This study was designed to test the hypothesis whether preemptive administration of rofecoxib, a novel selective COX-2 inhibitor, can prolong intraarticular bupivacaine analgesia after arthroscopic knee surgery. Sixty-two patients were randomly assigned to one of the three groups. Group 1 (n=21) was administered oral rofecoxib 50 mg 1 h before surgery plus intraarticular 0.5 % bupivacaine 20 ml postoperatively. Group 2 (n=21) was administered the same dose of bupivacaine. Group 3 (n=20) was administered saline 20 ml intraarticularly after surgery. Pain scores (VAS) were assessed at 30 min, 1, 2, 4, 6, 12 and 24 h postoperatively. Analgesia duration, analgesic (tramadol and tenoxicam) requirements, and adverse effects were recorded postoperatively for 24 h. Pain scores were significantly lower in the Group 1 at all time points (p<0.05, p<0.001) and were significantly lower in the Group 2 at 30 min (p<0.001), 1 and 4 h (p<0.05) compared to the Group 3. Pain scores were significantly lower in the Group 1 compared to the Group 2 during the first 4 h after surgery (p<0.05, p<0.001). Analgesia duration was longer in Group 1 than Group 2 or 3 (743.0 ± 480.5 min versus 262.4 ± 292.2 min and 17.0 ± 12.1 min; p<0.05, p<0.001 respectively), and in Group 2 than Group 3 (p<0.05). Tramadol requirements were significantly less in Group 1 than Group 2 and 3 (4.8 ± 15.0 mg versus 40.5 ± 43.6 mg and 67.5 ± 24.5 mg; p<0.05, p<0.001 respectively), and in Group 2 than Group 3 (p<0.05). There were no significant differences among the groups regarding the tenoxicam requirements and adverse effects. In conclusion, the combination of oral rofecoxib administered preemptively and intraarticular bupivacaine administered postoperatively provided a significant analgesic benefit and decreased the opioid requirements after arthroscopic knee surgery, when compared to bupivacaine alone or saline.