Theoretical value of deceleration capacity points to deceleration reserve of fetal heart rate

Objective: The interpretation of Average Acceleration and Deceleration Capacities (AC/DC), computed through Phase-Rectified Signal Averaging (PRSA), in intrapartum fetal heart rate (FHR) monitoring is still matter of investigation. We aimed to elucidate some behaviours of AC/DC. Methods: We derived the theoretical value of PRSA for stationary stochastic Gaussian processes and proved that for these time series AC and DC are necessarily identical in absolute value. The difference between DC and AC, termed Deceleration Reserve (DR), was introduced to detect signals' asymmetric trends. DR was tested on FHR signals from: near-term pregnant sheep model of labor consisting of chronically hypoxic and normoxic fetuses with both groups developing acidemia due to umbilical cord occlusions (UCO); and the CTU-UHB dataset containing fetal CTG recordings collected during labor of newborns that resulted acidotic and non-acidotic, respectively. DR was compared with AC and DC in terms of discriminatory power (AUC), between the groups, after correcting for signals' power or deceleration area, respectively. Results: DR displayed higher discriminatory power on the animal model during severe acidemia, with respect to AC/DC (p<0.05$ but also distinguished correctly all chronically hypoxic from normoxic fetuses at baseline prior to UCO. DR also outperformed AC/DC on the CTU-UHB dataset in distinguishing acidemic fetuses at birth (AUC: 0.65). Conclusion: Theoretical results motivated the introduction of DR, that proved to be superior than AC/DC for risk stratification during labor. Significance: DR, measured during labor, might permit to distinguish acidemic fetuses due to their different autonomic regulation, paving the way for new monitoring strategies.