BACKGROUND The aim of this study was to compare epidural infusion of bupivacaine and fentanyl and intermittent epidural morphine with regard to analgesic effect, and incidence and severity of side effects in children undergoing major abdominal or genito-urological surgery in order to improve the postoperative pain management of children. METHODS A double-blind, block-randomised study design was used. Thirty-one children aged 3 months to 6 years undergoing major abdominal or genito-urological surgery were studied. After induction of anaesthesia a lumbar epidural catheter was placed at L3-4 or L4-5. Postoperatively, the children received either 30 micrograms/kg of morphine every 8 h or a continuous infusion of fentanyl 2 micrograms/ml and bupivacaine 1.0 mg/ml at a rate of 0.25 ml.kg-1.h-1. All children additionally received rectal paracetamol in doses of 50-100 mg.kg-1.d-1 on a regular basis, and amol in doses of 50-100 mg.kg-1.d-1 on a regular basis, and if necessary supplementary intravenous morphine in doses of 50 micrograms/kg. Postoperatively, pain, administration of supplemental morphine and side effects were recorded 5 times by one observer during the day of surgery and the first postoperative day. All children had an epidural catheter throughout the study period. RESULTS Both regimens provided effective analgesia, but significantly better pain relief was obtained in children receiving the fentanyl/bupivacaine regimen. Sedation, pruritus, vomiting, and administration of antiemetics were seen in both treatment groups, and even though both the incidence and severity of side effects tended to be higher in children receiving morphine, no statistically significant difference was found. No episodes of respiratory depression or motor blockade were noticed. CONCLUSIONS Continuous epidural infusion of fentanyl and bupivacaine was found to be superior to intermittent epidural morphine. The initial regimen should be fentanyl 2 micrograms/ml and bupivacaine 1.0 mg/ml infused at a rate of 0.25 ml.kg-1.h-1.