This 4-week, randomized, double-blind, double-dummy, placebo-controlled, parallel-group, multicenter study was designed to compare the efficacy and safety of etodolac and nabumetone in the treatment of patients with active osteoarthritis (OA) of the knee. Ninety-one patients received etodolac 400 mg twice daily, 89 received nabumetone 1500 mg once daily, and 90 received placebo. Both active treatments significantly improved the patients' condition relative to baseline (P < or = 0.001) at all evaluations during treatment and relative to placebo (P < or = 0.05) by visit 4. Improvement relative to placebo in investigator's global assessments was earlier in the etodolac group (ie, by visit 3) than in the nabumetone group. At visit 4, improvement in investigator's and patient's global assessment scores, and in the distribution of investigator's assessment scores, was significantly (P < or = 0.05) greater in the etodolac group than in the nabumetone group. Other than hypokalemia, which occurred only in three patients in the nabumetone group (P = 0.035), there were no significant differences among the groups in the frequency of study events or premature discontinuation from the study as a result of study events. Study events considered at least possibly treatment related were reported for 26 patients in the etodolac group (28.6%), 20 in the nabumetone group (22.5%), and 23 in the placebo group (25.6%). The most frequently reported symptoms for all groups were dyspepsia, nausea, and headache. Four patients treated with nabumetone (4.5%) had elevations in aspartate aminotransferase or alanine aminotransferase during treatment. The results of this study show that etodolac 400 mg twice daily is at least as effective as nabumetone 1500 mg once daily and is equally well tolerated in the treatment of patients with active OA of the knee; etodolac may have an earlier onset of action and/or a relatively greater efficacy in patient and investigator global assessments than nabumetone.