Adaptive filtering for ventricular repolarization variability assessment

The purpose of this study was to assess ventricular repolarization variability canceling the influence of heart rate variability using adaptive filtering. Short-term recordings, at supine rest and after head-up tilt, were studied from forty-seven subjects (athletes, cardiac transplant recipient, and normal subjects). RR and RT intervals were extracted to obtain heart rate variability (HRV) and ventricular repolarization variability (VRV) time series respectively. A least mean squares adaptive filter was applied to extract the VRV unrelated to HRV (VRVu). Power spectral densities were estimated from HRV, VRV and VRVu series, and LF, HF were extracted for each power spectrum. Also, percentage power ratio between VRVu and VRV at LF and HF were calculated. Results showed that: i) the heart rate modulation on the ventricular repolarization phenomena showed to be more important in LF band, ii) HF oscillation in VRV is not modulated by the autonomic nervous system.