The purpose of this study was to determine the incidence of late potentials and their relation to QT prolongation in a family with a high incidence of sudden death during sleep at a young age and bradycardia-dependent QT prolongation (n = 9) and to compare the findings with those in consanguineous family members without QT prolongation (n = 13). Six (67%) of the 9 family members with QT prolongation had late potentials on the signal-averaged electrocardiogram (ECG) compared with 1 of the 13 normal subjects (p less than 0.007). Positive predictive accuracy of the signal-averaged ECG for the detection of subjects with QT prolongation was 86%; negative predictive accuracy was 80%. During exercise testing, the QT interval normalized, whereas late potentials did not change significantly. Exercise testing did not reveal the presence of coronary artery disease as a possible cause of late potentials. It is concluded that 1) compared with family members with a normal QT interval, patients with this type of bradycardia-dependent QT prolongation have a high incidence of late potentials; 2) late potentials persist despite normalization of the QT interval at high heart rates, indicating that there is no direct relation between late potentials and QT prolongation; and 3) late potentials are not caused by coronary artery disease in these subjects. Therefore, the detection of late potentials might be a new aid in the detection and risk stratification of patients with the long QT syndrome. Late potentials possibly indicate a substrate for ventricular tachyarrhythmias in this type of bradycardia-dependent QT prolongation.