Background: Osteolysis is an important complication associated with total knee arthroplasty. The purpose of this study was to compare the prevalence of osteolysis after failed total knee arthroplasty with a mobile-bearing prosthesis and after failed arthroplasty with a fixed-bearing prosthesis.

Methods: Eighty revision total knee arthroplasties performed between 1995 and 1998 were included in this study. All had radiographic evidence of advanced polyethylene wear. The mobile-bearing group consisted of thirty-four knees with a Low Contact Stress implant, and the fixed-bearing group included forty-six knees. The average time (and standard deviation) from the primary operation to the revision was 102.8 ± 26.5 months in the mobile-bearing group and 96.0 ± 30.1 months in the fixed-bearing group. The prerevision radiographs and operative findings were reviewed.

Results: The prevalence of osteolysis was significantly higher in the mobile-bearing group (47%; sixteen of thirty-four knees) than in the fixed-bearing group (13%; six of forty-six knees) (p = 0.003). The distal part of the femur was involved in thirteen knees in the mobile-bearing group and in four knees in the fixed-bearing group. Seventeen knees had osteolysis in the posterior aspect of the femoral condyle, which was the most common site of osteolysis; however, twelve of them had no evidence of osteolysis on prerevision radiographs.

Conclusions: The prevalence of osteolysis was higher in the knees with a mobile-bearing prosthesis than in those with a fixed-bearing prosthesis. The osteolysis was predominantly on the femoral side, adjacent to the posterior aspect of the condyle. Radiographic evaluation of osteolysis in the distal part of the femur may not be reliable and usually leads to an underestimation of the degree of osteolysis.