BACKGROUND Perispinal anaesthesia for Caesarean section allows injection of epidural (ED) or intrathecal (i.t.) morphine to provide long-lasting postoperative analgesia. To compare these two routes, a prospective, randomized, double-blinded study of 53 patients undergoing elective Caesarean section was performed. METHODS Combined spinal-epidural anaesthesia with 6 mg of i.t. hyperbaric bupivacaine plus sufentanil 5 microg, and additional ED lidocaine was used. Additionally, each patient received either 2 mg (2 ml) of ED morphine plus 1 ml of i.t. normal saline (ED group, n=28), or 0.075 mg (1 ml) of i.t. morphine plus 2 ml of ED normal saline (i.t. group, n=25). Additional postoperative analgesia was given in the form of propacetamol and ketoprofen, plus self-administered i.v. morphine. RESULTS No major respiratory depression occurred. Time to first demand of morphine was similar in the ED (307.5 min) and i.t. (310 min) groups, as was the incidence of side-effects such as sedation, pruritus, nausea, and vomiting. During the first 24 postoperative hours, VAS pain scores were greater in the i.t. group (P=0.032), as was additional morphine consumption (4 vs 1.5 mg) (P=0.03). CONCLUSIONS The ED protocol was more effective than the i.t. protocol, whilst side-effects were similar.