Summary of Results of Original Publication1
Of an original consecutive cohort of 245 modular posterior stabilized total knee prostheses (170 patients), 193 knees (131 patients) had a mean follow-up of seven years (range, five to fourteen years). The Insall-Burstein II prosthesis (Zimmer, Warsaw, Indiana) was the only knee prosthesis used for primary arthroplasty by one surgeon during this period.
According to the rating system of the Hospital for Special Surgery2, 112 knees (58%) had an excellent result, sixty knees (31%) had a good result, fifteen knees (8%) had a fair result, and six knees (3%) had a poor result. According to the rating system of the Knee Society3, the mean knee score was 89.3 points (range, 48 to 98 points) and the mean functional score was 52 points (range, 0 to 100 points) at the time of the latest follow-up. The mean postoperative amount of flexion was 112° (range, 75° to 130°).
There were three reoperations. One patient had an acute infection at thirteen months and was treated with débridement and polyethylene exchange as well as chronic antibiotic suppression; the patient had a good result at five years. The second patient had flexion instability develop at five years and had a polyethylene liner exchange at six years; the result was excellent at seven years. The third patient (who was on chronic hemodialysis) had chronic synovitis with an osteolytic lesion of the proximal medial tibia at 5.5 years. After synovectomy and liner exchange (little polyethylene wear was seen), the patient continued to have recurrent hemarthroses and a poor result. In the original publication, we did not report a reoperation with intramedullary nail fixation of a supracondylar fracture of the femur at six months. This patient underwent an above-the-knee amputation because of an unrelated cause at nine years.
Osteolysis was seen in the proximal end of the tibia in eight knees (4%) in six patients. These knees were asymptomatic except for the one knee mentioned above. In summary, at a mean of seven years, the overall clinical results were excellent or good in 89% of the knees, and all components appeared to be well fixed radiographically. Because of the small number of knees with osteolysis, statistical comparison of polyethylene thickness and sterilization method was not possible.
This updated review was approved by our institutional review board under the study entitled "Musculoskeletal Health Outcomes." The patients were evaluated clinically by one experienced clinical research nurse (E.S.S.) using the rating systems of the Hospital for Special Surgery2 and the Knee Society3.
The knees were evaluated radiographically with standing anteroposterior radiographs, supine lateral radiographs, and tangential radiographs of the patella. The radiographic scoring system of the Knee Society was used to determine the overall alignment of the knee, the presence of radiolucent lines in zones adjacent to the cement mantle, and the migration of the components4. The alignment of the femoral and tibial components in the sagittal plane was not measured. Serial radiographs were reviewed for changes in the bone that were consistent with osteolysis5 (circular or oval lytic lesions) and for progression of radiolucent lines. Fluoroscopic positioning was not done, and oblique views of the knee were not performed.
Kaplan-Meier survivorship analysis6 to fifteen years was performed with use of the entire cohort of 245 knees (170 patients). For one survival curve, the end point was mechanical failure of the knee, which was defined as aseptic loosening, instability, or polyethylene wear (with osteolysis). For the second survival curve, the end point was reoperation or removal of the prosthesis for any reason (infection, supracondylar femoral fracture, or amputation). Statistical analysis was performed with the use of the chi-square test for the categorical variables (for example, sex, diagnosis, polyethylene sterilization method, and polyethylene thickness) of the knees with osteolysis, and the Student t test was used for the continuous variables (age, body mass index, and years of follow-up). The level of significance was set at p = 0.05.
Source of Funding
There was no external funding source for this study.
The prospective consecutive cohort consisted of 245 modular posterior stabilized knee replacements performed in 170 patients by one surgeon (P.F.L.). At the time of the last report, 193 knees in 131 patients had been followed for a minimum of five years (mean, seven years; range, five to fourteen years). In the interim period following the original report, forty patients (fifty-four knees) had died and fourteen patients (twenty-one knees) declined to return for follow-up or were lost to follow-up. To our knowledge, none of the patients who died or declined further follow-up had a problem with the knee replacement at the last time they were seen. One patient (two knees) had bilateral above-the-knee amputation; one was done prior to ten years and one was done after the ten-year follow-up. One patient, who had staged bilateral arthroplasty, had complete ten-year follow-up of one knee, but died prior to the ten-year follow-up evaluation of the other knee. Two patients (two knees) who had a reoperation (one for infection and one for hemodialysis-related synovitis) in the first cohort died prior to the ten-year minimum duration of clinical and radiographic follow-up (Table I).
The demographic characteristics of the group with a minimum ten-year follow-up included the following data. There were eighty-two knees in fifty-six women and thirty-five knees in twenty-two men. The mean age at the time of the index arthroplasty was sixty-six years (range, twenty-two to eighty-two years). The preoperative diagnosis was osteoarthritis in ninety-one knees (78%), rheumatoid arthritis in twenty-three knees (20%), and other diagnoses in three knees. The mean patient weight was 82 kg (range, 47 to 129 kg). The mean duration of follow-up was twelve years (range, ten to eighteen years). Clinically, the mean Knee Society pain score at the final follow-up evaluation was 93 points (range, 58 to 100 points), and the mean Knee Society functional score was 46 points (range, 0 to 100 points). The mean Hospital for Special Surgery knee score was 84 points (range, 49 to 98 points). With use of the Hospital for Special Surgery score, seventy-three knees (fifty patients) were rated as excellent, twenty-eight knees (twenty-four patients) were rated as good, thirteen knees (eight patients) were rated as fair, and three knees (two patients) were rated as poor. Six patients had had bilateral arthroplasty, and each had one excellent and one good result. The mean amount of flexion was 112° (range, 60° to 130°), and the mean flexion contracture was 1.6° (range, 20° to 0°).
For the 117 knees with a minimum ten-year follow-up, the thickness of the polyethylene liner (and minimum thickness at the center) as stated by the manufacturer at the time of implantation was 8 mm (5.3 mm) in thirty-nine knees, 10 mm (7.3 mm) in forty-eight knees, 12 mm (9.3 mm) in twenty-seven knees, and 15 mm (12.3 mm) in three knees. The polyethylene stock and sterilization method was GUR 4150, which had been sterilized with gamma irradiation in air, in eighty-four knees (mean follow-up time, 12.2 years; range, ten to eighteen years) and GUR 1050, which had been sterilized with gamma irradiation in nitrogen, in thirty-two knees (mean follow-up time, 10.3 years; range, ten to eleven years). These data were not available for one knee.
Three knees had a reoperation for mechanical failure, and three had a reoperation for other reasons (Table I). Of the three revisions for mechanical failure, one revision, which was documented in the previous report, was a tibial liner exchange for flexion instability in a patient with inflammatory arthritis. One knee with tibial osteolysis had sudden loosening of the tibial component at 12.5 years and underwent revision of both the femoral and tibial components, with a good result at three years following the revision. One knee with tibial osteolysis noted at two years had synovitis with worsening osteolysis develop at ten years and had a liner exchange, synovectomy, and bone-grafting, with an excellent result at two years following the revision.
The overall femorotibial alignment was neutral (mean, 5.2° of valgus; range, 3° to 9° of valgus) in 115 knees, 2° of varus in one knee, and 10° of valgus in one knee. Thirty-seven nonprogressive radiolucent lines measuring 1 to 2 mm in width were seen in twenty-four knees (21%). The most common locations were zone 1 in the tibia as seen on the anteroposterior radiograph (ten knees) and zone 2 in the patella as seen on the axial radiograph (nine knees). One patient had a complete radiolucent line around the patellar component as seen on the lateral radiograph. Osteolysis was seen in eight knees in six patients. Six lesions were in the proximal end of the tibia, one was in zone 1 of the femur, and one was in zone 2 on the axial view of the patella. Three knees had asymptomatic stress fractures of the patella.
With a reoperation for mechanical failure as the end point, the estimated survival for the entire cohort was 96.8% (95% confidence interval, 92.6% to 100%) at fifteen years (Fig. 1). With any reoperation on the knee as the end point, the fifteen-year estimated survival was 90.6% (95% confidence interval, 85.1% to 96.1%) (Fig. 2).
With the small numbers evaluated, we could not identify a significant association between osteolysis and patient sex (p = 0.35), side (p = 0.71), diagnosis (p = 0.94), body mass index (p = 0.77), follow-up time (p = 0.80), polyethylene thickness (p = 0.18), or polyethylene sterilization method (p = 0.43). A significant association was detected between osteolysis and mean patient age (57.6 compared with 66.2 years; p = 0.04) and between osteolysis and reoperation for mechanical failure (p < 0.0001).