We reviewed 103 controlled clinical trials that compared the antiarthritic efficacy and tolerance of previously and currently marketed nonsteroidal antiinflammatory drugs (NSAIDs) and aspirin. Of 52 studies, 35 had data sufficient to calculate an NSAID efficacy index (the ratio of mean improvement in NSAID‐treated patients to that in aspirin‐treated patients) based on subjective and/or objective criteria. The mean indexes (obtained from all studies from which an index could be calculated) indicated no statistically significant difference in efficacy between aspirin and the NSAIDs as a group; the indexes tended to become less variable as the number of study subjects increased. Tolerance, assessed from the percentage of patients who discontinued the drug because of side effects, was significantly greater for NSAIDs than for aspirin. The NSAIDs had greater efficacy but not greater toxicity at increased doses. Efficacy differences described among NSAIDs in some studies were attributable either to comparisons at nonequivalent doses or to cha