Toxicity of anti-rheumatic drugs in a randomized clinical trial of early rheumatoid arthritis

Abstract

Objective. To evaluate the toxicity of slow-acting anti-rheumatic drugs (SAARDs) and non-steroidal anti-inflammatory drugs (NSAIDs) in early rheumatoid arthritis. Methods. Patients were randomized to receive a SAARD - hydroxychloroquine (HCQ; n = 120), i.m. gold (n = 114) or methotrexate (MTX; n = 118) - or a NSAID only (n = 67). Patients in the three SAARD groups were allowed to take NSAIDs. Follow-up included 545 patient-years (p-yr). Adverse effects were attributed to specific medications using the Naranjo scoring method. Results. Fifty-five per cent of the patients suffered from adverse effect(s). Adverse effects were most common during i.m. gold therapy (87 per 100 p-yr), which led to permanent discontinuation of this treatment in 31 cases. The incidences of adverse effects that were probably attributable to NSAIDs in patients treated simultaneously with a SAARD were similar for the three SAARD groups. The mean period until the first adverse effect was longer in the MTX group (39 weeks) than in the HCQ group (27 weeks). Baseline clinical and sociodemographic parameters were not predictive of the occurrence of adverse effects. Conclusion. No adverse effect could be classified as definitely related to either SAARDs or NSAIDs by the Naranjo scoring method. The incidence of possible adverse effects of NSAIDs and SAARDs was 72 per 100 p-yr, and adverse effects led to permanent discontinuation of the therapy in 56 cases (13%) (31 patients receiving i.m. gold, 12 receiving MTX, 10 receiving HCQ and three receiving NSAID only). © 2000 British Society for Rheumatology.