To assess the efficacy and safety of celecoxib and naproxen in the treatment of acute ankle sprain. Double-blind, parallel-group, randomized trial. Multicenter outpatient. Adult patients (n = 397) with acute first-degree or second-degree ankle sprain. Patients randomized to celecoxib 200 mg b.i.d. (n = 198) or naproxen 500 mg b.i.d. (n = 198) for 7 days. Primary measures of efficacy were Patient's Assessment of Ankle Pain Visual Analogue Scale (VAS) and Patient's Global Assessment of Ankle Injury. Secondary efficacy measures included Physician's Global Assessment of Ankle Injury, Patient's Return to Normal Function/Activity, and Patients' and Physicians' Satisfaction Assessments. Adverse events (AEs) were reported by investigators during the study. For the primary endpoints at day 4, the mean pain VAS scores were 31.9 mm +/- 1.96 for celecoxib and 29.0 mm +/- 1.91 for naproxen, and the responder rate for Patient's Global Assessment of Ankle Injury was 71% in the celecoxib group and 72% in the naproxen group, differences that were not statistically significant. In addition, noninferiority analysis demonstrated treatment differences that were within prespecified minimal clinical important differences. Gastrointestinal AEs were the most common AE, accounting for 14% in the celecoxib group and 21% in the naproxen group. The incidence of dyspepsia was 3% for celecoxib compared with 12% for naproxen (P = 0.032). Celecoxib is as effective as naproxen in treating acute first-degree or second-degree ankle sprains but causes significantly less dyspepsia.