790-2 Baseline Electrocardiogram Predicts 30-day Mortality Among 32,812 Patients with Acute Myocardial Infarction Treated with Thrombolysis

To determine the initial electrocardiographic variables predictive of survival among patients with acute myocardial infarction, we analyzed the baseline 12-lead ECGs in 32,812 patients enrolled into the GUSTO trial. All patients had≥0.1mV of ST segment elevation in at least one lead and received thrombolytic therapy. Those with LBBB or ventricular rhythm were excluded from analysis. Clinical follow-up was > 99.5% complete. 2218 (6.8%) patients died within 30 days of the initial ECG. Death within 30 days was more common in patients with RBBB (17%), LAFB (14%), and LPFB (17%), than in those with a normal conduction pattern (6%). Patients with ECG evidence of previous MI in a location distinct from the acute MI had a higher risk of death (9.8% vs. 5.9%) than those without prior infarction (p<0.0001). The variable having the greatest univariate predictive power for 30-day survival was the sum of the absolute ST-segment deviation in each lead (x2=341), as shown in the following mortality curve.Other ST segment variables that predicted 30-day survival were the sum of ST-segment elevation in each lead (x2=287). the maximum ST elevation in anyone lead (X2=257), and the number of leads with ST elevation (x2=250). When multivariate modeling was performed the sum of the absolute ST deviations, number of leads with ST elevation, prior ECG MI, RBBB, and LAFB each added independent prognostic information.We conclude that an ECG at the time of presentation contains substantial prognostic information which can be used to help stratify risk among thrombelytic-treated patients with acute myocardial infarction