SummaryWe investigated whether combined dexamethasone and ondansetron is more effective than ondansetron alone in preventing postoperative nausea and vomiting in patients with fentanyl-based intravenous patient-controlled analgesia. One hundred and thirty patients undergoing videoassisted thoracoscopic surgery were assigned to either an ondansetron group or a dexamethasone and ondansetron group. In all patients, ondansetron 4 mg was administered at the end of surgery and 12 mg was added to the patient-controlled analgesia solution. The dexamethasone and ondansetron group received dexamethasone 8 mg at the induction of anaesthesia. The overall incidence of nausea and vomiting during the first 48 h postoperatively did not differ between groups (34/61 (56%) vs 28/62 (45%) in the ondansetron group and dexamethasone and ondansetron groups, respectively). The incidence of severe nausea and vomiting ( ‡ 7 nausea on an 11-point verbal numerical rating scale, retching or vomiting) was higher in the ondansetron group than in the dexamethasone and ondansetron group (15/61 (25%) vs 6/62 (10%, respectively, p = 0.028). Combined dexamethasone and ondansetron is more effective in reducing severe nausea and vomiting than ondansetron alone in patients receiving fentanyl-based intravenous patient-controlled analgesia. Postoperative nausea and vomiting (PONV) is one of the most common and distressing complications after surgery. The incidence of PONV varies from 10% to 78% according to surgical, anaesthetic and patient factors [1,2]. Intravenous patient-controlled analgesia (PCA) with opioids is an effective, safe method of postoperative analgesia. Video-assisted thoracoscopic surgery is associated with less postoperative pain than thoracotomy [3], and PCA with fentanyl is an alternative to invasive regional analgesia such as epidural or paravertebral block [4]. However, to be safe and acceptable, PCA in patients undergoing video-assisted thoracoscopic surgery should be accompanied by effective prophylaxis of nausea and vomiting.While several investigators have reported that prophylactic administration of ondansetron reduces PONV in patients using morphine-based PCA [5,6], the incidence of PONV incidence in patients using