The goal of this randomized study of high-risk surgical patients was to determine whether intraoperative thoracic epidural anesthesia in combination with light general anesthesia alters postoperative morbidity when compared to a standard technique of "balanced" general anesthesia. A total of 173 patients scheduled for abdominal aortic reconstruction were admitted to the study; 86 were to receive "balanced" general anesthesia (group 1) and 87 thoracic epidural anesthesia in combination with light general anesthesia (group 2). Preoperative evaluation included standard clinical tools, dipyridamole thallium gammatomography, and radionuclide angiography. In these patients, all of whom had peripheral artery disease, there were no significant differences in associated coronary artery disease, hypertension, and cardiovascular treatment. The distribution of left ventricular ejection fraction and the number of patients with thallium redistribution were not statistically different between the two groups. During the postoperative period, group 1 received analgesia of subcutaneous morphine (n = 35), epidural fentanyl (n = 30), or epidural bupivacaine (n = 21). In group 2, 6 patients with a nonfunctioning epidural catheter due to technical failure received a balanced general anesthesia and were eliminated from the study. During the postoperative period, group 2 received analgesia of subcutaneous morphine (n = 26), epidural fentanyl (n = 25), or epidural bupivacaine (n = 30). Cardiovascular morbidity did not differ between the two groups: 22 patients in group 1 and 19 patients in group 2 had a major postoperative cardiac event.(ABSTRACT TRUNCATED AT 250 WORDS)