RESUMO
BACKGROUND: Atrial dysfunction or "cardiopathy" has been recently proposed as a mechanism in cryptogenic stroke. A prolonged PR interval may reflect impaired atrial conduction and thus may be a biomarker of atrial cardiopathy. We aim to compare the prevalence of PR interval prolongation in patients with cryptogenic stroke (CS) when compared with known non-cryptogenic non-cardioembolic stroke (NCNCS) subtypes. METHODS: We used prospective ischemic stroke databases of 3 comprehensive stroke centers to identify patients 18 years or older with a discharge diagnosis of ischemic non-cardioembolic stroke between December 1, 2013 and August 31, 2015. The main outcome was ischemic stroke subtype (CS versus NCNCS). We compared PR intervals as a continuous and categorical variable (<200 milliseconds; ≥200 milliseconds) and other clinical and demographic factors between the 2 groups and used multivariate regression analyses to determine the association between PR interval prolongation and CS. RESULTS: We identified 644 patients with ischemic non-cardioembolic stroke (224 CS and 420 NCNCS). Patients with CS were more likely to have a PR of 200 milliseconds or greater when compared with those with NCNCS (23.2% versus 13.8%, P = .009). After adjusting for factors that were significant in univariate analyses, a PR of 200 milliseconds or greater was independently associated with CS (odds ratio [OR] 1.70, 95% CI 1.08-2.70). The association was more pronounced when excluding patients on atrioventricular nodal blocking agents (OR 2.64, 95% CI 1.44-4.83). CONCLUSIONS: A PR of 200 milliseconds or greater is associated with CS and may be a biomarker of atrial cardiopathy in the absence of atrial fibrillation. Prospective studies are needed to confirm this association.