688 Nephrology © 2012 Asian Pacific Society of Nephrology