LAUSR

Introduction: Embolic stroke of undetermined source (ESUS) is suspected to be primarily of cardioembolic origin with ongoing trials evaluating cardiac biomarkers to confirm or refute this theory. Presently, it is unclear whether left ventricular ejection fraction (LVEF) is a contributor to ESUS and whether LVEF influences identification of atrial fibrillation (AF) on mobile telemetry. We hypothesized that lower LVEF is associated with the ESUS stroke subtype and higher likelihood of capturing AF. Methods: We retrospectively obtained data through the Comprehensive Stroke Program at Rhode Island Hospital. Those included were adults diagnosed with ESUS on discharge defined as an emboliform ischemic infarct, LVEF >30%, no ipsilateral stenoses > 50%, and cardiac telemetry for at least two weeks without contributory dysrhythmia. Data was compared with a non-cardioembolic (NCE) ischemic stroke population. A transthoracic echocardiogram was performed while hospitalized and interpreted by a cardiologist. A subset of those with ESUS completed cardiac monitoring whose data was retrospectively reviewed. LVEF was categorized as 30 - 54%, 55 - 65%, and > 65%. Standard descriptive and logistic regression analyses adjusted for age were used for statistical interpretation. Results: We identified 677 participants (ESUS: 324, NCE: 353); mean age was 66 years and 55% were male. In adjusted models, there was an association between ESUS subtype and LVEF 30 - 54% (vs. LVEF >65%) OR 2.51, 95%CI 1.27 - 4.95, p = 0.008 and LVEF 55 - 65% (vs. LVEF > 65%) OR 1.38, 95%CI 0.97 - 1.97, p = 0.077. With LVEF treated as a continuous variable, every 5% increase in LVEF lowered the likelihood of an ESUS subtype diagnosis (OR 0.85 95% CI 0.76-0.94 p=0.002). Those with a LVEF of 30 - 54% or LVEF > 65% had a lower rate of AF detection than patients with LVEF 55 - 65% (OR 0.37, 95%CI 0.16 - 0.85, p = 0.019) on prolonged monitoring. Conclusions: ESUS subtype was more common with lower LVEF. AF detection was lower amongst the lowest and highest LVEF ESUS subtype groups. This suggests that both LVEF and AF separately contribute to ESUS. Studies are needed to confirm our findings and improve stroke prevention strategies in patients with low LVEF and ESUS.