The Kinematic Stabilizer is a posterior-cruciatesubstituting design of total knee replacement. We have reviewed 109 primary total knee replacements in 95 patients at a mean follow-up time of 12.7 years (10 to 14). We used survival analysis with failure defined as revision of the implant. This gave a cumulative survival rate of 95% (95% CI± 5%) at ten years and 87% (± 10%) at 13 years.These results from an independent centre confirm the value of an established design of cemented total knee replacement and question the wisdom of the introduction of modifications and new designs without properly controlled trials. There are few reports from independent centres of total knee replacement with follow-up longer than ten years and we could find no long-term reports of knee replacement using the Kinematic Stabilizer design. We describe our 10-to 14-year results with this posterior stabilised total knee replacement. PATIENTS AND METHODSFrom 1980 to 1983 inclusive we performed 109 primary total knee replacements on 95 patients, most with rheumatoid arthritis, using the Kinematic Stabilizer prosthesis (Howmedica, Rutherford, New Jersey). In the first year, this implant was used for all knee replacements performed at this hospital. Subsequently, other types became available and the Kinematic Stabilizer was used only when the posterior cruciate ligament was absent or not functioning. All revisions were excluded from the series. All the operations were performed by the senior author (IMP) using the Universal-1 instrumentation and a standard operating technique. No patellae were resurfaced and all components were cemented.The Kinematic Stabilizer total knee replacement is a posterior-cruciate-substituting design (Walker 1980) in which a contoured central post on the tibial polyethylene tray engages a housing in the femoral component. The guiding surfaces provide anteroposterior stability from full extension to 30° of flexion. Posterior stability is maintained as flexion proceeds and the surfaces act to guide femoral rollback (Figs 1 and 2).Our series included 78 women and 17 men; their average age was 64 years (37 to 86) and 17 were under 55. The primary diagnosis was rheumatoid arthritis in 63, osteoarthritis in 29 and psoriatic arthritis in 3.All patients were reviewed annually. Those who did not attend at ten years were traced and contacted. Four patients (six knees) were unwilling to attend but provided information over the telephone. Patients were assessed according to the Knee Society clinical rating system , in which points are allocated for pain, range of motion and stability. Deductions are made for flexion contracture, extensor lag and poor coronal alignment. For the function score, points are gained for walking distance and stair-climbing, and deducted for the use of walking-aids. Each score is given out of a possible 100. Radiographs were obtained and assessed according to the Knee Society Radiological Evaluation Form (Ewald 1989).We performed survival analysis by constructing a life table (Armitage and Berry 1994) ...