Association between NT-proBNP and the risk of incident heart failure, overall and stratified by status of obesity.

<p>^a Person years are calculated from the sub-cohort (n = 1,163) only.</p><p>^b adjusted for sex, stratified for baseline age.</p><p>^c Model 1 further adjusted for educational degree, physical activity, smoking status, alcohol consumption, waist circumference, and prevalent diseases (diabetes, coronary heart disease, hypertension).</p><p>^d Model 2 further adjusted for biomarkers (hsCRP, creatinine, total cholesterol and HDL cholesterol).</p><p>*p for trend across tertiles was calculated by using tertiles of N-terminal pro brain natriuretic peptide (NT-proBNP) as a categorical variable in the respective Cox proportional hazards regression models.</p><p>^$the p value for nonlinearity was calculated by Wald chisquare test following restricted cubic spline Cox regression analysis.</p><p>^#in the non-stratified models, body mass index (BMI) was also included in the adjustment set.</p><p>Association between NT-proBNP and the risk of incident heart failure, overall and stratified by status of obesity.</p>