Background

Despite profound diagnostic assessment the causes of about 20% of strokes including embolic stroke of undetermined source (ESUS) remain elusive. Atrial fibrillation (AF) is an important possible cause; several AF risk factors are well characterized. However, the individual risk assessment after ESUS is insufficient. Here we aimed to establish a non-invasive approach to identify patients at high risk for AF as the presumed cause of ESUS. 

Methods

We performed a prospective, investigator-initiated, observational cohort study of patients with ESUS between 2018-2019 at our University Hospital. All patients received a detailed analysis of non-invasive AF risk factors, including age, sex, body-mass-index, hypertension, diabetes, CHA₂DS₂-VASc-Score, NT-proBNP and the newly developed rhythm irregularity burden derived from a 72h rhythm monitoring during hospitalization. We then performed uni- and multivariable adjusted analyses to assess the primary endpoint of incident AF during follow-up. Our results were independently validated using a second cohort from another University Hospital cohort.

Results

We included 297 patients with ESUS, of whom 46 (15.4%) developed AF during a median follow-up of 105 [25th; 75th percentile 31;338] days. Uni- and multivariable regression identified the rhythm irregularity burden as the most significant predictor of AF after ESUS (hazard ratio 3.12; 95% confidence interval 1.62-5.80; p<0001) even when compared to known AF risk factors including age, CHA₂DS₂-VASc-Score, and NT-proBNP. This result was independently validated in our second cohort (HR 2.20; 95% CI 1.45-3.33; p<0001).

Conclusion

We demonstrate that the rhythm irregularity burden obtained non-invasively from ECG-monitoring during the index hospitalization, may identify patients at higher risk for AF after ESUS. Further studies are needed to elucidate if high-risk patients benefit from intensified rhythm monitoring or modified treatment strategies for secondary stroke prevention.