We studied the preventive effect on postoperative nausea and vomiting (PONV) of ondansetron, metoclopramide and placebo associated with epidural anaesthesia. Sixty children, ASA I or II, 4 to 12 yr old underwent surgery for inguinal hernia repair (n = 30) or orchidopexy (n = 30). Children were randomly assigned to a postinduction intravenous medication group, ondansetron (5 mg.m-2), metoclopramide (0,12 mg.kg-1) or a saline solution placebo. After a general anaesthesia was obtained with halothane, nitrous oxide and oxygen delivered by mask, caudal or lumbar epidural anaesthesia was performed with plain mepivacaine. General performed with plain mepivacaine. General anaesthesia was interrupted and light narcosis maintained with diazepam during surgery. There were no significant differences in age and weight between the three groups. There were no adverse reactions to either ondansetron or metoclopramide. vomiting was not present in ondansetron group. The incidence of postoperative emesis in the metoclopramide group was 25%, whereas that of placebo group was 10%. The administration of ondansetron was associated with a lower (P = 0.017) incidence of postoperative vomiting if compared to the metoclopramide group. In conclusion ondansetron given preoperatively had proven to be an effective treatment for PONV after epidural block for lower abdominal surgery.