Review articles, systematic reviews and meta-analyses: which can be trusted?

A large number of scientific articles published every year requires from practicing physicians the ability to choose among them and to use secondary studies, such as guidelines, review articles, meta-analyses and systematic reviews. The aim of this article was to discuss basic differences between meta-analyses and systematic reviews. Meta-analysis is a mathematical method of pooling the results of several or more studies; a meta-analysis may be based on a systematic review, but this is not always the case. A systematic review is a multistage process aimed at the identification of all reliable evidence regarding a specific clinical problem. Systematic reviews make it possible to objectively address particular issues according to the current state of clinical knowledge and therefore constitute a reliable basis for clinical decision-making. An appropriate systematic review should include: 1) a defined clinical question, 2) pre-specified inclusion and exclusion criteria, 3) complex search for medical evidence sources according to a search strategy, 4) critical evaluation of reliability of identified clinical trials, 5) qualitative or quantitative data synthesis and 6) evidence based conclusions. These simple criteria, formulated by Cook et al. more than 10 years ago, allow to differentiate between a reliable systematic review and a "quasi-systematic" one, as well as between a reliable meta-analysis based on a systematic review and a potentially misleading meta-analysis without a systematic review.