Cerebral Blood Flow and Metabolism in the Developing Fetus

Acute and long-term oxygen substrate deprivation is a major cause of disrupted fetal brain development and long-term neurologic morbidity. Before truly informed and meaningful brain-oriented fetal care can become a clinical reality, major advances will be required in the understanding of fetal brain hazards and the mechanisms by which normal brain development is derailed. Although a myriad of potential insults may disturb brain development, this article focuses primarily on those intrinsic systems that reduce the risk of fetal cerebral energy deprivation by maintaining a positive balance in cerebral oxygen–energy substrate demand and supply. Because of the inability of the human fetal brain to direct measurements of hemodynamics and metabolism, current understanding is based in large part on data from experimental animal models and from studies of the premature infant, or ex utero fetus. Although both models have provided important insights, neither is ideal, and the understanding of the primary and compensatory support systems for in vivo brain metabolism in the human fetus remains poor. This article reviews the current status of this understanding.

ENERGY SUBSTRATE DEMANDS FOR NORMAL FETAL BRAIN DEVELOPMENT

The energy demands of the developing brain can be classified broadly as those required for its structural growth (accretion) and maintenance and those required for the functional activation of neuro-axonal and glial populations of the brain. These two sources of energy demand are of course wholly interdependent, and particularly during the later stages of development, the microstructural development of the immature brain depends heavily on activation-related trophic stimulation.^1–3 Given the rapid increase in brain mass, which is caused by explosive development of synaptic, dendritic, and axonal elements in the cortical and subcortical gray matter, the cerebral oxygen substrate demands increase exponentially during the later stages of pregnancy.

Structural Brain Development

Anatomic events in early brain development provide the structural substrate upon which later functionally driven changes in microstructural development are imposed. The critical events in brain development of the human fetus are reviewed in more detail in other articles in this issue; however, the broad stages of brain development may be summarized as follows. Primary neurulation leading to formation of the neural tube occurs before 4 weeks of gestation, and by 5 weeks, the neural tube has a well-defined rostral–caudal and dorsoventral organization. Development of the prosencephalon commences between 8 and 12 weeks of gestation, with neuronal proliferation and migration occurring between 8 and 20 weeks of gestation. These phases are followed by major events in cortical organization starting around 20 weeks of gestation and persisting well into postnatal development. Although myelination of the posterior fossa structures occurs during the fetal period, that of the supratentorial tissues occurs primarily after term gestational age.

Around the 11th week of gestation, formation begins of the transient subplate layer, which serves as a relay holding area for thalamocortical projections; this is the site of the first synapses, the activity of which is critical for developing cortical and thalamic projections.⁴ Between 24 and 28 weeks of gestation, there is a period of major refinement of cortical connections with an explosive increase in cortical synapse formation and remodeling, and the development of functionally coordinated outputs from the cerebral cortex.^5–7 In fact, the developing cortex has about 40% more synapses than the mature brain, with the excess synapses being trimmed back by active remodeling. These cortical activities coincide with a phase of major disassembly of subplate synapses and the energy-dependent programmed death of subplate neurons.

Between 28 weeks of gestation and term, these cortical events are largely responsible for an almost threefold increase in brain weight.⁸ The features of this period of accelerated brain growth have been described by quantitative in vivo magnetic resonance imaging (MRI) studies in premature infants⁹ and more recently in the fetus (see the article by Limperopoulos elsewhere in this issue). Such studies have shown that over the course of the third trimester, the cerebral cortex volume increases fourfold,^9,10 as does the volume of the cerebellar hemispheres.¹⁰ This phase of rapid cortical development demands a major increase in energy supply, because cerebral microstructural development depends on functional activation of neuroaxonal units, which in turn depends upon repeated restoration of transmembrane ionic gradients by energy-dependent enzymes (such as Na-K/adenosine triphosphatase [ATPase]).

Functional Brain Development

The functional development of the fetal brain can be viewed from both electrochemical and behavioral perspectives. The accelerated electrochemical maturation of the brain during the latter half of gestation is associated with an increasingly complex repertoire of fetal movement patterns and the emergence of behavioral states. The onset of spontaneous electrocortical activity in the developing human brain remains unclear. Scalp electroencephalography (EEG) recordings in premature infants, however, show bursts of activity alternating with periods of quiescence as early as 24 weeks of gestation.^11,12 These discontinuous patterns of electrocortical activity evolve with gestational age, becoming increasingly more continuous.^13,14 Although the energy requirements of these EEG patterns have not been defined clearly, the bursts of EEG activity have been associated with concurrent changes in cerebral oxygenation.¹⁵ the apparent coupling between these EEG bursts and cerebral hemodynamics suggests that the bursts are energy-dependent. Of note, in the mature brain, electrocortical activity consumes about 60% of all cerebral oxygen delivered to maintain the neuronal ionic gradients required for synaptic and neural activation.

The functional activation of the fetal brain is a critical stimulus for the development of brain structure, especially during the third trimester. Cortical development depends on early neuronal activation patterns, because these guide critical processes of neuronal differentiation, neuronal migration, synaptogenesis, and formation of neuronal networks.¹⁶ One proposed model for somatosensory cortex development is that spontaneous spinal and subcortical discharges elicit motor activity in the periphery, where sensory events associated with these movements then activate afferent signals back to the sensory cortex; in this manner, topographic representation of afferent input to the sensory cortex becomes established.¹⁷ This paradigm is supported by data showing that spindle-burst EEG oscillations (delta brushes) are triggered by sensory feedback from movements in the periphery activated by subcortical discharges.^16–18 Another example of the complexity of these developmental electro-behavioral phenomena relates to fetal heart rate changes coupled to fetal movements. These heart rate accelerations initially were thought to reflect an increase in cardiac output in response to the energy demands of the motor activity. Both the frequency and amplitude of these heart rate changes, however, persist even in paralyzed fetal animals, suggesting that both the movements and heart rate changes are mediated by concurrent central efferent discharges.¹⁹ Although much about these processes is not understood, it is likely that they are energy-dependent neural discharges that will be impaired by cerebral energy deprivation. Whether the time course and development of these EEG patterns in the preterm infant are similar to those in the gestational age-equivalent fetus remains unclear. The exciting novel techniques for assessing electrocortical development in the human fetus discussed by Lowery colleagues elsewhere in this issue may provide important insights into these complex developmental events.²⁰

Characterization of the development of fetal movement patterns and behavioral states has been advanced by fetal ultrasound studies, particularly the advent of four-dimensional ultrasound and the more recent development of fetal MRI (see the article by Prayer and colleagues elsewhere in this issue). These techniques have shown the evolution of fetal behavior patterns, which include an increasingly complex array of spontaneous and reflex movements of varying speed and amplitude, and the development of fetal state changes.²¹ Fetal movements start in the late first trimester and expand in repertoire and frequency throughout pregnancy. During the second and third trimesters, these movements become organized into increasingly complex and clearly distinct behavioral patterns. After midgestation, human fetal behaviors develop a periodic nature, in which movements alternate with periods of quiescence.^22,23 The initial simple movement patterns originate from spontaneous discharges in the spinal and brainstem circuitries; the subsequent emergence of complex and variable movements denotes modulation of these spinal–brainstem activities by descending input from higher brain centers. From around 28 weeks of gestation, the topographic organization of neural connections reaches a level that allows goal-directed behaviors to emerge.²⁴ Fetal sensorimotor reflex systems may be tested using vibroacoustic stimuli to elicit fetal startle responses.²⁵ Because such responses may originate at a brainstem–spinal level, however, a more reliable test for cerebral modulation of these behaviors is development of habituation to repeated stimuli.²⁶ From around 36 weeks of gestation, fetal behavioral states emerge, characterized by combinations of behavioral conditions that remain stable for periods of time.

DEVELOPMENTAL CONSEQUENCES OF IMPAIRED CEREBRAL ENERGY SUPPLY IN THE FETUS

The precise metabolic needs of the developing brain, especially in early pregnancy, remain poorly understood. Oxygen substrate deprivation, however, has been implicated in various developmental lesions in the fetal brain. More detailed discussion of these issues appears elsewhere in this issue. It suffices to say that oxygen substrate deprivation may cause long-term effects through disturbances in fetal programming, selective cell-specific injury, disruption of programmed brain development, or frank encephaloclastic lesions, depending on the nature and timing of the insult. The specific effects on brain development triggered by these disturbances in oxygen substrate supply depend on several factors, including the nature of the insult and the specific developmental events at the time of the insult.

In addition to oxygen–energy substrate deprivation, specific nutritional deficiencies have been implicated in specific developmental anomalies of the nervous system. For example, dietary folate deficiency has been associated with early first trimester disruption of neural tube development.²⁷ During later first trimester prosencephalic development, maternal diabetes and disturbed fetal cholesterol metabolism have been associated with malformations in the holoprosencephaly spectrum.^28,29 Later in gestation, interruptions in oxygen–energy substrate supply have been implicated in neuronal migration defects, such as lissencephaly and the more localized schizencephaly lesion and postmigrational disturbances such as layered polymicrogyria.^30,31 Disturbances in brain growth have been described by quantitative MRI studies in premature infants with intrauterine growth restriction (IUGR).^32,33 Furthermore, IUGR has been associated with impaired neuropsychological outcome.^34–36 Delayed brain development with structural immaturity recently was described in populations at risk for cerebral oxygen–glucose deprivation, such as fetuses with certain congenital heart lesions (see the article by Limperopoulos, elsewhere in this issue).³⁷ The relationship between hypoxia and the developmental stage is also evident on a cellular level. For example, recent studies suggest that hypoxia has different effects on the oligodendrocyte lineage at different stages of development. Akundi and colleagues³⁸ have suggested that very early oligodendrocyte precursors exposed to hypoxia undergo accelerated maturation, while hypoxia later in oligodendrocyte development triggers either degeneration or maturational arrest.³⁹ These are some examples of the interaction between cerebral energy deprivation and adverse fetal brain development. It is likely that a far wider range of cellular and subcellular disturbances will be elucidated in the future with more advanced research techniques.

SYSTEMS THAT SUPPORT OXYGEN SUBSTRATE SUPPLY TO THE DEVELOPING FETAL BRAIN

Normal fetal growth and development depend upon a supply system that keeps pace with the escalating oxygen substrate demands. From the perspective of brain development, such a supply system involves multiple complex arrangements along the path from the placental circulation (both maternal and fetal components) to the fetal systemic and cerebral circulations. Disturbances at any point along this path will jeopardize normal fetal brain development.

Development of the Placental Circulation

The placenta plays a complex and active role in fetal development, as discussed elsewhere in this issue in the article by Redline. Disturbances in placental function may cause a spectrum of sequelae, from fetal demise to cerebral developmental disruption, to the adverse effects of fetal programming, a mechanism whereby an insult at a critical period of fetal development exerts remote effects in later years.⁴⁰ Placental function evolves in a developmentally programmed sequence of steps across gestation, allowing it to support the 40-fold increase in the fetal/placental weight ratio between 6 and 40 weeks of gestation.⁴⁰ Insults such as hypoxia during this process may disturb future placental function, including disturbances in nutrient transfer and transporter protein expression.^40–42

The placenta is perfused by two circulations (ie, the maternal uteroplacental circulation and the fetal umbilicoplacental circulation). Although these two circulations share a close anatomic relationship, being separated by only a single tissue layer several cells thick, these two circulations are for the most part functionally distinct.⁴³ However, as the only sources of fetal oxygen substrate supply, alterations in either circulation may have important consequences for fetal growth and development. There is an exponential increase in placental perfusion during the second half of pregnancy, mediated by different mechanisms in the two circulations. Normally, the volume of uteroplacental blood flow far exceeds that required for fetal growth, creating a buffer zone or safety margin.⁴³ The uteroplacental blood flow increases by four- to fivefold without significant increases in arterial number. Instead, this uteroplacental blood flow occurs primarily because of vasodilation of uterine resistance vessels (spiral and more proximal arteries) during the later stages of pregnancy. Concurrently, there is an exponential volumetric increase in umbilicoplacental blood flow during the last trimester that primarily is caused by increased growth of the placental vascular bed and, to a lesser extent, vasodilation. Consequently, with advancing gestation, there is a progressive decrease in umbilicoplacental resistance as detected by Doppler ultrasound.^44,45

Umbilical and placental vessels lack autonomic innervation, and regulation of placental perfusion depends upon local or circulating vasoactive substances.^46–50 There are important functional changes in the placental response to endogenous vasoactive agents, with certain notable differences in the reactivity of the two circulations. During normal pregnancies, women develop changes in reactivity of their systemic and placental circulations that favor enhanced placental perfusion. For example, there is a decreased response of the systemic vasculature overall to the vasoconstrictor angiotensin-II; however, this effect is most striking in the umbilicoplacental circulation,⁵¹ creating a relative perfusion steal toward the placenta.⁵¹ Of note, this response does not occur in women who develop pregnancy-associated hypertension,⁵² in whom relative placental hypoperfusion may develop.⁵³ A similar attenuation in systemic vasoreactivity to the vasoconstrictor effects of the catecholamines develops in pregnant women. This attenuated vasoreactivity, however, is less prominent in the uteroplacental circulation compared with the systemic circulation; as a result, during periods of endogenous catecholamines, the uteroplacental circulation is not protected, and placental hypoperfusion may develop. The umbilicoplacental circulation differs from the uteroplacental circulation in that angiotensin-II is a potent vasoconstrictor,⁵⁴ while catecholamines at physiologic doses have minimal effect on umbilicoplacental blood flow. As a result, the increase in catecholamines during fetal hypoxemia causes significant vasoconstriction in the fetal peripheral circulation but not in the umbilicoplacental, myocardial, or cerebral circulations.

The placenta also plays a pivotal role in regulating the fetal endocrine and metabolic physiology. For example, glucocorticoids are important for the growth and development of fetal organ systems through several different pathways. Excess glucocorticoid exposure, however, may affect fetal development adversely, for example by decreasing the expression and function of glucose transporters.⁵⁵ The placenta regulates fetal glucocorticoid exposure by, for example, the placental enzyme 11β-hydroxysteroid dehydrogenase (11β-HSD), which converts active cortisol to inactive cortisone, protecting the fetus from excessive maternal cortisol production.⁵⁶ The levels of placental 11β-HSD production increase with fetal maturation,⁵⁷ but its expression and activity are decreased by hypoxia.⁵⁸

Development of the Fetal Systemic Circulation

Given the dependence of the fetus on the placenta rather than the lungs and intestine for oxygen and nutrient supply, there are of necessity major differences between the fetal and postnatal circulatory systems. The fetus has two arterial circulations operating in parallel and connected by intra- and extracardiac shunts, whose patency is critical for normal fetal development. The primary communication between these arterial systems is at the aortic isthmus, which is the watershed between the circulation to the brain and upper body on one side, and between the subdiaphragmatic and placental tissues on the other side.⁵⁹ Two special features of the fetal systemic circulation are important for the normal development of the fetal brain. First, under normal conditions, there is preferential oxygen substrate delivery to the aorta and cerebral circulation. Second, the fetal central circulation is very flexible, with adaptive mechanisms capable of maintaining preferential oxygen substrate supply to the developing fetal brain during periods of fetal hypoxia. Specifically, this is achieved by changes in regional vascular resistance in the two circulations, which in conjunction with fetal shunts (discussed later in this article) provide preferential oxygen substrate supply to the developing brain by changes in both the volume and oxygen substrate composition of regional tissue.^60,61

The highest fetal oxyhemoglobin saturations (85% to 90%) are in the umbilical vein as it enters the fetal abdomen. About half of this umbilical venous return enters the ductus venosus, bypassing the liver and thereby minimizing oxygen extraction. The oxygen substrate rich blood in the ductus venosus enters the inferior vena cava just proximal to the right atrium, from where it is diverted across the foramen ovale into the left heart and ascending aorta. In this manner, blood perfusing the coronary arteries and cerebral circulation has an oxyhemoglobin saturation significantly higher (65%) than that passing through the pulmonary artery (50% to 55%), ductus arteriosus, and into the descending aorta. These preferential streaming patterns have been demonstrated in the human fetus and in animals.⁶² Under normal circumstances, the vascular resistance of the placental circulation is significantly lower than that of the pulmonary circulation, enabling the right ventricle to drive systemic perfusion through the ductus arteriosus.⁶³ In fact, the low placental vascular resistance allows half the combined ventricular output to pass into the umbilical circulation. When placental resistance increases in pathologic conditions, however, the upstream effects may increase the amount of deoxygenated blood from the vena cavae that is diverted across the foramen ovale and into the ascending aorta.

The fetal cardiovascular system differs in other important ways from that of the term newborn infant. The fetal myocardium generates lower active tension to myocardial stretch but has a higher resting tension, which limits diastolic performance.^64,65 Earlier studies suggested that the immature fetal myocardium was limited in its ability to increase stroke volume, leaving it dependent on increasing heart rate to increase cardiac output.⁶⁶ Subsequent studies showed that the fetal myocardium follows the basic Frank-Starling model, and that fetal cardiac output can increase by mechanisms other than heart rate.⁶³ Autonomic, and particularly sympathetic, innervation of the fetal heart is immature and incomplete. The expression of myocardial β-adrenergic receptors is initially low and increases with gestation, while α-adrenergic receptor expression is initially higher and decreases with maturation. Although myocardial norepinephrine stores are low in the fetus,⁶⁷ the immature myocardium has an increased sensitivity to circulating norepinephrine,⁶⁸ which is important during labor and the early newborn period, particularly in the stressed newborn, when circulating catecholamine levels increase.

Development of the Fetal Brain Circulation

By the end of the first month of development, soon after closure of the neural tube, a primordial system of endothelium-lined vascular channels is present in the developing nervous system. By the end of the embryonic period, an extensive mesh of leptomeningeal arteries covers the cerebral hemispheres. Subsequent development of the brain vasculature is coupled tightly to the structural development of the brain. Arterial development of earlier developing structures proceeds more rapidly and is completed in the brainstem and cerebellum between 20 and 24 weeks of gestation, and in the basal ganglia and diencephalon by 24 to 28 weeks of gestation. The phase of accelerated third trimester cerebral hemispheric development is accompanied by rapid development of the hemispheric vasculature. Initially, a single vascular plexus covers the cerebral surface with extensive anastomoses between the major arteries of supply. From this surface network, superficial vessels begin to penetrate the brain parenchyma around 7 weeks of gestation, starting with large caliber basal penetrators that supply the basal ganglia and diencephalon and the germinal matrix in the subependymal periventricular zones. Thinner penetrating vessels descend from the cerebral surface; by 24 weeks of gestation, when the major phase of cerebral neuronal migration has been completed, these long penetrating vessels extend from the pial surface into the periventricular regions. The period between 24 and 28 weeks of gestation is one of particularly rapid cortical organization, axonal outgrowth, and synapse formation, and it is accompanied by rapid ingrowth of short penetrators between the earlier long penetrators.

Functional development of cerebral vasoreactivity follows a similar pattern. Specifically, the muscularis layer responsible for regulating cerebral vascular resistance initially is confined to the pial vessels and superficial penetrators, descending with maturation into the cerebral parenchyma.⁶⁹ Consequently, cerebral vasoregulation in the earlier fetal brain occurs largely in the superficial regions more remote from distal vascular beds than the precapillary arterioles of the mature brain. This fact may be of relevance to the vasoregulatory mechanisms that will be discussed.

Cerebral Metabolism in the Immature Brain

Very little is known about the specific energy requirements of the immature brain in the human fetus, particularly during the first half of gestation. This lack of knowledge is largely because of the inaccessibility of the fetal brain to currently available measurement techniques. In fact, current understanding is based largely on data extrapolated from animal studies, and studies in the premature infant (ex utero fetus).^70–73 Besides postconceptional age equivalency, however, the major differences in internal and external environment of the fetus and premature newborn seriously limit the applicability of data from premature infants to the fetus.

For several reasons, anaerobic glycolysis has been considered an important energy source for fetal cerebral metabolism. First, the partial pressure of oxygen in the umbilical vein is significantly lower than in the normal postnatal circulation. Second, earlier studies had demonstrated an inefficient coupling between oxygen availability and energy production in the immature brain, limiting its capacity for mitochondrial oxidative phosphorylation and making it less efficient at energy synthesis.⁷⁴ Third, the immature fetal brain is uniquely capable of using alternative energy substrates such as lactate and ketones during periods of glucose limitation.⁷⁵ The ability of the immature brain to metabolize alternative substrates relates to the maturational arrangement of cerebral membrane proteins that mediate transport of organic substrates into the developing brain. Specifically, the two distinct types of transporter proteins, namely glucose transporters (GLUT)^76,77 and monocarboxylic acid transporters (MCT), have opposite trajectories of developmentally regulated expression. In early gestation, GLUT expression in the immature brain is low but increases with brain development.^78,79 Conversely, MCT expression, which facilitates transport of lactate and ketones into the immature brain, decreases with brain maturation.^79,80

Although the immature brain is capable of using these alternative energy substrates, which in animal models may support up to 60% of cerebral energy demand,⁸¹ the principal substrate supporting normal fetal cerebral metabolism throughout all phases of development is glucose, and most of it is consumed aerobically, with little or no production of lactic acid. At no point in development is anaerobic glycolysis solely capable of supporting the energy requirements of the developing brain. Fetal tissue oxygen delivery, on the other hand, depends only on the partial pressure of oxygen but also upon the circulating oxygen content and blood flow. When considered in these terms, there is little difference in oxygen delivery between the fetus and ex utero animal. Several facts support the notion that the umbilical venous oxygen supply is sufficient to support aerobic metabolism in fetal tissues. First, during fetal hypoxemia, the umbilical venous–arterial oxygen extraction fraction increases from around 40% up to 50% to 60%. Conversely, under normal baseline conditions, supplemental maternal oxygen with fetal hyperoxia does not increase fetal oxygen extraction. Finally, under normal circumstances, the umbilical venous–arterial lactic is not increased; in fact, umbilical venous lactate may be higher, suggesting fetal lactate uptake. The notion that oxygen delivery to the normal fetal brain is adequate to support cerebral energy metabolism is supported further by the fact that energy use is substantially lower in the immature animal brain.⁸² For example, cerebral energy use in the newborn rat is 20 times lower than in the adult rat brain.⁸³ For these reasons, it has been argued that although immature animals are capable of using alternative energy sources, their ability to survive significantly longer periods of hypoxia than adult animals is largely because of their decreased cerebral energy requirements.⁸² Although cerebral energy demand starts off low in early fetal life, it escalates rapidly during the third trimester to support the function of enzymes, such as Na/K-ATPase, which are critical for maintaining electrocortical activity and the propagation of action potentials. With these escalating cerebral energy demands, there is a concurrent increase in the number and functional capacity of cerebral mitochondria, with increasingly efficient energy production.⁷⁴

Studies of cerebral metabolism in the immature human brain are confined to the ex utero premature infant. Positron emission tomography (PET) has been used to study in vivo oxygen and glucose metabolism in premature and term infants. Altman and colleagues⁷² studied the cerebral metabolic rate of oxygen (CMRO2) in third trimester newborns ranging between 26 to 40 weeks of gestation. The CMRO2 was reduced in all infants, but particularly in the smallest, presumably reflecting the functional immaturity of the brain.⁸⁴ In the fetal brain, both oxygen and glucose metabolism are highest in the brainstem and deep gray matter and decrease in a rostral direction, being lowest in the subcortical and other white matter structures. In premature infants between 25 and 37 weeks of gestation, cerebral glucose metabolism as measured by PET was lower than in full-term infants.^70,71 As with oxygen metabolism, glucose metabolism was highest in the brainstem and cerebellar structures.

COMPENSATORY SYSTEMS THAT PRESERVE OXYGEN SUBSTRATE AVAILABILITY IN THE FETAL BRAIN

Restriction in fetal oxygen substrate supply triggers several endogenous fetal compensatory responses at both systemic and cerebral levels, aimed at preserving a positive energy balance by decreasing energy demand and by increasing substrate supply.^85–89 The efficiency of these adaptive responses decreases as the insult dose (duration and severity) increases. In fetal sheep, these adaptive circulatory responses can be maintained for prolonged periods of moderate hypoxemia as long as metabolic acidosis does not develop.^90–93 In fact, in these animals, global and cerebral oxygen metabolism fails only when the circulating pH falls below 7.0.^91,93

Fetal Responses that Decrease Energy Utilization

Several physiologic responses triggered by hypoxemia aim to restrict fetal energy expenditure, both in the systemic and cerebral tissues.^94–98 For example, myocardial energy consumption is decreased by means of a neural reflex through which hypoxemia activates chemoreceptors, which in turn elicit a vagally mediated bradycardia^99–103 with an initial fall in blood pressure. This response is followed by a marked vasoconstriction in the peripheral circulation and an increase in blood pressure, a sympathetic response likely triggered in part by the hypoxemic stimulation of the chemoreceptors. With increased sustained hypoxemia, this bradycardia cannot be reversed by atropine, indicating that it no longer is mediated vagally but likely caused by hypoxic myocardial suppression.¹⁰³ At the level of the nervous system, the powering down of neuronal activity manifests as a decrease in fetal movements. This decrease in movement is accompanied by changes in the fetal EEG background to a high-voltage slow-wave pattern, which is less energy-demanding. In turn, it is likely that these EEG changes result from neuronal suppression. Adenosine is a breakdown product of adenosine triphosphate (ATP), which accumulates during failure of ATP resynthesis. Through actions at inhibitory presynaptic receptors and dilating vascular receptors, adenosine suppresses neuronal activation and increases perfusion.⁸⁹ These responses differ to some extent depending on the mechanism of fetal hypoxemia and level of fetal maturation. If hypoxemia develops very slowly, fetal behavior may not decrease until acidosis develops.⁹⁸ Severe cerebral hypoxia will limit cerebral oxygen metabolism directly during the terminal phases, leading to irreversible cellular injury.¹⁰⁴ When hypoxemia develops as a result of umbilical cord compression, the pattern of responses may be different. Specifically, umbilical arterial compression triggers an immediate increase in blood pressure with marked bradycardia, caused in this case by an early baroreceptor rather than chemoreceptor response.¹⁰² In the preterm sheep fetus, hypoxemia fails to trigger bradycardia or an increase in blood pressure, possibly because of a lack of chemoreceptor responses at this gestational age.¹⁰⁵

Fetal Responses that Increase Cerebral Oxygen Substrate Supply

An initial response to decreasing fetal oxygen delivery is an increase in fetal oxygen extraction, a mechanism capable of sustaining fetal oxygen delivery until the umbilical venous oxygen content falls to around 50% of normal. During fetal hypoxemia, flow to all organs is maintained relatively constant until the arterial oxygen content falls to around 50% of normal; before the development of metabolic acidosis, there are no significant changes in either fetal cardiac output or in umbilicoplacental perfusion. At circulating oxygen levels below about 50% of normal, oxygen extraction may persist but is unable to compensate fully for the decrease in oxygen delivery, and tissue oxygen consumption begins to fall.¹⁰² At this point, several compensatory changes in the fetal circulation are initiated to optimize oxygen substrate delivery to the fetal brain, heart, and adrenals. Hypoxemia results in an increased shunting of umbilical venous blood into the ductus venosus,^106–108 and from here through the foramen ovale into the left heart and up to the brain.⁸⁸ At the same time, there is a marked increase in sympathetic tone, mediated in part by the chemoreceptor pathways and an increase in circulating catecholamines, which results in peripheral vasoconstriction and an increase in blood pressure, but decreased perfusion of the kidneys, liver, intestine, and musculoskeletal system.^88,109 This response is known as circulatory centralization, or the brain-sparing effect.¹¹⁰ This centralization response is mediated by neural (mainly adrenergic), circulating (catecholamines, serotonin, and angiotensin-II), and local mechanisms (eg, nitric oxide [NO], prostaglandins).

Intrinsic Cerebral Compensatory Responses to Fetal Hypoxia

In addition to the previously mentioned compensatory responses in the systemic circulation, cerebral vasodilation in response to acute hypoxia has been described in fetal models less than 0.7 weeks of gestation.^111,112 Within the brain, hypoxemia triggers a redistribution of cerebral perfusion, such that blood is redistributed preferentially to the most actively developing brain regions at the particular gestational age. Most data for fetal cerebral responses to hypoxia are derived from animal fetal models, and to a lesser extent from the premature infant (or ex utero fetus). In considering these data, several important points are worth noting. First, because these responses aim to restore adequate oxygen substrate delivery, lower baseline demands as occur in the more immature brain will require less striking compensatory responses during cerebral hypoxia. There are important species differences in the level of brain maturation at different gestational ages, and the cerebral hypoxic responses must be considered in this context. In the fetal sheep, the most commonly used species in these studies, both cerebral blood flow and oxygen consumption are reduced significantly at 0.6 weeks of gestation, being only a third of that in the near-term fetus.^113,114 Using data from prematurely born infants is suboptimal, because there are significant differences between the fetal and postnatal circulation. Adaptive responses in development are rapid, and consequently within a very short time, the premature extrauterine circulation is no longer representative of the gestational age equivalent fetal circulation.

During fetal hypoxemia, there is not only a global increase in cerebral blood flow but also a redistribution of blood flow within the brain, such that brainstem perfusion exceeds cerebellar perfusion, which in turn exceeds blood flow to the cerebrum^88,115 in the fetal brainstem. This robust vasodilatory response in the fetal brainstem makes it significantly more resistant to hypoxic injury than other brain regions.¹¹⁶ These intrinsic cerebrovascular responses are part of a complex autoregulatory system aimed at maintaining metabolism in the most actively developing and physiologically critical brain structures. Although most if not all intrinsic vasoregulatory (autoregulatory) systems present in the mature brain begin to develop during the fetal period, their efficiency is related to the maturational level of the fetus.¹¹⁷ Furthermore, unlike the mature brain, in which the endothelium is a major source of vasoregulatory substances, the endothelium plays a lesser role in the cerebrovascular response to hypoxia in the fetus.^118,119 A review of the multiple different vasoregulatory mechanisms in the fetal brain is beyond the scope of this article; instead, the discussion will be confined to the intrinsic cerebrovascular responses to hypoxemia. Hypoxic vasodilation is mediated by an interplay between different endogenous substances, including NO, adenosine, endogenous opioids, and adrenomedullin. Hypoxia also may have direct vasodilating effects at the muscularis of the cerebral resistance vessels. Furthermore, the regional differences in cerebral perfusion during hypoxemia likely result from a complex interplay between regional vasodilator synthesis and the generation of regional vasoconstrictor substances.¹²⁰ In the fetus, the complex interplay between cerebral vasoactive substances during hypoxemia is characterized incompletely.¹²¹

NO is an important mediator of hypoxic vasodilation in the mature brain. The hemodynamic importance of NO in the developing fetal circulation, however, remains incompletely understood. In the fetus, it appears that NO plays a tonic vasodilator role during normoxia,¹²² but its role in hypoxic vasodilation remains unclear¹¹³ and appears to differ across species.^123,124 In the developing fetal cerebral cortex, there is a threefold increase in neuronal nitric oxide synthatase (NOS) activity during late gestation.¹²² As discussed previously, the ATP breakdown product adenosine mediates cerebral hypoxic vasodilation through an action on vascular adenosine (A2) receptors.^125,126 It additionally inhibits cerebral metabolism during hypoxia by stimulating presynaptic (A1) receptors.¹²⁷ Adenosine-mediated vasodilation is developed fully by 0.6 weeks of gestation in most species¹¹¹ and is thought to mediate about half the hyperemia during acute hypoxia.¹²⁷ Although hypoxia also increases prostaglandin levels,¹²⁸ the role for prostanoid vasodilation during hypoxia in the immature brain appears to be modest at best.¹²⁹

SUMMARY

Normal development of the fetal brain is heavily dependent on an oxygen substrate supply capable of supporting the energy demands of the complex structural and functional maturational processes. Consequently, oxygen substrate deprivation is a major cause of injury to the developing brain, with a broad spectrum of sequelae involving not only encephaloclastic lesions but also derailment of critical programmed developmental events. Although there are ongoing obstacles to the understanding of cerebral blood flow and metabolism in the human fetus, a growing number of innovative neurodiagnostic techniques are emerging (and are discussed elsewhere in this issue) that are likely to advance understanding and lead to meaningful brain-oriented fetal care in the future.