Among 500 patients consecutively undergoing primary aortic valve replacement combined with coronary bypass grafting from 1967 to 1981, there were 29 (5.8%) perioperative deaths. Follow-up of the late survivors ranged from 43 to 181 months (mean 85 months); actuarial survival rates were 88%, 77%, and 52% and event-free survival rates were 80%, 65%, and 32% at 2, 5, and 10 postoperative years. Cox proportional hazard analyses were used to identify determinants of late risk. Patient-related variables associated with decreased late survival rates included advanced age, cardiothoracic ratio 50% or greater, and preoperative New York Heart Association class III or IV symptoms. Moderate or severe impairment of left ventricular function as determined by angiography and advanced age were variables that decreased late event-free survival rates. Patients with bioprostheses had better survival rates (p less than 0.001) and event-free survival rates (p = 0.012) than did patients with mechanical valves. Analyses of subgroups according to the type of valve and postoperative anticoagulant management showed that both survival and event-free survival rates were decreased for patients with mechanical valves who were not taking warfarin and were enhanced for patients with bioprostheses who were not taking warfarin.