BACKGROUNDS: Epidural anesthesia before surgical stimulus may reduce or prevent subsequent pain by preemptive analgesia. We studied the effect of varied concentration of a local anesthetic agent administered through epidural catheter before operation on preemptive analgesia after an abdominal hysterectomy. METHODS: Fourty-five patients scheduled for abdominal hysterectomy were blindly randomized to receive 10 ml of 0.5% bupivacaine (group I), 0.25% bupivacaine (group II) or saline (group III) respectively before induction of anesthesia. For postoperative pain control all patients received the same analgesic regimen which was 10 ml of 0.125% bupivacaine and buprenorphine 0.15 mg after resection of the uterus. Then the epidural catheter was connected with infusion pump containing 100 ml of 0.125% bupivacaine and buprenorphine 0.45 mg. Its infusion rate was 2 ml/hour. Pain was assessed using the Visual Analogue Scale (0 = no pain to 10 = intolerable pain) on rest, motion, cough and verbal rating scale at 1, 3, 6, 24 and 48 hour after operation. Side effects were recorded at the same time intervals. RESULTS: Using the ANOVA on ranksum test, pain scores 1 hour after operation differed between groups I and III (P<0.05). In VAS in motion, the pain scores 24 hours after operation differed between groups I and III (P<0.05). In VRS, the pain scores 3 hours after operation differed between groups II and III (P<0.05), and those 24 hours after operation differed between groups I and III (P< 0.05). The incidence of motor weakness using Chi-square test in group I differed from groups II, III (P< 0.05). CONCLUSIONS: Among the 45 patients who had a postoperative pain control after abdominal hysterectomy with initial bolus injection and then continuous infusion of epidural bupivacaine and buprenorphine, postoperative pain scores of patients who had preoperative epidural injection of 10 ml of 0.5% bupivacaine (n = 15) were lower than those of saline injected patients (n = 15), but those of patients with a preoperative epidural injection of 10 ml of 0.25% bupivacaine (n = 15) were not lower than those of saline injected patients.