We sought to determine whether the extent of myocardial ischaemia on the admission electrocardiogram (ECG) has independent predictive value for short-term risk stratification of patients with non-ST-segment elevation acute coronary syndromes (NSTE ACS). Although the presence of ischaemic ECG changes on admission has been shown to predict outcome, the relationship between the extent of ECG changes and the risk of cardiac events is still ill defined. We analysed the admission ECGs of 5192 ACS patients enrolled in the GUSTO-IIb trial, without an ECG indication for thrombolysis. ECG tracings showing one or more of the following were eligible: ST-segment depression >0.5 mm, T-wave inversion >1 mm, and ST-segment elevation >0.5 mm but <1 mm. ECG variables associated with unfavourable 30 day outcomes in a univariable analysis were further assessed in a multivariable logistic regression model including independent clinical predictors. In the multivariable clinical, enzymatic, and ECG model, the sum of ST-segment depression (in millimetres) in all leads was a powerful independent predictor of 30 day death (P<0.0001), with a continuous increase in risk with the extent of ST-segment depression. The sum of ST-segment depression (P<0.0001) and the presence of minimal inferior ST-segment elevation (P<0.0001) or anterior ST-segment elevation (P=0.0182) were also independent predictors of the composite of death and myocardial infarction or reinfarction. The extent of ST-segment depression showed a highly significant correlation with the prevalence of three-vessel (P<0.0001) or left main coronary disease (P<0.0001), and also with the peak levels of creatine kinase (P<0.0001) during the index episode of ACS. In patients with NSTE ACS, the sum of ST-segment depression in all ECG leads is a powerful predictor of all-cause mortality at 30 days, independent of clinical variables and correlates with the extent and severity of coronary artery disease. The presence of even minimal (<1 mm) ST-segment elevation in anterior or inferior leads is independently associated with adverse outcomes.