This is a study of 199 consecutive patients who survived isolated mitral or aortic valve replacement with a Björk-Shiley (B-S) or St. Jude Medical (SJM) prostheses. Mean follow-up was 36 months. Preoperative status was similar for both groups (B-S or SMJ). Five year survival rates for mitral valve replacement (MVR) were 91% for B-S and SJM, and for aortic valve replacement (AVR) 96% vs 91% for SJM (p less than 0.06). Linearized incidence of thromboembolism in MVR was 5.62% patient year for BS and 5.23% patient year for SJM (p = NS). Classical risk factors for thromboembolism were evaluated on a univariate analysis: no single risk factor was identified. Only the association of two or more was possibly significant. The presence of inadequate anticoagulation (prothrombin index greater than 50%) was the main risk factor for thromboembolism, as the incidence rate was 19% per patient year when poorly anticoagulated versus 4.6% per patient year incidence in the whole series (p less than 0.01).