Blood Biomarkers of Heart Failure and Hypercoagulation to Identify Atrial Fibrillation–Related Stroke

Background and Purpose— Occult atrial fibrillation (AF) causes a relevant proportion of initially cryptogenic stroke (CS), but prolonged rhythm monitoring is difficult to apply to all such patients. We hypothesized that blood biomarkers indicating heart failure (NT-proBNP [N-terminal pro-brain natriuretic peptide]) and hypercoagulability (D-dimer, AT-III [antithrombin-III]) were associated with AF-related stroke and could serve to predict the likelihood of AF detection in CS patients early on. Methods— Over a 1-year period, we prospectively applied a defined etiologic work-up to all ischemic stroke patients admitted to our stroke unit. If no clear stroke cause was detected (CS), patients underwent extended in-hospital cardiac rhythm monitoring (≥72 hours). Blood to determine biomarker levels was drawn within 24 hours after admission. Results— Of 429 patients, 103 had AF-related stroke. Compared with noncardiac stroke patients (n=171), they had higher NT-proBNP (1867 versus 263 pg/ml) and D-dimer levels (1.1 versus 0.6 µg/ml), and lower AT-III concentration (89% versus 94%). NT-proBNP ≥505 pg/ml distinguished AF-related from noncardiac stroke with a sensitivity of 93% and a specificity of 72%. D-dimer and AT-III cutoffs had lower sensitivities (61% and 53%) and specificities (58% and 69%) for AF-related stroke. Of all initially 143 CS patients, 14 were diagnosed with AF during in-hospital monitoring. The preidentified NT-proBNP cutoff ≥505 pg/ml correctly predicted AF in 12 of them (86%, negative predictive value: 98%), while D-dimer and AT-III cutoffs were noncontributory. Conclusions— This study supports the association of NT-proBNP and to a lesser extent of hypercoagulation markers with AF-related stroke. NT-proBNP seems helpful in selecting CS patients for immediate extended cardiac rhythm monitoring to detect occult AF whereby levels <505 pg/ml seem to have a high-negative predictive value.