Abstract 3891: B-type Natriuretic Peptide, C-Reactive Protein, Urinary Albumin, and Glomerular Filtration Rate: Which is the Best Predictor for Onset of Heart Failure in the General Population?
Background: B-type natriuretic peptide (BNP), high sensitivity C-reactive protein (CRP), urinary albumin (UA) and glomerular filtration rate (GFR) have been suggested to be useful predictors for the development of cardiovascular events (CVE) including congestive heart failure (HF). However, it remains unknown which of these biomarker is the most useful predictor for HF in the general population.
Methods: Subjects of this community-based study were recruited from the general population. Blood levels of BNP, CRP, UA corrected by urinary creatinine (UA/Cr), and estimated GFR were determined in 14,151 participants (mean age=62.8 yrs, women=66%). HF admissions within in the cohort were prospectively recorded by the regional registration survey. For each biomarker, a multivariate Cox regression analysis was performed to determine the relative risk for new onset of HF. The predictive ability for HF was also assessed by the area under the curve (AUC) generated from receiver operating characteristic (ROC) analysis.
Results: The cohort was followed for 40,937 person-years. After adjustment for traditional cardiovascular risk factors (age, sex, hypertension, diabetes, smoking, hypercholesterolemia and atrial fibrillation), hazard ratios for HF development in terms of 1SD increments (as a continuous variable) and values above the 80th percentile (as a categorical variable) for each biomarker level were determined (Table⇓). In addition, the AUC of ROC analysis for prediction of HF was 0.842 (95% CI; 0.770 – 0.913) for plasma BNP but suboptimal for the other biomarkers (AUC <0.75).
Conclusion: Increased plasma BNP levels are the best predictor for the onset of HF in the general population.