- Селуянов В.Н., Биомеханика Двигательного Аппарата Человека

HIPS
KNEES
ANKLES
ELBOWS
SHOULDERS
PROMs
14th Annual Report
2017
National Joint Registry
for England, Wales,
Northern Ireland and
the Isle of Man
ISSN 2054-1821 (Print) Surgical data to 31 December 2016
Prepared by
The NJR Editorial Board
NJRSC Members
Michael Green
Mr Peter Howard
Mr Martyn Porter (Chairman, Editorial Board)
Professor Mark Wilkinson
Nick Wishart
NJR RCC Network Representatives

Mr Matthew Porteous (Chairman, RCC Sub-committee)
Mr Adam Watts
Orthopaedic Specialists
Mr Colin Esler
Mr Andy Goldberg
Professor Jonathan Rees
Healthcare Quality Improvement Partnership

NJR Management Team and NJR Communications
James Ludley
Elaine Young
Northgate Public Services

NJR Centre, IT and data management
Victoria McCormack
Anita Mistry
Dr Claire Newell
Dr Martin Pickford
Martin Royall
Mike Swanson
University of Bristol / University of Oxford

NJR Statistical support, analysis and research team
Professor Yoav Ben Shlomo
Professor Ashley Blom
Dr Emma Clark
Kevin Deere
Dr Celia Gregson
Dr Linda Hunt
Dr Andrew Judge
Sofia Mouchti
Professor Andrew Price
Adrian Sayers
Mr Michael Whitehouse
Pad Creative Ltd (design and production)
This document is available in PDF format for download from the NJR Reports website at www.njrreports.org.uk. Additional data and
information can also be found as outlined on pages 22-23.
National Joint Registry | 14th Annual Report
Chairman’s foreword
Laurel Powers-Freeling, National Joint Registry Chairman
The National Joint Registry’s Steering Committee

(NJRSC) oversees the work programme of the registry.
As Chairman it is always a pleasure to offer a foreword
to our Annual Report, now in its 14th edition, and to
update on the substantial developments during the
previous financial year.
Key work and developments

Improving data quality remains our number one overall
strategic priority. With our ever-maturing dataset, it is
critically important that the registry is collecting the most
relevant, high quality data in order to provide robust
evidence to support decision-making in regard to patient
safety, standards in quality of care and overall cost
effectiveness in joint replacement surgery. source of evidence for regulators, such as the Care
Quality Commission (CQC), to inform their judgements
The NJR’s national programme, which is aimed at about services, as well as being a fundamental driver
assessing data completeness and quality within the to inform improved quality of care for patients. As such,
registry, is now in its second year and underpins the this year we have further reviewed the NJR’s processes
NJR’s ‘Supporting Data Quality Strategy’. The data in monitoring implant and surgeon performance as part
quality audit has allowed the NJR to compare the of the development of the NJR’s ‘Accountability and
records in local hospitals’ databases to the registry’s Transparency Model’.
records, with the aim of ensuring the accuracy of the
number of arthroplasty procedures submitted when As part of the new model, ‘prevention’ is now a key
compared to the number carried out. element of the NJR’s monitoring process. Implemented
for the first time this year, ‘borderline outlier’ notifications
I’m delighted we were able to extend the programme were issued, acting to prevent surgeons from becoming
this year to check the quality of data in independent ‘outliers’ by alerting them to deteriorating outcomes and
healthcare organisations, giving all hospitals – now in thus enabling them to correct substandard practices and
both the NHS and independent sector – the opportunity reduce or eliminate poor outcomes. This new function
to demonstrate the highest possible standards of clinical should ensure even greater public confidence in the NJR
governance, which all are striving to achieve. monitoring process across the orthopaedic sector.
We are now able to fully report the audit’s findings in Looking ahead, ‘practitioner reflection’ will also
year one. These findings can be found on pages 18-20. become a key pillar in the NJR’s monitoring process.
I would like to offer my thanks on behalf of the NJR to all A bold new approach, which has the endorsement
staff who have worked to complete this important audit of the BOA and NHS Improvement, will see the NJR
and who I hope will enact upon its findings. monitoring surgeon engagement and reflection on
their own practice and performance data. This new
Elsewhere, monitoring continues to be a key function
process will allow joint replacement surgeons the unique
of the NJR. Registry data now provides an important
www.njrcentre.org.uk 3
opportunity to demonstrate and record, via the NJR’s • Undertake a complete redevelopment of the NJR’s
Clinician Feedback tool, that they have reviewed their main website (www.njrcentre.org.uk)
NJR data as part of their appraisal and revalidation, • Provide further analyses and investigation of NJR
and importantly reflected upon the data. At the time PROMs at 3 and 5 years
of writing, we are at an exciting juncture but once
implemented it will be ground-breaking for the NHS and Acknowledgements
for patient safety and reassurance.
During this reporting period, there have been some
An additional area of national policy that the NJR changes to the membership of the NJRSC. Professor
continues to support is the ongoing work surrounding Andrew Price stepped down from his role as a surgeon
the Getting It Right First Time (GIRFT) initiative. GIRFT member of the committee as he formally became part
essentially aims to bring about higher-quality care in of the NJR’s statistical analysis contractor team. My
hospitals, at lower cost, by reducing unwanted variations sincere thanks to Andrew for his valuable contribution in
in services and practices. The NJR’s implant price- his NJRSC role, we are fortunate to retain his services
benchmarking data has from the outset underpinned the and abilities in a new capacity. Work is underway with
initiative for orthopaedics. the Department of Health’s appointments team to find
Andrew’s replacement and also to appoint a senior
NJR pricing data gives providers the opportunity to specialist practitioner with an interest in orthopaedics. Mr
benchmark the price they pay for orthopaedic implants Martyn Porter, NJR Vice Chair and Medical Director, and
against the ‘best’ national prices achieved. Importantly, Sue Musson, patient representative member, were both
NJR data also helps ensure an important clinical context granted extensions to their membership of the NJRSC
is built into the initiative by providing surgeons access until March 2018 and October 2017, respectively.
to their individual-level price-benchmarking data. All
these services are now inclusive of the NJR’s annual This year I have also appreciated the significant
subscription charge. contribution made by both Mr Tim Wilton and Mr Ian
Winson. As immediate past-BOA president and BOA
A recent King’s Fund report1 into GIRFT highlighted that president respectively, Tim and Ian’s engagement
clinicians were engaging with the data and acting on as co-opted members of the NJRSC – particularly
the evidence provided. This is an area of work that the during this period of change and review of the NJR’s
NJR will continue to support and work closely with NHS monitoring process – has been invaluable. This
Improvement and the GIRFT team. September when a new president takes up post, I look
forward to welcoming him and hope to maintain the
Future plans for the coming year 2017/18 support and close links developed with the profession.
Patients and the public can be assured that the NJR
Importantly, I would like to end by thanking all members
is working hard to collect and report upon the most
of the NJRSC, and NJR sub-committees, for their
complete, accurate data possible across all hospitals in
valuable contribution and enthusiasm. In particular, my
England, Wales, Northern Ireland and the Isle of Man. In
thanks to Mr Martyn Porter, NJR Vice Chair and Medical
addition to our core schedule of activities, we will:
Director, and to the chairs of each of the NJR sub-
• Continue to develop NJR information systems, committees for their contribution, hard work and insight.
including enhanced Clinician Feedback to aid Without their dedication, the NJR would not be the world
surgeon appraisal, Supplier Feedback, Management leading and ground-breaking arthroplasty register that it
Feedback and Annual Clinical Reports is. I would encourage you to read and review the reports
from each committee chairman at www.njrreports.org.uk
• Roll out a dedicated NJR data access and research
where they provide strategic oversight and professional
portal to allow researchers to access the NJR
awareness into key work areas.
dataset via secure access
1 Timmins, N., ‘Tackling variations in clinical care: Assessing the Getting It Right First Time (GIRFT) programme’ (June 2017). www.kingsfund.org.uk (last
accessed 10 July 2017).
4 www.njrcentre.org.uk
Finally, my thanks to the NJR management team, in

particular our Director of Operations, Elaine Young.
Thanks also go to our NJR contractors, Northgate
Public Services (UK) Ltd., the University of Bristol, and
the communications team based at the Healthcare
Quality Improvement Partnership, for all their hard work
and efforts throughout the year in progressing the NJR
work agenda.
Laurel Powers-Freeling
Chairman, National Joint Registry Steering Committee
Contents
Chairman’s foreword 3
Executive summary 14
Part 1 NJR Data Quality Audit 17

1.1 NJR Data Quality Audit update 18
Part 1 Annual progress 21

1.2 Annual Report Introduction 22
1.3 Annual Progress 22
1.4 Summary of content for the NJR Annual Report 23
Part 2 Clinical activity 2016 and using the dedicated NJR Reports website 24
2.1 Clinical activity 2016 overview 25
2.2 Navigating the NJR Reports online facility 28
Part 3 Outcomes after joint replacement 2003 to 2016 29

3.1 Executive summary 29
3.2 Summary of data sources and linkage 35
3.3 Outcomes after primary hip replacement 40

3.3.1 Overview of primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3.2 Revisions after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.3.3 Revisions after primary hip surgery: effect of head size for selected bearing surfaces/fixation sub-groups. . . 61
3.3.4 Revisions after primary hip surgery for the main stem-cup brand combinations. . . . . . . . . . . . . . . . . . . . . . . 68
3.3.5 Revisions for different causes after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.6 Mortality after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.3.7 Primary hip replacement for fractured neck of femur compared with other reasons for implantation . . . . . . . 80
3.3.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.4 Revisions of a total hip replacement 84

3.4.1 Overview of hip revision procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.4.2 Rates of hip re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.4.3 Reasons for the hip re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.4.4 90-day mortality after hip revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.5 Outcomes after primary knee replacement 98

3.5.1 Overview of primary knee surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.5.2 First revision after primary knee surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.5.3 Mortality after primary knee surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.6 Revisions of knee replacements 143

3.6.1 Overview of knee revisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
3.6.2 Survival of first recorded knee revision to any subsequent re-revision procedure. . . . . . . . . . . . . . . . . . . . . 146
3.6.3 Reason for knee re-revision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3.7 Outcomes after primary ankle replacement 155

3.7.1 Overview of primary ankle surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
3.7.2 Revisions after primary ankle surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
3.7.3 Mortality after primary ankle replacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.7.4 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.8 Outcomes after primary shoulder replacement 160

3.8.1 Overview of primary shoulder replacement surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
3.8.2 Revisions after primary shoulder replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
3.8.3 PROMS Oxford Shoulder Scores (OSS) associated with primary shoulder replacement surgery . . . . . . . . . 175
3.8.4 Mortality after primary shoulder replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
3.8.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
3.9 Outcomes after primary elbow replacement 178

3.9.1 Overview of primary elbow replacement surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
3.9.2 Revisions after primary elbow replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
3.9.3 Mortality after primary elbow replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
3.9.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Part 4 Implant and unit-level activity and outcomes 187

4.1 Implant performance 188
4.2 Clinical activity 188
4.3 Outlier units for 90-day mortality and revision rates for the period 2003 to 2016 189
4.4 Better than expected performance 190
Glossary 193
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Part 3 tables
Table 3.1 Summary description of linked datasets used for main survivorship analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Table 3.2 Composition of person-level datasets for main survivorship analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Table 3.3 Numbers and percentages of primary hip replacements of each fixation type and by bearing surface . . . . . . . . . . . 42
Table 3.4 Percentages of primary hip replacements in each calendar year that use each fixation type and for each
fixation group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Table 3.5 Distribution of age at primary hip replacement (in years) and gender, for all procedures and for each type of fixation
and bearing surface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Table 3.6 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) after primary hip replacement,
by year from the primary operation, for all cases and by fixation and bearing surface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Table 3.7 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI), by gender and age, at
1, 3, 5, 7, 10 and 13 years from the primary hip replacement, for each fixation group and main bearing surface . . . . . . . . . . . 57
Table 3.8 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) at 1, 3, 5, 7, 10, and 13 years
after the primary hip replacement operation, for the most commonly used cup-stem brand combinations . . . . . . . . . . . . . . . . 68
Table 3.9 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) at 1, 3, 5, 7, 10 and
13 years after the primary hip replacement for the most commonly used cup-stem brand combinations with further sub-division
by main bearing surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Table 3.10 Revision rates after primary hip replacement for each indication, expressed as numbers per 1,000 patient-years
(95% CI), for all cases and by fixation and bearing surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Table 3.11 Revision rates after primary hip replacement for each indication, expressed as numbers per 1,000 patient-years
(95% CI), overall and by time interval from primary operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Table 3.12 Kaplan-Meier estimates of the cumulative percentage mortality (95% CI), at different time points after primary hip
replacement, for all cases and by age/gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Table 3.13 Proportions of primary total hip replacements for fracture of the neck of femur by year of primary operation . . . . . 80
Table 3.14 Comparison between primary hip replacements for fractured neck of femur and the remainder of cases with
respect to gender, age and type of primary hip received . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Table 3.15 Numbers of all hip revision procedures, by type of procedure, carried out each year . . . . . . . . . . . . . . . . . . . . . . . 85
Table 3.16 Reasons for the hip revision procedures: percentages indicating each reason, calculated separately for single and
two-stage revisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Table 3.17 (a) Kaplan-Meier estimates of cumulative percentage probability of a hip re-revision following the first revision . . . 95
Table 3.17 (b) Kaplan-Meier estimates of cumulative percentage probability of a hip re-revision following the first revision,
sub-divided by time since primary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Table 3.17 (c) Kaplan-Meier estimates of the cumulative percentage probability of a hip re-revision (95% CI) at 1, 3, and 5
years following the first revision in those with documented primaries in the NJR, by fixation and bearing surface . . . . . . . . . . . 95
Table 3.18 Reasons for the hip first revision and subsequent re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Table 3.19 Temporal changes in first hip revisions reported in the NJR and associated indications . . . . . . . . . . . . . . . . . . . . . 97
(a) Number of first hip revisions by year and proportions with an associated primary in the NJR . . . . . . . . . . . . . . . . . . . . . .97
(b) Numbers of first recorded hip revisions by stage and whether or not primary was in the NJR . . . . . . . . . . . . . . . . . . . . . 97
Table 3.20 Numbers and percentages of primary knee replacements by fixation method, constraint and bearing type . . . . . 103
Table 3.21 Percentage of all primary knee replacements performed each year by total and partial knee replacement types
by fixation method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Table 3.22 Reasons for primary knee replacement surgery; number and percentage of all NJR recorded primary knee
replacement surgeries carried out for each clinical reason broken down by gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Table 3.23 Descriptive statistics of total knee replacement, unicondylar and patellofemoral procedures performed
by consultant and unit by year of surgery in the last three years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Table 3.24 Age (in years) and percentage (%) male at primary operation for different types of knee replacement and
by fixation, constraint and bearing type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Table 3.25 (a) Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) at specified
times after primary knee replacement, by fixation, constraint and bearing type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Table 3.25 (b) Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) at specified
times after primary knee replacement, by age and gender, for each fixation, constraint and bearing group . . . . . . . . . . . . . . 124
Table 3.26 Revision rates (95% CI), expressed as number of revisions per 1,000 patient-years (PTIRs), for each
recorded reason for first knee revision. Rates shown are for all revised cases by total replacement fixation method and
by type of partial replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Table 3.27 Revision rates (95% CI), expressed as number of revisions per 1,000 patient-years (PTIRs), for each
recorded reason for first knee revision. Rates shown are broken down by constraint and bearing sub-group for each total
replacement fixation method and for unicondylar partial replacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Table 3.28 Revision rates (95% CI) broken down by time period in which primary was revised, expressed as number
of revisions per 1,000 patient-years (PTIRs), for each recorded reason for first knee revision . . . . . . . . . . . . . . . . . . . . . . . . . 135
Table 3.29 Kaplan-Meier estimated cumulative percentage probability of first revision (95% CI) of a primary total knee
replacement by main type of implant brand at the indicated number of years after primary operation. . . . . . . . . . . . . . . . . . . 136
Table 3.30 Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) of a primary
unicompartmental knee replacement by main type of implant brand at the indicated number of years after primary operation. . . 137
Table 3.31 Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) of a total knee
replacement or unicompartmental knee replacement at the indicated number of years after primary operation, by main
implant brands and, within brand, by type of fixation, constraint and bearing sub-group . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Table 3.32 Kaplan-Meier estimated cumulative percentage probability (95% CI) of a patient dying at the indicated
number of years after a primary knee joint replacement operation by age group and gender . . . . . . . . . . . . . . . . . . . . . . . . . 142
Table 3.33 Numbers of knee joint revision operations carried out each year, by revision operation type . . . . . . . . . . . . . . . . 144
Table 3.34 Percentage of all revision knee procedures of each stage type with the indicated reason for revision. . . . . . . . . . 145
Table 3.35 (a) Kaplan-Meier estimates of cumulative percentage probability of knee re-revision following the first revision
using different start points for time at risk of re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Table 3.35 (b) Kaplan-Meier estimates of cumulative percentage probability of knee re-revision following the first revision
broken down by whether a primary is on record in the NJR or not. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Table 3.35 (c) Kaplan-Meier estimates of the cumulative percentage probability of knee re-revision following the first
revision when the group of patient-sides with a primary record in the NJR are stratified by the time intervals in which the
first revision took place after the primary operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Table 3.36 Reasons given for first knee revision and re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Table 3.37 Temporal changes in first knee revisions reported in the NJR and associated indications . . . . . . . . . . . . . . . . . . 152
(a) Number of first knee revisions by year of surgery and proportions with an associated knee primary in the NJR . . . . . . . 152
(b) Numbers of first recorded knee revisions by stage and whether or not primary was in the NJR . . . . . . . . . . . . . . . . . . . 153
Table 3.38 (a) Descriptive statistics of ankle procedures performed by consultant and unit by year of surgery . . . . . . . . . . . 156
Table 3.38 (b) Numbers of primary ankle replacements by ankle brand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Table 3.39 Indications for the 153 (first) revisions following primary ankle replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Table 3.40 Kaplan-Meier estimates of the cumulative percentage mortality (95% CI), by gender and age at 90 days, and
1, 2, 3, 4 and 5 years after primary ankle replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Table 3.41 Numbers of all primary shoulder replacements (elective and acute trauma) by year and percentages of
each type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Table 3.42 Numbers of centres and consultant surgeons providing primary shoulder replacements over the last three
years, 2014-2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Table 3.43 Reasons for main types of primary shoulder replacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Table 3.44 Gender and age at primary for the main types of primary shoulder replacements, shown separately
for acute trauma and elective cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Table 3.45 Stem brands used in primary procedures (not resurfacing) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Table 3.46 Resurfacing brands used in primary resurfacing shoulder replacements, shown separately for acute trauma
and elective cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Table 3.47 Composition and fixation of standard glenoids used in total conventional shoulder arthroplasty . . . . . . . . . . . . . 169
Table 3.48 List of manufacturers of the standard glenoids used in total conventional shoulder arthroplasty . . . . . . . . . . . . . 169
Table 3.49 Kaplan-Meier estimates of cumulative percentage probability of revision (95% CI) by time from shoulder
primary procedure, shown separately for acute trauma and elective cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Table 3.50 Kaplan-Meier estimates of cumulative percentage probability of revision (95% CI) by time from elective
shoulder primary, by gender and age at 1, 2, 3 and 4 years from the primary operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
primary procedure, for all elective cases, sub-divided by the type of procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Table 3.52 Number of first revisions for each type of primary shoulder replacement and indications for revision . . . . . . . . . . 174
(i) Acute trauma cases only. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
(ii) Elective cases only. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Table 3.53 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from shoulder
primary, for acute trauma and elective cases at 90 days, 1, 2, 3 and 4 years from the primary shoulder replacement . . . . . . 176
Table 3.54 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from elective
shoulder primary, by age and gender at 90 days, 1, 2, 3 and 4 years from the primary shoulder replacement . . . . . . . . . . . . 176
Table 3.55 Numbers of primary elbow replacements by year and percentages of each stated type of procedure . . . . . . . . . 179
Table 3.56 Reasons for main types of primary elbow replacements, by year of primary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
Table 3.57 Types of primary elbow procedures used in acute trauma and elective cases . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
Table 3.58 Number of units and consultant surgeons providing any primary elbow replacements during each year from
2014 to 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Table 3.59 Number of units and consultant surgeons providing primary total prosthetic replacements . . . . . . . . . . . . . . . . . 182
Table 3.60 Brands used in total prosthetic replacements and lateral resurfacing replacements . . . . . . . . . . . . . . . . . . . . . . . 183
Table 3.61 Radial head brands used in total prosthetic replacements, radial head replacements and lateral resurfacing
replacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Table 3.62 Kaplan-Meier estimates of cumulative percentage probability of revision (95% CI) by time from elbow primary
procedure, shown separately for acute trauma and elective cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Table 3.63 Indications for first data linked revision after any primary elbow replacement. Acute trauma and elective cases
are shown separately, for all cases and for total prosthetic replacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Table 3.64 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from elbow primary,
for acute trauma and elective cases at 90 days, 1, 2 and 3 years from the primary elbow replacement . . . . . . . . . . . . . . . . . 186
Part 3 figures
Figure 3.1 (a) Initial numbers of procedures for analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Figure 3.1 (b) Total volume of uploads to the NJR, percentage of procedures consenting to be included in the NJR, and
percentage of patients traced in the NJR, in England and Wales by year of operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Figure 3.2 Temporal changes in percentages of each fixation method used in primary hip replacements . . . . . . . . . . . . . . . . 44
Figure 3.3 (a) Temporal changes in percentages of each bearing surface used in cemented primary hip replacements . . . . . 44
Figure 3.3 (b) Temporal changes in percentages of each bearing surface used in uncemented primary hip replacements . . . 45
Figure 3.3 (c) Temporal changes in percentages of each bearing surface used in hybrid primary hip replacements . . . . . . . . 45
Figure 3.3 (d) Temporal changes in percentages of each bearing surface used in reverse hybrid primary hip replacements . . 46
Figure 3.4 (a) Temporal changes in revision rates after primary hip replacement: Kaplan-Meier estimates of cumulative
percentage probability of revision for each year of primary operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Figure 3.4 (b) Temporal changes in revision rates after primary hip replacement: Kaplan-Meier estimates of cumulative
percentage probability of revision for each year of primary operation with failure rates at 1, 3, and 5 years indicated . . . . . . . . 50
Figure 3.5 Comparison of cumulative probability of revision (Kaplan-Meier estimates) for cemented primary hip replacements
with different bearing surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Figure 3.6 Comparison of cumulative probability of revision (Kaplan-Meier estimates) for uncemented primary hip replacements
with different bearing surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Figure 3.7 Comparison of cumulative probability of revision (Kaplan-Meier estimates) for hybrid primary hip replacements with
different bearing surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Figure 3.8 Comparison of cumulative probability of revision (Kaplan-Meier estimates) for reverse hybrid primary hip
replacements with different bearing surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Figure 3.9 (a) Cumulative probability of revision (Kaplan-Meier) for the whole cohort of primary hip replacements broken down
by age separately for each gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Figure 3.9 (b) Cumulative probability of revision (Kaplan-Meier) for the whole cohort of primary hip replacements broken down
by age separately for each gender, but excluding metal-on-metal (or uncertain) total hip replacement and resurfacings . . . . . 56
Figure 3.10 Effect of head size on cumulative revision rates after primary hip replacement using different bearing groups (only
head sizes used in >500 hips are shown). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
(a) Metal-on-polyethylene cemented monobloc cups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
(b.i) Metal-on-polyethylene uncemented metal shells with polyethylene liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
(b.ii) Metal-on-polyethylene with truncated data for head size 44mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
(c) Metal-on-metal uncemented metal cups or metal shells with metal liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
(d) Ceramic-on-polyethylene cemented monobloc cups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
(e) Ceramic-on-polyethylene uncemented metal shells with polyethylene liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
(f) Ceramic-on-ceramic uncemented metal shells with ceramic liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Figure 3.11 (a) Change in PTIR with time from primary hip replacement, for aseptic loosening for selected
fixation/bearing sub-groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Figure 3.11 (b) Change in PTIR with time from primary hip replacement, for pain for selected fixation/bearing sub-groups. . . 76
Figure 3.11 (c) Change in PTIR with time from primary hip replacement, for dislocation/subluxation for selected fixation/bearing
sub-groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Figure 3.11 (d) Change in PTIR with time from primary hip replacement, for infection for selected fixation/bearing
sub-groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Figure 3.11 (e) Change in PTIR with time from primary hip replacement, for adverse soft tissue reaction to particulate debris for
selected fixation/bearing sub-groups.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Figure 3.11 (f) Change in PTIR with time from primary hip replacement, for adverse soft tissue reaction to particulate debris for
selected fixation/bearing sub-groups including primaries since 2008 only.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Figure 3.12 Cumulative percentage revision rates (Kaplan-Meier) for hip primaries implanted for fractured neck of femur
compared with all other cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Figure 3.13 Cumulative percentage mortality (Kaplan-Meier) for hip primaries implanted for fractured neck of femur compared
with all other cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Figure 3.14 (a) Kaplan-Meier estimate of the cumulative probability of a hip re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Figure 3.14 (b) Kaplan-Meier estimates of the cumulative probability of a hip re-revision, shown separately for those with
documented primaries in the NJR and the remainder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Figure 3.14 (c) Kaplan-Meier estimates of the cumulative probability of a hip re-revision up to three years from the
first revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Figure 3.15 Kaplan-Meier estimates of the cumulative probability of a hip re-revision up to five years from the first
revision, shown separately for type of fixation used in the primary, with further sub-division by length of time from the
primary to the first revision (<1, 1-3, 3-5 and >5)
(a) Cemented . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
(b) Uncemented . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
(c) Hybrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
(d) Reverse hybrid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
(e) Resurfacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Figure 3.16 Exploring unit and consultant frequency of knee surgery and total caseload in the NJR between 2014 and 2016
(a) TKR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
(b) Unicondylar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
(c) Patellofemoral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Figure 3.17 (a) Changes in cumulative percentage chance of knee replacement failure by year of primary operation. Kaplan-Meier
estimates of cumulative percentage probability of a first revision grouped by year in which primary surgery took place . . . . . . . . 114
Figure 3.17 (b) Temporal changes in revision rates after primary knee replacement: Kaplan-Meier estimates of cumulative
percentage probability of revision for each year of primary operation with failure rates at 1, 3, and 5 years indicated.. . . . . . . 115
Figure 3.18 (a) Kaplan-Meier estimates of the cumulative percentage probability of a first revision of primary cemented knee
replacement broken down by age group (age at primary in years) at increasing years after the primary surgery. . . . . . . . . . . 116
Figure 3.18 (b) Kaplan-Meier estimates of the cumulative percentage probability of a first revision of primary unicondylar knee
replacement broken down by age group (age at primary in years) and gender at increasing years after the primary surgery. . 117
Figure 3.19 (a) Comparison of the Kaplan-Meier cumulative percentage probability estimates of a knee prosthesis first revision
for different constraint and bearing types at increasing years after the primary surgery when the primary arthroplasty method of
fixation is cemented only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Figure 3.19 (b) Comparison of the Kaplan-Meier cumulative percentage probability estimates of a knee prosthesis first
revision for different constraint and bearing types at increasing years after the primary surgery when the primary
arthroplasty method of fixation is uncemented or hybrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Figure 3.19 (c) Comparison of the Kaplan-Meier cumulative percentage probability estimates of a knee prosthesis first
revision for different constraint and bearing types at increasing years after the primary surgery when the primary is a
unicondylar or patellofemoral partial knee replacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Figure 3.20 (a) Kaplan-Meier estimate of the cumulative percentage probability of a knee re-revision, based on time from
the start date of the first revision episode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Figure 3.20 (b) Kaplan-Meier estimate of the cumulative percentage probability of a knee re-revision, based on time from
the last date of the first revision episode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Figure 3.20 (c) Kaplan-Meier estimates of the cumulative percentage probability of a knee re-revision, shown for those with
documented primaries in NJR and the remainder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Figure 3.20 (d) Kaplan-Meier estimates of the cumulative percentage probability of a knee re-revision up to three years from
the first revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Figure 3.21 Exploring unit and consultant frequency of primary ankle surgery, and total caseload in the NJR between
2014 and 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Figure 3.22 Gender and age distribution of elective primaries, for each type of primary procedure . . . . . . . . . . . . . . . . . . . . 166
Figure 3.23 Kaplan-Meier estimate of the cumulative percentage probability of revision after primary shoulder replacement with
acute trauma and elective cases shown separately . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Figure 3.24 Kaplan-Meier estimates of cumulative percentage probability of revision up to four years from primary shoulder
replacement surgery, by type of procedure, for elective cases only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Figure 3.25 Kaplan-Meier estimate of the cumulative percentage probability of revision after primary total prosthetic
replacement with acute trauma and elective cases shown separately . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Executive summary
Mr Martyn Porter, NJR Medical Director and Chairman, Editorial Board
The National Joint Registry’s maturing dataset, now

in its 14th year of reporting, offers the orthopaedic
community the invaluable ability to see important
determinants that influence the outcome and longevity
of joint replacement procedures. Monitoring and
reporting high quality, robust data suitable for decision-
making remains the registry’s core mission and I’m
delighted to present this year’s findings.
To ensure accurate annual reporting the NJR continues

to work with many stakeholders including hospitals,
industry, and individual surgeons. As outlined by the
NJR Chairman in her foreword, the NJR’s remit has
naturally broadened and the ability for the dataset to
drive forward change in other areas has grown – from
patient recorded outcome measures (PROMs) to
implant price-benchmarking, from research to surgeon
accreditation. As well as also being an important The increase in cases submitted during this period
source of evidence for regulators, such as the Care could also be in part due to the NJR’s sustained
Quality Commission (CQC), to inform their judgements programme to improve data quality and compliance
about services. in the registry. Namely, this has been the work
surrounding the data quality audits rolled-out across all
The NJR’s Annual Report is a fundamental pillar and eligible NHS hospitals, and for the first time during this
showcases how we monitor the performance of year, independent sector hospitals too. As such, the
implants, hospitals and surgical technique but also NJR has been able to work with hospitals to improve
how the registry is driving quality improvement in the their NJR processes, to ensure that all eligible primary
orthopaedic sector as a whole. and revision joint replacement operations are recorded
on the database and put forward for analysis.
Main headlines for 2016: procedures
and data quality Many hospitals work hard to ensure that they record all
eligible procedures. However, the completeness of data
During 2016/17, there were a total of 242,629 cases within the NJR is reliant on the input at the local level,
submitted to the NJR, an increase of over 20,000 on which the audit has highlighted is subject to variation
the previous year, which brings the total number of across hospitals.
records in the registry to approximately 2.35 million.
This is despite concern that the overall number of We are now able to fully report the audit’s findings in
joint replacement procedures being undertaken was year one. These findings can be found on pages 18-20.
decreasing. The constantly high number of cases I offered some very early analysis in last year’s Annual
submitted per year suggests continuing high levels of Report for year one of the audit, which highlighted
patient confidence and clinical performance, in what is a a really pleasing low overall level of missing records.
remarkably successful surgical intervention. However, the audit found that the proportion of missing
data is greater for revision procedures. The failure of
hospitals to upload revision procedures into the NJR the data, which highlights that the patient has an
is concerning, as linked revision procedures form the important effect on how long an implant will last. This
basis of the analyses of implant failure and surgical year’s analyses continue to show the increased risk of
performance – which fundamentally underpin the core revision associated with younger patients across all joint
purpose of the NJR. procedures recorded in the registry. This is particularly
important given the increase in total numbers of
Put another way, one of the NJR’s principal outcomes younger patients undergoing joint replacement.
of interest is revision surgery, an indication of implant
failure or surgical performance. This is determined As previously outlined by the NJR, if younger patients
by linking a primary arthroplasty procedure to a are most likely to need at least one revision surgery in
secondary procedure, which typically occurs a number their lifetime, then we must use the maturing dataset
of years after the primary procedure. Analyses of of the NJR to get the first-time surgery as right for the
revision estimates in this year’s Annual Report highlight patient as possible.
why compliance with reporting revision surgical
procedures is essential to estimate implant failure rates For example, the revision rate for total hip replacement
and surgical performance more accurately. Further increases at a faster rate over time for younger patients.
investigation is underway to ascertain whether these To explore this further, this year’s report examines the
are random events or a systematic under-reporting of effect of age and gender on hip revision rates across
revision procedures. the construct groups for the first time.
Main headlines for 2016: outcomes Elsewhere with hip replacement data, our analysis
confirms that choice of head size is an important factor
Across all joint procedures recorded in the registry, in determining revision outcome. For both metal-on-
revision estimates following primary joint replacement polyethylene and ceramic-on-polyethylene bearing
procedures remain low. For example, primary total choices, higher failure rates are seen with larger head
hip replacement revision estimates are less than five sizes. Importantly, the data indicates head sizes of
percent for the majority of procedures at thirteen years. 36mm and above are associated with increasingly
Knee replacement data in numerous ways mirrors that higher failure rates.
of hip replacement. Similarly, there are very positive
outcomes reflected in the ankle, shoulder, and now for Linked to the theme of younger patients, the knee
the first time, elbow joint replacement data too. replacement data again continues to show similar
trends. Given partial knee replacement surgery is used
These outcomes are extremely impressive and generally in younger patients, the importance of the
underpin the enormous success and reliability of joint effects of patient factors which influence the outcome
replacement surgery. These sorts of results should must be considered. For patients undergoing total knee
help drive greater confidence in the public and with replacement at the median age (69 years old), the 13-
commissioners of healthcare, that joint replacement year risk of revision is just over 4%. However, for total
is one of the most effective and cost effective knee replacement patients under the age of 60, the
interventions that the NHS has to offer. risk increases with decreasing age, reaching 10% for
those under 55 years old. This pattern is magnified in
Furthermore, data for this year outlines that osteoarthritis
unicondylar replacement, with patients under the age
is almost exclusively the diagnosis for both primary hip
of 55 facing a 25% chance of revision by 13 years. This
and knee joint replacements, in 90% and 99% of cases
has been a consistent finding across all annual reports.
respectively. Therefore, we should not lose sight of the
fact that joint replacement surgery offers significant Further improvements to the representation of shoulder
benefits – getting patients back to their chosen lifestyle replacement data have been made. We have made
sooner, free from pain with improved mobility. the distinction for the first time between stemmed
and stemless humeral implants, as well as improved
However, those in the orthopaedic community must
representation of data in stem branding to reflect which
continue to note an important trend emerging from
implants are being used less or more frequently year
on year. Naturally, trauma and elective procedures I would like end by acknowledging our NJR
continue to be separated out. Contractors: the hard work undertaken and led by
teams at the University of Bristol with support from
Elsewhere, we also report on ankle and elbow colleagues at the University of Oxford, who have once
replacements (Sections 3.7 and 3.9). As these are again provided excellent provision in terms of analysing
carried out less frequently and we have a shorter the outcomes following primary surgery and the many
follow-up period (since 2010 and 2012 respectively), peer reviewed publications which have been produced
data are still at a relatively early stage. However, I am from the registry data. I would particularly encourage
pleased that the British Elbow and Shoulder Society you to explore the research published since the last
(BESS) and the British Orthopaedic Foot and Ankle annual report on ethnicity and joint replacement1
Society (BOFAS) continue to work very closely with and the main cause of death following primary total
the NJR to take the collection and analysis of the hip and knee replacement for osteoarthritis2. Finally,
data forward. also to Northgate Public Services Ltd who provide
the IT support and expertise for the NJR to achieve
Concluding acknowledgements these outputs.
As well as the pages of this report, I would encourage
you to explore the NJR’s dedicated annual report
website at www.njrreports.org.uk. The website
offers a helpful interactive platform for Part 2 of the
report, which is the descriptive NJR data; supporting
appendices; and, when published, the latest NJR
Patient and Public Guides to the annual report. Mr Martyn Porter
The NJR continues to work with many stakeholders NJR Medical Director and Chairman,
including patients, regulators, hospitals, industry, Editorial Board
individual surgeons and procurement, to ensure
accurate annual reporting. To conclude, I would like
to thank NJR Chairman, Laurel Powers-Freeling, and
all members of the NJR Steering Committee, the
Editorial Board and other NJR sub-committees, and the
NJR Operational Management and Communications
team, all of whom have supported the production of
this report, and indeed all the orthopaedic surgeons
in hospitals that contribute data. The collective effort
ensures that the National Joint Registry maintains its
position as the largest and world-leading arthroplasty
registry, with a sharp focus on patient safety.
1 Smith MC, et al., ‘Rates of hip and knee joint replacement amongst different ethnic groups in England: an analysis of National Joint Registry data’,
Osteoarthritis and Cartilage (2017)
2 Hunt, et al., ‘Main Cause of Death Following Primary Total Hip and Knee Replacement for Osteoarthritis. A Cohort Study of 26,766 Deaths Following
332,734 Hip Replacements and 29,802 Deaths Following 384,291 Knee Replacements’, The Journal of Bone and Joint Surgery (JBJS), (2017).
Part 1
NJR Data
Quality Audit
1.1 NJR Data Quality CEO received correspondence from the NJR inviting
them to join the NJR’s data checking programme and
Audit update identify a data quality lead to help hospitals assess

data completeness and quality for hip and knee
procedures submitted to the NJR for the previous
In 2015, the NJR began a retrospective data quality financial year, 2014/15.
audit of hip and knee procedures performed during
the financial year 2014/15 (1 April 2014 to 31 March Once identified, the Trust or Health Board’s data quality
2015) in NHS hospitals. This was the inaugural year of lead was then contacted to obtain data on eligible
the programme. By comparing unit data from the local arthroplasty procedures from the local hospital Patient
hospital Patient Administration System (PAS) with the Administration System (PAS), which was subsequently
data entered in to the NJR, we aimed to investigate the linked to procedures uploaded from local Trusts and
compliance of NHS hospital Trusts and Health Boards’ Health Boards to the NJR, comparing record for record.
reporting of arthroplasty procedures to the registry.
Consistency between the NJR and hospital’s PAS was
Unlike many other national audits, there are two assessed. The process involved Trusts and Health
principal outcomes of interest to the NJR: 1) mortality, Boards returning a file of patients for whom OPCS4
and 2) revision surgery. Similar to other registries, codes had been locally recorded that suggested they
data on mortality is collected via the Office of National had had a primary or revision hip or knee replacement in
Statistics (ONS) and linked to individuals within the the financial year 2014/15. This was matched by the NJR
NJR. However, more consistent with the primary aims against all the joint replacements that the organisation
of the NJR, revision surgery (an indication of implant had submitted to the NJR for the same timeframe. The
failure or surgical performance) is determined by linking possible outcomes for each record were:
a primary arthroplasty procedure to a secondary
procedure, which typically occurs a number of years a. A full match by patient ID, operation date and
after the primary procedure. Therefore, compliance with procedure (recorded OPCS4 codes and in the NJR)
reporting revision surgical procedures is essential to
b. Recorded OPCS4 codes but no NJR record
estimate implant failure rates and surgical performance
identified
more accurately.
c. NJR record identified but no corresponding record or
Confidence in the NJR is based on the assumption
OPCS4 codes in PAS
of high quality data, robust analysis and strong
engagement with stakeholders (for example; surgeons, The total number of procedures identified between the
patients, healthcare providers, implant manufacturers, local hospital PAS extract and those uploaded to the
and the MHRA). Whilst the NJR is fully engaged with NJR was considered as the denominator for calculations.
stakeholders and conducts robust statistical analysis,
the completeness of data within the NJR is reliant Details of the unmatched records were returned to
on the input at unit level, which is subject to variation the participating hospitals for further analysis. The
across trusts and health boards. audit was completed for that organisation when the
outcome of this analysis, and any necessary corrections
It is clear that for surgeons and patients alike, the or submission of omissions, were received back by
necessity for having accurate and complete data is an the NJR. On completion of each audit, a NJR Audit
absolute requirement. Data quality and validation are Compliance Report was created and sent to the CEO
essential components of any audit or scientific research. which contained the key findings, recommendations
Quite simply, if the data is incomplete or incorrect, then and additional learning points from the audit process.
false conclusions may be drawn from any analysis. This report provided each Trust and Health Board with
their own key learning points to act upon.
Methodology
All (149) NHS Trusts and Health Boards who report to The NJR sent repeated communications to Trusts or
the NJR were selected for audit. In July 2015, each Health Boards which were slow or unengaged with
the audit and arranged external hospital visits where investigation by the units, procedures were found to have
necessary. The external visits allowed the NJR to been outsourced (25%), identified as not being an NJR
understand and rectify the blockage in communication procedure (15%), an incorrect patient identifier had been
and engagement with the audit, as well as supporting used (12%), or another reason (10%). The remaining
hospital colleagues with the audit process. Alongside 5,332 (38%) records were eligible for NJR entry.
this and in the first instance, NJR Regional Clinical
Coordinators worked to identify key individuals at each Regarding outcome (c), a total of 7,658 procedures
Trust or Health Board to resolve any on-going issues. were found to be on the NJR but not on the hospital
PAS system. This is of particular significance in the
Results current financial climate. As a simplified example, at
around £5,000 per procedure this equates to over
Five months after the audit commenced with the initial
£38million in potential lost revenue across the 115
contact with chief executives, only one Trust had failed
Trusts and Health Boards.
to engage in the audit process. A total of 119 (80%)
Trusts and Health Boards had completed the audit In summary, a total of 101,936 (50,550 hips, 49,686
or were due to receive their NJR Audit Compliance knees, and 1,700 which could not be defined)
Report. The remaining 29 (19%) were addressing procedures were identified between the NJR and local
unmatched records, and by April 2017 only eleven PAS extract. Of these, 89,956 (44,083 hips, 45,873
trusts had remaining unmatched records to resolve knees) were indicated to be primary procedures, and
from the FY14/15 audit. 10,280 (6,467 hips, 3,813 knees) were revisions.
Of the 115 Trusts and Health Boards that had 95.30% (94.88% hips, 95.70% knees) and 90.95%
completed their audit by the end of December 2016, a (91.33% hips, 90.32% knees) of primary and revision
total of 96,604 procedures were matched between the procedures were recorded in the NJR respectively.
NJR and local PAS extract (outcome (a)).
14,258 procedures were found in the hospital PAS

system but not on the NJR (outcome (b)). On further
Found in hospital PAS but Percentage of missing

Procedure All records not on the NJR records
© National Joint Registry 2017

Hip Primary 44,083 2,259 5.12%
Hip Revision 6,467 561 8.67%
Knee Primary 45,873 1,973 4.30%
Knee Revision 3,813 369 9.68%
Undefined 1,700 170
Total 101,936 5,332
Note: All records represent the total number of eligible procedures found between the NJR and local hospital PAS; missing NJR records represent cases where the
eligible PAS record provided could not be found on the NJR; undefined records refer to PAS records where the joint was not specified.
The audit has found data representing 5.23% of all Comparison between the published compliance figures
records missing (entered into PAS but not into the (based on Hospital Episode Statistics and Patient
NJR). However, it is also noted that the proportion of Episode Database for Wales data) and the compliance
missing data is greater for revision procedures (8.67% found during the data quality audit which compared
for hips and 9.68% for knees). record for record, shows the median dropping slightly
from 97% (IQR 84.5-101.5) to 95.37% (IQR 91.30- At the time of writing, the audit for the financial year
97.94). Compliance shows a significantly smaller range 2015/16 is concluding and was expanded to include
following the data quality audit which gives a greater the independent (private) sector whose submission
confidence in the reported figures. to the NJR has been mandatory since the registry’s
inception. Early responses to the FY2015/16 audit
The audit also enabled compliance to be measured showed a positive engagement by independent units
on a case by case basis (matching individual with strategic involvement at Independent Healthcare
records across NJR and HES/PEDW), and for this Provider Group level. Trusts and Health Boards have
to be compared to published compliance levels for been able to build upon their experience of the previous
each. Published compliance levels provide a crude year’s audit to improve their processes and datasets.
assessment as they simply compare total procedures
undertaken within a timeframe between HES/PEDW The NJR was also able to use the lessons learnt from the
and NJR. Compliance measured on a case by FY14/15 audit to better identify the data quality leads for
case basis provides a more accurate measure. This each unit and track these contacts and communications.
assessment highlights compliance for a number of Improvements were made to the previous audit tool and
Trusts and Health Boards to be significantly different the introduction of a data request template resulted in
using the case by case method, than that currently cleaner data being returned.
published, with some appearing better and some worse
than the published compliance levels would suggest. Conclusion
High quality complete data is essential for making
Outcomes robust inferences from the NJR. Systematic under-
Whilst the overall scale of missing records was found reporting of revision procedures is likely to bias
to be low (5.23%), the proportion of missing revision results and reduce the statistical power of the NJR to
records was found to be higher than that for primary quickly detect failing implants at higher than expected
procedures. The observed differences suggest rates. Although this is true, the large size of the NJR
systematic under-reporting of revision procedures in the somewhat compensates for this when assessing failure
audited Trusts and Health Boards. The ratio of missing rates at a national level. However, when attempting to
primaries for hip and knee replacements is approximately sub-divide data by surgeons, the reduction in statistical
20:1 and 23:1 respectively, whereas the ratio of missing power and systematic under-reporting of revisions, may
revision procedures is 12:1 and 10:1 respectively. be misrepresentative of individual surgeon performance.
When compliance is considered at a Trust or Health In other words, and to put the importance of this into
Board level, variation in compliance is substantial with context, if data is missing at random, then comparisons
high levels of under-reporting of revision procedures of the NJR data at the level of the implant may still be
by specific units. The failure of local hospitals to valid, but comparisons of sub-samples of the NJR, such
upload revision procedures into the NJR is especially as surgeon or hospitals, are much more problematic.
problematic, as linked revision procedures form the
basis of analyses which investigate implant failure and Further investigation is required to ascertain whether
surgical performance – which fundamentally underpins these are random events or a systematic under-
the primary aim of the NJR. reporting of revision procedures. Analysis of the
audit’s results in year two will help this and that work is
Audit FY2015/16 currently underway. Results of both the FY2014/15 and
FY2015/16 audits will be updated via the NJR’s main
The NJR believes that a minimum three year data
website – www.njrcentre.org.uk.
quality audit programme is required before the NJR and
its stakeholders have a greater degree of confidence in
the data and its quality.
Part 1
Annual progress
1.2 Annual Report • Linkability (the ability to link a patient’s primary
procedure to a revision procedure) was recorded as
Introduction 94.2%, a drop of 1% on the previous year.
Whilst a comparison of successive years will show

The 14th Annual Report of the National Joint Registry variation, the drop in the rates of the key indicators
for England, Wales, Northern Ireland and the Isle of consent and linkability may be attributable to the
of Man (NJR) is the formal public report for the outcomes of the data quality audits that have taken
period 1 April 2016 to 31 March 2017. The report place this year. This has resulted in the retrospective
consists of a number of parts which are outlined in submission of missing procedures for which some
the summary table. will not have had patient consent recorded. Linkability
is dependent on the submission rate of NHS and, in
As part of the continued approach to sharing Northern Ireland, HCN numbers. Please see the data
information about NJR progress, clinical activity and completeness and quality indicators section online for
hospital and implant activity, the NJR has updated further detail.
the data on its dedicated annual report website,
‘NJR Reports’, to showcase annual report data and Data quality has continued to be a primary focus
information. for the NJR in 2016/17 with the undertaking of the
second year’s data quality audit across all NHS units
Some of these data can also be found in this printed and, for the first time, independent sector units. The
report – in particular, the summaries and the full established NJR Data Quality and Clinical Leads at
detailed, statistical analysis of outcomes following joint all Trusts and Health Boards have worked with the
replacement surgery. audit team at unit level, resulting in a swifter response
and improved quality of data. Engagement with the
A short summary of the NJR’s progress over 2016/17
independent sector has also been very encouraging.
is included below and in both in the Chairman’s
Please visit www.njrreports.org.uk for further details of
Foreword and Annual Report Executive Summary.
the audit.
Additional information and reports are available online
Further enhancements to the NJR’s reporting services
via ‘NJR Reports’ at: www.njrreports.org.uk.
have been made in 2016/17. Surgeons are now able
to access more information through NJR Clinician
1.3 Annual Progress Feedback, monitor their patients through a report on
both primary and revision procedures and also, within
subscribing Trusts and Health Boards, gain access to
As at 31 March 2017, the total number of procedures
implant pricing reports. NJR Management Feedback
submitted to the NJR was approximately 2.35 million.
continues to issue an annual report to summarise
In the financial year 2016/17, a total of 242,629
activity and outcomes at each hospital within a Trust,
records were submitted which is an increase of
Health Board or organisation and offers a free price
20,857 over the previous year. This is despite
benchmarking service to units providing implant
concern that the overall number of joint replacement
pricing information.
procedures being undertaken was decreasing. Overall
key performance indicators demonstrated: Finally, the NJR remains committed to working for
patient safety and driving forward quality in joint
• Patient consent (to allow the recording of their
replacement surgery.
personal details in the NJR) was recorded as
92.1%, a decrease of 1.2% from the previous year. Further progress and updates will be available at
However, the consent rate for Northern Ireland www.njrreports.org.uk and also via the main NJR
increased to 96.1% from 94.5% in the previous year, website at www.njrcentre.org.uk.
while the overall consent rate for England, Wales and
the Isle of Man decreased by 1.2%.
1.4 Summary of content for the NJR Annual Report

Section Summary Content Full information can be found
Executive summaries, News and information in executive summaries,

Part One annual progress and committee reports and highlights about the www.njrreports.org.uk
FY2016/17 highlights progress of the NJR to 31 March 2017
Statistics on joint replacement activity for hip,

www.njrreports.org.uk through
Part Two Clinical activity 2016 knee, ankle, elbow and shoulder activity for the
interactive reporting
period 1 January to 31 December 2016
Detailed statistical analyses on hip and knee

replacement surgery using data from 1 April 2003
Outcomes after joint to 31 December 2016. Updated analyses of
In this printed report and via
Part Three replacement surgery primary ankles and shoulders representing data
www.njrreports.org.uk
2003-2016 collected since 1 April 2010 and 1 April 2012
respectively. Analyses on provisional data for
elbows using data collected since 1 April 2012
Indicators for hip and knee joint replacement

procedures by Trust, Local Health Board and
Implant and unit-level
Part Four unit. Plus commentary on implant performance www.njrreports.org.uk
activity and outcomes
and those that have higher than expected rates
of revision and were reported to the MHRA
Use of prostheses by Prostheses used in joint replacement surgery

Prostheses www.njrreports.org.uk
brand (implants) 2016 for hip, knee, ankle, elbow and shoulder
Information relating to the Composition, attendance, declarations of

NJR’s governance and interest for the NJR Steering Committee,
operational structure sub-committees and terms of reference
Appendices www.njrreports.org.uk
Published and approved research papers using
Research
NJR data
Part 2
Clinical activity
2016 and using
the dedicated NJR
Reports website
2.1 Clinical activity 2016 overview

Part Two of the NJR’s 14th Annual Report • Types of primary replacements undertaken
can now to be found online via the registry’s • Patient characteristics for primary replacement
dedicated NJR Reports website at: procedures, according to procedure type
www.njrreports.org.uk.
• Age and gender for primary replacement patients
Part Two presents data on clinical activity during • Patient’s physical status classification (ASA grades)
the 2016 calendar year. This includes information for primary replacement procedures
on the volumes and surgical techniques in relation • Body Mass Index (BMI) for primary replacement
to procedures submitted to the NJR, with the most patients
recent data being for the period 1 January 2016 to
31 December 2016. To be included in the report all • Indications for primary procedure based on age
procedures must have been entered into the NJR by groups
28 February 2017. • Age of patients undergoing primary joint replacement
• Surgical technique for primary replacement patients
The following double page spread offers a visual
summary of key facts relating to clinical activity during • Thromboprophylaxis regime for primary replacement
the 2016 calendar year. This can also be downloaded patients, prescribed at time of operation
as a waiting room poster via www.njrreports.org.uk. • Reported untoward intra-operative events for
primary replacement patients, according to
The information in Part Two now includes historical procedure type
data, going back to 2005 in most cases. Using the
• Patient characteristics for revision procedures,
dedicated website, readers are able to use interactive,
according to procedure type
filterable graphs to identify the key information and
trends associated with the following reports for hip, • Indication for surgery for revision procedures
knee, ankle, elbow and shoulder data (where sufficient • Trends in use of the most commonly used brands
data are available):
For hips specifically
• Total number of hospitals and treatment centres
in England, Wales and Northern Ireland able to • Components removed during hip revision procedures
participate in the NJR and the proportion actually • Components used during single-stage hip revision
participating procedures
• Number of participating hospitals and the number • Trends in femoral head size and hip articulation
and type of procedures performed
• Number of procedures undertaken as a proportion For knees specifically
of all procedures submitted annually • Implant constraint for primary procedures
• Procedure details by type of provider • Bearing type for primary procedures
• Primary procedure details by type of provider
Summary of key facts about joint replacement during the 2016
Hips
NJR Patien
Consent
t
101,651
replacement
60%
average ages:
NJR Patien
t
procedures
Consent
recorded on the NJR

3.5%
since April 2003 (98,211 in 2015) 67.6 69.8
Knees
NJR Patien
Consent
t
108,713
replacement
56%
average ages:
procedures
NJR Patien
t
Consent
3.8%
recorded on the NJR
since April 2003 (104,695 in 2015) 69.2 69.6
Shoulders
NJR Patien
t
6,967
replacement
70%
Consent
average ages:
procedures
12.9%
recorded on the NJR
since April 2012 (6,170 in 2015) 69.2 73.9
/nationaljointregistry @jointregistry
calendar year
Elbows
average
NJR PatBMI
Consen
ient
t
722
28.8
NJR Patien
t
90% Consent
replacement
osteoarthritis
procedures
= recorded on the NJR 10.2%
‘overweight’ since April 2012
Diagnosis (655 in 2015)
72% 31%
inflammatory
arthropathy
average ages:
average BMI 19%
osteoarthritis
99%
osteoarthritis 31.0 60.3 67.6 Diagnosis
= Ankles
Diagnosis
‘obese’
NJR Patient
Consent
839
replacement
procedures
15.6%
55% recorded on the NJR
since April 2010
(726 in 2015)
osteoarthritis
26% 61% 89%

osteoarthritis
cuff tear average ages: 7%

arthropathy
rheumatoid arthritis and
other inflammatory
Diagnosis 68.3 66.5
joint problems
Diagnosis
For more data on clinical activity during the 2016 calendar year visit www.njrreports.org.uk.
2.2 Navigating the NJR Reports online facility
What can you find at NJR Reports online?
As at 31 March 2017, the total number of procedures recorded in the NJR is now approximately
2.35 million.
The NJR has refreshed its dedicated online annual report website – NJR Reports – to showcase annual report data
and help users easily navigate the growing wealth of information collected about joint replacement procedures.
Part Two of the NJR’s 14th Annual Report presents data on clinical activity during the 2016 calendar year. Simply
navigate the left hand tabs to view information on the volumes and surgical techniques in relation to procedures
submitted to the NJR.
Top tabs: If you require

information about
specific procedures, go
Left hand tabs: Here, the straight to the data by
information is segregated clicking on the joint type
by report and information most relevant to you.
type. A wealth of updates
are available, from Executive
Reports including from the
NJR’s Steering Committee
Chairman, to Executive
There is also implant
Summaries on clinical
and hospital specific
activity and outcomes data, information available,
and highlights from the year. a glossary and
downloadable patient
guides to make all
the information as
Visit the NJR Reports website at: accessible as possible
www.njrreports.org.uk to all of our visitors.
Part 3
Outcomes
after joint
replacement
2003 to 2016
3.1 Executive
summary
Part Three of the 14th Annual Report provides percentage probability of revision of 4.3% at 13
outcome data in relation to hip, knee, shoulder, elbow years, with best results within group seen when a
and ankle replacements. It describes activity between ceramic-on-polyethylene bearing is used (3.8%). The
1 April 2003 and 31 December 2016. survival within the entire hybrid group is calculated at
5.1%, with ceramic-on-ceramic bearings providing
There were 2,284,416 procedures recorded in this best results of any sub-group at 3.3% at 13 years.
period, although 10% of these were excluded because Reassuringly the most commonly used cemented and
there were insufficient patient details to enable linkage. hybrid constructs by brand all perform well.
This relates predominantly to the early years of the
registry and was less of a feature in recent years as The total number of primary
data quality has improved.
hip replacements performed
The numbers of primary procedures available for continues to increase with 87,733
analysis were 890,681 total hip replacements, 975,739 performed in 2016, compared to
knee replacements, 3,899 ankle replacements, 23,608
86,496 the previous year. The vast
shoulder replacements and 2,196 elbow replacements.
majority continue to be performed
Hip replacement procedures for osteoarthritis
The total number of primary hip replacements
performed continues to increase with 87,733 For the uncemented construct group the pattern
performed in 2016, compared to 86,496 the previous of failure over time is different. The revision rate is
year. The vast majority continue to be performed for approximately double that of all cemented, calculated
osteoarthritis. In 2016, the ratio of women to men at 8.7%. Within group, the best survival figures are
receiving hip replacement was 60:40 and the median seen with a ceramic-on-polyethylene bearing, with
age at which primary surgery was performed is 69. survival rate improving to 4.5%. If metal-on-metal
bearings are excluded, the commonly used constructs
Uncemented fixation is still the most common by brand perform similarly well.
construct used by surgeons comprising 39% of
the total number, compared to 30% for cemented In this year’s analysis the effect of age and gender on
replacements. There has been a slight decrease in revision rates across the construct groups has been
both these construct types whilst hybrid fixation, presented for the first time. This is particularly relevant
mainly using a cemented stem and uncemented given the increase in total numbers of younger
cup, continues to grow in popularity, with surgeons patients undergoing joint replacement. Overall, as
using this method in 28% of cases. Metal-on- reported in previous annual reports, the revision rate
polyethylene is still the most commonly used bearing for total hip replacement increases at a faster rate
construct across cemented, uncemented and hybrid over time for younger patients. For female patients
hip replacements, but the usage of ceramic-on- under 55 the revision rate of 13.5% at 13 years is 2.5
polyethylene bearings continues to grow, reaching times greater than for women undergoing surgery
29% of all cases. Metal-on-metal bearings including between 65 and 74 years of age. However the choice
resurfacing is performed in very low numbers making of construct does affect revision rate in the younger
up less than 1% of all cases in 2016. age group and for women under 55 years a cemented
ceramic-on-polyethylene construct gives the best
In this year’s report, a total of 890,681 recorded hip results, with a revision rate of 3.8% at ten years. A
replacements were available for survival analysis, similar trend in the relationship of age to revision rates
with data collected over 13 years. The cumulative is seen for men, although at 13 years the revision
percentage probability of revision after primary hip rate for the under 55 group across all bearing types
replacement across all patients is 6.8% at 13 years. is 10%, approximately 3.5% lower than for women.
The lowest rate of revision continues to be seen in Again, the best performing construct for the younger
the all cemented construct group, with a cumulative patient is a cemented prosthesis, in this case using a
ceramic-on-polyethylene bearing, which provides half that revision rates are similar to those hip replacements
the revision rate of cementless fixation using metal-on- performed for other indications although, as expected,
polyethylene at all time points. Interestingly, for older mortality rates are higher.
patients all construct combinations have similarly good
revision outcomes. In 2016, 7,933 revision procedures were performed,
with the vast majority being single-stage procedures.
Presenting mortality data alongside revision outcomes The total number of revision procedures available
provides a greater understanding of the outcome of hip for analysis between 2003 and 2016 is now 97,341.
replacement, particularly in the older patient. In the vast The most commonly recorded indication for revision
majority of patients over the age of 75 at implantation, continues to be aseptic loosening, followed by
their hip implant will remain unrevised across their pain. Within the first year following primary surgery
remaining lifetime, with very low revision rates seen. dislocation, fracture and infection are the most
common indications for revision, whereas revision for
Our analysis confirms that choice of head size is an aseptic loosening increases in frequency over the first
important factor in determining revision outcome. ten years. The cumulative probability of hip re-revision
For both metal-on-polyethylene and ceramic-on- is approximately 17% at 13 years.
polyethylene bearing choices higher failure rates are
seen with larger head sizes, in particular 36mm for Knee replacement procedures
cemented and above 36mm for hybrid and cementless.
Between 2003 and 2016 a total of 975,739 knee joint
In contrast if a ceramic-on-ceramic bearing construct is
replacements were recorded and are available for
used then survival is improved with larger size.
analysis. Osteoarthritis remains the most common
indication for knee replacement across the whole
In this year’s analysis the effect of
cohort (96%), with the second most common
age and gender on revision rates indication being inflammatory arthritis at 2%.
across the construct groups has
been presented for the first time. During 2016, 104,079 knee joint replacements were
recorded in the NJR, with 98,147 primary and 5,932
This is particularly relevant given revision procedures. Within primaries, the most
the increase in total numbers common type of reconstruction performed was a total
of younger patients undergoing knee replacement, making up 89.7% of procedures.
joint replacement Of this group, the most widely used fixation method
remains cementing (84.9%). Uncemented total knee
replacement continues to decline in numbers, making
Metal-on-metal reconstructions, of either a resurfacing
up only 2% of the total number implanted. As seen
or stemmed variety, continue to fail at higher rates than
in other years, within the cemented group of total
other bearing choices, with revision rates ranging
knee replacements fixed bearing unconstrained
between 14% and 27% at ten years for the worst
(62.2%) and posterior-stabilised (19.8%) make up
performing implant types. However the survival profile
the vast majority of implantations performed. The
for the best performing resurfacing procedures by
proportion of unconstrained to posterior-stabilised
brand shows lower revisions rates of between 8% and
has remained steady over the last five years, at a ratio
9%. Overall the net effect of higher revision rates for
of 3:1. Unicompartmental knee replacement (medial
metal-on-metal procedures has been a dramatic and
and lateral) makes up 9.2% of all knee replacements
sustained reduction in their use.
performed in 2016, with this percentage remaining
The number of patients who are treated with primary fairly static over the last ten years. A mobile bearing
hip replacement after sustaining a fractured neck of construct is used in 5.1% of cases and fixed in 4.1%.
femur continues to grow with time. In 2016, 4,260 Patellofemoral replacement account for 1.1% of
were performed, representing 4.9% of all total hip all knee replacements and similarly this figure has
procedures. In this group of patients it is encouraging remained static for the last ten years.
Between 2003 and 2016 a total of fixed bearing choice. Patellofemoral joint replacement
continues to record the highest failure rate, with the
975,739 knee joint replacements
current estimate being 24.2%.
were recorded and are available
for analysis. Osteoarthritis remains For younger patients, the risk of revision is higher
the most common indication for with the same pattern seen for men and women. For
a patient at the median age of implantation (69), the
knee replacement across the 13-year risk of revision is just over 4%. However, for
whole cohort total knee replacement patients under the age of 60,
the risk increases with decreasing age, reaching 10%
for those under 55 years old. This pattern is magnified
Patient demographics showing the trend for more
in unicondylar replacement, with patients under the age
women than men to undergo knee replacement
of 55 facing a 25% chance of revision by 13 years. This
continues in all types of knee replacement. The median
has been a consistent finding across all annual reports.
age at which patients undergo replacement is 70 years
for total knee replacement, 64 for unicondylar knee
replacement and 58 for patellofemoral replacement. The median age at which patients
Over the last three years, 1,999 surgeons have undergo replacement is 70 years
undertaken total knee replacements and 820 performed
for total knee replacement, 64 for
unicondylar knee replacements. The median number
of each performed over a three year period is 104 unicondylar knee replacement and
(IQR 26-214) for total knee replacements and 12 (IQR 58 for patellofemoral replacement
3-35) for unicondylar replacements. This highlights the
continuing trend for some surgeons to perform very few In 2016, 5,932 revision knee replacements were
numbers of unicondylar replacements per year. performed, with the vast majority being single stage
revisions. The total number of revisions across all years
Survival analysis performed on the 975,739 knee was 60,680. The most common indications recorded
replacements in the NJR was completed out to 13 for first revision surgery in total knee replacement
years. Temporal changes over time show that the remain aseptic loosening, pain, infection and ‘other’
rate of change of cumulative percentage chance of (excluding dislocation, lysis, periprosthetic fracture,
revision has remained similar over the period between implant fracture, instability and malalignment).
2003 and 2013. Indications for first revision surgery in unicompartmental
knee replacement follow a broadly similar pattern, with
The cumulative risk of revision at 13 years for
aseptic loosening and pain remaining as the most
cemented total knee replacement is 4.2%, with
common specific reason, although rates are higher.
unconstrained fixed bearing total knee replacement
Considering all knee replacements within the first year
(the most common construct) recording 3.8% and
of implantation, infection remains the most common
posterior-stabilised total knee replacement 4.7%. In
cause of revision, with aseptic loosening becoming
the cementless class, the figure reached 5.4% but
more common in later years. The risk of subsequent
interestingly for uncemented fixed bearing posterior-
re-revision is approximately 16% at 13 years across this
stabilised total knee replacements the revision rate
entire group.
reaches 12.1% by 13 years, demonstrating that this
combination of implant choice puts patients at a Ankle replacement procedures
greater risk of revision.
In 2016, there were 690 primary ankle replacements
Unicondylar replacement revision rates are higher than entered into the NJR, compared to 602 the year
those for total knee replacements across all times before. Similar data has been collected from 2010 to
points, with a rate of 16% reached by 13 years post- 2016 and in total 3,899 primary ankle replacements
surgery. The trend is the same regardless of mobile or are available for analysis.
From the entire series the ratio of female to male In 2,038 cases the indication was acute trauma. In
patients was 60:40 and the median age at primary this group of patients, 77% were female and 23%
surgery was 68, with a range of 17 to 92 years. male, with a combined median age at surgery of 74
(IQR 67-80 years).
Of the 3,899 primary procedures, the vast majority
(89%) of implantations have been uncemented and, The most frequently used implant type is the reverse
with the exception of three recorded hybrid cases, the polarity total shoulder arthroplasty (42%), followed
remaining are cemented. by total conventional shoulder arthroplasty (30%)
and hemi-arthroplasty (13%). The vast majority of
A total of 229 consultants, working in 269 units these primary cases were stemmed. Resurfacing
carried out these procedures, with 44% of surgeons humeral hemi-arthroplasty or resurfacing total shoulder
performing over ten procedures and 56% less than arthroplasty was performed in 13% of cases.
ten, over the six year period.
The cumulative percentage probability of revision
Between 2010 and 2016, there were 153 revision at four years for elective primary cases was 4.2%
procedures, including 24 conversions to arthrodesis. and 3.9% for trauma cases. In elective cases, the
The estimated rate of revision at six years was 7.7% rate increases for patients under the age of 65 to
(95% CI 5.94-8.47). 7.6% in men and 6.4% in women. After four years,
total conventional shoulder arthroplasty and reverse
In 2016, the Infinity (30%), Box (18%) and Zenith (15%)
shoulder arthroplasty had the lowest revision rate for
were the most widely used brands, making up over
elective shoulder replacement, although caution in
half of all implantations.
interpretation is required as the differences seen may
Shoulder replacement procedures reflect threshold for revision and do not take account
of functional outcome. In both elective and trauma
There are now 23,608 primary shoulder replacements in procedures, the most common causes of revision
the NJR with 5,944 procedures performed in 2016, with were instability and cuff insufficiency.
the number performed each year continuing to increase.
In 2016 these procedures were performed in 338 units, In both elective and trauma
with 12 as the median number per unit (IQR 5-23). The
total number of consultants performing the procedures
procedures, the most common
was 476, with a median per consultant of 9 (IQR 4-18). causes of revision were instability
and cuff insufficiency
There are now 23,608 primary
shoulder replacements in the NJR Elbow replacement procedures
with 5,944 procedures performed in A total of 2,196 primary elbow replacement
2016, with the number performed procedures have been recorded in the NJR between
April 2012 and December 2016, including total, radial
each year continuing to increase head and lateral resurfacing replacements. In 2016,
a total of 513 procedures were performed, which is
A total of 21,570 cases were performed as part of a slight decrease from the year before, although the
elective care. The most common indications for surgery general trend since 2012 has been an increase in
were osteoarthritis and cuff tear arthropathy, sometimes overall numbers recorded.
combined in a small proportion of patients (522). For
elective cases the majority of the replacements were From the entire series of 2,196 procedures, women
performed on women (70%) and the median age at the (70%) undergo elbow replacement more often than
primary operation was 73 years (IQR 67-79 years), with men (30%) and the median age of patients undergoing
an overall range of 17-99 years. surgery was 68 (IQR 58-77 years). Trauma accounted
for 31% of all cases. In the 1,511 elective cases, the At three years, the cumulative percentage probability
most common stated indications for elective surgery of revision, across the entire group, was 4.4% (95% CI
were other inflammatory arthropathy, osteoarthritis 3.3-5.8). In trauma cases the probability of revision was
and sequelae of trauma. 2.1% (95% CI 1.1-4.2), but no radial head replacements
were revised and the revision rate for total replacement
was 3.2% (95% CI 1.6-6.3). This contrasts to a three-
A total of 2,196 primary elbow
year revision rate of 5.1% (95% CI 3.7-7.0) when total
replacement procedures have been replacement was performed in the elective setting. The
recorded in the NJR between April most frequently cited causes of revision in elective care
2012 and December 2016 were infection and aseptic loosening.
Total prosthetic (63%) and radial head replacement From the entire series of 2,196
(37%) were the prostheses used in trauma cases. In procedures, women (70%) undergo
elective care the vast majority of cases were total elbow replacement more often than
replacement (95%), with radial head replacement men (30%) and the median age of
performed in 4% of cases and lateral resurfacing in 1%.
patients undergoing surgery was 68
In 2016, there were 210 consultants working in 160
units. They undertook primary elbow replacements with
2 (IQR 1-4) as the median number of cases performed
per unit and 2 (IQR 1-3) the median per consultant.
Part 3
3.2 Summary
of data sources
and linkage
The main outcome analyses in this section relate to As of April 2016, 80% of Trusts and Health Boards
primary joint replacements. We included all patients had completed the audit, with the remaining actively
with at least one primary joint replacement carried out engaged in completing the audit. Although it is possible
between 1 April 2003 and 31 December 2016 inclusive, that some records may have been missed in the audit
whose records had been submitted to the NJR by 28 process, or subsequently entered, we believe this
February 2017. number is small.
Information governance and patient Whilst the proportion of missing data in the NJR is
confidentiality: relatively small, the propensity to not record revision
procedures is problematic and will lead to a reduction
NJR data is collected via a web-based data entry in power to detect trends. From a national perspective,
application and stored and processed in Northgate we believe selective under-reporting of revisions would
Public Services’ (NPS) data centre. NPS is ISO 27001 apply across all types of hip and knee replacements in
and ISO 9001 accredited, and compliant with the NHS’ a random pattern, and therefore would not affect the
Information Governance Toolkit. Data linkage to other group comparisons we make.
datasets is approved by the Health Research Agency
under Section 251 of the NHS Act 2006. Please visit Patient level data linkage:
www.hra.nhs.uk/about-the-hra/our-committees/
section-251 for more details. Documentation of implant survivorship and mortality
requires linkage of person-level identifiers, this enables
Data source: the identification of primary and revision operations
within the same individual.
In the early years of the registry, when reporting was
not mandated by the Department of Health, we know a Starting with a total of 2,284,416 NJR source records,
number of primary procedures were not recorded in the 9.3% were lost because no suitable person-level
NJR, as indicated by discrepancies between implant identifier was found (see Figure 3.1 (a)). In around half
levies and procedure rates. In the subsequent years, of these 213,441 procedures (47.8%), the patient had
selective reporting of primary and revision operations declined to give consent for details to be held or consent
may explain temporal increases in volume (primary and was not obtained, the remainder being attributable to
revision), and revision outcomes for hips and knees tracing and linkage difficulties. Cases from Northern
replacements (see sections 3.4 and 3.6). Ireland were excluded at this step because there was
no tracing service available for them. Although a person-
More recently primary procedures are less likely to level identifier was available for 95% of operations since
have been missed. The recent 2014/15 NJR data the beginning of 2008, in earlier years, the proportion
completeness and accuracy audit across 149 NHS trusts had been much lower (see Figure 3.1 (b)). In 2003/4 for
reporting to the NJR suggests we may have missed example, it was only 59%, rising to 79% in 2006 and
about 5% and 4% of hip and knee primaries respectively. 90% in 2007 (see Figure 3.1 (b)). Therefore, patients
with longer follow-up might be less representative of
What is of more serious consequence to our analyses; is
the whole cohort of patients undergoing primary joint
the differential and selective under-reporting of revision
replacement than those patients with shorter follow-up,
procedures associated with the primaries that have been
due to difficulties with data linkage.
entered, this could lead to reported revision outcomes
looking better than they actually are. This issue is being Among the linkable procedures with person-level
addressed by the NJR’s Data Quality Sub-committee. identifiers (2,070,781) there were 90,095 (4.4%) revision
Similarly, the 2014/15 data completeness and accuracy procedures within the analysis period (2003 to 2016)
audit suggested 9% and 10% of hip and knee revisions with no associated primary operation recorded in the
had been missed during this period respectively. NJR. This would have been either because the primary
had taken place at an earlier point in time (before the As in previous years, the unit of observation for all
NJR data collection period began in 2003) or was not sets of survivorship analysis has been taken as the
included for other reasons such as the operation being individual primary joint replacement. A patient with left
performed outside the geographical catchment area of and right replacements of a particular type, therefore,
the NJR, or consent for data linkage not being provided will have two entries, and an assumption is made
at the time of the primary procedure. At the joint level, that the survivorship of a replacement on one side is
some further revisions were excluded because they independent of the other. In practice, this would be
could not be matched to primary joint replacements, i.e. difficult to validate, particularly given that some patients
if a primary operation was recorded only for one side and did not have prior replacements recorded in the NJR.
there was only a documented revision for the other side, Established risk factors, such as age, are recorded
the latter was excluded. However, we have included at the time of primary operation and will therefore
these ‘unlinked’ revisions in our general overview of be different for the two procedures unless the two
outcomes after revision, see Sections 3.4 and 3.6. operations are performed at the same time. Patients
may also have more than one type of implant.
Linkage between primaries and any associated
revisions (the ‘linked files’): Within the NJR, a revision is defined as any operation
in which any prosthesis or part of a prosthesis is either
A total of 1,574,146 patients had at least one record removed, exchanged or inserted for any reason into
of a primary joint replacement within the NJR, i.e. hip, a joint in which there is an existing joint replacement.
knee, ankle, elbow or shoulder. At this stage, information This therefore not only includes complete replacement
about the primary procedures were linked to subsequent of one or both of the main components of any joint
associated revisions (i.e. for the same patient-joint-side). replacement, but also, for example, liner and/or head
Further data cleaning was carried out at this stage, for exchange at washout for suspected infection and
example, removal of duplicated primary information on secondary patella resurfacing of an existing total or
the same side or revision dates that appeared to precede unicondylar knee replacement.
the primary procedure, leading to the final numbers for
analysis shown in Tables 3.1 and 3.2.
In Table 3.2, of the 767,965 patients with primary hip

operations, 16.0% had documented primaries for
both hips (bilateral). Of the 800,477 patients with knee
operations, 21.9% were bilateral.
Implant survivorship is mainly described with respect

to the lifetime of the primary joint only, i.e. we have
looked only at the time to first revision, not the time
from a revision operation to any subsequent one. These
analyses are described in sections 3.3, 3.5, 3.7, 3.8
and 3.9 for hips, knees, ankles, shoulders and elbows.
In sections 3.4 and 3.6, we provide an overview of

further revisions following the first hip or knee revision
procedure. We have also included revisions to a joint
replacement where the associated primary had not
been documented in the NJR.
Figure 3.1 (a)
Initial numbers of procedures for analysis.
2,284,416 213,441 (9.3%)

no patient identifier
procedures
2,070,975 194 removed with errors

linkable procedures that hindered linkage

(1 missing side; 7 missing dates;
1 with unknown operation; 86 with
primary prior to 1 April 2003; 99
‘deaths before procedure’)
2,070,781
linkable procedures
1,574,146
patient identifiers
HIPS: KNEES: ANKLES: SHOULDERS: ELBOWS:

895,292 980,286 3,915 23,735 2,207
primaries primaries primaries primaries primaries
97,569 60,818 465 2,768 672
revisions revisions revisions revisions revisions
(+1,633 reoperations) (+1,421 reoperations)
Figure 3.1 (b)

Total volume of uploads to the NJR, percentage of procedures consenting to be included in the NJR, and
percentage of patients traced in the NJR, in England and Wales* by year of operation.
250,000
90 100
200,000
80
70
60
Percent (%)
N(Upload)
150,000
50
40
100,000
30
20
10
50,000
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Year
Procedures Uploaded % Consent Yes & Don’t Know

% Traced & Linked
* Northern Ireland statistics excluded from figure due to lack of patient tracing agreement.
Table 3.1 Summary description of linked datasets used for main survivorship analyses.
Summary of data NJR data (England and Wales only)

All NJR procedure-level data restructed to person-level
1 April 2003 - 31 December 2016 (hips and knees)

Time period
1 April 2010* - 31 December 2016 (ankles)
1 April 2012* - 31 December 2016 (shoulders and elbows)
Excludes data where person-level identifier is not present
Data exclusions Excludes patients where no primary operation is recorded in the NJR
Excludes any revisions after the first revision
890,681 975,739 3,899 23,608 2,196
Number of primary operations
hips knees ankles shoulders elbows
NJR identified primary-linked first revisions
Number of primaries that were
24,103 24,399 153** 582*** 55****
subsequently revised
hips knees ankles shoulders elbows
*These were the dates when data collection formally started however the analyses in this section include a small number of primaries in the database that took place
before these time points.
**Ankle revisions include 24 conversions to arthrodesis.
***Shoulder revisions include two excisions and one conversion to arthrodesis
****Elbow revisions includes one excision.
Table 3.2 Composition of person-level datasets for main survivorship analysis.
Joints
Hips Knees Ankles Shoulders Elbows
Number of patients 767,965 800,477 3,739 22,313 2,134
Number (%) of patients with only 645,249 625,215 3,579 21,018 2,072

one primary joint operation (84.0%) (78.1%) (95.7%) (94.2%) (97.1%)
Number (%) of patients with
118,410 164,665 156 1,275 61
both a left and right side primary
(15.4%) (20.6%) (4.2%) (5.7%) (2.9%)
operation but on different dates
Number (%) of patients with both a
4,306 10,597 4 20 1
left and right side operation on the
(0.6%) (1.3%) (0.1%) (0.1%) (<0.1%)
same date (bilateral operations)
Total number of primary joints 890,681 975,739 3,899 23,608 2,196
Number with at least one revision
24,103 24,399 153 582 55
operation linked to the primary
Number with more than one
3,522* 4,208* 13 (4)** 64 (46)** 11 (6)**
revision procedure
*Discussed more fully in later sections: the numbers shown include some stage two of two-stage revisions.
**In some cases the first revision was the stage one of a two-stage revision; the numbers in parenthesis exclude cases where subsequent revision procedures
appeared to relate only to that first (i.e. either were just other another stage one or the respective stage two).
Part 3
3.3 Outcomes
after primary hip
replacement
This section looks at revision and mortality outcomes Details of the patient cohort are given in Tables 3.1
for all primary hip operations performed between 1 and 3.2 of the preceding section; a total of 890,681
April 2003 and 31 December 2016. Patients operated hips were included in our analyses.
on at the beginning of the registry therefore had a
potential 13.75 years of follow-up. Osteoarthritis was given as a documented reason
in 820,818 (92.2% of the cohort) and was the sole
reason given in 815,257 (91.5%) hip procedures.
Methodological note (probability) of revision, or death, at different times

from the primary operation. Where possible, the
Survival analyses have been used throughout,
numbers at risk at each anniversary have been
first looking at the need for revision and then
added to the figures. These are particularly useful
looking at mortality. Only the first revision has been
where a group has appeared to plateau; it may
considered in this section. The majority of implants
simply be because the number of cases fell so low
did not require revision and survival analysis made
that occurrence of further revisions/deaths became
use of the information that was available for them,
unlikely. The Kaplan-Meier estimates shown have
i.e. that they had not been revised up to the end of
been multiplied by 100, therefore they estimate the
the follow-up period (the end of 2016) or prior to
cumulative percentage probability.
their death; these observations were regarded as
being ‘censored’ at those times. For mortality, the In the case of revisions, no attempt has been made
event was death, censoring only those cases that to adjust for the competing risk of death. The likely
were still alive at the end of 2016 (and not for any impact of mortality was reported in the 11th Annual
revision procedure). Report (published September 2014).
The survival tables in this report show ‘Kaplan-

Meier’ estimates of the cumulative chance
Terminology note to patients with a stemmed prosthesis and metal

bearing surfaces (a monobloc metal acetabular
The six main categories of bearing surfaces
cup or a metal acetabular cup with a metal liner).
for hip replacements are ceramic-on-ceramic
Although they have metal-on-metal bearing
(CoC), ceramic-on-metal (CoM), ceramic-on-
surfaces, resurfacing procedures, which have a
polyethylene (CoP), metal-on-metal (MoM), metal-
surface replacement femoral prosthesis combined
on-polyethylene (MoP) and resurfacing procedures.
with a metal acetabular cup, are treated as a
The metal-on-metal group in this section refers
separate category.
3.3.1 Overview of primary hip surgery The most commonly used operation type overall
remains cemented metal-on-polyethylene (87.1% of all
Table 3.3 shows the breakdown of cases by method cemented primaries, 30.4% of all primaries).
of fixation and within each fixation sub-group, by
bearing surface.
Table 3.3 Numbers and percentages of primary hip replacements of each fixation type and by bearing surface.
Bearing surface within
Fixation Number (%) fixation group Number (%)
All cases 890,681 (100%) 890,681 (100%)
MoP 270,476 (30.4%)
MoM 1,098 (0.1%)
All cemented 310,596 (34.9%)
CoP 33,041 (3.7%)
Others/unsure 5,981 (0.7%)

MoP 133,873 (15.0%)
MoM 28,816 (3.2%)
CoP 64,644 (7.3%)
All uncemented 347,587 (39.0%)
CoC 113,185 (12.7%)
CoM 2,155 (0.2%)
MoP 105,619 (11.9%)
MoM 2,188 (0.2%)
All hybrid 170,589 (19.2%) CoP 37,294 (4.2%)
CoC 23,206 (2.6%)
MoP 15,255 (1.7%)
All reverse hybrid 22,552 (2.5%) CoP 7,200 (0.8%)
Others/unsure 97 (<0.1%)
All resurfacing 39,318 (4.4%) (MoM) 39,318 (4.4%)
Unsure 39 (<0.1%) Unsure 39 (not applicable)
Table 3.4 and Figure 3.2 show the distributions across With regard to bearing surfaces, metal-on-polyethylene is
fixation groups for each year of primary operation and still the most widely used, with ceramic-on-polyethylene
Figures 3.3 (a) to (d) show distributions across bearing following close behind; while the use of ceramic-on-
surface of each fixation group. Trends of implant usage ceramic is declining. The use of metal-on-metal stemmed
are interesting in that the decline in cemented implants implants has virtually ceased, with the proportion of
between 2003 and 2009 has arrested and is now stable metal-on-metal resurfacing implants decreasing from
at around a third of cases. Conversely, although the use a peak of 10.8% in 2006 to account for only 0.7% of
of uncemented implants has decreased since 2010, implants in 2016.
they still remain the most widely used compared to other
implants. Hybrid implants continue to steadily increase in
popularity and now account for a fifth of cases.
Table 3.4 Percentages of primary hip replacements in each calendar year that use each fixation type and for each
fixation group*.
Fixation/ 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
bearing n=14,454 n=28,057 n=40,575 n=48,485 n=60,761 n=67,128 n=68,104 n=70,619 n=73,632 n=77,775 n=79,885 n=86,977 n=86,496 n=87,733
All
60.3 53.9 48.2 42.4 39.4 34.0 31.7 31.3 32.2 32.8 33.1 31.9 30.9 29.6
cemented
Cemented by bearing surface:
MoP 55.4 49.0 43.7 38.1 35.5 30.1 28.1 27.1 27.5 28.4 28.3 27.0 25.9 24.7
MoM 0.1 0.4 0.4 0.4 0.4 0.4 0.1 <0.1(31) <0.1(8) 0.0 0.0 0.0 <0.1(2) <0.1(1)
CoP 2.7 3.2 3.0 2.9 2.5 2.7 2.8 3.2 3.5 4.0 4.4 4.6 4.8 4.8
Others/
2.0 1.3 1.2 1.0 1.1 0.8 0.6 0.9 1.2 0.5 0.4 0.3 0.2 0.1
unsure
All
16.9 21.5 25.8 30.2 33.4 39.4 43.2 45.8 45.0 44.9 42.6 40.9 39.4 38.5
uncemented
Uncemented by bearing surface:
MoP 6.2 9.1 9.9 10.3 10.8 13.1 15.1 16.9 17.2 17.9 17.6 17.1 16.4 16.2
MoM 1.3 2.2 5.5 8.4 10.3 10.9 8.0 3.2 0.4 0.1 <0.1(4) <0.1(1) 0.0 <0.1(12)
CoP 5.0 5.1 5.1 4.4 4.0 3.9 4.7 5.6 6.1 7.3 8.3 9.6 11.4 12.6

CoC 3.5 4.2 4.4 6.2 7.3 10.1 13.6 18.1 20.1 19.3 16.4 14.0 11.4 9.6
CoM 0.0 <0.1(1) <0.1(1) <0.1(7) 0.1 0.4 0.9 1.1 0.5 0.1 <0.1(27) <0.1(6) <0.1(1) <0.1(2)
Others/
0.8 0.8 0.9 0.9 0.9 1.0 0.9 0.9 0.7 0.3 0.2 0.2 0.2 0.1
unsure
All hybrid 12.3 13.6 14.4 15.6 15.2 15.3 15.9 16.3 17.2 17.8 20.3 23.2 25.7 28.1
Hybrid by bearing surface:
MoP 8.3 9.3 9.5 10.0 10.0 10.0 10.5 10.9 11.6 11.7 12.3 13.6 14.6 15.5
MoM 0.7 0.5 0.5 0.7 0.8 0.8 0.4 0.2 <0.1(32) <0.1(4) 0.0 0.0 0.0 <0.1(7)
CoP 1.5 1.5 1.2 1.2 1.0 1.3 1.8 1.9 2.2 3.1 5.1 7.1 8.9 10.7
CoC 1.2 1.9 2.8 3.2 3.0 2.7 2.9 3.0 3.1 2.9 2.7 2.4 2.1 1.7
Others/
0.7 0.4 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.1 0.1 0.1 0.1 0.1
unsure
All reverse
0.6 0.8 0.9 1.0 1.6 2.4 2.6 2.8 3.1 3.1 3.0 3.1 3.1 3.0
hybrid
Reverse hybrid by bearing surface:
MoP 0.3 0.6 0.7 0.8 1.1 1.7 1.8 1.9 2.2 2.0 2.0 2.0 2.1 2.0
CoP 0.2 0.2 0.2 0.2 0.6 0.7 0.8 0.9 0.9 1.1 1.0 1.1 1.0 1.0
Others/
<0.1(1) <0.1(6) <0.1(4) <0.1(7) <0.1(10) <0.1(15) <0.1(14) <0.1(16) <0.1(6) <0.1(3) <0.1(5) <0.1(5) <0.1(3) <0.1(2)
unsure
All
resurfacing 9.8 10.1 10.6 10.8 10.3 8.9 6.6 3.9 2.5 1.4 1.1 0.9 0.9 0.7
(MoM)
All unsure <0.1(2) 0.1 <0.1(1) 0.0 <0.1(1) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
All types 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
* Percentages calculated as percentage of total yearly operations. Where percentage is less than 0.1 the actual number of procedures is included in parenthesis. 0.0 represents
no procedures with this bearing type.
Figure 3.2
Temporal changes in percentages of each fixation method used in primary hip replacements.
70
Percentage of primaries (%)
60
50
40
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Year of primary
Uncemented Cemented Hybrid Reverse Hybrid Resurfacing
Figure 3.3 (a)

Temporal changes in percentages of each bearing surface used in cemented primary hip replacements.
Percentage of cemented primaries (%)
100
90
80
70
60
50
40
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Year of primary
Cemented MoP Cemented CoP Cemented Other/Unsure Cemented MoM
Figure 3.3 (b)

Temporal changes in percentages of each bearing surface used in uncemented primary hip replacements.
Percentage of uncemented primaries (%)
60

50
40
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Year of primary
Uncemented MoP Uncemented CoP Uncemented CoC

Uncemented Other/Unsure Uncemented CoM Uncemented MoM
Figure 3.3 (c)

Temporal changes in percentages of each bearing surface used in hybrid primary hip replacements.
Percentage of hybrid primaries (%)
70
60 © National Joint Registry 2017
50
40
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Year of primary
Hybrid MoP Hybrid CoP Hybrid CoC Hybrid Other/Unsure Hybrid MoM
Figure 3.3 (d)
Temporal changes in percentages of each bearing surface used in reverse hybrid primary hip replacements.
Percentage of reverse hybrid primaries (%)
80
70
60
50
40
30
20
10
0
2003
2004
2005
2006
2007
2008
2009
2010
2011
2013
2014
2015
2016
2012
Year of primary
Reverse Hybrid MoP Reverse Hybrid CoP Reverse Hybrid Other/Unsure
Within the whole registry, all the 890,681 primary hip The majority of hip primary procedures were carried out
replacement procedures contributing to our analyses on women (females 59.8%: males 40.2%). The median
were carried out by a total of 3,331 consultant surgeons age at primary operation was 69 (IQR 61-76) years1,
working across 468 units. Over the last three years (1 overall range 10-105 years.
January 2014 to 31 December 2016), 261,206 primary
hip procedures (representing 29.3% of the current Table 3.5 provides a breakdown of fixation type by age
registry) were performed by 2,205 consultant surgeons and gender with further division by bearing surfaces
working across 416 units. Looking at caseload over within each fixation sub-group.
this three-year period, the median number of primary
Patients receiving resurfacing and ceramic-on-ceramic
procedures per consultant surgeon was 54 (inter-
bearings tended to be younger than the other groups
quartile range (IQR) 4-178) and the median number
but the age ranges were wide. Those receiving
of procedures per unit was 547.5 (IQR 257.5-867). A
resurfacings were more likely to be men.
proportion of consultants will have just qualified over
this period, and some may have retired, therefore their
apparent caseload would be lower.
1 Omitting 207 cases where the NHS number was not traceable, therefore the age was not verifiable.
Table 3.5 Distribution of age at primary hip replacement (in years) and gender, for all procedures and for each type
of fixation and bearing surface.
By bearing Age (years)*

surface within Percentage
Fixation fixation group n Median (IQR***) Minimum Maximum (%) males**
All cases 890,681 69 (61-76) 7 105 40.2
All cemented 310,596 74 (68-79) 7 103 33.7
Cemented and
MoP 270,476 75 (69-80) 15 103 33.0
MoM 1,098 64 (57-72) 25 98 46.9
CoP 33,041 65 (58-71) 14 101 38.7
Others/unsure 5,981 72 (65-78) 7 102 36.2
All uncemented 347,587 65 (58-72) 11 105 44.3
Uncemented and
MoP 133,873 71 (65-77) 12 101 40.9

MoM 28,816 64 (57-70) 13 105 50.6
CoP 64,644 64 (58-70) 13 100 44.8
CoC 113,185 60 (53-66) 11 100 46.6
CoM 2,155 63 (56-69) 20 92 42.4
Others/unsure 4,914 66 (58-73) 17 96 42.8
All hybrid 170,589 70 (63-77) 12 105 37.0
Hybrid and
MoP 105,619 73 (68-79) 12 105 34.9
MoM 2,188 64 (56-72) 18 95 48.0
CoP 37,294 66 (59-72) 14 97 39.7
CoC 23,206 60 (53-66) 13 93 40.9
Others/unsure 2,282 69 (61-77) 16 94 36.1
All reverse hybrid 22,552 71 (64-77) 13 100 35.9
Reverse hybrid and
MoP 15,255 73 (68-78) 13 100 34.2
CoP 7,200 64 (58-69) 16 94 39.6
Others/unsure 97 69 (61-76) 30 93 32.0
All resurfacing
39,318 55 (48-60) 12 95 71.3
(MoM)
Unsure 39 69 (56-75) 18 83 38.5
* Excludes 207 cases with unverifiable ages. **Excludes one with uncertain gender. *** IQR=interquartile range
3.3.2 Revisions after primary methods have been proposed to deal with this, they
typically give different values and, as yet, there is no
hip surgery clear consensus for the large datasets we have here.
Figures 3.4 (a) and (b) illustrate temporal changes in Kaplan-Meier estimates are not shown at all when the
the overall revision rates using Kaplan-Meier estimates; numbers at risk fell below ten.
procedures have been grouped by the year of the
Please note that the rates for ‘resurfacing’ throughout
primary operation. Figure 3.4 (a) plots each Kaplan-
Section 3.3 still include the ASR system unless
Meier survival curve with a common origin, i.e. time
explicitly stated otherwise.
zero is equal to the year of operation. Figure 3.4 (b)
shows the same curves plotted against calendar time, Figures 3.5 to 3.8 illustrate the differences
where the origin of each curve is the year of operation. between the various bearing surface sub-groups
In addition, the revision rate at 1, 3 and 5 years has for cemented, uncemented, hybrid and reverse
also been highlighted. Figure 3.4 (b) separates each hybrid hips, respectively. These continue to show
year allowing changes in failure rates to be clearly the worst outcome for metal-on-metal bearings,
identified. If revision surgery and timing of revision which, in uncemented hips (Figure 3.6), fared worse
surgery were static across time we would expect than resurfacings. The failure rates for ceramic-on-
all failure curves to be the same shape and equally polyethylene bearings were particularly low and it is
spaced, departures from this indicate a change in the encouraging that these are becoming more widely
number, and timing of revision procedures. It is also used with time.
very clear that the three- and five-year rate of revision
increases for operations occurring between 2003 In Table 3.6 and Figures 3.5 to 3.8, all age groups and
to 2008 and then reduces for operations occurring genders were combined. In Figures 3.9 (a) and 3.9
between 2009 and 2016. The differences may be (b), the whole cohort has been sub-divided by age at
partly a result of under-reporting in the earlier years primary operation and by gender. Across the whole
of the registry, but most probably reflect the usage of group, there was an inverse relationship between the
metal-on-metal, which peaked in 2008 and then fell probability of revision and the age of the patient. A
(see Table 3.4). Further investigation is needed. closer look at both genders (Figure 3.9 (a)) shows that
the variation between the age groups was greater
Table 3.6 provides Kaplan-Meier estimates of the in women than in men. Thus, for example, women
cumulative percentage probability of first revision, for under 55 years had higher revision rates than their
any cause, firstly for all cases combined and then by male counterparts in the same age band, whereas
type of fixation and by bearing surface within each women aged 80 years and older had a lower rate.
fixation group. The table shows updated estimates at 1, In Figure 3.9 (b), implants with metal-on-metal (or
3, 5, 7, 10, 11 and 13 years from the primary operation uncertain) bearing surfaces and resurfacings have
together with 95% Confidence Intervals (95% CI). been excluded. The revision rates for the younger
Results at 13 years have been added, but in general, women are much reduced; an age trend is seen in
the group sizes are too small for meaningful sub- both genders but rates for women are lower than for
division, hence many of these estimates are shown in men across the entire age spectrum.
blue italics. Estimates in blue italics indicate time points
where fewer than 250 cases remained at risk, meaning Where group sizes permitted (overall group
that the estimates are less reliable. Further revisions in size>10,000), Table 3.7 further expands Table 3.6 to
these groups would be highly unlikely and, when they show separate estimates for males and females within
do occur, they may appear to have a disproportionate each of four age bands, <55, 55-64, 65-74 and 75+
impact on the Kaplan-Meier estimate, i.e. the step years. Estimates are shown at 1, 3, 5, 7, 10 and 13
upwards may seem steeper. Furthermore, the upper years after the primary operation. These refine results in
95% Confidence Interval at these time points may our 2016 report, but now with larger numbers of cases
be underestimated. Although a number of statistical and therefore generally narrower Confidence Intervals.
Figure 3.4 (a)

Temporal changes in revision rates after primary hip replacement: Kaplan-Meier estimates of
cumulative percentage probability of revision for each year of primary operation.
7 2003
2004
Cumulative percentage probability (%)
6 2005
2006

2007
5
2008
2009
4
2010
2011
3
2012
2013
2
2014
2015
1
2016
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years since primary operation
50
Figure 3.4 (b)
Temporal changes in revision rates after primary hip replacement: Kaplan-Meier estimates of cumulative percentage probability of revision
for each year of primary operation with failure rates at 1, 3, and 5 years indicated.
www.njrcentre.org.uk
7
2
0
2003 2005 2007 2009 2011 2013 2015 2017
Year of primary surgery
Cumulative percentage probability of first revision (%)

Cumulative probability of revision 1 year after primary
Cumulative probability of revision 3 years after primary
Table 3.6 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) after primary hip
replacement, by year from the primary operation, for all cases and by fixation and bearing surface. Blue italics signify
that fewer than 250 cases remained at risk at these time points.
Cumulative percentage probability of revision (95% CI) at:
Fixation/
n 1 year 3 years 5 years 7 years 10 years 11 years 13 years
bearing types
0.78 1.56 2.41 3.47 5.21 5.74 6.83
All cases* 890,681
(0.76-0.79) (1.53-1.58) (2.37-2.44) (3.42-3.52) (5.13-5.29) (5.64-5.83) (6.67-6.99)
0.49 1.05 1.51 2.05 3.05 3.47 4.34
All cemented 310,596
(0.47-0.52) (1.01-1.09) (1.46-1.56) (1.98-2.11) (2.95-3.14) (3.35-3.59) (4.14-4.54)
Cemented by bearing surface
0.50 1.05 1.50 2.00 2.99 3.37 4.25
MoP 270,476
(0.47-0.52) (1.01-1.09) (1.44-1.55) (1.93-2.07) (2.89-3.09) (3.25-3.50) (4.04-4.46)
0.64 2.71 6.33 12.00 17.15 18.89 18.89
MoM 1,098
(0.31-1.34) (1.89-3.87) (5.01-7.99) (10.13-14.18) (14.80-19.84) (16.18-21.99) (16.18-21.99)
0.45 0.97 1.32 1.71 2.30 2.70 3.81
CoP 33,041
(0.38-0.53) (0.86-1.09) (1.18-1.48) (1.53-1.91) (2.03-2.61) (2.34-3.11) (3.07-4.72)
0.58 1.16 1.64 2.43 3.42 4.83 5.18
Others/unsure 5,981
(0.41-0.81) (0.91-1.48) (1.33-2.03) (2.01-2.94) (2.82-4.14) (3.90-5.97) (4.16-6.45)
All 0.99 1.93 2.99 4.39 6.77 7.36 8.66
347,587
uncemented (0.96-1.02) (1.88-1.98) (2.92-3.05) (4.30-4.48) (6.61-6.93) (7.17-7.55) (8.31-9.03)
Uncemented by bearing surface
1.06 1.78 2.28 2.90 4.18 4.66 5.90
MoP 133,873
(1.01-1.12) (1.71-1.86) (2.19-2.38) (2.78-3.02) (3.97-4.39) (4.41-4.93) (5.37-6.47)
1.04 3.41 7.52 12.20 18.20 19.40 22.14
MoM 28,816
(0.92-1.16) (3.21-3.63) (7.22-7.84) (11.81-12.60) (17.66-18.76) (18.76-20.05) (20.71-23.65)

0.87 1.52 2.04 2.49 3.40 3.67 4.49
CoP 64,644
(0.80-0.94) (1.42-1.63) (1.91-2.18) (2.33-2.67) (3.14-3.68) (3.38-3.99) (3.98-5.06)
0.95 1.80 2.36 2.88 3.99 4.46 5.69
CoC 113,185
(0.89-1.01) (1.72-1.88) (2.26-2.46) (2.76-3.00) (3.76-4.23) (4.16-4.78) (4.97-6.50)
0.65 2.83 4.84 6.17
CoM 2,155
(0.39-1.10) (2.20-3.63) (4.00-5.85) (5.18-7.35)
1.33 2.28 3.15 4.12 5.25 5.92 7.66
Others/unsure 4,914
(1.05-1.70) (1.89-2.75) (2.68-3.71) (3.56-4.77) (4.53-6.07) (5.03-6.97) (5.95-9.83)
0.74 1.29 1.86 2.47 3.62 4.07 5.05
All hybrids 170,589
(0.70-0.78) (1.23-1.35) (1.78-1.94) (2.37-2.57) (3.46-3.79) (3.87-4.29) (4.70-5.44)
Hybrids by bearing surface
0.78 1.32 1.83 2.29 3.40 3.96 4.94
MoP 105,619
(0.72-0.83) (1.25-1.40) (1.74-1.93) (2.17-2.41) (3.20-3.62) (3.70-4.24) (4.49-5.43)
0.78 3.00 6.52 11.32 16.08 16.60 19.46
MoM 2,188
(0.49-1.25) (2.35-3.82) (5.54-7.67) (10.01-12.79) (14.36-17.98) (14.81-18.60) (16.79-22.49)
0.71 1.18 1.53 1.87 2.48 2.64 4.21
CoP 37,294
(0.62-0.80) (1.06-1.32) (1.37-1.72) (1.65-2.12) (2.11-2.91) (2.22-3.13) (3.10-5.70)
0.59 1.04 1.57 2.04 2.78 3.13 3.31
CoC 23,206
(0.50-0.70) (0.91-1.19) (1.40-1.76) (1.83-2.27) (2.47-3.11) (2.75-3.56) (2.87-3.82)
1.16 1.55 1.96 2.62 3.52 3.52 3.87
Others/unsure 2,282
(0.79-1.70) (1.11-2.17) (1.45-2.66) (1.98-3.47) (2.68-4.62) (2.68-4.62) (2.85-5.24)
All reverse 0.78 1.50 2.03 2.55 4.00 4.32 5.73
22,552
hybrids (0.68-0.91) (1.34-1.68) (1.82-2.26) (2.28-2.85) (3.36-4.76) (3.57-5.23) (4.14-7.92)
Reverse hybrids by bearing surface
0.82 1.47 2.02 2.51 4.18 4.37 6.64
MoP 15,255
(0.69-0.99) (1.27-1.69) (1.76-2.30) (2.19-2.87) (3.37-5.19) (3.49-5.46) (4.29-10.21)
0.68 1.51 2.00 2.52 3.54 4.20 4.20
CoP 7,200
(0.51-0.91) (1.23-1.85) (1.65-2.43) (2.07-3.07) (2.65-4.72) (2.84-6.20) (2.84-6.20)
2.08 5.56 5.56 8.40 8.40
Others/unsure 97**
(0.53-8.08) (2.35-12.86) (2.35-12.86) (4.06-16.93) (4.06-16.93)
All
1.24 3.07 5.51 8.18 11.47 12.34 13.98
resurfacing 39,318
(1.13-1.35) (2.90-3.25) (5.28-5.75) (7.90-8.47) (11.11-11.84) (11.94-12.74) (13.41-14.57)
(MoM)
* Includes 39 with unsure fixation/bearing surface.

** Wide CI because based on very small group size (n=97).
Figure 3.5
Comparison of cumulative probability of revision (Kaplan-Meier estimates) for cemented primary hip
replacements with different bearing surfaces.
20
15
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Cemented MoP 270,476 242,126 214,673 186,554 159,779 134,548 111,865 91,418 72,403 54,218 37,084 23,605 12,163 4,184
Cemented MoM 1,098 1,079 1,048 1,020 976 934 873 793 694 494 321 173 78 14
Cemented CoP 33,041 28,423 23,988 19,737 16,129 12,934 10,301 8,022 6,164 4,488 3,131 1,924 937 259
Figure 3.6
Comparison of cumulative probability of revision (Kaplan-Meier estimates) for uncemented primary hip
25
20

15
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Uncemented MoP 133,873 116,605 100,644 84,375 69,229 54,834 42,228 30,868 21,562 14,098 8,860 5,043 2,189 541
Uncemented MoM 28,816 28,217 27,656 26,900 25,954 24,816 23,386 20,500 15,445 9,576 4,858 2,062 554 117
Uncemented CoP 64,644 52,666 42,380 33,774 26,897 21,086 16,552 12,657 9,584 7,200 5,025 3,209 1,596 522
Uncemented CoC 113,185 103,237 92,499 79,621 66,197 51,211 36,764 24,569 15,867 9,626 5,592 2,933 1,387 400
Uncemented CoM 2,155 2,125 2,087 2,034 1,959 1,860 1,523 840 290 47 7 1 1 0
Resurfacing 39,318 38,098 36,900 35,618 34,223 32,599 30,236 27,116 22,648 17,096 11,569 7,102 3,558 1,178
Figure 3.7
Comparison of cumulative probability of revision (Kaplan-Meier estimates) for hybrid primary hip
20
15
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Hybrid MoP 105,619 89,439 75,371 62,342 51,531 41,893 33,194 25,847 19,192 13,474 8,638 5,073 2,424 750
Hybrid MoM 2,188 2,144 2,089 2,023 1,942 1,879 1,763 1,574 1,330 898 525 306 170 65
Hybrid CoP 37,294 27,455 19,594 13,408 9,326 6,920 5,290 3,949 2,817 1,974 1,402 897 487 153
Hybrid CoC 23,206 21,481 19,487 17,298 14,990 12,680 10,391 8,318 6,380 4,661 3,011 1,623 647 149
Figure 3.8
Comparison of cumulative probability of revision (Kaplan-Meier estimates) for reverse hybrid primary hip

4
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Rev. Hyb. MoP 15,255 13,103 11,071 9,158 7,418 5,786 4,232 2,975 1,904 1,019 536 290 125 27
Rev. Hyb. CoP 7,200 6,245 5,303 4,297 3,490 2,677 2,008 1,440 937 527 222 129 65 26
Figure 3.9 (a)
Cumulative probability of revision (Kaplan-Meier) for the whole cohort of primary hip replacements broken
down by age separately for each gender.
Cumulative percentage probability of revision (%)
Males Females
15.0 15.0
<55 y
55−59 y
12.5 12.5
60−64 y
65−69 y
10.0 10.0
70−74 y
75−79 y
7.5 7.5
80+ y
5.0 5.0
2.5 2.5
0.0 0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years from primary operation
Figure 3.9 (b)

Cumulative probability of revision (Kaplan-Meier) for the whole cohort of primary hip replacements broken down by
age separately for each gender, but excluding metal-on-metal (or uncertain) total hip replacement and resurfacings.
Males Females
7 7
<55 y
55−59 y
6 6 60−64 y
65−69 y
5 5
70−74 y
4 4 75−79 y
80+ y
3 3
2 2
1 1
0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years from primary operation
Table 3.7 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI), by gender and age, at 1, 3, 5, 7, 10 and 13 years from the
primary hip replacement, for each fixation group and main bearing surface. Blue italics signify that fewer than 250 cases remained at risk at these time points.
Males Females
Age at Years from primary operation Years from primary operation
Fixation/ primary
bearing types (years) n 1 year 3 years 5 years 7 years 10 years 13 years n 1 year 3 years 5 years 7 years 10 years 13 years
0.91 2.24 3.78 5.51 8.26 10.14 0.92 2.40 4.27 6.83 10.48 13.52
All cases <55 53,803 53,757
(0.84-1.00) (2.11-2.38) (3.60-3.98) (5.27-5.77) (7.89-8.64) (9.54-10.77) (0.84-1.00) (2.27-2.55) (4.07-4.48) (6.55-7.12) (10.05-10.93) (12.75-14.34)
0.70 1.91 2.93 4.40 6.54 9.34 0.65 1.61 2.47 3.95 5.85 7.67
All cemented <55 4,377 6,588
(0.48-1.00) (1.51-2.41) (2.39-3.60) (3.64-5.32) (5.42-7.88) (7.64-11.39) (0.48-0.89) (1.30-1.98) (2.06-2.96) (3.34-4.66) (4.97-6.88) (6.28-9.33)
0.94 2.27 3.33 4.10 6.68 10.04 0.91 1.91 2.57 3.82 5.76 7.33
MoP <55 1,894 3,246
(0.59-1.51) (1.64-3.12) (2.51-4.41) (3.14-5.35) (5.16-8.63) (7.75-12.97) (0.63-1.31) (1.47-2.49) (2.02-3.27) (3.06-4.75) (4.67-7.10) (5.83-9.19)
0.55 1.35 2.01 2.87 3.48 4.67 0.38 1.15 1.88 2.83 3.79 6.92
CoP <55 2,136 3,013
(0.30-0.99) (0.89-2.04) (1.39-2.90) (1.99-4.13) (2.36-5.12) (2.97-7.31) (0.21-0.69) (0.79-1.67) (1.35-2.62) (2.05-3.89) (2.68-5.35) (3.90-12.13)
All 0.97 2.45 4.01 5.84 9.30 11.02 0.97 2.31 3.90 5.95 9.08 11.93
<55 28,561 31,058
uncemented (0.86-1.09) (2.26-2.66) (3.74-4.29) (5.47-6.23) (8.65-9.99) (10.03-12.11) (0.86-1.08) (2.14-2.50) (3.65-4.17) (5.60-6.33) (8.49-9.71) (10.66-13.35)
0.99 2.30 3.38 5.02 6.52 7.26 1.23 2.16 2.99 3.66 4.62 8.14
MoP <55 3,305 4,079
(0.69-1.40) (1.79-2.94) (2.69-4.25) (4.02-6.25) (5.19-8.16) (5.67-9.28) (0.93-1.63) (1.73-2.70) (2.43-3.68) (2.97-4.50) (3.68-5.80) (5.33-12.33)
0.68 3.52 7.41 11.35 18.26 22.04 1.78 5.71 12.59 19.94 26.96 35.27
MoM <55 3,241 2,365
(0.45-1.03) (2.94-4.22) (6.55-8.38) (10.28-12.51) (16.69-19.97) (19.02-25.47) (1.32-2.40) (4.84-6.72) (11.31-14.00) (18.36-21.63) (25.02-29.02) (29.50-41.79)
1.28 2.26 3.27 3.64 4.01 4.50 0.88 1.63 2.35 3.10 4.27 4.69
CoP <55 5,023 5,693
(0.99-1.65) (1.82-2.81) (2.63-4.07) (2.90-4.56) (3.15-5.09) (3.32-6.07) (0.67-1.17) (1.29-2.05) (1.89-2.93) (2.47-3.90) (3.24-5.62) (3.45-6.36)
0.93 2.19 3.04 3.79 5.05 6.43 0.81 1.86 2.52 3.26 4.95 5.96
CoC <55 16,417 18,192
(0.79-1.09) (1.96-2.45) (2.74-3.36) (3.42-4.21) (4.40-5.79) (5.17-7.98) (0.69-0.96) (1.66-2.08) (2.27-2.79) (2.94-3.62) (4.32-5.67) (4.86-7.31)
0.91 1.58 2.45 3.98 6.22 8.00 0.67 1.26 1.96 3.09 4.39 6.35
All hybrid <55 7,022 9,411
(0.71-1.17) (1.29-1.94) (2.03-2.94) (3.35-4.73) (5.23-7.38) (6.35-10.06) (0.52-0.86) (1.03-1.53) (1.65-2.34) (2.63-3.62) (3.71-5.19) (4.95-8.13)
1.30 2.31 3.45 4.62 7.65 11.05 0.87 1.75 2.51 3.42 5.11 7.88
MoP <55 1,304 1,946
(0.80-2.11) (1.57-3.39) (2.42-4.89) (3.27-6.49) (5.44-10.69) (7.24-16.67) (0.54-1.42) (1.21-2.51) (1.81-3.48) (2.51-4.65) (3.70-7.05) (5.18-11.91)
1.03 1.53 1.88 2.63 3.13 3.13 0.53 1.11 1.52 2.48 3.66 7.85
CoP <55 2,530 3,170
(0.69-1.53) (1.05-2.24) (1.26-2.79) (1.61-4.28) (1.87-5.22) (1.87-5.22) (0.33-0.87) (0.75-1.64) (1.03-2.25) (1.63-3.76) (2.29-5.82) (2.72-21.54)
0.69 1.28 1.84 2.59 3.33 3.33 0.57 0.91 1.36 2.16 3.04 4.08
CoC <55 2,802 3,927
(0.44-1.08) (0.91-1.79) (1.36-2.49) (1.95-3.44) (2.47-4.47) (2.47-4.47) (0.38-0.87) (0.64-1.27) (1.01-1.84) (1.65-2.83) (2.30-4.02) (2.90-5.72)
All reverse 0.80 1.82 1.82 1.82 5.10 0.83 1.57 2.69 3.78 5.29
<55 632 883
hybrid (0.33-1.92) (0.98-3.37) (0.98-3.37) (0.98-3.37) (2.03-12.49) (0.40-1.74) (0.89-2.77) (1.64-4.38) (2.34-6.09) (3.14-8.82)
2.86 2.86 2.86 2.86 0.87 3.16 5.09
MoP <55 129 0 197 0 0
(0.93-8.64) (0.93-8.64) (0.93-8.64) (0.93-8.64) (0.12-6.01) (1.02-9.57) (1.85-13.65)
All
0.88 2.24 4.12 5.87 8.30 10.10 1.33 4.97 9.33 14.38 20.41 24.51
resurfacing <55 13,207 5,813
(0.73-1.05) (1.99-2.51) (3.78-4.49) (5.45-6.31) (7.75-8.89) (9.20-11.07) (1.06-1.65) (4.44-5.56) (8.60-10.11) (13.49-15.33) (19.31-21.58) (22.89-26.23)
(MoM)
Note: Includes cases with unknown fixation/bearing but excludes the 207 cases where the ages were unveritable, plus a further one with uncertain gender.
57
58
Males Females
Fixation/ primary
0.90 1.94 3.04 4.40 6.57 8.60 0.73 1.67 2.85 4.42 6.74 8.61
All cases 55-64 89,239 108,096
(0.84-0.97) (1.84-2.04) (2.91-3.17) (4.23-4.58) (6.31-6.83) (8.12-9.11) (0.68-0.78) (1.59-1.75) (2.74-2.97) (4.27-4.59) (6.50-6.98) (8.20-9.04)
0.64 1.47 2.06 2.85 4.71 6.66 0.44 1.12 1.76 2.58 4.17 5.85
All cemented 55-64 15,008 24,386
(0.52-0.78) (1.28-1.69) (1.82-2.33) (2.54-3.20) (4.21-5.26) (5.76-7.71) (0.36-0.53) (0.98-1.27) (1.58-1.96) (2.34-2.84) (3.80-4.57) (5.21-6.55)
0.64 1.70 2.38 3.18 5.12 7.23 0.48 1.20 1.80 2.54 4.13 5.59
MoP 55-64 9,785 16,945
(0.50-0.83) (1.45-2.00) (2.07-2.74) (2.79-3.62) (4.52-5.79) (6.19-8.43) (0.39-0.60) (1.04-1.39) (1.59-2.04) (2.27-2.85) (3.72-4.59) (4.93-6.35)
0.62 0.91 1.12 1.44 2.20 2.72 0.28 0.76 1.19 1.67 2.47 5.23
CoP 55-64 4,593 6,685
(0.43-0.91) (0.66-1.25) (0.82-1.53) (1.06-1.97) (1.52-3.17) (1.68-4.39) (0.18-0.45) (0.56-1.04) (0.91-1.56) (1.28-2.16) (1.82-3.33) (3.33-8.17)
All 0.94 2.05 3.26 4.93 7.61 9.74 0.84 1.83 3.08 4.85 7.52 9.17
55-64 46,888 54,882
uncemented (0.85-1.03) (1.91-2.19) (3.08-3.45) (4.67-5.20) (7.19-8.06) (8.91-10.64) (0.76-0.92) (1.71-1.95) (2.91-3.25) (4.62-5.10) (7.14-7.92) (8.51-9.89)
1.03 2.13 2.83 3.73 5.33 7.72 0.81 1.74 2.27 3.04 4.56 6.76
MoP 55-64 11,172 14,516
(0.85-1.24) (1.85-2.44) (2.49-3.21) (3.29-4.22) (4.65-6.11) (6.28-9.47) (0.68-0.97) (1.52-1.98) (2.01-2.56) (2.70-3.41) (4.01-5.18) (5.53-8.26)
0.84 3.04 6.56 10.96 17.07 19.03 0.88 3.61 8.91 15.42 22.09 25.13
MoM 55-64 5,109 4,807
(0.63-1.14) (2.61-3.56) (5.91-7.29) (10.11-11.87) (15.84-18.38) (17.33-20.88) (0.65-1.18) (3.11-4.18) (8.13-9.76) (14.41-16.50) (20.77-23.47) (23.01-27.40)
0.87 1.52 2.16 2.77 3.56 5.48 0.69 1.44 2.05 2.59 3.92 4.58
CoP 55-64 9,842 12,037
(0.70-1.08) (1.27-1.83) (1.82-2.57) (2.32-3.30) (2.95-4.29) (4.25-7.06) (0.56-0.86) (1.22-1.70) (1.76-2.39) (2.23-3.02) (3.28-4.67) (3.75-5.61)
0.94 1.85 2.51 3.07 4.15 6.11 0.90 1.59 2.15 2.63 3.61 4.31
CoC 55-64 19,837 22,257
(0.81-1.09) (1.67-2.06) (2.27-2.76) (2.78-3.39) (3.63-4.74) (4.51-8.25) (0.78-1.03) (1.43-1.77) (1.95-2.37) (2.39-2.91) (3.18-4.09) (3.65-5.10)
All 0.79 1.55 2.31 2.94 4.50 6.71 0.56 1.15 1.81 2.51 3.59 5.15
55-64 13,962 21,467
hybrid (0.66-0.96) (1.34-1.79) (2.03-2.63) (2.59-3.33) (3.94-5.13) (5.55-8.11) (0.47-0.68) (1.00-1.31) (1.61-2.04) (2.25-2.81) (3.21-4.01) (4.35-6.09)
1.02 1.78 2.50 2.85 4.69 7.59 0.68 1.18 1.88 2.57 3.50 5.18
MoP 55-64 5,009 8,537
(0.77-1.34) (1.43-2.22) (2.05-3.04) (2.34-3.47) (3.83-5.73) (5.76-9.97) (0.52-0.88) (0.96-1.45) (1.58-2.24) (2.18-3.02) (2.97-4.12) (4.17-6.44)
0.67 1.29 1.63 1.89 2.85 5.31 0.57 1.09 1.36 1.71 2.59 6.25
CoP 55-64 4,651 6,700
(0.46-0.96) (0.95-1.75) (1.19-2.23) (1.35-2.63) (1.81-4.47) (2.98-9.35) (0.41-0.79) (0.83-1.43) (1.03-1.79) (1.27-2.31) (1.81-3.72) (3.17-12.15)
0.59 1.09 1.84 2.30 3.00 3.54 0.38 0.90 1.34 1.74 2.25 2.64
CoC 55-64 3,778 5,552
(0.39-0.89) (0.80-1.49) (1.42-2.39) (1.79-2.95) (2.31-3.90) (2.61-4.81) (0.25-0.59) (0.67-1.20) (1.05-1.71) (1.38-2.19) (1.78-2.85) (1.96-3.54)
All reverse 0.97 2.24 3.09 3.68 6.25 0.79 1.63 2.29 3.03 4.84 5.70
55-64 1,760 2,780
hybrid (0.59-1.58) (1.58-3.16) (2.23-4.26) (2.66-5.09) (3.92-9.89) (0.51-1.20) (1.19-2.23) (1.72-3.05) (2.28-4.00) (3.33-7.00) (3.70-8.74)
0.96 1.74 2.98 3.48 7.43 1.06 1.82 2.88 3.83 7.04
MoP 55-64 670 1,165
(0.43-2.12) (0.94-3.23) (1.73-5.11) (2.02-5.96) (3.78-14.33) (0.60-1.86) (1.16-2.85) (1.94-4.25) (2.62-5.59) (4.40-11.17)
All
1.22 2.42 3.96 5.60 7.54 9.36 1.62 4.49 8.59 12.91 17.58 21.06
resurfacing 55-64 11,617 4,577
(1.04-1.44) (2.15-2.72) (3.61-4.34) (5.18-6.06) (7.01-8.11) (8.45-10.37) (1.29-2.03) (3.93-5.13) (7.81-9.45) (11.96-13.94) (16.43-18.80) (19.38-22.86)
(MoM)
Males Females
Fixation/ primary
0.83 1.55 2.24 3.10 4.65 6.33 0.67 1.32 1.98 2.76 4.04 5.19
All cases 65-74 123,588 189,584
(0.78-0.88) (1.48-1.63) (2.15-2.34) (2.97-3.22) (4.46-4.85) (5.93-6.77) (0.64-0.71) (1.26-1.37) (1.91-2.06) (2.67-2.86) (3.89-4.18) (4.93-5.47)
0.56 1.14 1.65 2.31 3.58 5.43 0.41 1.01 1.47 1.95 2.82 3.90
All cemented 65-74 41,510 75,790
(0.49-0.64) (1.04-1.25) (1.52-1.80) (2.14-2.50) (3.32-3.87) (4.88-6.05) (0.37-0.46) (0.93-1.09) (1.38-1.57) (1.83-2.07) (2.65-3.00) (3.59-4.23)
0.58 1.18 1.69 2.35 3.64 5.54 0.40 0.99 1.47 1.95 2.84 3.92
MoP 65-74 36,086 67,003
(0.51-0.67) (1.07-1.31) (1.55-1.85) (2.17-2.55) (3.35-3.94) (4.94-6.21) (0.35-0.45) (0.91-1.07) (1.37-1.58) (1.83-2.08) (2.66-3.03) (3.60-4.27)
0.40 0.87 1.24 1.52 2.37 3.48 0.52 1.16 1.33 1.55 1.82 2.87
CoP 65-74 4,504 7,279
(0.25-0.64) (0.61-1.22) (0.90-1.70) (1.11-2.08) (1.71-3.26) (2.21-5.47) (0.38-0.73) (0.92-1.47) (1.06-1.67) (1.23-1.95) (1.39-2.36) (1.77-4.63)
All 0.97 1.82 2.64 3.71 5.70 7.69 0.92 1.70 2.64 3.91 5.97 7.38
65-74 52,498 67,779
uncemented (0.89-1.06) (1.70-1.94) (2.48-2.80) (3.50-3.93) (5.33-6.08) (6.75-8.77) (0.85-0.99) (1.60-1.81) (2.51-2.79) (3.72-4.11) (5.64-6.31) (6.74-8.08)
0.94 1.70 2.13 2.78 4.38 6.02 0.95 1.57 2.02 2.52 3.62 4.68
MoP 65-74 22,964 32,804
(0.82-1.07) (1.53-1.89) (1.93-2.36) (2.51-3.08) (3.88-4.94) (5.01-7.24) (0.85-1.06) (1.44-1.72) (1.85-2.20) (2.31-2.75) (3.26-4.02) (3.97-5.51)
1.06 2.99 6.00 9.28 13.79 14.93 1.10 3.46 8.51 13.73 20.08 25.21
MoM 65-74 4,538 4,655
(0.80-1.41) (2.53-3.53) (5.33-6.75) (8.44-10.20) (12.58-15.11) (13.09-17.01) (0.84-1.44) (2.97-4.03) (7.74-9.37) (12.75-14.79) (18.72-21.52) (20.96-30.14)
0.78 1.25 1.48 1.77 2.37 4.80 0.81 1.48 1.94 2.44 3.35 4.06
CoP 65-74 10,595 13,101
(0.62-0.97) (1.04-1.51) (1.24-1.78) (1.47-2.14) (1.90-2.95) (2.88-7.93) (0.66-0.98) (1.27-1.73) (1.67-2.25) (2.11-2.83) (2.85-3.95) (3.20-5.15)
1.14 1.86 2.34 2.64 3.57 5.98 0.86 1.48 1.80 2.10 2.66 4.10
CoC 65-74 13,366 15,901
(0.97-1.34) (1.63-2.11) (2.07-2.64) (2.34-2.98) (2.97-4.28) (3.98-8.94) (0.73-1.02) (1.30-1.69) (1.59-2.04) (1.85-2.39) (2.23-3.17) (2.47-6.78)
0.85 1.46 2.04 2.57 4.10 4.75 0.74 1.23 1.76 2.28 3.18 4.39
All hybrid 65-74 23,384 39,469
(0.74-0.98) (1.30-1.64) (1.83-2.27) (2.31-2.85) (3.65-4.61) (4.14-5.44) (0.66-0.83) (1.11-1.35) (1.61-1.93) (2.09-2.49) (2.88-3.51) (3.79-5.07)
0.83 1.45 1.99 2.45 4.07 4.88 0.73 1.27 1.76 2.21 3.14 4.23
MoP 65-74 15,060 26,915
(0.70-1.00) (1.26-1.67) (1.75-2.27) (2.15-2.79) (3.53-4.69) (4.13-5.77) (0.63-0.84) (1.13-1.42) (1.58-1.95) (1.99-2.45) (2.80-3.53) (3.60-4.96)
0.88 1.44 1.68 2.08 2.92 2.92 0.74 1.10 1.59 1.74 1.74 2.66
CoP 65-74 5,296 8,475
(0.65-1.18) (1.10-1.89) (1.26-2.23) (1.51-2.86) (1.94-4.41) (1.94-4.41) (0.57-0.96) (0.88-1.38) (1.25-2.03) (1.35-2.23) (1.35-2.23) (1.59-4.43)
0.76 1.34 2.03 2.03 2.85 0.75 0.94 1.38 1.87 2.82
CoC 65-74 2,442 3,293
(0.48-1.20) (0.94-1.91) (1.49-2.76) (1.49-2.76) (1.99-4.07) (0.50-1.12) (0.65-1.34) (1.01-1.89) (1.38-2.53) (2.00-3.97)
All reverse 0.93 1.83 2.35 3.20 4.32 0.55 1.02 1.53 1.87 3.63 3.63
65-74 3,188 5,713
hybrid (0.64-1.34) (1.38-2.42) (1.80-3.06) (2.44-4.18) (2.70-6.89) (0.38-0.78) (0.77-1.34) (1.19-1.97) (1.45-2.41) (2.41-5.44) (2.41-5.44)
1.12 1.99 2.65 3.25 4.49 0.55 0.96 1.43 1.90 4.22
MoP 65-74 2,185 4,192
(0.74-1.68) (1.44-2.75) (1.96-3.59) (2.39-4.41) (2.50-7.99) (0.36-0.83) (0.69-1.34) (1.06-1.94) (1.41-2.57) (2.67-6.65)
All
1.88 3.01 4.44 6.25 7.85 9.44 1.95 3.79 6.73 10.71 16.38 17.77
resurfacing 65-74 3,004 824
(1.45-2.44) (2.45-3.70) (3.74-5.27) (5.39-7.24) (6.82-9.02) (7.99-11.14) (1.20-3.16) (2.68-5.35) (5.20-8.70) (8.73-13.12) (13.72-19.50) (14.78-21.29)
(MoM)
59
60
Males Females
Fixation/ primary
0.90 1.49 1.98 2.47 3.42 4.38 0.67 1.10 1.48 1.90 2.55 3.33
All cases 75+ 91,320 181,085
(0.84-0.97) (1.41-1.58) (1.87-2.09) (2.33-2.61) (3.19-3.66) (3.77-5.09) (0.63-0.70) (1.05-1.15) (1.42-1.55) (1.82-1.99) (2.42-2.68) (2.99-3.71)
0.69 1.20 1.62 2.04 2.73 3.33 0.41 0.81 1.13 1.48 2.01 2.61
All cemented 75+ 43,902 98,969
(0.62-0.78) (1.10-1.32) (1.49-1.77) (1.88-2.23) (2.48-3.02) (2.90-3.83) (0.37-0.45) (0.76-0.88) (1.06-1.21) (1.39-1.58) (1.87-2.16) (2.29-2.97)
0.70 1.22 1.64 2.08 2.77 3.36 0.40 0.80 1.12 1.46 2.00 2.61
MoP 75+ 41,552 93,910
(0.62-0.79) (1.11-1.34) (1.50-1.79) (1.91-2.27) (2.51-3.07) (2.91-3.88) (0.37-0.45) (0.74-0.87) (1.05-1.21) (1.36-1.56) (1.86-2.16) (2.28-2.99)
0.54 1.09 1.41 1.41 1.86 0.39 0.72 1.04 1.36 1.36 1.36
CoP 75+ 1,565 3,258
(0.27-1.08) (0.64-1.85) (0.84-2.35) (0.84-2.35) (1.00-3.44) (0.22-0.68) (0.46-1.12) (0.69-1.57) (0.90-2.05) (0.90-2.05) (0.90-2.05)
All 1.29 1.95 2.59 3.20 4.86 7.13 1.26 1.80 2.32 3.00 4.19 6.70
75+ 26,028 39,806
uncemented (1.15-1.43) (1.78-2.14) (2.37-2.83) (2.91-3.50) (4.26-5.54) (4.01-12.52) (1.15-1.37) (1.67-1.94) (2.16-2.49) (2.79-3.22) (3.80-4.61) (4.56-9.77)
1.34 2.02 2.53 2.95 3.65 7.36 1.26 1.71 2.14 2.58 3.64 4.31
MoP 75+ 17,260 27,734
(1.18-1.53) (1.80-2.25) (2.27-2.82) (2.62-3.32) (3.17-4.22) (2.75-18.90) (1.13-1.40) (1.55-1.87) (1.95-2.34) (2.35-2.84) (3.18-4.16) (3.48-5.32)
1.08 1.97 3.74 5.61 10.16 1.30 3.01 4.84 7.15 9.51
MoM 75+ 1,693 2,404
(0.68-1.71) (1.40-2.78) (2.90-4.83) (4.52-6.97) (7.92-12.98) (0.91-1.84) (2.38-3.78) (4.03-5.81) (6.12-8.34) (7.96-11.35)
1.18 1.59 2.18 2.18 3.19 1.01 1.51 1.79 2.03 2.87 3.96
CoP 75+ 3,471 4,864
(0.87-1.62) (1.20-2.11) (1.65-2.88) (1.65-2.88) (1.91-5.29) (0.76-1.34) (1.18-1.92) (1.41-2.27) (1.59-2.60) (2.13-3.86) (2.21-7.02)
1.24 1.96 2.14 2.35 4.87 1.44 1.87 2.12 2.40 3.58
CoC 75+ 3,127 4,064
(0.91-1.70) (1.51-2.53) (1.66-2.77) (1.81-3.06) (2.80-8.43) (1.12-1.86) (1.49-2.35) (1.70-2.64) (1.91-3.01) (2.33-5.49)
0.82 1.47 1.90 2.44 3.55 5.57 0.70 1.09 1.49 1.84 2.36 2.96
All hybrid 75+ 18,675 37,157
(0.70-0.97) (1.29-1.68) (1.67-2.16) (2.13-2.79) (2.95-4.28) (4.02-7.69) (0.62-0.79) (0.98-1.21) (1.35-1.65) (1.66-2.05) (2.06-2.69) (2.30-3.81)
0.79 1.48 1.92 2.40 3.50 5.87 0.74 1.13 1.50 1.78 2.33 3.03
MoP 75+ 15,515 31,303
(0.65-0.94) (1.28-1.71) (1.67-2.22) (2.07-2.78) (2.85-4.30) (4.10-8.36) (0.65-0.85) (1.01-1.26) (1.34-1.67) (1.59-1.99) (2.01-2.70) (2.28-4.02)
0.90 1.34 1.60 1.60 1.60 0.52 0.82 1.21 1.21 1.21
CoP 75+ 2,312 4,148
(0.57-1.41) (0.88-2.03) (1.06-2.44) (1.06-2.44) (1.06-2.44) (0.33-0.81) (0.55-1.21) (0.79-1.84) (0.79-1.84) (0.79-1.84)
1.31 1.56 1.93 1.93 3.66 0.33 0.91 1.18 1.18 1.74
CoC 75+ 471 941
(0.59-2.89) (0.75-3.25) (0.96-3.89) (0.96-3.89) (1.36-9.66) (0.11-1.03) (0.43-1.92) (0.57-2.44) (0.57-2.44) (0.78-3.86)
All reverse 1.06 2.01 2.30 2.42 2.79 0.75 1.22 1.65 2.14 2.25
75+ 2,510 5,082
hybrid (0.72-1.56) (1.48-2.73) (1.70-3.10) (1.79-3.28) (1.93-4.02) (0.54-1.04) (0.94-1.59) (1.28-2.13) (1.65-2.77) (1.73-2.94)
1.15 2.17 2.41 2.41 2.82 0.76 1.24 1.69 2.11 2.24
MoP 75+ 2,227 4,487
(0.77-1.71) (1.59-2.95) (1.76-3.29) (1.76-3.29) (1.91-4.16) (0.54-1.07) (0.94-1.64) (1.29-2.21) (1.59-2.78) (1.68-2.98)
All
2.01 2.01 3.99 6.29 6.29 1.61 3.43 5.33 7.53 13.69
resurfacing 75+ 202 64
(0.76-5.27) (0.76-5.27) (1.91-8.25) (3.40-11.51) (3.40-11.51) (0.23-10.90) (0.87-13.08) (1.74-15.67) (2.87-18.98) (4.98-34.58)
(MoM)
3.3.3 Revisions after primary hip shell with polyethylene liners. Figure 3.10 (b.ii) shows
the same data but with the 44mm head data truncated
surgery: effect of head size for just prior to ten years from primary operation. This is to
selected bearing surfaces/fixation allow closer inspection of the difference between the
sub-groups other head sizes. There was a statistically significant
effect of head size (overall P<0.001), with head size
This section updates results from last year’s report on 44mm showing worse failure rates, but there were
the effect of prosthesis head size on the probability of small numbers after eight years.
revision following primary surgery. In total, six bearing
groups were defined: In Figure 3.10 (c) for metal-on-metal uncemented
metal cup / metal shell with liners, there was a similar
(a) Metal-on-polyethylene cemented monobloc cups effect of head size (overall P<0.001), with head size
n=282,044 46mm having the worst failure rate during the first ten
(b) Metal-on-polyethylene uncemented metal shells years of follow-up.
with polyethylene liners n=236,122
Results were similar for ceramic-on-polyethylene
(c) Metal-on-metal uncemented metal cups or metal
cemented monobloc cups shown in Figure 3.10 (d),
shells with metal liners n=30,983
with a statistically significant difference between the
(d) Ceramic-on-polyethylene cemented monobloc head sizes overall (P=<0.001) with head size 36mm
cups n=39,710 having the worst failure rate.
(e) Ceramic-on-polyethylene uncemented metal
shells with polyethylene liners n=100,409 For ceramic-on-polyethylene metal shells used with
polyethylene liners (Figure 3.10 (e)), whilst there was
(f) Ceramic-on-ceramic uncemented metal shells a statistically significant difference between the three
with ceramic liners n=133,227 head sizes shown (P=0.016), the best implant survival
was in the intermediate size group (32mm) with 28mm
Figures 3.10 (a) to 3.10 (f) show respective percentage
and 36mm showing similar worse outcomes.
cumulative probabilities of revision (Kaplan-Meier)
for various head sizes, for each of the above groups Figure 3.10 (f) showed statistically significant
with follow-up up to 13 years following the primary differences between all four head sizes shown
operation. (P=0.002) for ceramic-on-ceramic uncemented metal
shells used with ceramic liners. Head size 40mm
In Figure 3.10 (a), for metal-on-polyethylene cemented
showed the best survival rate, though there were small
monobloc cups, there was a statistically significant
numbers in this bearing group. Head sizes 28mm,
effect of head size (overall difference P<0.001 by
32mm, and 36mm showed similar failure rates but
logrank test) on revision rates. Overall, implants with
were worse than those of head size 40mm.
head size 32mm had the worst failure rates over the
entire duration of follow-up, but implants with head
size 36mm had the worst failure rates in the first six
years of follow-up.
Figure 3.10 (b.i) shows revision rates for different head

sizes for metal-on-polyethylene uncemented metal
Figure 3.10 (a)
Effect of head size on cumulative revision rates after primary hip replacement using different bearing groups
(only head sizes used in >500 hips are shown):
(a) Metal-on-polyethylene cemented monobloc cups

7
6
5
3
2
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Head size = 22.25mm 33,604 32,120 30,587 28,893 26,823 24,437 21,938 19,282 16,505 13,665 10,558 7,546 4,506 1,742
Head size = 26mm 18,584 17,931 17,259 16,324 15,176 13,853 12,367 10,968 9,208 7,384 5,561 3,889 2,125 724
Head size = 28mm 177,774 161,098 144,101 125,736 107,688 90,373 74,086 59,138 45,446 32,216 20,391 11,852 5,371 1,667
Head size = 30mm 722 683 629 524 412 346 305 258 188 128 82 47 13 4
Head size = 32mm 46,923 37,124 28,387 20,796 14,699 9,782 6,454 4,107 2,517 1,555 859 456 220 63
Head size = 36mm 4,365 3,454 2,672 1,875 1,199 647 299 117 18 1 1 1 1 0
Figure 3.10 (b)

(b.i) Metal-on-polyethylene uncemented metal shells with polyethylene liners

45
40
35
30
25
20
15
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Head size = 22.25mm 1,510 1,217 1,013 850 702 592 501 429 374 322 259 190 113 40
Head size = 26mm 866 827 783 737 682 628 560 483 418 339 255 173 92 22
Head size = 28mm 93,202 85,875 78,663 71,043 63,530 55,633 47,420 39,078 30,829 22,468 15,011 8,954 4,096 1,185
Head size = 32mm 88,377 71,243 56,065 41,871 30,272 21,056 14,036 8,741 4,890 2,485 1,197 458 154 13
Head size = 36mm 47,765 39,734 32,878 26,287 20,283 14,404 9,371 5,346 2,472 1,060 392 134 59 12
Head size = 40mm 3,403 3,255 3,082 2,872 2,624 2,170 1,690 1,174 701 230 16 8 6 0
Head size = 44mm 842 805 773 703 606 488 382 263 155 46 0 0 0 0
(b.ii) Metal-on-polyethylene with truncated data for head size 44mm

10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Figure 3.10 (c)
(c) Metal-on-metal uncemented metal cups or metal shells with metal liners
50
45
40
35
30
25
20
15
10
5
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Head size = 28mm 2,262 2,220 2,190 2,149 2,107 1,997 1,867 1,701 1,441 1,148 803 591 331 106
Head size = 36mm 12,939 12,623 12,355 12,049 11,633 11,191 10,599 9,147 6,541 3,862 1,831 703 162 25
Head size = 38mm 1,525 1,504 1,479 1,442 1,399 1,345 1,266 1,124 976 750 456 222 36 0
Head size = 40mm 892 870 842 816 779 740 706 625 431 183 95 33 11 1
Head size = 42mm 1,191 1,171 1,143 1,114 1,089 1,040 987 857 696 452 226 92 22 4
Head size = 44mm 1,964 1,932 1,900 1,834 1,748 1,639 1,529 1,364 1,045 585 298 115 29 2
Head size = 46mm 3,389 3,334 3,260 3,154 2,992 2,826 2,622 2,349 1,886 1,198 590 201 41 13
Head size = 48mm 2,263 2,230 2,190 2,111 2,029 1,949 1,815 1,582 1,206 698 323 126 22 3
Head size = 50mm 2,248 2,214 2,164 2,094 2,027 1,940 1,846 1,651 1,290 808 381 151 40 14
Head size = 52mm 1,010 988 972 944 908 883 823 720 537 328 170 57 9 2
Head size = 54mm 684 668 653 634 615 589 561 498 394 263 139 60 17 11
Figure 3.10 (d)

(d) Ceramic-on-polyethylene cemented monobloc cups


3
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Head size = 22.25mm 3,030 2,904 2,713 2,484 2,254 2,011 1,750 1,533 1,311 1,038 747 448 177 0
Head size = 28mm 25,760 22,779 19,822 16,742 13,995 11,435 9,201 7,137 5,372 3,787 2,488 1,548 804 278
Head size = 32mm 9,801 7,738 5,842 4,152 2,907 1,906 1,212 721 378 169 103 49 19 6
Head size = 36mm 1,114 833 602 431 302 182 97 32 6 0 0 0 0 0
Figure 3.10 (e)
(e) Ceramic-on-polyethylene uncemented metal shells with polyethylene liners

4
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Head size = 28mm 28,484 25,293 22,604 19,971 17,783 15,649 13,594 11,586 9,535 7,548 5,462 3,589 1,890 640
Head size = 32mm 41,126 30,729 22,237 15,640 10,959 7,709 5,269 3,247 1,923 1,087 663 375 141 30
Head size = 36mm 30,262 22,373 15,717 10,482 6,598 3,971 2,461 1,401 643 291 114 16 0 0
Figure 3.10 (f)

(f) Ceramic-on-ceramic uncemented metal shells with ceramic liners


4
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Head size = 28mm 17,751 16,988 16,185 15,185 14,081 12,732 11,066 9,379 7,780 6,049 4,300 2,708 1,381 390
Head size = 32mm 41,867 38,425 34,616 30,144 25,417 20,284 15,348 11,315 7,929 5,154 2,970 1,431 543 134
Head size = 36mm 69,507 62,738 55,236 46,331 37,390 27,800 18,959 11,597 6,446 3,036 1,307 402 103 21
Head size = 40mm 4,089 3,793 3,473 3,018 2,410 1,604 868 279 15 0 0 0 0 0
3.3.4 Revisions after primary hip Given that the sub-groups may differ in composition
with respect to age and gender, the percentage of
surgery for the main stem-cup brand males and the median (IQR) of the ages are also
combinations shown in these tables.
Table 3.8 shows Kaplan-Meier estimates of the Sub-groups with more than 10,000 procedures
cumulative percentage probability of revision (for any in Table 3.8 have been further divided by bearing
reason) for the main stem-cup brands. surface. Table 3.9 shows the estimated cumulative
percentage probabilities for the resulting fixation/
As in previous reports, we have only included
bearing sub-groups provided there were more than
stem-cup brand combinations with more than 2,500
1,000 procedures.
procedures for cemented, uncemented, hybrid and
reverse hybrid hips or more than 1,000 in the case Note: no further sub-divisions were made for Charnley
of resurfacings. Cemented Stem/Charnley Cemented Cup, as all the
procedures described in Table 3.8 were Cemented
The figures in blue italics are at time points where fewer
MoP. Similarly, the majority of the cemented CPT/ZCA
than 250 cases remained at risk; no results are shown
and Exeter V40/Exeter Duration combinations shown
at all where the number had fallen below ten cases.
in Table 3.8 were MoP.
Table 3.8 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) at 1, 3, 5, 7, 10, and 13 years
after the primary hip replacement operation, for the most commonly used cup-stem brand combinations (group sizes >2,500,
or >1,000 in the case of resurfacings). Blue italics indicate that fewer than 250 cases remained at risk at these time points.
Median Cumulative percentage probability of revision (95% CI) at:

(IQR) age at Percentage
Stem/cup brand n primary (%) males 1 year 3 years 5 years 7 years 10 years 13 years
Cemented
Charnley Cemented 0.37 1.18 1.88 2.49 3.93 5.28
10,076 73 (67-78) 38%
Stem / Charnley Ogee (0.27-0.52) (0.98-1.42) (1.62-2.18) (2.18-2.85) (3.48-4.43) (4.57-6.10)
Charnley Cemented
0.31 1.11 1.72 2.31 3.52 4.90
Stem / Charnley 4,510 72 (66-78) 38%
(0.19-0.53) (0.83-1.46) (1.37-2.16) (1.89-2.83) (2.94-4.21) (4.02-5.98)
Cemented Cup
Charnley Cemented
0.34 0.72 1.12 1.51 2.43 2.90
Stem / Charnley and 6,590 74 (68-79) 29%
(0.22-0.51) (0.53-0.96) (0.88-1.42) (1.22-1.86) (2.01-2.94) (2.38-3.52)
Elite Plus LPW
C-Stem Cemented 0.36 0.82 1.08 1.44 2.31 2.85
4,912 72 (66-77) 40%
Stem / Elite Plus Ogee (0.22-0.57) (0.60-1.14) (0.81-1.45) (1.10-1.89) (1.78-3.00) (2.13-3.80)
C-Stem Cemented 0.37 0.96 1.28 2.07

6,025 67 (59-75) 41%
Stem / Marathon (0.24-0.57) (0.71-1.30) (0.95-1.72) (1.44-2.97)
MS-30 / Original ME 0.22 0.49 0.81 1.07 1.65 2.57
3,164 74 (67-80) 31%
Muller Low Profile C (0.11-0.47) (0.29-0.83) (0.52-1.26) (0.69-1.65) (1.01-2.70) (1.19-5.50)
Muller Straight Stem
0.46 0.88 1.13 1.94 2.34 3.10
Original / ME Muller 2,644 74 (69-79) 30%
(0.26-0.81) (0.58-1.36) (0.76-1.68) (1.36-2.77) (1.64-3.33) (2.01-4.78)
Low Profile C
Stanmore Modular
0.43 1.11 1.59 1.95 2.45 4.10
Stem / Stanmore- 5,181 75 (70-80) 29%
(0.29-0.66) (0.85-1.45) (1.26-2.00) (1.56-2.43) (1.95-3.07) (3.00-5.58)
Arcom Cup
0.65 1.42 1.90 2.39 3.15 3.53
CPT / Elite Plus Ogee 2,955 73 (67-79) 36%
(0.42-1.02) (1.04-1.93) (1.44-2.51) (1.83-3.12) (2.36-4.21) (2.53-4.91)
0.78 1.34 2.01 2.57 3.58 4.42
CPT / ZCA 12,996 76 (71-81) 30%
(0.64-0.95) (1.14-1.57) (1.74-2.32) (2.24-2.94) (3.09-4.15) (3.65-5.34)
Exeter V40 / Exeter 0.40 0.86 1.22 1.57 2.27 3.54
69,842 74 (68-79) 34%
Contemporary Flanged (0.36-0.46) (0.79-0.94) (1.12-1.32) (1.45-1.69) (2.07-2.48) (2.94-4.25)
Exeter V40 / Elite Plus 0.34 0.77 1.11 1.51 2.15 2.71
23,535 74 (69-80) 35%
Ogee (0.28-0.43) (0.66-0.90) (0.97-1.27) (1.33-1.71) (1.89-2.44) (2.26-3.25)
Exeter V40 / Exeter 0.58 1.19 1.65 2.43 3.63 5.30
16,726 73 (67-79) 32%
Duration (0.48-0.71) (1.03-1.37) (1.46-1.87) (2.17-2.72) (3.26-4.05) (4.49-6.25)
0.40 0.81 1.16 1.56 3.18 5.24
Exeter V40 / Opera 2,820 74 (68-80) 32%
(0.22-0.71) (0.53-1.23) (0.81-1.66) (1.12-2.18) (2.27-4.44) (3.58-7.64)

Exeter V40 / Cenator 0.59 1.39 2.00 2.26 2.63 5.54
2,551 75 (69-80) 32%
Cemented Cup (0.36-0.98) (0.99-1.94) (1.51-2.66) (1.72-2.97) (2.00-3.45) (3.30-9.24)
Exeter V40 / Elite Plus 0.31 0.61 0.81 1.03 1.53 2.01
4,931 73 (66-78) 34%
Cemented Cup (0.19-0.51) (0.42-0.89) (0.58-1.13) (0.75-1.42) (1.09-2.16) (1.38-2.92)
0.44 0.95 1.36 1.66
Exeter V40 / Marathon 4,387 70 (63-77) 35%
(0.28-0.71) (0.66-1.37) (0.94-1.97) (1.12-2.46)
Exeter V40 / Exeter X3 0.51 0.97 1.16
20,960 69 (62-77) 36%
Rimfit (0.42-0.62) (0.82-1.15) (0.96-1.40)
Exeter V40 / Exeter 0.82 1.55 2.11 2.80 4.20 6.12
24,433 75 (69-80) 32%
Contemporary Hooded (0.71-0.94) (1.39-1.72) (1.91-2.32) (2.55-3.08) (3.79-4.65) (5.11-7.33)
Exeter V40 / Charnley 0.65 1.32 1.57 1.85 2.37 2.59
4,422 73 (67-78) 32%
and Elite Plus LPW (0.45-0.94) (1.00-1.73) (1.22-2.03) (1.43-2.39) (1.79-3.14) (1.91-3.50)
C-Stem AMT
0.20 0.80 1.09 1.73 1.73
Cemented Stem / Elite 3,153 77 (72-81) 31%
(0.09-0.45) (0.51-1.26) (0.71-1.67) (1.14-2.64) (1.14-2.64)
Plus Ogee
C-Stem AMT Cemented 0.38 0.90 1.30 1.30
6,140 75 (69-80) 33%
Stem / Marathon (0.25-0.59) (0.63-1.29) (0.84-1.99) (0.84-1.99)
C-Stem AMT Cemented
0.60 1.17 1.48 1.83 2.57
Stem / Charnley and 2,894 75 (71-79) 32%
(0.38-0.97) (0.83-1.67) (1.07-2.04) (1.30-2.56) (1.69-3.91)
Elite Plus LPW
Uncemented
0.92 1.91 2.65 3.23 4.57 5.36
Accolade / Trident 24,868 66 (59-73) 43%
(0.81-1.05) (1.74-2.10) (2.44-2.88) (2.97-3.51) (4.05-5.16) (3.95-7.24)

Corail / Duraloc 0.75 1.69 2.50 3.57 5.59 9.69
4,044 70 (64-75) 39%
Cementless Cup (0.52-1.06) (1.33-2.14) (2.05-3.04) (3.02-4.22) (4.81-6.49) (7.97-11.74)
0.81 1.64 2.56 3.93 6.52 8.17
Corail / Pinnacle 122,635 66 (59-73) 44%
(0.76-0.86) (1.57-1.72) (2.45-2.66) (3.78-4.09) (6.20-6.87) (6.91-9.64)
0.64 1.17 1.67 2.28 3.32 4.66
Corail / Trilogy 2,883 68 (62-74) 39%
(0.41-1.02) (0.83-1.65) (1.24-2.25) (1.73-2.99) (2.44-4.50) (2.95-7.33)
Corail / ASR 1.07 7.52 23.31 35.32 44.07
2,630 61 (54-67) 54%
Resurfacing Cup (0.74-1.54) (6.57-8.60) (21.72-25.00) (33.49-37.22) (41.95-46.26)
Corail / Pinnacle 1.11 1.99 2.36
4,220 67 (58-75) 40%
Gription (0.81-1.51) (1.50-2.63) (1.76-3.16)
Furlong HAC Stem / 1.04 1.73 2.11 2.65 3.59 4.59
16,907 69 (62-76) 40%
CSF (0.90-1.21) (1.54-1.94) (1.90-2.35) (2.40-2.92) (3.27-3.94) (4.10-5.14)
Furlong HAC Stem / 1.10 1.79 2.12 2.47
20,681 66 (59-73) 44%
Furlong HAC CSF Plus (0.97-1.25) (1.61-1.99) (1.91-2.34) (2.22-2.75)
Polarstem Cementless / 0.62 0.91 0.98 0.98
6,137 66 (59-73) 45%
R3 Cementless (0.45-0.87) (0.68-1.22) (0.72-1.32) (0.72-1.32)
SL-Plus Cementless 1.21 2.62 3.82 4.52 5.91
5,218 65 (59-73) 43%
Stem / EP-Fit Plus (0.94-1.55) (2.21-3.11) (3.31-4.42) (3.94-5.19) (5.15-6.78)
Synergy Cementless 1.02 1.50 2.10 4.75
2,999 65 (56-71) 50%
Stem / R3 Cementless (0.71-1.46) (1.09-2.07) (1.49-2.95) (3.02-7.44)
Taperloc Cementless 1.10 1.52 1.84 2.15 2.15
20,700 65 (58-72) 44%
Stem / Exceed ABT (0.96-1.25) (1.35-1.70) (1.64-2.06) (1.89-2.44) (1.89-2.44)
Anthology / R3 1.01 1.58 2.45 4.57
3,474 63 (55-71) 42%
Cementless (0.72-1.42) (1.17-2.12) (1.78-3.35) (2.90-7.15)
0.83 1.50 1.65
Metafix Stem / Trinity 3,467 64 (56-69) 45%
(0.56-1.21) (1.09-2.06) (1.18-2.32)
M/L Taper Cementless / 1.15 1.81 1.96
4,820 61 (53-68) 49%
Continuum (0.88-1.50) (1.44-2.27) (1.56-2.46)
M/L Taper Cementless / 1.00 2.50 2.50 (
2,889 63 (55-70) 52%
Trilogy IT (0.69-1.46) (1.84-3.39) 1.84-3.39)
Furlong Evolution
1.34 1.97
Cementless / Furlong 2,644 62 (52-70) 42%
(0.95-1.89) (1.41-2.74)
HAC CSF Plus
Hybrid
0.88 1.41 2.24 2.71 4.23 5.65
CPT / Trilogy 17,437 72 (65-78) 35%
(0.75-1.04) (1.23-1.61) (1.98-2.54) (2.40-3.07) (3.65-4.90) (4.42-7.21)
1.58 2.25 2.72 3.13
CPT / Continuum 4,627 68 (58-76) 37%
(1.24-2.00) (1.82-2.80) (2.15-3.45) (2.37-4.11)
Continued >
1.33 2.20 2.32
CPT / Trilogy IT 5,227 68 (61-75) 37%
(1.05-1.69) (1.72-2.82) (1.80-2.98)
0.81 1.18 1.61 1.84 2.17
Exeter V40 / Pinnacle 6,437 72 (65-78) 36%
(0.61-1.07) (0.92-1.50) (1.27-2.04) (1.43-2.38) (1.61-2.93)
0.55 1.03 1.41 1.86 2.61 2.89
Exeter V40 / Trident 60,984 69 (61-76) 39%
(0.50-0.62) (0.94-1.12) (1.29-1.52) (1.71-2.02) (2.36-2.89) (2.56-3.26)
0.59 0.98 1.35 1.67 2.40 3.10
Exeter V40 / Trilogy 12,905 70 (63-76) 40%
(0.47-0.74) (0.82-1.18) (1.16-1.58) (1.44-1.94) (2.06-2.79) (2.48-3.86)
Exeter V40 / ABG II 0.31 0.80 1.23 1.78 2.53 3.66
2,622 65 (59-72) 35%
Cementless Cup (0.15-0.62) (0.52-1.24) (0.85-1.76) (1.30-2.44) (1.90-3.37) (2.70-4.96)
1.13 1.92 2.34 2.61
Exeter V40 / Tritanium 2,909 67 (60-74) 44%
(0.80-1.61) (1.42-2.59) (1.70-3.23) (1.84-3.71)
C-Stem AMT
0.72 1.28 1.88 2.12 3.96

Cemented Stem / 7,821 71 (65-77) 38%
(0.55-0.94) (1.02-1.61) (1.48-2.39) (1.62-2.75) (2.73-5.73)
Pinnacle
Reverse hybrid
0.56 1.34 1.93 2.21 2.65
Corail / Elite Plus Ogee 2,543 71 (65-77) 37%
(0.33-0.94) (0.94-1.90) (1.42-2.63) (1.63-2.99) (1.93-3.63)
0.52 1.07 1.28 1.47
Corail / Marathon 9,276 70 (64-76) 38%
(0.39-0.70) (0.85-1.34) (1.02-1.61) (1.15-1.88)
Resurfacing
1.17 2.52 4.54 6.30 8.87
Adept Resurfacing Cup 3,602 54 (48-60) 72%
(0.87-1.58) (2.05-3.09) (3.89-5.29) (5.51-7.19) (7.71-10.19)
1.63 5.97 13.58 20.79 26.60 31.30
ASR Resurfacing Cup 3,060 55 (49-60) 68%
(1.24-2.15) (5.18-6.87) (12.41-14.85) (19.38-22.28) (24.99-28.29) (27.43-35.57)
1.05 2.37 3.76 5.40 8.16 10.30
BHR Resurfacing Cup 20,974 55 (49-60) 74%
(0.92-1.20) (2.17-2.59) (3.50-4.03) (5.08-5.73) (7.74-8.60) (9.67-10.96)
Cormet 2000 1.50 3.68 7.68 12.21 17.75 22.11
3,679 55 (48-60) 65%
Resurfacing Cup (1.15-1.94) (3.12-4.34) (6.86-8.59) (11.18-13.32) (16.45-19.14) (20.11-24.29)
1.33 3.60 5.57 7.61 8.83
Durom Resurfacing Cup 1,724 55 (49-60) 70%
(0.89-2.00) (2.82-4.60) (4.58-6.77) (6.43-8.99) (7.52-10.36)
1.83 3.38 5.56 8.00 10.29
Recap Magnum 1,754 54 (49-60) 73%
(1.29-2.57) (2.63-4.34) (4.57-6.76) (6.78-9.43) (8.74-12.08)
Conserve Plus 1.94 5.09 8.21 10.92 13.56 13.56
1,345 56 (50-61) 63%
Resurfacing Cup (1.32-2.83) (4.04-6.42) (6.85-9.82) (9.34-12.75) (11.68-15.71) (11.68-15.71)
Please note: Blank cells indicate that the number at risk at the time shown has fallen below ten and thus estimates have been omitted as they are highly unreliable.
Table 3.9 Kaplan-Meier estimates of the cumulative percentage probability of revision (95% CI) at 1, 3, 5, 7, 10 and 13
years after the primary hip replacement for the most commonly used cup-stem brand combinations (group size >10,000)
with further sub-division by main bearing surface; results are shown only for the bearing surface sub-groups with >1,000
procedures. Blue italics signify that fewer than 250 cases remained at risk at these time points.
(IQR)
Bearing age at Percentage

Stem/cup brand surface n primary (%) males 1 year 3 years 5 years 7 years 10 years 13 years
Cemented
74 0.39 0.85 1.21 1.57 2.28 3.50
Exeter V40 / MoP 64,603 34%
(69-80) (0.35-0.44) (0.77-0.93) (1.11-1.31) (1.45-1.71) (2.07-2.50) (2.89-4.24)
Contemporary
Flanged 65 0.52 0.99 1.31 1.42 2.10 4.00
CoP 4,838 37%
(60-70) (0.35-0.77) (0.73-1.35) (0.98-1.75) (1.07-1.90) (1.43-3.07) (1.85-8.53)
75 0.35 0.79 1.11 1.50 2.14 2.75
MoP 21,574 34%
Exeter V40 / Elite (70-80) (0.28-0.44) (0.67-0.92) (0.97-1.28) (1.32-1.70) (1.87-2.45) (2.26-3.34)
Plus Ogee 65 0.30 0.70 1.08 1.35 2.05 2.05
CoP 1,748 44%
(59-71) (0.12-0.72) (0.39-1.26) (0.65-1.81) (0.82-2.21) (1.24-3.39) (1.24-3.39)

(IQR)
Bearing age at Percentage
Stem/cup brand surface n primary (%) males 1 year 3 years 5 years 7 years 10 years 13 years
72 0.52 0.96 1.16
MoP 14,760 34%
Exeter V40 / Exeter (66-79) (0.41-0.66) (0.79-1.17) (0.91-1.48)
X3 Rimfit 62 0.47 0.99 1.16
CoP 6,144 40%
(56-67) (0.32-0.69) (0.73-1.35) (0.85-1.58)
75 0.82 1.53 2.09 2.79 4.12 6.21
Exeter V40 / MoP 23,024 32%
(70-80) (0.71-0.95) (1.37-1.71) (1.90-2.31) (2.54-3.07) (3.71-4.57) (5.10-7.55)
Contemporary
Hooded 67 0.71 2.03 2.58 3.14 5.82 5.82
CoP 1,259 39%
(61-72) (0.35-1.41) (1.31-3.14) (1.73-3.85) (2.13-4.60) (3.82-8.82) (3.82-8.82)
Uncemented
71 0.99 2.01 2.76 3.50 5.76
MoP 11,698 41%
(65-76) (0.82-1.19) (1.77-2.30) (2.46-3.11) (3.10-3.94) (4.21-7.86)
62 0.72 1.52 1.95 2.20 3.33
Accolade / Trident CoP 5,647 45%
(56-68) (0.53-0.98) (1.20-1.91) (1.55-2.44) (1.69-2.85) (2.28-4.85)
62 0.97 2.03 2.83 3.31 4.38 5.48
CoC 7,335 46%
(55-68) (0.77-1.22) (1.73-2.38) (2.46-3.25) (2.89-3.78) (3.73-5.14) (3.62-8.24)
71 0.82 1.37 1.70 2.17 3.05
MoP 48,744 41%
(65-77) (0.74-0.91) (1.26-1.48) (1.57-1.84) (1.99-2.36) (2.71-3.43)
67 0.87 2.44 5.17 8.78 13.98
MoM 11,938 47%
(60-74) (0.72-1.05) (2.17-2.73) (4.77-5.59) (8.26-9.32) (13.18-14.83)
64 0.72 1.21 1.74 2.16 2.78
Corail / Pinnacle CoP 21,533 45%
(57-70) (0.61-0.85) (1.05-1.39) (1.50-2.00) (1.83-2.55) (2.25-3.44)
60 0.83 1.79 2.40 2.93 3.90
CoC 37,846 48%
(53-66) (0.74-0.93) (1.65-1.93) (2.24-2.58) (2.73-3.15) (3.48-4.37)
63 0.45 2.67 4.39 5.76
CoM 1,784 41%
(57-70) (0.23-0.90) (2.02-3.54) (3.52-5.47) (4.72-7.03)
73 1.29 2.07 2.43 3.06 4.41 5.04
MoP 7,873 39%
(67-78) (1.06-1.56) (1.77-2.42) (2.10-2.81) (2.67-3.50) (3.86-5.04) (4.31-5.89)
Furlong HAC / Stem 67 0.71 1.26 1.65 2.07 2.65 3.62
CoP 7,097 41%
CSF (61-73) (0.54-0.94) (1.02-1.55) (1.37-1.98) (1.74-2.45) (2.25-3.12) (2.98-4.39)
59 1.28 2.08 2.59 3.19 4.36 5.97
CoC 1,646 44%
(53-66) (0.84-1.95) (1.49-2.90) (1.92-3.49) (2.43-4.18) (3.42-5.55) (4.47-7.94)
74 1.55 2.24 2.91 3.34
MoP 5,054 39%
(70-79) (1.24-1.93) (1.85-2.72) (2.42-3.50) (2.76-4.05)
Furlong HAC Stem /
67 1.00 1.84 2.02 2.46
Furlong HAC CSF CoP 2,496 47%
(62-71) (0.67-1.49) (1.35-2.51) (1.48-2.75) (1.77-3.40)
Plus
63 0.93 1.59 1.84 2.15
CoC 13,042 46%
(56-69) (0.78-1.12) (1.38-1.83) (1.60-2.11) (1.87-2.48)
72 1.23 1.75 2.07 2.52
MoP 6,489 41%
(66-77) (0.99-1.54) (1.44-2.12) (1.71-2.52) (2.01-3.15)
Taperloc Cementless 65 0.91 1.07 1.43 2.14
CoP 3,795 45%
Stem Exceed ABT (58-70) (0.65-1.27) (0.78-1.48) (1.01-2.02) (1.40-3.25)
61 1.08 1.52 1.83 2.02 2.02
CoC 10,227 46%
(54-67) (0.90-1.31) (1.29-1.79) (1.56-2.14) (1.71-2.39) (1.71-2.39)
Hybrid
73 0.83 1.37 2.20 2.69 4.29 5.70
MoP 12,415 34%
(67-78) (0.68-1.00) (1.17-1.60) (1.92-2.52) (2.35-3.07) (3.67-5.00) (4.41-7.36)
CPT / Trilogy
68 1.02 1.47 2.48 2.48 2.48
CoP 4,866 35%
(61-75) (0.76-1.37) (1.13-1.93) (1.56-3.94) (1.56-3.94) (1.56-3.94)
73 0.58 1.10 1.44 1.85 2.71 3.18
MoP 33,916 37%
(67-79) (0.50-0.66) (0.98-1.23) (1.29-1.61) (1.65-2.08) (2.31-3.17) (2.56-3.96)
64 0.48 0.88 1.16 1.46 2.52
Exeter V40 / Trident CoP 14,200 41%
(57-71) (0.38-0.62) (0.71-1.08) (0.92-1.45) (1.13-1.90) (1.59-3.97)
59 0.57 1.01 1.50 1.99 2.59 2.81
CoC 12,175 43%
(53-65) (0.45-0.72) (0.84-1.21) (1.28-1.74) (1.73-2.29) (2.23-3.01) (2.38-3.31)
71 0.58 0.92 1.33 1.67 2.49 3.18
MoP 10,418 40%
(65-77) (0.45-0.74) (0.75-1.13) (1.11-1.59) (1.41-1.98) (2.09-2.96) (2.46-4.10)
Exeter V40 / Trilogy
63 0.57 1.05 1.31 1.58 1.91 2.71
CoP 2,318 40%
(58-68) (0.33-0.98) (0.70-1.58) (0.91-1.90) (1.11-2.23) (1.34-2.72) (1.65-4.41)
Please note: Blank cells indicate that the number at risk at the time shown has fallen below ten and thus estimates have been omitted as they are highly unreliable.
3.3.5 Revisions for different causes after primary hip surgery
Methodological note been divided by the total of the individual patient-
years at risk. The figures shown are numbers of
The preceding sections looked at first revisions for
revisions per 1,000 years at risk.
any reason. Given that several indications may have
been given for a particular revision, these will not be This method is appropriate if the hazard rate (the
mutually exclusive and so cannot be regarded as rate at which revisions occur in the unrevised cases)
‘competing risks’. remains constant. The latter is further explored
by sub-dividing the time interval from the primary
Here we have calculated incidence rates for each
operation into intervals and calculating PTIRs for
reason using patient-time incidence-rates (PTIRs);
each interval.
the total number of revisions for that reason has
Overall, 24,065 (2.7%) of the 890,681 procedures PTIRs for both aseptic loosening and pain. Metal-on-
had an associated first revision. The most commonly metal bearings have the highest incidence of adverse
cited indications were aseptic loosening (cited in 5,841 reaction to particulate debris.
procedures), pain (4,298), adverse soft tissue reaction
to particulate debris (4,103, a figure that is likely to be In Table 3.11, the PTIRs for each indication are shown
an underestimate due to changes in MDS collection), separately for different time periods from the primary
dislocation/subluxation (4,038), and infection (3,325). Pain operation, within the first year from primary operation,
was not usually cited alone; in 3,010 out of the 4,298 and between 1-3, 3-5, 5-7, 7-10 and 10+ years after
instances, it was cited together with one or more other surgery. (Note the maximum follow-up for any implant
indications. Associated PTIRs for these, and the other is now 13.75 years.) The same overall time trends are
indications are shown in Table 3.10. Here, implant wear seen as before – revision rates due to aseptic loosening
denotes either wear of the polyethylene component, wear and pain both increased with time from surgery,
of the acetabular component or dissociation of the liner. whereas the rates due to subluxation/dislocation,
infection, peri-prosthetic fracture, and mal-alignment
The number of adverse reactions to particulate debris were all higher in the first year and then fell. Adverse
is likely to be under-estimated because this was not reaction to particulate debris increased with time, as did
solicited (i.e. not an option) on the revision report forms lysis, although the PTIRs for the latter were low.
in the early phase of the study, i.e. was missing for
MDSv1/2. Some of these cases may have been put Finally, Figures 3.11 (a) to 3.11 (f) show how PTIRs
under ‘other’ but we simply do not know. Adoption of for aseptic loosening, pain, dislocation/subluxation,
the later revision report forms (MDSv3) was staggered infection and adverse soft tissue reaction to particulate
over time and so revisions associated with a few debris changed with time in an arbitrary selection
primaries as late as 2011 had revisions reported of the cemented/uncemented bearing sub-groups
on versions 1 and 2 of the data collection form. By from Table 3.10. Only sub-groups with a total overall
restricting our analyses to primaries from 2008 onwards patient-years at risk of more than >150x103 have been
however, as we did in our previous three annual reports, included. With time from operation, PTIRs for aseptic
ensures that 99.3% of revisions had been recorded loosening and pain tended to rise in uncemented
on later forms. We noted, however, that only 1,814 of metal-on-metal replacements and resurfacings. These
the 4,103 instances of adverse reactions to particulate trends were not seen in the other groups shown
debris would thus be included, i.e. we are thereby (Figures 3.11 (a) and (b)). Conversely, there was a high
missing 2,289 of the earlier cases. Therefore, as we did initial rate for dislocation/subluxation in all fixation/
last year, we present two sets of PTIRs, one set for all bearing groups which later fell (Figure 3.11 (c)). Revision
primaries, which are likely to be underestimates, and the rates for infection were initially high and then fell in all
other set for all primaries performed since the beginning groups apart from uncemented metal-on-metal (Figure
of 2008, which has better ascertainment but does not 3.11 (d)). Revision rates due to adverse reaction to
include the cases with longer term follow-up. particulate debris increased with time up to five years in
uncemented metal-on-metal and resurfacings (Figures
Table 3.10 includes further breakdowns by hip fixation 3.11 (e) and (f)). Confidence Intervals have not been
and bearing. Metal-on-metal (irrespective of type of shown here for simplicity, but could be quite wide; these
fixation) and resurfacings seem to have the highest trends require more in-depth investigation.
Table 3.10 Revision rates after primary hip replacement for each indication, expressed as numbers per 1,000 patient-years (95% CI), for all cases and by fixation
and bearing surface.
Number of revisions per 1,000 patient-years (95% CI) for: Revisions

per 1,000
patient-years
(95% CI)
for adverse
Patient- reaction to
years at risk particulate
Patient- Head/ Adverse (x1,000) for debris
Fixation/ years Peri- socket reaction to primaries for primaries
bearing at risk Aseptic Dislocation/ prosthetic Mal- Implant Implant size Other particulate from from
types (x1,000) loosening Pain subluxation Infection fracture alignment Lysis wear fracture mismatch indication debris*** 1.1.2008**** 1.1.2008****
1.28 0.94 0.89 0.73 0.67 0.39 0.29 0.28 0.16 0.04 0.52 0.90 0.66
All cases* 4,553.4 2,766.3
(1.25-1.32) (0.92-0.97) (0.86-0.91) (0.71-0.76) (0.65-0.70) (0.37-0.41) (0.27-0.30) (0.27-0.30) (0.15-0.17) (0.03-0.05) (0.50-0.54) (0.87-0.93) (0.63-0.69)
All 1.00 0.36 0.83 0.68 0.43 0.21 0.19 0.16 0.08 0.02 0.16 0.09 0.07
1,679.1 879.8
cemented (0.95-1.04) (0.33-0.39) (0.79-0.87) (0.64-0.72) (0.40-0.46) (0.19-0.24) (0.17-0.21) (0.14-0.18) (0.06-0.09) (0.01-0.02) (0.14-0.18) (0.07-0.10) (0.05-0.09)
Cemented by bearing surface
1.01 0.34 0.85 0.66 0.44 0.22 0.19 0.16 0.07 0.02 0.15 0.03 0.03
MoP 1,478.5 759.6
(0.96-1.06) (0.32-0.38) (0.81-0.90) (0.62-0.71) (0.41-0.47) (0.19-0.24) (0.17-0.21) (0.14-0.18) (0.05-0.08) (0.01-0.03) (0.13-0.17) (0.02-0.04) (0.02-0.04)
3.43 3.32 1.00 1.11 1.22 0.22 1.88 0.22 0.89 0.11 2.77 8.74 10.56
MoM 9.0 2.6
(2.41-4.88) (2.32-4.75) (0.52-1.91) (0.60-2.06) (0.67-2.20) (0.06-0.88) (1.17-3.03) (0.06-0.88) (0.44-1.77) (0.02-0.79) (1.87-4.09) (7.01-10.90) (7.24-15.40)
0.72 0.33 0.66 0.74 0.29 0.18 0.11 0.12 0.08 0.01 0.13 0.05 0.06
CoP 152.8 99.4
(0.60-0.87) (0.25-0.43) (0.54-0.80) (0.61-0.89) (0.21-0.39) (0.12-0.26) (0.07-0.18) (0.07-0.19) (0.04-0.14) (0.00-0.05) (0.08-0.20) (0.02-0.10) (0.03-0.13)
Others/ 1.01 0.44 0.62 0.93 0.41 0.18 0.23 0.13 0.18 0.21 0.41 0.22
38.8 0 18.3
unsure (0.73-1.38) (0.27-0.71) (0.41-0.92) (0.67-1.29) (0.25-0.67) (0.09-0.38) (0.12-0.45) (0.05-0.31) (0.09-0.38) (0.10-0.41) (0.25-0.67) (0.08-0.58)
All 1.61 1.18 0.98 0.79 0.74 0.54 0.31 0.43 0.24 0.06 0.70 1.48 1.02
1,700.8 1,198.3
uncemented (1.55-1.67) (1.13-1.23) (0.93-1.02) (0.75-0.83) (0.70-0.78) (0.50-0.57) (0.28-0.33) (0.40-0.46) (0.21-0.26) (0.05-0.08) (0.66-0.74) (1.42-1.54) (0.97-1.08)
Uncemented by bearing surface
1.22 0.64 1.23 0.73 0.95 0.49 0.21 0.45 0.10 0.06 0.31 0.16 0.19
MoP 617.2 443.1
(1.13-1.31) (0.58-0.71) (1.14-1.32) (0.66-0.80) (0.88-1.03) (0.44-0.55) (0.17-0.25) (0.40-0.51) (0.08-0.13) (0.04-0.08) (0.27-0.36) (0.13-0.20) (0.15-0.23)
3.67 4.29 0.92 1.43 0.67 0.91 1.21 0.69 0.18 0.09 2.64 10.00 9.55
MoM 224.6 106.7
(3.43-3.93) (4.03-4.57) (0.80-1.06) (1.28-1.59) (0.57-0.79) (0.80-1.05) (1.08-1.36) (0.59-0.81) (0.13-0.24) (0.06-0.14) (2.44-2.87) (9.60-10.43) (8.98-10.15)
1.19 0.56 1.03 0.67 0.61 0.43 0.15 0.40 0.10 0.05 0.34 0.08 0.10
CoP 264.8 178.4
(1.06-1.32) (0.47-0.65) (0.91-1.16) (0.58-0.77) (0.52-0.71) (0.36-0.52) (0.11-0.21) (0.33-0.48) (0.07-0.14) (0.03-0.08) (0.27-0.41) (0.05-0.13) (0.06-0.16)
1.36 0.79 0.70 0.66 0.58 0.47 0.11 0.31 0.46 0.06 0.52 0.16 0.16
CoC 546.5 438.9
(1.26-1.46) (0.72-0.87) (0.63-0.77) (0.59-0.73) (0.52-0.65) (0.42-0.54) (0.09-0.15) (0.26-0.36) (0.41-0.53) (0.04-0.09) (0.46-0.58) (0.13-0.20) (0.13-0.21)
2.81 1.80 0.72 1.29 0.43 0.86 0.58 0.79 0.22 0.14 1.22 1.51 1.49
CoM 13.9 13.4
(2.05-3.84) (1.22-2.66) (0.39-1.34) (0.82-2.06) (0.19-0.96) (0.49-1.52) (0.29-1.15) (0.44-1.43) (0.07-0.67) (0.04-0.58) (0.76-1.97) (0.99-2.32) (0.96-2.32)
Others/ 1.89 1.06 0.89 0.56 0.86 0.56 0.30 0.38 0.38 0.09 0.53 1.06 0.79
33.8 17.7
unsure (1.48-2.42) (0.77-1.48) (0.62-1.27) (0.36-0.88) (0.60-1.23) (0.36-0.88) (0.16-0.55) (0.22-0.66) (0.22-0.66) (0.03-0.28) (0.34-0.85) (0.77-1.48) (0.47-1.33)
*Including 39 with unknown fixation/bearing.

**Based on a small group size (n=97), therefore estimates are unreliable.
***Rates likely to be underestimated: this reason not solicited in the early phase of the registry (i.e. revision report forms MDSv1/MDSv2).
****For primaries from 2008 onwards the majority of revision report forms were MDSv3/MDSv6 which explicitly stated this reason for revision.
Continued >
73
74
Table 3.10 (continued)

per 1,000
patient-years
(95% CI)
for adverse
Patient- reaction to
years at risk particulate
Patient- Head/ Adverse (x1,000) for debris
Fixation/ years Peri- socket reaction to primaries for primaries
bearing at risk Aseptic Dislocation/ prosthetic Mal- Implant Implant size Other particulate from from
types (x1,000) loosening Pain subluxation Infection fracture alignment Lysis wear fracture mismatch indication debris*** 1.1.2008**** 1.1.2008****
0.65 0.49 1.04 0.77 0.84 0.30 0.20 0.23 0.15 0.03 0.30 0.24 0.16
All hybrid 758.6 492.9
(0.59-0.71) (0.45-0.55) (0.97-1.12) (0.71-0.84) (0.78-0.91) (0.26-0.34) (0.17-0.23) (0.20-0.27) (0.13-0.18) (0.02-0.04) (0.26-0.34) (0.21-0.28) (0.13-0.20)
Hybrid by bearing surface
0.64 0.38 1.14 0.79 0.93 0.30 0.18 0.25 0.11 0.02 0.25 0.06 0.06
MoP 481.1 310.8
(0.57-0.71) (0.33-0.44) (1.05-1.24) (0.71-0.87) (0.85-1.02) (0.25-0.35) (0.15-0.23) (0.21-0.30) (0.09-0.15) (0.01-0.04) (0.20-0.29) (0.05-0.09) (0.04-0.10)
3.59 3.93 1.46 1.23 1.57 0.67 1.57 0.28 0.28 0.11 2.52 7.29 7.13
MoM 17.8 6.6
(2.81-4.59) (3.11-4.96) (0.99-2.14) (0.81-1.87) (1.08-2.27) (0.38-1.19) (1.08-2.27) (0.12-0.67) (0.12-0.67) (0.03-0.45) (1.88-3.38) (6.14-8.66) (5.36-9.49)
0.38 0.31 1.15 0.87 0.85 0.20 0.13 0.19 0.09 0.04 0.21 0.04 0.03
CoP 111.7 88.2
(0.29-0.52) (0.22-0.44) (0.97-1.37) (0.71-1.06) (0.70-1.04) (0.13-0.30) (0.07-0.21) (0.12-0.29) (0.05-0.17) (0.01-0.10) (0.14-0.32) (0.02-0.11) (0.01-0.11)
0.52 0.53 0.53 0.54 0.47 0.35 0.10 0.17 0.32 0.04 0.26 0.10 0.10
CoC 132.7 80.5
(0.41-0.66) (0.42-0.68) (0.42-0.68) (0.43-0.68) (0.36-0.60) (0.26-0.46) (0.06-0.17) (0.12-0.26) (0.24-0.44) (0.02-0.09) (0.19-0.37) (0.06-0.17) (0.05-0.20)
Others/ 0.66 0.92 0.99 1.05 0.33 0.13 0.26 0.26 0.20 0.20 0.20 0.15
15.2 0 6.8
unsure (0.35-1.22) (0.55-1.56) (0.60-1.64) (0.65-1.72) (0.14-0.79) (0.03-0.53) (0.10-0.70) (0.10-0.70) (0.06-0.61) (0.06-0.61) (0.06-0.61) (0.02-1.05)
All reverse 1.31 0.49 0.91 0.86 0.69 0.30 0.15 0.27 0.05 0.04 0.31 0.08 0.06
96.8 77.6
hybrid (1.10-1.56) (0.36-0.65) (0.74-1.12) (0.69-1.06) (0.54-0.88) (0.21-0.43) (0.09-0.26) (0.18-0.39) (0.02-0.12) (0.02-0.11) (0.22-0.44) (0.04-0.17) (0.03-0.15)
Reverse hybrid by bearing surface
1.21 0.32 1.01 0.81 0.83 0.31 0.14 0.21 0.05 0.03 0.31 0.09 0.06
MoP 65.2 52.3
(0.97-1.51) (0.21-0.49) (0.80-1.29) (0.62-1.06) (0.63-1.08) (0.20-0.48) (0.07-0.27) (0.13-0.36) (0.01-0.14) (0.01-0.12) (0.20-0.48) (0.04-0.20) (0.02-0.18)
1.48 0.81 0.71 0.94 0.42 0.26 0.19 0.39 0.06 0.06 0.29 0.03 0.04
CoP 31.0 25.0
(1.11-1.98) (0.54-1.19) (0.47-1.08) (0.65-1.35) (0.24-0.72) (0.13-0.52) (0.09-0.43) (0.22-0.68) (0.02-0.26) (0.02-0.26) (0.15-0.56) (0.00-0.23) (0.01-0.28)
Others/ 3.25 1.63 1.63 1.63 1.63 1.63 2.72
0.6 0 0 0 0 0 0 0.4
unsure** (0.81-13.01) (0.23-11.55) (0.23-11.55) (0.23-11.55) (0.23-11.55) (0.23-11.55) (0.38-19.34)
All
2.57 3.99 0.33 0.54 1.21 0.75 0.93 0.30 0.29 0.08 2.04 3.95 3.80
resurfacing 317.6 117.8
(2.40-2.75) (3.78-4.21) (0.28-0.40) (0.47-0.63) (1.09-1.33) (0.66-0.85) (0.83-1.04) (0.25-0.37) (0.23-0.35) (0.05-0.11) (1.89-2.20) (3.74-4.17) (3.46-4.16)
(MoM)
*Including 39 with unknown fixation/bearing.

**Based on a small group size (n=97), therefore estimates are unreliable.
***Rates likely to be underestimated: this reason not solicited in the early phase of the registry (i.e. revision report forms MDSv1/MDSv2).
****For primaries from 2008 onwards the majority of revision report forms were MDSv3/MDSv6 which explicitly stated this reason for revision.
Table 3.11 Revision rates after primary hip replacement for each indication, expressed as numbers per 1,000 patient-years (95% CI), overall and by time
interval from primary operation.

per 1,000
patient-
years (95%
CI) for
adverse
reaction to
Patient- particulate
years at risk debris
Time Patient- Head/ Adverse (x1,000) for for
from years Peri- socket reaction to primaries primaries
primary at risk Aseptic Dislocation/ prosthetic Mal- Implant Implant size Other particulate from from
operation (x1,000) loosening Pain subluxation Infection fracture alignment Lysis wear fracture mismatch indication debris*** 1.1.2008*** 1.1.2008***
1.28 0.94 0.89 0.73 0.67 0.39 0.29 0.28 0.16 0.04 0.52 0.90 0.66
All cases 4,553.4 2,766.3
(1.25-1.32) (0.92-0.97) (0.86-0.91) (0.71-0.76) (0.65-0.70) (0.37-0.41) (0.27-0.30) (0.27-0.30) (0.15-0.17) (0.03-0.05) (0.50-0.54) (0.87-0.93) (0.63-0.69)
Years
1.13 0.66 2.37 1.54 1.67 0.77 0.08 0.35 0.23 0.11 0.72 0.09 0.11
<1 year 836.2 646.4
(1.06-1.20) (0.61-0.72) (2.27-2.48) (1.46-1.63) (1.59-1.76) (0.71-0.83) (0.06-0.10) (0.31-0.39) (0.20-0.27) (0.09-0.13) (0.67-0.78) (0.07-0.11) (0.09-0.14)
1.13 0.91 0.64 0.75 0.34 0.35 0.16 0.16 0.13 0.03 0.42 0.29 0.36
1 - 3 years 1,374.4 1,006.2
(1.08-1.19) (0.86-0.96) (0.60-0.68) (0.70-0.79) (0.31-0.37) (0.32-0.38) (0.14-0.19) (0.14-0.18) (0.11-0.15) (0.03-0.05) (0.39-0.46) (0.26-0.32) (0.33-0.40)
1.14 1.09 0.47 0.48 0.41 0.27 0.27 0.24 0.13 0.02 0.50 1.02 0.95
3 - 5 years 1,003.3 653.7
(1.08-1.21) (1.03-1.16) (0.43-0.51) (0.44-0.52) (0.37-0.45) (0.24-0.30) (0.24-0.30) (0.21-0.27) (0.11-0.15) (0.01-0.03) (0.46-0.55) (0.96-1.08) (0.88-1.03)
1.40 1.20 0.46 0.40 0.52 0.28 0.44 0.32 0.16 0.02 0.54 1.97 1.54
5 - 7 years 679.5 353.0
(1.32-1.49) (1.12-1.28) (0.41-0.51) (0.36-0.45) (0.47-0.58) (0.25-0.33) (0.39-0.49) (0.28-0.37) (0.13-0.20) (0.01-0.04) (0.49-0.60) (1.87-2.08) (1.41-1.67)
7 - 10 1.80 0.98 0.61 0.40 0.61 0.28 0.66 0.42 0.17 0.02 0.53 2.10 1.98
524.7 107.0
years† (1.69-1.92) (0.90-1.07) (0.54-0.68) (0.35-0.46) (0.55-0.68) (0.23-0.33) (0.59-0.73) (0.36-0.47) (0.14-0.21) (0.01-0.03) (0.47-0.60) (1.98-2.23) (1.73-2.27)
10+ 2.19 0.48 0.60 0.38 0.84 0.24 0.76 0.78 0.26 0.01 0.24 1.29
135.3
years* (1.96-2.46) (0.38-0.61) (0.48-0.74) (0.29-0.50) (0.69-1.00) (0.17-0.33) (0.63-0.92) (0.65-0.95) (0.19-0.36) (0.00-0.06) (0.17-0.33) (1.12-1.50)
*Current maximum observed follow up is 13.75 years.

**Rates likely to be underestimated: this reason not solicited in the early phase of the registry (i.e. revision report forms MDSv1/MDSv2).
***For primaries from 2008 onwards the majority of revision report forms were MDSv3/MDSv6 which explicitly stated this reason for revision.
† 7 - 9 years for adverse reaction to particulate debris.
75
Figure 3.11 (a)
Change in PTIR with time from primary hip replacement, for aseptic loosening for selected
fixation/bearing sub-groups.
0−1y
1−3y
3−5y
(i) Cemented MoP 5−7y
7−10y
10+y
0−1y
1−3y
3−5y
(ii) Uncemented MoP 5−7y
7−10y
10+y
0−1y
1−3y
3−5y
(iii) Uncemented MoM 5−7y
7−10y
10+y
0−1y
1−3y
3−5y
(iv) Uncemented CoC 5−7y
7−10y
10+y
0−1y
1−3y
3−5y
(v) Resurfacing 5−7y
7−10y
10+y
0 1 2 3 4 5
PTIR (per 1,000 patient−years)
Figure 3.11 (b)

Change in PTIR with time from primary hip replacement, for pain for selected fixation/bearing sub-groups.
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0 2 4 6
Figure 3.11 (c)

Change in PTIR with time from primary hip replacement, for dislocation/subluxation for selected
0−1y
1−3y
3−5y
7−10y
10+y

0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0 1 2 3 4
Figure 3.11 (d)

Change in PTIR with time from primary hip replacement, for infection for selected
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0 0.5 1 1.5 2
Figure 3.11 (e)
Change in PTIR with time from primary hip replacement, for adverse soft tissue reaction to particulate
debris for selected fixation/bearing sub-groups.
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0−1y
1−3y
3−5y
7−10y
10+y
0 5 10 15 20
Figure 3.11 (f)

Change in PTIR with time from primary hip replacement, for adverse soft tissue reaction to particulate
debris for selected fixation/bearing sub-groups including primaries since 2008 only.
0−1y
1−3y
5−7y
0−1y
1−3y
5−7y
0−1y
1−3y
5−7y
0−1y
1−3y
5−7y
0−1y
1−3y
5−7y
0 5 10 15
3.3.6 Mortality after primary operations, with the left and right side operated on
the same day; here the second of the two has been
hip surgery excluded, leaving 886,168 procedures, of whom
This section describes the mortality of the cohort up to 113,030 had died before the end of 2016.
13 years from primary operation, according to gender
Table 3.12 shows Kaplan-Meier estimates of cumulative
and age group. Deaths were updated on 28 February
percentage mortality at 30 days, 90 days and at 1, 3, 5,
2017 using data from the NHS Personal Demographic
7, 10, 11 and 13 years from the primary operation, for
Service. A total of 207 cases were excluded because
all cases and by age and gender.
the NHS number was not traceable and, therefore,
the ages could not be verified. One additional record Note: These cases were not censored when further
was excluded as age was missing and one further revision surgery was undertaken. Whilst such surgery
record was excluded because of uncertainty in gender, may have contributed to the overall mortality, the
leaving 890,472. Amongst these, were 4,304 bilateral impact of this is not investigated in this section.
Table 3.12 Kaplan-Meier estimates of the cumulative percentage mortality (95% CI), at different time points after
primary hip replacement, for all cases and by age/gender.
Cumulative percentage probability of death (95% CI) at:
n 30 days 90 days 1 year 3 years 5 years 7 years 10 years 11 years 13 years

0.23 0.48 1.48 4.91 9.53 15.11 24.94 28.45 35.64
All cases 886,168*
(0.22-0.24) (0.47-0.50) (1.46-1.51) (4.86-4.95) (9.45-9.60) (15.01-15.21) (24.78-25.09) (28.27-28.64) (35.33-35.96)
Males
0.08 0.16 0.51 1.36 2.21 3.30 4.93 5.63 6.88
<55 years 53,052
(0.06-0.11) (0.13-0.20) (0.45-0.57) (1.26-1.47) (2.07-2.36) (3.11-3.50) (4.64-5.22) (5.28-5.99) (6.31-7.50)
55-59 0.06 0.19 0.63 1.86 3.28 5.05 8.36 9.96 13.30
36,410
years (0.04-0.09) (0.15-0.25) (0.55-0.71) (1.71-2.02) (3.07-3.50) (4.77-5.35) (7.92-8.83) (9.41-10.55) (12.29-14.39)
60-64 0.13 0.25 0.86 2.64 4.81 7.30 12.56 14.45 18.60
52,092
years (0.10-0.16) (0.21-0.30) (0.78-0.94) (2.50-2.79) (4.61-5.03) (7.02-7.59) (12.09-13.04) (13.89-15.03) (17.65-19.60)
65-69 0.17 0.37 1.14 3.61 6.92 11.07 18.78 21.67 29.51
61,922
years (0.14-0.21) (0.32-0.42) (1.06-1.23) (3.45-3.78) (6.68-7.16) (10.74-11.40) (18.26-19.32) (21.04-22.31) (28.36-30.70)
70-74 0.22 0.47 1.67 5.60 10.67 16.92 29.39 34.23 44.83
61,168

years (0.18-0.26) (0.42-0.53) (1.57-1.78) (5.41-5.81) (10.39-10.97) (16.54-17.32) (28.77-30.02) (33.49-35.00) (43.51-46.18)
75-79 0.42 0.78 2.51 8.63 16.96 27.75 46.23 52.96 66.38
50,628
years (0.37-0.48) (0.70-0.86) (2.38-2.66) (8.37-8.90) (16.57-17.36) (27.22-28.30) (45.41-47.06) (51.98-53.95) (64.66-68.09)
80-84 0.82 1.53 4.27 13.57 27.08 42.76 66.42 72.78 83.81
28,564
years (0.72-0.93) (1.39-1.68) (4.03-4.51) (13.13-14.02) (26.44-27.73) (41.93-43.60) (65.30-67.54) (71.53-74.02) (81.84-85.67)
1.73 3.09 7.85 23.82 44.00 63.34 85.77 90.30 95.72
85+ years 11,987
(1.51-1.98) (2.80-3.42) (7.37-8.36) (22.98-24.69) (42.90-45.13) (62.08-64.60) (84.42-87.05) (88.91-91.58) (94.05-97.03)
Females
0.06 0.21 0.67 1.63 2.49 3.45 4.93 5.39 6.37
<55 years 53,208
(0.04-0.09) (0.18-0.26) (0.60-0.74) (1.52-1.75) (2.34-2.65) (3.26-3.66) (4.65-5.22) (5.06-5.73) (5.85-6.93)
55-59 0.07 0.19 0.59 1.71 3.03 4.47 6.96 7.90 9.71
42,208
years (0.05-0.10) (0.15-0.24) (0.52-0.67) (1.58-1.84) (2.84-3.23) (4.22-4.73) (6.58-7.35) (7.46-8.37) (9.02-10.45)
60-64 0.07 0.17 0.60 2.02 3.76 5.68 9.44 11.06 14.87
65,199
years (0.05-0.10) (0.14-0.20) (0.54-0.66) (1.90-2.14) (3.59-3.94) (5.46-5.91) (9.08-9.82) (10.60-11.53) (14.02-15.76)
65-69 0.08 0.23 0.76 2.55 4.82 7.70 13.66 15.98 21.48
90,303
years (0.07-0.11) (0.20-0.27) (0.71-0.82) (2.44-2.67) (4.66-4.99) (7.47-7.94) (13.28-14.06) (15.51-16.46) (20.60-22.38)
70-74 0.12 0.29 0.96 3.53 7.19 11.79 21.63 25.58 34.74
98,606
years (0.10-0.14) (0.26-0.32) (0.90-1.03) (3.41-3.66) (7.00-7.38) (11.53-12.07) (21.18-22.09) (25.03-26.14) (33.71-35.79)
75-79 0.24 0.47 1.53 5.61 11.68 19.34 34.85 40.73 52.70
90,065
years (0.21-0.27) (0.43-0.52) (1.45-1.62) (5.45-5.78) (11.43-11.94) (18.99-19.69) (34.29-35.42) (40.05-41.41) (51.51-53.90)
80-84 0.37 0.84 2.58 9.07 18.69 31.48 53.57 61.47 74.63
59,862
years (0.32-0.42) (0.77-0.91) (2.45-2.71) (8.82-9.32) (18.31-19.07) (30.97-32.00) (52.82-54.32) (60.60-62.33) (73.24-75.99)
0.82 1.81 4.89 16.24 32.12 50.33 74.08 80.36 90.19
85+ years 30,894
(0.72-0.92) (1.66-1.96) (4.65-5.15) (15.79-16.70) (31.48-32.76) (49.55-51.12) (73.12-75.04) (79.29-81.41) (88.63-91.61)
* Excludes 208 cases where the age could not be verified (because NHS number was not traceable or age was missing) plus one further case with uncertain
gender; amongst the remainder, the second of 4,304 pairs of simultaneous bilateral operations were also excluded.
3.3.7 Primary hip replacement for as a result of fractured neck of femur compared to
cases implanted for other reasons. A total of 24,609
fractured neck of femur compared (2.8%) of the primary total hip replacements were
with other reasons for implantation performed for fracture of the neck of femur (#NOF)2.
As total hip replacement is becoming an increasingly Table 3.13 below shows that the proportion of primary
popular treatment option for fractured neck of femur; hip replacements due to fractured neck of femur has
this section updates results from last year’s annual continued to increase with time to a maximum of 4.9%
report (13th Annual Report 2016) on revision and in 2016.
mortality rates for primary hip replacements performed
Table 3.13 Proportions of primary total hip replacements for fracture of the neck of femur by year of primary operation.
Year of primary n Number (%) with fractured neck of femur

2003 14,452 142 (1.0%)
2004 28,057 292 (1.0%)
2005 40,573 390 (1.0%)
2006 48,470 529 (1.1%)

2007 60,751 773 (1.3%)
2008 67,124 863 (1.3%)
2009 68,101 1,074 (1.6%)
2010 70,618 1,361 (1.9%)
2011 73,631 1,706 (2.3%)
2012 77,775 2,433 (3.1%)
2013 79,885 3,115 (3.9%)
2014 86,977 3,716 (4.3%)
2015 86,496 3,955 (4.6%)
2016 87,733 4,260 (4.9%)
All years 890,643* 24,609 (2.8%)
*Excludes 38 with no data.
Table 3.14 compares the #NOF group with the in age and gender, as stratification by these variables
remainder with respect to gender and age composition left the result unchanged (P<0.001 using stratified
together and type of hip replacement received. A logrank test: 14 sub-groups of age <55, 55-59, 60-64,
significantly larger percentage of the #NOF cases 65-69, 70-74, 75-79, 80+ for each gender).
compared with the remainder were women (72.9%
versus 59.4%: P<0.001, Chi-squared test). The #NOF Finally Figure 3.13 shows a marked worse overall
cases were significantly older (median age 73 years survival in the #NOF cases compared to cases
versus 69 years at operation: P<0.001 by Mann- implanted for other reasons (P<0.001, logrank test).
Whitney U-test). Cemented and hybrid hips were used As in the overall mortality section above, 208 cases
more commonly in #NOF than in the other group. with untraced NHS numbers or missing age have been
excluded, together with 4,304 cases that were the
Figure 3.12 shows that the overall failure rate second of simultaneous bilateral procedures. Gender/
(cumulative revision) was higher in the #NOF group age differences did not fully explain the difference
compared with the remainder (P<0.001, logrank test). seen as a stratified analysis still showed a difference
This effect appeared not to be explained by differences (P<0.001) but the results warrant further exploration.
2 These comprised 2,224 cases with reasons for primary including fractured neck of femur in the early phase of the registry (i.e. 200,900 implants entered
using MDSv1 and v2) and 22,385 cases with reasons including acute trauma neck of femur in the later phase (i.e. 689,743 entered using MDSv3 and v6). 39
cases were omitted as no reasons were given.
Table 3.14 Comparison between primary hip replacements for fractured neck of femur and the remainder of cases
with respect to gender, age and type of primary hip received.
Reason for primary hip replacement

Fractured neck of femur Other reasons
(n=24,609) (n=866,034) Comparison
% Females* 72.9% 59.4% P<0.001 (Chi-squared test)

Median age (IQR)**
Both genders 73 (IQR 66-79) 69 (IQR 61-76) P<0.001 (Mann-Whitney U-test)
Males only 72 (IQR 65-79) 67 (IQR 59-75) P<0.001 (Mann-Whitney U-test)
Females only 73 (IQR 66-79) 70 (IQR 62-77) P<0.001 (Mann-Whitney U-test)
% Hip type***
Cemented 44.2% 34.6%
Uncemented 23.2% 39.5%
Hybrid 29.9% 18.8% Overall P<0.001 (Chi-squared test)
Reverse hybrid 2.5% 2.5%
Resurfacing 0.1% 4.5%
*Excludes one with uncertain gender.

**Excludes 208 whose NHS number was untraced whose ages, therefore, could not be verified.
***Excludes 39 with uncertain hip type.
Figure 3.12
Cumulative percentage revision rates (Kaplan-Meier) for hip primaries implanted for fractured neck of femur
compared with all other cases.
8
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Other 866,034 766,088 672,197 578,674 492,753 410,757 334,522 264,433 200,374 141,864 92,582 55,552 27,004 8,627
Fract. neck of femur 24,609 18,804 14,389 10,490 7,461 5,174 3,677 2,538 1,696 1,110 632 350 174 49
Figure 3.13
Cumulative percentage mortality (Kaplan-Meier) for hip primaries implanted for fractured neck of femur
compared with all other cases.
Cumulative percentage mortality (%)
60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Number at risk
Other 861,564 767,783 676,257 584,371 499,643 418,634 343,165 273,302 208,555 148,254 97,011 58,247 28,328 9,033
Fract. neck of femur 24,568 19,023 14,608 10,709 7,641 5,313 3,787 2,646 1,774 1,174 681 377 188 55
3.3.8 Conclusions Previous trends of implant usage have become more

pronounced with time. Since 2010 the use of ceramic-
As in previous annual reports, we have analysed on-polyethylene bearings has steadily increased with
implants by revision of the construct, rather than a corresponding decline in the use of ceramic-on-
revision of a single component, as the mechanisms of ceramic bearings. This is possibly in response to the
failure (such as wear, aseptic lymphocyte-dominated growing body of evidence from the NJR showing the
vasculitis-associated lesion (ALVAL) and dislocation) are very low failure rates associated with ceramic-on-
interdependent between different parts of the construct. polyethylene bearings.
We have also stratified revision by age and gender. The
highest failure rates are among young women and the The proportion of implants with uncemented fixation
lowest among older women. When data on metal-on- increased from 17% in 2003 to 46% in 2010 and
metal is excluded, young women have similar revision thereafter declined to 39% in 2016. Meanwhile the
rates to young men. Once again we must emphasise proportion of implants with hybrid fixation is steadily
that implant survivorship is only one measure of increasing from 12% in 2003 to 28% in 2016.
success and cannot be used as an indication of
This year, for the first time, we have presented data by
satisfaction, relief of pain, improvement in function
age and gender comparing combinations of fixation
and greater participation in society. Interestingly, the
and bearing. This will assist clinicians and patients in
breakdowns by age and gender show that cemented
choosing classes of prostheses that are effective for
fixation has the lowest implant revision rate at ten years
their particular age and gender and makes interesting
in all age bands and both genders.
reading. For example, in males under 55 years of
Overall the number of primary hip replacements age, cemented ceramic-on-polyethylene constructs
recorded annually in the NJR continues to increase and have nearly half the revision rate of cementless metal-
in 2016 over 87,000 were performed in England, Wales, on-polyethylene constructs at all time points. Hybrid
Northern Ireland and the Isle of Man. ceramic-on-polyethylene and hybrid ceramic-on-
ceramic implants have similar outcomes to cemented remains contrary to MHRA and manufacturers
ceramic-on-polyethylene. Uncemented hips with metal- guidelines for usage. For a more detailed analysis of
on-polyethylene, uncemented metal-on-metal bearings, this question please see Tucker et al. published open
hybrid metal-on-polyethylene, and resurfacings all have access in Acta Orthopaedica4.
statistically significantly worse survivorship at ten years.
It is encouraging that the most commonly used
In women under 55 years of age, cemented ceramic- constructs by brand in cemented and hybrid fixation
on-polyethylene gives excellent results with 3.79% have good results. This does not hold true for
(95% CI 2.26-5.35) revision rate at ten years. However, uncemented fixation, but further breakdown by bearing
cemented metal-on-polyethylene has a higher revision type for commonly used uncemented implants shows
rate, whilst results with uncemented constructs with that results are acceptable if metal-on-metal bearings
metal-on-polyethylene, ceramic-on-polyethylene are excluded.
and ceramic-on-ceramic are not statistically different
to those achieved by cemented ceramic-on- Metal-on-metal stemmed and resurfacing implants
polyethylene. In contrast for patients over 75 years old, continue to fail at higher than expected rates and their
all combinations except those with metal-on-metal use is now extremely rare. The best performing brands
bearings have good outcomes, with cemented and of resurfacing have failure rates greater than 8% at
hybrid ceramic-on-polyethylene possibly having the ten years. It is striking to note the high rates of revision
lowest failure rates. for adverse soft tissue reaction to particulate debris in
patients who have received metal-on-metal bearings.
Considering these ten-year implant survival rates Analysis of stemmed metal-on-metal bearings by head
alongside ten-year mortality rates in our patient size shows that 28mm heads have the best survivorship,
population shows that in older patients the vast majority but this is still poor compared to alternatives.
of treatment strategies will last the rest of the patients’
lives. We now have mortality data out to 13 years post- Revision rates by year of surgery for the entire cohort
surgery and this shows that the majority of patients increased dramatically from 2003 to 2008 and
aged over 75 years have died by 13 years regardless of then declined. This matches the use of resurfacing
gender. Even in those aged 65 to 69 years at the time arthroplasty and stemmed metal-on-metal with the peak
of surgery, only 70% of males and 79% of females are usage of these devices in 2008 corresponding with the
still alive 13 years later. highest failure rates by year of primary surgery. This
demonstrates the profoundly negative effect metal-on-
We have examined head sizes (bearing diameters) with metal has had on hip arthroplasty outcomes.
different fixation and bearing types and again these
results are interesting. With metal-on-polyethylene and Consistent with results from previous years’ reports,
ceramic-on-polyethylene, large head sizes appear to similar revision rates were observed for total hip
be associated with higher failure rates particularly with replacement performed as a result of fractured neck of
36mm heads used with cemented fixation and heads femur and those done for other causes. As expected,
>36mm used with hybrid and uncemented fixation. mortality rates were higher for the fractured neck of
Ceramic-on-ceramic bearings have lower failure rates femur group.
with larger bearings as predicted by Alison Smith’s
flexible parametric survival models published in the
Lancet in 20123.
With regard to specific branded stem-cup

combinations some of the best implant survivorship
are still achieved by “mix and match” cemented hard-
on-soft bearing constructs, although this practice
3 Smith AJ, Dieppe P, Vernon K, Porter M, Blom AW; National Joint Registry of England and Wales. Failure rates of stemmed metal-on-metal hip
replacements: analysis of data from the National Joint Registry of England and Wales. Lancet. 2012 Mar 31;379(9822):1199-204
4 Tucker K, Pickford M, Newell C, Howard P, Hunt LP, Blom AW. Mixing of components from different manufacturers in total hip arthroplasty: prevalence and
comparative outcomes. Acta Orthop. 2015;86(6):671-7
Part 3
3.4 Revisions
of a total hip
replacement
3.4.1 Overview of hip revision stage one and stage two are entered into the database
separately, whereas stage one and stage two revisions
procedures in practice have to be linked. In some cases, stage one
This section looks at all hip revision procedures revisions have been entered without stage two, and
performed since the start of the registry, 1 April 2003, up vice versa, making identification of individual revision
to 31 December 2016, for all patients with valid patient episodes difficult. An attempt has been made to do this
identifiers (i.e. whose data could therefore be linked). later in this section.
In total there were 97,341 revisions on 85,200 individual Table 3.15 below gives an overview of all revision
patient-sides5 (80,462 actual patients). In addition to procedures carried out each year since April 20035.
revisions on the 24,103 revised primaries described There were up to a maximum of nine documented
in section 3.3 of this report, there were revisions revision procedures associated with any individual
associated with 61,097 unrecorded primaries. patient-side (discussed later in this section). The
temporal increase reflects the increasing number of
Revisions are classified as single stage and stage one at-risk implants prevailing in the database.
and stage two of two-stage revisions. Information on
Table 3.15 Numbers of all hip revision procedures, by type of procedure, carried out each year.
Type of revision procedure

Year of revision Stage one Stage two
surgery Single stage of two-stage of two-stage All procedures
2003* 1,430 (100.0%) ---- ---- 1,430
2004 2,434 (90.0%) 117 (4.3%) 154 (5.7%) 2,705
2005 3,406 (87.0%) 206 (5.3%) 303 (7.7%) 3,915

2006 4,166 (86.6%) 267 (5.6%) 375 (7.8%) 4,808
2007 5,515 (87.2%) 347 (5.5%) 461 (7.3%) 6,323
2008 5,995 (86.0%) 424 (6.1%) 551 (7.9%) 6,970
2009 6,283 (84.2%) 523 (7.0%) 654 (8.8%) 7,460
2010 7,068 (86.6%) 500 (6.1%) 591 (7.2%) 8,159
2011 7,989 (87.6%) 529 (5.8%) 606 (6.6%) 9,124
2012 9,218 (88.1%) 602 (5.8%) 648 (6.2%) 10,468
2013 8,516 (87.8%) 564 (5.8%) 619 (6.4%) 9,699
2014 8,315 (87.0%) 658 (6.9%) 581 (6.1%) 9,554
2015 7,582 (86.2%) 645 (7.3%) 566 (6.4%) 8,793
2016 6,966 (87.8%) 483 (6.1%) 484 (6.1%) 7,933
All years 84,883 (87.2%) 5,865 (6.0%) 6,593 (6.8%) 97,341
*Incomplete year.
Note: MDSv1, in use in 2003, only defined operations as Primary or Revision. All revisions using MDSv1 have been listed as Single stage revisions in this table.
Table 3.16 shows the stated reasons for the revision stated, the reasons are not mutually exclusive and
surgery. Please note that, as several reasons can be therefore column percentages do not add up to 100%.
5 For 233 patient-sides, multiple procedures had been entered on the same operation date; 232 had two on the same date and one had three. Details of the
components that had been entered for these cases were reviewed. As a result of this, 237 of the 466 revision procedures have been dropped and 21 have
been reclassified.
Table 3.16 Reasons for the hip revision procedures: percentages indicating each reason, calculated separately for
single and two-stage revisions.
Type of revision procedure

Single stage Stage one of two-stage Stage two of two-stage
Reason (n=84,883) (n= 5,865) (n=6,593)
Aseptic loosening 50.1% 13.2% 12.3%
Pain 20.5% 13.5% 9.3%
Lysis 15.4% 9.6% 6.1%
Dislocation/subluxation 15.2% 4.1% 3.4%

Implant wear 14.1% 4.5% 3.1%
Periprosthetic fracture 10.1% 3.7% 4.0%
Other indication 7.6% 3.4% 8.3%
Malalignment 5.6% 1.5% 0.9%
Implant fracture 3.6% 1.1% 1.3%
Infection 3.5% 80.3% 72.5%
Head-socket size
0.8% 0.3% 0.2%
mismatch
Adverse reaction to
11.0% n=66,920 3.1% n=4,847 2.3% n=5,231
particulate debris*
*Not recorded in the early phase of the registry; MDSv3 and v6 only.
3.4.2 Rates of hip re-revision In Figure 3.14 (b) we sub-divided the first revisions into
those for whom a primary had been recorded in the NJR
For a given patient-side, we have looked at the survival (n=24,103) and the remainder. The survival of the former
following the first documented revision procedure in appeared much worse. This is interesting as primaries
the NJR (n=85,200). In most instances (91.3%), the not in the NJR are likely to have been performed prior to
first revision procedure was a single stage revision, 2003 and thus represent late failure. In contrast, revisions
however in the remaining 8.7% it was part of a two-stage linked to primaries in the NJR are more likely to represent
procedure. We have looked at the time from the first early failure. It thus appears that revision after late failure
documented revision procedure (of any type) to the time is less likely to need re-revision than revision after early
at which a second revision procedure was undertaken. failure. Figure 3.14 (c) and Table 3.17 (b) further exemplify
For this purpose, we regarded an initial stage one this; cumulative re-revision rates up to three years are
followed by either a stage one or a stage two as being shown separately for those with primaries in the NJR
the same revision episode and these were disregarded, according to their time intervals to first revision, less than
looking instead for the start of a second revision episode. 1 year, 1 to 3, 3 to 5 and more than 5 years.
(We counted the maximum number of distinct revision
episodes for any patient-side to be eight). There is a relationship between the indication for first
revision and time to first revision; earlier in this report
Kaplan-Meier estimates were calculated to estimate (section 3.3.5) we showed, for example, that revisions for
the cumulative probability of a subsequent revision dislocation/subluxation and pain were more prevalent in
(re-revision). These rates are plotted in Figure 3.14 (a) and the early period after the primary and aseptic loosening
tabulated in Table 3.17 (a). There were 7,522 re-revisions and pain later on. The relationship between (i) the time
and in 16,330 cases the patient died without having to first revision and the subsequent time to re-revision,
been revised; the censoring date for the remainder was and (ii) the indication for the first revision and the time to
the end of 2016. re-revision require further investigation.
For those with documented primaries within the NJR, 3 to 5 years of the initial primary hip replacement; though
Figures 3.15 (a) to (e) show cumulative re-revision rates the numbers were small and therefore the results should
up to five years from the first revision, according to the be interpreted with caution.
main fixation used in the primary. Each sub-group has
been further sub-divided according to the time interval Table 3.17 (c) shows cumulative re-revision rates at 1,
from the primary to the first revision, i.e. less than 1 year, 3, and 5 years following the first revision for those with
1 to 3, 3 to 5 and more than 5 years. For cemented, documented primaries within the NJR, broken down by
uncemented, hybrid, and resurfacing hip replacements, fixation types and bearing surfaces. Overall, the worst
those who had their first revision within one year of re-revision rates were demonstrated in those where
the initial primary hip replacement, experienced the the initial primary had been uncemented, with metal-
worst re-revision rates. However, for reverse hybrid on-metal bearings faring worse than other bearings
hip replacements, the worst re-revision rates were within the group. The failure rates for resurfacings were
experienced by those who had their first revision within comparatively low.
Figure 3.14 (a)

Kaplan-Meier estimate of the cumulative probability of a hip re-revision (shaded area indicate
point-wise 95% CI).
20
15

10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years since first revision
Number at risk
85,199 72,938 63,204 53,592 44,368 35,045 27,245 20,862 15,431 10,700 6,697 3,931 1,921 602
Figure 3.14 (b)
Kaplan-Meier estimates of the cumulative probability of a hip re-revision, shown separately for those with
documented primaries in the NJR* and the remainder (shaded area indicate point-wise 95% CIs).
25
20
15
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Primary not in the NJR 61,096 53,207 47,015 40,608 34,374 28,124 22,673 17,854 13,583 9,652 6,187 3,691 1,839 586
Primary in the NJR 24,103 19,731 16,189 12,984 9,994 6,921 4,572 3,008 1,848 1,048 510 240 82 16
*First documented revision in the NJR.
Figure 3.14 (c)

Kaplan-Meier estimates of the cumulative probability of a hip re-revision up to three years from the first
revision. Those with documented primaries in the NJR* are shown separately from the remainder and
have been sub-divided into those that had their first revision within <1, 1-3, 3-5 and >5 years from the
initial primary.
15

10
0
0 1 2 3
Numbers at risk
Primary not in the NJR 61,096 53,207 47,015 40,608
First rev. <1y 6,674 5,490 4,545 3,751
First rev. 1−3y 5,438 4,674 4,038 3,481
First rev. 3−5y 4,313 3,785 3,341 2,871
First rev. 5+y 7,678 5,782 4,265 2,881
Figure 3.15 (a)
Kaplan-Meier estimates of the cumulative probability of a hip re-revision up to five years from the first
revision, shown separately for type of fixation used in the primary, with further sub-division by length of
time from the primary to the first revision (<1, 1-3, 3-5 and >5).
(a) Cemented
15
10
0
0 1 2 3 4 5
Years since ﬁrst revision

Numbers at risk
First rev. <1y 1,478 1,173 967 781 630 490

First rev. 1−3y 1,378 1,141 961 802 665 515
First rev. 3−5y 841 695 592 493 397 284
First rev. 5+y 1,699 1,225 857 581 370 187
Figure 3.15 (b)

(b) Uncemented
15

10
0 1 2 3 4 5

Numbers at risk
First rev. <1y 3,337 2,791 2,321 1,914 1,564 1,196

First rev. 1−3y 2,584 2,252 1,955 1,685 1,398 1,077
First rev. 3−5y 2,081 1,841 1,614 1,371 1,029 572
First rev. 5+y 3,401 2,538 1,810 1,156 603 203
Figure 3.15 (c)
(c) Hybrid
15
10
0
0 1 2 3 4 5

Numbers at risk
First rev. <1y 1,206 933 709 550 442 360

First rev. 1−3y 666 527 428 340 282 219
First rev. 3−5y 462 389 314 247 192 140
First rev. 5+y 733 514 375 258 166 76
Figure 3.15 (d)

(d) Reverse hybrid

15

10
0
0 1 2 3 4 5

Numbers at risk
First rev. <1y 169 131 113 93 66 50
First rev. 1−3y 120 99 71 57 49 39
First rev. 3−5y 58 45 35 21 13 8
First rev. 5+y 67 45 28 20 10 8
Figure 3.15 (e)
(e) Resurfacing
15
10
0 1 2 3 4 5

Numbers at risk
First rev. <1y 484 462 435 413 392 359
First rev. 1−3y 690 655 623 597 556 488
First rev. 3−5y 871 815 786 739 620 425
First rev. 5+y 1,778 1,460 1,195 866 550 225
Table 3.17 (a) Kaplan-Meier estimates of cumulative percentage probability of a hip re-revision following the first revision.
Time point Cumulative re-revision rate (95% CI) at:

from which
time was
measured: Sub-group n 1 year 3 years 5 years 7 years 10 years 13 years
3.86 6.90 9.09 11.23 14.40 17.37
First revision* All 85,200
(3.73-3.99) (6.72-7.08) (8.87-9.32) (10.96-11.51) (14.01-14.80) (16.65-18.12)
Primary not
3.44 6.12 8.15 10.13 13.21 16.19
recorded in 61,097
(3.29-3.59) (5.92-6.32) (7.91-8.40) (9.84-10.43) (12.80-13.64) (15.45-16.97)
the NJR
First revision
Primary
4.95 8.98 11.74 14.72 18.56 22.01
recorded in 24,103
(4.67-5.24) (8.59-9.39) (11.25-12.24) (14.07-15.40) (17.44-19.75) (18.99-25.43)
the NJR
Table 3.17 (b) Kaplan-Meier estimates of cumulative percentage probability of a hip re-revision following the first revision,
sub-divided by time since primary.
Cumulative re-revision rate (95% CI) at:

n 1 year 3 years
Primary not in the NJR 61,097 3.44 (3.29-3.59) 6.12 (5.92-6.32)
Primary in the NJR where the
first revision took place:
<1 year after primary 6,674 6.41 (5.83-7.04) 11.34 (10.54-12.19)
1-3 years from primary 5,438 5.24 (4.67-5.89) 9.68 (8.87-10.55)
3-5 years from primary 4,313 4.46 (3.87-5.13) 8.22 (7.40-9.13)
5+ years from primary* 7,678 3.67 (3.26-4.14) 6.50 (5.89-7.17)
*Note: maximum interval was 12.9 years.
Table 3.17 (c) Kaplan-Meier estimates of the cumulative percentage probability of a hip re-revision (95% CI) at 1, 3, and
5 years following the first revision in those with documented primaries in the NJR, by fixation and bearing surface (group
size >1,000 in the case of the bearing surfaces).
Cumulative percentage probability of re-revision (95% CI)
Bearing following first revision at:
Fixation surface n 1 year 3 years 5 years
All types All 24,103 4.95 (4.67-5.24) 8.98 (8.59-9.39) 11.74 (11.25-12.24)
Cemented All 5,396 5.65 (5.05-6.33) 8.96 (8.15-9.83) 11.07 (10.10-12.12)

MoP 4,676 5.55 (4.91-6.27) 8.66 (7.81-9.59) 10.62 (9.60-11.74)
Uncemented All 11,403 5.16 (4.76-5.59) 9.57 (9.00-10.18) 12.12 (11.42-12.87)
MoP 3,046 5.31 (4.55-6.20) 9.31 (8.23-10.52) 11.56 (10.25-13.02)
MoM 4,268 4.87 (4.25-5.57) 9.36 (8.46-10.34) 12.34 (11.21-13.57)
CoP 1,144 6.00 (4.73-7.61) 11.99 (9.99-14.35) 13.57 (11.32-16.23)
CoC 2,612 4.90 (4.12-5.83) 9.01 (7.88-10.30) 11.29 (9.91-12.85)
Hybrid All 3,067 5.13 (4.38-6.01) 9.26 (8.16-10.50) 12.02 (10.64-13.56)
MoP 1,883 5.60 (4.62-6.79) 9.37 (8.00-10.96) 12.04 (10.31-14.02)
Resurfacing (MoM) 3,823 3.36 (2.83-4.00) 7.11 (6.29-8.03) 11.09 (9.98-12.31)
*Note: maximum interval was 12.2 years.
3.4.3 Reasons for the hip re-revision (i) shows indications for the first revision in the NJR,
(ii)/(iii) for the first revision but depending on whether or
Table 3.18 shows a breakdown of the stated indications not the implants were subsequently re-revised, and (iv)
for the first revision and for any subsequent revision for the re-revisions themselves.
(note the indications are not mutually exclusive). Column
Table 3.18 Reasons for the hip first revision and subsequent re-revision.
Reasons for the first recorded revision
for those who were:
(i)
Reasons for (ii) (iii) (iv)
first (recorded) Not subsequently Subsequently Reasons for
revision re-revised re-revised the re-revision
Number of cases 85,199 77,677 7,522 7,522

Aseptic loosening 41,077 37,677 3,400 2,416
Pain 17,231 15,726 1,505 1,131
Lysis 13,194 12,158 1,036 566
Implant wear 11,808 10,902 906 505
Dislocation/subluxation 11,172 10,134 1,038 1,793
Infection 7,832 6,873 959 1,532
Periprosthetic fracture 8,079 7,372 707 751
Malalignment 4,448 4,074 374 367
Implant fracture 2,862 2,610 252 282
Head-socket (size) mismatch 628 565 63 50
Other indication 6,399 5,784 615 503
Adverse reaction to particulate debris 7,095 n=66,157 6,592 n=61,029 503 n=5,128 453 n=6,752
Finally, Tables 3.19 (a) and 3.19 (b) provide additional with the remainder (Kaplan-Meier estimates 1.01
evidence that the 61,097 revised joints with no (95% CI 0.89-1.14) versus 1.64 (1.54-1.74)), which
associated primary in the NJR tended to be later may reflect the fact that this patient group were
revisions than the 24,103 joints that did have an younger at the time of their first revision, median
associated primary. The results also show that the age of 68 (IQR 60-75) years compared to the group
numbers of revisions with an associated primary in the without primaries documented in the NJR who had a
NJR increased with time. median age of 73 (IQR 65-80) years. The percentage
of males was similar in both groups (43.6% versus
3.4.4 90-day mortality after hip 42.1% respectively).
revision
The overall cumulative percentage mortality at 90
days after hip revision was lower in the cases with
their primaries documented in the NJR compared
Table 3.19 Temporal changes in first hip revisions reported in the NJR and associated indications.
(a) Number of first hip revisions by year and proportions with an associated primary in the NJR.
Year of first revision
in the NJR* Number of (first) revisions* Number (%) with the associated primary in the NJR
2003 1,404 43 (3.1%)
2004 2,620 141 (5.4%)
2005 3,707 301 (8.1%)
2006 4,466 450 (10.1%)

2007 5,835 803 (13.8%)
2008 6,300 1,132 (18.0%)
2009 6,560 1,492 (22.7%)
2010 7,121 1,934 (27.2%)
2011 7,978 2,627 (32.9%)
2012 9,032 3,305 (36.6%)
2013 8,228 3,001 (36.5%)
2014 8,017 3,028 (37.8%)
2015 7,304 3,008 (41.2%)
2016 6,627 2,838 (42.8%)
Total 85,199 24,103 (28.3%)
(b) Numbers of first recorded hip revisions by stage and whether or not primary was in the NJR.
Year of first revision Single stage First documented stage of two-stage

in the NJR* Primary not in the NJR Primary in the NJR Primary not in the NJR Primary in the NJR
2003 1,361 43 0 0
2004 2,270 120 209 21
2005 3,073 244 333 57

2006 3,633 363 383 87
2007 4,580 669 452 134
2008 4,686 929 482 203
2009 4,581 1,226 487 266
2010 4,771 1,703 416 231
2011 4,958 2,357 393 270
2012 5,343 2,976 384 329
2013 4,909 2,697 318 304
2014 4,643 2,732 346 296
2015 3,996 2,696 300 312
2016 3,570 2,590 219 248
All years 56,374 21,345 4,722 2,758
Part 3
3.5 Outcomes
after primary
knee replacement
This section reviews the outcome of primary knee time a patient could have for either outcome is 13.75
replacement surgery in terms of two key events that years, corresponding to a patient operated on at the
could happen post-operatively to a patient who has start of the registry.
undergone a knee replacement or to the knee joint;
the ’first revision’ of a knee implant and/or patient Tables 3.1 and 3.2 provide an overview of the primary
death or mortality. knee replacement patient cohort. Over the period of
2003 to 2016, a total of 975,739 knee joints were
Core to the analysis approach for both outcomes replaced for the first time (primary joint replacement).
is modelling the time until the event is observed to There were a total of 800,477 patients with a NJR
happen and giving due consideration to the time record of primary knee replacement on one or
the patient or joint is at risk of the event happening. both sides. Approximately four fifths of the patient
Further details of the statistical methods are given in cohort had just one record of a primary knee joint
statistical methodology notes I to III overleaf. replacement since the establishment of the NJR. The
remaining fifth of patients were those who had records
The outcomes of total and partial knee replacement of both left and right knees being replaced for the
procedures are discussed throughout this first time. The majority of this patient sub-group had
section, hereon referred to as total (TKR) and primary knee surgery at different times for each side
unicompartmental (UKR) replacement. Brief details (164,665 patients), but 10,597 patients had surgery
of the type of orthopaedic surgery involved for each for both knees on the same date (1.3% of all patients
form of replacement can be found in the terminology in the cohort).
note below. Of special note here is that the NJR data
collection process now collects separate information The predominant clinical reason recorded for primary
on medial and lateral unicondylar replacements, surgery was osteoarthritis (OA); it was the sole stated
although this was not the case in the past. reason in 938,349 (96%) of primary knee surgeries
and one of the reasons recorded in a further 1.1% of
The patient cohort described in this section is any primaries performed when multiple clinical reasons for
patient whose recorded primary knee replacement surgery were given on the data collection form.
surgery date fell on or after 1 April 2003 and up to 31
December 2016 (inclusive). The maximum follow-up
Terminology note: years monobloc all-polyethylene tibial components
have increased in popularity.
The knee is made up of three compartments: medial,
lateral and patellofemoral compartments. When a In modular tibial components, the tibial insert may
total knee replacement (TKR) is implanted, two out be mobile or remain in a fixed position on the
of the three compartments are always replaced tibial tray. This also applies to medial and lateral
(medial and lateral) and the patella is resurfaced if the unicompartmental knees. Many brands of total knee
surgeon considers this to be of benefit to the patient. implant exist in fixed and mobile forms with options for
If a single compartment is replaced then the term either CR or PS constraint.
unicompartmental is applied to the implant (UKR). The
medial, lateral or patellofemoral compartments can all The NJR now distinguishes between medial and
be replaced independently, if clinically appropriate. lateral unicondylar knee replacements during the data
collection process, however, this was not so in earlier
There is variation in the constraint of the tibial insert versions of the minimum dataset form (MDS). In
depending on whether the posterior cruciate ligament addition, there are other possible knee designs, such
is preserved (cruciate retaining; CR) or sacrificed as combinations of unicondylar and patellofemoral,
(posterior stabilised; PS) at the time of surgery. but these are not reported on here, as the numbers
Additional constraint may be necessary to allow the are too small.
implant to deal with additional ligament deficiency
or bone loss, where constrained condylar (CCK) or With regard to the use of the phrase constraint
hinged knee implants would be used, in a primary or here, for brevity, total knee replacements are
revision procedure. The tibial element may be modular termed unconstrained (instead of posterior cruciate-
with a metallic tibial tray and a polyethylene insert or retaining) or posterior-stabilised (instead of posterior
non-modular consisting of an all-polyethylene tibial cruciate-stabilised).
component (monobloc polyethylene tibia). In recent
Methodological note I: those which are no longer available to be observed

Survival analysis, time at risk and censoring until the end date of the observation period, termed
observations lost to follow-up. As a consequence of
Survival analyses have been employed to provide censoring, the total number of patients at risk of the
estimates of the two main outcomes of interest event at different points in time will vary over the whole
after primary knee replacement surgery; namely the observation period.
cumulative probability that an implant is revised for
the first time at different times after primary operation For mortality, the period of time at risk contributed
(revision outcome) and the cumulative probability that by a patient in the cohort is the length of time until
a patient dies at different lengths of time after primary they died post primary surgery or, if they do not
knee surgery (mortality outcome). die, the time from primary surgery until the last day
in December 2016, the last date of the period of
Key to these methods is correctly specifying the observation for this report.
period of time after primary surgery each replaced joint
is at risk of the event of being revised or the patient is Turning to the revision outcome, the time a joint is
at risk of dying. In addition, not all replaced joints will at risk of being revised for the first time is either the
be revised (or all patients will die) over the observation time until the joint is revised post primary surgery
period, i.e. the event of interest will not happen to (and before the end of 2016), the time until they
all joints/patients. When this is the case, the time die after surgery without being revised (and before
observations are censored. Censored observations the end of 2016) or the period of time they are not
occur for a number of reasons; they can be those revised after primary surgery up until the last date of
cases which have not experienced the outcome observation in 2016.
of interest by the end of the observation period or
Methodological note II: similar estimates in terms of actual numerical values

Use of Kaplan-Meier estimation for describing (see the glossary for further technical details). This is
mortality and revision no longer the case and we now solely use Kaplan-
Meier estimation throughout Part Three.
The main tables and figures shown in the text are
based on Kaplan-Meier estimates of the cumulative The Confidence Intervals (CI) found for the cumulative
probability of the joint being revised or the patient percentage probability estimates of revision or death,
dying at different times after the primary surgery. The based on the Kaplan-Meier method, become less
calculated probabilities have been multiplied by 100 reliable when the number at risk of revision or death
in all results presented here and so represent the falls below 250. Several methods have been proposed
cumulative percentage probability of having a first to calculate Confidence Intervals. These proposed
revision or of dying at different times after surgery. methods produce confidence intervals which are all
in agreement with one another when there are high
This is a change to previous NJR annual reports numbers at risk. However, they begin to give very
(prior to 2014) where a mixture of Kaplan-Meier different upper and lower limits once the numbers at
estimation of the cumulative probability of having a risk falls below 250. To date, there has been no clear
first revision (or of dying) and Nelson-Aalen estimation consensus on which method is to be preferred when
of cumulative hazard (the expected total number numbers at risk are small. For this reason, we highlight
of revisions or deaths up to a point in time) were the point estimate of the cumulative chance of
reported. Clearly, the two methods find different revision/death and the confidence interval throughout
quantities – one is a probability and the other is not – in blue italics once the number at risk drops below
but, under certain conditions, both methods provided 250 cases.
Methodological note III: in the modelling process. One commonly proposed

Competing risks considerations method is the use of the Cumulative Incidence
Function (CIF) adjusting for the competing risk of death
One assumption which underpins the use of the (see section 3.3.2.6 of the NJR Annual Report 2014
Kaplan-Meier method to estimate the cumulative where the impact of CIF on the probability estimates
chance of death or first revision is that the patients/ obtained was considered).
joints whose times are censored, have the same
chance of having the event of interest happen to them In the main analyses presented here, we have not
after censoring as those cases still at risk in the study. made adjustments for competing risks in the modelling
of first revision and death as outcomes.
This assumption could be compromised if the reason
they are censored is as a result of other events So, in the case of the revision outcome, no adjustment
happening to the patient or joint after primary knee for the competing risk of death has been made in the
surgery, but not the main one of interest, which main survival table and figure presentations. However, a
potentially change the likelihood of the main outcome simple assessment of the impact of the competing risk
(first revision or death) occurring afterwards. An event of death on the revision outcome estimates using the
like this is known as a competing risk. cumulative incidence function is presented in the text.
For example, if a patient dies before having a first For mortality, we have not accounted for the impact
revision, their observation will be treated as censored that having a first or further revision after primary
but the chance of the outcome revision happening surgery may have on the likelihood of a patient
after death is impossible. Death, here, is the dying subsequently, compared to the likelihood of
competing risk. The true effect of the event death on death for those who have not had a first or further
the Kaplan-Meier estimates for revision as the main revision surgery.
outcome can only be assessed if it is accounted for
3.5.1 Overview of primary knee surgery unconstrained and fixed, followed by 20.7% which
were all cemented, posterior stabilised and fixed.
3.5.1.1 Main types of primary knee surgery and Within each method of fixation, it can be seen
changes in type of operation over time that uncemented/hybrid prostheses are mostly
unconstrained (cruciate retaining) but almost equally
Table 3.20 shows the proportion of all main kinds likely to have a mobile or fixed bearing. Two-thirds
of primary knee operations carried out between (66.6%) of cemented implants are unconstrained
2003 and 2016, broken down by the method of (cruciate retaining) and have a fixed bearing.
fixation, constraint and bearing used for the implant Unicondylar knee surgery typically involves the use
in surgery. A breakdown within each method of of a mobile type of bearing/constraint. A number of
fixation of the percentage of constraint and bearing primary knee joint operations could not be classified
types used in surgery is shown in a separate column. according to their bearing/constraint (approximately
The vast majority of replacements performed were 1.1% of the total cohort).
total knee replacements (TKRs) with an all cemented
implant being the most common technique of fixation Table 3.21 shows the annual change in the usage of
used (84.9% of all primary knee operations). A further primary knee replacements. Overall, more than 80%
5.1% were either all uncemented or hybrid total knee of all primaries utilised an all cemented fixation method
replacements (where at least one component utilises and since 2003, the share of all implant replacements
cemented fixation and at least one component of this type has increased by about 6%. The main
utilises uncemented fixation). Most partial knee decline in the type of primary knee surgery carried out
replacements (UKRs) were unicondylar (8.7% of the has been in the use of all uncemented and hybrid total
total) with the remainder being patellofemoral knee knee replacements over time (now 2.4% of all knee
replacements (1.2%). replacements). Each implant of this type now used has
decreased proportionally to less than a third of those
More than half of all operations (56.6%) were total figures reported for 2003 (when they were 9.5% of all
knee replacements which were all cemented, knee replacements).
Table 3.20 Numbers and percentages of primary knee replacements by fixation method, constraint and bearing type.
Type of primary knee operation Percentage of each

constraint type used Percentage of
Constraint and Number of primary within each method all primary knee
Fixation method bearing type knee operations of fixation operations
Total knee replacement
All cemented 828,573 84.9
Cemented and
unconstrained, fixed 551,832 66.6 56.6
unconstrained, mobile 34,507 4.2 3.5
posterior-stabilised, fixed 202,284 24.4 20.7
posterior-stabilised, mobile 11,526 1.4 1.2
constrained, condylar 6,428 0.8 0.7
monobloc polyethylene
12,765 1.5 1.3
tibia

bearing type unknown 9,231 1.1 0.9
All uncemented 40,720 4.2
All hybrid 8,898 0.9
Uncemented/hybrid
and
unconstrained, fixed 21,625 43.6 2.2
unconstrained, mobile 23,395 47.2 2.4
posterior-stabilised, fixed 3,376 6.8 0.3
other constraint 632 1.3 0.1
bearing type unknown 590 1.2 0.1
Unicompartmental knee replacement
All unicondylar 85,312 8.7
Unicondylar and
fixed 27,901 32.7 2.9
mobile 56,523 66.3 5.8
bearing type unknown 888 1 0.1
All patellofemoral 12,191 n/a 1.2
Fixation unknown Bearing type unknown 45 n/a <0.01
All types 975,739 n/a 100.0
Table 3.21 Percentage of all primary knee replacements performed each year by total and partial knee replacement
types by fixation method1.
Percentage of primary knee replacements performed in each year by fixation method and percentage
breakdown by constraint/bearing type2
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
All cemented 81.5 80.8 81.7 81.3 81.9 81.8 82.6 84.0 85.4 86.7 87.7 87.4 87.4 87.3
Cemented and
unconstrained
53.2 52.8 52.8 50.4 50.3 51.2 52.8 54.2 56.3 59.0 59.8 60.7 61.7 62.2
fixed
unconstrained
4.0 4.2 5.4 6.5 6.4 5.7 4.8 4.1 3.0 2.4 2.2 2.0 1.7 1.8
mobile
posterior-
20.7 20.6 19.6 20.1 20.4 20.9 21.4 21.8 21.6 21.0 21.1 20.5 20.2 19.8
stabilised fixed
posterior-
stabilised 0.9 1.1 1.6 1.9 1.6 1.4 1.4 1.4 1.2 1.1 1.2 1.0 0.8 0.5
mobile
constrained
0.4 0.4 0.4 0.3 0.3 0.3 0.3 0.4 0.4 0.6 0.8 1.0 1.2 1.3
condylar
monobloc
polyethlene 0.3 0.2 0.4 0.6 0.9 0.8 0.7 1.0 1.6 2.0 2.1 1.9 1.5 1.5
tibia
bearing/
constraint 1.9 1.3 1.5 1.6 2.0 1.5 1.2 1.2 1.2 0.6 0.5 0.4 0.3 0.3
unknown
All uncemented 6.7 6.6 6.2 6.5 6.5 6.2 5.7 4.7 4.1 3.3 2.5 2.6 2.3 2.0
Uncemented and
unconstrained
2.6 2.4 2.4 2.7 3.0 2.8 2.6 1.8 1.4 1.0 0.7 0.6 0.7 0.7
fixed
unconstrained
3.2 3.3 3.2 3.1 3.0 2.9 2.6 2.6 2.4 2.0 1.6 1.6 1.4 1.1
mobile
posterior-
0.7 0.6 0.5 0.5 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.3 0.2 0.2
stabilised fixed
other
<0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
constraint
constraint
0.2 0.3 0.2 0.1 0.1 0.1 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
unknown
All hybrid 2.8 2.8 2.4 1.7 1.4 1.4 1.2 0.9 0.5 0.4 0.4 0.4 0.4 0.4
Hybrid and
unconstrained
2.3 2.3 1.9 1.2 1.0 1.1 1.0 0.7 0.3 0.2 0.2 0.1 0.1 0.1
fixed
unconstrained
0.3 0.3 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.3 0.3
mobile
posterior-
0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
stabilised fixed
other
<0.1 0.2 0.2 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
constraint
constraint
0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
unknown

Percentage of primary knee replacements performed in each year by fixation method and percentage
breakdown by constraint/bearing type2
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Unicompartmental knee replacement
All unicondylar 8.0 8.7 8.6 9.3 8.9 9.1 9.0 9.0 8.5 8.2 8.1 8.5 8.8 9.2
Unicondylar and
fixed 1.4 1.8 2.0 2.3 2.0 2.1 2.2 2.7 2.6 3.0 3.3 3.5 3.7 4.1
mobile 6.5 6.7 6.5 6.9 6.7 6.8 6.6 6.2 5.8 5.2 4.7 5.0 5.1 5.1
constraint
0.2 0.2 0.1 0.2 0.1 0.2 0.2 0.1 0.1 <0.1 0.1 <0.1 <0.1 <0.1
unknown
All
1.0 1.0 1.0 1.1 1.4 1.5 1.5 1.4 1.5 1.4 1.2 1.1 1.1 1.1
patellofemoral
All unknown 0.2 0.01
All types (n) 13,546 27,762 42,301 50,360 66,878 74,277 76,259 78,908 82,501 86,299 85,935 95,740 96,826 98,147
Note: 1 Percentage of all primary operations in a particular year which used one of the five fixation methods: cemented, uncemented, hybrid, patellofemoral
or unicondylar. 2 Percentages shown represent percentage of total procedures.
3.5.1.2 Reasons for primary knee diagnoses available on the particular version of the
replacement surgery data collection form filled out by the clinician. The total
number of indications, the percentage this forms of the
The diagnostic reason(s) for a patient undergoing total number of knee operations and a breakdown of
primary knee replacement surgery form part of the these by gender are shown separately for each reason.
clinical pre-assessment process and are recorded Reasons shown are all indications given for a primary
by the clinician on the MDS form. Of all reasons for surgery and in some cases multiple reasons have
primary surgery, the dominant diagnosis recorded in the been given for a primary operation. Therefore, reasons
registry is knee osteoarthritis; the number of joints with are not mutually exclusive of each other. In addition,
a sole diagnosis of knee osteoarthritis as the indication 144 knee procedures had no recorded reason for
for knee replacement is 938,349 (96%) of all 975,595 undergoing primary surgery.
knee replacements with a reason for primary surgery
recorded in the NJR. Other possible diagnoses include After osteoarthritis, the most frequently given indication
avascular necrosis, trauma, inflammatory arthritis for surgery was inflammatory arthritis (forming about
and infection (see Table 3.22 footnotes for primary 2% of reasons). There is some indication of gender
diagnoses details). differences in the primary reason given for carrying out
knee replacement, although for some diagnoses, the
Table 3.22 shows the main reasons cited by clinicians numbers of cases are small.
for primary surgery, as selected from the listed
Table 3.22 Reasons for primary knee replacement surgery; number and percentage of all NJR recorded primary knee
replacement surgeries carried out for each clinical reason broken down by gender.
Number (%) of knee joints with specified
primary diagnosis1
(n= 975,595)
All joints with this reason1

Reason for Knee Primary Male Female (% of all joints)
Osteoarthritis 413,516 (96.9) 535,727 (95.6) 949,243 (96.1)
Avascular necrosis 1,381 (0.3) 2,190 (0.4) 3,571 (0.4)
Previous infection 402 (0.1) 265 (<0.1) 667 (0.1)
Previous trauma 3,126 (0.7) 2,472 (0.4) 5,598 (0.6)
Inflammatory arthritis2 5,178 (1.2) 15,516 (2.8) 20,694 (2.1)
Trauma 16 (<0.1) 24 (<0.1) 40 (<0.1)
Other indication3 3,191 (0.7) 4,295 (0.8) 7,486 (0.8)
Note: 1 More than one diagnosis could be indicated by the clinician and results represent all reasons given by the surgeon. 2 Inflammatory arthritis for knees
combines diagnoses of rheumatoid arthritis, seronegative and seropositive rheumatoid arthritis, and other inflammatory arthropathy. 3 Other indication includes failed
internal fixation, previous arthrodesis, and other indicated reasons for primary knee replacement.
3.5.1.3 Summary of the types of primary Over this three-year period, there have been 261,842
knee surgery performed by consultant primary total knee replacements performed by 1,999
surgeons and units surgeons (median=95; IQR 25-194) in 403 separate
units (median=565 cases per unit; IQR 277-939). In
Within the whole registry, the 975,739 primary knee the same time period, there have been 25,718 primary
joint replacement procedures contributing to our unicondylar knee procedures performed by 820
analyses were carried out by a total of 3,124 consultant consultant surgeons (median=12; IQR 3-35) in 364
surgeons working across 460 units. Over the last three units (median=37 cases per unit; IQR 13.5-81.5). The
years (1st January 2014 to 31st December 2016), number of procedures per consultant over this period
290,713 primary knee procedures were performed by may be lower for newly qualified consultants and those
2,007 consultant surgeons working across 403 units. who may have retired during this period. Table 3.23
Looking at caseload over this three-year period, the shows how the caseload of TKR, unicondylar and
median number of primary procedures per consultant patellofemoral procedures for units and consultants has
surgeon was 104 (IQR 26-214) and the median number changed over the last three years.
of procedures per unit was 633 (IQR 318-1,006).
Table 3.23 Descriptive statistics of total knee replacement, unicondylar and patellofemoral procedures performed by
consultant and unit by year of surgery in the last three years.
Year of surgery: 2014 2015 2016

Number of procedures in year: 95,740 96,826 98,147
Number of consultants providing primary replacement each year 1,723 1,722 1,694
Consultant: Total Mean number of primary replacements per consultant 50 51 52
knee
Median (IQR) number of any primary replacement per consultant 40 (14-74) 40 (16-74) 41 (17-71)
replacements
(TKR) Number of consultants who entered >50 procedures each year 707 706 710
Number of consultants who entered >100 procedures each year 223 213 210
Number of consultants providing primary replacement each year 646 647 637
Consultant: Mean number of primary replacements per consultant 13 13 14
Unicondylar Median (IQR) number of any primary replacement per consultant 6 (3-14) 6 (2-15) 7 (2-17)
replacements Number of consultants who entered >10 procedures each year 212 234 242
Number of consultants providing primary replacement each year

317 295 282
Consultant: Mean number of primary replacements per consultant 3.3 3.6 3.7
Patellofemoral Median (IQR) number of any primary replacement per consultant 2 (1-4) 2 (1-5) 2 (1-4)
replacements Number of consultants who entered >10 procedures each year 14 12 14
Number of units providing primary replacement each year 391 386 394
Mean number of primary replacements per unit 221 226 224
Units: Total knee
Median (IQR) number of any primary replacement per unit 195 (97-315) 197 (99-317) 195 (97-312)
replacements
Number of units who entered >300 procedures each year 102 110 112
Units: Mean number of primary replacements per unit 24 26 27
Unicondylar Median (IQR) number of any primary replacement per unit 13 (5-29) 14 (6-29) 14 (6-32)
replacements Number of units who entered >10 procedures each year 196 202 195
Units: Mean number of primary replacements per unit 4.8 4.7 4.7
Patellofemoral Median (IQR) number of any primary replacement per unit 3 (2-6) 3 (1-6) 3 (1-6)
replacements Number of units who entered >10 procedures each year 23 25 26
Looking at recent annual unit caseload, in 2016, 8% 25% of consultants were performing one or two cases
of units performing primary total knee replacements a year accounting for only 2.4% of total unicondylar
(Figure 3.16 (a)) performed less than 25 total knee replacements. A further 37% of unicondylar consultants
replacements during the year. This compares to were performing between three and ten cases a year.
unicondylar knee replacements (Figure 3.16 (b)) where In total these 62% of unicondylar consultants were
62% of units performed 20 or less unicondylar knee performing 17.3% of all unicondylar procedures. For
replacements in the year. Figure 3.16 (c) shows that patellofemoral replacements (Figure 3.16 (c)), 54% of
99% of units performing patellofemoral replacements patellofemoral consultants were performing only one
performed 20 or less patellofemoral replacements in or two in the year which accounts for 21.2% of all
the year. patellofemoral replacements in the year. A further 41%
were performing between three and ten cases a year.
Looking at recent annual consultant caseload This means that 5% of patellofemoral consultants are
(Figure 3.16 (a)), in 2016, 34% of primary total knee doing 23.3% of all patellofemoral replacements. In total,
replacement consultants were performing 25 or less 99% of patellofemoral consultants are performing 20 or
total knee replacements a year. This accounts for less patellofemoral replacements in the year.
approximately 7% of primary total knee replacements.
For unicondylar knee replacements (Figure 3.16 (b)),
Figure 3.16 (a)

Exploring unit and consultant frequency of TKR, unicondylar and patellofemoral knee surgery, and total
caseload in the NJR between 2014 and 2016.
Exploring unit and consultant frequency of total knee replacement surgery,
(a) TKR and total caseload in the NJR between 2014 and 2016
Caseload of Units Caseload of Consultants
Distribution of unit annual Distribution of consultant annual

frequency of TKR procedures frequency of TKR procedures
2014 2015 2016 2014 2015 2016

Total caseload of TKR surgery by Total caseload of TKR surgery by
unit annual frequency consultant annual frequency
2014 2015 2016 2014 2015 2016
Annual frequency of unit procedures Annual frequency of consultant procedures

<25
25−100
101−200
201−300
301−400
401−500
501−750
>750
2−10
11−25
26−50
51−75
76−100
101−200
>200
Figure 3.16 (b)
Exploring unit and consultant frequency of unicondylar knee surgery,
(b) Unicondylar
and total caseload in the NJR between 2014 and 2016
Distribution of unit annual frequency Distribution of consultant annual frequency

of Unicondylar procedures of Unicondylar procedures
2014 2015 2016 2014 2015 2016
Total caseload of Unicondylar surgery by Total caseload of Unicondylar surgery by

2014 2015 2016 2014 2015 2016
1−2 3−5 6−10 11−20 21−50 51−90 >90 1−2 3−5 6−10 11−20 21−50 51−90 >90
Figure 3.16 (c)

Exploring unit and consultant frequency of patellofemoral surgery,
(c) Patellofemoral
and total caseload in the NJR between 2014 and 2016
Distribution of unit annual frequency Distribution of consultant annual frequency

of patellofemoral procedures of patellofemoral procedures

2014 2015 2016 2014 2015 2016
Total caseload of patellofemoral surgery Total caseload of patellofemoral surgery

by unit annual frequency by consultant annual frequency
2014 2015 2016 2014 2015 2016
1 2 3 4 5−10 11−20 >20 1 2 3 4 5−10 11−20 >20
3.5.1.4 Age and gender characterisation of the who underwent primary knee surgery as recorded on
primary knee patient cohort the NJR was aged 102 years.
Table 3.24 shows the age and gender distribution Over all operation types, a higher percentage of females
of patients undergoing a first replacement of their (56.8%) than males have had a knee joint replaced.
knee joint. The median age of a person receiving a Women are also more likely to have a primary total knee
cemented total knee replacement was 70 years (IQR replacement; 57.3%, 52.0% and 55.6% of cemented,
64-76 years). However, for unicompartmental primary uncemented and hybrid type procedures respectively
knee surgery, patients were typically six (unicondylar; are carried out on female patients. Conversely,
median age 64 years; IQR 57-70) and twelve years unicondylar surgery is performed on a higher proportion
younger (patellofemoral; median age 58 years; IQR 51- of males (53%). Patellofemoral surgery is predominantly
67). The 99th percentile of patient age for all types of carried out on females (77.6% of patients) who are
surgery ranged between 85 and 88 years, indicating typically younger than a TKR or unicondylar patient with
that surgery was rarely undertaken in a person aged 90 a median age at operation of 58.
years or older, although the maximum age of a patient
Table 3.24 Age (in years) and percentage (%) male at primary operation1, 2 for different types of knee replacement
and by fixation, constraint and bearing type.
Age of patient (years)

Constraint and Percentage
Fixation method bearing type (%) male1 Median (IQR)2 Minimum age Maximum age
All cemented 42.3 70 (64-76) 7 102
Cememented and unconstrained, fixed 42.7 70 (64-76) 13 102
unconstrained, mobile 42.7 69 (62-76) 22 98
posterior-stabilised,
41.0 70 (64-77) 15 102
fixed
44.9 66 (60-73) 22 95
mobile
constrained, condylar 36.0 71 (63-78) 18 97
bearing type unknown 42.1 70 (63-77) 7 99

monobloc polyethylene
40.7 74 (69-79) 25 96
tibia
All uncemented 48.0 69 (62-75) 20 101
All hybrid 44.4 69 (62-76) 23 96
Uncemented/
unconstrained, fixed 48.0 69 (62-76) 24 99
hybrid and
unconstrained, mobile 45.6 69 (62-75) 25 101
51.7 66 (59-74) 20 94
fixed
other type 64.5 66 (60-74) 33 93
All unicondylar 53.0 64 (57-70) 18 97
Unicondylar and fixed 53.9 63 (56-70) 18 97
mobile 52.6 64 (57-71) 23 95
All patellofemoral 22.4 58 (51-67) 21 93
Fixation unknown bearing type unknown 46.7 69 (59-77) 43 85
All types 43.2 69 (63-76) 7 102
Note: 1 The percentage male figures are based on a total number of 975,737 primary knee replacements after omitting two cases where gender was not specified.
2 Age distribution based on age at primary operation excluding 181 with age registered as less than or equal to zero or unverifiable age or gender. Figures are thus
based on a total of 975,558 replace primary knee joints. The interquartile range (IQR) shows the age range of the middle 50% of patients arranged in order of their age
at time of primary knee operation.
112
3.5.2 First revision after primary • Type of brand (section 3.5.2.5): The cumulative
percentage chance of revision for different implant
knee surgery brands at different points in time after primary
A total of 24,399 first revisions of a knee prosthesis surgery is looked at in Tables 3.29 to 3.31. These
have been linked to NJR primary knee replacement tables have additional columns detailing brand
surgery records of operations undertaken between specific summaries of patient age at primary
2003 and 2016. operation (median and IQR) and the proportion
of males receiving the particular implant brand at
This section explores how different surgical, clinical and primary surgery.
patient factors affect the estimated cumulative probability
of a knee prosthesis being revised for the first time at 3.5.2.1 Temporal trends in the cumulative
increasing time points after the primary surgery. probability of a first revision by year of primary
knee replacement
In brief, the main factors we consider, with references to
the main results associated with these, are: Figures 3.17 (a) and (b) illustrate temporal changes
in the overall revision rates using Kaplan-Meier;
• Year of primary operation (section 3.5.2.1): procedures have been grouped by the year of the
Formal submission of records of joint replacement primary operation. Figure 3.17 (a) plots each Kaplan-
surgery taking place in England and Wales to a Meier survival curve with a common origin, i.e. time zero
national database was not a mandatory requirement is equal to the year of operation. Figure 3.17 (b) shows
in the initial years of the NJR. Figures 3.17 (a) and the same curves plotted against calendar time, where
(b) review the chance of knee implant first revision the origin of each curve is the year of operation. Figure
by year of operation given the shift from optional to 3.17 (b) separates each year allowing changes in failure
mandatory record keeping. rates to be clearly identified.
• Age and gender (section 3.5.2.2): Figures 3.18 (a)
and (b) show age and age-gender stratified Kaplan- In addition, the revision rate at 1, 3 and 5 years has
Meier estimates of the cumulative percentage been highlighted. If revision rates and timing of revision
chance of revision after primary surgery. rates were static across time we would expect all failure
curves to be the same shape and equally spaced, a
• Fixation method and constraint (section
departure from this indicates a change in the number,
3.5.2.3): Implant survivorship up to 13 years after
and timing of revision procedures.
the primary operation date are presented in Tables
3.25 (a) and 3.25 (b) broken down by fixation The cumulative probability of a joint being revised at
method and then by constraint and bearing within three and five years increased for each operative year
fixation method. The latter table also gives age group between 2003 and 2008; the probability of
group and gender sub-divisions of survivorship, being revised at 3 and 5 years reduced for operations
when numbers are sufficient for these sub-groups. performed between 2009 and 2016. From the peak in
Figures 3.19 (a), (b) and (c) compare the implant 2008, the yearly survivorship curves are less divergent,
survivorship of different bearing/constraints when i.e. a slowing in the increasing trend.
the method of fixation used for the knee joint
was each of cemented, uncemented/hybrid or a Possible reasons for a peak in the probability of revision
unicompartmental replacement, respectively. in the 2008 cohort is: 1) the registry was not capturing
• Clinical reasons for revision (section 3.5.2.4): the full range and number of operations taking place
Revision rates for different reasons, broken down in units in England and Wales until 2008, and 2) there
by fixation method and by fixation/constraint and could be bias in terms of the general overall health, risk
bearing, are shown in Tables 3.26 and 3.27. Table of revision, and other key characteristics of the patients
3.28 considers whether revision rates for different on record in the NJR in the early years.
reasons change over various periods of time after
the date of primary surgery.
Figure 3.17 (a)
Changes in cumulative percentage chance of knee replacement failure by year of primary operation.
Kaplan-Meier estimates of cumulative percentage probability of a first revision grouped by year in which
primary surgery took place.
6
5
2003
3 2004
2005
2006
2007
2 2008
2009
2010
2011
1 2012
2013
2014
2015
0
2016
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years since primary surgery
Figure 3.17 (b)
Temporal changes in revision rates after primary knee replacement: Kaplan-Meier estimates of cumulative percentage probability
of revision for each year of primary operation with failure rates at 1, 3, and 5 years indicated.
2
0
2003 2005 2007 2009 2011 2013 2015 2017
Year of primary surgery

Cumulative probability of revision 1 year after primary
115
3.5.2.2 Revisions after primary knee surgery by higher in younger patient cohorts and that men were
grouped age at primary and gender slightly more likely, overall, to have a first revision
compared to women of comparable grouped age, if
Figures 3.18 (a) shows that the chance of revision they were under the age of 75 when they underwent
after primary cemented total knee replacement is far primary surgery.
Figure 3.18 (a)

Kaplan-Meier estimates of the cumulative percentage probability of a first revision of primary
cemented knee replacement broken down by age group (age at primary in years) at increasing
years after the primary surgery.
Males Females
12 12
10 10
8 8
6 6
4 4
Under 55 y
55−59 y
2 2 60−64 y
65−69 y
70−74 y
75−79 y
0 0
80+ y
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Figure 3.18 (b) shows that the risk of revision of patients in the risk of revision according to gender.
primary unicondylar knee replacement is, again, The risk of revision appears to be higher in females
substantially higher for younger patient cohorts but over the age of 75 compared to males.
that there are less marked differences in younger
Figure 3.18 (b)

Kaplan-Meier estimates of the cumulative percentage probability of a first revision of primary
unicondylar knee replacement broken down by age group (age at primary in years) and gender at
increasing years after the primary surgery.
Males Females
30 30

25 25
20 20
15 15
10 10
Under 55 y
55−59 y
60−64 y
5 5
65−69 y
70−74 y
75−79 y
0 0
80+ y
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13
3.5.2.3 Revisions after primary knee surgery by patellofemoral replacement is over four times higher
fixation method and constraint at 12 and 13 years although less than 250 remain at
risk at 13 years. First revision of an implant is slightly
Table 3.25 (a) shows Kaplan-Meier estimates of the less likely in females than males overall for the most
cumulative percentage probability of first revision, commonly used fixation method (cemented) but,
for any cause, for the cohort of all primary knee broadly, a patient from a younger age group is more
replacements. This is broken down for TKR by knee likely to be revised irrespective of gender, with the
fixation type and sub-divided further within each youngest group having the worst predicted outcome
fixation type by bearing/constraint type and for UKR, in terms of the risk of subsequent revision. Conversely,
by bearing/constraint type. Estimates are shown, female patients are more likely to have a unicondylar
together with 95% Confidence Intervals (95% CI), at implant revised compared to their male, age
each year after primary surgery. equivalent, counterpart. The reverse pattern is seen
in patellofemoral implant survivorship. It is clear that
Table 3.25 (b) shows gender and age stratified partial knee replacement surgery is used generally in
Kaplan-Meier estimates of the cumulative percentage younger patients. Younger patients may also be more
probability of first revision, for any revision cause, active which may put more strain on their implants and
firstly for all cases combined, then by knee fixation/ increase the risk of revision. However, for the first time
constraint sub-divisions. Estimates are shown, along in this report, we have reported revision by age group
with 95% CI, at 1, 3, 5, 7, 10 and 13 years after the and the pattern is consistent across age groups.
primary operation.
Figures 3.19 (a) and (b) explore the chance of knee
Estimates in blue italics indicate that the cumulative joint revision for different bearings and constraints
percentage probability of a first revision of a knee within a particular knee fixation type; that of cemented,
joint replacement estimate is less reliable as these are uncemented/hybrid. Figure 3.19 (c) looks at the chance
based on fewer than 250 at risk at that point in time. of revision for the most commonly used constraints
In addition, for a group at risk size of fewer than 250, in a unicondylar knee replacement and patellofemoral
the upper 95% CI limit tends to be underestimated implants. It should be noted that unknown constraint/
by the estimation method used here. Other methods fixation combinations are not shown.
have been proposed which take into account the
impact that censoring has on estimation of CIs when Overall, little difference is seen in implant survivorship
numbers at risk are small. However, the upper limit by constraint within a fixation type apart from:
values found differ considerably and as yet there is no
clear consensus as to which method provides the most • Cemented unconstrained, fixed bearing total knee
accurate upper limit. Estimates (and CIs) are not given replacement results in lower chances of revision
when the number at risk falls below ten. overall compared to other combinations of constraint
and bearing used in a cemented fixation of the joint
Unicompartmental knee replacements seem to fare with modular tibial components (Figure 3.19 (a))
worse compared to total knee replacements with • Uncemented/hybrid total knee replacements (Figure
the chance of revision at each estimated time point 3.19 (b)) with posterior stabilised constraint and
being more than double that of a TKR. The revision fixed bearings fare worse than their unconstrained
rate for unicondylar (medial or lateral UKR) is 2.8 bearing equivalents
times higher than the observed rate for all types of
knee replacement at 13 years and the revision rate for
Figure 3.19 (a)

Comparison of the Kaplan-Meier cumulative percentage probability estimates of a knee prosthesis first
revision for different constraint and bearing types at increasing years after the primary surgery when the
primary arthroplasty method of fixation is cemented only.
(a) Cemented
6
Cumulative probability of revision (%)

4
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Unconstrained, fixed 551,832 483,984 416,668 351,961 295,334 240,847 192,174 149,207 110,352 76,222 48,250 28,724 13,171 3,982
Unconstrained, mobile 34,507 32,293 30,015 27,525 25,073 22,512 19,655 16,279 12,687 8,770 5,203 2,649 1,003 296
Posterior−stabilised, fixed 202,284 179,843 157,107 134,678 114,208 94,398 75,749 58,502 42,969 29,360 18,289 10,700 5,073 1,547
Posterior−stabilised, mobile 11,526 10,891 9,940 8,845 7,703 6,678 5,609 4,484 3,452 2,510 1,541 821 298 88
Constrained, condylar 6,428 4,945 3,696 2,617 1,853 1,348 1,018 748 544 395 234 147 69 19
Monobloc polyethylene tibia 12,765 11,100 9,447 7,486 5,590 3,883 2,618 1,881 1,353 832 398 185 74 30
Figure 3.19 (b)
primary arthroplasty method of fixation is uncemented or hybrid.
(b) Uncemented/hybrid
12
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Unconstrained, fixed 21,625 20,438 19,323 18,187 17,005 15,594 13,897 11,734 9,120 6,525 4,182 2,600 1,236 373
Unconstrained, mobile 23,395 21,711 19,700 17,582 15,772 13,636 11,361 9,207 7,177 5,084 3,309 1,967 948 285
Posterior−stabilised, fixed 3,376 3,156 2,866 2,519 2,262 2,027 1,807 1,546 1,211 904 602 347 162 56
Figure 3.19 (c)

primary is a unicondylar or patellofemoral partial knee replacement.
(c) Unicondylar and patellofemoral partial knee replacements
24
22

20
18
16
14
12
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Unicondylar, fixed 27,901 23,627 19,608 15,942 12,963 10,293 8,097 6,080 4,511 3,127 1,994 1,080 443 116
Unicondylar, mobile 56,523 50,628 44,641 39,084 34,471 29,588 24,580 19,659 14,819 10,303 6,530 3,795 1,795 538
Patellofemoral 12,191 10,981 9,614 8,355 7,203 5,924 4,752 3,653 2,670 1,750 1,018 588 267 83
122
Table 3.25 (a) Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) at specified times after primary knee replacement, by
fixation, constraint and bearing type1,2.
Cumulative percentage probability of a first revision (95% CI) at time shown if time elapsed since primary operation is:
Fixation/constraint/bearing type n 1 year 3 years 5 years 7 years 10 years 13 years
0.47 1.81 2.63 3.31 4.34 5.62
All types 975,739
(0.46-0.49) (1.78-1.84) (2.59-2.67) (3.26-3.36) (4.27-4.40) (5.47-5.77)
2.63 2.63
Others/unknown 45 0 0
(0.37-17.25) (0.37-17.25)
0.39 1.49 2.13 2.63 3.35 4.16
All cemented 828,573
(0.38-0.41) (1.46-1.52) (2.10-2.17) (2.59-2.67) (3.29-3.42) (4.03-4.29)
0.35 1.37 1.93 2.39 3.02 3.82
unconstrained, fixed 551,832
(0.33-0.37) (1.34-1.41) (1.89-1.98) (2.34-2.44) (2.95-3.10) (3.66-3.98)
0.52 1.85 2.76 3.35 4.23 5.00
unconstrained, mobile 34,507
(0.45-0.61) (1.71-2.01) (2.58-2.95) (3.14-3.57) (3.96-4.51) (4.52-5.54)
0.45 1.65 2.43 2.99 3.87 4.69
posterior-stabilised, fixed 202,284
(0.42-0.48) (1.59-1.71) (2.35-2.51) (2.90-3.08) (3.74-4.01) (4.43-4.96)
0.67 2.10 2.78 3.41 4.28 5.58
posterior-stabilised, mobile 11,526
(0.53-0.84) (1.84-2.39) (2.47-3.12) (3.05-3.82) (3.80-4.81) (4.30-7.22)
0.82 2.11 2.78 2.95 4.01 4.69
constrained, condylar 6,428
(0.62-1.09) (1.72-2.59) (2.27-3.40) (2.40-3.63) (2.96-5.43) (3.19-6.87)
0.35 1.33 1.78 2.09 2.64 3.34
monobloc polyethylene tibia 12,765
(0.26-0.47) (1.13-1.57) (1.52-2.08) (1.77-2.46) (2.14-3.24) (2.39-4.65)
0.80 2.41 3.43 4.20 5.00 5.88
bearing type unknown 9,231
(0.63-1.01) (2.11-2.76) (3.05-3.85) (3.76-4.68) (4.45-5.60) (4.84-7.15)
0.58 2.09 2.86 3.39 4.12 5.37
All uncemented 40,720
(0.51-0.66) (1.95-2.24) (2.69-3.04) (3.19-3.59) (3.87-4.37) (4.78-6.03)
0.65 2.32 2.98 3.50 4.20 5.30

(0.53-0.79) (2.08-2.58) (2.71-3.28) (3.19-3.83) (3.82-4.61) (4.49-6.26)
0.53 1.88 2.67 3.17 3.73 4.43
(0.44-0.64) (1.69-2.08) (2.44-2.91) (2.91-3.45) (3.42-4.07) (3.80-5.16)
0.62 2.49 3.59 4.44 6.35 12.07
posterior-stabilised, fixed 2,860
(0.38-0.99) (1.95-3.18) (2.90-4.44) (3.63-5.41) (5.22-7.70) (8.06-17.87)
2.03 2.60 2.60 2.60
other constraint 246 0
(0.77-5.34) (1.09-6.15) (1.09-6.15) (1.09-6.15)
0.65 1.62 3.38 3.95 4.92 5.50
bearing type unknown 466
(0.21-2.00) (0.77-3.37) (2.01-5.65) (2.43-6.38) (3.06-7.85) (3.44-8.75)
0.57 1.84 2.44 3.03 3.56 4.41
All hybrid 8,898
(0.43-0.75) (1.57-2.16) (2.12-2.81) (2.66-3.45) (3.12-4.05) (3.73-5.21)
0.47 1.60 2.17 2.67 3.18 4.04
(0.32-0.68) (1.31-1.96) (1.82-2.59) (2.28-3.14) (2.72-3.72) (3.30-4.95)
1.03 1.90 2.40 2.91 3.64 4.45
(0.64-1.65) (1.31-2.75) (1.68-3.43) (2.04-4.15) (2.41-5.47) (2.73-7.21)
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. 2 Blank cells indicate the number at risk is below ten and thus estimates have been omitted as they are highly unreliable.
Table 3.25 (a) (continued)
Cumulative percentage probability of a first revision (95% CI) at time shown if time elapsed since primary operation is:
Fixation/constraint/bearing type n 1 year 3 years 5 years 7 years 10 years 13 years
0.20 2.73 3.93 4.85 5.33 6.99
posterior-stabilised, fixed 516
(0.03-1.41) (1.60-4.66) (2.49-6.17) (3.17-7.38) (3.49-8.09) (3.97-12.14)
0.28 2.35 3.01 4.31 4.89
other constraint 386
(0.04-1.97) (1.18-4.65) (1.63-5.53) (2.50-7.37) (2.88-8.26)
1.61 6.89 6.89 9.57 9.57
(0.41-6.29) (3.50-13.34) (3.50-13.34) (5.17-17.34) (5.17-17.34)
1.09 4.13 6.20 8.16 11.38 15.96
All unicondylar 85,312
(1.02-1.16) (3.98-4.28) (6.01-6.40) (7.93-8.40) (11.04-11.74) (15.10-16.87)
0.75 3.81 5.92 7.73 10.13 14.18
fixed 27,901
(0.65-0.86) (3.56-4.08) (5.58-6.28) (7.30-8.19) (9.49-10.80) (12.70-15.82)
1.26 4.29 6.37 8.36 11.82 16.54
mobile 56,523
(1.17-1.36) (4.11-4.47) (6.14-6.60) (8.09-8.65) (11.41-12.25) (15.53-17.61)

0.68 3.75 4.92 7.34 11.01 12.99
(0.31-1.51) (2.67-5.26) (3.65-6.63) (5.67-9.46) (8.50-14.20) (9.58-17.49)
1.11 5.78 9.54 12.89 18.18 24.21
All patellofemoral 12,191
(0.93-1.32) (5.35-6.26) (8.95-10.17) (12.15-13.68) (17.05-19.38) (21.84-26.79)
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. 2 Blank cells indicate the number at risk is below ten and thus estimates have been omitted as they are highly unreliable.
123
124
Table 3.25 (b) Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) at specified times after primary knee replacement, by age and
gender1, for each fixation, constraint and bearing group. Blue italics signify that fewer than 250 cases remained at risk at these time points.
Males Females
Cumulative percentage probability of a first revision (95% CI) at time shown if Cumulative percentage probability of a first revision (95% CI) at time shown if
Age at
time elapsed since primary operation is: time elapsed since primary operation is:
primary
Fixation group (years) n 1 year 3 years 5 years 7 years 10 years 13 years n 1 year 3 years 5 years 7 years 10 years 13 years
1.08 4.41 6.40 8.41 11.14 15.93 0.81 3.89 6.14 7.97 10.59 14.16
All types <55 30,095 42,413
(0.96-1.20) (4.16-4.68) (6.08-6.73) (8.01-8.83) (10.57-11.75) (14.33-17.69) (0.73-0.90) (3.69-4.11) (5.88-6.42) (7.63-8.31) (10.11-11.09) (13.02-15.39)
0.82 3.51 5.04 6.70 8.96 11.83 0.57 2.85 4.44 5.55 7.47 10.02
All cemented <55 19,966 28,336
(0.70-0.96) (3.24-3.81) (4.69-5.42) (6.25-7.17) (8.31-9.66) (10.13-13.78) (0.49-0.67) (2.64-3.08) (4.16-4.74) (5.21-5.92) (6.96-8.02) (8.87-11.30)
0.73 3.11 4.53 6.17 8.42 11.83 0.50 2.44 3.87 4.92 6.60 9.51
unconstrained, fixed <55 12,630 18,187
(0.59-0.90) (2.79-3.48) (4.11-5.00) (5.62-6.77) (7.60-9.32) (9.41-14.82) (0.41-0.62) (2.20-2.71) (3.54-4.23) (4.51-5.37) (6.00-7.27) (7.92-11.40)
1.09 4.27 5.84 7.38 9.96 13.84 0.70 3.27 5.41 6.82 8.57 10.39
unconstrained, mobile <55 1,220 1,491
(0.64-1.88) (3.23-5.63) (4.59-7.43) (5.89-9.22) (7.76-12.75) (8.45-22.21) (0.38-1.29) (2.44-4.37) (4.28-6.82) (5.49-8.46) (6.76-10.84) (7.48-14.34)
0.69 3.92 5.78 7.42 10.23 12.58 0.52 3.37 5.15 6.39 8.82 11.64
posterior-stabilised, fixed <55 4,810 6,986
(0.49-0.98) (3.35-4.59) (5.04-6.64) (6.49-8.47) (8.80-11.87) (10.00-15.76) (0.37-0.73) (2.92-3.88) (4.55-5.82) (5.67-7.20) (7.70-10.10) (9.53-14.17)
posterior-stabilised, 1.40 4.06 4.82 6.16 7.23 1.39 4.64 5.96 6.40 9.19
<55 650 726
mobile (0.73-2.68) (2.76-5.95) (3.37-6.87) (4.40-8.59) (5.09-10.23) (0.75-2.57) (3.30-6.50) (4.40-8.04) (4.75-8.58) (6.68-12.58)
3.67 5.83 6.85 6.85 6.85 0.73 1.69 1.69 1.69 5.47
constrained, condylar <55 225 315
(1.85-7.21) (3.22-10.45) (3.83-12.07) (3.83-12.07) (3.83-12.07) (0.18-2.90) (0.63-4.47) (0.63-4.47) (0.63-4.47) (1.38-20.38)
monobloc polyethylene 0.87 5.70 5.70 7.38 7.38 1.19 3.86 4.98 4.98 4.98
<55 124 177
tibia (0.12-6.01) (2.59-12.30) (2.59-12.30) (3.49-15.25) (3.49-15.25) (0.30-4.69) (1.75-8.41) (2.35-10.39) (2.35-10.39) (2.35-10.39)
2.00 5.54 8.70 12.70 13.68 1.81 5.86 9.15 11.09 12.68 12.68
bearing type unknown <55 307 454
(0.90-4.39) (3.43-8.89) (5.84-12.86) (8.95-17.86) (9.63-19.25) (0.91-3.59) (4.00-8.56) (6.70-12.43) (8.25-14.81) (9.27-17.23) (9.27-17.23)
0.66 4.14 5.53 6.84 8.15 15.22 0.86 4.36 6.36 7.06 8.38 8.38
All uncemented <55 1,562 1,658
(0.36-1.23) (3.21-5.34) (4.40-6.93) (5.52-8.45) (6.58-10.08) (9.46-24.01) (0.51-1.45) (3.44-5.52) (5.20-7.76) (5.81-8.56) (6.91-10.15) (6.91-10.15)
0.81 4.43 5.94 7.28 8.49 1.17 3.51 4.99 5.76 7.29 7.29
unconstrained, fixed <55 638 617
(0.34-1.94) (3.01-6.49) (4.23-8.32) (5.30-9.97) (6.20-11.57) (0.56-2.44) (2.28-5.39) (3.44-7.21) (4.04-8.16) (5.14-10.29) (5.14-10.29)
0.73 4.34 5.60 7.27 8.27 8.27 0.71 4.26 6.48 7.09 8.28
unconstrained, mobile <55 709 854
(0.30-1.74) (2.99-6.27) (4.00-7.81) (5.31-9.93) (6.00-11.34) (6.00-11.34) (0.32-1.58) (3.04-5.94) (4.90-8.56) (5.40-9.28) (6.32-10.81)
1.78 2.62 2.62 6.13 0.63 8.05 11.25 12.32 13.61
posterior-stabilised, fixed <55 183 0 163
(0.58-5.43) (0.98-6.93) (0.98-6.93) (2.53-14.49) (0.09-4.38) (4.65-13.76) (7.02-17.76) (7.78-19.20) (8.67-21.01)
7.14
other constraint <55 18 0 4
(1.04-40.92)
8.33 8.33 8.33 5.00 5.00 5.00
bearing type unknown <55 14 0 20
(1.22-46.10) (1.22-46.10) (1.22-46.10) (0.72-30.53) (0.72-30.53) (0.72-30.53)
0.90 3.41 6.17 7.76 8.71 0.51 3.05 4.87 7.04 7.46 8.93
All hybrid <55 338 417
(0.29-2.76) (1.90-6.07) (3.97-9.51) (5.21-11.48) (5.77-13.04) (0.13-2.01) (1.70-5.44) (3.05-7.73) (4.72-10.45) (5.04-10.98) (5.65-13.96)
1.01 3.09 5.85 7.10 7.10 0.80 3.78 5.54 7.00 7.00 8.94
unconstrained, fixed <55 198 255
(0.25-3.98) (1.40-6.74) (3.28-10.32) (4.18-11.94) (4.18-11.94) (0.20-3.17) (1.98-7.14) (3.25-9.36) (4.34-11.19) (4.34-11.19) (5.09-15.44)
1.92 4.65 7.72
unconstrained, mobile <55 61 0 87 0 0 0 0
(0.27-12.88) (1.16-17.70) (2.50-22.53)
Note: 1 Excludes 181 cases where either gender was not specified and/or ages were invalid and NHS trace not complete. Total sample on which results are based is 975,558 primary knee replacements.
Males Females
Age at
primary
3.23 7.43 2.44 5.08 8.14
posterior-stabilised, fixed <55 36 0 0 42
(0.46-20.77) (1.88-26.93) (0.35-16.08) (1.29-18.83) (2.68-23.30)
8.82 11.86 11.86 11.86 4.35 13.46 13.46
other constraint <55 34 24 0 0
(2.93-24.91) (4.62-28.60) (4.62-28.60) (4.62-28.60) (0.62-27.07) (4.55-36.18) (4.55-36.18)
1.70 6.33 9.21 12.16 16.62 25.07 1.49 6.29 9.93 13.18 17.35 26.49
All unicondylar <55 7,350 8,285
(1.42-2.03) (5.74-6.97) (8.46-10.02) (11.22-13.18) (15.19-18.16) (21.13-29.60) (1.24-1.79) (5.73-6.89) (9.19-10.73) (12.26-14.17) (16.03-18.77) (22.21-31.41)
1.31 5.42 8.32 10.69 13.43 1.10 6.24 9.33 12.51 16.37 21.82
fixed <55 3,019 3,084
(0.94-1.81) (4.56-6.43) (7.15-9.67) (9.18-12.42) (11.24-16.01) (0.78-1.57) (5.32-7.30) (8.11-10.71) (10.93-14.29) (14.09-18.97) (16.79-28.09)
1.96 6.86 9.74 12.89 17.94 25.97 1.73 6.33 10.24 13.60 17.80 28.20
mobile <55 4,242 5,106
(1.57-2.44) (6.08-7.73) (8.77-10.80) (11.69-14.20) (16.17-19.88) (21.26-31.49) (1.40-2.14) (5.65-7.10) (9.31-11.25) (12.47-14.83) (16.20-19.53) (22.78-34.60)
2.27 9.37 11.94 16.87 22.06 1.05 6.47 9.92 9.92 17.64
(0.57-8.78) (4.80-17.87) (6.60-21.08) (10.05-27.53) (12.16-38.07) (0.15-7.24) (2.96-13.84) (5.28-18.22) (5.28-18.22) (8.64-34.10)
2.46 9.56 14.85 18.67 23.40 1.09 6.25 10.33 14.53 20.49 24.76
All patellofemoral <55 876 3,715
(1.59-3.79) (7.63-11.95) (12.29-17.87) (15.61-22.24) (19.12-28.45) (0.80-1.50) (5.44-7.18) (9.21-11.57) (13.09-16.11) (18.25-22.97) (20.45-29.80)
0.68 2.60 3.78 4.71 6.24 8.30 0.48 2.27 3.43 4.36 5.71 7.27
All types 55-64 105,622 127,295
(0.63-0.73) (2.50-2.71) (3.65-3.92) (4.56-4.87) (6.02-6.47) (7.85-8.78) (0.44-0.52) (2.18-2.36) (3.31-3.55) (4.22-4.50) (5.51-5.91) (6.87-7.69)
0.60 2.28 3.26 3.99 5.10 6.49 0.39 1.88 2.78 3.44 4.37 5.43
All cemented 55-64 82,071 103,933
(0.55-0.65) (2.17-2.40) (3.12-3.40) (3.82-4.16) (4.88-5.34) (6.07-6.95) (0.35-0.43) (1.79-1.98) (2.67-2.90) (3.30-3.58) (4.18-4.56) (5.05-5.83)
0.51 2.09 2.96 3.64 4.60 6.17 0.36 1.74 2.49 3.06 3.93 4.83
unconstrained, fixed 55-64 54,985 69,362
(0.45-0.57) (1.97-2.23) (2.79-3.13) (3.44-3.84) (4.33-4.88) (5.61-6.78) (0.31-0.41) (1.64-1.85) (2.35-2.63) (2.90-3.23) (3.71-4.17) (4.43-5.26)
0.74 2.57 3.78 4.62 5.67 5.97 0.49 2.16 3.28 3.91 4.89 6.38
unconstrained, mobile 55-64 4,126 4,874
(0.52-1.06) (2.11-3.12) (3.21-4.45) (3.97-5.37) (4.87-6.60) (5.01-7.10) (0.32-0.73) (1.77-2.63) (2.78-3.86) (3.35-4.56) (4.19-5.70) (4.86-8.33)
0.74 2.63 3.90 4.77 6.19 7.53 0.44 2.16 3.37 4.17 5.19 6.47
posterior-stabilised, fixed 55-64 19,132 25,054
(0.62-0.87) (2.40-2.89) (3.59-4.23) (4.41-5.17) (5.68-6.75) (6.74-8.42) (0.37-0.54) (1.97-2.37) (3.11-3.64) (3.87-4.49) (4.78-5.63) (5.59-7.49)
posterior-stabilised, 0.76 2.26 2.70 3.31 4.91 5.31 0.37 1.62 2.84 3.81 4.84 7.72
55-64 1,731 1,944
mobile (0.44-1.31) (1.64-3.11) (2.01-3.63) (2.50-4.39) (3.64-6.61) (3.88-7.24) (0.17-0.77) (1.12-2.32) (2.14-3.77) (2.94-4.93) (3.74-6.25) (3.66-15.90)
0.96 2.79 4.13 4.13 4.13 0.43 1.95 1.95 1.95 5.22
constrained, condylar 55-64 589 764
(0.40-2.31) (1.52-5.10) (2.25-7.52) (2.25-7.52) (2.25-7.52) (0.14-1.34) (1.04-3.67) (1.04-3.67) (1.04-3.67) (2.08-12.80)
55-64 582 779
tibia (0.29-2.03) (1.14-3.91) (1.81-5.46) (2.14-6.53) (2.59-8.27) (0.07-1.09) (1.13-3.47) (1.65-4.62) (1.65-4.62) (2.74-8.45)
1.53 4.44 5.58 5.89 7.98 9.30 0.80 3.01 4.37 5.84 6.37 6.37
bearing type unknown 55-64 926 1,156
(0.91-2.57) (3.26-6.03) (4.23-7.34) (4.49-7.70) (5.97-10.63) (6.78-12.68) (0.42-1.52) (2.15-4.21) (3.29-5.79) (4.52-7.54) (4.94-8.18) (4.94-8.18)
0.59 2.30 3.27 4.02 5.16 7.03 0.60 2.34 3.51 4.36 5.43 6.79
All uncemented 55-64 5,389 5,311
(0.42-0.84) (1.91-2.76) (2.80-3.83) (3.47-4.65) (4.44-5.98) (5.51-8.96) (0.42-0.85) (1.95-2.80) (3.02-4.09) (3.79-5.02) (4.71-6.24) (5.48-8.41)
Note: 1 Excludes 181 cases where either gender was not specified and/or ages were invalid and NHS trace not complete. Total sample on which results are based is 975,558 primary knee replacements. Continued >
125
126
Table 3.25 (b) (continued)
Males Females
Age at
primary
0.45 2.42 3.39 4.19 5.43 6.23 0.72 2.63 3.42 4.49 5.41 6.55
(0.23-0.86) (1.82-3.22) (2.65-4.33) (3.34-5.25) (4.31-6.82) (4.74-8.16) (0.43-1.22) (2.00-3.47) (2.68-4.37) (3.60-5.60) (4.32-6.78) (5.00-8.56)
0.58 2.22 3.23 3.90 4.41 5.98 0.50 2.05 3.45 4.03 4.97 5.50
(0.35-0.97) (1.71-2.90) (2.58-4.04) (3.15-4.83) (3.56-5.46) (4.27-8.33) (0.30-0.85) (1.57-2.68) (2.79-4.27) (3.28-4.94) (4.06-6.07) (4.38-6.89)
0.96 1.80 2.32 3.44 7.43 0.82 3.45 5.18 6.80 9.81
posterior-stabilised, fixed 55-64 536 377
(0.40-2.28) (0.94-3.43) (1.28-4.16) (2.01-5.88) (4.65-11.78) (0.27-2.53) (1.97-6.01) (3.24-8.22) (4.46-10.29) (6.62-14.43)
2.86 2.86
other constraint 55-64 46 0 24 0 0 0
(0.41-18.60) (0.41-18.60)
2.68 4.18 7.27 7.27 7.27 1.72 1.72 1.72
bearing type unknown 55-64 75 63 0 0
(0.68-10.31) (1.37-12.43) (3.08-16.64) (3.08-16.64) (3.08-16.64) (0.24-11.62) (0.24-11.62) (0.24-11.62)
0.51 1.89 2.62 3.65 4.49 7.81 0.53 2.16 3.14 3.76 4.13 4.13
All hybrid 55-64 1,013 1,193
(0.21-1.22) (1.20-2.99) (1.76-3.89) (2.57-5.18) (3.15-6.39) (5.27-11.50) (0.24-1.17) (1.44-3.24) (2.23-4.41) (2.73-5.16) (3.01-5.66) (3.01-5.66)
0.31 1.25 2.28 3.06 3.76 6.57 0.79 2.42 3.29 3.91 4.39 4.39
unconstrained, fixed 55-64 651 769
(0.08-1.23) (0.63-2.49) (1.35-3.82) (1.94-4.83) (2.39-5.89) (4.02-10.64) (0.36-1.75) (1.53-3.82) (2.22-4.87) (2.72-5.62) (3.07-6.27) (3.07-6.27)
1.08 1.80 1.80 1.80 1.80 1.01 1.79 1.79 1.79
unconstrained, mobile 55-64 195 283 0
(0.27-4.27) (0.58-5.54) (0.58-5.54) (0.58-5.54) (0.58-5.54) (0.25-3.98) (0.57-5.62) (0.57-5.62) (0.57-5.62)
3.01 3.01 3.01 3.01 2.67 5.57 7.77 7.77
posterior-stabilised, fixed 55-64 73 0 76
(0.76-11.51) (0.76-11.51) (0.76-11.51) (0.76-11.51) (0.67-10.24) (2.13-14.19) (3.24-18.04) (3.24-18.04)
1.47 6.05 6.05 8.58 12.74 2.33 2.33 2.33 2.33

other constraint 55-64 75 54 0
(0.21-9.98) (2.31-15.32) (2.31-15.32) (3.56-19.93) (5.41-28.41) (0.33-15.38) (0.33-15.38) (0.33-15.38) (0.33-15.38)
5.26 5.26 19.62
bearing type unknown 55-64 19 11 0
(0.76-31.88) (0.76-31.88) (6.60-50.26)
1.06 4.20 6.33 8.12 11.53 16.15 1.03 4.36 6.74 8.96 12.47 16.32
All unicondylar 55-64 16,273 13,927
(0.91-1.24) (3.88-4.55) (5.91-6.78) (7.61-8.66) (10.77-12.34) (14.42-18.07) (0.87-1.22) (4.00-4.74) (6.27-7.24) (8.38-9.57) (11.64-13.36) (14.66-18.14)
0.58 3.26 5.83 7.47 10.61 15.44 0.78 4.31 6.46 8.81 11.06 14.13
fixed 55-64 5,442 4,464
(0.40-0.83) (2.76-3.85) (5.09-6.67) (6.55-8.50) (9.20-12.22) (12.27-19.34) (0.56-1.11) (3.68-5.04) (5.63-7.40) (7.73-10.03) (9.58-12.75) (11.53-17.25)
1.30 4.66 6.64 8.44 11.98 16.63 1.17 4.40 6.92 9.14 13.06 17.09
mobile 55-64 10,652 9,311
(1.10-1.54) (4.25-5.11) (6.13-7.20) (7.83-9.09) (11.08-12.95) (14.63-18.87) (0.96-1.41) (3.97-4.87) (6.36-7.52) (8.46-9.87) (12.07-14.13) (15.17-19.23)
1.13 3.46 4.09 7.62 8.59 3.55 3.55 3.55 5.90
bearing type unknown 55-64 179 152 0
(0.28-4.44) (1.57-7.53) (1.97-8.39) (4.38-13.09) (5.03-14.47) (1.49-8.32) (1.49-8.32) (1.49-8.32) (2.32-14.57)
1.60 5.73 9.87 13.53 20.28 0.76 5.53 9.40 13.06 17.88 25.00
All patellofemoral 55-64 871 2,926
(0.93-2.74) (4.26-7.69) (7.80-12.47) (10.85-16.82) (15.85-25.74) (0.49-1.16) (4.70-6.51) (8.26-10.70) (11.61-14.68) (15.83-20.16) (19.63-31.53)
Males Females
Age at
primary
0.48 1.71 2.39 2.93 3.79 4.74 0.35 1.48 2.18 2.74 3.56 4.31
All types 65-74 167,309 206,258
(0.45-0.52) (1.65-1.78) (2.31-2.48) (2.83-3.03) (3.65-3.94) (4.44-5.07) (0.33-0.38) (1.43-1.54) (2.11-2.26) (2.65-2.83) (3.43-3.69) (4.09-4.54)
0.42 1.52 2.12 2.57 3.26 3.90 0.31 1.28 1.88 2.33 2.90 3.39
All cemented 65-74 142,221 182,765
(0.39-0.46) (1.45-1.59) (2.03-2.21) (2.47-2.68) (3.12-3.41) (3.65-4.17) (0.28-0.34) (1.22-1.34) (1.80-1.95) (2.24-2.42) (2.79-3.02) (3.19-3.60)
0.40 1.39 1.96 2.38 2.92 3.69 0.25 1.18 1.70 2.13 2.64 3.00
(0.36-0.44) (1.31-1.47) (1.86-2.06) (2.26-2.50) (2.76-3.09) (3.35-4.05) (0.22-0.28) (1.11-1.25) (1.62-1.79) (2.02-2.23) (2.50-2.78) (2.80-3.22)
0.53 1.87 2.71 3.16 4.24 4.68 0.44 1.59 2.29 2.94 3.66 4.39
(0.37-0.75) (1.54-2.28) (2.30-3.21) (2.70-3.71) (3.61-4.99) (3.85-5.69) (0.31-0.62) (1.31-1.93) (1.95-2.70) (2.53-3.42) (3.14-4.26) (3.57-5.39)
0.47 1.72 2.34 2.91 3.92 4.25 0.41 1.41 2.15 2.61 3.34 4.01
posterior-stabilised, fixed 65-74 33,423 45,345
(0.40-0.56) (1.57-1.88) (2.16-2.54) (2.69-3.15) (3.60-4.26) (3.86-4.68) (0.36-0.48) (1.30-1.54) (2.00-2.31) (2.43-2.80) (3.09-3.60) (3.51-4.58)
posterior-stabilised, 0.52 1.93 2.58 3.04 3.53 4.68 0.58 1.97 2.52 3.16 3.31 4.20
65-74 1,797 2,125
mobile (0.27-0.99) (1.36-2.74) (1.89-3.53) (2.24-4.11) (2.55-4.87) (3.07-7.10) (0.33-1.02) (1.44-2.70) (1.89-3.36) (2.41-4.15) (2.51-4.35) (2.63-6.69)
0.42 2.65 4.56 4.56 5.89 0.76 2.17 2.86 3.22 3.22
constrained, condylar 65-74 815 1,410
(0.13-1.30) (1.56-4.49) (2.83-7.32) (2.83-7.32) (3.32-10.34) (0.41-1.41) (1.42-3.32) (1.87-4.36) (2.09-4.96) (2.09-4.96)
65-74 1,950 2,901
tibia (0.03-0.44) (0.92-2.17) (1.27-2.85) (1.42-3.37) (1.42-3.37) (0.20-0.71) (1.03-2.01) (1.48-2.78) (1.67-3.18) (1.96-4.06)
0.73 2.51 3.07 3.35 3.91 3.91 0.52 1.52 2.62 3.31 3.90 4.75
bearing type unknown 65-74 1,531 1,753
(0.41-1.32) (1.82-3.46) (2.28-4.12) (2.51-4.45) (2.91-5.23) (2.91-5.23) (0.27-1.00) (1.03-2.24) (1.94-3.55) (2.51-4.37) (2.89-5.24) (3.10-7.24)
0.58 1.78 2.36 2.82 3.47 4.35 0.52 2.26 2.99 3.27 3.93 4.40
All uncemented 65-74 7,684 7,749
(0.43-0.78) (1.49-2.12) (2.02-2.76) (2.43-3.27) (2.96-4.05) (2.99-6.29) (0.38-0.71) (1.94-2.63) (2.61-3.43) (2.86-3.73) (3.43-4.50) (3.77-5.14)
0.66 2.25 2.85 3.30 3.89 3.89 0.56 2.89 3.42 3.67 4.28 4.54
(0.42-1.03) (1.76-2.88) (2.28-3.56) (2.66-4.08) (3.11-4.87) (3.11-4.87) (0.34-0.92) (2.30-3.61) (2.77-4.21) (3.00-4.50) (3.49-5.24) (3.64-5.67)
0.49 1.42 1.84 2.34 2.96 4.47 0.53 1.89 2.75 3.04 3.61 4.00
(0.31-0.77) (1.08-1.86) (1.43-2.35) (1.85-2.96) (2.32-3.77) (2.26-8.75) (0.35-0.80) (1.51-2.36) (2.27-3.33) (2.52-3.66) (2.97-4.37) (3.19-5.01)
1.03 1.94 2.87 2.87 3.33 0.24 2.18 3.18 3.56 4.80 6.41
posterior-stabilised, fixed 65-74 510 423
(0.43-2.45) (1.01-3.71) (1.62-5.05) (1.62-5.05) (1.90-5.79) (0.03-1.67) (1.09-4.32) (1.76-5.69) (2.03-6.22) (2.77-8.24) (3.39-11.97)
1.56 1.56 1.56
other constraint 65-74 77 0 27 0 0 0
(0.22-10.58) (0.22-10.58) (0.22-10.58)
5.56 6.97 10.55 1.56 1.56 1.56 4.38
bearing type unknown 65-74 81 0 0 74 0
(2.12-14.14) (2.96-15.94) (4.46-23.85) (0.22-10.58) (0.22-10.58) (0.22-10.58) (1.05-17.29)
0.54 2.06 2.31 2.59 3.33 3.70 0.57 1.74 1.95 2.27 2.69 2.98
All hybrid 65-74 1,544 1,788
(0.27-1.07) (1.44-2.95) (1.64-3.26) (1.87-3.60) (2.42-4.59) (2.62-5.23) (0.31-1.06) (1.21-2.50) (1.38-2.75) (1.64-3.15) (1.96-3.68) (2.12-4.18)
0.27 1.88 2.09 2.32 3.06 3.52 0.33 1.16 1.24 1.64 2.16 2.50
(0.09-0.84) (1.21-2.89) (1.38-3.15) (1.56-3.45) (2.09-4.48) (2.32-5.32) (0.12-0.86) (0.69-1.94) (0.75-2.06) (1.05-2.57) (1.43-3.26) (1.60-3.90)
1.35 1.35 2.14 3.13 3.13 1.61 3.76 4.43 4.43 4.43
unconstrained, mobile 65-74 253 389
(0.44-4.13) (0.44-4.13) (0.78-5.83) (1.24-7.78) (1.24-7.78) (0.72-3.54) (2.13-6.60) (2.52-7.73) (2.52-7.73) (2.52-7.73)
1.27 4.01 4.01 4.01 6.75 1.22 2.50 2.50 2.50
posterior-stabilised, fixed 65-74 80 90 0
(0.18-8.65) (1.31-11.93) (1.31-11.93) (1.31-11.93) (2.42-18.09) (0.17-8.34) (0.63-9.64) (0.63-9.64) (0.63-9.64)
other constraint 65-74 65 0 0 0 0 0 43 0 0 0 0 0
5.00 17.18 17.18 17.18 9.57 9.57
bearing type unknown 65-74 20 23
(0.72-30.53) (5.77-44.98) (5.77-44.98) (5.77-44.98) (2.47-33.22) (2.47-33.22)
Note: 1 Excludes 181 cases where either gender was not specified and/or ages were invalid and NHS trace not complete. Total sample on which results are based is 975,558 primary knee replacements. Continued >
127
128
Males Females
Age at
primary
0.90 3.24 4.60 5.91 8.23 12.29 0.84 3.38 5.20 7.07 10.49 14.18
All unicondylar 65-74 15,261 12,095
(0.76-1.07) (2.95-3.57) (4.22-5.01) (5.44-6.41) (7.53-9.00) (10.28-14.66) (0.69-1.03) (3.04-3.76) (4.75-5.69) (6.50-7.68) (9.60-11.45) (12.40-16.18)
0.59 3.26 4.76 5.97 6.94 11.51 0.49 2.89 4.38 5.87 8.48 9.51
fixed 65-74 4,674 3,607
(0.40-0.87) (2.71-3.90) (4.05-5.61) (5.08-7.02) (5.80-8.29) (7.67-17.08) (0.30-0.80) (2.32-3.61) (3.61-5.32) (4.86-7.07) (6.84-10.48) (7.46-12.09)
11.12
1.05 3.27 4.58 5.91 8.65 12.44 1.00 3.61 5.55 7.50 15.52
mobile 65-74 10,455 8,357 (10.08-
(0.87-1.27) (2.92-3.66) (4.15-5.07) (5.37-6.49) (7.81-9.57) (10.17-15.16) (0.81-1.25) (3.20-4.08) (5.01-6.15) (6.83-8.24) (13.37-17.96)
12.26)
1.56 2.52 4.69 4.69 1.65 2.64 6.09 9.67
bearing type unknown 65-74 132 0 131 0
(0.39-6.08) (0.81-7.67) (1.95-11.01) (1.95-11.01) (0.42-6.45) (0.85-7.99) (2.76-13.18) (4.75-19.14)
2.33 7.78 11.26 13.56 18.84 0.85 4.94 8.11 10.94 16.99 23.03
All patellofemoral 65-74 592 1,854
(1.36-3.99) (5.74-10.52) (8.66-14.58) (10.53-17.37) (13.96-25.16) (0.51-1.41) (3.99-6.11) (6.83-9.61) (9.35-12.78) (14.37-20.03) (18.93-27.86)
0.38 1.08 1.43 1.76 2.16 2.48 0.36 0.99 1.35 1.61 1.98 2.46
All types 75+ 118,460 178,106
(0.35-0.42) (1.02-1.15) (1.35-1.51) (1.67-1.87) (2.03-2.31) (2.24-2.75) (0.33-0.39) (0.94-1.04) (1.29-1.41) (1.54-1.69) (1.88-2.08) (2.20-2.76)
0.34 0.98 1.29 1.56 1.87 2.12 0.32 0.88 1.18 1.40 1.69 2.00
All cemented 75+ 105,766 163,365
(0.31-0.38) (0.92-1.05) (1.21-1.37) (1.46-1.66) (1.74-2.01) (1.92-2.35) (0.29-0.35) (0.83-0.93) (1.12-1.25) (1.33-1.47) (1.59-1.79) (1.79-2.22)
0.33 0.95 1.22 1.48 1.76 1.98 0.28 0.84 1.09 1.28 1.56 1.74
unconstrained, fixed 75+ 71,166 106,373
(0.29-0.38) (0.88-1.04) (1.12-1.32) (1.36-1.60) (1.60-1.92) (1.75-2.24) (0.25-0.31) (0.78-0.90) (1.02-1.17) (1.20-1.37) (1.45-1.69) (1.56-1.95)
0.40 1.03 1.71 1.92 2.19 2.19 0.41 1.01 1.46 1.64 2.02 2.15
unconstrained, mobile 75+ 3,679 6,208
(0.24-0.67) (0.74-1.44) (1.30-2.25) (1.47-2.50) (1.66-2.87) (1.66-2.87) (0.28-0.61) (0.78-1.31) (1.17-1.83) (1.32-2.04) (1.61-2.53) (1.68-2.75)
0.37 1.02 1.37 1.66 2.07 2.50 0.34 0.93 1.31 1.58 1.88 2.16
posterior-stabilised, fixed 75+ 25,580 41,915
(0.30-0.45) (0.89-1.17) (1.21-1.55) (1.47-1.88) (1.79-2.40) (2.01-3.10) (0.29-0.40) (0.83-1.03) (1.19-1.44) (1.44-1.74) (1.69-2.08) (1.88-2.48)
posterior-stabilised, 0.62 1.54 1.73 2.02 2.02 0.60 1.06 1.41 1.69 1.88
75+ 998 1,553
mobile (0.28-1.38) (0.92-2.59) (1.04-2.88) (1.20-3.39) (1.20-3.39) (0.31-1.14) (0.64-1.76) (0.88-2.25) (1.07-2.66) (1.20-2.96)
0.68 1.73 1.73 1.73 1.73 0.87 1.30 1.71 2.13 2.13
constrained, condylar 75+ 683 1,625
(0.26-1.82) (0.85-3.50) (0.85-3.50) (0.85-3.50) (0.85-3.50) (0.51-1.50) (0.80-2.12) (1.04-2.79) (1.22-3.71) (1.22-3.71)
75+ 2,538 3,709
tibia (0.09-0.51) (0.80-1.78) (1.00-2.17) (1.32-2.99) (1.32-2.99) (0.26-0.72) (0.53-1.15) (0.69-1.48) (0.77-1.85) (0.88-2.60)
0.09 0.61 1.02 1.41 2.16 2.16 0.72 1.42 1.85 2.26 2.57 7.07
bearing type unknown 75+ 1,122 1,982
(0.01-0.65) (0.27-1.35) (0.53-1.97) (0.76-2.61) (1.14-4.09) (1.14-4.09) (0.43-1.21) (0.97-2.08) (1.31-2.62) (1.62-3.16) (1.83-3.59) (2.38-19.98)
0.53 1.37 1.71 2.14 2.33 2.33 0.56 1.28 1.61 1.87 1.94 3.81
All uncemented 75+ 4,908 6,449
(0.36-0.79) (1.07-1.77) (1.35-2.16) (1.71-2.68) (1.86-2.92) (1.86-2.92) (0.40-0.78) (1.02-1.60) (1.31-1.97) (1.53-2.28) (1.59-2.37) (1.93-7.45)
0.60 1.14 1.51 1.73 2.02 2.02 0.71 1.47 1.79 1.85 1.85 3.10
unconstrained, fixed 75+ 1,913 2,493
(0.33-1.08) (0.74-1.77) (1.02-2.23) (1.17-2.53) (1.37-2.96) (1.37-2.96) (0.44-1.13) (1.05-2.05) (1.31-2.43) (1.36-2.50) (1.36-2.50) (1.61-5.93)
0.56 1.43 1.79 2.20 2.34 2.34 0.44 1.17 1.41 1.76 1.84 2.15
unconstrained, mobile 75+ 2,622 3,473
(0.33-0.94) (1.02-2.00) (1.30-2.45) (1.62-2.98) (1.72-3.18) (1.72-3.18) (0.27-0.73) (0.84-1.61) (1.04-1.90) (1.32-2.34) (1.38-2.45) (1.48-3.13)
Males Females
Age at
primary
2.67 2.67 4.11 4.11 0.53 1.14 2.54 3.65 4.36 11.19
posterior-stabilised, fixed 75+ 278 0 390
(1.20-5.86) (1.20-5.86) (2.03-8.21) (2.03-8.21) (0.13-2.11) (0.43-3.03) (1.19-5.38) (1.87-7.08) (2.29-8.21) (3.35-33.88)
2.70 2.70 2.70
other constraint 75+ 40 0 9
(0.39-17.68) (0.39-17.68) (0.39-17.68)
3.45 3.45 1.20 1.20 1.20 1.20 1.20
bearing type unknown 75+ 55 0 0 0 84
(0.49-22.05) (0.49-22.05) (0.17-8.25) (0.17-8.25) (0.17-8.25) (0.17-8.25) (0.17-8.25)
0.50 1.04 1.49 1.68 2.31 2.31 0.67 1.32 1.58 2.04 2.22 2.22
All hybrid 75+ 1,052 1,553
(0.21-1.19) (0.56-1.93) (0.86-2.58) (0.98-2.86) (1.34-3.98) (1.34-3.98) (0.36-1.24) (0.84-2.06) (1.04-2.40) (1.38-3.01) (1.50-3.27) (1.50-3.27)
0.41 0.99 1.56 1.78 2.57 2.57 0.57 1.06 1.39 1.81 2.02 2.02
unconstrained, fixed 75+ 761 1,082
(0.13-1.27) (0.47-2.07) (0.84-2.89) (0.98-3.22) (1.42-4.64) (1.42-4.64) (0.25-1.25) (0.59-1.90) (0.83-2.35) (1.12-2.91) (1.26-3.24) (1.26-3.24)
1.15 1.15 1.15 1.15 1.15 1.28 1.75 1.75 2.91 2.91
unconstrained, mobile 75+ 175 344
(0.29-4.52) (0.29-4.52) (0.29-4.52) (0.29-4.52) (0.29-4.52) (0.48-3.37) (0.72-4.21) (0.72-4.21) (1.13-7.34) (1.13-7.34)
2.27 2.27 2.27 4.99 4.99 4.99
posterior-stabilised, fixed 75+ 50 0 69 0
(0.32-15.06) (0.32-15.06) (0.32-15.06) (1.63-14.69) (1.63-14.69) (1.63-14.69)
other constraint 75+ 52 0 0 0 0 0 39 0 0 0 0 0
bearing type unknown 75+ 14 0 19 0 0
0.86 2.32 3.29 4.48 6.14 7.84 1.19 3.22 4.65 6.00 8.18 10.31
All unicondylar 75+ 6,342 5,760
(0.66-1.13) (1.94-2.77) (2.80-3.85) (3.83-5.23) (5.18-7.26) (5.41-11.27) (0.94-1.52) (2.76-3.75) (4.06-5.32) (5.28-6.82) (7.15-9.35) (7.99-13.25)
0.51 1.71 2.94 3.86 5.49 0.85 2.57 3.70 4.36 5.34
fixed 75+ 1,908 1,697
(0.27-0.98) (1.15-2.53) (2.08-4.15) (2.71-5.47) (3.74-8.03) (0.49-1.46) (1.83-3.58) (2.74-4.98) (3.25-5.84) (3.86-7.35)
1.02 2.57 3.47 4.70 6.11 7.58 1.33 3.47 4.99 6.53 8.93 11.44
mobile 75+ 4,382 4,005
(0.76-1.38) (2.10-3.14) (2.89-4.15) (3.94-5.60) (5.05-7.39) (5.04-11.33) (1.01-1.75) (2.91-4.13) (4.28-5.81) (5.66-7.53) (7.70-10.36) (8.60-15.15)
1.96 1.96 4.93 18.00 1.82 3.67 5.60 8.46
bearing type unknown 75+ 52 58
(0.28-13.11) (0.28-13.11) (1.22-18.85) (7.49-39.71) (0.26-12.21) (0.93-13.89) (1.84-16.37) (3.18-21.46)
0.57 3.70 5.15 5.85 8.27 0.54 2.88 5.70 6.88 9.32
All patellofemoral 75+ 386 969
(0.14-2.28) (2.06-6.60) (3.05-8.62) (3.50-9.72) (4.21-15.89) (0.22-1.29) (1.93-4.27) (4.20-7.70) (5.14-9.17) (6.81-12.70)
Others/unknown 13 19
129
3.5.2.4 Revisions for different clinical causes This means that the reasons for revision are not
after primary knee replacement surgery mutually exclusive of each other. In addition, over the
last 13 years, there have been a number of versions
The Kaplan-Meier estimates of the cumulative of the MDS form and the reasons for revision options
probability of a first revision of an implant that have available have varied across these versions. As a result
been presented so far, have been shown irrespective of these inconsistencies, we opt to use person-time
of the clinical reason given for the revision surgery. incidence rates (PTIR) for each reason for revision on
This sub-section looks more closely at the various record so that the incidence rates for each reason,
reasons recorded for revision on the MDS form. taking into account the different time periods of
availability, can be compared.
Clinicians can indicate more than one diagnosis as
the indication for revision surgery on the MDS form.
Methodological note: Patient-time incidence The PTIRs are given in the tables as the number of
rate (PTIR) revisions for that reason per 1,000 patient-years at risk
for the period of time considered.
Incidence rates for each reason have been calculated
using patient–time incidence rates (PTIRs). This is found The PTIR method assumes that the hazard rate remains
by dividing the total number of times a revision for that constant over the whole time period. When this may not
specific reason has been given in a period of time by the be appropriate, PTIR incidence rates for sub-divisions of
total number of years all patients have been at risk of the whole time period of interest can be calculated to see
revision (for any reason) over the time period. whether the hazard rate holds constant across smaller
time intervals.
In the earliest version of the MDS form for revision, form of all revision surgeries for primary operations which
MDSv1, both arthritis and incorrect sizing were available took place from 2005 onwards. Thus, for the PTIR
as clinical reasons for revision surgery to be performed. calculation for stiffness, we have restricted the period
Subsequent forms, however, omitted these options. a primary replaced knee joint is at risk of revision for
Similarly, stiffness became available as a clinical reason stiffness to all primary knee joint replacement surgeries
for revision surgery on the later forms MDSv2, MDSv3 which took place from 1 January 2005 onwards. This
and MDSv6 but was not an option on the MDSv1 form. explains why fewer patient-years at risk are shown for
stiffness in the tables discussed in this section.
As the number of cases of incorrect sizing is small and
the MDSv1 form on which it was an option ceased to Table 3.26 shows the revision incidence rates, for each
be used after 2004, we have added incorrect sizing to reason recorded on the MDS forms for knee revision
the Other indication category for estimating PTIRs. surgery, for all cases and then sub-divided by fixation
type and whether the primary procedure was a TKR or
In the case of stiffness, an adjustment needs to be an UKR.
made to the total number of patients considered to be
at risk as any revisions occurring before the MDSv2 Table 3.27 shows these first knee revision PTIRs for
form was issued could not have been at risk of this each reason broken down further by fixation, constraint
reason for revision as it could not be selected by the and bearing type.
clinician. Checking the year of the primary operation
against all knee joints which have been revised over the For TKRs, the highest PTIRs, in descending order, were
life of the registry, the MDSv2 and later versions were for revision due to aseptic loosening, pain and infection.
being used to record reasons for revision in over 95% Revision incidences for pain and aseptic loosening were
slightly higher for implants which were uncemented for each revision reason have been calculated for the
compared to prosthesis implanted using a hybrid or following time periods; <1 year, 1 to 3 years, 3 to 5
cemented fixation. years, 5 to 7, 7 to 10 and 10 to 13 years after the
primary surgery took place.
For patellofemoral type unicompartmental
replacements, the top three reasons for revision were Table 3.28 shows the PTIR for each specified reason
for Other indication (including progressive arthritis), pain for first revision for different periods of time after primary
and aseptic loosening. The first two reasons had the surgery. It is clear that most of the PTIRs for a particular
highest incidence rates across all reasons by fixation reason do vary, especially for infection, aseptic
method breakdowns. Similarly, for unicondylar knee loosening and pain for different time intervals after
replacements (medial and lateral unicompartmental surgery. Infection is most likely to be the reason that
knee replacements), the highest three incidence a joint is revised in the first year but after seven years
rates for reasons for revising the implant were Other or more, is less likely than other reasons. Conversely,
indication, aseptic loosening and pain, respectively. revision between one and three years after surgery
is more likely for aseptic loosening and pain, with
Interest also lies in whether PTIRs for different reasons incidence rates dropping off for pain later on. Aseptic
remain the same for different time intervals after loosening PTIRs continue to remain relatively higher
primary surgery and whether certain reasons for than other indicated reasons for revision for implants
revision are more pronounced in the short, medium or surviving for longer periods after surgery.
longer term after primary surgery. To this end, PTIRs
132
Table 3.26 Revision rates (95% CI), expressed as number of revisions per 1,000 patient-years (PTIRs), for each recorded reason for first knee revision. Rates shown are for all
revised cases by total replacement fixation method and by type of partial replacement.

Patient- Patient- per 1,000
years Peri- years patient-
at risk Dislocation/ Aseptic prosthetic Implant Implant Mal- Other at risk years for
Fixation (x1,000) All causes Pain subluxation Infection loosening Lysis fracture fracture1 wear1 Instability alignment indication2 (x1,000) stiffness3
4.81 0.92 0.19 0.95 1.27 0.24 0.15 0.03 0.28 0.69 0.39 1.08 0.34
All types 4,951.1 4,521.9
(4.75-4.87) (0.89-0.94) (0.18-0.20) (0.92-0.98) (1.24-1.30) (0.23-0.25) (0.14-0.16) (0.02-0.03) (0.26-0.29) (0.67-0.72) (0.38-0.41) (1.06-1.11) (0.32-0.36)
TKR by fixation method
3.79 0.63 0.12 1.01 1.01 0.22 0.14 0.02 0.17 0.65 0.34 0.59 0.35
Cemented 4,130.6 3,784.3
(3.73-3.85) (0.61-0.66) (0.11-0.13) (0.98-1.04) (0.98-1.04) (0.20-0.23) (0.13-0.15) (0.02-0.03) (0.16-0.18) (0.62-0.67) (0.32-0.36) (0.57-0.61) (0.33-0.37)
4.78 1.03 0.20 0.71 1.67 0.29 0.15 0.06 0.28 0.79 0.43 0.77 0.38
Uncemented 257.6 228.7
(4.52-5.05) (0.92-1.16) (0.15-0.26) (0.61-0.82) (1.52-1.83) (0.23-0.37) (0.11-0.20) (0.04-0.10) (0.23-0.36) (0.69-0.91) (0.35-0.51) (0.67-0.89) (0.31-0.47)
4.00 0.72 0.14 1.00 1.15 0.21 0.13 0.05 0.37 0.69 0.33 0.43 0.24
Hybrid 62.7 50.5
(3.54-4.53) (0.54-0.96) (0.07-0.28) (0.78-1.29) (0.91-1.45) (0.12-0.36) (0.06-0.26) (0.02-0.15) (0.24-0.55) (0.51-0.92) (0.22-0.51) (0.30-0.63) (0.13-0.42)
UKR type
12.49 2.89 0.78 0.58 3.42 0.43 0.27 0.04 1.07 1.02 0.70 4.70 0.22
Unicondylar 436.8 399.8
(12.16-12.83) (2.74-3.06) (0.70-0.86) (0.51-0.66) (3.25-3.60) (0.38-0.50) (0.23-0.33) (0.03-0.07) (0.97-1.17) (0.93-1.12) (0.63-0.79) (4.50-4.91) (0.17-0.27)
19.58 5.54 0.83 0.44 2.46 0.19 0.22 0.16 1.76 1.02 1.48 10.45 0.55
Patellofemoral 63.0 58.5
(18.51-20.70) (4.99-6.15) (0.63-1.08) (0.31-0.64) (2.10-2.88) (0.11-0.34) (0.13-0.38) (0.09-0.30) (1.46-2.12) (0.80-1.30) (1.21-1.81) (9.68-11.28) (0.39-0.77)
Note: 1 The reason implant failure, as reported on in annual reports up to 2013, has been renamed implant wear as this reflects the wearing down of the implant but distinguishes from the implant itself breaking. The latter cause for revision is now
indicated separately as implant fracture. 2 Other indication now includes arthritis and incorrect sizing. Both these reasons were only asked in MDSv1 and so are associated with primaries which took place in the first few years of the registry with little
potential for long term follow-up of the incidence of revision for these specific clinical reasons. 3 This reason was asked in versions MDSv2, v3 and v6 of the clinical assessment forms for joint replacement/revision surgery and hence, for these reasons,
there are fewer patient-years at risk.
Table 3.27 Revision rates (95% CI), expressed as number of revisions per 1,000 patient-years (PTIRs), for each recorded reason for first knee revision. Rates shown are broken
down by constraint and bearing sub-group for each total replacement fixation method and for unicondylar partial replacements.
Number of revisions per 1,000 patient-years (95% CI) for: Revision per
By fixation, Patient- Patient- 1,000
constraint and years Peri- years patient-
bearing sub- at risk Dislocation/ Aseptic prosthetic Implant Implant Other at risk years for
groups (x1,000) All causes Pain subluxation Infection loosening Lysis fracture fracture1 wear1 Instability Malalignment indication2 (x1,000) stiffness3
Cemented
unconstrained, 3.45 0.62 0.10 0.92 0.84 0.19 0.10 0.02 0.15 0.60 0.33 0.58 0.34
2,685.6 2,458.3
fixed (3.38-3.52) (0.59-0.65) (0.09-0.11) (0.88-0.96) (0.81-0.88) (0.17-0.21) (0.09-0.12) (0.01-0.02) (0.14-0.17) (0.57-0.63) (0.31-0.36) (0.55-0.61) (0.31-0.36)
unconstrained, 4.60 0.87 0.23 1.05 1.39 0.33 0.15 0.04 0.27 0.98 0.47 0.47 0.50
221.3 203.8
mobile (4.32-4.89) (0.75-1.00) (0.18-0.30) (0.92-1.19) (1.24-1.55) (0.27-0.42) (0.11-0.22) (0.02-0.07) (0.21-0.35) (0.85-1.12) (0.39-0.57) (0.39-0.57) (0.41-0.61)
posterior- 4.32 0.59 0.11 1.22 1.32 0.26 0.21 0.02 0.18 0.67 0.33 0.60 0.33
1,023.3 935.3
stabilised, fixed (4.19-4.45) (0.54-0.63) (0.09-0.13) (1.15-1.29) (1.25-1.39) (0.23-0.29) (0.18-0.24) (0.01-0.03) (0.16-0.21) (0.62-0.72) (0.30-0.37) (0.55-0.64) (0.30-0.37)
posterior- 4.79 0.92 0.19 0.92 1.15 0.28 0.26 0.06 0.28 1.03 0.22 0.93 0.70
68.7 64.1
stabilised, mobile (4.30-5.33) (0.72-1.17) (0.11-0.33) (0.72-1.17) (0.92-1.43) (0.18-0.43) (0.16-0.42) (0.02-0.16) (0.18-0.43) (0.82-1.30) (0.13-0.36) (0.73-1.19) (0.52-0.94)
constrained, 5.59 0.39 0.53 2.75 1.01 0.19 0.39 0.29 0.72 0.24 0.48 0.26
20.7 0.0 19.2
condylar (4.66-6.71) (0.19-0.77) (0.29-0.96) (2.12-3.56) (0.66-1.55) (0.07-0.51) (0.19-0.77) (0.13-0.64) (0.44-1.20) (0.10-0.58) (0.26-0.90) (0.11-0.63)
monobloc 3.36 0.57 0.16 0.90 0.74 0.16 0.21 0.04 0.10 0.62 0.43 0.51 0.34
51.2 50.0
polyethylene tibia (2.89-3.90) (0.39-0.81) (0.08-0.31) (0.67-1.20) (0.54-1.02) (0.08-0.31) (0.12-0.39) (0.01-0.16) (0.04-0.23) (0.44-0.88) (0.28-0.65) (0.35-0.75) (0.21-0.55)
bearing type 5.65 1.04 0.18 1.17 1.50 0.25 0.18 0.10 0.33 0.79 0.45 1.20 0.26
59.8 53.5
unknown (5.08-6.29) (0.81-1.33) (0.10-0.33) (0.93-1.48) (1.22-1.85) (0.15-0.42) (0.10-0.33) (0.05-0.22) (0.22-0.52) (0.59-1.05) (0.31-0.66) (0.96-1.52) (0.15-0.44)
Uncemented
unconstrained, 4.90 0.87 0.12 0.67 1.88 0.26 0.15 0.05 0.26 0.81 0.42 0.89 0.38
103.6 92.9
fixed (4.50-5.35) (0.71-1.07) (0.07-0.20) (0.53-0.84) (1.64-2.17) (0.18-0.38) (0.09-0.25) (0.02-0.12) (0.18-0.38) (0.65-1.00) (0.31-0.56) (0.72-1.09) (0.27-0.52)
unconstrained, 4.39 1.04 0.24 0.74 1.43 0.24 0.12 0.06 0.26 0.70 0.37 0.62 0.35
131.5 117.0
mobile (4.04-4.76) (0.88-1.23) (0.17-0.34) (0.60-0.90) (1.24-1.65) (0.17-0.34) (0.07-0.20) (0.03-0.12) (0.18-0.36) (0.57-0.86) (0.28-0.49) (0.50-0.77) (0.26-0.48)
posterior- 6.95 1.77 0.46 0.85 1.94 0.80 0.28 0.11 0.57 1.14 0.91 1.25 0.53
17.5 15.1
stabilised, fixed (5.82-8.30) (1.24-2.51) (0.23-0.91) (0.52-1.42) (1.38-2.71) (0.47-1.35) (0.12-0.68) (0.03-0.46) (0.31-1.06) (0.74-1.77) (0.56-1.49) (0.83-1.90) (0.27-1.06)
4.18 4.18 0.84 0.84 0.84 0.84 1.72
other constraint 1.2 0.0 0.0 0.0 0.0 0.0 0.0 1.2
(1.74-10.05) (1.74-10.05) (0.12-5.94) (0.12-5.94) (0.12-5.94) (0.12-5.94) (0.43-6.87)
bearing type 4.88 0.81 0.27 2.99 0.81 0.27 0.54 1.90 0.54 0.54 0.38
3.7 0.0 0.0 2.6
unknown (3.08-7.75) (0.26-2.52) (0.04-1.93) (1.65-5.39) (0.26-2.52) (0.04-1.93) (0.14-2.17) (0.91-3.98) (0.14-2.17) (0.14-2.17) (0.05-2.71)
Hybrid
unconstrained, 3.58 0.63 0.17 0.93 0.99 0.19 0.08 0.04 0.38 0.57 0.34 0.36 0.19
47.5 37.4
fixed (3.08-4.16) (0.44-0.90) (0.08-0.34) (0.69-1.24) (0.74-1.32) (0.10-0.36) (0.03-0.22) (0.01-0.17) (0.24-0.60) (0.39-0.83) (0.21-0.55) (0.22-0.58) (0.09-0.39)
unconstrained, 4.62 0.62 0.12 1.00 1.62 0.37 0.12 0.50 0.75 0.62 0.62 0.30
8.0 0.0 6.7
mobile (3.35-6.38) (0.26-1.50) (0.02-0.89) (0.50-2.00) (0.94-2.80) (0.12-1.16) (0.02-0.89) (0.19-1.33) (0.34-1.67) (0.26-1.50) (0.26-1.50) (0.07-1.19)
posterior- 6.08 1.11 1.94 1.94 0.28 0.55 1.38 0.32
3.6 0.0 0.0 0.0 0.0 0.0 3.1
stabilised, fixed (4.01-9.24) (0.42-2.95) (0.92-4.06) (0.92-4.06) (0.04-1.96) (0.14-2.21) (0.58-3.32) (0.05-2.29)
4.43 1.85 0.74 0.37 0.37 0.37 1.48 0.74 0.76
other constraint 2.7 0.0 0.0 0.0 0.0 2.6
(2.52-7.81) (0.77-4.44) (0.18-2.95) (0.05-2.62) (0.05-2.62) (0.05-2.62) (0.55-3.94) (0.18-2.95) (0.19-3.05)
bearing type 11.99 1.20 2.40 4.80 1.20 1.20 3.60
0.8 0.0 0.0 0.0 0.0 0.0 0.7 0.0
unknown (6.45-22.28) (0.17-8.51) (0.60-9.59) (1.80-12.78) (0.17-8.51) (0.17-8.51) (1.16-11.15)
Continued >
Note: 1 The reason implant failure, as reported on in annual reports up to 2013, has been renamed implant wear as this reflects the wearing down of the implant but distinguishes from the implant itself breaking. The latter cause
for revision is now indicated separately as implant fracture. 2 Other indication now includes arthritis and incorrect sizing. Both these reasons were only asked in MDSv1 and so are associated with primaries which took place in the
first few years of the Registry with little potential for long term follow-up of the incidence of revision for these specific clinical reasons. 3 This reason was asked in versions MDSv2, v3 and v6 of the clinical assessment forms for joint
replacement/revision surgery and hence, for these reasons, there are fewer patient-years at risk.
Number of revisions per 1,000 patient-years (95% CI) for: Revision per
By fixation, Patient- Patient- 1,000
constraint and years Peri- years patient-
bearing sub- at risk Dislocation/ Aseptic prosthetic Implant Implant Other at risk years for
groups (x1,000) All causes Pain subluxation Infection loosening Lysis fracture fracture1 wear1 Instability Malalignment indication2 (x1,000) stiffness3
Unicompartmental knee replacemental
Unicondylar
11.46 3.01 0.11 0.66 3.37 0.37 0.25 0.06 0.93 0.81 0.63 4.18 0.28
fixed 121.8 114.6
(10.88-12.08) (2.72-3.34) (0.06-0.18) (0.53-0.82) (3.06-3.71) (0.28-0.49) (0.18-0.36) (0.03-0.12) (0.77-1.12) (0.67-0.99) (0.51-0.79) (3.83-4.56) (0.20-0.39)
12.93 2.83 1.05 0.56 3.45 0.47 0.29 0.04 1.13 1.09 0.73 4.91 0.19
mobile 308.7 279.6
(12.53-13.33) (2.65-3.03) (0.94-1.17) (0.48-0.65) (3.25-3.66) (0.40-0.55) (0.23-0.35) (0.02-0.06) (1.01-1.25) (0.98-1.21) (0.64-0.83) (4.67-5.16) (0.15-0.25)
bearing type 11.12 3.50 0.16 0.32 2.86 0.16 0.16 0.79 1.43 0.64 4.45
6.3 0.0 5.6 0.0
unknown (8.80-14.06) (2.30-5.31) (0.02-1.13) (0.08-1.27) (1.80-4.54) (0.02-1.13) (0.02-1.13) (0.33-1.91) (0.74-2.75) (0.24-1.69) (3.07-6.44)
2.14 2.14 2.14
Others/unknown 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
(0.30-15.21) (0.30-15.21) (0.30-15.21)
Note: 1 The reason implant failure, as reported on in annual reports up to 2013, has been renamed implant wear as this reflects the wearing down of the implant but distinguishes from the implant itself breaking. The latter cause for revision is now
indicated separately as implant fracture. 2 Other indication now includes arthritis and incorrect sizing. Both these reasons were only asked in MDSv1 and so are associated with primaries which took place in the first few years of the Registry with little
potential for long term follow-up of the incidence of revision for these specific clinical reasons. 3 This reason was asked in versions MDSv2, v3 and v6 of the clinical assessment forms for joint replacement/revision surgery and hence, for these reasons,
there are fewer patient-years at risk.
Table 3.28 Revision rates (95% CI) broken down by time period in which primary was revised, expressed as number of revisions per 1,000 patient-years (PTIRs), for each
recorded reason for first knee revision.
Time Revision
period Patient- Number of revisions per 1,000 patient-years (95% CI) for: Patient- per 1,000
(years) years Peri- years patient-
since at risk Dislocation/ Aseptic prosthetic Implant Implant Mal- Other at risk years for
primary (x1,000) All causes Pain subluxation Infection loosening Lysis fracture fracture1 wear1 Instability alignment indication2 (x1,000) stiffness3
4.81 0.92 0.19 0.95 1.27 0.24 0.15 0.03 0.28 0.69 0.39 1.08 0.33
All cases 4,951.1 4,734
(4.75-4.87) (0.89-0.94) (0.18-0.20) (0.92-0.98) (1.24-1.30) (0.23-0.25) (0.14-0.16) (0.02-0.03) (0.26-0.29) (0.67-0.72) (0.38-0.41) (1.06-1.11) (0.32-0.35)
4.56 0.59 0.38 1.64 0.62 0.11 0.27 0.01 0.19 0.57 0.35 0.69 0.32
<1 920.5 900.3
(4.42-4.70) (0.54-0.64) (0.34-0.42) (1.56-1.73) (0.57-0.67) (0.09-0.13) (0.23-0.30) (0.01-0.02) (0.16-0.22) (0.52-0.62) (0.31-0.39) (0.64-0.74) (0.29-0.36)
6.61 1.58 0.21 1.25 1.63 0.26 0.12 0.03 0.23 0.97 0.59 1.41 0.56
1-3 1,507.6 1,468.3
(6.48-6.74) (1.52-1.65) (0.19-0.23) (1.20-1.31) (1.56-1.69) (0.24-0.29) (0.10-0.14) (0.02-0.04) (0.21-0.26) (0.92-1.02) (0.55-0.63) (1.35-1.47) (0.52-0.60)
4.16 0.88 0.10 0.62 1.31 0.25 0.11 0.02 0.22 0.63 0.35 0.98 0.27
3-5 1,098.4 1,061.5
(4.04-4.28) (0.82-0.93) (0.08-0.12) (0.58-0.67) (1.24-1.38) (0.22-0.28) (0.09-0.13) (0.02-0.03) (0.20-0.25) (0.58-0.67) (0.32-0.39) (0.93-1.04) (0.24-0.30)
3.48 0.53 0.09 0.50 1.20 0.26 0.11 0.03 0.31 0.50 0.27 0.98 0.16
5-7 739.2 704.9
(3.34-3.61) (0.48-0.59) (0.07-0.12) (0.45-0.55) (1.13-1.29) (0.23-0.30) (0.09-0.13) (0.02-0.04) (0.28-0.36) (0.45-0.55) (0.24-0.31) (0.91-1.05) (0.13-0.19)
3.51 0.39 0.11 0.38 1.30 0.32 0.17 0.04 0.47 0.54 0.19 1.13 0.12
7-10 554.4 508.5
(3.36-3.67) (0.34-0.44) (0.09-0.14) (0.33-0.44) (1.20-1.39) (0.28-0.38) (0.14-0.21) (0.02-0.06) (0.41-0.53) (0.48-0.61) (0.16-0.23) (1.04-1.22) (0.09-0.15)
4.43 0.27 0.14 0.36 1.80 0.39 0.25 0.07 0.94 0.70 0.23 1.37 0.09
10-13 131.0 90.6
(4.09-4.81) (0.19-0.37) (0.09-0.22) (0.27-0.48) (1.59-2.05) (0.30-0.51) (0.18-0.35) (0.04-0.13) (0.79-1.12) (0.57-0.86) (0.16-0.33) (1.19-1.59) (0.04-0.18)
Note: 1 The reason implant failure, as reported on in annual reports up to 2013, has been renamed implant wear as this reflects the wearing down of the implant but distinguishes from the implant itself breaking. The latter cause for
revision is now indicated separately as implant fracture. 2 Other indication now includes arthritis and incorrect sizing. Both these reasons were only asked in MDSv1 and so are associated with primaries which took place in the first few
years of the registry with little potential for long term follow-up of the incidence of revision for these specific clinical reasons. 3 This reason was asked in versions MDSv2, v3 and v6 of the clinical assessment forms for joint replacement/
revision surgery and hence, for these reasons, there are fewer patient-years at risk.
135
3.5.2.5 Revisions after primary knee replacement the chance of revision so the figures are unadjusted
surgery by main brands for TKR and UKR probabilities. In addition, simple indicators of the age
profile and proportion of male patients who typically
Tables 3.29 and 3.30 show the Kaplan-Meier receive that implant brand are shown.
estimates of the cumulative percentage probability
of first revision, for any reason, of a primary TKR Table 3.31 shows Kaplan-Meier estimates of the
(Table 3.29) and primary UKR (Table 3.30) by implant cumulative percentage probability of first revision of
brand. We have only included those brands that have a primary TKR or primary UKR by implant brand and
been used in a primary knee procedure in 1,000 or bearing/constraint type for those brands/bearing
more operations. Figures in blue italics indicate those types which were implanted on at least 1,000
time points where fewer than 250 primary knee joint occasions. Again, patient summaries of age and
replacements remain at risk. No attempt has been gender by brand are also given.
made to adjust for other factors that may influence
Table 3.29 Kaplan-Meier estimated cumulative percentage probability of first revision (95% CI) of a primary total knee
replacement by main type of implant brand at the indicated number of years after primary operation1.
Cumulative percentage probability of first revision (95% CI) if time elapsed
Median
since primary operation is:
Number of (IQR) age Percentage
Brand2 knee joints at primary (%) male 1 year 3 years 5 years 7 years 10 years 13 years
All total knee 70 0.40 1.52 2.17 2.68 3.39 4.24
878,191 43%
replacements (63-76) (0.39-0.42) (1.50-1.55) (2.14-2.21) (2.63-2.72) (3.33-3.45) (4.11-4.37)
68 0.69 3.20 3.76 4.45 5.45
ACS 1,996 47%
(62-74) (0.40-1.18) (2.44-4.18) (2.92-4.84) (3.48-5.69) (4.13-7.15)
70 0.50 2.10 2.94 3.63 4.35 4.35
Advance MP 7,941 47%
(63-76) (0.36-0.69) (1.79-2.47) (2.55-3.39) (3.17-4.16) (3.72-5.10) (3.72-5.10)
Advance MP 69 0.08 1.82 3.18 3.18
1,317 15%
Stature (62-75) (0.01-0.57) (1.18-2.81) (2.19-4.61) (2.19-4.61)
72 0.55 2.39 3.19 4.29 5.90 5.90
Advance PS 1,122 45%
(66-77) (0.25-1.23) (1.59-3.58) (2.22-4.58) (3.06-6.00) (4.10-8.47) (4.10-8.47)
71 0.30 1.50 2.11 2.66 3.57 4.91
AGC 65,138 43%
(64-77) (0.26-0.34) (1.40-1.60) (1.99-2.23) (2.53-2.80) (3.38-3.77) (4.48-5.38)
68 0.30 1.00
Attune 9,878 44%
(61-75) (0.20-0.46) (0.66-1.52)
70 0.46 1.84 2.63 2.89 3.58
Columbus 11,143 43%
(64-76) (0.35-0.62) (1.57-2.16) (2.26-3.05) (2.47-3.38) (2.71-4.73)
E-Motion
67 0.69 2.39 3.19 4.08 4.40
Bicondylar 3,111 44%
(61-74) (0.45-1.06) (1.88-3.04) (2.56-3.97) (3.30-5.05) (3.52-5.48)
Knee
Endo Rotating 76 1.51 3.82 5.35 6.08 8.82
1,164 28%
Hinge (68-83) (0.93-2.45) (2.76-5.27) (3.98-7.18) (4.54-8.12) (5.84-13.22)
71 0.39 1.41 1.95 2.39 2.96 3.14
Genesis 2 58,549 42%
(65-77) (0.34-0.45) (1.30-1.52) (1.82-2.10) (2.23-2.57) (2.71-3.22) (2.82-3.50)
Genesis 2 59 0.56 2.26 3.40 4.19 5.51 5.51
8,254 41%
Oxinium (54-64) (0.42-0.75) (1.93-2.64) (2.97-3.90) (3.67-4.78) (4.74-6.40) (4.74-6.40)
†Insall-Burstein 71 0.27 1.64 2.90 3.76 5.36 6.52
2,588 45%
2 (65-77) (0.13-0.57) (1.21-2.22) (2.31-3.65) (3.07-4.60) (4.48-6.41) (5.34-7.94)
71 0.25 1.76 2.68 3.51 4.68 5.79
†Kinemax 10,958 43%
(64-77) (0.17-0.36) (1.53-2.03) (2.39-3.01) (3.17-3.89) (4.27-5.13) (5.24-6.39)
70 0.64 1.79 2.37 2.65 3.09 3.40
†LCS 2,050 41%
(63-76) (0.37-1.10) (1.30-2.47) (1.78-3.14) (2.03-3.47) (2.40-3.97) (2.65-4.34)
70 0.46 1.69 2.59 3.15 3.72
LCS Complete 25,297 44%
(63-76) (0.38-0.55) (1.53-1.87) (2.38-2.82) (2.91-3.42) (3.42-4.05)
70 0.37 1.79 2.66 3.38 4.92 7.74
Maxim 2,191 42%
(63-77) (0.18-0.74) (1.31-2.46) (2.05-3.46) (2.66-4.29) (3.93-6.14) (4.47-13.22)
70 0.28 1.19 1.62 2.26 2.88 2.88
MRK 10,534 42%
(64-77) (0.20-0.41) (0.98-1.45) (1.36-1.94) (1.90-2.68) (2.36-3.51) (2.36-3.51)
70 0.32 1.30 2.22 3.36 3.99 6.35
Natural Knee II 2,840 42%
(64-76) (0.17-0.61) (0.94-1.80) (1.72-2.87) (2.69-4.19) (3.20-4.97) (4.27-9.39)
70 0.38 1.42 2.15 2.80 3.63 4.50
Nexgen 133,343 43%
(63-76) (0.35-0.42) (1.35-1.50) (2.05-2.24) (2.69-2.93) (3.46-3.81) (4.12-4.93)
† Denotes a brand that has been discontinued/withdrawn/not implanted in last three years.
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. Blank cells indicate the number at risk is below ten and therefore
estimates are omitted as they are unreliable. 2 Brands shown have been used in at least 1,000 primary total knee replacement operations. Excludes 7,202 primary
operations where the knee brand was not recorded.
Cumulative percentage probability of first revision (95% CI) if time elapsed

Median
Number of (IQR) age Percentage
Brand2 knee joints at primary (%) male 1 year 3 years 5 years 7 years 10 years 13 years
70 0.39 1.58 2.39 2.87 3.49
NRG 12,843 43%
(64-76) (0.30-0.52) (1.37-1.82) (2.12-2.70) (2.55-3.24) (3.03-4.01)
69 0.70 2.86 4.31 5.23 7.45
Optetrak 2,447 43%
(63-76) (0.44-1.12) (2.26-3.61) (3.55-5.23) (4.35-6.27) (6.02-9.20)
PFC Sigma
70 0.38 1.37 1.88 2.20 2.65 3.07
Bicondylar 296,366 43%

(64-76) (0.35-0.40) (1.32-1.41) (1.82-1.93) (2.14-2.27) (2.57-2.74) (2.93-3.21)
Knee
73 0.38 1.34 1.89 2.33 2.87 3.15
Profix 3,983 44%
(67-78) (0.23-0.63) (1.02-1.75) (1.51-2.38) (1.89-2.87) (2.35-3.51) (2.45-4.04)
61 0.80 2.82 3.23 3.66 4.12 4.68
Profix Oxinium 1,003 43%
(56-67) (0.40-1.59) (1.95-4.06) (2.30-4.54) (2.66-5.04) (3.02-5.61) (3.40-6.43)
71 0.43 2.18 3.47 4.19 4.43
Rotaglide 1,472 39%
(63-77) (0.19-0.96) (1.51-3.15) (2.56-4.70) (3.12-5.60) (3.30-5.95)
70 0.62 3.01 3.93 4.73 6.35 7.90
†Rotaglide + 2,115 44%
(63-76) (0.36-1.06) (2.35-3.84) (3.17-4.87) (3.88-5.76) (5.29-7.61) (6.14-10.15)
71 0.43 1.80 2.58 3.20 3.98 5.44
Scorpio 25,288 42%
(64-77) (0.35-0.52) (1.64-1.97) (2.39-2.79) (2.98-3.43) (3.71-4.26) (4.74-6.25)
70 0.67 1.75 2.37 2.78 3.42 4.12
TC Plus 15,430 45%
(64-76) (0.55-0.81) (1.55-1.97) (2.13-2.63) (2.52-3.07) (3.10-3.78) (3.51-4.82)
70 0.48 1.56 2.15 2.60 3.66
Triathlon 78,098 43%
(63-76) (0.43-0.53) (1.46-1.67) (2.02-2.29) (2.43-2.79) (3.15-4.24)
70 0.32 1.43 2.09 2.49 3.22
Vanguard 52,768 42%
(63-76) (0.28-0.38) (1.31-1.55) (1.92-2.27) (2.26-2.73) (2.50-4.14)
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. Blank cells indicate the number at risk is below ten and therefore
estimates are omitted as they are unreliable. 2 Brands shown have been used in at least 1,000 primary total knee replacement operations. Excludes 7,202 primary
operations where the knee brand was not recorded.
Table 3.30 Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) of a primary
unicompartmental knee replacement by main type of implant brand at the indicated number of years after primary operation1.
Cumulative percentage probability of first revision (95% CI) if time elapsed since
Number Median primary operation is:
of knee (IQR) age at Percentage
Brand2 joints primary (%) males 1 year 3 years 5 years 7 years 10 years 13 years
All unicompartmental 63 1.09 4.34 6.64 8.78 12.25 16.99
97,503 49%
knee replacements (56-70) (1.03-1.16) (4.20-4.49) (6.46-6.83) (8.55-9.01) (11.92-12.60) (16.18-17.85)
Unicondylar
64 2.12 5.80 7.36 9.58 11.65 14.63
AMC/Uniglide 2,848 50%
(57-72) (1.65-2.73) (4.97-6.76) (6.41-8.44) (8.42-10.88) (10.20-13.28) (11.84-18.02)
62 0.93 3.94 5.94 7.56 10.09 12.22
†MG Uni 2,381 54%
(56-70) (0.61-1.40) (3.22-4.80) (5.05-6.97) (6.55-8.71) (8.84-11.50) (10.16-14.67)
64 1.14 4.06 6.10 8.06 11.53 15.73

Oxford Partial Knee 55,447 52%
(57-71) (1.05-1.24) (3.88-4.24) (5.87-6.33) (7.78-8.34) (11.11-11.96) (14.76-16.77)
63 0.37 2.48 3.95 5.29 6.66
*Physica ZUK 10,246 55%
(55-69) (0.26-0.51) (2.13-2.88) (3.45-4.52) (4.60-6.07) (5.54-8.01)
62 2.32 7.73 11.34 14.24 17.09 26.23
†Preservation 1,515 55%
(56-69) (1.67-3.21) (6.49-9.20) (9.83-13.06) (12.55-16.13) (15.19-19.19) (22.38-30.59)
62 0.80 3.57 5.01 5.68
Sigma HP 7,587 57%
(55-69) (0.61-1.04) (3.10-4.10) (4.38-5.74) (4.88-6.61)
Patellofemoral
59 0.77 4.24 7.47 10.08 14.43 20.22
Avon 5,277 22%
(50-68) (0.56-1.05) (3.69-4.87) (6.71-8.32) (9.14-11.10) (13.12-15.87) (17.47-23.33)
59 0.90 6.74 9.61 12.47
FPV 1,587 23%
(51-68) (0.53-1.51) (5.56-8.15) (8.15-11.31) (10.65-14.57)
58 2.08 7.39 12.63 17.84 22.37
Journey PFJ Oxinium 1,572 23%
(50-67) (1.47-2.94) (6.11-8.91) (10.87-14.64) (15.55-20.43) (19.35-25.77)
58 2.46 8.74 12.85 16.90
Sigma HP 1,164 22%
(51-66) (1.69-3.57) (7.09-10.75) (10.64-15.48) (13.60-20.90)
57 0.64 4.41 6.98 9.54
Zimmer PFJ 1,774 22%
(50-65) (0.35-1.19) (3.39-5.75) (5.43-8.95) (7.09-12.78)
* Denotes that this brand is now marketed by Lima
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. Blank cells indicate the number at risk is below ten and
therefore estimates are omitted as they are unreliable. 2 Brands shown have been used in at least 1,000 primary total knee replacement operations. Excludes 141
primary operations where the knee brand was not recorded. 137
Table 3.31 Kaplan-Meier estimates of the cumulative percentage probability of first revision (95% CI) of a total knee
replacement or unicompartmental knee replacement at the indicated number of years after primary operation, by main
implant brands and, within brand, by type of fixation, constraint and bearing sub-group1,3.
Cumulative percentage probability of a first revision (95% CI) if time elapsed since
Brand2 joints primary (%) male 1 year 3 years 5 years 7 years 10 years 13 years
Total knee replacements
AGC
Cement, unconstrained 71 0.26 1.43 2.03 2.56 3.44 4.76
61,690 42%
fixed (64-77) (0.22-0.31) (1.33-1.53) (1.92-2.15) (2.42-2.71) (3.25-3.65) (4.32-5.24)
Uncemented hybrid, 70 1.15 3.28 4.11 4.76 6.22 9.07
2,115 50%
unconstrained fixed (63-76) (0.77-1.70) (2.59-4.14) (3.33-5.06) (3.91-5.80) (5.07-7.62) (5.76-14.15)
Advance MP
7,705 47%
fixed (63-76) (0.35-0.68) (1.73-2.41) (2.41-3.24) (3.04-4.05) (3.61-5.01) (3.61-5.01)
Advance MP Stature
Cement, unconstrained 69 0.08 1.83 3.20 3.20
1,311 14%
fixed (62-75) (0.01-0.57) (1.18-2.82) (2.20-4.64) (2.20-4.64)
Advance PS
Cement, posterior- 72 0.55 2.39 3.19 4.29 5.90 5.90
1,122 45%
stabilised fixed (66-77) (0.25-1.23) (1.59-3.58) (2.22-4.58) (3.06-6.00) (4.10-8.47) (4.10-8.47)
Attune
Cement, unconstrained 68 0.27 1.23
5,490 44%
fixed (61-75) (0.15-0.49) (0.61-2.48)
Cement, posterior- 69 0.49 1.17
2,759 42%
stabilised fixed (61-76) (0.27-0.89) (0.60-2.29)
Columbus
Cement, unconstrained 70 0.43 1.78 2.51 2.79 3.59
9,931 44%
fixed (64-76) (0.31-0.59) (1.50-2.12) (2.14-2.95) (2.37-3.30) (2.60-4.94)

E-Motion Bicondylar Knee
1,079 35%
mobile (61-74) (0.20-1.17) (2.07-4.38) (2.76-5.57) (2.97-6.05) (2.97-6.05)
Uncemented hybrid, 67 0.83 2.03 2.78 3.80 4.16
1,974 49%
unconstrained mobile (61-74) (0.51-1.34) (1.47-2.79) (2.09-3.70) (2.91-4.95) (3.18-5.44)
Endo Rotating Hinge
Cement, bearing/ 76 1.33 3.38 4.95 5.69 8.44
1,066 29%
constraint unknown (68-83) (0.77-2.27) (2.37-4.82) (3.61-6.78) (4.17-7.74) (5.47-12.89)
Genesis 2
42,087 43%
fixed (65-77) (0.28-0.39) (1.16-1.40) (1.61-1.91) (1.98-2.36) (2.35-2.88) (2.43-3.05)
Cement, posterior- 71 0.58 1.74 2.47 2.96 3.85
14,377 39%
stabilised fixed (65-77) (0.46-0.72) (1.52-2.01) (2.17-2.82) (2.59-3.39) (3.12-4.76)
Genesis 2 Oxinium
5,137 41%
fixed (54-64) (0.32-0.72) (1.57-2.41) (2.53-3.63) (2.94-4.18) (4.08-5.99) (4.08-5.99)
2,695 43%
†Insall-Burstein 2
2,394 46%
†Kinemax
10,704 43%
fixed (64-77) (0.17-0.36) (1.54-2.05) (2.40-3.03) (3.19-3.92) (4.28-5.16) (5.22-6.36)
†LCS
1,364 41%
unconstrained mobile (63-76) (0.40-1.37) (1.26-2.74) (1.71-3.39) (1.78-3.49) (1.93-3.71) (2.20-4.15)
LCS Complete
10,775 42%
mobile (64-76) (0.33-0.59) (1.39-1.90) (2.41-3.10) (3.06-3.87) (3.76-4.87)
14,389 46%
MRK
10,358 42%
fixed (64-77) (0.20-0.42) (0.98-1.46) (1.37-1.96) (1.92-2.71) (2.39-3.55) (2.39-3.55)
† Denotes a brand that has been discontinued/withdrawn/not implanted in the last three years.
* Denotes that this brand is now marketed by Lima.
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. 2 Brands shown have been used in at least 1,000 primary knee
replacement operations for that type of fixation and bearing type. 3 Excludes 6,062 joint replacements with no record of main brand.
138
Maxim
1,322 43%
fixed (63-76) (0.04-0.61) (0.94-2.30) (1.47-3.10) (2.21-4.22) (3.21-5.81) (3.71-17.30)
NRG
Cement, unconstrained 70 0.34 1.48 2.36 2.83
7,886 43%
fixed (64-76) (0.23-0.50) (1.22-1.79) (2.00-2.77) (2.41-3.32)
4,728 44%
Natural Knee II
2,695 41%
fixed (64-76) (0.18-0.65) (0.99-1.90) (1.70-2.87) (2.53-4.03) (3.07-4.88) (3.92-6.55)
Nexgen
60,528 43%
fixed (63-76) (0.25-0.34) (1.01-1.20) (1.47-1.73) (1.96-2.29) (2.34-2.81) (2.41-2.98)
62,539 41%
67 1.03 2.83 3.60 5.12 7.13 9.73
Cement, PS mobile 1,110 39%
(60-74) (0.57-1.84) (1.96-4.07) (2.58-5.02) (3.81-6.87) (5.43-9.32) (6.76-13.90)
5,126 55%
2,204 54%
ps fixed (59-73) (0.19-0.75) (1.21-2.39) (1.67-3.09) (2.05-3.70) (2.45-4.40)
Optetrak
1,620 41%

PFC Sigma Bicondylar Knee
191,293 43%
fixed (64-76) (0.32-0.37) (1.19-1.30) (1.64-1.77) (1.93-2.08) (2.27-2.47) (2.63-2.98)
7,842 48%
mobile (58-72) (0.46-0.81) (1.64-2.28) (2.37-3.14) (2.75-3.60) (3.36-4.50) (3.45-4.77)
74,909 41%
65 0.68 2.12 2.79 3.36 4.06 4.19
(59-72) (0.51-0.91) (1.79-2.50) (2.41-3.24) (2.91-3.87) (3.47-4.75) (3.56-4.94)
Cement, bearing/ 71 0.34 1.53 2.29 2.90 3.11 3.11
2,092 47%
constraint unknown (64-77) (0.16-0.71) (1.07-2.18) (1.70-3.08) (2.20-3.81) (2.37-4.09) (2.37-4.09)
monobloc polyethylene 75 0.34 1.30 1.64 1.92 2.17
9,763 41%
tibia (70-79) (0.24-0.48) (1.07-1.58) (1.36-1.99) (1.54-2.39) (1.61-2.92)
1,727 46%
1,052 49%
Profix
2,311 45%
Rotaglide
1,393 39%
mobile (63-77) (0.11-0.80) (1.39-3.03) (2.39-4.53) (2.72-5.06) (2.89-5.44)
†Rotaglide +
1,711 43%
mobile (64-77) (0.24-0.94) (2.13-3.74) (2.84-4.67) (3.34-5.33) (4.51-6.96) (5.04-7.82)
Scorpio
10,765 41%
fixed (64-77) (0.33-0.58) (1.61-2.13) (2.32-2.93) (2.82-3.51) (3.44-4.25) (4.47-7.58)
1,173 43%
mobile (63-75) (0.13-0.91) (1.77-3.63) (2.68-4.89) (3.41-5.87) (4.24-7.29)
6,107 41%
68 0.37 1.49 2.19 2.54 3.27 3.75
(61-76) (0.15-0.88) (0.96-2.30) (1.53-3.14) (1.81-3.56) (2.38-4.48) (2.70-5.20)
4,824 45%
139
TC Plus
7,947 46%
fixed (64-76) (0.58-0.96) (1.61-2.21) (2.23-2.93) (2.62-3.38) (3.08-4.05) (3.42-5.22)
4,954 44%
mobile (64-76) (0.36-0.78) (1.14-1.83) (1.64-2.45) (2.03-2.95) (2.63-3.74)
2,125 40%
Triathlon
61,067 43%
fixed (63-76) (0.38-0.49) (1.37-1.60) (1.87-2.17) (2.30-2.72) (2.79-3.66)
Cement, posterior- 70 0.62 1.73 2.53 2.87
14,666 41%
stabilised fixed (63-76) (0.50-0.77) (1.50-1.99) (2.22-2.87) (2.51-3.28)
Uncemented hybrid, 69 0.83 2.76 3.37 3.90
1,528 48%
unconstrained fixed (62-76) (0.46-1.50) (1.92-3.98) (2.38-4.76) (2.61-5.80)
Vanguard
42,667 42%
fixed (63-76) (0.26-0.37) (1.24-1.51) (1.84-2.21) (2.12-2.62) (2.25-4.62)
6,798 40%
Cement, constrained 69 0.27 1.08 1.66
1,760 38%
condylar (63-76) (0.10-0.71) (0.56-2.06) (0.74-3.68)
Unicondylar knee replacements
AMC/Uniglide
67 0.30 2.98 4.41 6.44 8.49
Unicondylar, fixed 1,379 47%
(60-75) (0.11-0.80) (2.16-4.11) (3.35-5.79) (5.00-8.27) (6.45-11.15)
62 3.84 8.47 10.16 12.42 14.43 15.46

Unicondylar, mobile 1,453 53%
(56-68) (2.96-4.98) (7.12-10.06) (8.67-11.89) (10.71-14.37) (12.47-16.66) (12.81-18.59)
†MG Uni
63 0.90 3.96 5.95 7.55 10.03 12.24
(56-70) (0.59-1.38) (3.24-4.84) (5.06-7.00) (6.53-8.72) (8.78-11.45) (10.12-14.76)
Oxford Partial Knee
64 1.15 4.07 6.13 8.09 11.54 15.77
Unicondylar, mobile 54,376 53%
(57-71) (1.06-1.25) (3.90-4.26) (5.90-6.37) (7.81-8.38) (11.12-11.98) (14.78-16.82)
*Physica ZUK
62 0.37 2.45 3.94 5.31 6.67
(55-69) (0.26-0.52) (2.10-2.85) (3.44-4.52) (4.60-6.11) (5.51-8.07)
†Preservation
63 1.81 6.89 10.21 12.95 15.09 20.04
(57-70) (1.20-2.74) (5.60-8.48) (8.62-12.07) (11.16-15.00) (13.09-17.37) (16.54-24.17)
Sigma HP
62 0.80 3.56 5.01 5.67
(55-69) (0.61-1.04) (3.09-4.09) (4.37-5.73) (4.87-6.60)
Patellofemoral knee replacements
Avon
59 0.77 4.24 7.47 10.07 14.43 20.21
Patello-femoral 5,278 22%
(50-68) (0.56-1.05) (3.69-4.87) (6.70-8.32) (9.14-11.10) (13.12-15.86) (17.47-23.33)
FPV
59 0.90 6.74 9.61 12.47
(51-68) (0.53-1.51) (5.56-8.15) (8.15-11.31) (10.65-14.57)
Journey PFJ Oxinium
58 2.08 7.39 12.63 17.84 22.37
(50-67) (1.47-2.94) (6.11-8.91) (10.87-14.64) (15.55-20.43) (19.35-25.77)
Sigma HP
58 2.46 8.74 12.85 16.90
(51-66) (1.69-3.57) (7.09-10.75) (10.64-15.48) (13.60-20.90)
Zimmer PFJ
57 0.64 4.41 6.97 9.53
(50-65) (0.34-1.19) (3.38-5.74) (5.43-8.94) (7.08-12.77)
3.5.3 Mortality after primary Furthermore, 179 were excluded as they did not have
an NHS number (176) and therefore any record of their
knee surgery death could not be traced or had missing information
This section looks at differences in the likelihood of a on their age (one) or gender (two).
patient dying at increasing lengths of time after primary
This identified a mortality analysis sample of 964,963
operation according to a patient’s gender and age
distinct patients who had a primary operation to replace
at primary. Kaplan-Meier estimates of the cumulative
one or both knees within the NJR and 116,504 of these
percentage probabilities of a patient undergoing knee
patients died in the postoperative time period up to 31
surgery dying in the short term (30 or 90 days after
December 2016.
the primary operation) and in the longer term, up to
13 years after their primary operation are shown. For Table 3.32 shows the Kaplan-Meier estimated
simplicity, we do not take into account whether the cumulative percentage probability of a patient
patient had a first (or further) joint revision after the dying at the indicated number of years after surgery
primary operation when calculating the cumulative stratified by age group and gender. Fewer men than
probability of death (see statistical methodology note III). women have had a primary knee replacement and,
proportionally, more women than men undergo
Of the 975,739 records of a primary knee replacement
surgery above the age of 75.
operation over the period 1 April 2003 to 31
December 2016, 10,597 were bilateral operations Males, particularly in the older age groups, had a higher
in which the patient had both knees replaced on the cumulative percentage probability of dying in the short
same day. Patients identified as having a bilateral or longer term after their primary knee replacement
operation have had the second recorded joint operation than females in the equivalent age group.
excluded from the sample used for mortality analysis.
Table 3.32 Kaplan-Meier estimated cumulative percentage probability (95% CI) of a patient dying at the indicated number
of years after a primary knee joint replacement operation by age group and gender.
Number Cumulative percentage probability of patient death (95% CI) if time elapsed since primary operation is:
Age group of
(years) patients 30 days 90 days 1 year 3 years 5 years 7 years 10 years 13 years
Males
0.04 0.06 0.25 1.01 1.87 3.08 5.29 8.33
<55 29,490
(0.02-0.07) (0.04-0.10) (0.20-0.31) (0.89-1.14) (1.70-2.07) (2.82-3.36) (4.84-5.78) (7.15-9.71)
0.06 0.11 0.39 1.45 2.82 4.78 8.24 12.83
55-59 37,346
(0.04-0.09) (0.08-0.15) (0.33-0.46) (1.32-1.59) (2.62-3.03) (4.49-5.08) (7.76-8.74) (11.67-14.11)
0.07 0.12 0.47 1.95 3.94 6.46 11.31 19.12
60-64 66,256
(0.05-0.09) (0.10-0.15) (0.42-0.53) (1.84-2.07) (3.77-4.12) (6.21-6.71) (10.90-11.74) (18.02-20.28)
0.10 0.18 0.67 2.80 5.88 9.88 17.76 28.37
65-69 82,979
(0.08-0.12) (0.16-0.21) (0.61-0.72) (2.68-2.93) (5.69-6.08) (9.61-10.17) (17.28-18.24) (27.21-29.57)
0.15 0.29 1.11 4.59 9.64 16.22 28.61 44.60
70-74 82,155
(0.13-0.18) (0.26-0.33) (1.04-1.19) (4.43-4.75) (9.40-9.89) (15.88-16.57) (28.06-29.17) (43.36-45.86)
0.30 0.54 1.88 7.24 15.26 25.37 44.85 64.90
75-79 67,283
(0.26-0.34) (0.49-0.60) (1.78-1.99) (7.03-7.46) (14.94-15.59) (24.92-25.82) (44.16-45.54) (63.53-66.27)
0.66 1.09 3.22 12.21 24.65 40.15 63.95 83.18
80-84 36,657
(0.58-0.74) (0.99-1.21) (3.04-3.42) (11.84-12.58) (24.12-25.19) (39.46-40.85) (63.01-64.89) (81.52-84.78)
1.23 2.12 5.85 20.26 39.28 59.09 82.37
85+ 13,681
(1.06-1.43) (1.89-2.38) (5.46-6.27) (19.52-21.02) (38.27-40.31) (57.88-60.30) (80.93-83.76)
Females
0.02 0.04 0.14 0.71 1.40 2.19 4.04 5.90
<55 41,673
(0.01-0.04) (0.03-0.07) (0.11-0.19) (0.62-0.81) (1.27-1.55) (2.01-2.40) (3.70-4.42) (5.09-6.82)
0.02 0.05 0.22 0.88 1.95 3.43 6.18 9.71
55-59 48,806
(0.01-0.04) (0.03-0.07) (0.18-0.27) (0.79-0.98) (1.80-2.10) (3.22-3.65) (5.82-6.57) (8.91-10.58)
0.04 0.08 0.33 1.34 2.77 4.54 8.74 13.49
60-64 76,832
(0.03-0.06) (0.07-0.11) (0.29-0.37) (1.25-1.43) (2.64-2.91) (4.35-4.74) (8.39-9.12) (12.65-14.38)
0.07 0.12 0.43 1.88 3.85 6.39 12.61 20.63
65-69 99,398
(0.05-0.09) (0.10-0.15) (0.39-0.48) (1.79-1.97) (3.71-4.00) (6.19-6.61) (12.22-13.00) (19.72-21.57)
0.10 0.19 0.67 2.79 6.11 10.69 20.91 33.44
70-74 105,165
(0.08-0.12) (0.16-0.22) (0.62-0.72) (2.69-2.91) (5.94-6.29) (10.44-10.94) (20.47-21.36) (32.46-34.45)
0.17 0.32 1.16 4.76 10.40 18.10 34.31 53.22
75-79 95,671
(0.14-0.19) (0.29-0.36) (1.10-1.24) (4.61-4.91) (10.17-10.63) (17.77-18.43) (33.77-34.85) (52.06-54.39)
0.31 0.62 1.96 7.75 16.92 28.76 52.30 73.12
80-84 57,999
(0.27-0.36) (0.56-0.69) (1.84-2.07) (7.52-7.99) (16.56-17.29) (28.27-29.27) (51.54-53.06) (71.73-74.50)
0.66 1.28 3.70 13.93 28.88 47.25 72.47 88.77
85+ 23,572
(0.56-0.77) (1.15-1.44) (3.46-3.95) (13.45-14.43) (28.18-29.59) (46.37-48.15) (71.35-73.58) (86.69-90.66)
0.17 0.31 1.03 4.07 8.55 14.32 25.43 37.82
All cases 964,963
(0.16-0.18) (0.30-0.32) (1.01-1.05) (4.03-4.11) (8.48-8.61) (14.22-14.41) (25.27-25.58) (37.47-38.16)
Note: 1 Estimates in blue italics indicate that fewer than 250 cases remain at risk at the time shown. 2 Excluded 10,597 bilateral operation performed on the same
day and a further 179 with unverifiable age or gender.
Part 3
3.6 Revisions of
knee replacements
3.6.1 Overview of knee revisions surgical episodes difficult. An attempt to link these multiple
stages and/or other information to identify an overall
This section looks at knee revision procedures revision episode is made later in this section.
performed since the registry began on 1 April 2003 up
to the end of December 2016, for all patients with valid An outline of the main revision themes explored in this
patient identifiers. section are as follows: we look at numbers of knee
revision operations recorded in the NJR over time by type
In total there were 60,680 joint revision operations of revision operation (single stage or part of a two-stage
recorded for 48,960 individual patients on 51,241 procedure), the reasons given for knee joint revision by
individual patient-sides. As well as the 24,339 first stage of operation and the survival of the first documented
revisions of primary patient sides reported on earlier in revision of the joint to re-revision. The sensitivity of model
section 3.5 there are 26,842 additional revisions for a survival estimates for re-revision in relation to the choice of
patient-side for which we have no associated primary the starting point of the first revision episode and resulting
operation record. survival times to the next re-revision is explored. Reasons
for re-revision are also presented.
Revisions are classified as single-stage, Stage one or
Stage two of two-stage revisions. Information about Stage An overview of all knee joint revision procedures carried
one and Stage two of two-stage revisions are entered out each year since April 20036 is given in Table 3.33.
into the database separately, whereas Stage one and There were up to a maximum of nine documented
Stage two revisions in practice will be linked when both revision procedures associated with any individual patient-
records have been properly recorded in the NJR. Stage side (discussed later in this section). The increase in the
one procedures have been entered without Stage two, number of operations over time reflects the increasing
and vice versa, making identification of individual revision number of at-risk implants prevailing in the database.
Table 3.33 Numbers of knee joint revision operations carried out each year, by revision operation type. The percentages
of each revision operation type for each year is shown in brackets.
Number of revision joint operations of each revision stage type per
year (% of all revision joint operations in a year)
Year of Stage one of Stage two of Total revision joint
revision surgery Single stage two-stage two-stage operations
2003* 630 (100.0) ---- ---- 630
2004 980 (80.0) 80 (6.5) 165 (13.5) 1,225
2005 1,471 (73.6) 211 (10.6) 316 (15.8) 1,998
2006 1,945 (75.2) 285 (11.0) 358 (13.8) 2,588

2007 2,597 (74.8) 387 (11.1) 490 (14.1) 3,474
2008 3,286 (75.4) 477 (10.9) 596 (13.7) 4,359
2009 3,656 (75.9) 527 (10.9) 631 (13.1) 4,814
2010 4,129 (76.9) 573 (10.7) 670 (12.5) 5,372
2011 4,268 (77.2) 616 (11.1) 647 (11.7) 5,531
2012 4,932 (78.3) 628 (10.0) 742 (11.8) 6,302
2013 4,616 (78.1) 628 (10.6) 664 (11.2) 5,908
2014 4,964 (77.6) 736 (11.5) 694 (10.9) 6,394
2015 4,868 (79.1) 648 (10.5) 637 (10.4) 6,153
2016 4,784 (80.6) 566 (9.5) 582 (9.8) 5,932
All years 47,126 6,362 7,192 60,680
* Incomplete year. Note: MDSv1, in use in 2003, only defined operations as Primary or Revision. All revisions using MDSv1 have been listed as Single stage revisions in
this and subsequent tables.
6 A second procedure had been entered on the same operation date for 139 patient-sides. For these cases, a review of both the components entered for the
surgery and information on all remaining revision surgeries linkable to the patient and side was carried out by one of the orthopaedic surgeons in the NJR
Bristol team. This led to a decision to drop 139 of the duplicated patient side records with the same operation date and to a reclassification of 18 of the
remaining revision operations which had been duplicated originally. In addition, the nine knee joint revision procedures which had been misclassified as a hip
revision procedure in the original raw data set were reclassified as a knee revision after checking records of the type of components used during the surgery.
Table 3.34 below shows the stated reasons for the for approximately two fifths of single stage revision
revision surgery. Note that, as several reasons can operations and pain almost a fifth. Of the two-stage
be stated for the same operation, the reasons are not revision operations, infection is the main reason
mutually exclusive and so the column percentages recorded for revision surgery in over 75% of either
do not add up to 100%. Aseptic loosening accounts Stage one or Stage two procedures.
Table 3.34 Percentage of all revision knee procedures of each stage type with the indicated reason for revision.
Percentage of all revision joint operations of each stage type with
the stated reason for revision
Single stage Stage one of two-stage Stage two of two-stage

Reason for revision (n= 47,122)1 (n= 6,362) (n= 7,187)2
Aseptic loosening 39.5 12.0 11.4

Other indication 19.8 4.1 5.8
Pain 17.8 5.1 4.0
Instability 17.7 4.4 4.1
Implant wear 14.8 3.6 2.0
Lysis 9.9 10.3 6.3
Malalignment 8.1 1.5 1.6
Infection 5.6 83.8 78.7
Dislocation/subluxation 4.3 1.6 1.1
Periprosthetic fracture 3.9 1.5 1.4
Implant fracture 1.3 0.4 0.3
Stiffness3 5.9 n=46,218 2.7 n=6,362 1.9 n=7,187
Notes: 1 Four single stage procedures had a missing entry for the reason for revision and have not been included in the percentage calculations. 2 Five stage two of a
two-stage procedure had a missing entry for the reason for revision and have not been included in the percentage calculations. 3 This reason was not recorded in the
earliest phase of the registry; only in MDS v2, v3 and v6. The number of joints on which the percentage is based is stated beside the percentage figure.
3.6.2 Survival of first recorded estimates of the cumulative percentage chance of
having a re-revision after the first revision (and 95% CI)
knee revision to any subsequent for these two groups are shown in Table 3.35 (b). The
re-revision procedure survival of the first revisions without a linked NJR primary
was much better than those with a linked NJR primary.
For a given patient-side, we have looked at the survival
Those without primaries in the NJR are likely to have
following the first NJR documented revision procedure
been performed before 2003 and so imply a long period
(n=51,241). The majority of first revision procedures
between the original primary or previous revision surgery
(84.7%) were carried out as a single stage revision,
that was not recorded in the NJR and the recorded
however, in the remaining 15.3% of first revisions, the
episode of revision surgery. On the other hand, revisions
process of first revision involved either stage of a two-
linked to primaries in the NJR are likely to represent
stage procedure. We have looked at the time from
shorter times to the first revision of the joint.
the first documented revision procedure (of any type)
to the time at which a second revision procedure was Figure 3.20 (d) and Table 3.35 (c) illustrate this
undertaken. For this purpose, we took an initial Stage difference in early (within the first three years) risk of
one followed, subsequently, by either a Stage one or re-revision for those with primaries in the NJR and
a Stage two as being the same revision episode and those without a recorded primary in the NJR. The
any interim stages were disregarded, looking instead 24,339 with a NJR primary on record have been
for the start of a second revision episode. On this grouped by time interval to the first failure (less than 1
premise, the maximum number of distinct revision year, 1 to 3 years, 3 to 5 years and 5 years or more).
episodes for any patient-side was found to be eight. It is clear that the risk of re-revision is higher for those
primaries which have already failed for the first time
Kaplan-Meier estimates of the cumulative percentage
in the first few years (under 3 years after the primary
probability of having a subsequent revision (re-
replacement) compared to those which were revised
revision) were found. There were 4,518 re-revisions
at later times after the primary and the group without
and, for 7,238 cases, the patient died without having
a known primary on record. The risk of re-revision is
been revised. The censoring date for the remainder
similar for both the first revision after 3 to 5 and 5+
was the end of 2016. Estimates were found for
year groups with a primary procedure recorded in the
two approaches to modelling the start-time to next
NJR and the group of first revisions without a primary
failure: (i) taking the start time as the time of the first
procedure recorded in the NJR. A more in-depth
revision episode and (ii) taking the start time to be the
future investigation of the reasons for first revision
end of the first revision episode. This would make a
and the next re-revision of the joints with linked NJR
difference only for those whose first revision was not
primaries and those without and the patient case mix
a single stage revision, by shortening their follow-up
for each type may yield further insights into why there
time. A plot comparing the cumulative percentage
are the differences described above.
probabilities for the two methods of re-revision is
shown in Figures 3.20 (a) and (b). The rates at 1, 3, In an earlier section of this report, a link between time
5, 7, 10 and 13 years after first revision along with to first revision and the cited reason for revision was
their associated 95% Confidence Intervals are given found (see section 3.5.2.4). It was shown there that
in Table 3.35 (a). The effect on the overall failure rates if a knee joint was revised within the first year after
was negligible as is illustrated in Figures 3.20 (a) and primary surgery, infection was the most likely reason
(b) and shown in Table 3.35 (a). for this, followed by pain, aseptic loosening and then
other reasons for revision. The most common reasons
The first revisions in Figure 3.20 (c) have been divided
given for first revision (of the primary) between one and
into those with a primary recorded in the NJR (n=24,339)
three years were found to be aseptic loosening, pain,
and the remainder (n=26,842). The Kaplan-Meier
other reasons and instability respectively.
Future work will explore the relationship between (i) the end of 2016, 66.3% of all first time records of
the time to first revision and the subsequent time to revision surgery for a joint could be linked to a NJR
re-revision and (ii) the reason for the first revision and primary operation (see Tables 3.37 (a) and (b)). This is
the resulting time to re-revision. a further indication that the first revisions with a linked
primary in the NJR could be failing sooner than the
The number of recorded first revisions in the NJR group of revisions without a linkable primary within the
with an associated NJR primary record has increased NJR dataset.
each year since the start of the registry in 2003. By
Figure 3.20 (a)

Kaplan-Meier estimate of the cumulative percentage probability of a knee re-revision, based on time from
the start date of the first revision episode. The shaded area indicate point-wise 95% CIs.
Cumulative percentage probability of re−revision (%)
20
15

10
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Numbers at risk
51,241 44,307 37,818 31,623 26,220 20,786 16,311 12,129 8,637 5,680 3,448 1,954 902 282
Figure 3.20 (b)
Kaplan-Meier estimate of the cumulative percentage probability of a knee re-revision, based on time from
the last date of the first revision episode*. The shaded area indicate point-wise 95% CIs.
20
15
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Years since last date associated with first revision

Numbers at risk
51,241 44,112 37,606 31,450 26,074 20,645 16,203 12,019 8,554 5,615 3,413 1,937 893 282
*First documented revision recorded in the NJR.
Figure 3.20 (c)

Kaplan-Meier estimates of the cumulative percentage probability of a knee re-revision, shown for those
with documented primaries in NJR* and the remainder. Shaded areas are point-wise 95% CIs for the
rate estimates.
20

15
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Number at risk
Primary not in NJR 26,842 24,136 21,485 18,667 16,099 13,408 10,973 8,624 6,536 4,561 2,919 1,736 847 278
Primary in NJR 24,399 20,171 16,333 12,956 10,121 7,378 5,338 3,505 2,101 1,119 529 218 55 4
Figure 3.20 (d)
Kaplan-Meier estimates of the cumulative percentage probability of a knee re-revision up to three
years from the first revision*. Those with recorded primaries in the NJR are shown separately from the
remainder and have been split into those that had their first revision within 1 year, 1 to 3, 3 to 5 years or
more than 5 years after the initial primary.
16
12
0
0 1 2 3
Primary not in NJR 26,842 24,136 21,485 18,667

1st revision <1 year 4,342 3,595 2,996 2,495
1st revision 1−3 years 10,262 8,871 7,453 6,149
1st revision 3−5 years 4,656 3,933 3,282 2,632
1st revision 5+ years 5,139 3,772 2,602 1,680
Table 3.35 (a) Kaplan-Meier estimates of cumulative percentage probability of knee re-revision following the first
revision using different start points for time at risk of re-revision.

Number Cumulative percentage probability of a re-revision (95% CI) at time shown if time elapsed since
of revised first revision is:
Time point from which joints at
time to re-revision was risk of re-
measured revision 1 year 3 years 5 years 7 years 10 years 13 years
(i) At start of first 2.79 7.29 9.82 11.66 14.44 16.59
51,241
revision episode (2.65-2.95) (7.05-7.54) (9.53-10.13) (11.31-12.02) (13.94-14.96) (15.55-17.70)
(ii) End of first revision 2.89 7.35 9.85 11.67 14.45 16.58
51,241
episode (2.74-3.04) (7.11-7.60) (9.55-10.15) (11.32-12.03) (13.95-14.97) (15.53-17.68)
Table 3.35 (b) Kaplan-Meier estimates of cumulative percentage probability of knee re-revision following the first
revision broken down by whether a primary is on record in the NJR or not.

Number Cumulative percentage probability of a re-revision (95% CI) at time shown if time elapsed since
of first first revision is*:
revised
joints at
Revised risk of
patient-sides re-revision 1 year 3 years 5 years 7 years 10 years 13 years
Primary not recorded in 2.15 5.65 7.86 9.58 12.53 14.85
26,842
the NJR (1.98-2.33) (5.36-5.95) (7.50-8.23) (9.17-10.02) (11.94-13.15) (13.72-16.07)
Primary recorded in 3.53 9.29 12.33 14.40 16.71 17.09
24,399
the NJR (3.30-3.78) (8.89-9.71) (11.83-12.85) (13.80-15.02) (15.79-17.68) (16.04-18.21)
*Estimates in blue italics are based on the number at risk falling below 250 patient-sides (see methodological notes in earlier sections). The number at risk for the
year 13 estimate for those with primary recorded in the NJR is only five.
Table 3.35 (c) Kaplan-Meier estimates of the cumulative percentage probability of knee re-revision following the first
revision when the group of patient-sides with a primary record in the NJR are stratified by the time intervals in which the
first revision took place after the primary operation.
Cumulative percentage probability of a re-revision (95% CI)
Number of first revised at time shown if time elapsed since first revision is:

joints at risk of
Revised patient-sides re-revision 1 year 3 years
Primary not in the NJR 26,842 2.15 (1.98-2.33) 5.65 (5.36-5.95)
Primary in the NJR where the first
revision took place:
<1 year after primary 4,342 7.59 (6.82-8.44) 14.99 (13.88-16.17)
1-3 years after primary 10,262 3.12 (2.79-3.49) 9.55 (8.95-10.20)
3-5 years after primary 4,656 2.22 (1.82-2.70) 6.73 (5.96-7.59)
5+ years after primary* 5,139 1.96 (1.60-2.41) 5.26 (4.53-6.09)
* The maximum of this interval was 13.5 years.
3.6.3 Reason for knee re-revision each type of reason for revision when the revision was
(i) the first recorded revision in the NJR, (ii) the first
Table 3.36 shows a breakdown of the stated reasons revision and the implant was not subsequently revised,
for the first revision and for any subsequent revision. (iii) the first revision and the implant was subsequently
The reasons are not mutually exclusive. The four re-revised and (iv) the re-revision of the first revision.
columns show the number of joints which indicated
Table 3.36 Reasons given for first knee revision and re-revision.
Reasons for the first recorded revision for those
(i) who were:
Number of (iv)
cases for each (ii) (iii) Number of cases for
given reason for Not subsequently Subsequently each given reason for
first revision re-revised re-revised re-revision
Reason for revision n=51,2331 n=46,7172 n=4,5163 n=4,518
Aseptic loosening 18,393 17,013 1,380 1,200

Pain 8,245 7,430 815 527
Infection 8,370 7,241 1,129 1,550
Instability 7,844 7,149 695 783
Implant wear 6,973 6,496 477 237
Lysis 5,176 4,817 359 283
Malalignment 3,684 3,384 300 253
Dislocation/subluxation 1,905 1,702 203 205
Periprosthetic fracture 1,806 1,681 125 137
Implant fracture 563 517 46 54
Stiffness 4
2,641 n=50,345
2,402 n=45,965
239 n=4,380
285 n=4,380
Other indication 9,291 8,745 546 442
Note: 1 Reasons for revision for eight first (recorded) revisions were missing. 2 Reasons for first revision for six joints not re-revised were missing. 3 Reasons for
first revision for two subsequently re-revised joints were missing. 4 Stiffness as a reason for revision was not recorded in MSDv1. The denominator number of
joints on which stiffness is based is stated beside the total figure.
Table 3.37 Temporal changes in first knee revisions reported in the NJR and associated indications.
(a) Number of first knee revisions by year of surgery and proportions with an associated knee primary in the NJR.
Number of first revisions (%) with the
Year of first revision in the NJR* Number of (first) revisions associated primary in the NJR
2003 622 11 (1.8)
2004 1,172 83 (7.1)
2005 1,842 275 (14.9)
2006 2,343 499 (21.3)

2007 3,108 852 (27.4)
2008 3,800 1,358 (35.7)
2009 4,155 1,767 (42.5)
2010 4,579 2,156 (47.1)
2011 4,641 2,290 (49.3)
2012 5,245 2,899 (55.3)
2013 4,842 2,772 (57.2)
2014 5,151 3,122 (60.6)
2015 4,938 3,130 (63.4)
2016 4,803 3,185 (66.3)
Total 51,241 24,399 (47.6)
(b) Numbers of first recorded knee revisions by stage and whether or not primary was in the NJR.
Single stage First documented stage of two-stage

Primary not in the Primary in the Primary not in the Primary in the
Year of (first) revision NJR total per year NJR total per year NJR total per year NJR total per year
2003 611 11 0 0
2004 905 62 184 21

2005 1,239 196 328 79
2006 1,497 379 347 120
2007 1,856 638 400 214
2008 2,046 1,056 396 302
2009 1,999 1,444 389 323
2010 2,070 1,758 353 398
2011 2,057 1,862 294 428
2012 2,080 2,441 266 458
2013 1,829 2,340 241 432
2014 1,812 2,624 217 498
2015 1,643 2,691 165 439
2016 1,473 2,792 145 393
All years 23,117 20,294 3,725 4,105
3.6.4 Conclusions Unicondylar replacements are far more common (8.7%

of the total) than patellofemoral replacements (1.2% of
The current year’s analysis demonstrates an extension the total).
of the trends observed in previous years. In general,
total knee replacements have excellent implant In terms of choice of bearing/constraint in cemented
survivorship out to 13 years whilst unicompartmental, TKR surgery and the cumulative chance of revision
and patellofemoral knee replacements have higher of the implant, the majority of these perform similarly
implant revision rates. However, implant survivorship over time (Figures 3.19 (a) and Table 3.25 (a)). The
is not the only metric of success and patients and best 13-year survivorship is observed in the cemented
surgeons need to consider patient demographics, unconstrained (cruciate retaining) fixed bearings
disease pattern and severity, pain relief, function, compared to the unconstrained mobile, posterior
participation in society and post-operative mortality stabilized fixed and mobile and constrained condylar
when making choices about whether to undergo implants although the numbers are small at the
surgical intervention and the type of surgical intervention longest term follow-up so estimates are less reliable.
that is appropriate for them. Promising survivorship results are seen in the monobloc
polyethylene tibia implants but the numbers at risk are
Cementation of the primary prosthesis in total knee small beyond the medium term. The risk of revision is
replacements continues to be the most commonly used higher in the uncemented (particularly for posterior-
method of fixation (84.9%) in total knee replacement stabilised designs) and hybrid fixation groups.
surgery. Conversely uncemented fixation for primary
TKR continues to decline in use making up only 2.0% Unicondylar fixed and mobile constraints again
of all surgeries last year. UKR (medial and lateral perform similarly overall but, compared to any TKR
unicondylar and patellofemoral knee replacement) constraint choice, fare worse in terms of the need for
still represents one in ten of all primary knee surgeries revision surgery. The use of a patellofemoral implant
(10.3%) and this proportion overall has remained incurs the highest risk of revision over all surgical
relatively consistent over the 2003 to 2016 period. choices. Although patients that receive this type of
implant are typically younger (by about ten years) and The most common clinical reasons for revision cited
it has therefore previously been assumed that the for TKR were aseptic loosening, pain, infection and
difference in revision rates may be because patients other indication (excluding dislocation/subluxation,
receiving these implants are more active than those lysis, periprosthetic fracture, implant fracture, implant
receiving a TKR. We have this year analysed revision wear, instability, malalignment and stiffness), each of
rates by age group in each gender (Table 3.25 (b)). which account for approximately one revision per 1,000
It can be seen that the revision rates are higher for patient-years or more across all cases. However, for
unicondylar knee replacements and higher still for UKRs, the incidence rates of revision for pain, aseptic
patellofemoral replacements across all age groups in loosening and other indication each account for around
males and females. four revisions per 1,000 patient-years. The indicated
reasons for revision of a primary patellofemoral knee
The volume of procedures of different types performed resemble those of unicondylar indications for revision
by consultants shows interesting trends. 34% of total surgery, but PTIRs are even higher than those reported
knee replacement consultants in 2016 performed for revision of a unicondylar implant with pain and other
25 or less during the year, accounting for only 7% indication having PTIRs of approximately 6 and 10
of total replacements. This compares to unicondylar revisions per 1,000 patient-years respectively.
knee replacement consultants where 82% were
performing 20 or less a year (accounting for 37.4% of In the first year after primary surgery, revision due to
all unicondylar replacements) and 25% of consultants infection has the highest PTIR. Between one and three
performed only one or two cases in the year. This trend years post primary surgery, aseptic loosening and
was similar for patellofemoral replacements where pain become more prevalent as reasons for revision
95% of consultants were performing ten or less a year surgery and in the longer term, aseptic loosening is the
accounting for 76% of all patellofemoral replacements. dominant reason for revision.
54% of patellofemoral consultants performed only one
or two cases in the year. The effect of the volume of The risk of death remains higher in men than women
procedures on the risk of revision is not yet clear and in the same age group in the short, medium and long
requires further exploration. term after primary knee surgery, and the risk of dying
increases the older the patient is when they initially
Unlike the hip surgery findings in the last section, undergo primary surgery. The risk of death within 90
gender differences in the chance of needing revision days of surgery in primary knee replacement is 0.31%,
surgery following total knee replacement are only with the death rate rising to 1.03% at 1 year, 8.55% at
small, with males at slightly higher risk than females 5 years, 25.4% at 10 years and 37.8% at 13 years.
for all ages. However, as also seen in hip replacement
surgery, younger patients are at far higher risk of
requiring first knee revision surgery than patients
belonging to older age groups.
Part 3
3.7 Outcomes
after primary ankle
replacement
3.7.1 Overview of primary (88.7%) used uncemented and 439 (11.3%) used
cemented fixation methods for the implant. There were
ankle surgery three (0.8%) joints where the fixation type was hybrid.
This section looks at revision and mortality for
A total of 229 consultants carried out these primary
all primary ankle operations performed up to 31
procedures; 106 (44.3%) of them entered ten or more
December 2016. There were 3,899 primary ankle
procedures over the seven-year period of data capture.
operations in total (see Tables 3.1 and 3.2), including
The maximum number of procedures for any consultant
four bilateral operations (both sides done on the same
was 239 over the same time period. Similarly, the total
date). Although ankles were entered routinely from
number of units involved was 269; 89 (38%) of which
2010, 14 primary operations have been entered that
carried out ten or more over the seven-year time period.
had been carried out before this date.
The maximum number of procedures carried out by any
The median age at primary surgery was 68 years (IQR unit was 234. Table 3.38 (a) shows how the caseload
61-74 years), with an overall range of 17 to 92 years. of ankle surgery for units and consultants has changed
More procedures were performed in men (59.1%) during the seven-year period.
than women. Of the 3,899 primary procedures, 3,457
Table 3.38 (a) Descriptive statistics of ankle procedures performed by consultant and unit by year of surgery.
Year of surgery
Number of primary replacements
during each year 2010* 2011 2012 2013 2014 2015 2016
Number of procedures in year 416 521 581 546 543 602 690
Number of units providing primary
111 127 145 132 137 140 134
replacements each year
Mean number of primary replacements

3.7 4.1 4.0 4.1 4.0 4.3 5.1
per unit
Median (IQR) number of any primary
2 (1-4) 2 (1-5) 2 (1-4) 2 (1-5) 2 (1-4) 2 (1-5) 2 (1-6)
replacements per unit
Number of units who entered
10 7 10 9 10 10 16
>10 procedures each year
Number of units who entered
3 3 3 3 4 5 5
Number of consultants providing primary

114 126 144 131 126 139 132
replacements during each year
Mean number of primary replacements
3.6 4.1 4.0 4.2 4.3 4.3 5.2
per consultant
2 (1-4) 3 (2-5) 2 (1-5) 3 (1-5) 3 (2-5) 2 (1-6) 3 (1-7.5)
replacements per consultant
Number of consultants who entered
9 10 10 11 8 13 15
Number of consultants who entered
2 2 2 2 2 4 5
*Includes 14 operation dates prior to 2010.
Figure 3.21 further illustrates how a large proportion The data for units resembles that of the consultants,
of ankle arthroplasty procedures are performed by with 12% of units performing 315 procedures which
a minority of consultants or units over the last three accounts for 46% of all ankle procedures in 2016. The
years. For example, in the last year (2016) 15 of consultant data and unit data are very similar as 87
132 (11%) consultants submitting ankle arthroplasty units (65%) have just one consultant performing ankle
procedures performed 286 cases. This accounts for procedures. There were only four units (3%) where
41% of all ankle arthroplasty procedures that year. three or more consultants operated in 2016.
Figure 3.21
Exploring unit and consultant frequency of primary ankle surgery, and total caseload in the NJR between
2014 and 2016. Exploring unit and consultant frequency of primary ankle surgery, and
total caseload in the NJR between 2014 and 2016
Distribution of unit annual Distribution of consultant annual

frequency of ankle procedures frequency of ankle procedures
2014 2015 2016 2014 2015 2016

Total caseload of ankle surgery by Total caseload of ankle surgery by
2014 2015 2016 2014 2015 2016
1 2 3 4 5−10 11−20 21−30 >30 1 2 3 4 5−10 11−20 21−30 >30
Table 3.38 (b) shows an overall breakdown of brands overall was Mobility, which was used in just under half
used and further breakdowns by year of primary of the procedures overall but whose usage since 2012
operation. Please note that 14 procedures had dates declined and in June 2014 was withdrawn from the
of operation before 2010 (one in 2006, four in 2008 market. In 2016, the most common brand used was
and nine in 2009) and these have been combined with the Infinity (30.1%), followed by the Box (18%) and the
those performed in 2010. The most common brand Zenith (14.8%).
Table 3.38 (b) Numbers of primary ankle replacements by ankle brand.
Number (%) of each brand, for each year of operation
Brand Number (%) ≤2010* 2011 2012 2013 2014 2015 2016
Mobility 1,125 (28.9) 258 (62.0) 295 (56.6) 285 (49.1) 200 (36.6) 87 (16.0) 0 (0.0) 0 (0.0)
Zenith 853 (21.9) 78 (18.8) 109 (20.9) 126 (21.7) 133 (24.4) 150 (27.6) 155 (25.7) 102 (14.8)
Box 486 (12.5) 23 (5.5) 29 (5.6) 45 (7.7) 50 (9.2) 84 (15.5) 131 (21.8) 124 (18.0)
Salto 289 (7.4) 23 (5.5) 29 (5.6) 39 (6.7) 44 (8.1) 56 (10.3) 54 (9.0) 44 (6.4)
Hintegra 258 (6.6) 15 (3.6) 18 (3.5) 35 (6.0) 63 (11.5) 45 (8.3) 53 (8.8) 29 (4.2)
Star 328 (8.4) 16 (3.8) 29 (5.6) 31 (5.3) 35 (6.4) 60 (11.0) 81 (13.5) 76 (11.0)
Rebalance 53 (1.4) 0 (0.0) 4 (0.8) 13 (2.2) 13 (2.4) 6 (1.1) 4 (0.7) 13 (1.9)
Inbone 104 (2.7) 0 (0.0) 0 (0.0) 2 (0.3) 4 (0.7) 22 (4.1) 20 (3.3) 56 (8.1)
Infinity 330 (8.5) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 28 (5.2) 94 (15.6) 208 (30.1)
AKILE 10 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 4 (0.7) 6 (0.9)
TARIC 1 (0.0) 0 (0.0) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Not known 62 (1.6) 3 (0.7) 8 (1.5) 4 (0.7) 4 (0.7) 5 (0.9) 6 (1.0) 32 (4.6)
Total 3,899 (100.0) 416 (100.0) 521 (100.0) 581 (100.0) 546 (100.0) 543 (100.0) 602 (100.0) 690 (100.0)
*Includes 14 operation dates prior to 2010.
3.7.2 Revisions after primary were: 0.13 (95% CI 0.05-0.31) at 90 days; 0.81 (95%
CI 0.56-1.17) at 1 year; 2.74 (95% CI 2.21-3.39) at 2
ankle surgery years; 3.93 (95% CI 3.26-4.73) at 3 years; 4.9 (95%
CI 4.11-5.84) at 4 years; 6.61 (95% CI 5.57-7.84) at 5
The definition of revision accepted by the British
years; and 7.71 (95% CI 5.94-8.47) at 6 years.
Orthopaedic Foot and Ankle Society (BOFAS) is the
removal or exchange of any component of the ankle BOFAS believes that the small number of revisions
replacement, except in the case of an incidental reported may indicate under-reporting of the revision
exchange of a polythene liner in a mobile bearing implant. procedures as these figures are lower than published
In situations where this definition is met, the surgeon data in the literature. BOFAS and the NJR encourage
should complete a NJR A2 MDS form. Only 153 (3.9%) surgeons to complete A2 MDS forms where relevant
of the 3,899 procedures had a NJR A2 MDS form and wishes to remind surgeons and hospitals that
completed to indicate revision before the end of 2016. this is a mandated requirement and that all revisions,
The first revisions shown here include 24 conversions to conversion of an ankle replacement to an arthrodesis,
arthrodesis but no amputations have been recorded. and amputations require the completion of a NJR A2
MDS form.
The estimated cumulative percentage probabilities of
(first) revision overall (using Kaplan-Meier estimation)
Table 3.39 Indications for the 153 (first) revisions following primary ankle replacement. Note: these are not
mutually exclusive.
Indication Number
Infection High suspicion (e.g. pus or confirmed micro) 9
Low suspicion (awaiting micro/histology) 27
Aseptic loosening* Tibial component 49
Talar component 47
Lysis** Tibia 12
Talus 15
Malalignement 17
Implant fracture*** Tibial component 3
Talar component 4
Implant fracture Meniscal component 3
Wear of polyethylene component 11
Meniscal insert dislocation 3
Component migration/dissociation 9
Pain (undiagnosed) 47
Stiffness 20
Soft tissue impingement 13
Other indications for revision 21
*29 patients had aseptic loosening of both tibial and talar component. ** Six patients had lysis of both tibial and talar component. *** Two patients had implant
158 fracture of both tibial and talar component.
3.7.3 Mortality after primary 0.24) at 90 days; 0.64 (95% CI 0.42-0.97) at 1 year; 1.76
(95% CI 1.35-2.29) at 2 years; 3.07 (95% CI 2.48-3.8) at
ankle replacement 3 years; 4.92 (95% CI 4.10-5.89); 5.97 (95% CI 5-7.12)
Our analysis excluded two procedures where the NHS at 5 years; and 9.14 (95% CI 7.59-10.99) at 6 years.
number was untraceable (and hence the age could not Estimates at five and six years were unreliable as too few
be validated) plus the second of each of the four bilateral patients remained at risk.
procedures. Among the remaining 3,893, a total of 155
Table 3.40 shows the estimated cumulative percentage
patients had died before the end of 2016.
probability of death at different times after surgery by
The estimated cumulative percentage mortality (based gender and age at primary.
on Kaplan-Meier estimates) were: 0.08 (95% CI 0.03-
Table 3.40 Kaplan-Meier estimates of the cumulative percentage mortality (95% CI), by gender and age at 90 days,
and 1, 2, 3, 4 and 5 years after primary ankle replacement. Figures in blue italics signify time points where fewer than
250 patients remain at risk.
Cumulative percentage probability of patient death (95% CI) if time elapsed

Age at primary Number of
Gender (years) patients 90 days 1 year 2 years 3 years 4 years 5 years
1.12 1.81 3.07 3.07
Male <65 750 0* 0*
(0.50-2.48) (0.94-3.46) (1.76-5.31) (1.76-5.31)
0.20 1.00 2.31 4.19 7.34 9.52
65+ 1,553
(0.06-0.61) (0.60-1.69) (1.61-3.32) (3.14-5.59) (5.76-9.33) (7.55-11.97)
0.19 0.90 1.46 1.77 1.77
Female <65 621 0*
(0.03-1.35) (0.34-2.38) (0.65-3.25) (0.84-3.71) (0.84-3.71)
0.82 1.98 3.49 5.27 5.87
65+ 969 0*
(0.39-1.72) (1.20-3.28) (2.34-5.18) (3.71-7.45) (4.16-8.24)
*No events recorded after surgery.
3.7.4 Conclusions submitted less than ten primary procedures in the

seven years the NJR has been capturing data. Since
The collection of data relating to ankle primary the withdrawal of the market leading brand (Mobility)
operations only began in 2010 and hence total in 2014, the use of other brands such as Zenith and
number of primaries remain small and numbers of Box has increased accordingly. The fixed bearing
first revisions even smaller, although we believe Infinity implant has gained rapid popularity over the
that there is under-reporting of revision procedures, last three years and is now the market leader. The
making outcome analysis difficult. A total of 54% cumulative percentage probability of death following
of consultant surgeons and 62% of units have primary ankle surgery is very low.
Part 3
3.8 Outcomes after
primary shoulder
replacement
3.8.1 Overview of primary shoulder had been entered, (ii) designated as resurfacings or
reverse polarity total prosthetic replacements but for
replacement surgery which a uni-polar or a bi-polar head had been entered,
The registry has recorded shoulder replacements since (iii) designated as total prosthetic replacements,
1 April 2012. This section contains an overview of the hemiarthroplasty or reverse polarity total prosthetic
(linked) primary shoulder replacements performed up to replacement but in which information on a humeral
31 December 2016 and documents the first revision and resurfacing head component had been entered and (iv)
mortality for these primaries. designated as hemi-arthroplasty but glenoid component(s)
had been entered.
A total of 23,608 linked primary replacements were
available for analysis for a total of 22,313 patients. Of The proportion of resurfacings (both total and hemi-
these patients, 1,295 had documented replacements arthroplasty) has continued to decline with time and the
on both left and right sides, 20 of which were bilateral proportion of reverse polarity total replacements has
operations (left and right on the same day), see Table 3.2 increased again this year. For the 20,123 non-resurfacing
in section 3.2, summary of data sources and linkages. implants, the distinction has been made this year
between stemmed and stemless humeral components.
Due to the rapid expansion of new shoulder arthroplasty Table 3.41 demonstrates this breakdown. Stemmed is
designs, the classification system for shoulder defined as any part of the humeral component entering
arthroplasty will be updated this year to allow for the the diaphysis, while Stemless is defined as being
future accurate data collection of what is a rapidly completely confined to the metaphysis with no part of
changing product area. the stem entering the diaphysis. 19,060 of these were
classified directly according to their stated stem brand.
Table 3.41 demonstrates that the number of primary A further 490 had no stem brand entered but were able
shoulder replacements has continued to increase to be classified on the basis of the catalogue numbers
year by year and gives a breakdown by the stated of the humeral proximal component; the remaining 573
7
type of replacement . could not be further sub-divided.
A number of cases (401) had discrepancies between The majority of the replacements were performed on
the stated type of procedure and the components women (men 29.0%; women 71.0%). The median age
entered and these are shown under the final row headed at the primary operation was 73 years (IQR 67-79 years)
Uncertain. This final column comprises cases that were (i) overall, with a range of 17-99 years.
designated as resurfacings but information about either a
stem component or a metaphyseal proximal component
7 Provisional breakdown using the primary procedure as stated, without further validation by the actual components used.
Table 3.41 Numbers of all primary shoulder replacements (elective and acute trauma) by year and percentages of
each type.
Year of primary operation:

All years 2012* 2013 2014 2015 2016
All cases 23,608 (100%) 2,544 (100%) 4,345 (100%) 5,224 (100%) 5,551 (100%) 5,944 (100%)
3,050 384 696 719 661 590
Humeral hemiarthroplasty
(12.9%) (15.1%) (16.0%) (13.8%) (11.9%) (9.9%)
2,490 338 575 579 537 461
Stemmed
(10.5%) (13.3%) (13.2%) (11.1%) (9.7%) (7.7%)
485 31 95 124 112 123
Stemless
(2.1%) (1.2%) (2.2%) (2.4%) (2.0%) (2.1%)
75 15 26 16 12 6
Uncertain
(0.3%) (0.6%) (0.6%) (0.3%) (0.2%) (0.1%)
Resurfacing humeral 2,251 461 565 525 363 337
hemiarthroplasty (9.5%) (18.1%) (13.0%) (10.1%) (6.5%) (5.7%)
Total conventional shoulder 7,105 678 1,240 1,597 1,778 1,812
arthroplasty (30.1%) (26.7%) (28.5%) (30.6%) (32.0%) (30.5%)
5,009 503 915 1,148 1,250 1,193

Stemmed
(21.2%) (19.8%) (21.1%) (22.0%) (22.5%) (20.1%)
1,732 107 230 377 453 565
Stemless
(7.3%) (4.2%) (5.3%) (7.2%) (8.2%) (9.5%)
364 68 95 72 75 54
Uncertain
(1.5%) (2.7%) (2.2%) (1.4%) (1.4%) (0.9%)
Resurfacing total shoulder 833 149 219 189 149 127
arthroplasty (3.5%) (5.9%) (5.0%) (3.6%) (2.7%) (2.1%)
Reverse polarity total 9,968 806 1,531 2,100 2,516 3,015
shoulder arthroplasty (42.2%) (31.7%) (35.2%) (40.2%) (45.3%) (50.7%)
9,748 772 1,467 2,056 2,471 2,982
Stemmed
(41.3%) (30.3%) (33.8%) (39.4%) (44.5%) (50.2%)
86 15 20 13 21 17
Stemless
(0.4%) (0.6%) (0.5%) (0.2%) (0.4%) (0.3%)
134 19 44 31 24 16
Uncertain
(0.6%) (0.7%) (1.0%) (0.6%) (0.4%) (0.3%)
401 66 94 94 84 63
Uncertain
(1.7%) (2.6%) (2.2%) (1.8%) (1.5%) (1.1%)
*Includes 36 in the registry with primary operation dates before 2012.
The primary shoulder replacements over the last the numbers of units and consultants for each year,
three years were undertaken by 613 consultant together with their number of primaries, is shown in
surgeons working across 371 units. A breakdown of Table 3.42 below.
Table 3.42 Numbers of units and consultant surgeons providing primary shoulder replacements over the last three
years, 2014-2016.

Number of
Number of Median (IQR) consultants Median (IQR)
Number units providing number of primary providing number of primary
of primary the primary replacements the primary replacements per
Year of primary replacements replacements per unit replacements consultant
2014 5,224 338 9 (4-20) 453 8 (3-17)
2015 5,551 343 10 (4-22) 479 8 (3-17)
2016 5,944 338 12 (5-23) 476 9 (4-18)
Table 3.43 details the indications for the primary Table 3.45 lists the main stem brands used in the
operation, for the cases overall and with further sub- non-resurfacing primary procedures. The table shows
division by type of procedure. the totals in the registry since April 2012 as well as
the numbers within the last twelve months (2016).
Acute trauma accounts for 2,038 cases. These have The latter are further sub-divided into acute trauma
been separated from the remaining 21,570 elective and elective cases. The numbers of elective cases are
cases. Please note, 76 of the 2,038 acute trauma further divided into the types of implant. Not all cases
cases had another reason(s) stated in addition to had the stem information recorded and a number
acute trauma; the most common reasons being had multiple stems entered (shown in the bottom
osteoarthritis (29) and trauma sequelae (21). row of the table). The total numbers of stemmed and
stemless implants in this table are fewer than reported
The reasons given for the elective cases are
in the earlier table (Table 3.43) because some of
documented in Table 3.43. The reasons entered were
the non-resurfacing implants had no humeral stem
not all mutually exclusive with some surgeons entering
entered (hence no brand) but could be classified into
more than one indication. Amongst these 21,570
stemmed or stemless by further inspection of the
cases, 1,243 (5.76%) had two or more reasons stated,
catalogue details available for their humeral proximal
the most common combinations included osteoarthritis
component.
together with cuff tear arthropathy (522).
Finally, Table 3.46 shows a similar table for the
Table 3.44 summarises the age and gender
resurfacing brands used in resurfacing shoulder
distributions of the acute trauma and elective cases
replacements. Note that Tables 3.45 and 3.46
according to their main primary procedure. Where
exclude the 401 cases where the type of procedure
numbers permit (elective cases only), the non-
was uncertain.
resurfacings have been further divided into stemmed
or stemless implants.
Figures 3.22 (a) to (e) illustrate the distributions by

gender and age groups of the elective patients,
according to the primary patient procedure.
164
Table 3.43 Reasons for main types of primary shoulder replacements.
Acute trauma Elective

Number (%)* for each reason (amongst elective cases only):
Other
Number of Number Cuff tear Trauma inflammatory Avascular
cases of cases Osteoarthritis arthropathy sequelae arthropathy necrosis Other cause**
All cases 2,038 21,570 13,442 (62.3%) 5,877 (27.3%) 1,393 (6.5%) 995 (4.6%) 700 (3.3%) 471 (2.2%)
Humeral hemiarthroplasty 957 2,093 1,376 (65.7%) 156 (7.5%) 278 (13.3%) 159 (7.6%) 233 (11.1%) 65 (3.1%)
Stemmed 941 1,549 938 (60.6%) 143 (9.2%) 248 (16.0%) 127 (8.2%) 181 (11.7%) 47 (3.0%)
Stemless 5 480 387 (80.6%) 11 (2.3%) 27 (5.6%) 31 (6.5%) 44 (9.2%) 13 (2.7%)
Uncertain 11 64 51 (79.7%) 2 (3.1%) 3 (4.7%) 1 (1.6%) 8 (12.5%) 5 (7.8%)
Resurfacing humeral hemiarthroplasty 8 2,243 1,892 (84.4%) 149 (6.6%) 56 (2.5%) 122 (5.4%) 80 (3.6%) 40 (1.8%)
Total conventional shoulder arthroplasty 29 7,076 6,415 (90.7%) 91 (1.3%) 161 (2.3%) 288 (4.1%) 180 (2.5%) 140 (2.0%)
Stemmed 28 4,981 4,543 (91.2%) 77 (1.6%) 100 (2.0%) 191 (3.8%) 122 (2.5%) 86 (1.7%)
Stemless 1 1,731 1,574 (90.9%) 4 (0.2%) 39 (2.3%) 82 (4.7%) 49 (2.8%) 33 (1.9%)
Uncertain 0 364 298 (81.9%) 10 (2.8%) 22 (6.0%) 15 (4.1%) 9 (2.5%) 21 (5.8%)
Resurfacing total shoulder arthroplasty 1 832 731 (87.9%) 26 (3.1%) 19 (2.3%) 47 (5.7%) 15 (1.8%) 13 (1.6%)
Reverse polarity total shoulder
1,041 8,927 2,676 (30.0%) 5,440 (60.9%) 866 (9.7%) 366 (4.1%) 177 (2.0%) 205 (2.3%)
arthroplasty
Stemmed 1,032 8,716 2,617 (30.0%) 5,322 (61.1%) 842 (9.7%) 356 (4.1%) 173 (2.0%) 194 (2.2%)
Stemless 1 85 29 (34.1%) 55 (64.7%) 4 (4.7%) 0 (0.0%) 1 (1.2%) 3 (3.5%)
Uncertain 8 126 30 (23.8%) 63 (50.0%) 20 (15.9%) 10 (7.9%) 3 (2.4%) 8 (6.4%)
Uncertain 2 399 352 (88.2%) 15 (3.8%) 13 (3.3%) 13 (3.3%) 15 (3.8%) 8 (2.0%)
*Percentages based on the total numbers of elective cases; note the listed reasons are not mutually exclusive in the sense that more than one reason could have been stated but this was only 5.76% for elective cases.
**Includes 22 metastatic cancer/malignancies that have only been documented separately since November 2014, when MDSv6 was introduced.
Table 3.44 Gender and age at primary for the main types of primary shoulder replacements. These are shown
separately for acute trauma and elective cases†.
Age at primary in years:

Shoulder type Number of cases Number (%) male* Median (IQR**), Range***
All cases 2,038 458 (22.5%) 74 (67-80), 37-99
Humeral hemiarthroplasty 957 254 (26.5%) 70 (62-78), 37-96
Resurfacing humeral
8 2 74, 51-82
hemiarthroplasty
Total conventional
Acute 29 11 (37.9%) 70 (64-77), 40-86
shoulder arthroplasty
trauma
Resurfacing total shoulder
1 1
arthroplasty
Reverse polarity total
1,041 190 (18.3%) 76 (71-81), 51-99
Uncertain 2 0 57, 54-60

All cases 21,570 6,384 (29.6%) 73 (67-79), 17-99
Humeral
2,093 685 (32.7%) 70 (61-77), 17-95
hemiarthroplasty
Stemmed 1,549 461 (29.8%) 71 (63-78), 20-95
Stemless 480 198 (41.3%) 67 (58-75), 17-91
Uncertain 64 26 (40.6%) 66 (51-75), 31-90
Resurfacing humeral
2,243 665 (29.7%) 72 (64-78), 20-95
hemiarthroplasty
Total conventional
7,076 2,070 (29.3%) 71 (64-76), 22-96
Elective Stemmed 4,981 1,366 (27.4%) 71 (66-77), 24-96
Stemless 1,731 580 (33.5%) 69 (62-75), 23-93
Uncertain 364 124 (34.1%) 68 (60-75), 22-96
Resurfacing total
832 253 (30.4%) 70 (62-76), 20-95
8,927 2,563 (28.7%) 76 (70-80), 18-99
Stemmed 8,716 2,484 (28.5%) 76 (70-80), 24-99
Stemless 85 26 (30.6%) 75 (69-78), 50-92
Uncertain 126 53 (42.1%) 71 (63-76), 18-92
Uncertain 399 148 (37.1%) 70 (62-77), 18-92
† Cells are blank when there are too few data for meaningful analysis.
*Percentages not shown where n<10.
**IQR=Inter-quartile range, i.e. 25th to 75th centile – not given where number is small.
***Range is lowest – highest.
Figure 3.22
Gender and age distribution of elective primaries, for each type of primary procedure.
(a) Humeral hemiarthroplasty (b) Resurfacing humeral hemiarthroplasty

Male Male
600 600
400 400
200 200
Frequency
Frequency
0 0
Female Female
600 600
400 400
200 200
0 0
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
Age at primary Age at primary
(c) Total conventional shoulder arthroplasty (d) Resurfacing total shoulder arthroplasty
Male Male
3000 300
2000 200
1000 100
Frequency
Frequency
0 0
Female Female
3000 300
2000 200
1000 100
0 0
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
Age at primary Age at primary
(e) Reverse polarity total shoulder arthroplasty

Male
3000
2000
1000
Frequency
Female
3000
2000
1000
0
10 20 30 40 50 60 70 80 90 100
Age at primary
Table 3.45 Stem brands used in primary procedures (not resurfacing).
Numbers of primaries implanted within the last year (2016)

Type of Elective procedure
Reverse
Total Total polarity
Number of Number Humeral conventional total
primaries in Number in the for Acute Number for hemi- shoulder shoulder
Stem brands the registry* last year trauma Elective arthroplasty arthroplasty arthroplasty
Stemmed
Oxford Modular 78 12 0 12 11 1 0
Aequalis Ascend 38 1 0 1 1 0 0
Aequalis stem 534 64 20 44 22 22 0
Affiniti Stem 14 0 - - - - -
Comprehensive 1,823 555 61 494 23 126 345
Delta Xtend 3,055 836 89 747 9 7* 731
Global Unite 418 194 37 157 20 134 3
Global FX 167 18 13 5 4 1 0
Global AP humeral stem 1,108 203 1 202 34 168 0
Global Advantage stem 781 108 10 98 17 80 1
RSP 140 48 4 44 0 0 44
Vaios stem 423 73 10 63 6 11 46
Lima SMR stem 1,339 414 51 363 25 89 249
Affinis stem 110 26 1 25 6 19 0
Arrow 259 48 2 46 3 30 13

Equinoxe Stem 1,566 463 8 455 13 136 306
Mosaic 1 0 - - - - -
Anatomical shoulder 951 257 34 223 6 70 147
B/F 110 8 1 7 3 4 0
TM reverse 381 118 14 104 4 5 95
EPOCA 603 103 6 97 8 89 0
Verso 271 101 12 89 0 1 88
Bio-Modular shoulder 13 0 - - - - -
METS Shoulder 13 3 0 3 1 1 1
Polarus 5 0 - - - - -
Nottingham 49 0 - - - - -
Aequalis Ascend Flex 1,022 469 3 466 67 181 218
SMR 9 0 - - - - -
NEER 3 25 0 - - - - -
Affinis Fracture 175 40 36 4 1 0 3
Affini Inverse 377 122 7 115 1 0 114
Aglion Stem 4 0 - - - - -
Humelock II 7 0 - - - - -
Univers Reverse 20 12 0 12 0 2 10
Equinoxe Fracture 245 75 60 15 3 0 12
Aequalis Reversed II 820 181 4 177 3 1 173
Aequalis Reversed 173 62 43 19 0 1 18
Stemless
TESS 61 0 - - - - -
UNIC 7 0 - - - - -
Simpliciti 398 106 0 106 17 89 0
Eclipse Stem 379 117 0 117 14 103 0
Affinis Short 1,088 336 0 336 78 258 0
Multiple brands entered 22 0 - - - - -

Missing
1,041 244 3 241 20 175 46
(no brand entered)
Total 20,123 5,417 530 4,887 420 1,804 2,663
*Possible misclassifications that are being investigated further; excludes the 401 primaries where the type of
procedure was uncertain. www.njrcentre.org.uk 167
Table 3.46 Resurfacing brands used in primary resurfacing shoulder replacements, shown separately for acute trauma
and elective cases.
Numbers of primaries implanted within the last year (2016)

Number by Elective group
Resurfacing Number of Resurfacing Resurfacing
humeral head primaries in Number in the Number for Number for humeral hemi- total shoulder
brand the NJR* last year Acute trauma Elective arthroplasty arthroplasty
Aequalis head 276 37 0 37 32 5
Copeland 1,497 170 0 170 151 19
Global CAP 566 96 1 95 78 17

Vaios head 48 23 0 23 20 3
Lima SMR head 134 35 0 35 28 7
Arrow
resurfacing 42 6 0 6 5 1
head
Sidus 6 1 0 1 1 0
EPOCA
458 78 0 78 14 64
resurfacing
Hemicap 6 1 0 1 1 0
Equinoxe
13 11 0 11 2 9
humeral head
Multiple brands
1 0 - - - -
entered
Missing (no
37 6 0 6 4 2
brand entered)
Total 3,084 464 1 463 336 127
Note: Excludes the 401 primaries where the type of procedure was uncertain.
Glenoids used in total conventional Of the glenoids used in the 7,105 total conventional
shoulder arthroplasty shoulder arthroplasties; 5,560 of these had
information entered about their type which were
Glenoid components continue to be developed then sub-divided according to the composition and
and many are now available on the market. Some fixation of the glenoids.
manufacturers have more than one glenoid type
that can be used as part of their total conventional In this registry, if cement is used even partially as
shoulder arthroplasty brand. This means branding it is in some new ‘hybrid’ fixation glenoids then
is important in the future so that the performance of that component is considered cemented. As such
these different glenoid implant types can be analysed. Table 3.47 has three groupings of glenoids and their
Some are metal backed with modular polyethylene fixation methods.
inserts, while others are metal backed but have fixed
polyethylene bearing surfaces. Most others are all
polyethylene and are usually either pegged or keeled
and require cement for fixation.
Table 3.47 Composition and fixation of standard glenoids used in total conventional shoulder arthroplasty.

Fixation
Composition Cementless Total
Cemented Cementless HA coated non-HA coated
Metal (modular) 3 13 33 49
Metal polyethylene
144 0 233 377
(fixed poly)
All polyethylene 4,521 0 613 5,134
Total 4,668 13 879 5,560
Table 3.48 List of manufacturers of the standard glenoids used in total conventional shoulder arthroplasty.
Manufacturer Number
DePuy 1,637
JRI Orthopaedics Ltd 3
Zimmer Biomet 524
Stanmore Implants Worldwide 24

Tornier 1,060
Mathys Orthopaedics Ltd 854
Synthes 613
Exactech (UK) Ltd 523
Implantcast GmbH 2
FH Orthopedics 107
Lima 65
Arthrex 145
Innovative Design Orthopaedics 1
FX Solutions 1
Multiple manufacturers entered 1
Total 5,560
3.8.2 Revisions after primary shoulder operation, together with 95% Confidence Intervals (CI),
for all cases are shown in Table 3.49, together with
replacement surgery a separation into acute trauma and elective cases.
A total of 582 linked shoulder were subsequently revised. Figure 3.23 further compares the acute trauma and
elective cases for all time points up to four years,
Kaplan-Meier estimates of the cumulative percentage after which time point there were too few cases for
revision at 1, 2, 3 and 4 years after the primary meaningful summary.
primary procedure, shown separately for acute trauma and elective cases. Figures in blue italics signify time points
where fewer than 250 patients remain at risk.
Cumulative percentage probability of revision (Kaplan-Meier estimates),

together with 95% CI, by years from shoulder primary procedure
n 1 year 2 years 3 years 4 years

1.27 2.54 3.45 4.20
All cases 23,608
(1.12-1.43) (2.31-2.78) (3.16-3.77) (3.82-4.62)
1.29 2.68 3.03 3.91
Acute trauma only 2,038
(0.86-1.94) (1.95-3.66) (2.23-4.12) (2.59-5.88)
1.27 2.52 3.48 4.23
Elective cases only 21,570
(1.12-1.43) (2.29-2.78) (3.18-3.82) (3.83-4.67)
Figure 3.23
Kaplan-Meier estimate of the cumulative percentage probability of revision after primary shoulder
replacement with acute trauma and elective cases shown separately.
5.0
4.0
3.0
2.0
1.0
0
0 1 2 3 4
Numbers at risk
Acute trauma 2,038 1,423 886 457 119

Elective 21,570 15,743 10,476 5,768 2,067
A further breakdown by gender and age of the in younger patients of either gender. Revision rates at
cumulative percentage revisions in the elective cases, four years in patients under 65 is 7.6% for men and
shown in Table 3.50, suggests a worse outcome up 6.4% for women. The acute trauma group remains too
to four years for men and a trend to worse outcome small for a similar breakdown.
Table 3.50 Kaplan-Meier estimates of cumulative percentage probability of revision (95% CI) by time from elective
shoulder primary, by gender and age at 1, 2, 3 and 4 years from the primary operation.


Males Females
Age at Years from primary operation: Years from primary operation:

primary
(years)* n 1 year 2 years 3 years 4 years n 1 year 2 years 3 years 4 years
2.38 4.77 6.95 7.62 1.35 3.14 4.64 6.39
<65 1,943 2,202
(1.76-3.23) (3.76-6.04) (5.60-8.61) (6.11-9.48) (0.92-1.98) (2.38-4.14) (3.63-5.93) (4.90-8.30)
1.99 2.86 3.52 4.78 0.96 2.22 3.34 3.86
65-74 2,418 5,629
(1.48-2.68) (2.19-3.72) (2.73-4.54) (3.60-6.34) (0.72-1.27) (1.81-2.73) (2.77-4.02) (3.18-4.70)
2.08 3.34 4.12 4.40 0.72 1.62 2.09 2.59
75+ 2,021 7,348
(1.52-2.85) (2.55-4.38) (3.13-5.41) (3.31-5.83) (0.54-0.96) (1.31-2.00) (1.71-2.56) (2.08-3.22)
*Excludes nine cases for whom the NHS number was not traced and therefore age was not validated.
In Figure 3.24 and Table 3.51, the elective cases Using revision rate alone, Figure 3.24 may lead
have been sub-divided by the type of procedure. The readers to only consider total conventional shoulder
cumulative revision rate was worse for the reverse arthroplasty or reverse shoulder arthroplasty for
polarity replacement during the first two years after elective shoulder replacements. However, it is worth
the primary replacement, after which it seems to noting that sensible options for revision of these two
stabilise and demonstrate the lowest revision rate groups are limited and challenging, where revision
between years two to four. Total conventional of hemiarthroplasty and resurfacing implants is more
shoulder arthroplasty seems to perform relatively straightforward and often influenced by failing and
well in terms of revision over the same period while tearing of the patients rotator cuff shoulder tendons.
hemiarthroplasty operations (including resurfacing) It therefore does remain difficult to evaluate the true
and resurfacing total shoulder arthroplasty have higher outcomes of shoulder arthroplasty on the basis of
revision rates in years three and four. revision rates alone and patient reported outcome
measures (PROMs) remain a critical adjunct in
The early increased revision rates for reverse shoulder assessing implant performance and failure. Shoulder
arthroplasty may represent issues with instability PROMs are discussed later in this report.
and the need to revise modular parts of the implant,
however infection was also a common cause in this
group (see Table 3.48).
Figure 3.24
Kaplan-Meier estimates of cumulative percentage probability of revision up to four years from primary
shoulder replacement surgery, by type of procedure, for elective cases only.
6.0
5.0
4.0
3.0
2.0
1.0
0.0
0 1 2 3 4
Numbers at risk
Hemi-arthroplasty 2,093 1,625 1,142 672 242
Resurfacing hemi-arthroplasty 2,243 1,886 1,467 926 401
Total prosthetic replacement 7,076 5,181 3,358 1,791 613
Resurfacing total arthroplasty 832 695 530 341 131
8,927 6,029 3,741 1,895 625
prosthetic replacement
primary procedure, for all elective cases, sub-divided by the type of procedure. Figures in blue italics signify time
points at which fewer than 250 patients remain at risk.

Elective cases n 1 year 2 years 3 years 4 years

All cases 21,570 1.27 (1.12-1.43) 2.52 (2.29-2.78) 3.48 (3.18-3.82) 4.23 (3.83-4.67)
Humeral
2,093 0.81 (0.49-1.34) 2.51 (1.82-3.45) 4.18 (3.20-5.45) 5.01 (3.79-6.59)
hemiarthroplasty
Stemmed 1,549* 1.01 (0.60-1.71) 2.53 (1.75-3.64) 4.26 (3.14-5.78) 5.01 (3.67-6.84)

Stemless 480 0.24 (0.03-1.69) 2.59 (1.29-5.14) 3.00 (1.56-5.73) 4.43 (2.05-9.1)
Resurfacing humeral
2,243 0.44 (0.23-0.84) 2.63 (1.98-3.50) 4.28 (3.37-5.43) 5.17 (4.09-6.54)
hemiarthroplasty
Total conventional
7,076 0.99 (0.77-1.27) 2.08 (1.71-2.51) 2.81 (2.35-3.36) 3.53 (2.89-4.31)
Stemmed 4,981* 1.19 (0.90-1.55) 2.33 (1.88-2.87) 3.15 (2.58-3.84) 3.92 (3.17-4.84)
Stemless 1,731 0.46 (0.22-0.98) 1.27 (0.74-2.20) 1.78 (1.07-2.96) 2.78 (1.28-5.95)
Resurfacing total
832 0.52 (0.19-1.38) 2.01 (1.17-3.45) 3.93 (2.58-5.96) 6.02 (4.00-9.02)
8,927 1.84 (1.56-2.16) 2.70 (2.34-3.12) 3.17 (2.74-3.66) 3.39 (2.91-3.96)
Stemmed 8,716* 1.77 (1.50-2.09) 2.64 (2.28-3.06) 3.06 (2.63-3.55) 3.30 (2.81-3.87)
Stemless 85 2.71 (0.68-10.41) 2.71 (0.68-10.41) 2.71 (0.68-10.41) 2.71 (0.68-10.41)
*Groupings as in Table 3.43, but note that cases where there was uncertainty about the groupings have been excluded here.
Table 3.52 gives a breakdown of the number of note the indications for revision were not mutually
primaries that were subsequently revised together with exclusive and, for 79 of the 582 first revisions, more
the indications for the first revision procedure. Please than one reason was stated.
Table 3.52 Number of first revisions for each type of primary shoulder replacement and indications for revision.
Acute trauma and elective cases are shown separately.
(i) Acute trauma cases only
Type of primary procedure:

Resurfacing Total Reverse
Humeral humeral conventional Resurfacing polarity total
All hemi- hemi- shoulder total shoulder shoulder
Acute trauma cases arthroplasty arthroplasty arthroplasty arthroplasty arthroplasty Uncertain
Number of cases 2,038 957 8 29 1 1,041 2
Number revised 45 32 2 0 0 11 0
Reasons for
first revision*:
Infection 4 2 0 0 0 2 0
Instability 11 4 1 0 0 6 0
Cuff insufficiency 15 15 0 0 0 0 0
Aseptic loosening 3 0 1 0 0 2 0
Periprosthetic
1 1 0 0 0 0 0
fracture
Conversion hemi-
13 13 0 N/A N/A N/A 0
to total-**
Conversion total-
0 N/A N/A 0 0 0 0
to hemi-
Other indications 9 8 0 0 0 1 0
*Note the reasons are not mutually exclusive; more than one could be stated; MDSv5 refers to these as “Indications for or findings at the time of revision”.
Conversions have been italicised to differentiate from actual reasons for revision.
**Listed as ‘conversions hemi- to total’ but six were revised to reverse polarity total prosthetic replacements and one to a further hemi-arthroplasty
(ii) Elective cases only
Type of primary procedure:

Resurfacing Total Reverse
Humeral humeral conventional Resurfacing polarity total
All hemi- hemi- shoulder total shoulder shoulder
Elective cases arthroplasty arthroplasty arthroplasty arthroplasty arthroplasty Uncertain
Number of cases 21,570 2,093 2,243 7,076 832 8,927 399
Number revised 537 59 80 139 28 139 20

Reasons for
first revision*:
Infection 62 5 6 7 2 40 2
Instability 136 2 4 54 7 66 3
Cuff insufficiency 124 15 27 63 10 5 4
Aseptic loosening 49 5 4 15 1 24 0
Periprosthetic
29 0 1 4 0 23 1
fracture
Conversion hemi-
75 27 37 N/A N/A N/A 11
to total-
Conversion total-
12 N/A N/A 1 0 11 0
to hemi-
Conversion -
13 1 0 2 8 2 0
uncertain**
Other indications 120 14 17 26 3 57 3
*Note the reasons are not mutually exclusive; more than one could be stated; MDSv5 refers to these as “Indications for or findings at the time of revision”.
Conversions have been italicised to differentiate from actual reasons for revision.
**Listed as ‘conversions’ but of a type that would be incompatible with the primary implant.
3.8.3 PROMS Oxford Shoulder Scores was recorded at the pre-operative time point was
predictive of subsequent revision.
(OSS) associated with primary
shoulder replacement surgery As the pilot period has ended, there are no new 2016
PROMs data to add to last year’s report but the pilot
Last year we presented the results of a three successfully demonstrated the critical importance
year NJR pilot to collect the OSS on shoulder of collecting PROMs for shoulder replacements. As
replacement patients. a consequence, a full programme of on-going OSS
collection has been approved by the NJR Steering
This pilot has provided information on the feasibility
Committee and is fully supported by the British Elbow
and usefulness of PROMs for shoulders. In last year’s
and Shoulder Society and the British Orthopaedic
report, 7,131 patients completed the Q1 (pre-op)
Association and commenced in July 2017.
while 7,503 completed the Q2 (at six months).
3,411 patients completed both the Q1 and Q2
questionnaires. While this cohort did not necessarily
3.8.4 Mortality after primary shoulder
represent all 17,199 primaries, 275 of 3,331 elective replacement surgery
cases (8%) had worse scores at six months post-
For this analysis, the second procedure or side of
surgery than they did pre-surgery (i.e. they were
the 20 pairs of bilateral operations performed on
worse off). Further analysis is on-going to see if having
the same day (see Table 3.2) were deleted. Of the
a worse score at six months post-operatively than
remaining 23,588 implants, 1,088 of the recipients probability of mortality shown separately for acute
had died by the end of December 2016. Estimates trauma and the elective cases and shows higher rates
of the cumulative percentage probability of mortality in the acute trauma group.
in this cohort were 0.40 (95% CI 0.32-0.48) at
90 days and 1.55 (95% CI 1.39-1.73), 3.79 (95% However this shows all-cause mortality and in
CI 3.51-4.09), 6.44 (95% CI 6.02-6.88) and 9.96 extended follow-up beyond the immediate post-
(95% CI 9.28-10.69) respectively at 1, 2, 3 and 4 operative period, we would expect higher rates in
years after the primary operation. older age groups, and also in men. In the subsequent
table, Table 3.54, the larger elective group has been
It is important to separate mortality rates following sub-divided in to gender and age sub-groups; the
acute trauma from mortality rates after elective surgery number remains too small for further breakdown in the
due to the different populations and risks involved. acute trauma cases.
Table 3.53 shows the overall cumulative percentage
Table 3.53 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from
shoulder primary, for acute trauma and elective cases at 90 days, 1, 2, 3 and 4 years from the primary shoulder
replacement. Figures in blue italics denote time points where fewer than 250 cases remained at risk, hence the
95% CI are not reliable.

Cumulative percentage probability of death (Kaplan-Meier estimates),
together with 95% CI, by time from shoulder primary procedure
Number 90 days 1 year 2 years 3 years 4 years
1.73 4.08 8.37 12.67 18.10
Acute trauma 2,031
(1.24-2.41) (3.26-5.10) (7.05-9.92) (10.82-14.80) (14.97-21.79)
0.27 1.32 3.36 5.87 9.26
Elective 21,557
(0.21-0.35) (1.16-1.49) (3.09-3.67) (5.45-6.32) (8.57-10.00)
Table 3.54 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from elective
shoulder primary, by age and gender at 90 days, 1, 2, 3 and 4 years from the primary shoulder replacement. Figures
in blue italics denote time points where fewer than 250 cases remained at risk, hence the 95% CI are not reliable.

together with 95% CI, by time from shoulder primary procedure
Males Females
Years from primary operation: Years from primary operation:

Age at
primary
(years)* n 90 days 1 year 2 years 3 years 4 years n 90 days 1 year 2 years 3 years 4 years
0.21 0.87 2.25 3.40 4.50 0.05 0.41 1.38 2.31 4.15
<65 1,942 2,201
(0.08-0.57) (0.53-1.44) (1.57-3.20) (2.46-4.69) (3.18-6.36) (0.01-0.33) (0.20-0.82) (0.91-2.10) (1.61-3.33) (2.89-5.93)
0.25 1.13 2.98 4.38 6.59 0.16 0.67 1.94 3.41 5.86
65-74 2,418 5,625
(0.11-0.57) (0.76-1.68) (2.27-3.92) (3.41-5.61) (5.03-8.61) (0.09-0.32) (0.47-0.94) (1.55-2.43) (2.81-4.13) (4.80-7.16)
0.71 3.45 7.19 12.53 17.77 0.32 1.70 4.47 8.30 13.54
75+ 2,018 7,344
(0.42-1.20) (2.69-4.43) (5.96-8.66) (10.63-14.74) (14.97-21.02) (0.21-0.48) (1.41-2.05) (3.93-5.09) (7.45-9.25) (12.12-15.11)
*Excludes nine cases whose NHS number was not traced therefore the age could not be validated.
3.8.5 Conclusions on-going PROMs programme is important. This was

highlighted last year by the fact that some patients in
We continue to build and improve the shoulder section the elective group (8%) had a worse PROMs score six
of the NJR annual report to produce more granular months after surgery than they did prior to surgery.
data and to present it in a useful and informative
format to all stakeholders. Due to their fundamental Descriptive data on glenoid replacements used as
differences, we continue to present shoulder part of conventional total shoulder arthroplasty are
replacements for acute trauma and elective indications also presented for the first time. It should be noted
separately. This approach has been supported by that some manufacturers have more than one
feedback from the shoulder surgical community. brand of glenoid that can be used with their humeral
components and there is a need to ensure that such
There are now 23,608 primary shoulder replacements implants are correctly branded within the NJR to allow
in the NJR after 4.75 years and it continues to grow for future sub-analysis of glenoid types. We plan to
rapidly. In response to the rapid expansion of shoulder expand this section of the report next year.
implant types on the market we also present for
the first time a more detailed breakdown of stem It should again be noted that overall revision rates are
types to include stemless implants. These in fact do much higher in younger patients, particularly males,
have a short stem but that stem remains within the and these rates are higher than in similar patients
metaphysis of the humeral neck and does not enter undergoing other joint replacements. As this data
the diaphysis. Some surgeons are starting to use grows it is likely to influence decision making by
these implants in response to the industry promotion patients and surgeons on joint replacement in younger
of bone preservation, but the majority at present age groups.
continue to use the standard stemmed implants.
Finally, while the three-year pilot of NJR PROMs
We will continue to observe patterns of use and any
collection was completed and presented in last year’s
variable patterns in revision rates
annual report, the importance of shoulder PROMs has
The use of some shoulder replacements continues been recognised and acknowledged. The collection of
to expand from their original intentions and be used the OSS as part of the NJR shoulder data is now not
across all shoulder pathologies. It remains important just set to continue, but to improve further with Q3 and
to monitor the performance of these implants in these Q4 collection being planned at years three and five post-
different sub-groups both through revision rates and shoulder replacement. Combining the growth of the NJR
using PROMs. With data now out to four years, some with this new comprehensive shoulder PROMs collection
revision rate patterns can be observed, but again will make this an unparalleled shoulder registry.
the need to assess these revision rates alongside an
Part 3
3.9 Outcomes
after primary
elbow replacement
3.9.1 Overview of primary elbow left and right sides, and in one patient these were both
performed on the same day (bilateral), see Table 3.2 in
replacement surgery section 3.2.
This section contains an overview of the primary elbow
The number of primary elbow replacements entered
replacements with linked8 revision and mortality data
into the NJR has continued to increase year by year,
entered into the registry since recording began (1 April
see Table 3.559.
2012) up to the end of 31 December 2016, and
documents the first revision and mortality for these This table also gives a breakdown by the stated type
primaries. Primary elbow replacement in this section of replacement; the Uncertain group here contains
refers to total prosthetic replacements, radial head one radial head replacement that also had an ulnar
replacements and lateral resurfacing replacements. component recorded in the same procedure and
two lateral resurfacings that had ulnar components
A total of 2,196 primary replacements were available
recorded. Some other inconsistencies at the
for analysis for a total of 2,134 patients. Of these
component level have not yet been investigated.
patients, 62 had documented replacements on both
Table 3.55 Numbers of primary elbow replacements by year and percentages of each stated type of procedure.
Year of primary operation

Procedure type All years 2012* 2013 2014 2015 2016
Total 2,196 260 449 448 526 513
Total prosthetic 1,862 227 407 389 438 401
replacement (84.8%) (87.3%) (90.6%) (86.8%) (83.3%) (78.2%)
Radial head 314 23 36 57 88 110
replacement (14.3%) (8.8%) (8.0%) (12.7%) (16.7%) (21.4%)
17 9 5 2 0 1
Lateral resurfacing
(0.8%) (3.5%) (1.1%) (0.4%) (0.0%) (0.2%)
3 1 1 0 0 1
Uncertain
(0.1%) (0.4%) (0.2%) (0.0%) (0.0%) (0.2%)
*Includes one primary operation date given as 2010.
The majority of replacements were performed on have been stated. A total of 685 (31.1%) were carried
women (70.5%). The median age at the primary out for acute trauma (25 of which also had a second
operation was 68 years (IQR 58-77), with an overall reason stated). In this table, and in subsequent tables,
range of 14 to 98 years. 79 of the total prosthetic these acute trauma cases have been separated
replacements had only a humeral component entered, out from the 1,511 remaining elective cases. The
which either means incomplete data entry or possible indications for the elective cases are listed, the
distal humeral hemiarthroplasty replacements. most common reason being Other inflammatory
arthropathy. More than one indication for surgery was
Table 3.56 details the indications for the primary stated in 74 (4.9%) of the elective cases.
operation. Please note that the reasons are not
mutually exclusive as more than one reason could
8 The term “linked” here refers to data linkage not to Linked total elbow arthroplasty.
9 The number of primaries in 2016 was slightly lower than 2015 but this may reflect that some 2016 primaries had not been entered by the time the date was
cut in February 2017.
Table 3.56 Reasons for main types of primary elbow replacements, by year of primary (includes total prosthetic
replacements, radial head replacements and lateral resurfacing replacement).
Elective
Acute
trauma
Number (%)** for each reason (amongst elective cases only):
Number Other
Year of of elbow Number Number Osteo- inflammatory Trauma Essex Avascular Other
primary primaries of cases of cases arthritis arthropathy sequelae Lopresti necrosis cause(s)
All years 2,196 685 1,511 527 (34.9%) 724 (47.9%) 249 (16.5%) 5 (0.3%) 5 (0.3%) 77 (5.1%)
2012* 260 65 195 75 (38.5%) 84 (43.1%) 33 (16.9%) 1 (0.5%) 0 (0.0%) 12 (6.2%)
2013 449 119 330 121 (36.7%) 159 (48.2%) 44 (13.3%) 1 (0.3%) 1 (0.3%) 20 (6.1%)
2014 448 122 326 120 (36.8%) 163 (50.0%) 42 (12.9%) 1 (0.3%) 0 (0.0%) 14 (4.3%)
2015 526 191 335 113 (33.7%) 160 (47.8%) 58 (17.3%) 2 (0.6%) 3 (0.9%) 17 (5.1%)
2016 513 188 325 98 (30.2%) 158 (48.6%) 72 (22.2%) 0 (0.0%) 1 (0.3%) 14 (4.3%)
*Includes one primary operation date given as 2010.

**Percentages based on the total numbers of elective cases; note the listed reasons are not mutually exclusive in the sense that more than one reason could have
been stated.
Table 3.57 summarizes the type of procedures used in the acute trauma and elective cases.
Table 3.57 Types of primary elbow procedures used in acute trauma and elective cases.
Type of elbow primary procedure
Total prosthetic Radial head Lateral

Year of primary replacement replacement resurfacing Uncertain Total
All years 433 (63.2%) 252 (36.8%) 0 (0.0%) 0 (0.0%) 685 (100%)
2012 50 (76.9%) 15 (23.1%) 65 (100%)
Acute 2013 90 (75.6%) 29 (24.4%) 119 (100%)
trauma 2014 71 (58.2%) 51 (41.8%) 122 (100%)
2015 120 (62.8%) 71 (37.2%) 191 (100%)
2016 102 (54.3%) 86 (45.7%) 188 (100%)
All years 1,429 (94.6%) 62 (4.1%) 17 (1.1%) 3 (0.2%) 1,511 (100%)
2012 177 (90.8%) 8 (4.1%) 9 (4.6%) 1 (0.5%) 195 (100%)
2013 317 (96.1%) 7 (2.1%) 5 (1.5%) 1 (0.3%) 330 (100%)
Elective
2014 318 (97.5%) 6 (1.8%) 2 (0.6%) 0 (0.0%) 326 (100%)
2015 318 (94.9%) 17 (5.1%) 0 (0.0%) 0 (0.0%) 335 (100%)
2016 299 (92.0%) 24 (7.4%) 1 (0.3%) 1 (0.3%) 325 (100%)
Over the last three years (from 2014), 1,487 of all types year for all primary elbow replacements performed is
of primaries have been entered into the registry (see shown in Table 3.58 below, together with the number of
Table 3.56). These procedures were performed by 320 units and consultants entering only acute trauma cases,
consultants in total, working across 226 units. only elective cases, and both types within that year.
A breakdown of unit and consultant caseload for each
Table 3.58 Number of units and consultant surgeons providing any primary elbow replacements during each
year from 2014 to 2016 (includes total prosthetic replacements, radial head replacements and lateral resurfacing
replacements).
Year of primary
2014 2015 2016

448 526 513
during each year
Number of units providing any

148 161 160
primary replacement types each year
Mean number of any primary
3 3.3 3.2

2 (1-4) 2 (1-4) 2 (1-4)
Units replacements per unit
Number of units who entered:
(i) only acute trauma cases 28 32 29
(ii) only elective cases 82 74 75
(iii) both acute trauma and
38 55 56
elective cases
Number of consultants providing any

189 210 210
primary replacement types each year
Mean number of any primary
2.4 2.5 2.4
2 (1-3) 2 (1-3) 2 (1-3)
Consultants replacements per consultant
Number of consultants who entered:
(i) only acute trauma cases 34 54 46
(ii) only elective cases 109 95 106
(iii) both acute trauma and
46 61 58
elective cases
A breakdown of unit and consultant caseload for each and consultants entering only acute trauma cases,
year for primary total elbow replacements performed only elective cases, and both types within that year is
is shown in Table 3.59. Data on the number of units also shown.
Table 3.59 Number of units and consultant surgeons providing primary total prosthetic replacements.
Year of primary
2014 2015 2016

389 438 401
during each year
Number of units providing total

132 144 141
prosthetic replacements
Mean number of total prosthetic
2.9 3 2.8
Median (IQR) number of total
2 (1-4) 2 (1-4) 2 (1-3)
prosthetic replacements per unit
Number of units who entered:

Units
(i) total prosthetic replacements
15 19 18
for acute trauma only
(ii) total prosthetic replacements
85 80 79
for elective cases only
(iii) total prosthetic replacements
for both acute trauma and 32 45 44
elective cases
Number of consultants providing total

167 181 170
prosthetic replacements
Mean number of total prosthetic
2.3 2.4 2.4
Median (IQR) number of total
prosthetic replacements per 2 (1-3) 2 (1-3) 2 (1-3)
consultant
Consultants Number of consultants who entered:
(i) total prosthetic replacements
17 33 19
for acute trauma only
(ii) total prosthetic replacements
115 101 104
for elective cases only
(iii) total prosthetic replacements
for both acute trauma and 35 47 47
elective cases
A total of 205 units had entered at least one primary primary total prosthetic replacements in that year
total prosthetic replacement (either elective or acute were 57 and 73 respectively. The numbers of units
trauma) over the three-year period; the maximum and surgeons doing fewer than five total prosthetic
number entered over this three-year period by any replacements per year were 119 and 151 respectively.
one unit was 57, with five units entering 20 or more.
However, 115 units (56%) had entered fewer than five Table 3.60 lists the humeral brands used in total
elective cases over this same period. prosthetic replacements and lateral resurfacings.
Acute trauma and elective cases are shown separately
In 2016, taking elective and trauma cases together, with a further sub-division by type of procedure
the numbers of units and surgeons doing only one amongst the elective cases only.
Table 3.60 Brands used in total prosthetic replacements and lateral resurfacing replacements.
Elective cases
Total prosthetic Lateral

Brand Total number Acute trauma Elective cases replacement resurfacing

Latitude Humeral 230 63 167 167 0
Discovery 500 94 406 406 0
K Elbow 4 0 4 4 0
IBP 9 0 9 9 0
Coonrad Morrey 994 249 745 745 0
GSB 111 39 3 36 36 0
LRE 13 0 13 0 13
NES 2 0 2 2 0
Mutars Elbow 1 0 1 1 0
Custom made part 2 0 2 2 0
No humeral part entered 85 24 61 57 4
Total 1,879 433 1,446 1,429 17
Table 3.61 lists the radial head brands used in total cases are shown separately with a further sub-division
prosthetic replacements, radial head replacements by type of procedure amongst the elective cases only.
and lateral resurfacings. Acute trauma and elective
Table 3.61 Radial head brands used in total prosthetic replacements, radial head replacements and lateral
resurfacing replacements.
Elective cases
Total
Elective prosthetic Radial head Lateral
Brand Total number Acute trauma cases replacement replacement resurfacing
Latitude 6 0 6 5 1 0
RHS 15 7 8 0 8 0
ExploR 21 19 2 0 2 0
Corin Radial Head 19 15 4 0 4 0
Evolve 45 37 8 0 8 0
Anatomic Radial Head 161 137 24 0 24 0
rHead 6 3 3 0 3 0
MoPyC 7 5 2 0 2 0
LRE Radial 14 0 14 0 0 14
Ascension 23 16 7 0 7 0
Liverpool Radial Head 4 3 1 0 1 0
Uni Radial Elbow 1 0 1 0 1 0
Custom made part 1 0 1 0 0 1
Radial stem entered
but no head (thus 51 1 50 50 0 0
unbranded)
No radial implant
1,819 442 1,377 1,374 1 2
included (stem or head)
Total* 2,193 685 1,508 1,429 62 17
*Excludes the uncertain procedures.
3.9.2 Revisions after primary elbow and elective cases. Generally the group sizes were too
small for meaningful sub-division by type of procedure.
replacement surgery However, amongst the 252 radial head replacements
A total of 55 elbow primaries (nine acute trauma cases carried out for acute trauma, no revisions had been
and 46 electives) were revised up to the end of 2016, reported up to the end of 2016. The total prosthetic
including one excision arthroplasty. replacements performed for acute trauma cases,
however, had similar cumulative revision rates to those
Kaplan-Meier estimates of the cumulative percentage for elective cases, as further illustrated in Figure 3.25.
probability of revision up to three years after the
primary operation, together with 95% Confidence At the current time, there are too few cases for further
Intervals, are shown in Table 3.62. sub-division into age/gender sub-groups, but we hope
to do this in future reports as the numbers increase.
The table also shows separate results for acute trauma
Table 3.62 Kaplan-Meier estimates of cumulative percentage probability of revision (95% CI) by time from elbow
primary procedure, shown separately for acute trauma and elective cases. Figures in blue italics signify time points
at which fewer than 250 patients remain at risk.
Cumulative percentage probability of revision

(Kaplan-Meier estimates), together with 95% CI,
by years from elbow primary procedure
Number Number
Elbow primaries of cases revised 1 year 2 years 3 years
All cases 2,196 55 0.90 (0.56-1.45) 2.20 (1.57-3.09) 4.35 (3.27-5.77)
All acute
685 9 0.64 (0.24-1.70) 1.42 (0.66-3.03) 2.14 (1.08-4.23)
trauma cases
Acute Radial head
252 0 0 0 0
Trauma replacements
Total prosthetic
433 9 1.00 (0.37-2.64) 2.16 (1.01-4.57) 3.23 (1.64-6.32)
replacements
All elective cases 1,511 46 1.01 (0.59-1.74) 2.49 (1.71-3.64) 5.12 (3.76-6.96)
Elective
Total prosthetic
cases 1,429 43 0.91 (0.50-1.63) 2.46 (1.66-3.64) 5.08 (3.68-6.98)
replacements
Figure 3.25
Kaplan-Meier estimate of the cumulative percentage probability of revision after primary total prosthetic
replacement with acute trauma and elective cases shown separately.
6.0
5.0

4.0
3.0
2.0
1.0
0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Number at risk
Acute trauma 433 375 309 250 190 149 117 67
Elective 1,429 1,268 1,095 931 755 577 433 280
Table 3.63 gives a breakdown of the indications for more than one reason was stated. It is interesting that
the first data linked revision procedure, the most aseptic loosening is so common within only four years
common reasons being for infection and for aseptic of a primary elective procedure. A few cases once
loosening. Please note, the indications for revision were revised had gone on to have more revision procedures
not mutually exclusive; in five out of the 55 revisions, (other than planned two-stage revisions for infection).
Table 3.63 Indications for first data linked revision after any primary elbow replacement. Acute trauma and elective
cases are shown separately, for all cases and for total prosthetic replacements.
Acute trauma Elective

Total prosthetic Total prosthetic

All cases replacement only All cases replacement only
All cases 685 433 1,511 1,429
Total revised 9 9 46 43
Infection 3 3 18 18
Periprosthetic fracture 2 2 4 4
Instability 1 1 6 4
Other indications 2 2 4 3
Aseptic loosening 2 2 18 18
3.9.3 Mortality after primary elbow Table 3.64 shows the overall cumulative percentage
probability of mortality shown separately for acute
replacement surgery trauma and the elective cases, and shows higher rates
For this analysis, the second procedure of the pair of in the acute trauma group.
bilateral operations performed on the same day (see Table
However this is all-cause mortality and in extended
3.2) were removed. Among the remaining 2,195 implants,
follow-up beyond the immediate post-operative period,
142 of the recipients had died by the end of December
we would expect higher rates in older age groups, and
2016. Estimates of the cumulative percentage probability
also in men. As the size of the dataset increases, we
of mortality in this cohort were 0.46 (95% CI 0.25-0.86) at
will present further sub-divisions by age and gender, as
90 days and 2.53 (95% CI 1.91-3.33), 5.11 (95% CI 4.14-
we have done for other types of joint.
6.32) and 9.10 (95% CI 7.61-10.88) respectively at 1, 2
and 3 years after the primary operation.
Table 3.64 Kaplan-Meier estimates of cumulative percentage probability of mortality (95% CI) by time from elbow primary,
for acute trauma and elective cases at 90 days, 1, 2 and 3 years from the primary elbow replacement. Figures in blue italics
denote time points where fewer than 250 cases remained at risk, hence the 95% CI are not reliable.

together with 95% CI, by time from elbow primary procedure
Number 90 days 1 year 2 years 3 years

0.59 4.14 7.68 15.06
All cases 685
Acute (0.22-1.58) (2.79-6.13) (5.59-10.49) (11.51-19.56)
trauma Total prosthetic 0.70 5.54 10.34 19.70
433
replacements only (0.23-2.16) (3.64-8.39) (7.44-14.28) (15.05-25.54)
0.40 1.83 4.05 6.80
All cases 1,510
(0.18-0.90) (1.24-2.70) (3.04-5.38) (5.34-8.64)
Elective
Total prosthetic 0.43 1.93 4.25 7.15
1,428
replacements only (0.19-0.95) (1.30-2.84) (3.20-5.65) (5.61-9.08)
3.9.4 Conclusions 102 were performed for acute trauma and 299 were
performed for elective indications. The median number
This is the first year we have presented a report on per surgeon and per unit was only two cases.
primary elbow replacements because the numbers
performed annually are small. However, the numbers Besides infection, aseptic loosening was a common
in the registry within 4.75 years from inception are cause of revision within four years and this highlights a
already greater than most other national arthroplasty potentially important issue with elbow replacements and
registers. The data collection compliance rates for their long term performance. With revision surgery being
primary and revision elbows is still not known, and difficult and with the options for revision surgery being
some anomalies in the report may be explained by limited, we will continue to monitor this failure rate and
incomplete reporting by surgeons not familiar with whether there are differences between patient groups.
NJR data entry.
The revision total elbow replacements that took place in
As expected, radial head replacements are more 2016 were performed by 64 surgeons across 52 units.
common in acute trauma procedures and total
Since the start of the registry, distal humeral
prosthetic replacement more common in elective
hemiarthoplasty has also been introduced and
surgery. Over the 4.75 years these joint replacements
become more commonly used, especially in trauma
took place in a total of 246 units but included all types
cases. The minimum dataset classification is due for
of replacement.
modification to include this implant type and they will
In 2016, 401 primary total elbow replacements were begin to appear as a separate implant type in this
performed in 141 units by 170 surgeons. A total of report in future years.
Part 4
Implant and
unit-level activity
and outcomes
Part Four of the annual report gives performance • 20% of NHS providers reported less than 80%
and data entry quality indicators for Trusts, Local
Health Boards (many of whom comprise more than Of those hospitals submitting data, the proportion
one hospital) and independent (private) providers in of patients who gave permission (consent) for their
England, Wales, Northern Ireland and the Isle of Man details to be entered into the NJR were:
for the 2016 calendar year. Outcomes analysis after
hip and knee replacement surgery is also provided for
NHS hospitals
the period 2003 to 2016. • 48% of NHS hospitals achieved a consent rate of
greater than 95%
This section now also provides data for implant outliers
since 2003 and further information on notification and • 34% achieved a consent rate of 80% to 95%
last usage date.
• 18% recorded a consent rate of less than 80%
The full analysis for both units and implants can
be found in the Part Four online document at Independent hospitals
www.njrreports.org.uk – ‘Implant and unit-level • 70% of independent hospitals achieved a consent
activity and outcomes’. rate greater than 95%
4.1 Implant performance • 21% achieved a consent rate of 80% to 95%
• 8% recorded a consent rate of less than 80%

The Implant Scrutiny Committee reports Level 1
outlier implants to the MHRA. Since the committee’s Similarly, the proportion of entries in which there is
formation in 2009 there have been four hip stems, five significant data to enable the patient to be linked to an
hip acetabular (cup) components and 19 hip stem/cup NHS number (linkability) are listed below:
combinations reported. Six knee brands have been
notified. Implants notified in the last year are still in use
NHS hospitals
and none of the hip implants are metal-on-metal. • 87% achieved a proportion of patients with a linkable
NHS number greater than 95%
4.2 Clinical activity • 11% achieved a proportion of 80% to 95%
Overall in 2016, 150 NHS Trusts and Local Health • 2% recorded a proportion of linkable records of less
Boards (comprising 247 separate hospitals) and 177 than 80%
independent hospitals were open and eligible to report
patient procedures to the NJR. All units except for
Independent hospitals
three trauma units submitted data in 2016. • 72% achieved a proportion of patients with a linkable
NHS number greater than 95%
The proportion of all hip and knee joint replacements
entered in to the NJR against those carried out • 21% achieved a proportion of 80% to 95%
(compliance) is only available by NHS Trust and Local
Health Board. No data on this is currently available from • 6% recorded a proportion of linkable records of less
private providers and figures would also exclude units in than 80%
Northern Ireland as compliance is not available.
Note: Independent hospitals might be expected to
• 26% of NHS providers reported 95% or more of the have lower linkability rates than NHS hospitals, as a
joint replacements they undertook proportion of their patients may come from abroad
and not have an NHS number. Linkability figures are
• 54% of NHS providers reported between 80%
not currently available for Northern Ireland.
and 95%
4.3 Outlier units Important note about the outlier hospitals

listed below
for 90-day mortality In previous annual reports, the NJR has reported
outlying hospitals based on all cases submitted to the
and revision rates for NJR since 1 April 2003. To reflect changes in hospital
practices and component use, the NJR now also
the period 2003 to 2016 reports outlying hospitals based on the last five years
of data (1 March 2012 to 1 March 2017 inclusive, the
The observed numbers of revisions of hip and knee latter date being when the dataset was cut). This five
replacements for each hospital were compared to year cut of data excludes from the analysis the majority
the numbers expected given the unit’s case-mix in of withdrawn outlier implants, and metal-on-metal
respect of age, gender and reason for primary surgery. total hip replacements, and thus better represents
Hospitals with a much higher than expected revision contemporary practice.
rate for hip and knee replacement have been identified.
These hospitals had a revision rate that was above Outlier for Hip mortality rates since 20031
the upper of the 99.8% control limits (these limits None identified
approximate to +/-3 standard deviations). We would
expect 0.2% (i.e. one in 500) to lie outside the control Outlier for Knee mortality rates since 20031
limits by chance, with approximately half of these (one Redwood Diagnostic Treatment Centre [closed in 2013]
in 1,000) to be above the upper limit.
Outliers for Hip revision rates, all linked primaries
When examined over the life of the registry, a total of from 20031
28 hospitals reported higher than expected rates of Ashtead Hospital (Surrey)
revision for knee replacement and 41 hospitals had Basingstoke and North Hampshire Hospital
higher than expected rates of revision for hip surgery. BMI Esperance (East Sussex)
However, revisions taken only from the last five years
BMI Gisburne Park Hospital (Lancashire)
of the registry showed only twelve hospitals reporting
BMI Sarum Road Hospital (Hampshire)
higher than expected rates for knees, and five for hips.
BMI The Somerfield Hospital (Kent)
The 90-day mortality for hip and knee replacement Clifton Park Hospital (North Yorkshire)
was calculated for all hospitals by plotting standardised Dunedin Hospital (Berkshire)
mortality ratios for each hospital against the expected Homerton University Hospital
number of deaths. One hospital (closed in 2013) had a Llandough Hospital
higher than expected mortality rate for knee replacement Maidstone District General Hospital
while none were identified for hip replacement. Medway Maritime Hospital
Musgrove Park Hospital
Note: The case mix for mortality includes age, gender
and ASA grade. Trauma cases have been excluded Nevill Hall Hospital
from both the hip and knee mortality analyses together New Hall Hospital (Wiltshire)
with hips implanted for failed hemi-arthroplasty or for North Downs Hospital (Surrey)
metastatic cancer (the latter only from November 2014 North Tyneside General Hospital
when recording of this reason began). Also, where both Northampton General Hospital (Acute)
left and right side joints were implanted on the same Nuffield Health Brighton Hospital (East Sussex)
day, only one side was included in the analysis. Nuffield Health Haywards Heath Hospital (West Sussex)
Nuffield Health Tees Hospital (Cleveland)
Note: Any units identified as potential outliers in Part
Nuffield Health Wessex Hospital (Hampshire)
Four have been notified. All units are provided with an
Annual Clinical Report and additionally have access to Nuffield Health York Hospital (North Yorkshire)
an online NJR Management Feedback system. Pilgrim Hospital
Note: 1 Date range 1 April 2003 to 1 March 2017 inclusive. 2 Date range 1 March 2012 to 1 March 2017 inclusive.
Outliers for Hip revision rates, all linked primaries Outliers for Knee revision rates, all linked primaries
from 20031 from 20031
Prince Charles Hospital New Hall Hospital (Wiltshire)
Rotherham District General Hospital Peterborough City Hospital
Royal Cornwall Hospital (Treliske) South Tyneside District Hospital
Salisbury District Hospital Southampton General Hospital
Shepton Mallet Treatment Centre (Somerset) Southmead Hospital
Spire Cardiff Hospital (Glamorgan) Spire Alexandra Hospital (Kent)
Spire Gatwick Park Hospital (Surrey) Spire Clare Park Hospital (Surrey)
Spire Tunbridge Wells Hospital (Kent) Spire Southampton Hospital (Hampshire)
St Albans City Hospital St Albans City Hospital
St Michael's Hospital St Richard's Hospital
Sussex Orthopaedic NHS Treatment Centre University College Hospital
The Royal London Hospital Withybush General Hospital
University Hospital (Coventry)
University Hospital Of Hartlepool Outliers for Knee revision rates, all linked primaries
University Hospital Of North Tees from 20122
Ashford Hospital
Watford General Hospital
BMI The London Independent Hospital (Greater London)
York Hospital
BMI The Meriden Hospital (West Midlands)
Outliers for Hip revision rates, all linked primaries from Broadgreen Hospital
20122 County Hospital Louth
Homerton University Hospital Ealing Hospital
Southampton General Hospital King Edward VII Hospital Sister Agnes (Greater London)
St Richard's Hospital North East London NHS Treatment Centre (Essex)
Watford General Hospital Spire Southampton Hospital (Hampshire)
Weston General Hospital St Richard's Hospital
University College Hospital
Outliers for Knee revision rates, all linked primaries West Cumberland Hospital
from 20031
BMI Bishops Wood Hospital (Middlesex)
BMI Goring Hall Hospital (West Sussex) 4.4 Better than
BMI The Meriden Hospital (West Midlands)
Bradford Royal Infirmary expected performance
Cannock Chase Hospital
Charing Cross Hospital This year we have also listed hospitals where revision
Conquest Hospital rates are statistically better than expected. These are
units that lie below the 99.8% control limit.
County Hospital Louth
Good Hope Hospital Better than expected for Hip revision rates, all linked
Hinchingbrooke Hospital primaries from 20031
Horton NHS Treatment Centre (Oxfordshire) Addenbrooke's Hospital
Hospital Of St Cross Alexandra Hospital
James Paget University Hospital Bedford Hospital South Wing
King Edward VII Hospital Sister Agnes (Greater London) Bishop Auckland Hospital
Llandough Hospital BMI Mount Alvernia Hospital (Surrey)
Better than expected for Hip revision rates, all linked Better than expected for Hip revision rates, all linked
primaries from 20031 primaries from 20031
BMI Sandringham Hospital (Norfolk) The Cheshire and Merseyside NHS Treatment Centre
BMI The Edgbaston Hospital (West Midlands) [closed 2011]
The Great Western Hospital
Chapel Allerton Hospital
The Princess Royal Hospital
Claremont Hospital (South Yorkshire)
West Suffolk Hospital
Emersons Green NHS Treatment Centre (Avon)
Wrightington Hospital
Euxton Hall Hospital (Lancashire)
Glenfield Hospital [closed 2012]
Better than expected for Hip revision rates, all linked
Goole and District Hospital (Acute)
primaries from 20122
Harrogate District Hospital Addenbrooke's Hospital
Hereford County Hospital Calderdale Royal Hospital
Ipswich Hospital Ipswich Hospital
Kidderminster Treatment Centre Royal Devon & Exeter Hospital (Wonford)
Leicester General Hospital Royal Surrey County Hospital
London Road Community Hospital [closed 2009]
New Cross Hospital Better than expected for Knee revision rates, all linked
Northern General Hospital primaries from 20031
Nottingham Woodthorpe Hospital (Nottinghamshire) Bishop Auckland Hospital
Nuffield Health Brentwood Hospital (Essex) Blackpool Victoria Hospital
Nuffield Health Cambridge Hospital (Cambridgeshire) BMI Beardwood Private Hospital (Lancashire)
Nuffield Health Derby Hospital (Derbyshire) BMI Huddersfield (West Yorkshire)
Nuffield Health Exeter Hospital (Devon) BMI Three Shires Hospital (Northamptonshire)
Nuffield Health Hereford Hospital (Herefordshire) Bronglais General Hospital
Nuffield Health Ipswich Hospital (Suffolk) Chapel Allerton Hospital
Nuffield Health Leicester Hospital (Leicestershire) City Hospital
Nuffield Health North Staffordshire Hospital (Stafordshire) Clifton Park Hospital (North Yorkshire)
Nuffield Health Wolverhampton Hospital (West Midlands) Darlington Memorial Hospital
Prince Philip Hospital Glenfield Hospital [closed 2012]
Princess Alexandra Hospital Ipswich Hospital
Queen Alexandra Hospital London Road Community Hospital [closed 2009]
Queens Hospital Burton Upon Trent New Cross Hospital
Queens Medical Centre Nottingham University Hospital Nuffield Health Derby Hospital (Derbyshire)
Royal Derby Hospital Nuffield Health Ipswich Hospital (Suffolk)
Royal Devon & Exeter Hospital (Wonford) Nuffield Health Leeds Hospital (West Yorkshire)
Royal Hospital at Haslar Treatment Centre [closed 2012] Nuffield Health York Hospital (North Yorkshire)
Royal Stoke University Hospital Princess Alexandra Hospital
Royal Surrey County Hospital Queen Elizabeth Hospital Woolwich
Russells Hall Hospital Queens Hospital Burton Upon Trent
Spire Parkway Hospital (West Midlands) Rivers Hospital (Hertfordshire)
Spire Portsmouth Hospital (Hampshire) Royal Bournemouth Hospital
St Mary's Hospital Royal Derby Hospital
Royal Orthopaedic Hospital
Better than expected for Knee revision rates, all linked
Royal Stoke University Hospital
Russells Hall Hospital
Sandwell General Hospital
Spire Hartswood Hospital (Essex)
St Woolos Hospital
Stepping Hill Hospital
Wansbeck Hospital
Worcestershire Royal Hospital
Wrightington Hospital
Better than expected for Knee revision rates, all linked

Bishop Auckland Hospital
Burnley General Hospital
Calderdale Royal Hospital
New Cross Hospital
North Tyneside General Hospital
Princess Alexandra Hospital
Queens Hospital Burton Upon Trent
Royal Derby Hospital
Glossary
A
Acetabular component The portion of a total hip replacement prosthesis that is inserted into the acetabulum - the socket part
of a ball and socket joint.
Acetabular cup See Acetabular component.
Acetabular prosthesis See Acetabular component.
Antibiotic-loaded bone cement See cement.
Arthrodesis A procedure where the bones of a natural joint are fused together (stiffened).
Arthroplasty A procedure where a natural joint is reconstructed with an artificial prosthesis.
ABHI Association of British Healthcare Industries - the UK trade association of medical device suppliers.
ALVAL Aseptic Lymphocyte-dominated Vasculitis-Associated Lesion. This term is used in the Annual Report
to describe the generality of adverse responses to metal debris, but in its strict sense refers to the
delayed type-IV hypersensitivity response.
ASA American Society of Anaesthesiologists scoring system for grading the overall physical condition of the
patient, as follows: P1 – fit and healthy; P2 – mild disease, not incapacitating; P3 – incapacitating
systemic disease; P4 – life threatening disease; P5 – expected to die within 24 hrs without an operation.
Bearing type The two surfaces that articulate together in a joint replacement. Options include metal-on-polyethylene,
metal-on-metal, ceramic-on-polyethylene, ceramic-on-metal and ceramic-on-ceramic.
Beyond Compliance A system of post market surveillance initiated in 2013. Under this system a scrutiny committee closely
monitors the usage and performance of implants which are new to the market in order that any
problems may be quickly indentified and that the necessary corrective actions are undertaken in order
to protect patient safety.
Bilateral operation Operation performed on both sides, e.g. left and right knee procedures, carried out during a
single operation.
BMI Body mass index. A statistical tool used to estimate a healthy body weight based on an individual’s
height. The BMI is calculated by dividing a person’s weight (kg) by the square of their height (m2).
BOA British Orthopaedic Association - the professional body representing orthopaedic surgeons.
Bone cement See cement.
Brand (of prosthesis) The brand of a prosthesis (or implant) is the manufacturer’s product name, e.g. the Exeter V40 brand
for hips, the PFC Sigma brand for knees, the Zenith brand for ankles, the Delta Xtend brand for
shoulders and the Coonrad Morrey for elbows.
CQC Care Quality Commission. Regulators of care provided by the NHS, local authorities, private
companies and voluntary organisations.
Case ascertainment Proportion of all relevant joint replacement procedures performed in England, Wales, Northern
Ireland and the Isle of Man that are entered into the NJR.
Case mix Term used to describe variation in surgical practice, relating to factors such as indications for surgery,
patient age and gender.
Cement The material used to fix cemented joint replacements to bone - polymethyl methacrylate (PMMA).
Antibiotic can be added to bone cement to try and reduce the risk of infection.
Cemented Prostheses designed to be fixed into the bone using cement.
Cementless Prostheses designed to be fixed into the bone by bony ingrowth or ongrowth, without using cement.
Compliance The percentage of all total joint procedures that have been entered into the NJR within any given
period compared with the expected number of procedures performed. The expected number of
procedures is based on the number of procedures submitted to HES and PEDW.
Compliance Confidence Interval (CI) A ‘Confidence Interval’ (CI) is calculated to accompany anything being estimated from just a random
sample of cases, for example the cumulative probability of revision; a CI tells us something about the
range of values that the ‘true’ (population) value can take. Whilst calculated Confidence Intervals by
their very nature will vary from sample to sample, calculation of a ‘95% Confidence Interval’ (95% CI)
means that 95% of all such calculated intervals should actually contain the ‘true’ value.
Confounding Can occur when an attempt to quantify how a particular variable of interest affects outcome is
hampered by another variable(s) being related to both the variable of interest and the outcome. For
example a comparison of the revision rates between two distinct types of implant may be hampered
by the fact that one implant has been used on an older group of patients than the other; age here
is a ‘confounder’ for the relationship between implant type and outcome because revision rate also
depends on age. Statistical methods may help to ‘adjust’ for such confounding variables.
Cox ‘proportional hazards’ model A type of multivariable regression model used in survival analysis to look at the simultaneous effects of
a number of variables (‘predictors’) on outcome (first revision or death). The effect of each variable is
adjusted for the effects of all the other ‘predictor’ variables in the model so the Cox model can be
used to adjust for ‘confounders’ (see above). Some regression models used in survival modelling make
assumptions about the way the hazard rate changes with time (see ‘hazard rate’). The Cox model
doesn’t make any assumptions about how the hazard rate changes however it does assume that
the predictor variables affect the hazard rates in a ‘proportional’ way; the latter requiring some careful
model checking when this method is used.
Cross-linked polyethylene See modified polyethylene.
Cumulative incidence function (CIF) Used instead of Kaplan-Meier to estimate the cumulative probability of ‘failure’ in the presence of a
‘competing risk(s)’. A competing risk event can prevent the event of interest from occurring; ‘death’
for example is a ‘competing risk’ for revision because once unrevised patients die they can no longer
experience revision. Instead of ‘censoring’ for death (which technically assumes that such patients
might still be at risk of revision but that no further information is available), cumulative incidence
functions make appropriate adjustment.
Cup See Acetabular component.
Data collection periods for annual The NJR Annual Report Part One reports on data collected between 1 April 2016 and 31 March
report analysis report analysis 2017 – the 2016/17 financial year. The NJR Annual Report Parts Two and Four analyse
data on hip, knee, ankle, elbow, and shoulder procedures undertaken between 1 January and 31
December 2016 inclusive – the 2016 calendar year. The NJR Annual Report Part Three reports on hip,
knee, ankle and shoulder joint replacement revision rates for procedures that took place between 1
April 2003 and 31 December 2016.
DDH Developmental dysplasia of the hip. A condition where the hip joint is malformed, usually with a shallow
socket (acetabulum), which may cause instability.
DH Department of Health.
DVT Deep vein thrombosis. A blood clot that can form in the veins of the leg and is recognised as a
significant risk after joint replacement surgery.
Excision arthroplasty A procedure where the articular ends of the bones are simply excised, so that a gap is created
between them, or when a joint replacement is removed and not replaced by another prosthesis.
Femoral component (hip) Part of a total hip joint that is inserted into the femur (thigh bone) of the patient. It normally consists of a
stem and head (ball).
Femoral component (knee) Portion of a knee prosthesis that is used to replace the articulating surface of the femur (thigh bone).
Femoral head Spherical portion of the femoral component of the artificial hip replacement.
Femoral prosthesis Portion of a total joint replacement used to replace damaged parts of the femur (thigh bone).
Femoral stem The part of a modular femoral component inserted into the femur (thigh bone). Has a femoral head
mounted on it to form the complete femoral component.
Funnel plot A graphical device to compare unit or surgeon performance. Measures of performance (e.g. a ratio
of number of observed events to the expected number based on case-mix) are plotted against an
interpretable measure of precision. Control limits are shown to indicate acceptable performance. Points
outside of the control limits suggest ‘special cause’ as opposed to ‘common cause’ variation (see for
example D Spiegelhalter, Stats in Medicine, 2005).
Glenoid component The portion of a total shoulder replacement prosthesis that is inserted into the scapula – the socket
part of a ball and socket joint in conventional shoulder replacement or the ball part in reverse
shoulder replacement.
Glenoid head Domed head portion of the glenoid component of the reverse shoulder replacement attached to
the scapula.
Hazard rate Rate at which ‘failures’ occur at a given point in time after the operation conditional on ‘survival’ up
to that point. In the case of first revision, for example, this is the rate at which new revisions occur in
those previously unrevised.
Head See Femoral head and/or Humeral head.
Healthcare provider NHS or independent sector organisation that provides healthcare; in the case of the NJR, orthopaedic
hip, knee, ankle, elbow or shoulder replacement surgery.
HES Hospital Episode Statistics. Data on case mix, procedures, length of stay and other hospital statistics
collected routinely by NHS hospitals in England.
HQIP Healthcare Quality Improvement Partnership. Manages the NJR on behalf of NHS England.
Promotes quality in health and social care services and works to increase the impact that clinical audit
has nationally.
Humeral component (elbow) Part of a total elbow joint that is inserted into the humerus (upper arm bone) of the patient to replace
the articulating surface of the humerus.
Humeral component (shoulder) Part of a total or partial shoulder joint that is inserted into the humerus (upper arm bone) of the patient.
It normally consists of a humeral stem and head (ball) in conventional shoulder replacement or a
humeral stem and a humeral cup in a reverse shoulder replacement.
Humeral cup The shallow socket of a reverse shoulder replacement attached to the scapula.
Humeral head Domed head portion of the humeral component of the artificial shoulder replacement attached to the
humeral stem.
Humeral prosthesis Portion of a total joint replacement used to replace damaged parts of the humerus (upper arm bone).
Humeral stem The part of a modular humeral component inserted into the humerus (upper arm bone). Has a humeral
head or humeral cup mounted on it to form the complete humeral component.
Hybrid procedure Joint replacement procedure in which cement is used to fix one prosthetic component while the other
is cementless. For hip procedures, the term hybrid covers both reverse hybrid (cementless stem,
cemented socket) and hybrid (cemented stem, cementless socket).
Image/computer-guided surgery Surgery performed by the surgeon, using real-time images and data computed from these to assist
alignment and positioning of prosthetic components.
Independent hospital A hospital managed by a commercial company that predominantly treats privately-funded patients but
does also treat NHS-funded patients.
Index joint The primary joint replacement that is the subject of an NJR entry.
Indication (for surgery) The reason for surgery. The NJR system allows for more than one indication to be recorded.
ISTC Independent sector treatment centre (see Treatment centre).
Kaplan-Meier Used to estimate the cumulative probability of ‘failure’ at various times from the primary operation.
‘Failure’ may be either a first revision or a death, depending on the context. The method properly takes
into account ‘censored’ data. Censorings arise from incomplete follow-up; for revision, for example,
a patient may have died or reached the end the analysis period (end of 2016) without having been
revised. The estimates do not adjust for any confounding factors.
Lateral resurfacing (elbow) Partial resurfacing of the elbow with a humeral surface replacement component used with a lateral
resurfacing head inserted with or without cement.
Linkable percentage Linkable percentage is the percentage of all relevant procedures that have been entered into the NJR,
which may be linked via NHS number to other procedures performed on the same patient.
Linkable procedures Procedures entered into the NJR database that are linkable to a patient’s previous or subsequent
procedures by the patient’s NHS number.
Linked total elbow Where the humeral and ulnar parts of a total elbow replacement are physically connected.
LHMoM Large head metal-on-metal. Where a metal femoral head of 36mm diameter or greater is used in
conjunction with a femoral stem, and is articulating with either a metal resurfacing cup or a metal liner
in a modular acetabular cup. Resurfacing hip replacements are excluded from this group.
LMWH Low molecular weight Heparin. A blood-thinning drug used in the prevention and treatment of deep
vein thrombosis (DVT).
MDS Minimum dataset, the set of data fields collected by the NJR. Some of the data fields are mandatory
(i.e. they must be filled in). Fields that relate to patients’ personal details must only be completed where
informed patient consent has been obtained.
MDSv1 Minimum dataset version one, used to collect data from 1 April 2003. MDS version one closed to new
data entry on 1 April 2005.
MDSv2 Minimum dataset version two, introduced on 1 April 2004. MDS version two replaced MDS version
one as the official dataset on 1 June 2004.
MDSv3 Minimum dataset version three, introduced on 1 November 2007 replacing MDSv2 as the new official
dataset.
MDSv4 Minimum dataset version four, introduced on 1 April 2010 replacing MDSv3 as the new official dataset.
This dataset has the same hip and knee MDSv3 dataset but includes the data collection for total ankle
replacement procedures.
MDSv5 Minimum dataset version five, introduced on 1 April 2012 replacing MDSv4 as the new official dataset.
This dataset has the same hip, knee and ankle MDSv4 dataset but includes the data collection for total
elbow and total shoulder replacement procedures.
MDSv6 Minimum dataset version six, introduced on 14 November 2014 replacing MDSv5 as the new official
dataset. This dataset includes the data collection for hip, knee ankle, elbow and shoulder replacement
procedures.
MHRA Medicines and Healthcare Products Regulatory Agency – the UK regulatory body for medical devices.
Minimally-invasive surgery Surgery performed using small incisions (usually less than 10cm). This may require the use of special
instruments.
Mixing and matching Also known as ‘cross breeding’. Hip replacement procedure in which a surgeon chooses to implant a
femoral component from one manufacturer with an acetabular component from another.
Modified Polyethylene Any component made of polyethylene which has been modified in some way in order to improve its
performance characteristics. Some of these processes involve chemical changes, such as increasing
the cross-linking of the polymer chains or the addition of vitamin E and/or other antioxidants. Others
are physical processes such as heat pressing or irradiation in a vacuum or inert gas.
Modular Component composed of more than one piece, e.g. a modular acetabular cup shell component with a
modular cup liner, or femoral stem coupled with a femoral head.
Monobloc Component composed of, or supplied as, one piece, e.g. a monobloc knee tibial component.
NHS National Health Service.

NICE National Institute for Health and Care Excellence.
NICE benchmark See ODEP ratings.
NJR National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. The NJR has
collected and analysed data on hip and knee replacements since 1 April 2003, on ankle replacements
since 1 April 2010 and on elbow replacements and shoulder replacements since April 2012. It covers
both the NHS and independent healthcare sectors to ensure complete recording of national activity in
England, Wales, Northern Ireland and the Isle of Man.
NJR Centre National coordinating centre for the NJR.
NJR StatsOnline Web facility for viewing and downloading NJR statistics on www.njrcentre.org.uk.
ODEP Orthopaedic Data Evaluation Panel of the NHS Supply Chain. www.odep.org.uk.
ODEP ratings ODEP ratings are the criteria for product categorisation of prostheses for primary total hip and knee
replacement against benchmarks. An ODEP rating consists of a number and a letter and a star. The
number represents the number of years for which the product’s performance has been evidenced.
The letter represents the strength of evidence (data) presented by the manufacturer. The star has
been added to the rating system following revised guidelines from NICE in February 2014, in which a
benchmark revision rate of less than 5% at 10 years was defined. The star is awarded where products
are evidenced to comply with this benchmark. A* represents evidence above A and B. Ratings without
a star signify compliance with the prior NICE guidance of a replacement rate of less than 10% at 10
years. The same benchmark has been adopted by ODEP for knees. All implants that are used without
a 10-year benchmark should be followed up closely. See www.odep.org.uk.
OPCS-4 Office of Population, Censuses and Surveys: Classification of Surgical Operations and Procedures, 4th
Revision – a list of surgical procedures and codes.
Outlier Data for a surgeon, unit or implant brand that falls outside of acceptable control limits. See also
‘Funnel plot’.
Pantalar (ankle) Affecting the whole talus, i.e. the ankle (tibio talar) joint, the subtalar (talo calcaneal) joint and the
talonavicular joint.
Patella resurfacing Replacement of the surface of the patella (knee cap) with a prosthesis.
Patellofemoral knee Procedure involving replacement of the trochlear and replacement resurfacing of the patella.
Patellofemoral prosthesis Two-piece knee prosthesis that provides a prosthetic (knee) articulation surface between the patella
and trochlear.
Patient consent Patient personal details may only be submitted to the NJR where explicit informed patient consent has
been given or where patient consent has not been recorded. If a patient declines to give consent, only
the anonymous operation and implant data may be submitted.
Patient physical status See ASA.
Patient procedure Type of procedure carried out on a patient, e.g. primary total prosthetic replacement using cement.
Patient-time The total of the lengths of time a cohort of patients were ‘at risk’. In the calculation of PTIRs for
revision, for example, each individual patient’s time is measured from the date of the primary operation
to the date of first revision or, if there has been no revision, the date of patient’s death or the last
observation date. The individual time intervals are then added together.
PDS The NHS Personal Demographics Service is the national electronic database of NHS patient
demographic details. The NJR uses the PDS Demographic Batch Service (DBS) to source missing
NHS numbers and to determine when patients recorded on the NJR have died.
PEDW Patient Episode Database for Wales. The Welsh equivalent to Hospital Episode Statistics (HES) in
England.
Primary hip/knee/ankle/elbow/ The first time a total joint replacement operation is performed on any individual joint in a patient.
shoulder replacement
Prosthesis Orthopaedic implant used in joint replacement procedures, e.g. a total hip, a unicondylar knee, a total
ankle, a reverse shoulder or a radial head replacement.
PROMs Patient Reported Outcome Measures.
PTIR Patient-Time Incidence Rate. The total number of events (e.g. first revisions) divided by the total of the
lengths of times the patients were at risk (see ‘patient-time’).
Pulmonary Embolism A pulmonary embolism is a blockage in the pulmonary artery, which is the blood vessel that carries
blood from the heart to the lungs.
Radial head component (elbow) Part of a partial elbow joint that is inserted into the radius (outer lower arm bone) of the patient to
replace the articulating surface of the radial head. May be monobloc or modular.
Resurfacing (hip) Resurfacing of the femoral head with a surface replacement femoral prosthesis and insertion of a
monobloc acetabular cup, with or without cement.
Resurfacing (shoulder) Resurfacing of the humeral head with a surface replacement humeral prosthesis inserted, with or
without cement.
Reverse shoulder replacement Replacement of the shoulder joint where a glenoid head is attached to the scapula and the humeral
cup to the humerus.
Revision burden The proportion of revision procedures carried out as a percentage of the total number of surgeries on
that particular joint.
Revision hip/knee/ankle/elbow/ Operation performed to remove (and usually replace) one or more components of a total joint
shoulder replacement prosthesis for whatever reason.
Shoulder hemi-arthroplasty Replacement of the humeral head with a humeral stem and head or shoulder resurfacing component
which articulates with the natural glenoid.
Single-stage revision A revision carried out in a single operation.
SOAL Lower Layer Super Output Areas. Geographical areas for the collection and publication of small area
statistics. These are designed to contain a minimum population of 1,000 and a mean population size
of 1,500. Please also see Office for National Statistics at www.ons.gov.uk.
Subtalar The joints between the talus and the calcaneum, also known as the talocalcaneal joints.
Surgical approach Method used by a surgeon to gain access to, and expose, the joint.
Survival (or failure) analysis Statistical methods to look at time to a defined failure ‘event’ (for example either first revision or death);
see Kaplan-Meier estimates and Cox ‘proportional hazards’ models. These methods can take into
account cases with incomplete follow-up (‘censored’ observations).
Talar component Portion of an ankle prosthesis that is used to replace the articulating surface of the talus at the
ankle joint.
TAR Total ankle replacement (total ankle arthroplasty). Replacement of both tibial and talar surfaces, with or
without cement.
TED stockings Thrombo embolus deterrent (TED) stockings. Elasticised stockings that can be worn by patients
following surgery and which may help reduce the risk of deep vein thrombosis (DVT).
THR Total hip replacement (total hip arthroplasty). Replacement of the femoral head with a stemmed femoral
prosthesis and insertion of an acetabular cup, with or without cement.
Thromboprophylaxis Drug or other post-operative regime prescribed to patients with the aim of preventing blood clot
formation, usually deep vein thrombosis (DVT), in the post-operative period.
Tibial component (knee) Portion of a knee prosthesis that is used to replace the articulating surface of the tibia (shin bone) at
the knee joint. May be modular or monobloc (one piece).
Tibial component (ankle) Portion of an ankle prosthesis that is used to replace the articulating surface of the tibia (shin bone) at
the ankle joint.
TKR Total knee replacement (total knee arthroplasty). Replacement of both tibial and femoral condyles (with
or without resurfacing of the patella), with or without cement.
Total condylar knee Type of knee prosthesis that replaces the complete contact area between the femur and the tibia of a
patient’s knee.
Treatment centre Treatment centres are dedicated units that offer elective and short-stay surgery and diagnostic
procedures in specialties such as ophthalmology, orthopaedic and other conditions. These include
hip, knee, ankle, elbow, and shoulder replacements. Treatment centres may be privately funded
(independent sector treatment centre – ISTC). NHS Treatment Centres exist but their data is included
in those of the English NHS Trusts and Welsh Local Health Boards to which they are attached.
Trochanter Bony protuberance of the femur, found on its upper outer aspect.
Trochanteric osteotomy Temporary incision of the trochanter, used to aid exposure of hip joint during some types of total
hip replacement.
Two-stage revision A revision procedure carried out as two operations, often used in the treatment of deep infection.
Type (of prosthesis) Type of prosthesis is the generic description of a prosthesis, e.g. modular cemented stem (hip),
patellofemoral joint (knee), talar component (ankle), reverse shoulder (shoulder) and radial head
replacement (elbow).
Ulnar component (elbow) Part of a total elbow joint that is inserted into the ulna (inner lower arm bone) of the patient to replace
the articulating surface of the ulna. May be linked or unlinked.
Uncemented See cementless.
Unicondylar arthroplasty Replacement of one tibial condyle and one femoral condyle in the knee, with or without resurfacing of
the patella.
Unicondylar knee replacement See Unicondylar arthroplasty.
Unilateral operation Operation performed on one side only, e.g. left hip.
Unlinked total elbow Where the humeral and ulnar parts of a total elbow replacement are not physically connected.
Data collection For research and analysis purposes, NJR data is annually linked to data from
The National Joint Registry (NJR) produces this report using data collected, other healthcare systems using patient identifiers, principally a patient’s NHS
collated and provided by third parties. As a result of this the NJR takes no number. These other datasets include the Hospital Episodes Statistics (HES)
responsibility for the accuracy, currency, reliability and correctness of any data service, the Patient Episode Database Wales (PEDW), data from the NHS England
used or referred to in this report, nor for the accuracy, currency, reliability and Patient Reported Outcomes Measures (PROMs) programme, and data from the
correctness of links or references to other information sources and disclaims all Office of National Statistics. The purpose of linking to these data sets is to expand
warranties in relation to such data, links and references to the maximum extent and broaden the type of analyses that the NJR can undertake without having to
permitted by legislation. collect additional data. This linkage has been approved by the Health Research
Agency under Section 251 of the NHS Act 2006 on the basis of improving patient
The NJR shall have no liability (including but not limited to liability by reason of safety and patient outcomes: the support provides the legal basis for undertaking
negligence) for any loss, damage, cost or expense incurred or arising by reason the linkage of NJR data to the health data sets listed above.
of any person using or relying on the data within this report and whether caused
by reason of any error, omission or misrepresentation in the report or otherwise. Once the datasets have been linked, patient identifiable data are removed from
This report is not to be taken as advice. Third parties using or relying on the the new dataset so that it is not possible to identify any patient. This data is then
data in this report do so at their own risk and will be responsible for making their made available to the NJR’s statistics and analysis team at the University of Bristol
own assessment and should verify all relevant representations, statements and whose processing of the data is also subject to strict guidelines set out in an
information with their own professional advisers. approved System Level Security Policy. The work undertaken by the University of
Bristol is directed by the NJR’s Steering Committee and the NJR’s Editorial Board
Information governance and patient confidentiality and the results of the analyses are published in the NJR’s Annual Report and in
The NJR ensures that all patient data is processed and handled in line with professional journals. All published work is based on aggregated data, rather than
international and UK standards and within UK and European legislation: protecting individual record level data. This means that no patient could be identified.
and applying strict controls on the use of patient data is of the highest importance.
NJR data is collected via a web-based data entry application and stored and
processed in Northgate Public Services’ (NPS) data centre. In addition to being
accredited to ISO 27001 and ISO 9001, NPS is also compliant with the NHS’
Information Governance Toolkit.
NJR Centre contact details
National Joint Registry

Produced by Pad Creative Ltd www.padcreative.co.uk
based at Northgate Public Services (UK) Ltd

Peoplebuilding 2
Peoplebuilding Estate
Maylands Avenue
Hemel Hempstead
Hertfordshire
HP2 4NW
Telephone: 0845 345 9991

Fax: 0845 345 9992
Email: enquiries@njrcentre.org.uk
Website: www.njrcentre.org.uk
www.njrreports.org.uk
Every effort was made at the time of

publication to ensure that the information
contained in this report was accurate. If
amendments or corrections are required
after publication, they will be published on
the NJR website at www.njrcentre.org.
uk and on the dedicated NJR Reports
website at www.njrreports.org.uk.
At www.njrreports.org.uk, this document

is available to download in PDF format
along with additional data and information
on NJR progress and developments,
clinical activity and implant and
unit-level activity and outcomes.
/nationaljointregistry @jointregistry

- Селуянов В.Н., Биомеханика Двигательного Аппарата Человека

Uploaded by

Copyright:

Available Formats

- Селуянов В.Н., Биомеханика Двигательного Аппарата Человека

Uploaded by

Document Information

Original Title

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

- Селуянов В.Н., Биомеханика Двигательного Аппарата Человека

Uploaded by

Copyright:

Available Formats

HIPS

14th Annual Report

The NJR Editorial Board

NJR RCC Network Representatives

Healthcare Quality Improvement Partnership

Northgate Public Services

University of Bristol / University of Oxford

Pad Creative Ltd (design and production)

The National Joint Registry’s Steering Committee

Key work and developments

Finally, my thanks to the NJR management team, in

Part 1 NJR Data Quality Audit 17

Part 1 Annual progress 21

1.3 Annual Progress 22

1.4 Summary of content for the NJR Annual Report 23

2.2 Navigating the NJR Reports online facility 28

Part 3 Outcomes after joint replacement 2003 to 2016 29

3.2 Summary of data sources and linkage 35

3.3 Outcomes after primary hip replacement 40

3.3.2 Revisions after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3.3.5 Revisions for different causes after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

3.3.6 Mortality after primary hip surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

3.4 Revisions of a total hip replacement 84

3.4.2 Rates of hip re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

3.4.3 Reasons for the hip re-revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

3.4.4 90-day mortality after hip revision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

3.5 Outcomes after primary knee replacement 98

3.5.2 First revision after primary knee surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

3.5.3 Mortality after primary knee surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

3.6 Revisions of knee replacements 143

3.6.3 Reason for knee re-revision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

3.6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

3.7 Outcomes after primary ankle replacement 155

3.7.2 Revisions after primary ankle surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.7.3 Mortality after primary ankle replacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

3.7.4 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

3.8 Outcomes after primary shoulder replacement 160

3.8.2 Revisions after primary shoulder replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

3.8.4 Mortality after primary shoulder replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

3.8.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

3.9 Outcomes after primary elbow replacement 178

3.9.2 Revisions after primary elbow replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

3.9.3 Mortality after primary elbow replacement surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

3.9.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

Part 4 Implant and unit-level activity and outcomes 187

4.4 Better than expected performance 190

Table 3.2 Composition of person-level datasets for main survivorship analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

(i) Acute trauma cases only. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

(ii) Elective cases only. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

(a) Metal-on-polyethylene cemented monobloc cups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

(b.i) Metal-on-polyethylene uncemented metal shells with polyethylene liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

(b.ii) Metal-on-polyethylene with truncated data for head size 44mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

(d) Ceramic-on-polyethylene cemented monobloc cups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

(e) Ceramic-on-polyethylene uncemented metal shells with polyethylene liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

(f) Ceramic-on-ceramic uncemented metal shells with ceramic liners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

(d) Reverse hybrid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

(a) TKR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

(b) Unicondylar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

(c) Patellofemoral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111