This 6-week study was conducted to test the efficacy, safety, and tolerability of rofecoxib (a selective COX-2 inhibitor) compared to nabumetone (a non-selective NSAID) and placebo in osteoarthritis (OA) patients aged 80 and older. Three hundred forty-one patients, mean age 83 years, were randomized. Allocations were made in an approximately 1:2:1:2 ratio (placebo: 12.5 mg rofecoxib: 25 mg rofecoxib: 1500 mg nabumetone). Least square mean changes from baseline in the primary efficacy endpoint, Patient Global Assessment of Disease Status, were as follows (with negative numbers indicating improvement): -14.85 mm for placebo; -25.34 mm for 12.5 mg rofecoxib; -25.40 mm for 25 mg of rofecoxib; and -25.95 mm for nabumetone (p<0.001 for all active treatments vs placebo.) Results from secondary endpoints, including the 3 WOMAC sub-scales (pain, stiffness, and disability) and the Investigator Global Assessment of Disease Status, were consistent with those for the primary endpoint. No significant between-group differences were observed in the proportions of patients who discontinued treatment due to either clinical or laboratory adverse experiences. Renal safety (edema and hypertension adverse experiences) was similar for rofecoxib and nabumetone. No gastroduodenal ulcers occurred; however, the demonstration of gastrointestinal risk with rofecoxib or nabumetone was beyond the scope of this trial. We conclude that in patients 80 years and older, rofecoxib, 12.5 mg and 25 mg once daily, demonstrated clinical efficacy for the treatment for OA as did 1500 mg of nabumetone. Rofecoxib and nabumetone were generally well tolerated in this elderly population.