A network meta-analysis of randomized controlled trials for comparing the effectiveness and safety profile of treatments with marketing authorization for relapsing multiple sclerosis

What is known and objectiveThe relative effectiveness and safety profile of the treatments with marketing authorization for relapsing multiple sclerosis (MS) are not well known because randomized controlled trials with head-to-head comparisons between these treatments do not exist. Thus, a network of multiple-treatments meta-analysis was performed using four clinical outcomes: patients free of relapse', patients without disease progression', patients without MRI progression' and patients with adverse events'. MethodsRandomized controlled trials (RCTs) on MS were systematically searched in PubMed and Cochrane Central Register of Controlled Trial. The network analysis performed pairwise comparisons between the marketed treatments (Betaferon 250mcg, Avonex 30mcg, Rebif 44mcg, Rebif 22mcg, Aubagio 7mg, Aubagio 14mg, Copaxone 20mg, Tysabri 300mg, Gilenya 0<bold>5</bold>mg and Novantrone 12mg/m(2)) using direct and indirect analyses. Results and discussionThe analysis included 48 articles, involving 20455 patients with MS. The direct analysis showed better response for more than one outcome for Gilenya compared with Avonex (patients free of relapse' and patients without MRI progression') and for Betaferon compared with Avonex (patients without disease progression' and patients without MRI progression'). The indirect analysis indicated that Tysabri may have better relative effectiveness compared with the other treatments for two outcomes: patients free of relapse' and patients without MRI progression'. Regarding patients with adverse events', no data were available for all comparisons to make fair inferences. What is new and conclusionThis was an attempt, for the first time, to compare the efficacy and safety profile of existing approved treatments for relapsing MS. Although some treatments have shown better response, the results of the network analysis should be interpreted with caution because of the lack of RCTs with head-to-head comparisons between treatments.