Background Therapeutics generate non-specific effects (the placebo effect), and specific ones, like the 5 tumour necrosis factor (TNF) inhibitors used in the treatment of Rheumatoid Arthritis (RA). Few abstracts deal with the placebo effect. Its role in clinical responses to treatments has not been clearly accounted for. A systematic review must henceforth be conducted to appraise its rate in TNF-blockers treatments. This rate seems quite important in the first 3 months of treatment.
Objectives The aim of this study was to estimate the placebo effect in ACR 20, 50 and 70 responders in patients suffering from RA and being treated by TNF-blockers after MTX failure.
Methods A systematic literature review was conducted using PubMed, Embase and Cochrane library databases. The articles selected till March 2013 reported on double-blind RCTs of TNF-blockers versus placebo with patients receiving concomitant MTX. These patients, suffering from RA according to ACR 1987 criteria, were naive of biotherapies. MTX therapy with a minimum dosage of 10 mg/week had proved unsuccessful for at least 6 months. The data collected dealt with RA patients and the rates of ACR 20, 50 and 70 responders in each group at week 24. A meta-analysis was conducted using Methodomics for global responders, each molecule and injection method.
Results Twenty-two RCTs out of 1,386 were included in the systematic review and meta-analysis: 6 studied infliximab, 5 etanercept, 4 adalimumab, 3 certolizumab, and 4 golimumab. ACR 20 responses at W24 were available in 13 RCTs. The placebo effect was 24.70% on the whole (IC95% [0.192;0.301] p=0) with significant statistical results for adalimumab: 26.70% (IC95% [0.161;0.373] p=0.001), certolizumab: 16.40% (IC95% [0.068;0.26] p=0), infliximab: 31.90% (IC95% [0.154;0.485] p=0.001). The subgroup analysis concerning the injection method showed 31.90% vs. 23.30% for the IV route (IC95% [0.154;0.485] p=0.001). The ACR 50 non-specific response at W24 was 10.90% (IC95% [0.074;0.144] p=0) for the 14 RCTs. It was only significant for adalimumab, estimated at 15.40% (IC95% [0.026;0.281] p=0). It was not increased by the IV relative to the SC routes. The ACR 70 placebo effect response at W24 was estimated at 3.40% (IC95% [0.019;0.049] p=0) in 13 studies. No result was statistically significant. The more the ACR criteria were stringent, the more the placebo effect decreased. Heterogeneity was important due to both MTX and TNF-blockers different dosages and administration frequencies. The rates were analysed by the random model.
Conclusions Even if RCTs showed that TNF-blockers were more efficient than MTX alone, this meta-analysis shows that the non-specific effect must be taken into account. It is estimated at about 20% even if patients are in failure with the treatment. Adalimumab, certolizumab and infliximab rose above others in placebo effect responses, as well as for the IV route. This placebo effect seems to decrease with strict evaluation criteria. It would be interesting to better consider its role in clinical responses aspects and to take into account the cost margin.
Disclosure of Interest None declared