The circadian changes in the activity of the autonomic nervous system in a group showing long QT duration in electrocardiogram (ECG) were studied in order to differentiate symptomatic congenital long QT syndrome from asymptomatic. The asymptomatic group presented only long QT duration (QTc > 0.46). Seven girls and two boys, including two subjects experiencing syncope of non-neurological origins, were examined by using heart rate (HR) power spectrum analysis. In three subjects, the peak of the high frequency band, indicating the parasympathetic activity, disappeared during night-time (sleep), which means the possibility of a high risk of sudden cardiac death. In two of three subjects, moreover, the averaged sympathetic activity during daytime was significantly increased compared to that during night-time, in addition to the abrupt increase of sympathetic activity in the morning. These two subjects, in which one boy had a family history of Romano-Ward syndrome, were the same people experiencing the syncope attack. We succeeded in objectively elucidating that congenital long QT syndrome is adrenergic-dependent, and suggest that HR power spectrum analysis may be of value in distinguishing symptomatic congenital long QT syndrome from asymptomatic showing only long QT duration.