Predictive Factors for New-Onset Atrial Fibrillation in Patients with Embolic Stroke of Undetermined Source and Left Atrial Enlargement — Development of a Risk Score

BACKGROUND:In patients with embolic stroke of undetermined source (ESUS), detecting new-onset atrial fibrillation (AF) is crucial because it leads to treatment change and reduces stroke recurrence and mortality. However, the high cost and relatively low AF detection rate of implantable cardiac monitors (ICM) necessitate the identification of predictive factors to better select candidates. This study aimed to identify AF predictors in ESUS patients with left atrial enlargement (LAE) to improve ICM selection. METHODS:We conducted a retrospective observational study with consecutive patients with ESUS and LAE admitted to a tertiary hospital. Echocardiographic, electrocardiographic, and 24-hour Holter electrocardiogram (ECG) findings were collected and analyzed using multivariate logistic regression and cross-validation techniques. Independent predictors were incorporated into a risk model classifying patients into low-, moderate-, and high-risk categories for new-onset AF. RESULTS:A total of 100 patients were included. After an 18-month follow-up period, new-onset AF was detected in 19 patients (20%). Independent predictors included severe LAE (odds ratio [OR] 4.80, 95% confidence interval [CI] 1.32-17.37; p = 0.017), interatrial block (OR 6.22, 95% CI 1.27-30.47; p = 0.024), and atrial tachycardia ≥ 20 beats on 24-hour Holter-ECG (OR 7.62, 95% CI 1.21-47.74; p = 0.03). A predictive risk score (area under the curve [AUC] 0.733; p < 0.001) was developed, classifying patients into low (9.67% risk), moderate (36.6%), and high (100%) risk categories. CONCLUSIONS:Interatrial block (IAB), atrial tachycardia (AT) ≥ 20 beats, and severe LAE are strong new-onset AF predictors in ESUS patients with LAE. A new scoring system effectively stratifies AF risk, optimizing ICM selection.