National Joint Registry

Part 1 of the report focuses on the performance and progress of the Registry for 2007/08, highlighting key developments. It provides an account of the work undertaken by the NJR, its funding and governance arrangements and planned developments

Progress during the year 2007/08 included:

  • recording 151,496 hip and knee joint replacement operations, representing 95% of the total carried out in England and Wales in both the NHS and independent healthcare sector. This is the highest ever annual total and brings the total number of procedures recorded by the NJR, as at 31st March 2008, to 583,724
  • continuing to improve the number of records submitted to the NJR with patient consent to 86% by the end of Quarter 4 2008
  • increasing the number of the records submitted with both patient consent and the patient's NHS number from 69% on 31st March 2007, as recorded in the 4th Annual Report, to 83% by 31st March 2008
  • increasing the overall quality of the data held in the database by linking NJR records to data held by the Hospital Episode Statistics (HES) service to find patient NHS numbers.

Key developments and achievements in the year 2007/08 included:

  • publishing the 4th Annual Report which provided a detailed analysis of the data held by the NJR and included, for the first time, survivorship analysis for certain types of procedure
  • continuing to develop links to data held by HES to improve the type of analysis of hip and knee joint surgery that can be undertaken
  • the launch of version 3 of the Minimum Dataset (MDSv3) accompanied by changes to the application which have improved data quality and reduced the amount of time needed to enter operation details
  • the development of a statistical methodology for the identification of higher than expected revision rates. Agreement on how to investigate higher than expected failure rates have been agreed by the NJR Steering Committee following consultation with the British Orthopaedic Association (BOA), the Medicines and Healthcare products Regulatory Authority (MHRA), and the suppliers of orthopaedic implants
  • providing data to the British Hip Society as part of a Parliamentary Sub-Committee review into the use of chemical thromboprophylaxis following surgery
  • working with the MHRA, units, and the orthopaedic suppliers to identify quickly those patients who had received implants which were subsequently the subject of a device alert or product recall, significantly reducing the period of time between the issuing of the notice and clinical review
  • improving the Bulk Upload process in order to make it easier for units to submit data in bulk from other computer systems and avoid the need for entering the same data into different systems. Future, additional developments include:
  • the development of a NJR Clinician Feedback system that will enable surgeons to assess their performance against defined comparators at hospital, sector (NHS or independent healthcare), and national levels
  • the re-development and re-launch of the NJR website to ensure that NJR stakeholders can find relevant content more easily
  • the provision of support into a study on metal on metal hip articulations.

Consideration is also being given to:

  • extending the NJR to include ankle, elbow, shoulder, and wrist joint replacement surgery
  • extending the NJR to Northern Ireland.

Part 2 of the NJR 5th Annual Report summarises the data for hip and knee procedures carried out between 1st January 2007 and 31st December 2007 in England and Wales, and entered into the NJR by 29th February 2008. It also includes a description of outcomes after hip and knee replacements that were entered into the NJR since the start of data collection in April 2003. Revision rates, mortality and pulmonary embolism are also included. The analysis of these outcomes is based on linkage of procedures entered into the NJR with records from HES and the Patient Episode Database Wales (PEDW). This is the first time that PEDW data has been used.

NJR Compliance in 2007

  • The total number of levies collected in England and Wales from 1st January to 31st December 2007 was 164,053 and the total number of hip and knee procedures reported to the NJR was 141,394. This suggests that the compliance rate for 86% for procedures undertaken during the 2007 calendar year. This is an improvement over previous years with compliance at 79% in 2006.

Overview of Hip and Knee Procedures

  • In 2007, 409 hospitals and treatment centres were open and, of these, 383 (94%) submitted data to the NJR. On average, 181 hip replacements and 190 knee replacements were recorded per participating unit.
  • The NJR recorded 68,950 hip procedures, of which 9.7% were revisions or re-operations, and 72,480 knee replacement procedures, of which 5.6% were either revision or re-operation procedures undertaken between 1st January and 31st December 2007.
  • Of all the hip procedures, 66% were carried out in NHS hospitals, 24% in independent hospitals, 5% in NHS treatment centres, and 5% in independent sector treatment centres. For all knee operations, 67% were undertaken in NHS hospitals, 21% in independent hospitals, 5% in NHS treatment centres, and 7% in independent sector treatment centres (ISTCs).
  • 93% of patients at independent hospitals and ISTCs, compared with 82% at NHS units, were graded as fit and healthy or with mild disease, according to the ASA grade system. 98% of procedures undertaken at ISTCs were primary procedures. Revision procedures undertaken at NHS hospitals account for 81% of all revision procedures.
  • The trend towards more cementless total hip replacement procedures has continued, with cementless procedures now accounting for 33% of all total hip replacements. The reduction in the proportion of cemented procedures has also continued, down from 48% in 2006 to 43% in 2007. Hip resurfacing accounts for 8% of total hip replacements.
  • In 2007, 129 different brands of acetabular cups and 144 different brands of femoral stems were recorded in the NJR. The total number of acetabular cups and femoral stems recorded in the NJR database is 165 and 187, respectively.
  • 20% mixed and matched cup-stem combinations were used, a decrease of 2% from 2006.
  • The types of primary knee replacements procedures have, largely, remained the same, with 83% being total knee replacements and 8% being unicondylar knee replacements.
  • 48 brands of condylar knee prostheses were recorded in the NJR, along with 15 brands of unicondylar prostheses, 9 brands of patello-femoral replacement prostheses and 14 brands of hinged prostheses. Of the 48 brands of condylar knee prostheses, all were used in primary procedures and 29 were used in revision procedures. The total of 65 brands recorded by the NJR and available for use, represents a 20% increase from 2006.

Outcome of hip and knee procedures carried out between 1st April 2003 and 30th September 2007

  • Revision rate, mortality, and pulmonary embolism rate analyses were based on procedures in the NJR linked to data in HES. Focus has been placed on hip resurfacing, unicondylar knee replacement and outcomes of various hip bearing surface combinations. The HES database records only those procedures funded by the NHS in England and includes NHS patients being treated in independent sector hospitals.
  • The overall revision rate following primary hip replacement was 0.6% (95% CI 0.5% – 0.6%) at 1 year and 1.2% (95% CI 1.1% – 1.3%) at 3 years. The 3 year revision rates were lowest in patients who received cemented prosthesis (0.7%, 95% CI 0.6% – 0.8%) and highest after hip resurfacing (2.8%, 95% CI 2.3% – 3.5%). Overall revision rates were lower in women (1.1%) than in men (1.4%). However, women have higher revisions rates for resurfacing (3.6%) but lower revision rates for all other prostheses.
  • The calculated hazard ratios suggest that prosthesis type is the most influential factor on implant survival following primary hip replacement. Hip resurfacing procedures have statistically significantly higher revision rates than cemented total hip procedures after adjusting for age, sex, ASA grade and provider type.
  • Data on bearing surfaces was available from 99,397 primary hip replacement procedures. The 3 year revision rates for the four most frequently entered articulation combinations (excluding all hip resurfacing procedures and implants with heads larger than 32mm) show that the highest revision rates are for ceramic on polyethylene and ceramic on ceramic (p<0.001). Multivariable analysis adjusting for age, sex, physical status and prosthesis type did not confirm this observation (p=0.9).
  • The overall revision rate following primary knee replacement was 0.3% (95%CI 0.3% – 0.4%) at 1 year and 1.2% (95%CI 1.1% – 1.3%) at 3 years after surgery. The 3 year revision rates were lowest in patients who receive a cemented prosthesis (1.2%, 95%CI 1.1% – 1.3%) and highest in those who received a unicondylar knee replacement (2.0%, 95%CI 1.5% – 2.7%). Overall, revision rates were lower in women 1.1% 95% CI 1.0% – 1.3%) than in men (1.3% 95% CI 1.2% – 1.5%) (p=0.003). The statistical evidence for an interaction between prosthesis type and sex is weak (p=0.3).
  • Of the 3,377 knee revision procedures reported, 74% were single stage procedures, 11% were stage 1 of a 2 stage process, and 14% were stage 2 of a 2 stage process. The remainder (less than 1%) were conversions to arthrodesis and amputations.
  • The multivariable analysis to investigate the association between prosthesis type and revision in the first three years after surgery, adjusted for age, sex, physical status and provider type, showed that prosthesis type seems to be an important factor determining revision rates. The risk of revision in the first 3 years after surgery was approximately 1.4 times higher in patients who received a unicondylar knee replacement than in patients who received a cemented prosthesis.
  • The mortality rate for all 102,179 first primary hip replacements carried out between 1st April 2003 and 30th September 2007 that could be linked to a HES record was 0.7% (95%CI 0.6% – 0.7%) at 3 months after surgery and 6.2% (95%CI 5.9% – 6.4%) at 3 years after surgery. The mortality rate for all 111,723 first primary knee replacements (excluding patellofemoral replacements) for the same period was 0.5% (95%CI 0.5% – 0.6%) at 3 months after surgery and 6.1% (95%CI 6.0% – 6.4%) at 3 years after surgery. Mortality and Pulmonary Embolism after Hip and Knee Replacement Mortality and pulmonary embolism rates (for both fatal and non-fatal events) after hip and knee replacement were examined more closely for patients receiving aspirin and low molecular weight heparin (LMWH). The crude rates were then subjected to a multivariable analysis adjusting for age, sex, physical status, and the use of mechanical prophylaxis.
  • Of the 55,210 patients prescribed LMWH after hip replacement surgery, the 3 month pulmonary embolism rate was 0.6% (95%CI 0.5% – 0.7%). For the 16,662 patients prescribed aspirin, the corresponding rate was 0.5% (95%CI 0.4% – 0.6%) (p=0.01). The multivariable analysis suggests that the pulmonary embolism rate with LMWH is about 1.2 (95%CI 0.9 – 1.6) times that for those patients prescribed aspirin. However, the effect was no longer statistically significant (p=0.1).
  • For patients prescribed LMWH, the 3 month mortality rate following hip replacement surgery was 0.6% (95%CI 0.5% – 0.7%) with the corresponding rate for aspirin being 0.7% (95%CI 0.6% – 0.9%). After applying the multivariable analysis, the hazard ratio comparing LMWH to aspirin is 0.8 (95%CI 0.6 – 0.9) (p=0.01). This indicates a statistically significantly lower mortality rate in those that receive LMWH, but should be interpreted with caution.
  • The 3 month pulmonary embolism rate after knee surgery was 0.4% (95%CI 0.4% – 0.5%) in the 58,871 patients prescribed LMWH, and 0.5% (95%CI 0.4% – 0.6%) in the 21,343 patients prescribed aspirin. The multivariable analysis suggests that the pulmonary embolism rate with LMWH is about 0.9 (95%CI 0.7 – 1.1) times that with aspirin (p=0.3). This is not statistically significant.
  • The 3 month mortality rate following knee surgery was 0.6% (95%CI 0.5% – 0.6%) in patients prescribed LMWH, with the corresponding rate for those patients prescribed aspirin 0.5% (95%CI 0.4% – 0.6%). The multi-variable analysis suggests that the mortality rate for those prescribed LMWH is about 1.2 (95%CI 1.0 – 1.5 ) times higher than in those prescribed with aspirin (p=0.08). This is not statistically significant.
  • The mortality rates at 3 months following hip and knee replacement for patients receiving mechanical prophylaxis only were 0.7% (95% CI 0.6 – 0.9%) and 0.4% (95% CI 0.4% – 0.6%) respectively. For Pulmonary Embolism at 3 months the rates were 0.7% (95% CI 0.5% – 0.8%) and 0.4% (95% CI 0.4% – 0.5%) respectively.
  • In conclusion, there were very small differences in mortality according to the chemical thromboprophylaxis treatment that patients received. Patients receiving aspirin had statistically significantly higher mortality after hip surgery and lower mortality after knee surgery than those receiving LMWH. These differences should be interpreted with caution, as they are in opposite directions for hip and knee replacements. Also, the risk adjustment that could be applied is likely to be incomplete and therefore 'confounding by indication' cannot be excluded as a contributing factor.

About Author