Seventy-nine patients with idiopathic dilated cardiomyopathy were assessed and followed up to evaluate 9 variables that might predict duration of survival after assessment for cardiac transplantation. Patients with ischemic heart disease, alcoholic and peripartum cardiomyopathy were excluded. There were 38 deaths (48%) during the 18-month (mean) follow-up. Patients underwent determination of left ventricular ejection fraction by radionuclide scan, echocardiography, cardiac catheterization and myocardial biopsy. Only left ventricular ejection fraction determined by radionuclide study correlated significantly with time to death in nonsurvivors (r = 0.38, p less than 0.05). Multivariant analysis and Cox multivariate regression analysis revealed that the single consistent determinant of prognosis was radionuclide-determined ejection fraction. It was an excellent predictor of survival to 3 months (p less than 0.0001) and a reasonable predictor of survival to 6 months (p less than 0.05). There was no variable that efficiently predicted survival for any period greater than 6 months. In 15 of 70 patients (21% of the entire group), clinical status and radionuclide ejection fraction improved after assessment but only one of these had an ejection fraction less than or equal to 0.10. No patient with a radionuclide ejection fraction greater than or equal to 0.20 died within 6 months of assessment. For those with ejection fraction between 0.11 and 0.19, survival after cardiac transplantation exceeded that of the natural history of their disease; this suggests that transplantation should be undertaken within 6 to 12 months of assessment. Left ventricular ejection fraction less than or equal to 0.10 predicts an extremely poor prognosis (6-month survival was 17%) and such patients should be transplanted with minimal delay.