Reliability of spectral analysis of the heart rate variability signal from a commercial Holter system

The statistical characterization of the measurement error (ME) of some commonly used spectral parameters, taking into account the overall chain of recording, playing, and analyzing Holter tapes, is addressed. Two electrocardiograms (ECGs) with known modulation of the R-R interval were generated into a PC and fed to three Holter recorders successively equipped with four different tapes. One ECG signal had normal variability (var = 2376 ms²) with the other reduced variability (var = 86 ms²). Twenty-four 10 min recordings were analyzed on each tape. For each investigated parameter the ME was computed subtracting its true value obtained before Holter recording from its value estimated from the Holter generated R-R list. It was found that the investigated Holter equipment can affect power spectral estimates on the reduced variability tachogram.