We identified 148 patients who had undergone a revision total knee replacement using a single implant system between 1990 and 2000. Of these 18 patients had died, six had developed a peri-prosthetic fracture and ten had incomplete records or radiographs. This left 114 with prospectively-collected radiographs and Bristol knee scores available for study. The height of the joint line before and after revision total knee replacement was measured and classified as either restored to within 5 mm of the pre-operative height or elevated if it was positioned more than 5 mm above the pre-operative height. The joint line was elevated in 41 knees (36%) and restored in 73 (64%).

Revision surgery significantly improved the mean Bristol knee score from 41.1 (sd 15.9) pre-operatively to 80.5 (sd 15) post-operatively (p < 0.001). At one year post-operatively both the total Bristol knee score and its functional component were significantly better in the restored group than in the elevated group (p < 0.01). Overall, revision from a unicondylar knee replacement required less use of bone graft, fewer component augments, restored the joint line more often and gave a significantly better total Bristol knee score (p < 0.02) and functional score (p < 0.01) than revision from total knee replacement.

Our findings show that restoration of the joint line at revision total knee replacement gives a significantly better result than leaving it unrestored by more than 5 mm. We recommend the greater use of distal femoral augments to help to achieve this goal.

We report the mid-term results of a new patellofemoral arthroplasty for established isolated patellofemoral arthritis. We have reviewed the experience of 109 consecutive patellofemoral resurfacing arthroplasties in 85 patients who were followed up for at least five years.

The five-year survival rate, with revision as the endpoint, was 95.8% (95% confidence interval 91.8% to 99.8%). There were no cases of loosening of the prosthesis. At five years the median Bristol pain score improved from 15 of 40 points (interquartile range 5 to 20) pre-operatively, to 35 (interquartile range 20 to 40), the median Melbourne score from 10 of 30 points (interquartile range 6 to 15) to 25 (interquartile range 20 to 29), and the median Oxford score from 18 of 48 points (interquartile range 13 to 24) to 39 (interquartile range 24 to 45). Successful results, judged on a Bristol pain score of at least 20 at five years, occurred in 80% (66) of knees. The main complication was radiological progression of arthritis, which occurred in 25 patients (28%) and emphasises the importance of the careful selection of patients. These results give increased confidence in the use of patellofemoral arthroplasty.

There have been several reports of good survivorship and excellent function at ten years with fixed-bearing unicompartmental knee replacement. However, little is known about survival beyond ten years.

From the Bristol database of over 4000 knee replacements, we identified 203 St Georg Sled unicompartmental knee replacements (174 patients) which had already survived ten years. The mean age of the patients at surgery was 67.1 years (35.7 to 85) with 67 (38.5%) being under 65 years at the time of surgery. They were reviewed at a mean of 14.8 years (10 to 29.4) from surgery to determine survivorship and function. There were 99 knees followed up for 15 years, 21 for 20 years and four for 25 years. The remainder failed, were withdrawn, or the patient had died.

In 58 patients (69 knees) the implant was in situ at the time of death. Revision was undertaken in 16 knees (7.9%) at a mean of 13 years (10.2 to 21.6) after operation. In seven knees (3.4%) this was for progression of arthritis, in three (1.5%) for wear of polyethylene, in four (2%) for tibial loosening, in two (1%) for fracture of the femoral component and in two (1%) for infection. Two knees (1%) were revised for more than one reason.

The mean Bristol knee score of the surviving knees fell from 86 (34 to 100) to 79 (42 to 100) during the second decade. Survivorship to 20 years was 85.9% (95% CI 82.9% to 88.9%) and at 25 years was 80% (95% CI 70.2% to 89.8%). Satisfactory survival of a fixed-bearing unicompartmental knee replacement can be achieved into the second decade and beyond.

Following total hip arthroplasty (THA) and total knee arthroplasty (TKR) only the ‘visible’ measured blood loss is usually known. This underestimates the ‘true’ total loss, as some loss is ‘hidden’. Correct management of blood loss should take hidden loss into account.

We studied 101 THAs and 101 TKAs (with re-infusion of drained blood). Following THA, the mean total loss was 1510 ml and the hidden loss 471 ml (26%). Following TKA, the mean total loss was 1498 ml. The hidden loss was 765 ml (49%). Obesity made no difference with either operation.

THA involves a small hidden loss, the total loss being 1.3 times that measured. However, following TKA, there may be substantial hidden blood loss due to bleeding into the tissues and residual blood in the joint. The true total loss can be determined by doubling the measured loss.

Nine patients underwent arthrodesis of the knee using a customised coupled nail (the Mayday arthrodesis nail), five after infected arthroplasty, one following failed arthrodesis, one for intractable anterior knee pain, one for Charcot instability and one after trauma. Comparison was made with 17 arthrodeses, eight undertaken using external fixation, four with dual compression plates, and five with long Küntscher nails. Union was achieved in all patients (100%) at a mean time of ten months using the customised implant. There were no complications despite early weight-bearing. No further procedures were required. This contrasted with a rate of union of 53% and a complication rate of 76% with alternative techniques. Of this second group, 76% required a further operative procedure.

We compared the Mayday arthrodesis nail with other techniques of arthrodesis of the knee. The differences in the need for further surgery and occurrence of complications were statistically significant (p < 0.001), and differences in the rate of nonunion and inpatient stay of less than three weeks were also significant (p < 0.05) using Fisher’s exact test.

We conclude that a customised coupled intramedullary nail can give excellent stability allowing early weight-bearing, and results in a high rate of union with minimal postoperative complications.

We describe 88 knees (79 patients) with lateral unicompartmental osteoarthritis which had been treated by the St Georg Sled prosthesis.

At a mean follow-up of nine years (2 to 21) 15 knees had revision surgery, nine for progression of arthritis, six for loosening, four for breakage of a component and four for more than one reason. Six patients complained of moderate or severe pain at the final follow-up. Only five knees were lost to follow-up in the 21-year period.

We performed survivorship analysis on the group using revision for any cause as the endpoint. At ten years the cumulative survival rate was 83%, and at 15 years, when ten knees were still at risk, it was 74%.

Based on our clinical results and survival rate the St Georg Sled may be considered to be a suitable unicompartmental replacement for isolated lateral compartment osteoarthritis.

We have studied the long-term outcome of 408 primary medial St George Sled unicompartmental arthroplasties of the knee and 531 primary Kinematic total knee arthroplasties using survivorship analysis. The operations were performed by a number of surgeons under the supervision of two consultants at one orthopaedic centre. Prospective clinical assessment was carried out before and at 2, 5, 8, 10, 12 and 15 years after operation. Failure was defined as follows: revision or removal of the implant; the presence of moderate or severe pain; or ‘worst-case’ with all patients lost to follow-up.

Cumulative survival rates at ten years were calculated using life tables. The follow-up rate was 97%. At ten years, 25 medial sled arthroplasties and 20 Kinematic knee arthroplasties had been revised. With revision or removal as the survivorship endpoint at ten years there was a success rate of 87.5% for the medial sled and 89.6% for the Kinematic knee arthroplasty. When moderate or severe pain was included these rates became 79.4% for both arthroplasties. There was no statistically significant (p > 0.05) difference between the rates of survival for the two arthroplasties using either of the endpoint criteria.

Good or excellent results were recorded for 77.9% of the medial sled knees and 75.1% for the Kinematic knees. The former had 93.8% of cases with a final range of movement in excess of 90° compared with 83.7% for the Kinematic knees (p < 0.01). We conclude that at a single orthopaedic centre in the UK, the St Georg Sled medial compartment arthroplasty for appropriate specific indications offers predictable survivorship at ten years which is comparable with that of the Kinematic total knee arthroplasty.

We present a prospective review of the outcome of 76 Lubinus patellofemoral arthroplasties carried out in 59 patients between 1989 and 1995. At a mean follow-up of 7.5 years, 62 knees in the 48 patients were reviewed; 11 patients (14 knees) had died. None was lost to follow-up.

The clinical outcome using the Bristol Knee Scoring system was satisfactory in 45% of the cases. Maltracking of the patella, resulting in lateral tilt, subluxation and polyethylene wear, was the most common complication (32%). Revision surgery was carried out in 21 knees (28%) giving a cumulative survival rate of 65% (confidence interval (CI) 49 to 77) at eight years. The survival rate for revision and moderate pain was 48% (CI 36 to 59) at six years. Progression of arthritis was seen in seven cases (9%). In five of these (6.5%), the symptoms were severe enough to need revision surgery. Due to the high proportion of unsatisfactory results, we have discontinued the use of this prosthesis.

Failure of a unicompartmental knee replacement (UKR) may be caused by progressive osteoarthritis of the knee and/or failure of the prosthesis. Limb alignment can influence both of these factors. We have examined the fate of the other compartments and measured changes in leg alignment after UKR.

A total of 50 UKRs was carried out on 45 carefully selected patients between 1989 and 1992. At operation, deliberate attempts were made to avoid overcorrection of the deformity. Four patients died, one patient was lost to follow-up and two knees were revised before review which was at a minimum of five years. Standard long-leg weight-bearing anteroposterior views of the knee and skyline views of the patellofemoral joint were taken before and at eight months and five years after operation. The radiographs of the remaining 43 knees were reviewed twice by blind and randomised assessment to measure the progression of osteoarthritis within the joints. Overcorrection of the deformity in the coronal plane was avoided in all but two knees. Only one showed evidence of progression of osteoarthritis within the patellofemoral joint, and this was only identified in one of the four assessments. Deterioration in the state of the opposite tibiofemoral compartment was not seen. Varus deformity tended to recur. Recurrent varus of 2° was observed between eight months and five years after operation. There was no correlation between the postoperative tibiofemoral angle and the extent of recurrent varus recorded at five years. Changes in alignment may be indicative of minor polyethylene wear or of subsidence of the tibial component.

The incidence of progressive osteoarthritis within the knee was very low after UKR. Patients should be carefully selected and overcorrection of the deformity be avoided.

Patella infera may occur after reconstruction of the anterior cruciate ligament (ACL), high tibial osteotomy and total knee replacement (TKR). Restriction of movement of the knee and pain may result. Our aim was to compare the incidence and to assess the effects of patella infera after TKR and unicompartmental knee replacement (UKR).

We reviewed radiographs of the knees of 84 patients who had had either TKR or UKR as part of a randomised, controlled trial. The length of the patellar tendon was measured on serial radiographs taken before, at eight months and at five years after operation.

There was no significant change in the length of the patellar tendon after UKR, but a significant reduction was observed after TKR. Five years after the operation, the shortening of the tendon had increased to a mean of 3.5 mm. Of the knees with TKR reviewed at five years, 34% developed patella infera, defined as 10% or more of shortening, compared with 5% of those with UKR. Shortening was greatest in those knees which had required a lateral release; in this subgroup the mean shortening was 7.2 mm. Shortening correlated with restriction of movement and pain in the knee.

Our study has shown that patella infera develops in most patients after TKR with lateral release, and in approximately 25% of patients after TKR without this additional procedure. Patella infera rarely occurs after UKR. It is associated with restriction of movement and pain in the knee. It may be an effect of the more extensive exposure required to perform TKR and may, in part, explain the better clinical results of UKR.

We randomised 102 knees suitable for a unicompartmental replacement to receive either a unicompartmental (UKR) or total knee replacement (TKR) after arthrotomy. Both groups were well matched with a predominance of females and a mean age of 69 years.

Patients in the UKR group showed less perioperative morbidity, but regained knee movement more rapidly and were discharged from hospital sooner.

At five years, two UKRs and one TKR had been revised; another TKR was radiologically loose. All other knees appeared to be clinically and radiologically sound.

Pain relief was good in both groups but the number of knees able to flex ≥ 120° was significantly higher in the UKR group (p < 0.001) and there were more excellent results in this group.

Our findings have shown that UKR gives better results than TKR and that this superiority is maintained for at least five years.

We have carried out a randomised, controlled trial on 70 patients having unilateral total knee replacement in which transfusion was either with homologous bank blood or by reinfusion of unwashed blood salvaged after operation.

No complications or adverse effects were observed from reinfusion. The need for bank blood was reduced by 86% in the reinfusion group but, more importantly, the number of infective episodes was significantly less when the use of bank blood was avoided. The mean length of stay in hospital was also reduced by more than two days.