Abstract P667: Clinical Variables, Electrocardiogram Parameters and Blood Biomarkers Associated With Atrial Fibrillation Detection in the First Year After Ischaemic Stroke - A Systematic Review and Meta-Analysis

Abstract

Introduction: Cardiac monitoring is performed to detect atrial fibrillation (AF) after stroke. Identifying patients at high or low risk of AF may allow cardiac monitoring approaches to be tailored on a more personalised basis. We performed a systematic review and meta-analysis to identify variables associated with AF detection after ischaemic stroke. Methods: We followed the Cochrane Collaboration Guidelines and retrieved 8503 studies. After screening, 35 studies were selected and 68 variables were assessed. We assessed 41 clinical variables, 20 ECG parameters and 7 blood biomarkers associated with AF detection >30 seconds duration in the first year after stroke. Comprehensive Meta-analysis software was used to generate an odds ratio and Forest plot for each variable. Studies were assessed for quality using the Quality in Prognostic Studies (QUIPS) tool. Results: The 35 studies included 12010 patients and AF was detected in 1551 patients (13%). Of the 68 variables assessed, 20 were associated with increased odds of AF, 5 were associated with reduced odds of AF and 43 were not associated with AF (Figure 1). The variables most strongly associated with AF detection (odds ratio >3.00) were older age, patients who received IV thrombolytic therapy, maximum P-wave duration, premature atrial complexes, P-wave dispersion, P-wave index, QTc interval and brain natriuretic peptide. Risk of bias was low in 3 studies, moderate in 24 studies and high in 8 studies. Conclusions: We have identified clinically applicable variables that can stratify the probability of AF detection after stroke. Our results will help guide more personalised approaches to cardiac monitoring for AF detection after stroke.