B-type natriuretic peptide (BNP) tends to decrease in response to successful treatment of decompensated heart failure (HF). We identified characteristics and outcomes of patients whose BNP levels rise during admission despite appropriate decongestive therapy.

Patients enrolled in the ESCAPE trial admitted with acute systolic HF were divided into two groups according to whether or not BNP increased during the period from admission (T0) to discharge (T1). The study endpoint was 6‑month all-cause mortality.

Of 245 patients (age 56 years, 75 % men) admitted with acute systolic HF, 67 (27.3 %) had a higher BNP at T1 relative to T0. Despite similar degrees of congestion at T0, patients with BNP rise at T1 had less degree of decongestion from T0 to T1 as evident in the lower frequency of patients who had resolution of jugular venous distension (39.7 vs. 59.5 %, P = 0.01) and orthopnea (32.2 vs. 48.8 %, P = 0.029) at T1, in addition to lower reduction in IVC diameter during inspiration (P = 0.001) and expiration (P = 0.002) and less weight loss (P = 0.04). Patients with BNP rise at T1 were more likely to die (29.9 vs. 15.7 %, univariate OR 2.28, 95 % CI 1.177–4.414, P = 0.015) despite a lower BNP at T0 (492 vs. 1260 pg/ml, P < 0.001). Cox proportional hazard analysis revealed that a higher BNP at T1 independently predicts 6‑month mortality (hazard ratio 1.95, 95 % CI 1.067–3.578, P = 0.03) after adjustment for age, sodium, creatinine, and NYHA class—all at discharge. Kaplan–Meier analysis comparing survival in patients with or without BNP rise on discharge showed a significant intergroup difference (log-rank P value = 0.017).

Higher BNP levels on discharge identifies a subset of patients with lower degree of decongestion from T0 to T1 and higher 6‑month mortality.