The influence of 35 preoperative and intraoperative characteristics on operative mortality risk after 1,479 isolated aortic valve replacement procedures (1967 to 1981) was investigated utilizing univariate and multivariate logistic regression analyses. Mean age at operation was 58 +/- 13 years; 72% of patients were men. Physiology was classified as aortic stenosis (58%), regurgitation (30%), or both (9%). The overall operative mortality rate was 7% +/- 1%, but there were substantial differences in operative mortality rates among physiological subgroups (aortic regurgitation, 10% +/- 2%; aortic stenosis, 6% +/- 1%; stenosis/regurgitation, 5% +/- 2%). Independent determinants of operative mortality rate in the entire group were advanced New York Heart Association functional class, renal dysfunction, physiological subgroup, atrial fibrillation, and older age. In the aortic regurgitation subgroup, functional class, atrial fibrillation, and operative year were independent predictors. In the aortic stenosis subgroup, the significant determinants were functional class, renal dysfunction, age, prosthetic valve dysfunction, and absence of angina. Concomitant coronary bypass grafting, previous operation, endocarditis, and ascending aortic replacement had no independent predictive effect on operative mortality rate. Thus, the early results of aortic valve replacement can be related to several specific variables describing the functional and physiological status of the patient. Operative mortality rate is not independently related to previous operation or concomitant operative procedures. Specific differences in risk factors exist among the various physiological subgroups, probably reflecting the pathophysiology of the different hemodynamic lesions. This information should provide for a more rational approach to aortic valve replacement, at least in terms of early risk/benefit deliberations.