Methodological Quality Assessment of Meta-Analyses of Hyperthyroidism Treatment

Abstract

Hyperthyroidism is a common condition that is associated with increased morbidity and mortality. A number of meta-analyses (MAs) have assessed the therapeutic measures for hyperthyroidism, including antithyroid drugs, surgery, and radioiodine, however, the methodological quality has not been evaluated. This study evaluated the methodological quality and summarized the evidence obtained from MAs of hyperthyroidism treatments for radioiodine, antithyroid drugs, and surgery. We searched the PubMed, EMBASE, Cochrane Library, Web of Science, and Chinese Biomedical Literature Database databases. Two investigators independently assessed the meta-analyses titles and abstracts for inclusion. Methodological quality was assessed using the validated AMSTAR (Assessing the Methodological Quality of Systematic Reviews) tool. A total of 26 MAs fulfilled the inclusion criteria. Based on the AMSTAR scores, the average methodological quality was 8.31, with large variability ranging from 4 to 11. The methodological quality of English meta-analyses was better than that of Chinese meta-analyses. Cochrane reviews had better methodological quality than non-Cochrane reviews due to better study selection and data extraction, the inclusion of unpublished studies, and better reporting of study characteristics. The authors did not report conflicts of interest in 53.8% meta-analyses, and 19.2% did not report the harmful effects of treatment. Publication bias was not assessed in 38.5% of meta-analyses, and 19.2% did not report the follow-up time. Large-scale assessment of methodological quality of meta-analyses of hyperthyroidism treatment highlighted methodological strengths and weaknesses. Consideration of scientific quality when formulating conclusions should be made explicit. Future meta-analyses should improve on reporting conflict of interest.

Protocol registration or development may benefit the design, conduct and reporting of dose-response meta-analysis: empirical evidence from a literature survey

“…Given the significantly imbalanced distribution in the geographic region and year of publication between DRMAs with protocol registration or development versus those not, we used propensity score matching method, with nearest neighbor approach, in a 1:2 ratio of the two types of DRMAs [27]. We divided geographic region as Asian and Non-Asian (European, America, and Australia) for the matching since previous literatures demonstrated that meta-analyses and randomized controlled trials by author from Asian tends to poorly conducted [4, 28]. We used 1:2 ratio due to the consideration of both the available sample size and the optimal matching (i.e.…”