The proportion of younger patients undergoing total knee arthroplasty (TKA) is increasing and predictions state that the 45–54 age group will be the fastest growing group by 2030. We aim to collate data across studies to assess functional outcomes following total knee replacement in patients under 55 years of age. Secondary outcomes include implant survival and complications will also be reviewed.

The search identified 980 studies for title and abstract review. 43 full-texts were then assessed. 13 studies underwent quality assessment and data extraction from them. PRISMA guidelines were followed throughout. Outcomes extracted included pre- and post-operative functional scores, range of motion (ROM) and patient satisfaction. Clinical complications and survival were also recorded.

Across 13 studies we were able to demonstrate 54-point improvement in clinical Knee Society Score and a 2.9° improvement in range of motion. Satisfaction rate was 85.5%. Revision rate was 5.4% across 1323 TKAs. Risk of revision for aseptic loosening was 0.37% per year.

Total knee arthroplasty is an excellent treatment option for the young osteoarthritic knee with a >50% improvement in functional knee scores. Satisfaction is high and the revision rate remains under 1% per year.

Introduction

The optimal treatment for isolated patello-femoral arthritis is unclear. Patello-femoral arthroplasty (PFA) may offer superior knee function in isolated patello-femoral osteoarthritis compare to TKA. The literature is controversial for patient outcomes in PFA. Some reports showed improved outcomes while others were disappointing. We assessed our outcomes to try to identify causes for poor outcomes.

Purpose of the Study

To identify trends in patient satisfaction of their knee arthroplasty, total and unicompartmental, one year post primary knee arthroplasty surgery, with reference to age, gender and primary diagnosis, from 1990 to 2008.

Purpose

To identify if the age of the patient was related to satisfaction rates 1 year following unicompartmental knee replacement, in a generalist setting in the U.K.

Only 0.8% of arthroplasties registered on the National Joint Registry in 2006 are patello-femoral. The Leicester patello-femoral replacement (Corin) has been in use for over 10 years with satisfactory initial results. The indication for use is isolated patello-femoral osteoarthritis with the theoretical benefits of bone preservation, maintenance of normal knee mechanics and easy revision. The implant was only available in one size and the femoral component was uncemented.

We reviewed 49 patients managed with this prosthesis with a median follow up of 10 years (range 4-16). The mean age of the patients at time of surgery was 64. 62 arthroplasties were performed.51 were in females and 11 in males giving a 5:1 ratio. Thirteen patients had bilateral procedures. Thirty-nine revisions (62.9%) were performed for progressive tibio-femoral arthritis or prosthetic failure. Mean time to revision was 5 years 3 months. The knees were revised to total knee replacements without the need for stems, wedges or constraint. The unrevised knees had a mean survivorship of 8 years 6 months with a range of 4-14 years. The mean Oxford score for these surviving implants was 22.5.

Results of other implants from the literature included the Avon prosthesis with 80% satisfaction rate at 5 years and the Lubinus with 45% satisfaction rate at 7.5 years. The Leicester device showed a pattern of progressive failure with up to 40% revision at 5 years. However those with surviving implants were reasonably happy as shown by the Oxford scores.

We concluded that patello-femoral replacement has a role though this is not as well defined as TKR or even UKR. We posed the question as to whether these results were due to poor patient selection or design failings of the Leicester prosthesis. This prosthesis has been successful at reducing pain and improving function and may have been more successful with more sizes of implant, better instrumentation and more rigid patient selection.

NICE published the guidelines ‘Selection of prostheses for primary hip replacement’ in 2000. Essentially these guidelines made two recommendations: firstly to use prostheses which had attained the ‘10 year benchmark’ of a revision rate of 10% or less at 10 years, or had a minimum of three years revision rate experience that was on target to reach this benchmark; and secondly to use cemented hip prostheses to the exclusion of uncemented and hybrid prostheses.

The information from the Trent Regional Arthroplasty Study (TRAS) has been used to retrospectively examine the types of hip prostheses used from 1990 – 2005, and assess the impact that the NICE guidelines have had on orthopaedic practice.

This study revealed that the percentage of prostheses used which attained the ‘ten year benchmark’ has increased since the guidelines were published. In 2001, of the ten cups, which constituted 80% of the acetabular components used, only three attained this NICE benchmark, but by 2005 this number had risen to eight. Similarly in 2001, of the eight stems, which constituted almost 80% of the femoral components used, only five attained this NICE benchmark. In 2005 seven out of these eight stems had attained the minimum standard.

However contrary to the recommendation made by NICE in 2000, to use cemented prostheses, the results indicate the use of uncemented prostheses has trebled (from 6.7% to 19.2%, n= 137 and 632 respectively), and the use of hybrid prostheses has more than doubled (from 8.8% to 22% of all prostheses, n= 181 and 722 respectively) since the guidelines were published.

Therefore the recommendations made by NICE are not being followed, which calls the value of NICE guidelines into question.

Introduction: To report the 11-year follow-up of patients of contaminated femoral heads donated at primary total hip replacement from a cohort study published in 2004.

Methods: Of the 440 donors, 266 patients (61%) were still alive at a mean of 11.7 years post primary total hip replacement and 170 patients (64%) agreed to review. The patients had either a positive or negative culture from their donated femoral head at primary total hip replacement. The Oxford hip score, European Quality of Life (EuroQOL) questionnaire, rates of complication and revision surgery were used as outcome measures.

Results: At long-term follow-up, the Oxford hip score was not different between the positive (mean = 36, standard deviation = 12.3) and sterile (mean = 39, standard deviation = 10.9) groups (p = 0.40). The EuroQOL questionnaire was also not different between the positive and sterile groups. There was no statistically significant difference in the rate of complications (p = 0.93) and revision surgery (p = 0.11) of the age-matched patients whose femoral heads had a positive culture compared with those whose femoral heads were sterile.

Discussion: The results of this 11-year follow-up of contaminated donated femoral heads cohort study were similar to those at short-term follow-up. Culture results of donated femoral heads play no significant role in predicting failure of primary total hip replacement in the donor at long-term follow-up.

This study presents the intraoperative findings of a cohort of 201 cases of failed Unicompartmental knee arthroplasties (UKA) from the Trent Wales arthroplasty audit group (TWAAG) register from 1990 to 2008. The main objectives of the study were to determine the common modes of failure and trends in implant systems used using sex and age matching criteria.

Results demonstrate the varying reasons for revision, use of augmentation and surgical preference in revision system. Results include survival rates and revision rates of UKA from the Trent Wales arthroplasty audit group. The average patient age at revision surgery with the average times from primary UKA to total knee arthroplasty are demonstrated. Aseptic loosening was the commonest reason for revision in both younger and older age groups, closely followed by Polyethylene wear in the younger age group versus progression of osteoarthritis in the other compartments in the older age group. The commonest implant used was Oxford unicompartmental knee system at primary surgery with the PFC implant used in almost 50% of all cases that were revised. This study demonstrates the survival rate of UKAs to be significantly higher in female patients and in those patients with primary UKAs at a younger age. The trends in revision systems have changed over the years. In the early years, over 50% used the PFC knee systems, compared to the latter eight years where the majority used revision knee systems, (e.g. LCCK and Legion). This trend is due to increased availability and ease of use of revision systems. The commonest site of augmentation was for tibial bone defects. Approximately 50% of all augmented cases required tibial blocks or wedges.

Although current thinking suggests most UKAs can be revised to a primary total knee system without difficulty, a significant proportion required revision implant systems with associated implications.

The aim of this study was to evaluate temporal trends in the prevalence of primary total hip and knee replacements (THR and TKR) throughout the Trent region from 1991 through 2004.

The Trent Regional Arthroplasty Study (TRAS) records details of primary THR and TKR prospectively. TRAS data in conjunction with age-gender population data from the National Office of Statistics was used to quantify the rates of primary THR and TKR as a function of age (45–55, 56–65, 66–75, 76–85 and greater than 85 years), gender and diagnosis (osteoarthritis, rheumatoid arthritis and trauma). Poisson regression analysis was used to evaluate the procedural rate over time in primary THR and TKR as a function of age, gender and diagnosis.

A total of 26,281 THR and 23,606 TKR were recorded during this period. The overall prevalence for primary THR did not change significantly over time (IRR = 1.0, 95% CI: 0.99 to 1.0, p = 0.875), whereas, the overall prevalence for primary TKR increased significantly by 2.5% during the fourteen year period (IRR = 1.025, 95% CI: 1.021 to 1.028, p < 0.001). Analysis showed that females had an increased incidence rate ratio (IRR) for both primary THR (IRR = 1.29, 95% CI: 1.26 to 1.33, p < 0.001) and TKR (IRR = 1.17, 95% CI: 1.14 to 1.20, p < 0.001). Patients aged 74–85 years had the largest IRR for both primary THR (IRR = 6.7, 95% CI: 6.4 to 7.0, p < 0.001) and TKR (IRR = 15.3, 95% CI: 14.4 to 16.3, p < 0.001).

The prevalence of primary TKR increased significantly over time whereas THR increased steadily in the Trent region between 1991 and 2004. These trends have important ramifications to the number of joint replacements expected to be performed in the future.

Purposes of the study: Demographic s Description and Pattern Recognition.

Methods: We reviewed the revision total knee replacements performed between 2001 and 2005, which were recorded on the Trent and Welsh knee register. We looked at the total number recorded in the register, the number of operations entered every year and the performance of each of the hospitals on the register through the 5 years period. We looked as well at the male: female ratio, the median age of revisions, the side of the operation, whether the implants were cemented or not and the seniority of the operating surgeon.

Results: Total number of revisions between 2001–2005 was 662. In the year 2001 we had 69 revisions, in 2002 we had 129 revisions, in 2003–2004 we had 201 revisions and in 2005 we had 263 revisions.

Out of the grand total of 662, 3 were form Bangor hospital, 10 were from Barnsley, 20 were from Basset-law, 2 were from Cardiff, 1 was from Claremont, 38 were from Derby, 18 were from Doncaster, 138 were from Leicester, 1 was from Glan Glwyd, 17 were from Grantham, 121 were from KMH, 2 were from Lincoln, 23 were from Llandough Hospital, 131 were from NGH, 86 were from Nottingham City, 2 were from Parkhill Hospital, 1 was from Pilgrim Hospital, 2 were from Prince Phillip Hospital, 4 were from QMC, 28 were from Royal Gwent, 2 were from Rotherham, 1 was from Thornbury Hospital, 4 were from WWG, 3 were from WWQ. The rest of the revisions were anonymous to place.

344 cases were males, 304 were females and the rest were anonymous.

The average age for males was 74 years and the average age for females was 75 years.

330 operations were carried out on right knees, while 302 were on the left side. The rest were anonymous.

491 revisions were carried out by Consultants, 19 were carried out by a Staff Grade, 37 were carried out by Registrars, 28 were carried out by others and the rest were anonymous. To be noted is that in 414 revisions the surgeons had no assistant.

253 knees were cemented, 10 were marked as uncemented and the rest were anonymous. Bone Grafts were used in 71 patients and augments were used in 107.

Conclusions: The input of the numbers of revisions has improved over the years. Poor documentation of whether the implant was cemented or not. 84 revisions were performed by Trainees/Staff Grades, does this have an impact on the outcome?. In 414 revisions the consultants didn’t have an assistant, this should not occur as revisions are complex operations and the lack of an assistant might have an impact on the outcome.

Purpose: To evaluate the fifteen year survivorship of primary Total Knee Replacements in a single UK health region.

Methods: Since the beginning of 1990, and with the agreement of all consultant orthopaedic surgeons in the region, all primary total knee replacements (TKR) performed throughout Trent were recorded prospectively. At the time of operation the surgeon completes a questionnaire, which records demographic, medical and operative details for each patient and implant.

In this study we have traced all the patients, who had a primary total knee replacement between 1990 and 1992. We issued a validated, self administered questionnaire to all surviving patients, at a mean of fifteen years post arthroplasty. This questionnaire examines the patient’s level of expectation and satisfaction with their TKR, and also measures their quality of life (using EQ-5D and visual analogue score). Using a similar register, containing information of all revision TKR in the region, we have measured the survivorship of these primary TKR at 10 and 15 years.

Results: 4,665 primary TKR were performed on 4,448 patients. At fifteen year follow-up 1,408 patients were alive. The questionnaire response rate was 57.1% (n=912). Of our responders, 87.8% were satisfied with the result of their TKR at 15 years post-arthroplasty, and 82% felt their TKR had met their expectations.

Survivorship analysis revealed that 94.7% (+/−0.4%) of implants survive to 10 years, and 92.7% (+/−0.5%) to 15 years. Survivorship was significantly affected by gender of the patient, age at time of primary, and type of prosthesis used. Infection rate at 15 years was 0.9%.

Discussion: This is one of the first long term studies of primary TKR, which assesses survivorship of primary TKR beyond 10 years. This study shows that survivorship at 5 and 10 years compares favourably to the results of similar studies from other countries.

Approximately 10% of primary hip replacements performed each year for osteoarthritis are in patients aged 55 or less. These patients have a longer life expectancy and a higher activity level than an elder cohort, which may translate to higher revision rates.

We utilized a regional hip register (Trent and Welsh Arthroplasty Audit Group (TWAAG)) to review current surgical practice in this age group. The TWAAG group comprises 118 surgeons working in 31 different hospitals covering a population of 8 million (14.2% of the population).

1 January 2000 to 31 December 2002, we were notified of 7,678 primary THRs for osteoarthritis. 911 (11.7%) were performed on patients aged 55 or less. Age, gender, grade of lead operating surgeon, type of femoral and acetabular prosthesis implanted, fixation method, femoral head size and bearing surfaces were recorded. There were 434 males, 477 females, with an age range of 16–55. Thirty-five femoral and thirty-three acetabular components were identified. 61.7% of femoral prostheses were cemented. 67.4 % of acetabular prostheses were uncemented. 30% of THRs implanted in the group over the study period were hybrid. 50% of implants had a metal/UHMWPE bearing. Other bearing surfaces comprised ceramic/UHMWPE 28.7%, metal/ metal resurfacing 13.8% and ceramic/ceramic 7.5%. Consultants performed 84.5% of procedures.

Femoral prostheses with little or no published data are used and, unless closely monitored, such practices will not be compliant with NICE recommendations. 40% of THRs performed had components implanted that were produced by different manufacturers. At the present time there does not appear to be a clear picture as to what is the ‘gold’ standard for young patients. Continued monitoring of these implants is essential to provide feedback and drive choice.

The Trent Arthroplasty Audit Group has been prospectively collecting data on primary hip and knee arthroplasties since 1990. Details of 61,000 primary and 4,00 revision arthroplasties have been registered. The Royal College of Surgeons of England. Capital Hip Report (July 2001) concluded that a national joint register could have detected failures of an implant at an earlier stage. We examined data on the register to ascertain why we had been unable highlight a problem with this implant.

The Trent Arthroplasty Register was unable to detect the poor results with Capital hips at an earlier stage than surgeons. A scientific presentation had raised concern before our register could detect a problem. The hips had been listed for revision but were still on a waiting list. Additionally some of the failed hips were not revised as patients were insufficiently fit for surgery.

The stated reason for revision on revision forms was vague and not sufficient to draw conclusions as to the mechanism of failure. Radiological studies have identified a higher radiological failure rate than expected (Charnley & Elite +) but we have shown that outcome scores (Oxford Scores) were not successful at identifying these failures. Since the implementation of the Data Protection Act (1998) consent must be obtained before details are registered, which may lead to further inaccuracy in the creation of survivorship curves.

Joint registers can contain the problem once it is detected but are not a substitute for regular follow-up. Surgical vigilance and a scientific approach is required to ascertain the reason for failure. Revision should not be the only endpoint for registration. Joint registers may be part of the solution but need to be backed up with adequate resources, financial and intellectual, to analyse clinical information, if valid conclusions are to be drawn.

We sought to determine the incidence of complications and re-operation up to one year following primary total knee replacement in a single health region.

The Trent Arthroplasty Audit group collects prospective data on all knee replacements performed within this health region (population 5.2 million). All patients are sent a validated, self-administered questionnaire one year after surgery. The questionnaire addresses patient satisfaction and any complications and re-operations following surgery. We analysed the returned questionnaires of 5352 patients [5896 knees] who had their primary knee arthroplasty between 1998 to 2000. Responses were received from 4169 patients [4592 knees] (response rate 80%). Clinical records were also examined to gain further information.

516 patients reported complications in 546 knees. Complication rate of (12%) and 3.5% had a further operation on the joint within one year. Complications were highest following knee replacement for trauma (36%) followed by osteoarthritis (12%) and lowest for rheumatoid arthritis (8%). We have no knowledge of the complexity of the surgery but 60% of the complications occurred in patients operated on by a Consultant, 29% by a Specialist Registrar and 10% by an Associate Specialist & Staff Grade. On the whole Consultants performing fewer than 10 joint replacements per year registered a higher complication rate (21%) as compared to Consultants performing more than 25 joints per year (12%), but they performed 18% of the arthroplasties. The incidence of complications, as stated by the patient was as follows: Pain 7%, Stiffness 2%, Superficial infection 1%, Swelling 0.7%, Deep infection 0.7%, DVT 0.4%. 1.2% (infection 0.4%: Instability 0.7%: Patellar resurfacing 0.2%), Manipulation (1.3%), Arthroscopy (0.7%), ORlF of Peri-prosthetic fracture (0.06%). 12 % of the patients who had a primary knee replacement in Trent region between 1998 and 2000 considered that they had a complication. Complications rates appear to be higher for surgeons performing less than 10 joint replacements per year. Only 43% of Consultants performed more than 10 knee arthroplasties themselves in any one of these three years. The deep infection rate was 0.4% and one-year post surgery the revision rate, for all causes was 1.2% and the manipulation rate was 1.3%.

The Trent Arthroplasty Audit Group has been prospectively collecting data on primary knee arthroplasty since 1990 and revision procedures since 1992. Details of 27 000 primary and 1300 revision knee arthroplasties have been registered. In 2001 hospitals in Wales joined the group, increasing the catchment population to 8 million (14% of the UK population). The register has enabled evaluation of changes in the demography and surgical practice of knee arthroplasty in the Trent region over the past 13 years.

Over this period there has been a steady increase in the number of arthroplasties registered, from 1330 cases in 1990 to 2855 in 2002. Whilst there has been a slight increase in the proportion of men undergoing surgery, the age distribution remains consistent (mean age 69 years). PFC/Sigma is currently the most commonly used prosthesis in the region.

Since 1990 the number of patients registered with rheumatoid arthritis has fallen by almost 50%. During this period there has also been a slight decrease in the proportion of uncemented joints and a decline in the number of bilateral simultaneous procedures taking place.

The ratio of primary to revision knee arthroplasty has not changed significantly since 1992 but there is some evidence of specialisation of revision knee surgery. There has been no significant change in patient satisfaction rates since the start of the registry, with 80% of patients reporting that they are satisfied with their joint replacement at 1 year.

Conclusion: It is reasonable to assume that these findings reflect practice across the UK as a whole, given the diversity of hospitals contributing and the large population base of the Trent & Wales register.

Using the Trent regional arthroplasty register, we analysed the survival at ten years of 1198 consecutive Charnley total hip replacements carried out across a single health region of the United Kingdom in 1990. At ten years, information regarding outcome was available for 1001 hips (83.6%). The crude revision rate was 6.2% (62 of 1001) and the cumulative survival rate with revision of the components as an end-point was 93.1%. At five years, a review of this series of patients identified gross radiological failure in 25 total hip replacements which had previously been unrecognised. At ten years the outcome was known for 18 of these 25 patients (72%), of whom 13 had not undergone revision.

This is the first study to assess the survival at ten years for the primary Charnley total hip replacement performed in a broad cross-section of hospitals in the United Kingdom, as opposed to specialist centres. Our results highlight the importance of the arthroplasty register in identifying the long-term outcome of hip prostheses.

Introduction and Aims: ‘External’ snapping hip, secondary to a tight thickened portion of the iliotibial band snapping over the greater trochanter is usually treated by Z-plasty. We report experience of managing painful paediatric snapping hips by a novel alternative, proximal TFL release.

Method: Patients referred between 1989–2002 with painful snapping hips to the paediatric orthopaedic unit were retrospectively reviewed. Diagnosis was made on the basis of history and physical examination. All the patients had a plain radiograph of the hip and other causes of an audible snap were excluded. Most cases had failed to improve with physiotherapy pre-operatively. Fifteen patients (21 hips) had proximal tensor fascia lata release. All the patients had regular follow-up and were assessed for recurrence and complications.

Results: There were two boys and thirteen girls. The mean age at the time of operation was 15 years and minimum post-operative follow-up was six months. Two hips required revision surgery. Eighty-one percent had a pain-free non-snapping hip at the time of discharge. The remainder were pain-free, with a reduced frequency of snapping.

Conclusion: For painful snapping hips when conservative methods have failed, proximal release can offer good results with low morbidity in a paediatric population.

Introduction and Aims: To assess the 10-year outcome of knee arthroplasty surgery in a generalist setting, across a single UK health region.

Method: The Trent Arthroplasy Audit Group has been collecting prospective data on primary knee replacements since 1990. 4677 primary knee replacements were performed in the period 1990–92. A validated self-administered questionnaire, including a EuroQol health evaluation questionnaire, was mailed to all surviving patients. The mean age at the time of surgery was 70 years.

Results: Thirty-nine percent of the patients had died by 10 years. Responses were received from 87% of patients. Eighty-two percent of patients had been satisfied with the result of their knee replacement. Twenty-one percent of patients felt that it was not as good as they had been led to believe before surgery. Forty percent of patients had had problems with their knee in the 10 years since surgery. Twenty-one percent had had some form of surgical intervention. Eighty-two percent of patients were taking medication to alleviate pain in their replaced knee.

Eighteen percent complained of constant pain in this knee. When the pain occurred it was described as being severe in 16% and moderate in 40%. Most patients were on prescribed medication, but had not been referred to a surgeon to identify if there was a problem with their implant. The revision rate, for all causes at 10 years was 13%.

Conclusion: The results of knee arthroplasty at 10 years are disappointing. Patients’ responses indicate that previously successful knees are becoming painful. At the present time there appears to be inadequate long-term follow-up/re-referral of patients and surgeons have a high threshold for revision surgery. More detailed clinical examination is planned for this group of patients.

Introduction and Aims: To audit the practice and evaluate the outcomes of knee arthroplasty surgery in a general setting in the UK.

Method: The Trent & Wales Arthroplasty Audit group collects prospective data on all knee replacements performed within these UK health regions (population eight million). All patients are sent a validated, self-administered questionnaire to assess outcomes. Data has been collected on 27,500 primary and 1400 revision knee procedures.

Results: In the period from 1990 to 2002 there has been a marked decrease in the proportion of knees being replaced for rheumatoid arthritis (21% to 5%). More arthroplasties are being performed in men (35% to 46%). There has been an increase in knee arthroplasty in patients aged 85+ yrs (2.1% to 3.7%), but a decrease in patients aged 55yrs and under (8.6% to 5.3%). Uncemented implants are used less frequently (7.7% to 4.4%). Unicompartmental knee are implanted more frequently (2.8% to 4.1%). In 2002, 10% of the joints implanted had mobile bearings, 8% were PCL sacrificing, 28% had the patella resurfaced, 0.4% were patellofemoral replacements.

The percentage of patients satisfied with their arthroplasty one-year post-surgery has not improved over the 13 years. (O.A 81%, R.A 87.5% , trauma 63%). The incidence of reoperation (for all causes) in the first year is 1.2% (0.5% infection). The incidence of revision for implant failure appears to be increasing. Outcome data for revision surgery is being collected at present.

Conclusion: Over the past 13 years there has been a marked increase in knee arthroplasty surgery and significant demographic trends. However, validated outcome measures have not shown an improvement of outcome over this period.

The Trent Arthroplasty Audit Group (TAAG) has been prospectively collecting data on primary knee arthroplasty since 1990 and revision procedures since 1992. To date details of 27 000 primary and 1300 revision knee arthroplasties, from both teaching and district general hospitals in the Trent region, have been contributed.

This provides us with a unique opportunity to look at the outcome of total knee arthroplasty across an English region, at 10 or more years of follow-up.

A questionnaire was designed, with assistance from the department of epidemiology and public health, including both generic (EuroQol) and knee specific questions. Patients who underwent primary knee arthroplasty from 1990–92 were included in the study (4420 patients). Surviving patients (65%) were traced using the NHSIA patient tracing service and the questionnaire was administered by post with a response rate was 55%.

The mean length of follow-up was 11 years 7 months and the mean age at the time of follow-up was 78 years. The vast majority (86%) of patients were satisfied with their joint replacement. Although 52 % took prescribed analgesia for pain in their knee, 22% still complained of pain often or all the time. 18% felt their knee totally or greatly interfered with their usual activities. 14% of patients reported re-operation on their knee, of which 11% were revision procedures.

This suggests that a significant number of patients experience pain and/or disability at 10 or more years following knee replacement. Since most GPs have a high threshold for re-referral in this group of patients, it is likely that the degree of morbidity has previously been underestimated in the orthopaedic community.