The Complex Interpretation of Heart Rate Variability Components in Mechanically Ventilated Patients

Sofia Perez-Gracia1, Pablo Armañac-Julián2, Alba Martin1, Josefina López Aguilar3, Francesc Suñol Galofre3, Candelaria de Haro3, Leonardo Sarlabous4, Rafael Fernández5, Montserrat Batlle Solà5, Lluís Blanch3, Raquel Bailón6
1Biomedical Signal Interpretation and Computational Simulation Group (BSICoS), Universidad de Zaragoza, 2BSICoS, University of Zaragoza, 3Institut d'Investigació Parc Taulí I3PT, 4Institut d'Investigació i Innovació Parc Taulí (I3PT-CERCA). CIBER de Enfermedades Respiratorias, Instituto de Salud Carlos III., 5Department of Intensive Care, Fundació Althaia, 6I3A, IIS, Universidad de Zaragoza, CIBER-BBN

Abstract

Aim: Mechanical ventilation (MV) can modulate the cardiopulmonary interaction, potentially altering heart rate variability (HRV) beyond the conventional low (LF) and high-frequency (HF) bands. This study aims to investigate how MV shapes HRV through the appearance of additional spectral components—particularly at harmonic frequencies of the respiratory rate—and to quantify the impact of ventilator-induced modulations and aliasing effects using novel spectral indices that assess the coupling between respiration and HRV in MV-ICU patients.

Materials and methods: A total of 54 recordings from ICU patients under MV were studied during the 24 hours preceding a spontaneous breathing trial (SBT). RR intervals were obtained from ECG signals using a wavelet-based detector, and HRV was reconstructed using the integral pulse frequency modulation (IPFM) model. Power spectral densities (PSD) of both HRV and tidal volume (TV) signals were estimated over 5-minute windows. Peaks at the respiratory frequency (F_R), its second harmonic (2F_R), and potential alias (F_{HR}-2F_R) were identified. Several spectral indices were computed: the relative power at 2F_R with respect to F_R in the respiratory and HRV spectra, depending on the presence of aliasing, and coupling indices quantifying the transfer of power from respiration to HRV.

Results: In controlled ventilation mode, second harmonics in the respiratory signal were observed in over 93.39% of the analyzed time. These harmonics translated into additional components in HRV, including aliasing effects. The HRV spectrum exhibits a higher proportion of harmonic-related power relative to the respiratory signal, supporting the interpretation of cardiopulmonary coupling as a dynamic system with distinct amplification patterns across ventilation modes.

Conclusion: MV induces harmonic components in respiration that manifest in HRV, often through aliasing. These findings underscore the importance of considering additional frequency components for a more comprehensive understanding of HRV and its role in assessing the success of weaning.