Study of Heart Rate Variability as a Marker of Asphyxia/Hypoxia

The onset of labour represents the starting point of a perilous challenge in life, as a new born must adapt to an unknown environment. During this adaptation there are several risks: hypoxia, asphyxia, trauma, intervention and, in worst case scenario, death. These risks can be reduced trough electronic fetal monitoring. During this delicate period the study and analysis of the variability in beat-to-beat intervals of fetal heart rate plays a fundamental role in the pursuit of fetal wellbeing, reduction of fetal morbidity and mortality. Given that the use of an animal model allows direct experimental manipulation of the subjects and their environment and considering the ethical issues and difficulties to acquire data related to asphyxia during labour and delivery, linear techniques (time domain and frequency domain) and non-linear techniques (detrended fluctuation analysis, complexity analysis and Poincarè indices and plots) have been initially implemented for the study of heart rate variability (HRV) using data from the animal model. Data was acquired from experiments in which rats were submitted to controlled episodes of asphyxia (0, 1, 3, 5 and 7 min). Linear and non-linear methods highlighted significant differences in HRV before, during and after the insult. We show how, through a multiparametric analysis, it is possible to detect the onset of asphyxia. Furthermore, tracking the changes in heart rate variability along time, we suggest a novel non-invasive way to assess the amount of injury suffered. With this background we applied HRV analysis to data collected during labour and delivery. We have obtained fetal beat-to-beat heart intervals from non-invasive Doppler ultrasound using Wavelet transform, Hilbert-Huang transform and Autocorrelation function, and these were compared with beat to beat heart intervals extracted from invasive scalp fetal ECG used as a gold standard. For the autocorrelation approach the results of HRV obtained from Doppler ultrasound, using both linear and non-linear analysis, correlate very well with those obtained using fetal scalp ECG. We also modelled and measured the recovery time (from nadir to baseline) following a deceleration of fetal heart rate to study the recovery behaviour and its relation with the development of hypoxic scenarios. The results presented here provide a framework to detect and assess asphyxia by means of linear and non-linear techniques. These techniques have been tested first in an animal model and later in data collected during labour and delivery, showing that Doppler ultrasound provides a reliable alternative for assessing fetal heart rate variations non-invasively during pregnancy and delivery, when fetal scalp ECG is not available. Nevertheless more data needs to be collected and studied using the multiparametric HRV analysis described here to fully validate this approach.