Effectiveness of direct oral anticoagulants in obese adults with atrial fibrillation: an overview examining the evidence from international systematic reviews and meta-analyses

Abstract Background Obesity may influence the pharmacology of direct oral anticoagulants (DOAC) that are recommended by all international guidelines for stroke prevention in adults with atrial fibrillation (AF). Purpose To evaluate the safety and efficacy of DOACs in obese adults with AF. Methods Medline, CINAHL, Scopus, Web of Science, Cochrane Database, Johanna Briggs Institute, Embase, Psych Info and ProQuest were searched till April 2020. Only systematic reviews with meta-analyses, that were published after 2005 and were in the English language were included. Articles were screened by title and abstract, followed by full text assessment using the Covidence systematic review software. Data was extracted using a standardised extraction tool. AMSTAR-2® and ROBIS® tools were used for quality and risk of bias assessment. The entire process was undertaken by two investigators at each stage of the study selection, appraisal, and data extraction. Disagreements were resolved through consensus discussion with a third arbitrary investigator. Statistical analyses were performed using the DerSimonian and Laird method for random effects. Meta-analysis was performed using only randomised controlled trials from eligible systematic reviews at both 12 months and across the entire trial. Primary outcomes assessed was stroke (ischemic or haemorrhagic) or systemic or pulmonary embolism. Secondary outcomes assessed included all-cause mortality, transient ischemic attack, myocardial infarction, major bleed, all cause-hospitalisation, and cardiovascular mortality. Results Of the 8162 articles screened, a total of five systematic reviews were included in this overview. There was disagreement within the published reviews on the effect of DOAC in obesity. Four of the five reviews were of either “low” or “critically low” quality, with inconsistencies in data extraction and appropriateness of the included studies and statistical methods used in analysis. Data from only the RE-LY, AVERROES and ENGAGE AF-TIMI 48 trials, were available for the meta-analysis, which did not find any significant difference between all BMI groups for all outcomes, at both time points. However, analysis of the different weight groups versus normal weight, highlighted non-significant differences between the different DOACs. Conclusion There was no difference between the BMI classes in any of the outcomes assessed. However, the non-significant trends that were seen, suggests individual superiority of DOACs