Detection of atrial fibrillation (AF) following embolic stroke of undetermined source (ESUS) has therapeutic implications, as it may alter the secondary prevention strategy. However, implantable cardiac monitor (ICM) use is limited by cost and the need for improved patient selection.

A retrospective observational study published in the Cardiology Journal included 100 consecutive patients with ESUS and left atrial enlargement (LAE) admitted to a tertiary hospital. Baseline echocardiographic, electrocardiographic (ECG), and 24-hour Holter-ECG parameters were evaluated.

Multivariable logistic regression with cross-validation identified independent predictors of new-onset AF, which were incorporated into a risk stratification model. The primary outcome was detection of new-onset AF during follow-up.

Over an 18-month follow-up period, AF was detected in 19 patients (20%). Severe LAE was independently associated with AF occurrence (odds ratio [OR] 4.80; 95% confidence interval [CI] 1.32–17.37; p=0.017). Interatrial block (IAB) was also independently associated with AF (OR 6.22; 95% CI 1.27–30.47; p=0.024), as were atrial tachycardia (AT) episodes ≥20 beats on Holter-ECG (OR 7.62; 95% CI 1.21–47.74; p=0.03).

A composite risk score incorporating these variables demonstrated moderate discrimination (area under the curve [AUC] 0.733; p<0.001). Stratified risk estimates showed AF incidence of 9.67% in the low-risk group, 36.6% in the moderate-risk group, and 100% in the high-risk group.

The findings suggest severe LAE, IAB, and AT were associated with a higher likelihood of new-onset AF after ESUS. Risk stratification may help identify patients more suitable for ICM monitoring.