Between January 1978 and June 1992, 798 patients underwent isolated AVR for hemodynamically significant calcific aortic stenosis. Preoperative coronary angiography was performed in 687 patients, of whom 229 had coronary stenosis > 40% and were divided into three (I, IIa and IIb) groups. Group I consisted of 144 patients undergoing AVR plus coronary artery bypass grafts (CABG); 39 patients in Group IIa had AVR only in the presence of coronary stenoses < 60%, and 46 patients in Group IIb had AVR only in the presence of coronary stenoses > 60%. Group III consisted of 144 matched patients selected from the remaining 458 patients with no coronary disease, or stenoses less than 40%, according to five matching criteria (age, sex, functional status, ejection fraction and year of surgery) with patients in Group I. Early mortality was 10.4% in Group I, 7.7% in Group IIa, 13% in Group IIb and 4.9% in Group III. Although the differences in operative mortality are apparent, they did not reach statistical significance. A difference in long term survival could only be detected between Groups I and III (actuarial survival at nine years 66% and 78.9%, respectively, p < 0.01). Similarly, late coronary events were more frequent in Group I than in Group III (13.9% vs. 5.1%, p < 0.03). It is concluded that revascularization should be as complete as possible for severe coronary stenoses coexisting with significant calcific aortic stenosis. However, bypassing of moderate coronary lesions (stenoses in the order of 50%) in association with AVR does not appear justified on current evidence. Further studies on this important aspect are clearly required.