Most Trial Eligibility Criteria and Patient Baseline Characteristics Do Not Modify Treatment Effect in Trials Using Targeted Therapies for Rheumatoid Arthritis: A Meta-Epidemiological Study.

To determine if variations in trial eligibility criteria and patient baseline characteristics could be considered effect modifiers of the treatment response when testing targeted therapies (biological agents and targeted synthetic disease modifying antirheumatic drugs (DMARDs)) for rheumatoid arthritis (RA).We conducted a meta-epidemiological study of all trials evaluating a targeted therapy approved by regulatory authorities for treating RA. The database search was completed on December 11th 2013. Eligible trials reported ACR20 data at months 3-6 and used an add-on design. Odds ratios (ORs) were calculated from the response rates and compared among the trial eligibility criteria/patient baseline characteristics of interest. Comparisons are presented as the Ratio of Odds Ratios (ROR).Sixty-two trials (19,923 RA patients) were included in the primary analyses using ACR20 response. Overall, targeted therapies constituted an effective treatment (OR 3.96 95% confidence interval (CI) 3.41 to 4.60). The majority of the trial eligibility criteria and patient baseline characteristics did not modify treatment effect. The added benefit of targeted therapies was lower in trials including "DMARD-naïve" patients compared with trials including "DMARD inadequate responders" (ROR = 0.45, 95%CI 0.31 to 0.66) and trials including "targeted therapy inadequate responders" (0.50, 95%CI 0.29 to 0.87), test for interaction: p = 0.0002. Longer mean disease duration was associated with a higher likelihood of responding to treatment (β = 1.05, 95%CI 1.00 to 1.11 OR's per year; p = 0.03). Analyses conducted using DAS28-remission as the outcome supported the above-mentioned findings.Our results suggest that a highly selective inclusion is not associated with greater treatment effect, as might otherwise be expected. The added benefit of a targeted therapy was lower in trials including patients who were DMARD-naïve and trials including patients with shorter disease durations.