Circulating B-type natriuretic peptides (BNP and Nterminal part of the pro-B-type natriuretic peptide [NT-proBNP]) increase in proportion to cardiac dysfunction, and plasma concentrations are powerful independent predictors of outcome across the spectrum of cardiovascular disease. Most reports underpinning the prognostic utility of plasma B-type peptides rely on analyses derived from single measurements per study participant. Such isolated measurements improve risk stratification at any time during progression of chronic heart failure (CHF). Less is known about additional information to be gained from serial peptide measurements. Changes in either BNP or NTproBNP over brief periods are related to subsequent outcomes in both acute coronary syndromes and acute decompensated heart failure (HF) (1,2). Changes in NT-proBNP over months predict outcome in advanced HF (3). The relationship between changes in peptide levels over longer intervals with outcomes in a large group with stable CHF has been explored for BNP (4) but not NT-proBNP.
See page 997In this issue of the Journal, Masson et al. (5) examined the prognostic value of changes in NT-proBNP between baseline and 4 months in 1,742 patients randomized to the placebo arm of the Val-HeFT (Valsartan Heart Failure Trial) study. Changes in NT-proBNP expressed as absolute, percent, or categorical change were related to subsequent mortality. Absolute change in plasma NT-proBNP in quartiles (Q) showed a U-shaped relationship to subsequent mortality (15.1%, 8.3%, 11.5%, and 26.4% in Q1 to Q4, respectively), whereas mortality increased incrementally across quartiles of percent change in NT-proBNP (from 9.2% in Q1 to 21.5% in Q4). In a third approach, outcomes were assessed according to categorical changes of NT-proBNP. In patients with results below the nominated threshold (1,078 pg/ml) at baseline and 4 months (52% of patients), with values above decreasing to below (7.2%), below increasing to above (6.5%), and with both values above the threshold (34%), mortality rates were 8.9%, 7.2%, 21.1%, and 25.7%, respectively. Multivariable analysis including adjustment for age, body mass index, renal function, etiology, New York Heart Association functional class, left ventricular function, and baseline NT-proBNP indicated categories were independently associated with mortality.This report is the most recent in a series arising from the Val-HeFT neurohormonal substudy. The Val-HeFT study showed BNP to be prognostically superior to several other recognised neurohormonal markers of risk in HF, including norepinephrine, renin activity, aldosterone, and endothelin (6). In a head-to-head comparison of BNP and NTproBNP including 3,916 Val-HeFT study participants (7), baseline BNP and NT-proBNP were powerfully and similarly related to both mortality and risk of admission with decompensated HF independent of, and more strongly than, any of a range of pertinent predictive demographic, clinical, and echocardiographic variables.From results in 4,305 patients, Anand et al. (8) reported an ass...