Abstract
PURPOSE: Type 2 diabetes mellitus (T2DM) is the subject of numerous randomized controlled trials (RCTs). The validity of RCTs may be threatened by attrition bias due to the discontinuation of the study. The aim of this systematic review is to evaluate the reasons of patient's withdrawal from these RCTs. METHODS: A systematic literature search on PubMed, Cochrane Library, Web of Science, and Scopus databases was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. The aim was to obtain all relevant blinded RCTs published before January 2017 in which the effectiveness of synthetic drugs, vitamins/minerals were compared to that of placebo or active control in T2DM. The quality of RCTs was assessed using the Jadad score. The frequency of withdrawal reasons was presented based on treatments with placebo/active control, national/international level of the studies, and publication year. Meta-analysis was not performed due to the heterogeneity. RESULTS: Overall, 1368 articles comprising of 640,780 subjects were included. In the majority of the RCTs (75.0%), the intervention and the placebo arms were compared. Most of the included studies (96%) were classified in the high-quality category (Jadad score≥3). The highest proportion of reported withdrawal cases was found in international studies, national RCTs conducted in Japan, and RCTs published in 2011. The withdrawal reasons were reported for 91,669 (63.75%) of the total 143,794 participants who had withdrawn from these studies. The main reported reasons were "adverse effects" (24.04%), "withdraw consent" (16.10%), and "missing data" (11.08%). Variations in the reported withdrawal reasons were based on the country or published year. RCTs with triple blinded design as well as those in which anti-hyperlipidemia and anti-obesity medications were applied, showed significantly higher probability of reported the withdrawal. CONCLUSION: High proportion of reported discontinuation in blinded RCTs on patients with T2DM was related to drug adverse effects. Overall, the total number and reason of drop out were unsatisfactory.