Nausea and vomiting remain unpleasant side effects of intrathecal (i.t.) morphine and of the numerous therapies tried, only prophylactic intravenous (i.v.) metoclopramide has been reported to be promising. Seventy-three patients, scheduled for orthopaedic prosthesis surgery of the hip or knee were studied. They received 4mL of plain bupivacaine and 0.3 mg of preservative-free morphine i.t. for anaesthesia. The test drugs given in a double-blind and randomized fashion, were either metoclopramide 20 mg, three times, at 6 h intervals (23 patients), ondansetron, 8 mg, twice, at 12 h intervals (25 patients), or 0.9% saline three times, at 6 h intervals (25 patients). The occurrence of nausea, vomiting and pain was followed for 24 h. The incidences of nausea and vomiting were 60% (15/l25) and 56% (14/25) in the saline group, 52% (12/23) and 48% (11/23) in the metoclopramide group, and 52% (13/25) and 40% (10/25) in the ondansetron group. Incidences of severe vomiting were 24, 35 and 12%, respectively. Eight patients in the saline group, seven in the metoclopramide and 10 in the ondansetron group did not need additional opioids for post-operative pain relief. We conclude that, metoclopramide and ondansetron were not better than saline in the prevention of post-operative emesis induced by intrathecal morphine.