Between 1970 and 1984,1138 patients underwent the insertion of 1300 prosthetic heart valves in Western Australia; 56% received an aortic‐valve replacement; 34% received a mitral‐valve replacement and 10% had more than one valve replaced. Mechanical valves were used in 93% of patients before 1977, in 20% of patients between 1978 and 1982 and in 70% of patients from 1983 onwards. The 30‐day mortality was 18% before 1973 and has been below 6% since 1974. The over‐all, 15‐year actuarial survival rate was 67%; this was not affected by age, sex, race, valvular position or the type of prosthesis. Both the 30‐day mortality and 15‐year survival rates were significantly worse in patients who underwent multiple valvular replacements (13% and 54%, respectively) or reoperation (16% and 58%, respectively). The major causes of death were cardiac failure and myocardial infarction (65%); endocarditis (13%); cancer (6%); and thromboembolism and bleeding (6%). The hazard rate for reoperation was low and fairly constant in patients with mechanical valves, but increased markedly after four years in patients with tissue valves. Although our experience so far suggests that survival rates are not affected by the choice of prosthesis, this may not be so in the future, as more patients with tissue valves undergo reoperation and so become exposed to an increased risk of mortality.