Stroke risk associated with atrial fibrillation: a Burden of Proof study.

Atrial fibrillation (AF) is increasingly recognized as a significant predictor for stroke, carrying profound implications for morbidity and mortality globally. This study aims to explore the relationship between AF and stroke incidence and/or stroke mortality (in 30 days) by employing the Burden of Proof methodology. Utilizing a systematic review and meta-analysis of published observational studies from PubMed and EMBASE databases and employing the Burden of Proof approach, this study identified a large dataset to estimate conservatively the mean AF-stroke risk functions and the Burden of Proof risk function (BPRF), estimating pooled mean relative risks (RR), evaluating systematic biases, quantifying heterogeneity, and assessing publication bias. The MR-BRT (Meta-regression-Bayesian, Regularized, Trimmed) model was utilized for its advanced capability to address systematic biases and heterogeneity. After reviewing 6,447 de-duplicated articles from 8,527, we included and extracted 61 reported risk effect size measures from 49 articles and run the MR-BRT model. This analysis indicate that AF presents an increased risk factor for overall stroke, estimated mean relative risk (RR) for overall stroke incorporating between-study heterogeneity (γ) is 1.99 (95% uncertainty interval (UIs) 1.91, 2.08). The comprehensive Burden of Proof analysis also revealed the Risk-outcome score (ROS = 0.32), which suggests that AF contributes to an average increase in stroke risk by 38.29% compared to person with no-AF (the percentage of mean relative risk increases here revealed using (Exp (ROS) – 1) x 100) %. As per the ROS, the interpreted strength of this harmful relationship to be a moderate and three-star rating out of five. This analysis also considered publication biases (not found any, in this analysis) and interaction of significant bias covariates in the analysis for reporting conservatively estimated the BPRF (most conservative estimate of AF-stroke harmful association, 5th percentile of mean RR) which is 1.92 (95% UIs 1.90, 1.99). This study confirms the substantial role of AF in stroke outcome; however, the Burden of Proof framework facilitated a nuanced analysis that enhances our comprehension of the AF-stroke association. We anticipated this relationship to be stronger than a three-star rating (intuitive from the mean relative risk), emphasizing the need for synthesizing evidence for implementing public health strategies and informed policymaking to alleviate the burden of stroke attributable to AF.