Purpose: This study evaluated perioperative variables to predict death in nonruptured and ruptured abdominal aortic aneurysm (AAA) surgery. Methods: A consecutive review of all patients who underwent AAA surgery from January 1984 to December 1993 was carried out. Perioperative variables were analyzed with univariate and multivariate statistical models to predict mortality rates. Results: Four hundred seventy-eight patients with nonruptured AAAs and 157 patients with ruptured AAAs were studied. In patients with nonruptured AAAs, the mortality rate was 3.8%. Using stepwise logistic regression analysis, independent predictors of death were perioperative myocardial infarction (odds ratio [OR], 5.0; p < 0.01), prolonged postoperative ventilation (OR, 4.0; p < 0.01), history of peripheral vascular disease (OR, 2.9; p < 0.01), preoperative renal dysfunction (OR, 2.7; p < 0.01), and history of congestive heart failure (OR, 2.6; p < 0.03). In patients with ruptured AAAs, the mortality rate was 46%. Analysis of preoperative variables using multivariate stepwise logistic regression found predictors of death to be preoperative unconsciousness (OR, 3.1; p < 0.01), advanced age (OR, 1.9; p < 0.01), and cardiac arrest (OR, 1.8; p < 0.05). In patients who survived the initial surgery for ruptured AAA, a second stepwise logistic regression model found independent predictors for subsequent postoperative death to be coagulation disorder (OR, 7.9; p < 0.01), ischemic colitis (OR, 6.4; p < 0.01), inotropic support beyond 48 hours (OR, 4.8; p < 0.01), delayed transport to operating room (OR, 4.6; p < 0.01), advanced age (OR, 4.4; p < 0.01), perioperative myocardial infarction (OR, 4.0; p < 0.05) and postoperative renal dysfunction (OR, 3.7; p < 0.01). Conclusion: Prolonged ventilation, perioperative myocardial infarction, a history of peripheral vascular disease, preoperative renal dysfunction, and a history of congestive heart failure are independent predictors of perioperative death in patients with nonruptured AAAs. For patients with ruptured AAA, mortality rates can be estimated before surgery using age, level of consciousness, and cardiac arrest. For patients who survive the initial surgery for ruptured AAA, subsequent mortality rates can also be predicted.