Background: Occult paroxysmal atrial fibrillation (AF) is detected in up to 30% of patients with Embolic Stroke of Unknown Source (ESUS). We aim to identify predictors of AF on outpatient cardiac monitoring in patients with ESUS.
Methods: We included ischemic stroke patients with ESUS subtype between January 1st, 2013 and December 31 1st, 2016 who underwent an inpatient diagnostic evaluation, including 24 hour telemetry, followed by outpatient cardiac monitoring with 30-day cardiac monitors and/or implantable cardiac monitor (ICM). Patients were divided into two groups based on detection of AF or atrial flutter during monitoring. We compared demographic data, clinical risk factors, NIHSS, and cardiac biomarkers (left atrial enlargement on echocardiography, PR interval on ECG) between the two groups. Multivariable models were used to identify independent predictors of AF and to develop an AF prediction-risk scoring system.
Results: We identified 296 consecutive patients during the study period; 38 (12.8%) demonstrated previously undetected AF on outpatient cardiac monitoring. Compared to non-AF patients, AF patients were older (72.0 ± 11.0 vs. 62.7 ± 14.9; p <0.001), less likely to be current smokers (10.5% vs. 27.2%, p = 0.027), had lower NIHSS (3, IQR 1-7 vs. 3, IQR 1-5; p=0.002), and larger left atrial diameter in mm (39.3 ± 6.0 vs. 35.2 ± 6.4; p=0.001). In multivariable regression analysis, advanced age (ages 65-74: odds ratio 2.36, 95% confidence interval 0.85-6.52; ages ≥75: odds ratio 4.08, 95% confidence interval 1.58-10.52) and moderate-to-severe LAE (odds ratio 4.66, 95% confidence interval 1.79-12.12) were predictive of AF detection on outpatient monitoring. NIHSS, current smoking status, and mild LAE did not contribute significantly to AF prediction. We developed the Brown ESUS-AF score: age (65-74 years: 1 point, ≥ 75 years: 2 points) and left atrial enlargement (moderate/severe: 2 points) with good prediction of AF (AUC 0.725). The percentage of patients in each score category were: 0: 4.2%, 1: 14.8%, 2: 20.8%, 3: 22.2%, 4: 55.6%.
Conclusion: In this large sample of patients with ESUS, the Brown ESUS-AF score predicts AF detection on prolonged outpatient monitoring. More studies are needed to externally validate our findings.