Abstract 3384: Amino-terminal pro-Brain Natriuretic Peptide Testing is Superior to New York Heart Association Classification for Prediction of One-Year Mortality in Patients with Acute Heart Failure: Results from the International Collaborative of NT-proBNP Study
Background: Estimating risk in patients with acute heart failure (HF) may be challenging. A time-tested and validated method for estimating risk has been use of the New York Heart Association (NYHA) functional classification, a subjective measure of disease severity, based on symptoms. Amino-terminal pro-brain (NT-proBNP) testing may be useful for predicting risk in acute HF, and is a more objective measure than symptom assessment. The importance of NYHA classification for prognosis estimation relative to NT-proBNP results remains unclear, however.
Methods: 720 subjects with acute heart failure enrolled in the 4 centers from the International Collaborative of NT-proBNP (ICON) study were analyzed. Bootstrapping statistical methods were used to evaluate for candidate predictors of death by one year and entered into a Cox Proportional Hazards model to evaluate for their independent value. Analyses were performed for the group as a whole (n=720); as well, for those with ejection fraction data, stratified analyses were performed in those with systolic (n=387) and non-systolic HF (defined as an ejection fraction <50%; n=267). Hazard ratios (HR) with 95% confidence intervals (CI) were estimated.
Results: Bootstrapping confirmed an NT-proBNP >5180 pg/ml to be the strongest predictor of one-year death in all 720 subjects, as well as in those with systolic and/or non-systolic HF. In multivariable Cox analyses, an NT-proBNP >5180 pg/ml was the strongest predictor of death among all 720 subjects with acute HF (HR=2.14; 95% CI=1.64–2.81; p<0.00005), and remained so for those with systolic (HR=2.44; 95% CI=1.49–3.97; p<0.00005) and non-systolic HF (HR=2.19; 95% CI=1.32–3.64; p=0.002). In contrast, though a significant predictor of death in univariable analyses, NYHA symptom classification was no longer an independent predictor of death once NT-proBNP >5180 pg/ml was entered into each model: for all 720 subjects (HR=0.87; 95% CI=0.62–1.22; p=0.41), those with systolic HF (HR=01.34; 95% CI=0.97–1.87; p=0.08), or those with non-systolic HF (HR=0.87; 95% CI=0.62–1.21; p=0.41).
Conclusion: NT-proBNP is superior to NYHA functional classification for estimating risk for death by one year in patients with acute systolic and non-systolic HF.