Plasma catecholamine levels were measured to evaluate postoperative pain relief either with epidural morphine or systemic analgesics in sixteen patients who underwent gastrectomy. Eight patients (epidural morphine group) obtained postoperative analgesia with continuous epidural morphine with a pump (CADD-PCA, Model 5200P, Pharmacia). A bolus of morphine was administered through an indwelling thoracic (Th8 X 9) epidural catheter 3 hrs prior to the proposed end of the surgery, which was followed with continuous epidural infusion of morphine at a rate of 0.167-0.042mg.hr-1 by the pump during and after anesthesia and surgery with gradual decrease in dose until the third postoperative day. The remaining eight patients (systemic analgesics group) received repeatedly intravenous or intramuscular pentazocine and buprenorphine when needed. Plasma epinephrine levels increased significantly at the end of surgery in both groups, and were higher in the systemic analgesics group than those in the epidural morphine group. In the epidural morphine group, the catecholamine levels decreased to the previous day's levels on the first postoperative day and afterward, but remained high during three postoperative days in the systemic analgesics group. Plasma norepinephrine levels increased significantly at the end of surgery and afterward in both groups. However, they were significantly higher in the systemic analgesics group than in the epidural morphine group. Plasma dopamine levels were unchanged in the epidural morphine group during the surgical procedures, but they increased significantly on the first postoperative day and thereafter in the systemic analgesics group. Our study suggests that continuous epidural infusion of morphine is adequate for postoperative pain relief and exerts a suppressing effect on plasma catecholamine levels as compared with systemic analgesics regimen.