Duration of paroxysmal atrial fibrillation in cryptogenic stroke is not associated with stroke severity and early outcomes

Abstract

The current definition of paroxysmal atrial fibrillation (PAF) requires an arbitrary cut-off of >30s, but in clinical practice cryptogenic stroke (CS) patients with PAF duration of ≤30s are not usually excluded from anticoagulation therapy. We sought to evaluate the clinical relevance of short-duration (≤30s) PAF in CS. Consecutive CS patients with no prior AF history and sinus-rhythm on baseline electrocardiography (ECG) were prospectively evaluated over a three-year period. Baseline stroke severity was assessed by NIHSS-scores. All patients underwent 24-hour Holter-ECG during hospitalization. ECG recordings were analyzed by two blinded investigators using dedicated analysis software. Total time in AF was calculated as the sum of each individual AF episode for patients with multiple episodes during monitoring. Patients were dichotomized in two groups using PAF total duration (≤30s & >30s). Early recurrent stroke and favorable functional outcome (FFO, defined as mRS-grades of 0-1) were evaluated during a three-month follow-up period. A total of 184 patients (66% men, mean age 57±11years) with CS (median NIHSS-score 4, IQR: 2-7) were evaluated. PAF of any duration was detected in 23 individuals (13%; 95%CI: 8%-18%). Among these patients the prevalence of brief PAF was 57% (n=13). The two groups did not differ (p>0.2) in terms of demographics, vascular risk factors and NIHSS-scores. Early recurrent stroke and FFO rates were similar (p>0.4) in the two groups. Duration of PAF is not associated with baseline stroke severity and early outcomes in patients with CS and should not influence anticoagulation decision in these patients.

Keywords: Cryptogenic stroke; Duration; Outcomes; Paroxysmal atrial fibrillation; Severity.