We evaluated the effect of ketamine as an adjunct to a fentanyl-based i.v. patient-controlled analgesia (IV-PCA) on postoperative nausea and vomiting (PONV) in patients at high risk of PONV undergoing lumbar spinal surgery. Fifty non-smoking female patients were evenly randomized to either the control or ketamine group. According to randomization, patients received either ketamine 0.3 mg kg−1 i.v. or normal saline after anaesthetic induction with fentanyl-based IV-PCA either with or without ketamine mixture (3 mg kg−1 in 180 ml). The incidence and severity of PONV, volume of IV-PCA consumed, and pain intensity were assessed in the postanaesthesia care unit, and at postoperative 6, 12, 24, 36, and 48 h. The overall incidence of PONV during the first 48 h after surgery was similar between the two groups (68 vs 56%, ketamine and control group, P=0.382). The total dose of fentanyl used during the first 48 h after operation was lower in the ketamine group than in the control group mean (sd), 773 (202) μg vs 957 (308) μg, P=0.035. The intensity of nausea (11-point verbal numerical rating scale) was higher in the ketamine group during the first 6 h after operation median (interquartile range), 6 (3–7) vs 2 (1.5–3.5), P=0.039, postoperative 12–24 h 5 (4–7) vs 2 (1–3), P=0.014, and postoperative 36–48 h 5 (4–7) vs 2 (1–3), P=0.036. Pain intensities were similar between the groups. Ketamine did not reduce the incidence of PONV and exerted a negative influence on the severity of nausea. It was, however, able to reduce postoperative fentanyl consumption in patients at high-risk of PONV.