Aims: Plasma levels of individual neurohormones (NH) have been proposed as reliable indicators for risk stratification of patients with heart failure (HF). Mainly because of small sample size, the predictive value of different NH has never been compared, while taking into account demographic, clinical and echocardiographic markers of risk in HF.
Methods and results: Plasma brain natriuretic peptide (BNP), norepinephrine (NE), renin activity (PRA), aldosterone (aldo) and endothelin were measured in 4300 patients before randomization in Val-HeFT. Univariate and multivariate Cox proportional hazard analyses were performed to investigate the relationship between NH and two primary study outcomes, mortality and combined mortality and morbidity (M/M). Higher baseline values for all NH were related to mortality and M/M, with univariate hazard ratios ranging from 1.13 [95% CI 0.99-1.30] (aldo) to 2.47 [2.13-2.87] (BNP) for mortality, and from 1.24 [1.11-1.39] (aldo) to 2.56 [2.28-2.89] (BNP) for M/M. In multivariate analyses, BNP had the strongest association with outcome, followed by NE and PRA. Patients with more activation of renin-angiotensin-aldosterone system tended to show greater benefit from valsartan; but the trend was not statistically significant.
Conclusion: All the NHs evaluated in 4300 patients with stable moderate to severe HF were found to be significant markers of outcome, despite therapy with ACEi, BB and randomization to an angiotensin receptor blocker or placebo. Several of these markers have been implicated as contributors to progression of HF, but BNP, which is thought to be protective, was the most powerful indicator for poor outcome.