BACKGROUND Some authors reported that a small dose of intra-articular morphine (1-5 mg) injection provided a potent and long-acting analgesic effect on pain after arthroscopic knee surgery. However, many in other reports did not agree to this result. Therefore, the characteristic of the analgesic effect of intra-articular morphine is worth evaluation. In order to rule out the systemic action of intra-articular morphine, we designed a study to compare the efficacy of 3 mg intra-articular morphine with that of 3 mg intravenous morphine in providing analgesia after arthroscopic knee surgery. METHODS Sixty patients undergoing arthroscopic knee surgery under spinal anesthesia were randomly divided into two groups (n = 30, each). At the end of surgery, group 1 received 3 mg intra-articular morphine in 15 ml isotonic saline and intravenous isotonic saline 10 ml; group 2 received 15 ml intra-articular isotonic saline and 3 mg intravenous morphine in 10 ml isotonic saline. Postoperative pain was evaluated using visual analogue scale (VAS) during an active or passive flexion of the operated knee joint. Also, additional analgesic consumption requested by patient postoperatively (morphine, 2 mg, iv, each) was recorded. RESULTS Pain scores were lower in group 1 than in group 2 from 4 to 36 h postoperatively (Mann-Whitney rank-sum test). Also, the consumption of supplemental analgesic morphine was lower in group 1 than in group 2 from 6 to 36 h postoperatively. CONCLUSIONS Intra-articular morphine (3 mg) provided long-lasting analgesia from 6 to 36 h postoperatively and this effect was not due to systemic absorption of morphine.