From 1986 to 1991, 74 consecutive hybrid total knee arthroplasties (in 65 patients) of a single design were performed at the authors' institution. Seven patients (eight knees) died with well functioning replacements during the surveillance period. One patient refused to participate in any followup efforts because of billing disputes. This left 65 total knee arthroplasties (57 patients) for review. There were 35 men and 22 women with an average age of 60 years (range, 27-83 years) at the time of arthroplasty. The underlying diagnosis was osteoarthritis in 46 knees, rheumatoid arthritis in six knees, and other in four knees. The average length of followup was 7.4 years (range, 5-10 years), and no patients were lost to followup. The Knee Society scores improved from an average of 37 preoperatively to 84 at latest followup. The functional score improved from 49 to 69 at latest followup. Ten (13.8%) knees required revision surgery. Eight of the nine knees that were revised had femoral component problems, including a loose component in six and a fractured component in two. One knee was revised for polyethylene wear with secondary osteolysis and one was revised elsewhere in which the operative details are unknown. The implant survival to revision at 5 years was 89% and 85% at final followup. Femoral component fixation in hybrid total knee arthroplasty is unreliable with this component design. In light of the excellent 10- to 15- year results of cemented condylar knee designs it is thought that hybrid fixation should be abandoned.