To analyse available evidence on the efficacy and safety of anti-TNFα drugs (infliximab, etanercept and adalimumab) for treating rheumatoid arthritis (RA).
We searched systematically for randomised controlled clinical trials on treatment of RA with anti-TNFα drugs, followed by a systematic review with metaanalysis. Trials were searched from MEDLINE, EMBASE and Cochrane Library databases. The American College of Rheumatology (ACR) efficacy response criteria were used. Safety parameters provided by the trials were also assessed. Positive and undesired effects were estimated using combined relative risks (RR), number needed to treat (NNT) and number needed to harm (NNH). Heterogeneity was evaluated by Cochrane's Q and I2 statistics.
Thirteen trials (7087 patients) met the inclusion criteria. The combined RR to achieve a therapeutic response to treatment with recommended doses of any anti-TNFα drug was 1.81 (95% CI 1.43–2.29) with a NNT of 5 (5–6) for ACR20. NNT for ACR50 [5 (5–6)] and ACR70 [7 (7–9)] were similar. Overall therapeutic effects were also similar regardless of the specific anti-TNFα drug used and when higher than recommended doses were administered. However, lower than recommended doses elicited low ACR70 responses (NNT 15). Comparison of anti-TNFα drugs plus methotrexate (MTX) with MTX alone in patients with insufficient prior responses to MTX showed NNT values of 3 for ACR20, 4 for ACR50 and 8 for ACR70. Comparison of anti-TNFα drugs with placebo showed a similar pattern. Comparisons of anti-TNFα drugs plus MTX with MTX alone in patients with no previous resistance to MTX showed somewhat lower effects. Etanercept and adalimumab administered as monotherapy showed effects similar to those of MTX. Side effects were more common among patients receiving anti-TNFα drugs than controls (overall combined NNH 27). Patients receiving infliximab were more likely to drop out because of side effects (NNH 24) and to suffer severe side effects (NNH 31), infections (NNH 10) and infusion reactions (NNH 9). Patients receiving adalimumab were also more likely to drop out because of side effects (NNH 47) and to suffer injection site reactions (NNH 22). Patients receiving etanercept were less likely to drop out because of side effects (NNH for control versus etanercept 26) but more likely to experience injection site reactions (NNH 5).
Anti-TNFα drugs are effective in RA patients, with apparently similar results irrespective of the drug administered. Doses other than those recommended are also beneficial. The main factor influencing therapeutic efficacy is the prior response to DMARD treatment. The effect of treatment with etanercept or adalimumab does not differ from that obtained with MTX. The published safety profile for etanercept is superior but the fact that no patients are treated with higher than recommended doses requires explanation.