SummaryA double-blind randomized trial was carried out in 60 patients with osteoarthrosis of the hip or knee to assess the relative efficacy and tolerance of treatment with diflunisal twice daily (maximum 750 mg per day) or with aspirin given 4-times daily (maximum 3 g per day) over a period of 12 weeks. Clinical assessments were made on entry and at regular intervals of weight-bearing pain, night pain, specific function pain, inactivity stiffness, and range of joint movement. A considerable proportion of patients in both groups showed improvement in all parameters except for limitation of movement. The difference in response between the two groups was not statistically significant, neither was the patients' overall opinion of response to or clinician's assessment of therapeutic efficacy of either drug treatment at the end of the trial. Ten of the 29 patients in the aspirin group had to withdraw because of adverse reactions, mainly gastro-intestinal, compared with 4 of the 31 patients in the diflunisal group. The overall incidence of side-effects in all patients was lower in the diflunisal group, and those that were reported were less disturbing. At the end of the double-blind study, patients were given the option to continue with the particular drug treatment for a further two 13-week periods. More patients chose to remain on diflunisal than on aspirin at the end of each of the three periods and the difference was statistically significant.