Previous studies of the survivorship of unicompartmental knee replacements have shown mixed results1,2. Recent studies have described ten-year survivorship in the mid-90% range, which is comparable with that of total knee replacement3-5. Improved results have been attributed to patient selection, improvement in surgical technique, and better prosthetic design. A design feature of some unicompartmental implants6,7 is a mobile meniscal tibial bearing, designed in an attempt to achieve congruency throughout the range of motion between the articulating surfaces. Both experimental and clinical evidence has shown that congruous articulating surfaces reduce the rate of polyethylene wear8,9. The mobile meniscal bearing requires restoration of normal soft-tissue tension and symmetrical balance throughout the range of motion to allow for optimal bearing function. This is accomplished by removal of bone rather than by releasing of ligaments.
The purposes of this study were to determine the postoperative alignment of the limb, the survivorship, and the modes of failure in a series of unicompartmental knee arthroplasties with a medial compartment Oxford implant, which had a completely unconstrained polyethylene mobile bearing.
From 1989 to 1994, as part of an investigational device exemption study, fifty-five consecutive medial compartment Oxford phase-2 implants (Biomet, Warsaw, Indiana) were placed in fifty-one patients by a single surgeon (R.H.E.Jr.). This study was approved by the hospital internal review board, and the patients signed an informed consent for participation in the study. Thirty-two knees were in women and twenty-three knees were in men. The average age at the time of the index surgery was sixty-four years (range, thirty-eight to eighty-five years), and the average weight was 78.1 kg (range, 50.9 to 109.1 kg). Clinical and radiographic data were collected prospectively.
The phase-2 Oxford implants were introduced in 1987 (Figs. 1-A and 1-B). Compared with the phase-1 implant, the phase-2 implant had a redesigned femoral component, which used a spherical mill for femoral bone removal rather than a saw blade, allowing for bone removal in 1-mm increments, along with gap-measuring gauges, which enabled symmetrical balancing of the soft-tissue envelope of the knee in both flexion and extension. Achievement of flexion and extension symmetry was based solely on bone removal. On the basis of the flexion and extension space measurements, a corresponding tibial bearing was placed. The bearing was large enough to be stable but not so large that the soft tissues were tighter than physiologically normal. There was no attempt to achieve a specific degree of limb alignment.
Patient selection followed the criteria of Carr et al.10. These consisted of patients with osteoarthritis, a correctable varus deformity, an intact anterior cruciate ligament, an absence of degenerative findings in the lateral compartment of the knee on full-length (90-cm) standing radiographs, an absence of lateral compartment tenderness, and no more than minimal patellofemoral abnormalities, on clinical and radiographic assessment. This grouping of findings was termed anteromedial osteoarthritis by White et al.11.
The final decision to proceed with unicompartmental arthroplasty was made at the time of surgery. Age, weight, and activity level specifically were not determining factors in the decision for unicompartmental replacement. A median parapatellar arthrotomy was used, and all cartilage surfaces were evaluated. Small areas of cartilage loss on the patella, usually seen on the medial facet, were acceptable. Full-thickness cartilage loss on the weight-bearing surface of the lateral compartment and notable inflammatory synovitis were considered contraindications to proceeding with unicompartmental replacement. A total knee replacement was performed in such cases.
Plain radiographs and the Knee Society clinical scoring instrument were used to assess the outcome at each follow-up visit. The Knee Society instrument consists of two separate scales: a knee score and a function score12. All data were collected prospectively. All knees were evaluated with plain radiographs with standard views, with use of the modified Knee Society zones9, to assess for lucencies and osteolysis. In addition, the radiographs were assessed to detect component subsidence, loosening, and progressive degeneration in the remaining joint compartments. Knee alignment was assessed with use of a preoperative and an immediate postoperative 90-cm standing anteroposterior radiograph, in order to determine the location of the mechanical axis with respect to the center of the tibial surface, as described by Kennedy and White13. Zones 1 and 2 are on the medial side of the tibial eminence, and zones 3 and 4 are on the lateral side of the tibial eminence. Zone C is the central part of the tibial plateau. At the final follow-up evaluation, patients who were not able or were unwilling to return for an examination were interviewed by telephone with use of a standard questionnaire, which was based on the knee scoring instrument.
Preoperative and postoperative alignment was compared with the Student t test. Correlation of bearing size to final alignment was determined by the Pearson coefficient. For all tests a p value of <0.05 was considered significant. Survivorship analysis was by the Kaplan-Meier method.
For the fifty-five knees, the mean preoperative Knee Society knee and function scores were 43 and 56 points, respectively. In addition, the mean preoperative alignment was 1.6° of varus (range, 8° of varus to 7° of valgus alignment), and the mean postoperative alignment was 5.6° of valgus (range, 2° of varus to 13° of valgus alignment). Immediately postoperatively, thirty-three of the fifty-five knees had the mechanical axis crossing through the intercondylar notch of the tibia (zone C), and forty-seven knees had the axis crossing the central 50% of the proximal aspect of the tibia (zones 2, C, and 3) (Fig. 2).
At the time of the final follow-up of the original series of fifty-one patients (fifty-five knees), fifteen patients (sixteen knees) had died with the prosthesis intact. These fifteen patients had been followed for an average of 5.1 years (range, 1.2 to 6.2 years) prior to death. Two patients (three knees) were lost to follow-up at 4.2 and 5.2 years, with the implants functioning well. Seven knees in six patients had been revised, at an average of 7.6 years (range, 1.0 to 11.2 years). The remaining twenty-nine knees in twenty-eight patients continued to function at the time of the latest follow-up. This final group of living patients had an average duration of follow-up of 10.4 years, (range, 2.6 to 14.3 years). Of the twenty-nine knees, thirteen had no recent follow-up radiograph and the patients were contacted by telephone, extending the average period of clinical follow-up to 11.8 years (range, 8.8 to 14.3 years). None of the contacted patients reported any changes in implant function. Clinical scoring was updated by combining the subjective symptoms with the most recent functional status contained in the medical record. At the time of the final follow-up, twenty-eight patients had mean Knee Society knee and function scores of 75 and 90 points, respectively.
Six of the knees that had a revision arthroplasty underwent conversion to a total knee prosthesis. Of the six knees, five were revised for progression of arthritis; these included one knee that was revised for acute inflammatory polyarthritis at one year and four that were revised for progression of arthritis in the lateral compartment, with progressive valgus, at an average of 10.2 years. The four knees with lateral compartment progression all had small tibial bearings, with an average preoperative alignment of 4° of varus and an average postoperative alignment of 5° of valgus (range, 2° to 8° of valgus). Immediately postoperatively, the mechanical axis crossed the center zone (zone C) in three of the knees, and it crossed in zone 2 in the fourth, on the medial side of the tibia, and was therefore in slight varus. None of the knees that had lateral compartment arthritis progression were overcorrected into valgus alignment at the time of the initial surgery.
One knee had femoral component loosening that required revision at 3.7 years. Only one size of femoral component was available for the phase-2 components, and the femoral component that loosened was undersized by current sizing recommendations.
No failure was due to subsidence, tibial component loosening, or polyethylene wear. Radiographic zone analysis at the time of the final follow-up, with the radiographic methods used, showed five femora with one radiolucent zone each. On the tibial side, with five zones for analysis, eleven components had a radiolucent line in one zone, four components had a lucent line in two zones, three components had a lucent line in four zones, and two components had a complete radiolucent line around the tibial component but without migration. Only one of these two patients had mild pain but did not desire revision. None of the radiolucent lines demonstrated a divergent pattern, and no osteolysis was seen around any of the components in this study, suggesting a stable interface appearance, and no tibial components were thought to be clinically loose.
One bearing had to be replaced at three years because of osseous impingement in full extension, causing pain and an effusion. The anterior edge of the bearing was striking the femur in full extension. Bearing replacement, with removal of the impinging bone, eliminated the effusion and the pain, with the original metal components intact at eleven years. There were no bearing dislocations in this series.
With the small numbers of subjects in this study, no correlation could be identified between the size of the bearing and the alignment of the knee (Pearson coefficient, r = 0.009). Therefore, by restoring soft-tissue tension of the collateral and cruciate ligaments, optimal mechanical alignment of the knee was consistently achieved.
With use of the Kaplan-Meier method, survivorship at ten years was 85% with revision surgery for any reason as the end point (Fig. 3), 90% with revision for progression of the arthritis in the lateral side as the end point, and 96.3% with component loosening as the end point.