The common causes of postoperative liver dysfunction during anesthesia and surgery are due to the decreased cardiac output and/or systemic blood pressure as well as poor regional distribution of hepatic blood flow. It is well known that the surgical manipulation and the kinds of operation, especially around the hepatic area, are more important factors about the hepatic blood flow than anesthetics and anesthesia techniques. We postulated that the activity of sympathetic reflex initiated by surgical manipulation could be abolished by instituting epidural block on the peripheral operative field and consequently hepatic blood flow might be improved. We randomly selected 40 patients for upper abdominal surgery and divided them into 4 groups as follows: Group 1: Halothane anesthesia (10 cases, normal liver function test (LFT); contol) Group 2: Halothane anesthesia with epidural block (10 cases, normal LFT) Group 3: Enflurane anesthesia (10 cases, abnormal LFT; control) Group 4: Enflurane anesthesia with epidural block (10 cases, abnormal LFT) Epidural block was performed just before general anesthesia with 1% lidocaine, 2 ml/segment. Mean arterial pressure (MAP) was measured before anesthesia and at 5, 10, 15, 20, 30 and 60 min during surgery. And serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT), alkaline phosphatase, total protein and albumin were measured before anesthesia and on the 1st, 3rd, 5th and 7th postoperative day. The results were as follows: 1) MAP was changed significantly at 20 min in group 2 (p < 0. 01) and at 15 min in group 4(p < 0. 05). 2) The values of postoperative LFT were changed similarly but no statistical difference between group 1 and 2. 3) The values of postoperative LFT were declined significantly (esp. SGOT and SGPT), but no statistical difference between group 3 and 4. 4) Even though the postulation is correct, the effects of epidural block combined with general anesthesia (group 2, 4) on postoperative liver function were not different statistically compared with that of general anesthesia alone (group 1, 3). It might be suggested that the amount of decreased blood flow caused by surgical manipulation would be similar with that by the epidural block. 5) Now we hope that the further evaluation will be expected about the hepatic blood flow such as direct measurement.