The therapeutic equivalence and safety of treatment for 21 days with 400 mg t.i.d. oxaceprol (n = 132) and 50 mg t.i.d. diclofenac (n = 131) were assessed in a multicentre, randomised, double-blind study of a mixed population of patients with osteoarthritis of the knee and/or hip. In a per-protocol analysis of efficacy, the mean Lequesne index decreased by 2.5 points in the oxaceprol group (n = 109) and by 2.8 points in the diclofenac group (n = 109). The 95% confidence interval for the end-point difference revealed therapeutic equivalence. This was confirmed by assessments (visual analogue scale) of pain at rest, weight-bearing pain, pain on standing and pain on movement, all of which decreased to a similar extent under both treatments. The pain-free walking time increased in both groups from 10 min to 25 min by the end of the treatment period. Mobility was also increased to a similar extent by both drugs. The physicians assessed treatment as good or very good in 45-46% of patients in both groups. In all patients who received treatment, 28 and 37 adverse events were reported by 25 out of 132 (18.9%) and 33 out of 131 (25.2%) patients treated with oxaceprol and diclofenac, respectively. In 15 patients (11.4%) with 15 adverse events in the oxaceprol group and 25 patients (19.1%) with 27 adverse events in the diclofenac group, a relation to the medication was considered probable. The difference between the groups was statistically significant (p = 0.04106) for the number of these adverse events. Oxaceprol is therapeutically equivalent to diclofenac, but better tolerated than diclofenac in the treatment of osteoarthritis.