Background:We sought to assess reproducibility, effects of various lead systems, and influence of physical activity on the results of signal‐averaged ECG obtained from the digital Holter system. Late potentials are routinely recorded using specially designed electrocardiographic devices. It has been shown recently that late potentials can also be recorded and analyzed from Holter tapes, which enables to examine possible dynamic changes of late potentials in relation to transient ischemia and spontaneous ventricular arrhythmias. However, there are some technical limitations of the signal‐averaged ECG analysis from Holter tapes that may be overcome by the use of digital, solid‐state Holter recorders where the ECG signal is acquired directly by the computer system and analyzed in real time.Methods:The signal‐averaged ECG was recorded at rest from XYZ leads and from lead system suggested by the manufacturer (Holter leads), and during moderate physical activity in 34 postin‐farction patients, using a new solid‐state Holter recorder. From each of these 20 minutes of recording, four consecutive ECG segments, lasting 5 minutes, were averaged, combined into vector magnitude, and analyzed using a Butterworth bidirectional filter between 40 and 250 Hz.Results:Reproducibility of signal‐averaged ECG was high; the P values (ANOVA) ranged from 0.93–0.99. Of the individual signal‐averaged ECG variables, reproducibility of the total QRS duration was superior to that of the terminal QRS measurements (the standard deviations of the mean relative errors were 6.3% vs 22.3% and 24.4%, respectively). There were no signifcant differences in signal‐averaged ECG variables when results obtained using orthogonal XYZ leads were compared with those obtained from Holter leads. Modest physical activity (slow walk) did not alter significantly signal‐averaged ECG variables. The values of the total QRS duration were more consistent than the terminal QRS measurements when two lead confgurations and recordings obtained during rest and physical activity were compared.Conclusions:Reproducibility of signal‐averaged ECG analysis obtained from a digital Holter system is high, the results are similar using XYZ or Holter leads, and the signal‐averaged ECG can be recorded and analyzed durin