Initial risk stratification in patients with acute heart failure (AHF) is poorly validated. Previous studies tended to evaluate the prognostic significance of only one or two selected ECG parameters. The aim of this study was to evaluate the impact of multiple ECG parameters on mortality in AHF. The Acute Heart Failure Database (AHEAD) registry collected data from 4,153 patients admitted for AHF to seven hospitals with Catheter Laboratory facilities. Clinical variables, heart rate, duration of QRS, QT and QTC intervals, type of rhythm and ST-T segment changes on admission were collected in a web-based database. 12.7 % patients died during hospitalisation, the remainder were discharged and followed for a median of 16.2 months. The most important parameters were a prolonged QRS and a junctional rhythm, which independently predict both in-hospital mortality [QRS > 100 ms, odds ratio (OR) 1.329, 95 % CI 1.052-1.680; junctional rhythm, OR 3.715, 95 % CI 1.748-7.896] and long-term mortality (QRS > 120 ms, OR 1.428, 95 % CI 1.160-1.757; junctional rhythm, OR 2.629, 95 % CI 1.538-4.496). Increased hospitalisation mortality is predicted by ST segment elevation (OR 1.771, 95 % CI 1.383-2.269) and prolonged QTC interval >475 ms (OR 1.483, 95 % CI 1.016-2.164). Presence of atrial fibrillation and bundle branch block is associated with increased unadjusted long-term mortality, but mostly reflects more advanced heart disease, and their predictive significance is attenuated in the multivariate analysis. ECG in patients admitted for acute heart failure carries significant short- and long-term prognostic information, and should be carefully evaluated.