Large vessel occlusion is independently associated with atrial fibrillation detection.

Abstract

Covert paroxysmal atrial fibrillation (pAF) is the most frequent cause of cardiac embolism. Our goal was to discover parameters associated with early pAF detection with intensive cardiac monitoring. Crypto-AF was a multicentre prospective study (four Comprehensive Stroke Centres) to detect pAF in non-lacunar cryptogenic stroke continuously monitored within the first 28days. Stroke severity, infarct pattern, large vessel occlusion (LVO) at baseline, electrocardiography analysis, supraventricular extrasystolia in the Holter examination, left atrial volume index and brain natriuretic peptide level were assessed. The percentage of pAF detection and pAF episodes lasting more than 5h were registered. Out of 296 patients, 264 patients completed the monitoring period with 23.1% (61/264) of pAF detection. Patients with pAF were older [odds ratio (OR) 1.04, 95% confidence interval (CI) 1.01-1.08], they had more haemorrhagic infarction (OR 4.03, 95% CI 1.44-11.22), they were more likely to have LVO (OR 4.29, 95% CI 2.31-7.97) (P<0.0001), they had a larger left atrial volume index (OR 1.03, 95% CI 1.01-1.1) (P=0.0002) and they had a higher level of brain natriuretic peptide (OR 1.01, 95% CI 1.0-1.1). Age and LVO were independently associated with pAF detection (OR 1.06, 95% CI 1.00-1.16, and OR 4.58, 95% CI 2.27- 21.38, respectively). Patients with LVO had higher cumulative incidence of pAF (log rank P<0.001) and more percentage of pAF>5h [29.6% (21/71) vs. 8.3% (12/144); OR 4.62, 95% CI 2.11-10.08; P<0.001]. In a mean follow-up of 26.82months (SD 10.15) the stroke recurrence rate was 4.6% (12/260). Large vessel occlusion in cryptogenic stroke emerged as an independent marker of pAF.