Early reports on the excellent hemodynamic function and low thromboembolic rates of the Ionescu-Shiley bovine pericardial bioprosthetic valve (BPV) encouraged us to use it as our choice for valve replacement in 2680 patients from 1978 through 1983. Analysis of these patients at 5-year follow-up (mean 21.6 months) demonstrated the following important trends. Despite anticoagulation therapy in 48%, thromboembolism occurred in 88 patients for a linearized rate of 1.87% emboli per patient-year and was not time-related. The highest incidence of thromboembolism was in mitral valve replacement (MVR) (2.76% per patient-year). The actuarial freedom from reoperation resulting from valve failure at 5 years was 82% in aortic valve replacement (AVR), 87.1% in MVR, and 92.6% in AVR/MVR. The most distressing causes for reoperation were valve calcification (33 patients, 0.68% per patient-year) and leaflet disruption (11 patients, 0.23% per patient-year). Valve calcification was related to age, small valve size, and AVR position and increased with time, especially at the 4- to 5-year intervals. In patients under 30 years of age, calcification occurred in 18.7% at a mean time of 40.8 months in AVR and in 8.2% at 44 months in MVR, for an overall rate of 11.6%. Over the age of 30 years, it occurred in 14 patients (0.6%) at a mean time of 44 months. Leaflet disruption was not related to age and occurred later in AVR (50 to 58 months) than MVR (1.5 to 61 months). Events increased with time (mean range 37 to 58 months). Because of calcification and leaflet disruption, valve failure causing reoperation has increased significantly at the 4- to 5-year intervals even when valve replacement in patients under 30 years of age is excluded. If this trend continues, the valve failure rate will be exceedingly high on further follow-up. Thus we have limited the use of the BPV to a selected group of patients in whom valve longevity is less important than effective orifice size, thromboembolic rate, and freedom from anticoagulation.