In a multicenter, double blind, parallel group study, 233 patients with classical or definite rheumatoid arthritis (RA) were randomized to 12 weeks of either salsalate (salicylsalicylic acid, nonacetylated salicylate) or aspirin following disease flare. One hundred-fifty patients completed, 83 taking salsalate and 67 taking aspirin. Patients received initial doses of 3 g/day of salsalate or 3.6 g/day of aspirin. Doses were adjusted during the first 5 weeks for efficacy and tolerance. Both treatments were equally effective as measured by all the usual variables, but there was a higher incidence of severe gastrointestinal problems among patients taking aspirin. Thus, this study demonstrated that the acetyl group of aspirin does not enhance the antiinflammatory efficacy of salicylic acid in RA.