Network analysis of randomized controlled trials in multiple sclerosis

Abstract

Background: The optimal treatment of multiple sclerosis (MS) is not yet well-defined.

Objective: To estimate the relative effectiveness of treatments in MS, we performed a network of multiple-treatments meta-analysis of randomized controlled trials (RCTs) for relapsing MS using three main efficacy outcomes: relapse-free patients, patients without disease progression, and patients without magnetic resonance imaging progression.

Methods: We systematically searched PubMed and the Cochrane Central Register of Controlled Trials to identify English-language articles with RCTs that compared pharmaceutical treatments using the terms multiple sclerosis and randomized controlled trial. All RCTs that involved patients with definite relapsing MS and provided data for calculating the odds ratios for the main outcomes were considered. First, comparative effectiveness relative to placebo was assessed using direct analysis. Then, each therapy was compared with interferon beta-1b (250 μg)* in direct and indirect analyses. Effect sizes were estimated by applying a random-effects model.

Results: We identified 4165 titles; after screening, 109 articles were eligible for inclusion. In total, 26,828 patients were included. The network consisted of 145 treatments involving 59 direct comparisons with placebo and 8 direct comparisons with interferon beta-1b (250 μg). Two treatments showed better response compared with placebo (direct analysis) for all three efficacy outcomes: natalizumab (300 mg)(†) and fingolimod (0.5 mg). In comparing treatments with interferon beta-1b (250 μg), the network analysis revealed that no therapy shows better response for all 3 efficacy outcomes and alemtuzumab, 12 and 24 mg, have better response for 2 of the outcomes (relapse-free patients and patients without disease progression).

Conclusions: Although some treatments seem to have better efficacy, the results should be interpreted with caution because the network was dominated by indirect comparisons. Data from the selected studies included in the network cannot be extrapolated beyond them. Large RCTs that make head-to-head comparisons between treatments are needed to draw safe conclusions for the optimal treatment of MS.