Meta-analysis on the dissection table

Abstract

Introduction

Determination of the best evidence involves a systematic, comprehensive review and synthesis of the research literature as performed in high-quality meta-analysis. In a recently published multiple-treatment meta-analysis by Cipriani and collegues (Lancet, 2009) the authors finally concluded to favour sertralin over escitalopram due to lower costs.

Objectives/aims

Based on the example published by Cipriani et al. we tried to highlight possible pitfalls, the clinican with limited time might not be able to spot easily whilst skimming the results of meta-analysis.

Results

Escitalopram had the best efficacy and tolerability profile in the study. Sertraline was recommanded as first choice because of the most favourable balance between benefits, acceptability and acquisition costs. Analyzing study-design, statistical procedures and knowledge such as the early launching of generics of escitalopram lead to a different view of the interpretation by Cipriani and collegues.

Conclusions

We concluded that the findings of Cipriani et al. have limited generalizability due to problematic study design issues including statistical concerns. Despite the fact that meta-analysis might provide best evidence-based information, the findings of the latter should always be interpreted with caution as the quality of them can only be as good as the studies included and the study design implemented.