Preoperative cardiac testing in patients undergoing vascular surgery remains controversial. We have advocated selective use of dipyridamole-thallium scans based on clinical markers of coronary artery disease before aortic surgery. The present study assessed both the efficacy of this policy and the role of surgical factors in the current morbidity of aortic reconstruction. Two hundred two elective aortic reconstructions (151 abdominal aortic aneurysms, 51 aortoiliac occlusive disease) performed in the period from January 1989 to June 1990 were reviewed. Preoperative dipyridamole-thallium scanning was performed in 29% of all patients, prompting coronary angiograms in 11% and coronary artery bypass grafting/percutaneous transluminal coronary angioplasty in 9% of patients before aortic reconstruction. The overall operative mortality rate was 2%, with one cardiac-related death. Major cardiac (nonfatal myocardial infarction, unstable angina) and pulmonary complications occurred in an additional 4% and 6%, respectively, of patients. Coronary artery disease clinical markers and surgical factors were analyzed with stepwise logistic regression for the prediction of operative mortality rates and major cardiopulmonary complications. Variables retaining significance in predicting postoperative death or cardiopulmonary complications included prolonged (more than 5-hour) operative time (p less than 0.004), operation for aortoiliac occlusive disease (p less than 0.010), and a history of ventricular ectopy (p less than 0.002). Prolonged operative time (p less than 0.006) and the detection of intraoperative myocardial ischemia (p less than 0.030) were predictive of major cardiac complications after univariate analysis.(ABSTRACT TRUNCATED AT 250 WORDS)