Pre-discharge, but not admission, levels of NT-proBNP predict adverse prognosis following acute LVF.

Abstract

Circulating natriuretic peptide levels provide prognostic information following acute coronary syndromes and in chronic heart failure. Little evidence exists of their utility following hospitalisation with acute left ventricular failure (LVF). To examine the relative prognostic value of admission and pre-discharge plasma N-terminal pro B-type natriuretic peptide (NT-proBNP) following hospitalisation with acute heart failure. NT-proBNP was measured at admission in 96 patients hospitalised with acute LVF. In a subset of 34 patients, NT-proBNP was also measured prior to discharge. Multivariate analysis was performed of the clinical and serological predictors of a combined primary endpoint of death or heart failure (hospitalisation or as an outpatient). During follow up (median 350 days, range 2-762), 37 (38.5%) patients died (n=16, 16.7%), or experienced at least 1 heart failure event (n=21, 21.9%). For the entire cohort of 96 patients, only a prior history of heart failure was associated with the primary endpoint (OR 3.5 [1.10-11.08], P=0.034). Admission plasma NT-proBNP was not predictive (OR 1.84 [0.75-4.51], P=0.185). In the 34 patients for whom both admission and pre-discharge NT-proBNP was available, 19 (55.9%) died (n=8, 23.5%) or experienced heart failure (n=11, 32.4%). Only pre-discharge plasma NT-proBNP (OR 15.30 [95% CI: 1.4-168.9], P=0.026) was independently predictive of the composite endpoint. The area under the receiver-operator-characteristic (AUC ROC) curve for pre-discharge NT-proBNP was superior to that for admission NT-proBNP for prediction of death or heart failure (AUC ROC 0.87 cf 0.70), for death (0.79 cf 0.66), LVF hospitalisation (0.78 cf 0.70) or heart failure as an outpatient (0.71 cf 0.61). Plasma NT-proBNP measured pre-discharge provides useful prognostic information following hospitalisation with acute LVF.