Introduction: Patients hospitalized for heart failure have a very high in-hospital as well as one-year mortality. Natriuretic peptides play both a diagnostic and a prognostic role in this disease. Changes of natriuretic peptide levels in response to therapy are a well-known prognostic marker. Regarding in-hospital mortality, however, little is known about the prognostic value of extremely high levels of natriuretic peptides measured on admission. Aim: To decide whether extremely high levels of B-type natriuretic peptide have a prognostic value with regard to in-hospital mortality. Method: NT-proBNP levels on admission and in-hospital mortality were extracted retrospectively from the data of patients treated with heart failure in the cardiology department of the Hospital of St. John of God in Budapest. We separately analyzed the data of patients hospitalized for heart failure in 2015 with extremely high initial NT-proBNP levels. The cut-off value in this regard was 10 000 ng/l. We also analyzed the comorbidities of these patients. Results: The median NT-proBNP level of those patients who survived beyond the index hospital stay in the last 10 years was 4842 ng/l, whereas the median NT-proBNP level of those 182 patients who died during their hospital stay was 10 688 ng/l (p10 000 ng/l on admission have a significantly higher in-hospital mortality. The difference of NT-proBNP levels of surviving versus deceased patients in the group with admission NT-proBNP levels >10 000 ng/l is no longer significant. We could not identify any etiological factors that would explain these extremely high NT-proBNP levels or the excess in-hospital mortality. Orv Hetil. 2017; 158(20): 779–782.