Seventy-one patients presenting for day-case arthroscopy were randomly allocated to receive either intramuscular diclofenac 75 mg or ketorolac 30 mg immediately after induction of anaesthesia. One hour after operation visual analogue pain scores were significantly lower in the ketorolac group compared with those receiving diclofenac. Pain scores at 2 and 4 h postoperatively were not significantly different between the two groups although six of those receiving diclofenac required opioid analgesia compared with only one in the ketorolac group. Discomfort in the operated knee was similar for both groups on the day following surgery, but pain from the intramuscular injection site was significantly greater in the diclofenac group. Intramuscular ketorolac 30 mg provided better postoperative analgesia and less pain at the injection site than diclofenac 75 mg.