A double-blind, double-dummy controlled study to compare the clinical efficacy and gastric tolerability of salsalate and piroxicam in the treatment of rheumatoid arthritis was performed. A total of 23 patients were treated with 1.5 g salsalate twice daily and 20 with 20 mg piroxicam (after the evening meal) for 4 weeks. Patients were submitted to gastroscopy at the start and end of treatment; only patients who presented a normal baseline gastroscopy were admitted to the trial. At the end of the planned treatment period, a statistically significant improvement of all clinical variables was observed in both treatment groups, the difference between the two drugs not being statistically significant. Five of 20 (25%) piroxicam treated patients and only 2/19 (11%) salsalate treated patients showed gastric lesions at final endoscopy. No relationship was found between dyspeptic symptoms and endoscopic lesions. The results show that salsalate and piroxicam have equal efficacy in relieving arthritic symptoms, but salsalate causes fewer gastric lesions.