The introduction of aortic balloon valvuloplasty has prompted review of established techniques to treat aortic valve disease. To this end, the surgical results were reviewed for 1,148 patients undergoing aortic valve replacement, with and without associated coronary artery bypass grafting, for aortic stenosis between January 1, 1975 and September 1, 1987. The overall death rate was 4.6%. The neurological event rate (stroke or reversible ischemic event) was 4.2%, the myocardial infarction rate was 3.5%, and 30.7% had a hospital stay in excess of 10 days. Mortality for isolated elective aortic valve replacement rose from 2.5% for patients under age 70 to 7.3% for those over age 70. Mortality for elective aortic valve replacement plus coronary bypass was 3.8% for patients under age 70 and 11.6% for patients over age 70. The neurological event rate rose from 3.0% to 8.4% and the prolonged length of stay from 28% to 42%. By stepwise logistic regression analysis, the predictors of in-hospital mortality were age and emergent surgery. Age was the only multivariate predictor of neurological events. The year of surgery and emergent status were the multivariate predictors of prolonged length of hospital stay. Aortic valve replacement for aortic stenosis is a well-established technique with low neurological event rates and mortality, which may be predicted in advance. Newer techniques must be considered in light of data such as these.