Biomarkers of Cardiac Dysfunction as Risk Factors in Cryptogenic Stroke.

Background: It is unclear whether biomarkers of cardiac dysfunction are associated with cryptogenic stroke (CS).

Methods: We retrospectively evaluated consecutive ischemic stroke patients. Patients underwent transthoracic echocardiography to evaluate left atrial diameter and the peak transmitral filling velocity/mean mitral annular velocity during early diastole (E/e'). Patent foramen ovale (PFO) and left atrial appendage flow velocity were evaluated by transesophageal echocardiography. We compared clinical characteristics and biomarkers of cardiac dysfunction (brain natriuretic peptide [BNP], left atrial diameter, E/e', and left atrial appendage flow velocity) between CS or CS without large PFO and other causative stroke subtypes.

Results: Among 1,514 patients with ischemic stroke, 264 patients were classified as having CS. Of these, transesophageal echocardiography revealed 27/158 (17%) large PFOs. In comparison, for the noncardioembolic stroke group, which consisted of large artery and small vessel subtypes, patients with CS without large PFO had higher log10 BNP (adjusted OR 2.70; 95% CI 1.92-3.78; p < 0.001), higher log10 E/e' (3.41; 1.21-13.15; p = 0.019), and lower left atrial appendage flow velocity (0.98; 0.97-1.00; p = 0.031). Left atrial diameter was similar for noncardioembolic stroke and CS without large PFO (p = 0.380). Cutoff values of BNP, E/e', and left atrial appendage flow velocity capable of distinguishing CS without large PFO from noncardioembolic stroke were 65.0 pg/mL (sensitivity 55.3%; specificity 70.9%), 13.0 (45.5%; 68.0%), and 46.0 cm/s (37.1%; 87.5%), respectively.

Conclusion: Patients with CS without large PFO could have biomarkers of cardiac dysfunction.