The aim of this study was to document the survivorship and patient reported outcome of the Avon patello-femoral replacement in a consecutive series with follow up of 10 years or more.

All cases performed in Bristol from 1996 onwards were prospectively recorded. Follow up was at 1,2,5,7,10,12 and 15 years with the Bristol Patella Score, the Oxford and WOMAC scores and SF12. Implant survival was analysed using the Kaplan-Meier method.

There were 323 PFJ replacements (280 individuals). Follow up was available for 286 cases in 250 patients (89% follow up). The 10 year survival rate was 77%, falling to 67% at 15 years. The most common reason for revision was tibio-femoral progression (45/74 revisions), with loosening or polyethylene wear recorded in 8 cases. The best results were seen in the youngest and the oldest patients. Good improvements were seen in PROMs, with the mean OKS improving from 19.5 to 34.1 at 2 years and 32.7 at the 15 years.

The Avon patello-femoral knee replacement is a successful long-term treatment for isolated patello-femoral knee osteoarthritis, although further improvements are expected in subsequent series, particularly as indications for surgery have evolved over time.

Introduction

The Bristol Knee Group has prospectively followed the results of over 500 isolated patellofemoral arthroplasties. Initial experience with the Lubinus prosthesis was disappointing. The main causes of failure were mal-tracking and instability leading to excessive polythene button wear and disease progression in the tibio femoral joint.

This experience resulted in the design of a new prosthesis to correct the tracking problems and improve the wear. We have now performed over 425 Avon arthroplasties with a maximum follow-up of 9 years.

Although good long term results for fixed bearing uni-compartmental knee replacements (UKRs) have been reported mobile bearings predominate in some parts of the world. Three prospective studies have been undertaken comparing the short and medium term outcomes of fixed and mobile UKRs.

A 5 year comparative cohort study of 47 Oxford mobile bearing and 57 St Georg Sled fixed bearing UKRs.

In all groups the preoperative sex mix, average age and knee scores were extremely similar.

All patients were assessed both pre and postoperatively by a research nurse and radiographs were taken; the results were entered on the Bristol Knee database.

Results:

Multiple problems were encountered, perhaps because of the introduction of MIS, but at 5 years 11 Oxford and 4 Sleds had failed. The major problem with the mobile bearing implant was instability though tibial fractures were also seen. Both groups had three cases of arthritic progression and loose cement was seen twice in the fixed bearing group.

– Amongst the remaining patients the median scores for the Sled were better. Bristol Knee Score (Max 100) 95:90; Oxford (Max 48) 39:37; and reduced WOMAC (Best score 12) 18:24.

Conclusion. Although theoretically mobile bearings will give greater longevity there is frequently a short term price to pay particularly when the procedures are performed by inexperienced surgeons and trainees. Since the wear properties of polyethylene have improved in recent years these studies suggest that a fixed bearing option can be used more safely in many patients.

Introduction: The Bristol Knee group has performed over 600 patellofemoral arthroplasties in the last 18 years. Experience with the Lubinus prosthesis led to the development of the Avon Prosthesis. In the last 11 years we have prospectively recorded the results of over 470 consecutive Avon arthroplasties. The main cause of the 9.5% failure is arthritic disease progression in the tibiofemoral compartments.

Patients and Method: Over the last 11 years we have identified 21 patients from our own series and from tertiary referrals that have persistent pain, which was mainly due to technical error rather than arthritic disease progression. The causes can be classified into three main reasons: First, an incorrect anterior cut in the saggital plane which was cut in either too much flexion or extension. Second, the anterior cut had inadequate external rotation, which should be 3–6 degrees to lateralise the groove and facilitate correct tracking. Third, the prosthesis was oversized in several cases leading to retinacular impingement.

Results: The overall results followed up to 10 years showed excellent and consistent improvement in both pain and function as judged by the WOMAC 12 scale. Of those patients with persistent pain, seven had the femoral component revised to either resize the prosthesis or revise the alignment of the anterior resection and correctly inset the prosthesis, with good results. Ten cases were revised to a total knee replacement. In the remaining cases, two had an Insall realignment, one a patella distalisation, and in one no treatment was required. As a permanent solution new instruments were designed to reduce the incision size and increase the accuracy of the saggial alignment and to create an exact amount of external rotation.

Conclusion: The lessons from 11 years experience with the Avon arthroplasty has led to the development of improved instrumentation which should reduce the failures.

Purpose of Study: Fixed bearing unicompartmental knee replacement (UKR) has become popular since several series have now shown good 10 year survivorship and excellent function. However little is known about survival during the second decade.

Method: From the Bristol database of over 4000 knee replacements 203 St. Georg Sled UKR’s which had already survived 10 years were identified. The mean age at surgery was 67 years (48–85), with 64% being female. This cohort has been further reviewed at an average of 14.8 years (10–30) from surgery to determine survivorship and function.

Results: Survivorship during the second decade was 87.5%. 58 patients (69 knees) had died with implant in situ and only 2 after revision. A further 15 UKR’s have been revised at an average of 13 years post op; 7 for progression of disease in the opposite compartment, 4 for tibial loosening, 3 for polyethylene wear, 2 for femoral component fracture and 2 for infection.

99 knees were followed for 15 years, 21 knees for 20 years and four for 25 or more years. The average Bristol knee score of the surviving knees fell from 86 to 79 during the second decade, largely as a result of aging.

A previous study of the St Georg. Fixed bearing UKR showed an 89% 10 year survivorship and this is now extended to 82% at 15 years and 76.5% at 20 years.

Conclusion: Satisfactory survival of fixed bearing UKR can be achieved in the second decade suggesting that the indications for mobile bearings require careful definition since there is a higher incidence of complications in many people’s hands.

Since 1996 over 400 isolated patellofemoral replacements using the Avon prosthesis have been undertaken in Bristol. As a result of the usually gratifying outcomes seen in patients over 55 years old, the indications for this surgery were soon extended to include the many younger patients who have severely disabling patellofemoral disease. We present results from a prospective cohort study of patients under 55 years of age at the time of surgery.

110 knees in 86 patients (median age 47 years, range 24–54) have been treated with Avon patellofemoral replacement (88 in females and 22 in males). Diagnoses included lateral facet OA (59 knees), patella dislocation (25 knees), trochlear dysplasia (41 knees) and post patellectomy instability (7 knees). 79 knees had undergone previous surgery. 17 knees required additional intra-operative procedures including 11 lateral releases and 2 patella realignments. All patients were assessed pre-operatively and at regular intervals using the Oxford, Bristol and WOMAC scores.

All knees were scored preoperative and only one knee has been lost to follow-up due to the patient’s death, which was unrelated to surgery. Post-operative Oxford knee scores have been obtained for 106 knees with follow-up between 8 months and 8 years (mean follow-up 24 months). The mean Oxford score improved from 18 preoperatively to 31 at latest review. Bristol and WOMAC scores showed similar improvements. 16 knees required post-operative additional procedures including 6 lateral releases, 3 patella realignments and 11 revisions. Of the revisions 5 were for progression of tibiofemoral OA but none of these were knees with trochlear dysplasia. Equally good mean scores were seen when comparing patients with the 3 main underlying pathologies (trochlear dysplasia, patellar dislocation and lateral facet OA). At least 37% of the patients studied had pre-existing trochlear dysplasia and the majority of these patients report onset of symptoms, often patellar dislocation, in the first three decades of life. More than 90% of patients were overweight or obese according to their BMI at the time of surgery.

Many of this type of patient, with disabling symptoms, wish to “live now”. The short-term improvements are frequently dramatic. As yet there is no suggestion of prosthetic failure. Revision presents little difficulty since minimal bone is resected in the primary proceedure. Radiological deterioration of the tibio femoral joint is seen in some cases of primary OA but not with trochlear dysplasia.

Functional outcome after patellofemoral joint replacement (PFA) for osteoarthritis remains inconsistent. It is believed that functional outcome for joint replacement is dependent upon postoperative joint kinematics. Minimal disruption of the native joint, as in PFA, should produce more normal kinematics and improved outcome. No previous studies have examined joint kinematics after isolated PFA.

Aim: To investigate the sagittal plane kinematics of patellofemoral replacement and compare with the normal knee.

Twelve patients who had undergone successful PFA at least two years previously were recruited. Patients performed flexion/extension against gravity, and a step up. Video fluoroscopy of these activities was used to obtain the Patellar Tendon Angle (PTA), the angle between the long axis of the tibia and the patella tendon, as a function of knee flexion. This is a previously validated method of assessing sagittal plane kinematics of a knee joint. The kinematic profile of the PFA joints was compared to the profiles for fourteen normal knees.

Overall, the kinematic plot obtained for PFA reflected similar trends to that for normal knees; but the PTA was slightly but significantly increased throughout the entire range of flexion (two degrees). This is equivalent to an average displacement of the lower pole of the patella of 1.5mm.

Sagittal plane knee kinematics after PFA are much more normal than after TKR and this should give improved functional outcome. The observed increase in PTA through range may result from increased patella thickness or a shallow trochlear groove and may influence patellofemoral contact forces.

Introduction Unicompartmental Knee Replacement (UKR) has now become an accepted and widely used treatment for unicompartmental arthritis. Our unit has performed over 1000 UKRs in the past 22 years. The optimal mechanical design of the implant has yet to be determined.

Methods After gaining ethical approval a prospective randomised trial was commenced in 1999 to compare the 2–5 year results of a fixed bearing with a mobile bearing prosthesis. 104 knees in 91 patients underwent a UKR, the mean age of the group was 65 years and a mean weight of 80kg. 57 had a St Georg Sled fixed bearing prosthesis and 47 an unconstrained mobile bearing Oxford UKR. All were prospectively reviewed using the Oxford and Bristol Knee Scores.

Results All 104 knees have been reviewed at 2 years, with none lost to follow-up. 3 patients in the Oxford group suffered a dislocated meniscus and a further 4 required revision, as well as 3 in the St Georg Sled group. The overall function of the 2 groups was the same, but the Oxford mobile bearing group had significantly more persistent pain (p=0.013).

Conclusion The results in both groups were less satisfactory than previous series from this unit probably due to the efforts being made to use minimal incision. However the early complication rate was higher with the mobile bearing devise. This must be balanced against the possible better long-term survival.

Introduction: Isolated patello-femoral arthritis occurs in up to 10% of patients suffering osteoarthritis of the knee. Reports of patello-femoral replacements have given indifferent results. We report our experience of 350 cases of the Avon Arthroplasty.

Method: In 1994 after experience with the Lubinus prosthesis which had a 50% failure rate at eight years, we defined the design criteria for a new prosthesis. Since September 1996, 350 knees have been treated with this design. Prospective review was undertaken and 150 knees have reached 2 years and 80 knees are at five years with 10 knees at eight years. The outcome was assessed using pain scores, Bartlett’s patella score and the Oxford knee score.

Results: The patients recovered function rapidly and 20 cases (6%) suffered early complications which resolved. The median pain score improved from 15/40 points to 38 at five years. The movement increased from 114 to 120 at five years. The Bartlett patella score improved from 11/30 points to 25 at five years. The Oxford knee score improved from 19/48 points to 40 points at five years.

The functional results are similar or better than those of a total knee replacement. Fourteen patients developed mal-alignment (4%) two of which required distal realignment. There have been no cases of deep infection, fracture, wear or loosening. Twenty seven knees (7%) developed evidence of disease progression, twenty two of which (6%) have required revision to a total knee replacement. Nineteen patients (5%) complained of some persistent anterior knee pain of uncertain cause.

Conclusions: Results to date show a large improvement in pain and function and this improved design has reduced the problems of mal-alignment and polythene wear. There is a low complication rate and an excellent range of movement. Disease progression remains a problem which is not predictable. This type of prosthesis offers an alternative to total knee replacement in this small group of patients with isolated patello-femoral disease with a low morbidity.

Introduction. There has been a recent major increase in the use of unicompartmental knee replacement (UKR) but few studies exist comparing its long term efficiency with total knee replacement (TKR)

Method. Ethical approval was obtained.

Between 1989 and 1992 a randomised prospective study was undertaken in which 102 cases adjudged suitable for UKR were allocated to receive either a St Georg Sled UKR or a Kinematic Modular TKR. Both cohorts had a median age of 68 and a similar sex distribution and preoperative knee score. Regular follow up has been maintained. As reported the early results favoured UKR. All cases have now been assessed after a minimum of 10 years using modified WOMAC, Oxford and Bristol Knee Scores (BKS) as well as radiographs.

Results. 28 cases had died with their knees known to be intact, 2 cases were untraceable. 33 knees in each group were reviewed. 3 in each group had been revised, no impending failures were identified on the radiographs.

At 10 years the UKR group had better Oxford and WOMAC scores as well as significantly more excellent results (19:14) and fewer fair and poor results on the BKS. Both groups averaged over 105′ of flexion but 61% of the UKR and only 15% of the TKR group had 120′ or more of flexion.

Conclusion. The faster rehabilitation and better early results with UKR are maintained for 10 years with few failures occurring. The average BKS of the UKR group only fell from 91 to 88 between 5 and 10 years suggesting minimal evidence of functional deterioration in either the prosthesis or the remainder of the joint.

Introduction and Aims: Single compartmental replacement procedures are increasingly preferred over total knee replacement (TKR) for single compartment osteoarthritis of the knee joint. Theoretically, reduced disruption of the native joint should produce more normal kinematics. This study aimed to describe and compare the sagittal plane kinematics of four different, commonly used devices.

Method: Four groups of patients who had undergone successful single compartment replacement at least two years previously were recruited. Fifteen following Oxford medial UKA, 12 following medial St Georg Sled UKA, five following Oxford lateral UKA, and 12 following Avon PFJ replacement. Patients performed flexion/extension against gravity, and a step-up during video fluoroscopy. The Patellar Tendon Angle (PTA), the angle between the long axis of the tibia and the patella tendon, was obtained as a function of knee flexion. This relationship provides indication of sagittal movement between femur and tibia through range and has been validated as a reliable measure of joint kinematics.

Results: The kinematic profile for each group was compared to that of the profile for 12 normal and 30 TKR (AGC) knees. All three tibiofemoral devices produced knee kinematics similar to the normal knee. The PTA was found to have a linear relationship to flexion angle, decreasing with increasing knee flexion angle. No such linear relationship exists for the TKR joint, which display abnormal kinematics. The PF device also reflected similar trends to that for normal knees except that the PTA was moderately increased throughout the entire range of flexion (three degrees).

Conclusion: In contrast to TKR, all single compartmental knee replacements provided kinematics similar to the normal joint. The kinematic pattern of the PFJ replacement may be of most interest as the observed increase in PTA through range could influence patello-femoral contact forces

Introduction: Uni-compartmental Knee Replacement (UKR) has now become an accepted and widely used treatment for uni-compartmental arthritis. Our unit has performed over 1000 UKRs in the past 22 years. The optimal mechanical design of the implant has yet to be determined.

Method: A prospective trial was commenced in 1999 to compare the short-term results of a fixed bearing with a mobile bearing prosthesis. One hundred and three knees in 95 patients underwent a UKR. Fifty-three had a St Georg Sled fixed bearing prosthesis and 50 had unconstrained mobile bearing Oxford UKR. All were prospectively reviewed using the Oxford and Bristol Knee Scores.

Results: Eighty-four knees have been reviewed at two years with one lost to follow-up. One fixed (2%) and six mobile (12%) bearing knees had been revised. Three fixed bearing and eight mobile bearing knees were dissatisfied because of stiffness or persistent pain and seven mobile bearing knees had undergone further minor procedures, usually for dislocation. At two years, the Bristol and Oxford knee scores for fixed bearing were better than the mobile bearing knees at 90:83 /100 and 38:33 /48 respectively.

Conclusion: There were more early complications in the mobile bearing UKR group, but the functional results were slightly better than in the fixed bearing knees.

Introduction and Aims: There has recently been an increase in the use of uni-compartmental knee replacement (UKR), but there are few studies comparing its long-term efficacy and survival with total knee replacement (TKR). We reported the five-year results of a randomised trial which showed that uni-compartmental replacement had a faster rehabilitation and better functional results than total replacement. We now report the 10-year results.

Method: After obtaining ethical approval, between 1989 and 1992, 102 cases judged suitable for UKR were randomly allocated to receive either a St Georg Sled UKR or a Kinematic Modular TKR. Both cohorts had a median age of 68 and a similar sex distribution and pre-operative knee scores. Regular follow-up has been maintained. All cases have now been assessed after a minimum of 10 years using modified the WOMAC, Pain, Oxford and Bristol Knee (BSK) scoring systems as well as radiographs.

Results: At 10 years, 29 cases had died with their knees known to be intact, two totals cases were untraceable. Thirty-two uni`s, and 34 totals were available for review. Three uni`s (6%) and two totals (4%) had been revised; there was one impending total failure on the radiographs.

The UKR group had better scores with Oxford: 38 v 34 /48 and WOMAC: 17 v 21 /60 and more excellent results (19 v 14) and fewer fair and poor results on the BKS (4 v 6). The range of movement improved in UKA`s from 107 degrees to 117 degrees, whereas the range decreased in TRK`s from 107 degrees to 104 degrees. Sixty-one percent of the UKR and only 16% of the TKR group had more than 120 degrees of flexion.

Conclusion: The St Georg Uni-compartmental arthroplasty continues to function well at 10 years, with better functional scores and a better range of movement than the Kinematic Total arthroplasty. There were similar failures in both groups.

Introduction: Isolated patello-femoral arthritis occurs in up to 10% of patients suffering osteoarthritis of the knee. Reports of patello-femoral replacements have given indifferent results so we wish to report our experience of 326 cases using two different protheses.

Method: In 1989 we started a prospective series of the Lubinus prosthesis. The results of 76 cases at eight years showed a 50% failure rate. The main reasons for failure were malalignment, wear and disease progression. The Avon patello-femoral arthroplasty was designed to solve some of these problems.

Since September 1996, 250 knees have been treated. Prospective review was undertaken and 120 knees have reached two years and 40 are at five years. The outcome was assessed using pain scores, Bartlett’s patella score and the Oxford knee score.

Results: The median pain score improved from 15/40 points to 35 at five years. The movement increased from 114 to 120 degrees at five years. The Bartlett patella score improved from 10/30 points to 26 at five years. The Oxford knee score improved from 19/48 points to 40 points at five years.

The functional results are similar to those of a total knee replacement. Two patients developed malalignment (1%) one of which required distal soft tissue realignment. There have been no cases of deep infection, fracture, wear or loosening. Sixteen knees (6%) developed evidence of disease progression, 14 of which (6%) have required revision to a total knee replacement.

Conclusions: Results to date suggest that this improved design has reduced the problems of malalignment and early wear. There is a low complication rate and an excellent range of movement. Disease progression remains a problem that is not predictable. This type of prosthesis offers an alternative to total knee replacement in this small group of patients with isolated patello-femoral disease with a lower morbidity.

Introduction and Aims: Our experience with over 200 cases of Avon patellofemoral replacement (PFR) in the elderly with severe arthritis has been satisfactory, so we have extended the indications to younger patients with severe symptoms and various pathologies.

Method: Sixty-six knees in 53 patients (10M: 46F) under the age of 55 years (average 47 years) have been treated with an AVON PFR for the following conditions:

PATHOLOGY NUMBER of Knees

Isolated lateral facet OA 34

Failed realignment 12

Persistent subluxation/dislocation 5

Trochlear dysplasia 5

Pure chondral disease 3

Failed carbon fibre implant 3

Post-patellectomy instability 3

Post-traumatic pain 1

All patients were recorded prospectively and have been regularly reviewed using the modified Oxford, Bartlett & Bristol Knee scoring systems. The mean follow-up of the group is 24 months.

Results: No patient has been lost to follow-up, all knee scores improved substantially.

Most of the patients retained their range of flexion and the mean range of movement increased from 112 to 122 degrees. Patients with persistent subluxation were the most dramatically improved. There have been no cases of deep infection, loosening or wear.

Conclusion: Avon PFR provides a reliable short-term solution to some of the more difficult and disabling problems of the patello-femoral joint as long as demonstrable pathology is present.

Aim: To study the sagittal plane kinematics of the Avon patello-femoral replacement (Stryker-Howmedica), PTA.

Introduction: Replacement of the patello-femoral joint for end stage osteoarthritis has previously been associated with inconsistent results. Retention of the cruciate ligaments is likely to be important in maintaining normal kinematics and hence improved functional outcome.

Methodology: Twelve patients who had undergone Avon PFR least two years previously were recruited following ethical approval. American Knee Society, Bristol and Oxford knee scores were obtained. Patients performed open chain flexion and extension against gravity, in addition to closed chain step up. Video fluoroscopy of these activities was used to obtain the Patellar Tendon Angle (PTA), which is the angle between the long axis of the tibia and the patella tendon, at specific angles of knee flexion. This is a previously validated method of assessing the kinematic profile of a knee joint. These measurements were used to determine the kinematic profile of each knee and they were then compared to a group of twelve normal knees.

Results: A one way ANOVA revealed no significant differences between the kinematic profile following Avon PFR and that of the normal knee. All patients had good or excellent knee scores.

Conclusion: The kinematic profile after Avon PFR is similar to that of the normal knee. In contrast all TKRs we have studied have abnormal kinematics, which are associated with abnormal patello-femoral joint loading. This suggests that isolated PFR should have a functional advantage over TKR.

Introduction: Lateral uni-compartmental knee replacement (UKR) is seldom indicated and remains controversial. It has been considered a technically challenging operation with variable success. Aim: The aim of this paper is to assess the results of unicompartmental knee replacement using a þxed noncongruous prosthesis. Methods: Eighty-eight knees in 79 patients with lateral unicompartmental osteoarthritis treated with a þxed bearing, non-congruous, fully unconstrained prosthesis. (St Georg Sled UKR) over a period of 21 years by different consultants and there trainees. Surgical approach and rehabilitation protocol varied over this period but there was no change in the surgical indications. Results: At a mean follow up of 9 years (range 2 to 21 years). Fifteen knees had revision surgery (9 for progression of arthritis and 6 for component loosening, 4 for implant breakage and 4 for more than one reason). Six patients complained of moderate or severe pain at þnal follow up. Only þve knees were lost to follow up in the 21-year period. Survivorship analysis was performed on the group using revision for any cause as the end point. At ten years the cumulative survival rate was 83% and at 15 years, when ten knees were still at risk, the cumulative survival rate was 74%. Conclusion: Based on our clinical results and survival rate the St Georg Sled UKR can be considered a suitable treatment for isolated lateral compartment osteoarthritis.

Isolated patello-femoral arthritis occurs in up to 10% of patients suffering osteoarthritis of the knee. Previous reports of several different patello-femoral designs have given indifferent results. The Lubinus prosthesis has a reported 50% failure rate at eight years in a study of 76 cases. The main reasons for failure were mal-alignment, wear, impingement and disease progression. The Avon patello-femoral arthroplasty was designed to solve some of these problems.

The first cases were implanted in September 1996 and entered into a prospective review. The outcome was assessed using pain scores, Bartlett’s patella score and the Oxford knee score. To date 207 knees have been treated and 95 knees have reviewed at two to five years. The median pain score improved from 15/40 points to 35 at five years. The movement increased from 114° to 120° at five years. The Bartlett patella score improved from 10/30 points to 26 at five years. The Oxford knee score improved from 19/48 points to 40 points at five years. Two patients developed mal-alignment (1%) one of which required distal soft tissue realignment. There have been no cases of deep infection, fracture, wear or loosening. Fifteen knees (7%) developed evidence of disease progression, twelve of which (6%) have required revision to a total knee replacement. The functional results are similar to those of a total knee replacement.

Results to date suggest that this improved design has all but eliminated the previous problems of mal-alignment and early wear. The functional results are as good as or better than those of a total knee replacement. There is a low complication rate and an excellent range of movement. Disease progression remains a potential problem. This type of prosthesis offers a reasonable alternative to total knee replacement in this small group of patients with isolated, early patello-femoral disease.

The Bristol Knee Replacement Registry was established in the 1970’s and contains prospectively recorded data on 3024 patients. The present study examines the group of 812 patients for whom complete pre-operative and five year post-operative data is available in order to relate their eventual health status to the pre-operative demographic data and to disease severity. The group comprised 593 women and 219 men who had undergone either Kinematic, Total or Sled unicompartmental knee replacement.

Pre-operatively, the average American Knee Society Score (AKSS) was 89 with the elderly, rheumatoid patients and women having significantly lower scores.

Five years later the average AKSS had risen to 161 with patients of all ages, (including the over 80’s) gaining considerably. However, the rheumatoid patients remained more disabled as did women who had a final AKSS of 157 as opposed to 171 for the men (p< 0.01). In addition, a statistically significant finding was that those with the lowest pre-operative scores also had the lowest 5 year scores - they never catch up.

173 patients underwent bilateral knee replacement at separate times. Their pre-operative AKSS was significantly higher at the time of the second knee replacement (90) than the first (82) p< 0.01.

We conclude that since patients present earlier for their second TKR and those with most disability fail to catch up the procedure should be performed earlier in the natural history of the disease, especially in women.