A clinical trial was conducted to evaluate the postoperative analgesic efficacy and the safety of intrathecal neostigmine in patients undergoing anterior and posterior vaginoplasty under spinal anaesthesia. Thirty-six patients were randomly divided into three groups to receive: normal saline (1 ml), morphine (100 micrograms in 1 ml of saline) or neostigmine (100 micrograms in 1 ml of saline) intrathecally just before a spinal injection of hyperbaric bupivacaine (0.5%, 4 ml). The mean [SD] time to the first analgesic (nonsteroidal anti-inflammatory drug) administration was significantly prolonged by intrathecal neostigmine (10.7 [4.3] h) and morphine (15.3 [3.0] h) compared with saline (4.5 [1.0] h). The three groups also differed in the number of patients requiring subcutaneous morphine to complement the analgesia provided by the intramuscular nonsteroidal anti-inflammatory drugs and the mean [SD] times for their administration: eight patients in the saline group (8.0 [3.8] h), one patient in the morphine group (18 h) and two patients in the neostigmine group (8 and 12.9 h). The morphine and neostigmine groups showed similar analgesic effectiveness. The characteristics of spinal anaesthesia were not modified by intrathecal morphine or neostigmine. Severe nausea and vomiting, sweating and distress during surgery were the most obvious adverse effects of intrathecal neostigmine. On the other hand, less hypotension was observed in the neostigmine group. The usefulness of intrathecal neostigmine as the sole postoperative analgesic may be restricted by the severity of its adverse effects.