We sought to compare the prognostic utility of growth hormone (GH) with NT-proBNP) and the ADHERE score in a large cohort of acute heart failure (HF) patients, subcategorized into HF with reduced EF <50% (HFrEF) and preserved EF ≥50% (HFpEF). GH and NTproBNP levels were measured in 537 patients (HFrEF n = 415; HFpEF n = 122) with acute HF recruited into this prospective cohort study. The main outcome measure was death or HF readmission at 1 year. GH levels were higher in both HFrEF [1.26 (0.54-2.62) vs. 0.8 (0.26-1.94) ng/mL, P < 0.001] and HFpEF [1.04 (0.48-2.92) vs. 0.53 (0.18-1.94) ng/mL, P = 0.020] patients with the outcome compared with event-free survivors. GH levels were independently predictive for the outcome at 1 year in the entire cohort [HR 1.47, 95% confidence interval (CI) 1.16-1.86, P = 0.001] and those with HFrEF (HR 1.54, 95% CI 1.19-1.99, P = 0.001) in multivariate Cox hazard analysis. GH improved risk classification as measured by continuous net reclassification improvement (NRI) when added to the ADHERE multivariate logistic model of age, sex, urea, heart rate, and systolic blood pressure, for all patients [NRI 29.6 (12.1-47.1), P = 0.001] and HFrEF NRI 21.7 (1.9-41.6), P = 0.034] patients, as well as in addition to the ADHERE model combined with NT-proBNP for all patients [NRI 25.4 (7.8-43.1), P = 0.005]. GH offers incremental prognostic information over the ADHERE score clinical predictors and NT-proBNP for risk stratification of acute HF patients.