BACKGROUND: All currently available nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit both cyclooxygenase-1 and cyclooxygenase-2 and exhibit many complications. It has been suggested that the anti-inflammatory and also most of the analgesic effects of NSAlDs result from an inhibition of arachidonic acid metabolites synthesised via cyclooxygenase-2. In the present study, the extent of analgesic and anti-inflammatory effects of ibuprofen (a non-selective cyclooxygenase inhibitor), indomethacin (a selective cyclooxygenase-1 inhibitor) and NS-398 (a selective cyclooxygenase-2 inhibitor) are investigated in on acute model of arthritis in rats by a behavior test and pathologic examination. METHODS: Arthritis was induced with 2% kaolin and 3% carrageenan into the right knee joint cavity under enflurane anesthesia (2 - 4%). Before and after the injection, rats were allowed to walk freely through a pathway, constructed to record weight load by means of 8 weight sensors attached to 8 plates which function independently. Weight bearing, the weight of rat and the diameter of the knee joint were measured serially before and after the injection. At 6 hours after the injection, ibuprofen, indomethacin and NS-398 were injected intraperitoneally (1, 5 and 25 mg/kg/ml). RESULTS: In the control group, weight bearing decreased maximally and the weight bearing ratio increased maximally at 6 hours after inflammation and the diameter ratio increased maximally 1 day after inflammation. There were improvements in weight bearing with ibuprofen, indomethacin and NS-398 in a dose-dependent manner at 8, 10 and 12 hours after induction of arthritis. NS-398 demonstrated better analgesic and anti-inflammatory effects than ibuprofen or indomethacin at a low dose (1 mg/kg). In contrast to NS-398, significant analgesic effects of indomethacin on the behavior test was not shown at a low dose. CONCLUSIONS: These results suggest that the selective cyclooxygenase-2 inhibitor plays an important role as an analgesic and anti-inflammatory drug.