BACKGROUND: Ketamine hydrochloride, NMDA receptor antagonist is a potent analgesic and anesthetic. Other analgesics, like opioid, have been shown to effectively relieve postoperative pain when infused into epidural space, but effects of ketamine hydrochloride infused into epidrual space for postoperative pain control is still controversial, and therefore the present study was undertaken. METHODS: Ninety adult patients (ASA I or II) scheduled for upper abdominal and chest surgery were randomized into ketamine and fentanyl groups. For all patients, informed consent was obtained preoperatively. Anesthesia was induced with thiopental sodium/succinylcholine and maintained with nitrous oxide/oxygen/enflurane. Skeletal muscle relaxation was maintained with vecuronium. Epidural catheterization was done after operation. Ketamine group received epidural bolus of 0.1% bupivacaine 10 ml followed by continuous epidural infusion of 0.1% bupivacaine 100 ml containing ketamine 200 mg. Fentanyl group received epidural bolus of 0.1% bupivacaine 10 ml containing fentanyl 100 microgram followed by continuous epidural infusion of 0.1% bupivacaine 100 ml containing fentanyl 600 microgram. Continuous infusion rate was 2 ml/hr in both groups. Analgesic effects were assessed using VAS (visual analogue score), PHS (Prince Henry score) and PRS (pain relief score). Side effects and number of patients using additional analgesics were evaluated. RESULTS: Analgesic effects were significant in both group after drug administration. But fentanyl group had greater analgesic effects than ketamine group. Fentanyl group experienced side effects such as pruritus (27 cases), nausea and vomiting (9 cases). Ketamine group had side effects such as nausea and vomiting (13 cases). Number of patients using additional analgesics were seven and twenty-four in the fentanyl and ketamine groups, respectively. CONCLUSIONS: We conclued continuous epidural infusion of ketamine had fewer analgesics effect at early state of postoperative pain than fentanyl.