Background and Purpose- NT-proBNP (N-terminal pro-B-type natriuretic peptide) is a risk factor for atrial fibrillation and a marker of cardiac function used in the detection of heart failure. Given the link between cardiac dysfunction and stroke, NT-proBNP is a candidate marker of stroke risk. Our aim was to evaluate the association of NT-proBNP with stroke and to determine the predictive value beyond a panel of established risk factors. Methods- Based on the Biomarkers for Cardiovascular Risk Assessment in Europe-Consortium, we analyzed data of 58 173 participants (50% men; mean age 52 y) free of stroke from 6 community-based cohorts. NT-proBNP measurements were performed in the central Biomarkers for Cardiovascular Risk Assessment in Europe laboratory. The outcomes considered were total stroke and subtypes of stroke (ischemic/hemorrhagic). Results- During a median follow-up time of 7.9 years, we observed 1550 stroke events (1176 ischemic). Increasing quarters of the NT-proBNP distribution were associated with increasing risk of stroke ( P for trend <0.0001; multivariable Cox regression analysis adjusted for risk factors and cardiac diseases). Individuals in the highest NT-proBNP quarter (NT-proBNP >82.2 pg/mL) had 2-fold (95% CI, 75%-151%) greater risk of stroke than individuals in the lowest quarter (NT-proBNP <20.4 pg/mL). The association remained unchanged when adjusted for interim coronary events during follow-up, and though it was somewhat heterogeneous across cohorts, it was highly homogenous according to cardiovascular risk profile or subtypes of stroke. The addition of NT-proBNP to a reference model increased the C-index discrimination measure by 0.006 ( P=0.0005), yielded a categorical net reclassification improvement of 2.0% in events and 1.4% in nonevents and an integrated discrimination improvement of 0.007. Conclusions- In European individuals free of stroke, levels of NT-proBNP are positively associated with risk of ischemic and hemorrhagic stroke, independently from several other risk factors and conditions. The addition of NT-proBNP to variables of established risk scores improves prediction of stroke, with a medium effect size.
