Ketoprofen, 50+50+100 mg, was compared to naproxen, 250+250+250 mg, in a double-blind, cross-over twice 4-weeks' study on patients with RA. There was no significant difference in the effect on morning stiffness, pain at rest, joint count, grip strength or ESR. Among 28 patients 10 preferred ketoprofen and 7 naproxen. Two ketoprofen and 1 naproxen periods were interrupted owing to intolerable side-effects. Twenty of the patients experienced some side-effect from at least one drug. The most common complaints were gastrointestinal, seen in 12 patients on ketoprofen and 9 on naproxen. Most side-effects were mild. No abnormality in blood morphology, liver function tests, serum creatinine or fasting blood glucose was observed. No occult bleeding was detected on routine stool examination.