The potential benefits of a nonsteroidal antiinflammatory drug to 67 patients undergoing knee arthroscopy were evaluated in a prospective, randomized, placebo-controlled, double-blinded study. Group A received the drug (diclofenac, 75 mg twice daily) for 3 to 5 days before and for 7 days after surgery. Group B received a placebo preoperatively and the drug postoperatively. Group C received a placebo at both times. Codeine was available postoperatively for all patients if needed. Outcomes reported by the subjects included pain, crutch use, and return to activities. Outcomes assessed by physicians included knee effusion, range of motion, and gait. Knee flexion and extension strengths were measured isokinetically pre- and postoperatively. Pain scores on the 1st postoperative day were higher in Group C than in Group A. Pain scores at all other time points were not significantly different in the three treatment groups. Groups A and B required less codeine during the first 72 hours after surgery than Group C (mean, 2.9 +/- 1.0 versus 6.8 +/- 1.0 pills). Recovery of function, recovery of strength, and physical examination parameters were not significantly different in the three treatment groups. Diclofenac was an effective analgesic in the immediate postoperative period. Recovery from arthroscopy, however, was not enhanced by taking the drug.