Objectives: The primary aim of this work was to compare the efficacy of valdecoxib 10, 20, and 40 mg QD with that of placebo and naproxen 500 mg BID in patients with rheumatoid arthritis (RA). The overall safety and tolerability profiles of valdecoxib and naproxen were also compared. Methods: A 12-week, multicenter, randomized, double-blind, parallel-group, placebo- and active-controlled study was performed in patients with adult-onset RA whose disease was in a flare state after discontinuing NSAIDs or other analgesics. Patients were randomly assigned to valdecoxib 10, 20, or 40 mg QD, naproxen 500 mg BID, or placebo. The primary efficacy measures were the American College of Rheumatology (ACR) 20% responder index (ACR-20), physicians' assessments of tender/painful joint count and swollen joint count, and patients' and physicians' global assessments of disease activity. Adverse events, clinical laboratory data, and vital signs were assessed by the investigator and compared between treatment groups to evaluate overall tolerability and safety. Results: A total of 1093 patients were randomized to receive either valdecoxib 10 mg QD (n=226), valdecoxib 20 mg QD (n=219), valdecoxib 40 mg QD (n=209), naproxen 500 mg BID (n=219), or placebo (n=220). At all time points, the proportion of ACR-20 responders was significantly higher in the valdecoxib groups than the placebo group at weeks 2 (10 mg, P<0.001; 20 mg, P=0.008; 40 mg, P= 0.004), 6 (all, P<0.001), and 12 (10 mg, P=0.006; 20 mg, P=0.004; 40 mg, P<0.001). Similarly, at all time points, the proportion of ACR-20 responders was significantly higher in the naproxen 500-mg group than the placebo group (all time points, P<0.001). In addition, mean changes in the number of tender/painful joint counts were significantly greater in the valdecoxib groups than the placebo group at weeks 2 (all, P<0.001), 6 (10 mg, P=0.002; 20 and 40 mg, P<0.001), and 12 (10 mg, P=0.004; 20 mg, P= 0.012; 40 mg, P<0.001). Naproxen treatment was also associated with greater reductions in tender/painful joint count than placebo (all, P<0.001). Mean changes in swollen joint count decreased at all time points in all groups, with significantly greater changes in the valdecoxib and naproxen treatment groups than the placebo group (valdecoxib 20 and 40 mg: week 6, P= 0.014 and P=0.003, respectively; naproxen: week 2, P=0.014; week 6, P=0.015; week 12, P=0.030). Physicians' global assessments of disease activity scores were significantly lower in the valdecoxib (10 mg: weeks 2 and 6, P<0.001; week 12, P=0.001; 20 and 40 mg: all weeks, P<0.001) and naproxen (all time points, P<0.001) treatment groups than the placebo group. Adverse events were reported by 45.5% patients in the placebo group, 51.8% in the valdecoxib 10 mg QD group, 58.0% in the valdecoxib 20 mg QD group, 56.9% in the valdecoxib 40 mg QD group, and 62.6% in the naproxen 500 mg BID treatment group. Conclusions: Valdecoxib 10, 20, and 40 mg QD were efficacious for treating the signs and symptoms of RA in these patients. The efficacy of valdecoxib 20 and 40 mg QD was not significantly different from that of naproxen 500 mg BID. Valdecoxib was generally well tolerated in this study.