A system approach to stillbirth, perinatal brain injury and unexpected deterioration of fetal heart rate tracing.

 This is a brainstorming idea. I need criticism and more ideas. Like most brainstorming, it is a broad idea that needs to be narrowed and focused.

            As a pathologist who has performed over 1,000 autopsies on stillbirth infants, I am frustrated that at least one third of those autopsies do not reveal a convincing chain of causation accounting for the infant’s to death. As an obstetrical pathology medico-legal consultant reviewing more than 500 placentas from infants who meet the criteria for birth asphyxia and have subsequent brain damage, I am more often than not, unable to find a definitive explanation as to why the infant suffered such a tragic outcome. Every week, I look at several placentas sent for examination because of an unplanned Cesarean section that was required due to non-reassuring fetal heart rate tracings. In many cases, there is no clear explanation. In these examples, there may be some pathological lesions in the placenta, but these same lesions are present in infants without serious complications. 

            Of course, in some of the above situations, there is an explanation for the obstetrical complications that are understandable. These explanations include a wide range of pathologies such as large abruptions, tight umbilical cord knots, fetal to maternal hemorrhage, and infection with disease causing microorganisms. However, even with a discoverable anatomic cause of death, questions often remain. How large of a placental separation in abruption is lethal for a given infant? When and how did an umbilical cord knot tighten? How did a fetal hemorrhage occur and how rapid a blood loss is needed to produce death or brain damage?  The exact degree of compromise may vary for each placenta and fetus.

            These questions underly a research project that I proposed to find better explanations for stillbirth. The first part of the study simply provides an MRI before delivery of the infant which will show anatomic positions that might compromise fetal umbilical blood flow but would not be detectable after delivery. The second part was directed to the general question of cumulative factors that might lead to fetal death by tipping the fetus into a lethal feedback loop that would lead to death. For example, if a certain event led to a just enough decrease in oxygen level to produce fetal lactic acidosis, this acidosis could decrease cardiac contractility, hence causing the heart to pump less blood to the placenta, thus further reducing oxygen, creating a spiral to fetal death. The idea of a tipping point as a mechanism of stillbirth without one single cause seemed probable from the lack of definitive findings at autopsy in so many cases. The kinds of events we are asking about in the stillbirth study, such as supine sleeping or possible sleep apnea, would not normally cause a tipping point in themselves, unless other factors were also present such as a compromised umbilical cord, or a placenta that was not adapting fast enough to a growing fetus. This idea a tipping point forced me to think of stillbirth as a placental system failure. 

            This system includes all the components needed to provide oxygen for brain and heart survival, such as the functional placenta reserve for oxygen diffusion, fetal metabolic level, the fetal and maternal cardiac output, fetal and maternal hematocrit, maternal blood gases, etc, etc. Many of the individual components have been studied. The same idea of placental system failure in stillbirth, would also apply to birth asphyxia or non-reassuring fetal heart tones. The components that lead to failure in any individual case might be very different, but the final pathway would be the same. Within the feed forward loop to hypoxia, rescue might be possible if the onset of placental system failure were detected in time. Prediction of stillbirth or asphyxia during labor might might depend on different factors in the individual cases. Many current tests of fetal well-being as a result do not have high sensitivity and specificity.  A model using multiple inputs to the placental system could better predict the risk that each factor contributes. This broader understanding could help avoid catastrophic obstetrical complications, and improve specificity to avoid unnecessary emergency operative deliveries. 

            Because of the many relationships between the different factors, there is unlikely to be a simple solution to predicting placental system failure. This kind of system prediction, like a weather forecast, requires a model of the system components and their known interactions. This approach to the complexity of placental system failure views fetal survival states as attractor states in a complex chaotic system. There are recursive interactions between fetal blood pressure, fetal pH, fetal cardiac output, utero-placental blood flow, maternal oxygenation, etc. The inputs on some components can be measured before labor, such as measures of fetal metabolic requirement for oxygen (heart and brain size) or placental efficiency of oxygen transfer between contractions (placental functional volume, or even better functional MRI estimates). I think the anatomic degree of umbilical wrapping and kinking is also important. These measures change during the pregnancy, but they can be measured near term. Other measures that could be obtained during labor include uterine pressure tracings, heart rate tracings, and perhaps a continuous doppler measure of cardiac output (umbilical vein diameter in a good cross section times the flow rate). The outcome could include the umbilical artery and vein gases, the cord hematocrit, etc. The quantitative descriptions of the interactions of variables on fetal heart rate (includes carotid chemoreceptor, CNS vagal effects, cord occlusion effects, which include effects on cardiac dynamics, not just gas exchange), and on fetal acidosis such as metabolic needs (brain and heart size, inflammatory or anemic elevations of heart rate, etc. and on fetal cardiac output (pressures, ductal flow modulation, contractility based on substrate, oxygen and blood pH) have been studied individually, usually in sheep but some in humans and other primates. The goal is to build a model from all this data that will demonstrate what happens when different stresses (changes in the parameters) are put on the system, such as maternal supine position, prolonged contractions or close together contractions, or fetal stresses such as hemorrhage, or inflammation. The effects in the model can then be compared to patient results and eventually improve predictive value in patient care and ideally improve our ability to predict placental system failure.