Signal-averaged electrocardiograms were obtained in 62 consecutive patients with advanced congestive heart failure (CHF) undergoing evaluation for possible heart transplantation to determine if late potentials: (1) provide unique information compared to assessment of ventricular ectopic activity on ambulatory electrocardiogram, and (2) identify a subgroup of CHF patients with higher sudden death risk. Patients with a history of cardiac arrest or sustained ventricular tachycardia were excluded. CHF was due to old myocardial infarction in 40 patients and idiopathic dilated cardiomyopathy in 22 patients. Late potentials were present in 16 of 40 (40%) patients with old infarction but in only 3 of 22 (14%) patients with nonischemic CHF (p = 0.03). Twenty-four-hour ambulatory electrocardiograms were obtained in 34 patients (55%). Total ventricular ectopic activity and repetitive forms of ectopy were similar in patients with and without late potentials. Nine patients died suddenly, 9 had nonsudden death, 15 underwent heart transplantation and 29 were alive and well after a mean follow-up of 218 +/- 154 days. At 1 year, the actuarial risk of death was 37% and of sudden death was 20%. Sudden death risk was 12% in patients with late potentials versus 21% in those without (p = 0.73). Thus, the incidence of the arrhythmia substrate producing late potentials depends on the CHF etiology. The signal-averaged electrocardiogram and ambulatory electrocardiogram provide independent information for possible risk assessment in CHF. However, late potentials are poor predictors of sudden death risk when CHF is advanced, possibly due to the heterogeneity of causes of sudden death--ventricular tachycardia being only 1 of many possible mechanisms.