Ferreira JP, Ouwerkerk W, Tromp J, et al. - In this analysis with heart failure (HF) patients (n = 2,309) from the BIOSTAT-CHF (a systems BIOlogy Study to TAilored Treatment in Chronic Heart Failure) study, researchers characterized patients by cardiovascular (CV) vs non-CV death, constructed predictive models for the respective events, and tested the performance of models to distinguish CV from non-CV death. The best combination of variables related to each cause-specific death was evaluated using competing-risk models. For the prediction of CV death, the model that appeared the best incorporated low blood pressure, estimated glomerular filtration rate ≤ 60 mL/min, peripheral oedema, prior HF hospitalization, ischaemic HF, chronic obstructive pulmonary disease, raised N-terminal pro-B-type natriuretic peptide (NT-proBNP), and troponin. Age > 75 years, anaemia and elevated NT-proBNP were included in the non-CV death model. The recognition of cases with a higher absolute risk of dying from CV causes (vs non-CV ones) was enabled by the risk models developed in this study. Only CV death prediction was aided by troponin, whereas NT-proBNP aided the prediction of both CV and non-CV death.