The risk associated with surgical revascularization for unstable angina is critically dependent upon the clinical presentation of the patient. For this study, between January 1982 and December 1987, clinical, angiographic, operative and hospital outcome data were collected prospectively for 6539 patients undergoing surgery for unstable angina. Urgent surgery was performed in 1523 patients (23.3%), while 5016 (76.7%) underwent semielective revascularization. The mean age was 58.9 +/- 9.2 years and 805 patients (12.3%) were aged 70 years or above. The male:female ratio was 3.6:1. Depressed left ventricular function (left ventricular ejection fraction less than 40%) was present in 27.2% of the population. Preoperative myocardial infarction (within 30 days of surgery) had occurred in 588 patients (9.0%). Operative mortality was 4.6% (301 deaths). Stepwise logistic regression analysis was performed to determine the independent predictors of operative mortality. The following variables were selected in descending order: urgent surgery (P less than 0.001), coronary reoperation (P less than 0.001), depressed left ventricular (P less than 0.001), female gender (P less than 0.001), increasing age (P less than 0.001), left main stenosis (P = 0.002), and preoperative myocardial infarction (P less than 0.001). Predicted operative mortality varied between 0.5 +/- 0.3% and 82.6 +/- 12.7%. The most important determinant for patients with a preoperative myocardial infarction was left ventricular dysfunction, whereas urgent surgery for unstable angina was the most important risk variable in those without preoperative necrosis.