Efficacy of biologics in the treatment of moderate to severe psoriasis: a network meta-analysis of randomized controlled trials

BACKGROUND: Ustekinumab, a novel monoclonal antibody for the treatment of moderate to severe plaque-type psoriasis, has recently received regulatory approval in Europe, bringing the total number of biologic agents licensed in this indication to five. To assist treatment selection in daily practice it is essential to understand the benefit/risk profile of these agents and in the absence of a clinical trial comparing all available biologics a number of reviews have used statistical techniques to generate estimates of the comparative effectiveness of these therapies through the available network of randomized clinical trials. These estimates have previously been published for a limited range of psoriasis biologic treatments, although, to date no review has compared all the currently available agents in Europe. OBJECTIVES: To estimate the comparative effectiveness of all biologic agents indicated in the treatment of moderate to severe psoriasis currently available in Europe based on the primary trial endpoints. METHODS: A number of databases were searched for details of randomized controlled trials of available biologics in the treatment of plaque-type psoriasis in adults. Comparative effectiveness was estimated based on the reported Psoriasis Area and Severity Index (PASI) 50, 75 and 90 response rates. A network meta-analysis conducted on the ordered probit scale and implemented as a Bayesian hierarchical model provided estimates for the probability of response and relative risk vs. placebo, based on all observed comparisons. RESULTS: Twenty trials were included in the meta-analysis including patients with a mean disease duration of 18-22years. Based on the indirect comparison and given a placebo PASI 50 response of 13%, infliximab had the highest predicted mean probability of response at PASI levels 50 (93%), 75 (80%) and 90 (54%), followed by ustekinumab 90 mg at 90%, 74% and 46%, respectively, and then ustekinumab 45mg, adalimumab, etanercept and efalizumab. CONCLUSIONS: The ordered probit model allowed a quantitative comparison of all currently licensed biologics, providing estimates on comparative effectiveness and a suggested ranking of treatments that is of potential use to decision-makers. However, the analysis is based on indirect comparisons of the primary endpoint reported from short-term randomized trials.