A mathematical model for simulation of fetal heart rate decelerations in labor

Abstract

Fetal wellbeing during labor and delivery is commonly monitored through the cardiotocogram (CTG), the combined registration of uterus contractions and fetal heart rate (FHR). From the CTG, the fetal oxygen state is estimated as the main indicator of the fetal condition. However, this estimate is difficult to make, due to the complex relation between CTG and oxygen state. Mathematical models can be used to assist in interpretation of the CTG, since they enable quantitative modeling of the flow of events through which uterine contractions affect fetal oxygenation and FHR. This thesis describes the development of a model that can be used to reproduce FHR response to uterine contractions during several clinical scenarios. First, a model was developed that describes the relation between uterine contractions, maternal and fetal hemodynamics, oxygen distribution within the feto-maternal circulation and cardiovascular (reflex) regulation in the fetus in response to deviations in blood- and oxygen pressures. The model is partly based on previously presented models for cardiac function, chemoreceptor control in adults and oxygen distribution in the fetal circulation. These modules are coupled and scaled to meet requirements for the (pregnant) maternal and fetal condition. The model is completed with a module for uterine contractions and a module of the vascular system of both mother and fetus. A first clinical scenario was simulated with the model to test model response to changes in cerebral blood flow during the descent of the fetal head in the birth canal. A validation pilot was performed to investigate the quality of model outcome via expert opinion. Experts were unable to discriminate between real and simulated signals, suggesting that the model can be used for educational training. Second, the model was extended with the baroreceptor reflex. This allowed simulation of a second clinical scenario, where both chemo- and baroreflex pathways lead to a FHR deceleration in response to uterine flow reduction during contractions. Results for the uncompromised fetus show that partial oxygen pressures reduce in relation to the strength and duration of the contraction. Furthermore, decelerations during several scenarios of uteroplacental insufficiency were studied. Results for reduced uterine blood supply or reduced placental diffusion capacity, demonstrated lower baseline FHR and smaller decelarations during contraction. Reduced uteroplacental blood volume was found to lead to deeper decelerations only. The model response in several nerve blocking simulations is similar to experimental findings. Third, the model was used to simulate a third type of decelerations, i.e. variable heart rate decelerations, originating from umbilical cord compression. Different degrees of compression were investigated. An increase in contraction amplitude and duration leads to increased umbilical cord compression grade and thus affects the extent of blood pressure increase, flow redistribution and FHR response. There is a clear relation between fetal oxygenation, blood pressure and the resulting FHR. The extent of umbilical compression and thus FHR deceleration is positively related to increased contraction duration and amplitude, and increased sensitivity of the umbilical resistance to uterine pressure. Fourth, gynaecologists, midwives and residents were asked to rate a set of both model-generated CTGs and real CTGs for the three clinical scenarios. Although real tracings were more likely to be recognized correctly, the suitability for use in simulation training was found to be almost equal for real and computer-generated tracings. Due to limited numbers for early and variable deceleration evaluation, statistical analysis turned out to be valid only for the CTG’s with late decelerations. Additional comments from the respondents revealed that variability and regularity of the simulated signals greatly influence the perception of a tracing. Clinicians agreed that a tracing is suitable for use in simulation training when it is clear and free of physiological incompatibilities, which is the case for all simulated tracings. Fifth, the model was used to test the clinical hypothesis that administration of oxygen to the mother may increase FHR during variable fetal heart rate decelerations. The model was used to test the response of fetal oxygenation and heart rate to maternal oxygen increase following 100% oxygen administration. Model outcome suggests that FHR benefits from oxygen administration as the duration and depth of FHR decelerations and fetal oxygenation improves. However, the beneficial effect of maternal hyperoxygenation on FHR and oxygenation reduces during more severe variable decelerations. In conclusion, a model was developed to simulate the physiologic cascade from uterine contraction to changes in fetal heart rate. Model outcome for various scenarios is in correspondence with findings from animal experiments. The model can be used in an educational setting for the simulation of short-term changes in fetal hemodynamics and oxygenation status in response to uterine contractions to increase insight into the complex physiology. In addition, it can be integrated in a full-body delivery simulator to enhance obstetric team training.