Prediction of heart failure by serum amino-terminal-pro-b-type natriuretic peptide (nt-probnp): an interim analysis of the screening evaluation of the evolution of new heart failure (screen-hf) study

Abstract

Serum NT-proBNP level predicts heart failure. The SCREEN-HF study is a community-based cohort study that aims to identify an appropriate threshold NT-proBNP level for stratification of individuals into high and low risk for heart failure. Aim: To assess whether serum NT-proBPN level can predict heart failure risk in an at-risk population. Methods: We recruited people with at least one risk factor for heart failure: age ≥60 years with one or more of self-reported myocardial infarction or other ischemic or valvular heart disease, arrhythmia, cerebrovascular disease, renal impairment, or treatment for hypertension or diabetes for ≥2 years. Exclusion criteria were known heart failure or left ventricular dysfunction on previous cardiac imaging. Blood was collected from all participants at baseline for measurement of electrolytes, creatinine and NT-proBNP. Median age of the 3938 participants (2171 men and 1767 women) was 70 years (interquartile range 65–75), 83% were receiving treatment for hypertension, 18% were diabetic, 23% had ischemic heart disease (IHD), 11% had cerebrovascular disease, 10% had atrial fibrillation (AF), 32% had body mass index (BMI) >30 kg/m2, 7% had obstructive sleep apnoea (OSA), and 23% had glomerular filtration rate <60 mL/min/1.73 m2.Results: At the time of this interim analysis there were 77 cases of incident heart failure (49 men and 28 women) during a median follow-up of 6 years (incidence rate 3.3 per 1000 person years). Relative to NT-proBNP tertile 1, the odds ratio for incident heart failure was 4.0 (95% confidence interval: 1.1–14.4) for tertile 2 and 21.6 (6.8–69.0) for tertile 3. The C-statistic from receiver operating characteristic analysis was 0.81 (0.77–0.86), with similar values for men and women. NT-proBNP >18 pmol/L (the highest 35%) predicted incident heart failure with 80.5% sensitivity, 66.2% specificity, positive predictive value 4.5% and negative predictive value 99.4%. Although age, diabetes, IHD, AF, BMI and OSA were significant predictors of incident heart failure in a multivariable logistic regression model including NT-proBNP, none improved classification of heart failure risk beyond NT-proBNP alone. Among 3046 participants who had echocardiography, NT-proBNP >18 pmol/L predicted left ventricular ejection fraction (LVEF) <45% with 74% sensitivity and LVEF <40% with 79% sensitivity. Conclusions: Serum NT-proBNP level assists stratification of heart failure risk among a community population with risk factors for heart failure. Improved identification of individuals at increased risk of heart failure will enable targeting of preventative therapies.