Abstract 12621: B-Type-Natriuretic Peptide and C-Reactive Protein in the Prediction of Atrial Fibrillation Risk: The CHARGE-AF Consortium

Abstract

Introduction The role of biomarkers in atrial fibrillation (AF) predictive models is mostly unknown. B-type natriuretic peptide (BNP), marking atrial distension and volume overload, and the inflammatory marker C-reactive protein (CRP) are associated with AF. We assessed improvement of AF prediction by inclusion of BNP and CRP levels. Methods We followed 18556 whites and African-Americans free of AF at baseline, pooled from the Atherosclerosis Risk in Communities Study (ARIC), Cardiovascular Health Study (CHS), and Framingham Heart Study (FHS), for 5 years (prediction horizon). Cox models predicted AF incidence (covariates: age; race; smoking; height; weight; systolic and diastolic blood pressure; diabetes; hypertension medications; history of heart failure or myocardial infarction). We added BNP (ARIC / CHS: NT-proBNP; FHS: BNP), CRP, or both and assessed model calibration (Hosmer-Lemeshow statistic), and biomarker predictive ability (C-statistic, integrated discrimination improvement [IDI], net reclassification improvement [NRI]). Results BNP and CRP were significantly associated with AF incidence [n=1186]: BNP: hazard ratio per SD ln-transformed biomarker 1.66 (1.56-1.76), p<0.0001; CRP: Hazard ratio 1.18 (1.11-1.25), p<0.0001 (Figure). Model calibration was sufficient (Hosmer-Lemeshow BNP, p=0.05; CRP, p=0.31; BNP + CRP, p=0.16). BNP improved the C-statistic from 0.740 to 0.765. BNP yielded an IDI of 0.027 (0.022-0.032), a relative IDI of 41.5%, and a continuous NRI of 0.389 (0.322-0.455); corresponding numbers for CRP were 0.003 (0.002-0.005), 5.0%, and 0.154 (0.081-0.228). Addition of BNP + CRP was similar to addition of BNP alone. Overall, BNP (per SD) was the strongest predictor of AF apart from age (per 5 years). The predictive ability of CRP was weak. Conclusion BNP significantly improved AF risk prediction beyond clinical factors in a heterogeneous population, and could serve as a benchmark to evaluate novel putative AF risk biomarkers.