To determine which groups of patients are at highest risk for operative or late mortality, 259 consecutive patients who underwent operation between 1978 and 1984 were studied; 170 underwent aortic valve replacement and 89 underwent aortic valve replacement combined with coronary artery bypass grafting. Multivariate analysis of risk factors selected emergency operation and patient age older than 70 years as the strongest predictors for operative death. Although patients having aortic valve replacement and coronary artery bypass grafting had a higher operative mortality rate (13.5 versus 3.5%), the combined operation had no independent predictive effect on early or late results. At a mean follow-up time of 48 months after surgery, 72% of the survivors of operation were living, 10% were lost to follow-up and 18% were dead. Seventy-seven percent of long-term survivors were in New York Heart Association functional class I or II. The incidence of thromboembolism, paravalvular leak, bacterial endocarditis and hemorrhage each occurred at a rate of less than 1% per patient-year. The factors associated with late death were preoperative age, male sex, left ventricular end-diastolic pressure, cardiac index and functional class. Despite an increase in operative mortality, patients undergoing emergency operation were not at higher risk of late death. Operative mortality is concentrated among several high risk groups. For patients undergoing elective operation, operative mortality is low, especially if the patient is less than 70 years old. Late results are good for all groups of patients undergoing operation, including those who are at greater risk of dying at operation.