BACKGROUND Opioids can produce potent antinociceptive effects by interacting with local opioid receptors in inflamed peripheral tissue. In this study we examined the analgesic effects of the intraarticular, as compared with intravenous, administration of morphine after arthroscopic knee surgery. METHODS In a double-blind, randomized trial, we studied 52 patients who had received one of four injections at the end of surgery. The patients in group 1 (n = 18) received 1 mg of morphine intraarticularly and saline intravenously; those in group 2 (n = 15), saline intraarticularly and 1 mg of morphine intravenously; those in group 3 (n = 10), 0.5 mg of morphine intraarticularly and saline intravenously; and those in group 4 (n = 9), 1 mg of morphine and 0.1 mg of naloxone intraarticularly and saline intravenously. The volume of the intraarticular injections was 40 ml, and that of the intravenous injections was 1 ml. After 1, 2, 3, 4, 6, and 24 hours, postoperative pain was assessed with a visual-analogue scale, a numerical-rating scale, and the McGill pain questionnaire. The need for supplemental analgesic agents, the patients' vital signs, and the occurrence of side effects were monitored. RESULTS All pain scores were lower in group 1 than in group 2 at all times. The differences were significant (P less than 0.05) at three, four, and six hours (mean [+/- SD] visual-analogue score at six hours, 9 +/- 13 mm vs. 37 +/- 31 mm). The mean (+/- SD) consumption of supplemental analgesic medication per 24 hours was significantly lower in group 1 (36 +/- 51 mg of diclofenac and 1.2 +/- 3.4 mg of meperidine) than in group 2 (75 +/- 42 mg of diclofenac and 14 +/- 18 mg of meperidine, P less than 0.05). The visual-analogue scores in group 3 were slightly but not significantly higher than those in group 1 at all times except 6 and 24 hours after injection. The visual-analogue scores were significantly higher in group 4 than in group 1 one to four hours after injection (P less than 0.05), indicating that the analgesic effect of intraarticular morphine was reversible by naloxone. CONCLUSIONS Low doses of intraarticular morphine can significantly reduce pain after knee surgery through an action specific to local opioid receptors that reaches its maximal effect three to six hours after injection.