This study was designed to assess whether intra- and postoperative epidural analgesia would diminish the overall rate of postoperative complications after major abdominal operations when compared to a standard anesthetic and postoperative analgesic regimen. A total of 214 patients undergoing infrarenal aortic bypass operations, gastric resection, gastrectomy, Whipple's operation, or duodenum-preserving pancreatic resection were randomly divided into two groups. Patients in the epidural group (n = 98) were operated on under light general anesthesia (midazolam, low-dose fentanyl, N2O/O2, pancuronium bromide). In addition, a mixture of bupivacaine (0.25%) and fentanyl (2 micrograms/ml) was infused (6-10 ml/h) via a thoracic epidural catheter intra- and postoperatively for 76:1.45 h (logarithmic normal distribution). Patients in the control group (n = 116) were operated on under a standard general anesthesia (midazolam, fentanyl, N2O/O2, isoflurane, pancuronium-bromide). Piritramid was injected for postoperative pain relief, either i.v. (recovery room, intensive care unit) or i.m. (surgical ward). In the epidural group the quality of analgesia and ability to cough were significantly better (2 P less than 0.0071) than in the control group (four observations each on the 1st and 2nd postoperative days). Heart rate and mean arterial pressure were lower in the epidural group at the same points of observation (2 P less than 0.01), as was the plasma glucose on the 1st postoperative day. The time up to the first postoperative defecation was shorter in the epidural group (79:1.51 h) as compared to the control group (93:1.38 h; 2 P less than 0.0167). The time to hospital discharge was equal in both groups (epidural group 19:1.6 days, control group 18:1.6 days).(ABSTRACT TRUNCATED AT 250 WORDS)