The results of 1753 consecutive aortic valve replacements performed during a 15-year period (1969 to June 1983) are reported. All patients received one of five models of the Björk-Shiley valve. The follow-up was 99.6% complete and covered 10,658 patient-years (mean 6.4 years). An autopsy was performed in 72% of all fatalities. Early mortality was 5.5%. The early mortality rate for patients undergoing isolated, elective aortic valve replacement between 1979 and 1983 was 2.3%. Five-, 10-, and 15-year actuarial survival rates (early mortality excluded) were 85%, 70%, and 54%, respectively, and complication-free survival rates at the same time intervals were 78%, 59%, and 42%. Eighteen percent of the late deaths were considered valve related. Anticoagulant-related hemorrhage was the single most common valve-related complication (1.4/100 patient-years) and the most common valve-related cause of death (0.3/100 patient-years). The incidences of embolism and valve thrombosis were closely related to the efficacy of the anticoagulant program. Fourteen of 19 valve thromboses occurred among 180 patients without anticoagulants. The incidences of embolism and valve thrombosis among 1573 patients with continuous anticoagulant treatment were 0.7 and 0.06/100 patient-years, respectively. Six strut fractures occurred, all in convexo-concave valves. There was no mechanical failure among the Monostrut valves. The incidence of "prosthetic failure" (ie, valve-related complications [not leakage] necessitating reoperation or causing the patient's death) was 0.6/100 patient-years. The incidence of embolism and prosthetic failure was similar for the different valve models, and these long-term results might therefore be considered representative for all Björk-Shiley aortic valves.