Although interferon (IFN) shows cardiotoxicity and arrhythmogenesis, the influence of IFN on signal-averaged electrocardiography remains to be clarified. The aim of this study was to test a clinical hypothesis that IFN therapy for hepatitis C virus may induce ventricular late potentials (LPs) and related arrhythmias in patients with chronic active hepatitis. Signal-averaged and ambulatory electrocardiograms were recorded sequentially in patients with chronic active hepatitis C (n = 22) throughout the entire period of IFN therapy. The filtered QRS duration (fQRS) and low amplitude (< 40 microV) signal duration (LAS40) were significantly increased (95.5 +/- 8.5 to 99.6 +/- 9.4 msec, P < 0.0001, and 32.8 +/- 3.1 to 36.3 +/- 3.0 msec, P < 0.0001, respectively), whereas the root mean square voltage in the terminal 40 msec of the fQRS (RMS40) was significantly decreased (25.5 +/- 5.4 to 22.3 +/- 5.2 microV, P < 0.005) 1 month after starting the IFN therapy. The ventricular LP was negative in all subjects before starting therapy, but became positive in 7 patients after the therapy commenced. There were no differences in clinical baseline characteristics between the LP-positive (n = 7) and LP-negative (n = 15) groups. Significant increases in mean heart rate, fQRS, and LAS40 were observed after starting the therapy, irrespective of the appearance of the ventricular LP, whereas a decrease in RMS40 was observed only in the LP-positive group. No sustained ventricular arrhythmias were documented in the ambulatory electrocardiography and no cardiac events were encountered in the follow-up period. Therefore, the results indicate a reversible and subclinical risk of IFN-induced arrhythmogenesis.