Abstract 12656: Atrial Speckle Tracking Echocardiography for Detecting Atrial Fibrillation in Cryptogenic Stroke: A Meta-Analysis

Abstract

Introduction: Cryptogenic stroke (CS) accounts for approximately twenty-five percent of ischemic stroke cases with atrial fibrillation (AF) accounting for thirty percent of CS cases. We aimed to investigate the utility of atrial speckle tracking (STE) in identifying patients at high risk of AF after CS and potentially guiding patients who will benefit from long term rhythm monitoring devices. Methods: Cochrane Library, MEDLINE, and EMBASE were searched for relevant studies. We included studies that examined patients with new CS without a history of AF, and further examined LA strain parameters (peak and/or reservoir strain). Continuous data were pooled as a mean difference (MD) comparing patients who developed AF versus(vs) no AF. We used inverse variance method with Paule-Mandel estimator for tau2 and Hartung-Knapp adjustment, for random effect analysis. Hazard ratios were pooled using a similar statistical method. I2 was used to assess heterogeneity. Results: Eleven observational studies met our criteria, and included 2834 patients with new CS. Of those, 377 patients developed AF on follow up and 2457 patients did not develop AF. The AF group vs no AF had significantly reduced LA reservoir strain (LARS) [MD: -8.95; 95% CI: -11.33, -6.57) I2= 85%, p<0.001] at presentation. Further, a higher LARS was associated with a significantly lower incidence of AF after CS[HR: 0.93; 95%CI:0.89, 0.97)I2=0%]. Conclusion: Atrial reservoir strain is significantly lower in patients who developed AF after CS and a lower value is predictive of AF. More studies are needed to validate this data.