The influence of epidural neural blockade on postoperative insulin resistance was studied using the euglycaemic insulin clamp technique. Eighteen patients undergoing elective upper abdominal surgery of moderate severity were allocated to two groups: group G patients underwent operation under general anaesthesia, and postoperative pain was relieved by systemic administration of analgesia; and group E patients received epidural analgesia during surgery and epidural morphine postoperatively. In each patient the euglycaemic insulin clamp test was performed twice: several days before surgery and on postoperative day 1. Peroperative catecholamine and cortisol responses were also measured to investigate possible endocrine mechanisms of the insulin resistance. Glucose disposal (M) decreased in both groups on postoperative day 1 at plasma insulin concentrations ranging from 1.2 to 10.0 milliunits ml-1, resulting in the downward shift of dose-response curves. However, this downward shift was significantly smaller in group E than in group G patients. Urinary adrenaline excretion increased markedly on the day of operation in group G, but was significantly inhibited in group E. Urinary noradrenaline excretion increased mainly on postoperative day 1 in group G, but was significantly inhibited in group E. Plasma cortisol response was lower in group E than in group G during and shortly after operation, and was significantly inhibited in group E on postoperative day 1. These results indicate that insulin resistance after elective abdominal surgery is due to a postreceptor deficit in glucose utilization, as indicated by the downward shift of the dose-response curves. This disturbance in glucose metabolism was reduced by epidural analgesia, the results being associated with inhibited catecholamine and cortisol responses.