Executive Summary

Professor Mike Reed - Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

We now report on some 73,720 primary shoulder replacements with the growth in recent years coming from reverse polarity total shoulder replacements.

Overall activity is now above pre-COVID levels, bolstered by the independent sector (Figure 3.S2). Since 2020 we can see that the recovery of shoulder replacement is due to an expansion of provision within the independent sector (both NHS and independently-funded). Notably, there has been a substantial increase in the number of independently-funded procedures compared to pre-2020 data, but the NHS hospitals are still performing substantially fewer shoulder replacements than prior to the pandemic.

As detailed in Figure 3.S3, there has been a decrease in the use of stemmed humeral hemiarthroplasty and stemmed total shoulder replacements. Surgeons carrying out stemmed humeral hemiarthroplasties tend to not be doing many each year. The growth in stemless total shoulder replacements and stemmed reverse polarity total shoulder replacements appears to be occurring in higher-volume shoulder surgeons. The majority of procedures, by some margin, were stemmed reverse total shoulder replacements. 

Table 3.S3 demonstrates that the average number of primary replacements per unit (now 17) and per surgeon (now 12) is growing again since COVID and is the highest it has been. There has been relatively little growth in the number of surgeons performing these procedures over the last decade while the number of cases per surgeon per annum has grown by around 50%.

As always, the revision data in the shoulder section is very hard to interpret and the reader is referred to the main text and tables. The data needs to be interpreted in the context of revisability, cuff insufficiency, patient frailty, and PROMs scores. Revision rates at five years for patients under 55 however, do remain high in relation to hip and knee replacement at around 10%.

For PROMs, in Table 3.S15 we see that less than a quarter of patients complete the pre-operative Oxford shoulder PROMs at the appropriate time. Some 30% complete the questionnaire at six months but the utility of collecting post-op scores without pre-op is limited. Also, as overall compliance with the completion of pre-op PROMs is poor, those patients that have completed them cannot be seen as representative of the whole cohort. In 2022 less than 12% of elective patients had pre-op and the 6-month PROMs both recorded (Table 3.S17). 

Clearly having both pre-op and post-op scores for the same patients allows the calculation of health gain rather than simply a spot measure of the general pre-op or post-op status. Nonetheless, one of the main questions remaining to be answered from the data is whether the low revision rate in reverse shoulders reflects better overall performance or whether it reflects the relative difficulty in performing a revision procedure on a reverse shoulder prosthesis. It is possible that following post-op PROMs in large groups of patients may help us to resolve this issue even in the absence of individual pre-op scores and these post-op scores will therefore continue to be collected. That said valuable data is available. The vast majority of patients appear to improve with surgery and in the paired PROMs at six months only 5.1% of patients score worse outcomes than they had reported pre-operatively (Figure 3.S11).

The NJR collects pre-operative PROMs as part of the minimum data set to be entered by all hospitals carrying out shoulder surgery. These pre-operative measures are vital in assessing the benefits of shoulder surgery to the patient, and low compliance by hospitals continues to be of concern.
 

Professor Mike Reed - Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

We have now been collecting data on elbow replacement for over ten years and data on first linked revision episodes and mortality are presented. There are over 10,000 elbow replacements in the registry available for analysis, with the majority in women (66%), and this is the largest dataset of elbow replacements globally. Almost half of the recorded primary elbow replacements are radial head replacements, and overall, some 56% of primary elbow replacements are performed for trauma. From Table 3.E2, we can see that in 2023 almost 70% of elbow replacements for trauma indications are radial head replacements, with the rest being equally split between distal humeral hemiarthroplasty and total elbow replacement. In elective primary elbow replacement, total elbow replacement dominates (73%). It is very important that the reader should appreciate this preponderance of trauma indications in the data when considering the analysis of outcomes.

Recovery from the pandemic has been slow. The increase in numbers of other joint replacements being performed recently in the independent sector has not been seen with elbow replacements. Overall, in the last three years, there have been more surgeons performing more than 13 cases per year but there is still a significant proportion of elbow replacements being performed by low volume surgeons performing one or two per year. The median number of elbow replacements per surgeon per year, over the last three years, has remained consistent at around three per annum. Networks are being set up and funded imminently in England, and this may improve the situation.

In terms of brands of total elbow replacement in the registry dataset, Coonrad Morrey and Discovery have had the highest numbers but there is an increase in the use of Latitude EV. In radial head replacement, the monopolar Anatomic dominates.

Within the registry we have 443 data-linked revisions within these elbow primaries. Single-stage is the most common type of revision procedure, although DAIR and one of two-stage procedures make up significant numbers. There is concern that radial head replacement revision may be underreported and excision joint replacement may be poorly recorded by units. This should be reported on NJR forms as with any other revision, and now relevant reoperation procedure.

Figure 3.E8 shows Kaplan-Meier estimates of the cumulative revision of total elbow replacements by implant brand within the elective cases. Broadly the market leaders have a revision rate at seven years of around 7 to 9%.

Mortality for acute trauma cases is 0.27% at 30 days, increasing to almost 24% at ten years. In elective cases 30-day mortality is 0.07%, rising to almost 30% at ten years.

Overall elective activity remains below that seen in 2013 but this may represent other factors such as alternative evolving non-surgical treatments.

It may therefore be necessary to carefully consider the implications of COVID as regards elbow replacement as the background trend may not be so clearly one of increasing numbers year-on-year because the proportion of cases performed for inflammatory arthritis is higher than for many other joints.
 

Professor Mike Reed - Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

The NJR reports on ankle revision and mortality, with almost 10,000 primary ankle replacements available for analysis. Similar to hip and knee replacement, the average age of patients is 69.

Although designed as uncemented implants, some 4% of ankle replacements have been cemented; it is presumed that these operations were associated with poor bone stock or were in low demand patients. Although numbers of primaries performed increased after 2015, they dropped significantly during the COVID pandemic and have only just recovered. The recent recovery in numbers has been achieved in part by surgeons doing higher volumes of ankle replacement – more than 13 per year, but it seems unfortunate that a similar increase in numbers has been seen in those surgeons performing fewer than six cases per year. Similar trends are seen in both the NHS and the independently-funded sector but only in the latter has the number increased to above pre- COVID levels (Figure 3.A3). Expansion has been seen in the fixed bearing design, whereas the last three years have seen a partial recovery in the numbers of mobile bearing implants used against a background of steady decline of these implants over the last 12 years. Some 73.5% of all ankle replacements now have an Infinity tibial component, but 11.1% of these (8.2% of the total) use an Inbone talar component. This is reported separately by the NJR.

Over the years the mean number of primary procedures per surgeon has increased from 3.8 (2010) to 6.2 (2023) while the median has risen from 2 to 4. However, in 2023, 18% of surgeons performing the procedures were doing more than 10 per year, although over 45% of primary ankle replacement procedures were done by surgeons performing 13 or more per year. Figure 3.A4 shows there is still a high number of surgeons completing one to two procedures per year. The British Orthopaedic Foot and Ankle Society (BOFAS) has not given guidance on absolute minimum numbers but do encourage higher volumes per surgeon, and for surgeons to work in networks.

When it comes to revision, we have small numbers available for analysis, but it appears that younger people and particularly younger women have higher revision rates (Table 3.A3). The reason for this difference between males and females is not clear but merits further attention. Overall the revision rate is around 10% at 13 years. The device with Infinity talar and tibial components, now in common use, has a revision rate of <4% at seven years.
 

Mr Andrew Porteous - Incoming Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

There are now over 1.66 million primary knee joint replacement procedures available for analysis within the registry, performed by 3,721 consultant surgeons in 479 hospitals.

We can see 116,845 primary knee replacement procedures were performed in 2023 with a patient mean age of 69.3, an average BMI of 30.8 and 55% were female. Of these procedures, 98% were performed for osteoarthritis. Over the last three years contributing consultant surgeons have performed a median of 111 knee procedures (approximately 37 per annum), and each hospital performed a median of 665.5 knee procedures (approximately 222 per annum).

Figure 3.K1(e) shows an increase in elective primary knee procedures to a new annual high. However, NHS hospitals only performed approximately 80% of the volume they achieved in 2019 and the growth in volume has been in the independent sector, where there has been an increase in both privately-funded and NHS-funded cases. This poses questions about available facilities, staff or productivity in the NHS.

By fixation type, over 90% of all primary knee replacements in the registry are all cemented, with uncemented or hybrid types making up only 2.2% of all primary knee replacements in 2023. Uncemented fixation has only increased in mobile bearing unicondylar knees. The most common total knee replacement (TKR) configuration is a cemented unconstrained fixed bearing implant representing about 68% of TKRs.

There are now multiple total knee replacement implants that are performing extremely well with very large numbers being tracked. Surgeons can choose from a wide number of brands with failure rates of less than 5% at 15 years. Patella resurfacing reduces revision rates for most, but not all, brand constructs.

Unicondylar knee replacement usage has risen to its highest actual and proportional annual usage (14%) since the start of the registry (see Table 3.K2). Most unicondylar knee replacement procedures are now performed by surgeons performing more than 25 cases per year. A steady but small proportion of cases continue to be done by very low volume surgeons, a finding which warrants further analysis. Lateral unicondylar replacement made up less than 4% of all unicondylar knee replacements and shows higher revision estimates than medial unicondylar replacement (Table 3.K5). Patellofemoral replacements remained less than 1% of knee procedures in 2023, with the majority being performed by surgeons doing more than five per year, with 77% of these being performed on women. Multicompartmental replacements remain exceedingly rare making up 0.1% of primary knee replacements and they show revision rates at five years that are over five times higher than cemented TKR and about double that of cemented unicondylar replacements (Table 3.K5).

Despite predictions in the literature of a massive increase in knee revision volume, numbers have remained relatively stable in recent years (Table 3.K13). Before COVID, approximately 7,000 revisions were performed annually, but this dropped to 4,142 in 2020. In 2023, 5,783 knee revisions were performed, an increase of 82 on the year before. Revision rates for total and unicondylar knee replacements both increase with increasing age bands (see Figures 3.K5 (a) and (b)). This is a consideration when comparing revision rates as the mean age of implantation varies considerably (see Table 3.K3). Table 3.K16 (a) shows that aseptic loosening/lysis remains the most common reason for revision (33%), with infection second (24%). Analysis of the timing of the first revision after total or unicondylar replacement, confirms that the need for an early revision correlates with a higher re-revision rate. A first revision within one year of the primary has an almost 9% re-revision rate within a year, whereas if the first revision is at ten years, the re-revision rate at one year is only 2% (Table 3.K15 (b)). Getting the primary procedure right and making it last as long as possible has a significant effect on the subsequent risk of re-revisions.

There is wide variation in the revision rates between the best and worst performing unicondylar implants and surgeons should consider their implant choices carefully (see Table 3.K8). Unicondylar revision rates have improved over time and now the best performing brands (cemented fixed bearing) can achieve revision rates of approximately 1.98% at five years and 5.3% at ten years now well under the 1% per year that used to be considered the standard for total knee replacements historically. This is still almost double the current revision rate of total knee replacement at ten years (3.11%) and 15 years (4.20%) (Table 3.K7 (a)). There is now literature to support that revision rates of unicondylar replacements are influenced by a number of factors including threshold for revision, patient age and surgeon volume. Furthermore, there is evidence for reduced peri-operative morbidity and mortality, improved function and PROMs, and reduced complexity of revision with unicondylar replacements. In patients who are eligible for unicondylar replacement, it is important that the surgeon consider revision rate as well as medical risks and PROMs in the shared decision-making and consenting process with patients.

Table 3.K12 (a) shows mortality after primary knee replacement for the average 70-year-old woman is 20% at ten years, while men over 85 have a 2% mortality at 90 days. Knee replacements in the elderly are very likely to outlast the patient and therefore avoid the need for revision.

The registry now reports 20 years of data. Cohorts at 20 years are relatively small and many original implants are no longer on the market. Nonetheless, it is encouraging that 20-year revision rates for unconstrained cemented fixed bearing TKR appears to be less than 6%.
.

Professor Mike Reed - Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

This year’s Annual Report is based on over 1.5 million primary hip replacements performed by over 4,000 surgeons in almost 500 hospitals.

We are now at 21 years since the NJR’s data collection commenced and we are reporting a maximum of 20 years of follow-up, although the sizes of some of the groups at longer follow-up are modest.

Looking at caseload over the latest three-year period, as we continue to recover from the effects of COVID, the median number of primary procedures per consultant surgeon was 68 (around 22 per annum) and the median number of procedures per unit was 626. Pre-COVID, surgeons performed a median of around 21 primary hip replacement per annum. Hybrid fixations are the most popular choice in 2023, making up over 40% of primary procedures (see Table 3.H2).

In terms of bearing surface, ceramic-on-polyethylene (CoP) is again dominating in both hybrid and uncemented fixations (see Table 3.H2). Metal-on-polythene still dominates in cemented fixations, although fully cemented fixation is now used in less than 18% of cases – its lowest proportion ever in the registry. However, across the whole life of the registry around 24% of hip primaries have been cemented metal-on-polythene. Ceramic-on-ceramic bearings are now infrequently used, contributing about 3% of the total in 2023.

The number of dual mobility implants recorded in the registry has increased this year. This is in part due to better coding and capturing of this articulation, but is also due to a year-on-year increase in use. They are categorised by the material of each part of the bearing surface e.g. metal-on-polyethylene-on-metal (MoPoM) and ceramic-on-polyethylene-on-metal (CoPoM).

This year, 2023, saw the highest number of primary total hip replacements (THR) performed since the registry began, and this is almost 10% above 2019 levels (Table 3.H2). If surgical teams can continue at those levels, it will take until approximately 2031 to eliminate the backlog of hip replacement surgery (French et al 2024). An increase in volumes beyond this will improve the picture further.

Figure 3.H1 (b) shows that after declining substantially in popularity, resurfacing has remained relatively stable over the past five years, with a slight increase in absolute numbers in 2021 and 2022, but this reduced in 2023. In 2023, around half of the resurfacing procedures were performed by consultants who used resurfacing in more than 25 cases per year.

In terms of the volume of primary hip replacements performed, we are just above 2008 levels for procedures in NHS hospitals, despite overall activity having hugely increased in that time (Figure 3.H1(d)). The independent sector numbers have increased hugely, mainly through NHS-funded procedures.

In 2023, the three most common head sizes are 32mm (1st), 36mm (2nd) and 28mm (3rd). Only ten procedures used 26mm heads in 2023. Ceramic-on-polyethylene bearings with 32mm and 36mm heads now dominate.

A total of 47,090 first revisions of a hip replacement have been linked to a previous primary hip replacement recorded in the registry between 2003 and 2023. Figure 3.H4 (b) illustrates that revision rates increased between 2003 and 2007/8 and then declined between 2007/8 and 2023. A similar effect is shown in knees (which didn’t suffer from metal-on metal bearing revisions) and this improvement is felt to be, at least in part, due to the introduction of NJR Clinician Feedback reporting. A hip replacement performed in 2012 has a 10-year revision estimate of 3%, already better than the NICE threshold of 5% at 10 years (NICE 2014).

The revision rate of metal-on-polyethylene-on-metal dual mobility bearings appears higher up to five years across all fixation types than that of most of the unipolar bearing combinations, except metal-on-metal and ceramic-on-metal. The ceramic-on-polyethylene-on-metal dual mobility bearings show lower revision estimates for cemented and uncemented THRs than the metal-on-polyethylene-on-metal combinations but with overlapping confidence intervals (Table 3.H5).

The 1- and 3-year revision rates for ceramic-on-ceramic resurfacing appear similar to those for metal-on-metal resurfacing which are generally higher than for other unipolar variants. The revision rates at five years appear lower (3%), but the numbers at risk at all time points in the ceramic-on-ceramic resurfacing group are low, so this should be interpreted with caution. This should also be weighed against unipolar hip replacement which for the first time has a combination with an estimated revision rate of less than 1% at ten years. This is the cementless Polar Stem with an R3 shell, and a ceramic-on-ceramic bearing (Table 3.H8 (b)). The same table shows that there are 11 cemented, ten uncemented, two hybrid and two reverse hybrid stem / head / cup (or liner / shell) / bearing combinations with revision rates of less than 2% at ten years, where there are at least 250 cases left at risk at that time point; and more combinations are on track to achieve the same.

For the 25% of primary replacements now performed as hybrid ceramic-on-polyethylene hip replacements, the analysis by head size (Figure 3.H10 (h)) shows bearings with 28mm heads had higher revision rates than those with 32mm and 36mm heads (P<0.001).

For the 21% of hip replacements now performed as uncemented ceramic-on-polyethylene (CoP), Figure 3.H10 (e) shows that 28mm and 40mm heads have higher revision estimates than 32mm and 36mm heads.

For the 11% of hip replacement now performed as cemented metal-on-polyethylene (MoP), Figure 3.H10 (a) shows the effect of head size on revision estimates over the follow-up period. Overall, implants with head size 22.25mm had the worst revision estimates over the entire duration of follow-up, but implants with head size 36mm had marginally worse revision estimates in the first six years of follow-up.

The use of total hip replacement for hip fracture appears to have peaked in 2019. The proportion of THRs for a fractured neck of femur using a dual mobility bearing has increased in comparison with previous reports. This may at least partly be due to the increased granularity of the data in the new component database which has been introduced for this year s report and allows better identification of the bearings used. For patients with a hip fracture the use of dual mobility hip replacement has continued to increase with just over 18% of hip replacements being dual mobility in 2023. Last year saw a stabilising in the proportion of dual mobility after 15 years of increase. The results of total hip replacement in this group appear broadly similar with unipolar and dual mobility implants (Figure 3.H12 (b)). Hip replacement for trauma indications still has over a quarter of dual mobility hip replacements being performed by surgeons with an annual volume of four or fewer procedures (Figure 3.H1 (c)).
.

Professor Mike Reed - Chair, Editorial Committee
Mr Tim Wilton - NJR Medical Director

The year commenced 20 years after the first patient details were submitted to the NJR in 2003 and shortly after our 21st anniversary we have celebrated the submission of the four millionth case. As usual we attended the BHS, BASK and BESS specialist society events and in addition we have also celebrated at an event for our wide range of stakeholders recently in Westminster, at which many important development milestones were highlighted by a variety of members of the NJR team and our guests. The registry is the largest such database of joint replacement cases in the world, but more important is that the database is so complete, especially over the last 15 years. There have of course been some challenges along the way.

One such challenge has been the long-planned development of the new NJR Data Warehouse which has now been completed by NEC and which should afford a great deal of improvement in many of the NJR activities. This enhanced system puts many of the associated data streams in the same computer environment so that they can be analysed and integrated in a much more straight-forward way. It is difficult, though, to make an omelette without breaking a few eggs and there have been inevitable complexities and delays involved in overcoming some of the teething difficulties with the new system. We believe that the new system will now allow us to do things more quickly, more smoothly and to add new mechanisms to our repertoire, so the disruption should prove to have been worthwhile. The introduction of the new system has, however, contributed to the staggered appearance of parts of this year’s report because of the delays in the analysis of hip and knee data, while some of the database changes were ironed out. 

We have been developing an elaborate classification system with the German Arthroplasty Registry for some years and this year has not only seen this completed but the classification has also been accepted and adopted by the International Society of Arthroplasty Registries. This will now form the basis for specifying hip and knee implant characteristics for this whole international group. This is a major step forward which should enable all registries using the system to be very precise about exactly which variant of an implant they are describing. Hitherto this has been difficult because of the variety of types of implant available in different geographical areas and the fact that sometimes different names have been used for the same device.

We should now be able to be confident that an implant which has either much better (or much worse!) outcomes than another implant is clearly defined. Similarly, this system will allow the pooling of cases from different registries to enable more powerful analysis of outcomes in the knowledge that those pooled implants are all the same variant. This will be welcomed by the regulatory bodies, as well as clinical and patient organisations, as a significant step forward. 

Embedding this new classification system has also been the cause for some of the delay in our analysis of hip and knee data this year with the result that the production of our Annual Clinical Reports (ACRs) for hospitals, and Consultant Level Reports (CLRs) for surgeons have been delayed somewhat, as well as those respective joint sections of the Annual Report. On a more positive note, we look forward to seeing the wider variety of data in those and other reports as a result of this development.

Every few years we have a review of the Minimum Data Set (MDS), and this most recent review going on during 2023, has now been completed. The resulting version MDSv8 has now been implemented and allows us to collect more details about aspects of the patients, more complexity details for revision cases, and to collect data about complications and re-operations other than revision procedures. Although our primary outcome measure continues to be revision, with the same definition maintained, it is clearly important for patients when they have some other operation in relation to their replaced joint and we have greatly increased the scope of the data collected about these other re-operations.

An important development over the last year has also been to start collecting the same sort of data about hemiarthroplasty for hip fractures that we have previously collected for planned joint replacement. Previously we collected data about Total Hip Replacement for hip fractures but were not collecting hemiarthroplasty cases. These have a similar potential to have variation in the outcome depending upon the specific implant used and it is therefore a welcome development that we should now be able to perform suitable analysis to enable the improved treatment of these many thousands of frail and elderly patients.

A very significant problem over the last two years has been the breakdown in the provision of data feeds from the NHS data sources which has affected Hospital Episode Statistics (HES) data, civil registration data (death information) and national hip and knee Patient Reported Outcome Measures (PROMs) data. These are collected by the central NHS systems but are integral to the work of the NJR, as without these other data feeds the data collected specifically by the NJR can only be partially analysed. Some of these data have been supplied but have arrived very late so could not be included in some of the reports. 

The PROMs data have been an even greater problem as not only have they been greatly delayed but the data files have not been usable or complete when they have finally arrived. These matters have been completely outside the control of the NJR, but a solution is being actively pursued by the team at NEC Software Solutions and the NHS England Medical Directorate. In the meantime, the analysis of PROMs outcomes, which is so important to provide a comprehensive assessment of implants and hospital services alike, has not been possible for the last two years and we are therefore relying on primary outcomes such as revision rates and mortality.

There had been a delay to our development programme caused by the pandemic, as the funding of those projects was put on hold for two years. The programme has now been restarted and some of these developments will be specifically targeted at the data feeds and other issues outlined above.

Our continuing work in specific areas where the data were less complete have included: audits of elbow replacements to achieve full coverage particularly of trauma cases; audit of dual mobility hips to capture those with unusual combinations of components; and audit of shoulder replacements where historic missing cases are being picked up. These joint specific audits have greatly improved volumes on the registry and therefore our ability to analyse the relevant implant ranges and the accuracy of those analyses.

Acknowledgements 

The NJR continues to work collaboratively with our many stakeholders; the most important, of course, are the patients we serve, and whom we would like to thank for allowing us to use their data.

The NJR operational collaboration is a huge team effort. Elaine Young, NJR Director of Operations, has demonstrated the great versatility of her leadership and her team.

Many thanks also to the following without which the NJR could not function:

All members of the NJR Board and members of the NJR committees: 

Executive 
Data Quality 
Editorial 
Implant Scrutiny 
Medical Advisory 
Regional Clinical Coordinators 
Research 
Surgical Performance 

Members of the Data Access Review Group 

Members of the NJR Patient Network

Other organisations: 

Medicines and Healthcare products Regulatory Agency (MHRA) 
Care Quality Commission (CQC) 
NHS England (NHSE) 
Welsh Government 
Northern Ireland Executive 
Isle of Man Department of Health 
States of Guernsey 
Independent Healthcare Providers Network Services 
Getting It Right First Time (GIRFT) 
British Orthopaedic Association (BOA) 
British Hip Society (BHS) 
British Association for Surgery of the Knee (BASK) 
British Elbow and Shoulder Society (BESS) 
British Orthopaedic Foot and Ankle Society (BOFAS) 
European Orthopaedic Research Society (EORS) 
Healthcare Quality Improvement Partnership (HQIP) 
Confidentiality Advisory Group (CAG) 
Association of British HealthTech Industries (ABHI)

We are most grateful to our NJR delivery contractors for their very valuable input into the NJR Annual Report and their many other functions. NEC Software Solutions, University of Bristol and University of Oxford teams help us refine and improve each year.

We offer our personal thanks to Vicky McCormack, Report Project Manager, NEC; and Deirdra Taylor, Associate Director of Communication and Stakeholder Engagement, for getting the final report into shape.