To determine the reliability of the admission electrocardiogram in predicting outcome in patients hospitalized for chest pain at rest, 90 patients were randomized into a trial of aspirin versus heparin in unstable angina or non-Q-wave myocardial infarction, and prospectively followed for 3 months. The emergency room admission electrocardiogram was analyzed for ST-segment deviation greater than or equal to 1 mm/lead and T-wave changes. Unfavorable outcomes were recurrent ischemic pain, myocardial infarction and coronary revascularization with angioplasty or surgery. In patients who underwent coronary arteriography, a myocardium in jeopardy score ranging from 0 to 10 was assigned, based on the number of vessels with a diameter stenosis greater than or equal to 70% and the location of the stenoses. Considering all 90 patients, an admission electrocardiogram with ST-segment deviation in greater than or equal to 2 leads had a positive predictive value for adverse clinical events of 79% and a negative predictive value of 64%. In the subset of patients without left ventricular hypertrophy and whose admission electrocardiograms were recorded during chest pain (62 of 90), the positive predictive value of ST deviation in greater than or equal to 2 leads improved to 89% and the negative value to 72%. Of the 62 patients, 53 underwent coronary arteriography. There was a positive linear correlation between the total number of leads with ST-segment deviation and the myocardium in jeopardy score (r = 0.80, p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)