The Oxford unicompartmental knee arthroplasty:
A review from an independent centre

K.A. GIANNIKAS, R.A. BUXTON
Orthopaedic Department, Victoria Hospital, Scotland

Mailing address:
K. Giannikas FRCS
312 Wakefield Road
Stalybridge
SK15 3BY
United Kingdom

ABSTRACT
Background: The aim of this study is to evaluate whether the medium-term results produced by the originators of the Oxford Unicompartmental knee replacement can be reproduced in an independent centre.
Methods: A retrospective study of 71 consecutive Oxford unicompartmental knee prostheses in 57 patients, performed by one surgeon between 1987 and 1995, with a mean follow-up of 6.3 years is presented. All operations were performed for osteoarthrosis of the knee by a minimal approach technique, using a short vertical incision over the affected compartment. Of these, 14 patients (17 knees) had meanwhile died. A further 6 patients (9 knees) did not wish to be reviewed or had left the area. The remaining 37 (45 knees) patients were further reviewed, the Oxford Knee score and the American Knee Score were completed and current radiographs were obtained.
Results: Four medial compartment and one lateral compartment arthroplasty were revised. Eight-year cumulative survival was 94.1% for all patients and 95.4% for patients that underwent medial compartment replacement. 88% of the reviewed patients were enthusiastic or satisfied about the operation. Mean flexion for non-revised arthroplasties was 122.2° and mean flexion deformity 3.8°.
Conclusions: The results in our study compare well with equivalent series produced by the originators of the prosthesis, suggesting that, contrary to some reports in the literature, the Oxford Knee Replacement can be performed equally well in the setting of a District General Hospital.

INTRODUCTION
Unicompartmental knee arthroplasty was first described by Marmor in 1973[14]. Early published series were considered unsatisfactory, since they were inferior to the results of Total Knee Replacements[9,12].
However, subsequent improvements in the indications, instrumentation and implants, established the value of unicompartmental replacement for the treatment of localised gonarthrosis[11].
The Oxford Meniscal Bearing Unicompartmental Knee was first introduced in 1978, and was designed to allow large areas of contact between the femoral and tibial component in an attempt to lower the rate of wear and creep of the polyethylene[5]. Between 1985 and 1988 the originators published the criteria for this arthroplasty; giving specific emphasis to the presence of a functional anterior cruciate ligament, the presence of full thickness articular cartilage in the other compartment and passively correctable varus or valgus deformity of the knee depending on the affected compartment[3,4,6].
In 1998, the 10-year survival analysis of the Oxford Unicompartmental Knee Replacement for medial osteoarthrosis was published by their originator[15]. The claimed 97.7% cumulative survival rate can easily match the best results of any tri-compartmental prosthesis, establishing, therefore, the value of the prosthesis for the treatment of medial gonarthrosis of the knee. There has, however, been a lot of concern regarding the reproducibility of these results from other Orthopaedic surgeons. Most of the non-Oxford data regarding the Oxford Unicompartmental Replacement comes from the Swedish registry[10] and the papers derived from it. Knutson et al[10] present a cumulative revision rate for medial Oxford unicompartmental replacement of 12% in 7 years.
Lewold et al[13] in a study comparing the Oxford with the Marmor knee suggests that the Oxford knee is not yet suitable for full-scale use. The authors believe that the outcome is very dependant on patient selection, implantation technique, skill and rehabilitation, not allowing the average surgeon to achieve the originator's results.
This study is a retrospective survey, measuring the outcome of the Oxford Meniscal Bearing Knee over 3 to 11 years, for the treatment of unicompartmental osteoarthrosis, performed by one surgeon using a minimal approach technique and working in a District General Hospital.

1a 1b
Picture 1. Preoperative X-ray of patient with arthrosis of medial part of the knee joint.

2a 2b
Picture 2. 10-year follow-up radiographs of a currently 84-year-old patient. (Same patient of figure 1). Note that appearance of subchondral sclerosis and the marginal osteophytes in the lateral compartment have remained unchanged.

MATERIAL - METHODS
Patients
From 1987 up to and including 1995, 57 patients were identified from the theatre records that had undergone an Oxford knee replacement for osteoarthrosis of the knee, performed by or under the direct supervision of the senior author (R.A.B.). Of these, 43 were unilateral replacement and 14 bilateral; a total of 71 arthroplasties. Thirty-one patients were female and 26 were male. The case notes / microfilms of all these patients were retrieved and the indications, the operative notes and their immediate and long term postoperative course, together with all documented complications were noted.
Unfortunately, for most patients, it was not possible to retrieve preoperative radiographs as it is the policy of the hospital to destroy films older than 5 years. The mean age of the patients at the time of the operation was 71.8 years (range: 4988 years). Follow up ranged between 3 and 11 years (mean 6.3 years).
Two patients had secondary osteoarthrosis due to previous trauma, one of which had required internal fixation of a depressed lateral plateau fracture 6 years before the knee replacement. All the remaining suffered from primary unicompartmental osteoarthrosis.
Of the 57 patients, 14 (17 knees) are deceased. Four patients (6 knees) did not wish to be reviewed, however, following communication with their general practitioner, it was established that they have not received any orthopaedic treatment elsewhere. Finally, 2 more patients (3 knees) had left the area. The remaining 37 patients (45 knees; 39 being medial and 6 being lateral compartment replacements) were reviewed in the clinic. While at the review clinic they were examined by an independent surgeon according to the guidelines of the British Orthopaedic Association1, and the "American Knee Score" (AKS)[8] was completed. The AKS was modified to the evaluation of unicompartmental knee replacements, in that the alignment of the knee, even though calculated, was not added to the score.
All patients had a weight bearing anteroposterior and lateral radiograph of their replaced knee(s). The radiographs were examined for radiolucency around the femoral and tibial components, and for the presence of osteoarthritic changes in the non-replaced compartment and in the patello-femoral joint.

Indications and operative approach
The indication for the operation was based on clinical criteria, which included localised unicompartmental pain unresponsive to other modalities of treatment, associated with a good range of movement. Cases with clinical evidence of laxity of the anterior cruciate ligament were not offered unicompartmental knee replacement. The radiographic appearance of the contralateral compartment was taken into consideration, but a minor degree of sclerosis and marginal osteophytes were ignored as long as they did not influence the clinical presentation. Varus or valgus deformities in all patients were passively correctable.
The operative procedure was performed under tourniquet, and involved a minimal approach technique, with a vertical incision centred over the diseased compartment, enough to give adequate exposure. The incision performed was approximately 2-3 cm longer than the one currently advocated by the Oxford group[17], as it was felt that this allows a better exposure of the diseased compartment. The patella was retracted and not averted, permitting only partial inspection of the contralateral compartment. All patients had the Phase II prosthesis and instrumentation. Care was taken during the preparation of the femur to avoid anterior impingement of the polyethylene. During the early years a drain was used, which was abandoned during the subsequent years. Patients were placed on a continuous passive motion machine, and were encouraged to start weight bearing as soon as they felt confident to cope at home.
Patients were allowed to go home once their immediate postoperative pain had settled and they were able to flex to 90°. No form of prophylactic anticoagulation was used, but T.E.D. stockings were worn for the first two to three weeks.

Statistical analysis
Survival of the prosthesis has been presented in life tables as calculated by Armitage and Berry[2]. Confidence intervals were calculated by the method of Peto et al[16].

RESULTS
The interval between knee symptoms and the knee arthroplasty as documented or questioned during the review averaged 9.6 years (223 years). During this period, 8 knees had undergone an arthroscopy and wash out, while 19 knees had undergone at least one steroid injection. Peri-operatively, in three cases a fracture of the tibial plateau was documented in the notes, one of which required 2 cancellous screws to stabilise the fragments. The remaining two were considered stable once the prosthesis was inserted. These cases were further treated with protected weight bearing and, in the long term, achieved excellent results. Tourniquet time ranged between 50 and 105 minutes (mean: 66 minutes).
In the immediate post-operative period, there was one case of Deep Vein Thrombosis and a further case with non-fatal pulmonary embolism. One patient developed a haemarthrosis that required hospitalisation but settled with conservative treatment. A further eight patients were discharged from the hospital with antibiotics due to wound erythema or persistent wound leakage. No deep infections were documented in this series.
In the long term, mean flexion was 122.2° (range: 90140°), and mean fixed flexion deformity 3.8° (range: 018°). From the interviewed patients who did not require revision surgery, 26 (72%) claimed that they were enthusiastic with the operation, 6 (16%) claimed they were satisfied, 3 (8.3%) were non-committal and 1 (2.7%) was disappointed

MEDIAL COMPARTMENT ARTHROPLASTY
Four knees required revision surgery to a total knee replacement. One revision followed an unnoticed chronic dislocation of the meniscus, which was initially pain-free. A second revision was due to loosening of the tibial component combined with painful progression of lateral compartment osteoarthrosis. There was a further patient that sustained a fracture of the medial plateau following direct injury and eventually required a revision to a TKR. In the fourth case, the patient was referred to the pain clinic due to continuing pain at the medial compartment, and eventually had a revision to a TKR. This patient subsequently developed a haematoma and a superficial infection and is currently disappointed with the TKR. A further patient is disappointed with the outcome. This is an 80-year-old lady who had her arthroplasty 5 years previously, and is currently complaining of diffuse pain around the replaced knee and is wheel-chair bound. Radiographicaly, there is no evidence of loosening of her prosthesis. The patient does not want to proceed with any further surgical intervention due to her currently severely compromised health.
Three patients complained of tenderness in their surgical scar that required local steroid injection and symptoms eventually settled. Finally, a patient developed pain over her medial compartment, which settled after attending the pain clinic.

LATERAL COMPARTMENT ARTHROPLASTY
Of the total of 9 knees that underwent lateral compartment replacement, one required revision surgery to a TKR three months after the primary operation due to repeated dislocation of the meniscus. A further case can be considered as a failure. This was performed 4 years after a severely comminute lateral plateau fracture and is currently 5 years post arthroplasty. Even though the prosthesis is technically satisfactory, the patient is experiencing diffuse pain around her knee, but has not currently agreed to proceed with a revision to a TKR.
The American Knee Score was completed for all interviewed patients that did not undergo revision surgery. Analytical presentation for medial and lateral compartment arthroplasty is given in table I.

RADIOGRAPHIC EVALUATION
Radiographic examination of all reviewed patients, revealed lucency under the peripheral part of the tibial component in 32 knees (78.4%%). In only one patient was there a radiolucent line affecting most of the length of the tibial component. This patient, who had the prosthesis inserted 7 years before, had no pain, a full range of movement, and there was no progression of the radiolucency from a film obtained 3 years before. Two further patients had radiolucent lines, at the posterior aspect of the femoral component, 3 and 8 years post-arthroplasty, without, however any clinical symptoms. Thirty-one knees (75.6%) had minor osteoarthritic changes affecting the contralateral compartment, such as subcondral sclerosis or marginal osteophytes. Twenty-eight knees (68.3%) had osteoarthritic changes affecting the patello-femoral joint. Only two of them complained of some degree of peripatellar pain, which was easily controlled with non-steroid anti-inflammatories.

SURVIVAL ANALYSIS
Table III and table IV present the life tables for the total number of prostheses and for the medial compartment arthroplasties respectively. A life table for the lateral compartment arthroplasties was not created, as the numbers in this series are too limited for consideration. The obtained 95% confidence level permits the calculation of the "worse-case scenario" in which the cases lost to follow-up subsequently required revision surgery. We consider the 8th year cumulative survival the longest reliable; as the small numbers of cases thereafter influence considerably the 95% confidence levels.

DISCUSSION
This is the only one-surgeons reported series of Oxford Unicompartmental prosthesis, as well as the only series form a centre in which the minimal approach is used without the need to dislocate the patella. This approach has only recently been reported in the Orthopaedic literature by the Oxford group[17], according to whom, the early rehabilitation period is shortened significantly without affecting the long-term survivorship of the prosthesis. The results in our series support this evidence.
Regarding the medial compartment arthroplasty, there are only two papers with long-term results reported in the literature. In 1998 the originators reported a series of 144 prosthesis with mean follow-up of 7.3 years and predictive cumulative survival at 10 years of 97.7 (95% CI: 92.7 to 100.0%)15. In 2001 an independent retrospective study from Sweden was published investigating 124 prosthesis, in which all cases had been followed-up more than 10 years (mean: 12.5 years)[18]. They claim a 10-year cumulative survival of 95.0% (95% CI: 90.899.3%). These two studies are potentially comparable as they both used similar inclusion criteria. In both studies patello-femoral arthrosis was not considered a contraindication.
Our series is not strictly comparable with the above as the selection criteria for the patients were slightly wider than the ones suggested by the originator. More specifically, a minor degree of clinically-silent arthrosis at the contralateral compartment was not considered to be a contraindication. It is of notice that in only one of the reviewed cases there was progression of the contralateral osteoarthritic changes causing clinical painful symptomatology. This finding supports the current orthopaedic literature according to which it is suggested that medial gonarthrosis is a focal disorder, which if treated, will not cause progression of the disease in the lateral compartment[20]. Our limited number of revisions and clinical failures together with the 94.1% cumulative survival for 8 years permit us to conclude with confidence that the Oxford Knee can safely be performed for medial compartment replacement in the setting of a General District Hospital with results close to the ones achieved by the originator. Figure 1 and 2 demonstrate the preoperative and 10-year follow-up radiograph of a currently 84-year-old patient who is highly satisfied with the outcome of her operation.
With regard to lateral compartment arthroplasty, the longest published series has a follow up of 5-years and claims eleven revisions from a total of 53 replacements. Six of these revisions were due to early dislocation of the meniscus. The authors also claim that changes in the instrumentation and their surgical technique resulted in a reduction of the dislocation rate of the meniscus. As a result, during the last 14 cases, they experienced only one early dislocation. In our series there was one revision (due to early dislocation) in 9 replacements. We should however, include as a failure the case of the depressed lateral plateau fracture that was treated with delayed unicompartmental arthroplasty. The radiographic appearance of the specific prosthesis is highly satisfactory, however, the patient complains of pain affecting all compartments, which we cannot easily explain. Even though our results are favourable if compared with the ones published, we feel that our numbers are too small to allow us to make any safe conclusions.
The significance of the radiolucent line under the tibial component has already been described from Tibrewal et al from Oxford[19]. They noticed a radiolucent line with a radiodense line in the immediately adjoining bone in 88 out of 91 consecutive Oxford knee replacements. Histologicaly this radiolucency coincided with dense fibrous connective tissue. The authors concluded that there was no increased risk for loosening of the prosthesis. These observations are in agreement with our own, as we were unable to identify any relevance to clinical symptoms in any of our cases. We remain, however, unsure of the significance of the radiolucent lines noticed at the posterior aspect of two femoral components. Both cases are asymptomatic, but only further follow up will provide us with information regarding the importance of this finding.

CONCLUSIONS
We acknowledge that this series has the limitations of a retrospective study. We believe that his results show that the Oxford Unicompartmental knee replacement may be performed in a district general hospital with results close to those of the originator. The minimal approach technique permits less dissection with minimal interference to the extensor mechanism and a shortened rehabilitation period. Further independent follow-up studies are required to establish the long-term reproducibility of the Oxford Knee Replacement.

References
1. Aichroth P., Freeman M.A.R., Smillie I.S., Souter W.A. A knee function assessment chart. J Bone Joint Surg. [Br] 1978; 60-B, 308-9.
2. Armitage P., Berry G. Survival analysis. In: Armitage P, Berry G, ed. Statistical methods in medical research 3rd ed. Oxford: Blackwell science, 1994, 469-92.
3. Gibson P.H., Goodfellow J.W. Stress radiograph in degenerative arthritis of the knee. J Bone Joint Surg. [Br] 1986; 68-B, 608-9.
4. Goodfellow J., Kershaw C.J., Benson M.K.D., OΥConnor J.J. The Oxford knee for unicompartmental osteoarthritis. J Bone Joint Surg. [Br] 1988; 70-B, 692-701.
5. Goodfellow J., OΥConnor J. The mechanics of the knee and prosthesis design. J Bone Joint Surg. [Br] 1978; 60-B, 358-69.
6. Goodfellow J., Tibrewal S.B., Sherman K.P., OΥConnor J.J. Unicompartmental Oxford knee arthroplasty. J Arthroplasty. 1987; 2, 1-9.
7. Gunther T.V., Murray D.W., Miller D.A., et al. Lateral unicompartmental arthroplasty with the Oxford meniscal knee. The Knee. 1996; 3: 33-9.
8. Insall J.N., Dorr L.D., Scott R.D., Scott W.N. Rational of the Knee Society Clinical Rating System. Clin Orthop. 1989; 248, 13-4.
9. Insall J., Aglietti P. A five-to-seven-year follow-up of unicondylar arthroplasty. J Bone Joint Surg. [Am] 1980; 62-A, 1239.
10. Knutson K., Lewold S., Robertsson O., Lidgren L. The Swedish knee arthroplasty register. A nation-wide study of 30,003 knees 1976Π1992. Acta Orthop Scand. 1994; 65, 4, 375-86.
11. Kozinn S., Scott R. Current concepts review. Unicondylar knee arthroplasty. J Bone Joint Surg. [Am] 1989; 71-A, 145.
12. Laskin R.S. Unicompartmental tibiofemoral resurfacing arthroplasty. J Bone Joint Surg. [Am] 1978; 60-A, 182.
13. Lewold S., Goodman S., Knutson K., Robertsson O., Lidgren L. Oxford Meniscal Bearing Knee versus the Marmor Knee in Unicompartmental Arthroplasty for Arhrosis. A Swedish Multicenter Survival Study. J Arthroplasty. 1995; 10, 722-31.
14. Marmor L. The modular knee. Clin Orthop. 1973; 94, 242.
15. Murray D.W., Goodfellow J.W., OΥConnor J.J. The Oxford medial unicompartmental arthroplasty. A ten-year survival study. J Bone Joint Surg. [Br] 1998; 80-B, 983-9.
16. Peto R., Pike M.C., Armitage P., et al. Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br J Cancer. 1977; 35, 1-40.
17. Price A.J., Webb J., Topf H., Dodd C.A., Goodfellow J.W., Murray D.W. Rapid recovery after oxford unicompartmental arthroplasty through a short incision. J Arthropasty. 2001; 16, 8, 970-6.
18. Svard U.C., Price A.J. Oxford medial unicompartmental knee arthroplasty. A survival analysis of an independent series. J Bone Joint Surg. [Br] 2001; 83, 2, 191-4.
19. Tibrewal S.B., Grant K.A., Goodfellow J.W. The radiolucent line beneath the tibial components of the Oxford meniscal knee. J Bone Joint Surg. [Br] 1984; 66-B, 523-7.
20. Weale A.E., Murray D.W., Crawford R., et al. Does arthritis progress in the retained compartment after "Oxford" medial unicompartmental arthroplasty? A clinical and radiographical study with a minimum ten-year follow-up. J Bone Joint Surg. [Br] 1999; 81-B, 783-9.