The prognostic significance of electrocardiographic variables was retrospectively investigated in 88 patients with dilated cardiomyopathy and with normal coronary arteriograms. During an average follow-up of 3.7 +/- 2.9 years, 43 patients died, 26 of progressive heart failure, 15 patients with sudden death, and one due to cerebral embolism. Excluding one patient, who died of esophageal cancer, the cumulative survival rate was 73% at 2 years and 60% at 5 years. By univariate life table analysis, abnormal Q-waves, a QRS duration > or = 0.12s, a cardiothoracic ratio > or = 60%, systolic blood pressure < 110mmHg, and left ventricular end-diastolic pressure > or = 15mmHg were significantly associated with increased mortality within 5 years. Multivariate analysis using Cox's proportional hazards model determined the major independent risk factors in the following order: (1) for all patients; the presence of abnormal Q-waves, left bundle branch block or intraventricular conduction disturbances, left ventricular end-diastolic pressure, systolic blood pressure and the cardiothoracic ratio; (2) for patients without left bundle branch block or intraventricular conduction disturbances; abnormal Q-waves, left ventricular end-diastolic pressure and systolic blood pressure. The present study thus demonstrated that the electrocardiogram could provide independent prognostic predictors in patients with dilated cardiomyopathy, possibly reflecting the severity of myocardial damage.