Appraising the quality of meta-analysis for breast cancer treatment in the adjuvant setting: A systematic review

Abstract

IntroductionBreast cancer is the tumor with highest incidence in women worldwide and adjuvant treatment is extremely important to achieve disease control. Given the relevance of systematic reviews, their rigor should be warranted to avoid biased conclusions. Our objective was to investigate the methodological quality of meta-analysis of early breast cancer adjuvant treatment. Material and MethodsComprehensive searches were performed using electronic databases from 1/1/2007 to 11/12/2018. All studies identified as a systematic review with meta-analysis investigating the efficacy of breast cancer adjuvant treatments were included. Two reviewers independently assessed titles and abstracts, then full-texts for eligibility. Quality was assessed using the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) version 2 tool. ResultsOf 950 citations retrieved, 66 studies (7.0%) were deemed eligible. Methodological quality was highly variable, median AMSTAR score 8.5 (IQR 7–9.5) and range 0–16. There was a weak positive correlation between journal impact factor and AMSTAR score (r = 0.17) and citation rate and AMSTAR score (r = 0.16). Cochrane Systematic Reviews were of higher quality than reviews from other journals. Overall confidence was critically low for 61 (92.4%) studies, and the least well-reported domains were the statement of conflict of interest and funding source for the included studies (4.6%), the report of a pre-defined study protocol (15.2%), and the description of details of excluded studies (6.1%). ConclusionsOur findings reinforce concerns about the design, conduction and interpretation of meta-analysis in current literature. Methodological quality should be carefully considered and journal editors, decision makers and readers in general, must follow a critical approach to this studies.