Hypoxic-ischemic encephalopathy (HIE) following severe perinatal asphyxia (also described in the literature as perinatal hypoxia-ischemia or asphyxia neonatorum) has an incidence of 1 to 2 per 1000 live births in the western world, and is far more common in developing countries (see Chapter 10). Although metabolic disorders may mimic perinatal asphyxia, and genetic and placental factors may contribute to the clinical picture, brain imaging techniques have demonstrated acute changes in the term neonatal brain following perinatal asphyxia. The chance of irreversible damage or death following perinatal asphyxia is high, up to 65% of patients enrolled in trials of neuroprotective strategies. Therapeutic hypothermia is neuroprotective as has been demonstrated in several trials, and is standard therapy for full-term or near-term neonates with severe perinatal asphyxia and encephalopathy.47 Ongoing studies will aim at additive strategies to augment the neuroprotection of hypothermia.87 Experiments in animals have demonstrated that the immature brain is more resistant to hypoxia-ischemia than the brain of the term neonate. The several reasons to explain this difference are a lower cerebral metabolic rate; lower sensitivity to neurotransmitters with potential neurotoxicity; and the greater plasticity of the immature central nervous system. Nevertheless, in the fetus and newborn, cerebral hypoxia-ischemia is a major cause of acute mortality and morbidity in survivors. However, the neuropathology will be different from that of the full-term neonate (see Chapters 59 and 60). The severity of HIE can be defined as mild, moderate, or severe depending on clinical findings as described by Sarnat and Sarnat93; this classification is widely used and is summarized in Table 61-1. More recently, other clinical scoring systems have been developed to assess the severity of HIE (Table 61-2).21,104 TABLE 61-1 Clinical Staging of Hypoxic-Ischemic Encephalopathy Modified from Sarnat HB, et al. Hypoxic-ischemic encephalopathy following fetal distress: a clinical and electroencephalographic study. Arch Neurol. 1976;33:695. TABLE 61-2 From Thompson CM, et al. The value of a scoring system for hypoxic ischaemic encephalopathy in predicting neurodevelopmental outcome. Acta Paediatr. 1997;86:757-761. Severe fetal hypoxic-ischemic injury affects the entire organism, and these effects have been well studied in animal models. In particular studies, instrumented fetal sheep and monkeys in the 1960s and 1970s had been used to describe the physiologic and pathologic changes in the brain following hypoxia.22,74 Hypoxic-ischemic injury may occur at any time during pregnancy, the birth process, or the neonatal period. The pattern of brain damage is reflected by the gestational age of the fetus at the time that the injury occurs. Fetal hypoxic-ischemic injury may result from maternal, uteroplacental, or fetal problems (Box 61-1). The fetus may survive maternal hypoxia-ischemia such as transient hypoxia or hypotension. Correctable placental factors include hyperstimulation with oxytocic agents or intermittent cord compression, but these may cause irreversible brain damage before recovery occurs. Survival with brain damage as the result of early in utero hypoxic-ischemic insult does occur, but in the majority of cases severe maternal hypoxia in the second trimester of pregnancy will result in fetal death.100 Occasionally there is a history of a catastrophic maternal illness such as suffocation, anaphylaxis, or major physical trauma. In other situations, the antecedent pathogenic event is confined to the fetus, with or without an associated abnormality of the uteroplacental unit. Whatever the cause of the cerebral hypoxia-ischemia, the neuropathologic consequence is often devastating. In the more mature fetus, a period of mild to moderate hypoxemia produces a consistent pattern of responses (Figure 61-1). Initially there is fetal bradycardia with an immediate rise in blood pressure and, in particular, an increase in perfusion to the brain and other vital organs at the expense of the rest of the body. With ongoing hypoxemia, fetal heart rate will decrease further, apnea will occur, and permanent brain injury occurs after 10 to 15 minutes. The fetus is very resistant to milder hypoxemia, and normal cardiovascular function will be maintained for up to an hour even with a Pao2 of 15 mm Hg (normal fetal Pao2 is 25 mm Hg). With prolonged moderate hypoxia, cerebral perfusion will remain normal, but (asymmetric) fetal growth restriction will occur. In fetuses with such prolonged moderate hypoxia, lactate levels may be elevated, indicating that anaerobic glycolysis has occurred in some tissues. The rapid changes in blood gas values have been documented in numerous experiments (Figure 61-2). As the hypoxic insult becomes more severe, changes in regional cerebral blood flow occur. The brainstem is able to extract sufficient oxygen to maintain metabolism despite very low Pao2 at the expense of the cerebrum. Failing myocardial function may cause a fall in cardiac output, and the watershed areas of the cerebral hemispheres are most exposed to damage (see Vascular Territories). An acute hypoxic-ischemic insult, often referred to as a sentinel event, as may occur during cord prolapse or uterine rupture, is likely to damage the basal ganglia and brainstem, in contrast to the more chronic insult, which leads to damage in the cerebrum (see Neuropathology).74 Preconditioning describes reduced sensitivity of the immature brain to injury depending on whether it has been exposed to previous nondamaging hypoxic events some hours before the main hypoxic-ischemic insult. Preconditioning of immature rat pups by exposure to moderate hypoxia or inflammation before hypoxia-ischemia appears to be neuroprotective.64 Preconditioning may work through stabilizing hypoxia-inducible factor-1α (HIF-1α) during hypoxia. When dimerized with HIF-1β to HIF-1, it acts on hypoxia response elements in the promoter of hypoxia-responsive genes, which will lead to induction of genes encoding erythropoiesis, angiopoiesis and antiapoptosis.31 During the last decades, the processes leading to neuronal death in the neonate have been described in more detail. This knowledge is important when considering neuroprotective strategies (see Specific Neuroprotective Strategies). In contrast to adult ischemic stroke, neonatal hypoxia-ischemia is characterized in most cases by a combination of cerebral hypoxia (and ischemia during bradycardia), followed by reperfusion and potential excessive distribution of oxygen. The contribution of reperfusion to cerebral injury is recognized, and has led to the reduced use of supplemental oxygen during neonatal resuscitation.116 An acute hypoxic-ischemic insult leads to events that can be broadly categorized as early (primary) and delayed (secondary) neuronal death. Previously, two different patterns of cell death are reported in neonates.79 Necrosis is the lytic destruction of cells, whereas apoptosis is the programmed cell death, which is ATP-driven. More recently, these cell death patterns appear to be in continuum. The final pathway is dependent on tissue circumstances such as oxygen content. It has been demonstrated in experimental settings that secondary energy failure starts within 6 to 8 hours after the primary insult. The term energy failure reflects the fact that high energy phosphates are reduced as can be demonstrated in vivo using phosphorus magnetic resonance spectroscopy.59 This process is a result of changes in mitochondria, and may last up to 72 hours or even longer after the acute insult. Timing of hypothermia for neuroprotection is based on this “therapeutic window” of 6 hours before the onset of secondary energy failure. Excessive neuroexcitatory activity occurs as a result of the asphyxial event, and this is mediated through glutamate toxicity.57 Glutamate activates N-methyl-d-aspartate (NMDA) receptors, which in turn cause calcium channels to open in an unregulated manner with excess entry of intracellular Ca2+. The high concentrations of this ion activate lipases, proteases, endonucleases, and phospholipase C, which in turn break down organelle membranes. This sets up a variety of abnormal processes with release of free radicals, including NO• and superoxide ions. This has further adverse effects on cell membranes and leads to mitochondrial failure with the release of caspase-3 and eventual DNA fragmentation, poly(ADP-ribose) polymerase, which causes further energy failure of intracellular membrane function. This process also triggers an apoptotic response in the cell. The process of cell death is different from that seen in adults,122 requiring specific neonatal animal models in the research of perinatal asphyxia. Furthermore, cell death pathways differ between male and female rat pups (Figure 61-4).123 Some compounds like erythropoietin appear to be more protective in female than in male pups. The relevance for human neonates is not yet established. Oxygen free radicals cause peroxidation of unsaturated fatty acids, and because the brain is especially rich in polyunsaturated phospholipids, it is especially susceptible to free radical attack. Mechanisms for quenching and inhibiting free radical production exist within the brain, but in the immature organ these mechanisms may be underdeveloped. Consequently the human neonatal brain is at particular risk for oxygen free radical–induced injury.12 Brain arteriole endothelium is the main source of free radical production by the action of xanthine oxidase, but free radicals are also produced by activated neutrophils, microglia, and intraneuronal structures. During reperfusion, free radical production from the arteriolar endothelium results in blood-brain barrier leakage and release of platelet-activating factor, platelet adhesion, and neutrophil accumulation, which may contribute to cellular damage. Resuscitation of human neonates with 100% oxygen has led to prolonged changes in oxidized glutathione as a result of excessive production of oxygen free radicals.116 Apoptosis (see Chapter 58), or programmed cell death, is perhaps the most important cause of neuronal death in the neonate following hypoxia-ischemia and resuscitation. It can be distinguished histologically from necrosis by shrinkage of affected cells with retention of the cell membrane. By contrast, necrosis is associated with cell rupture, which induces secondary inflammatory processes. DNA degradation develops in the apoptotic cell, giving a characteristic ladder appearance on gel electrophoresis. Apoptotic pathways have been shown to differ between males and females. Apoptosis via an apoptosis inducing factor-dependent pathway was demonstrated in cultured XY neurons, whereas a cytochrome c-dependent pathway was seen in XX neurons.27,78,123 Studies in human neonates have confirmed the presence of secondary energy failure after perinatal asphyxia. 31P magnetic resonance spectroscopy (MRS) spectra in affected term infants were usually normal within the first 6 hours after birth, suggesting that mitochondrial phosphorylation had initially recovered with resuscitation. After 8 hours there was a significant decline in the high energy phosphates such as phosphocreatine and ATP, with a further decline in the most severely affected infants at 48 to 72 hours. In some infants, recovery occurred to normal values within 7 days. The delayed fall in phosphocreatine/inorganic phosphate (PCr/Pi) ratio represents secondary energy failure.45 In addition, proton MRS did not show lactate in the neonatal brain on very early scans, whereas high levels of lactate could be demonstrated after 48 hours.52 Studies using near-infrared spectrophotometry have shown reduced oxygen uptake, and higher brain oxygen saturations in neonates with perinatal asphyxia, suggesting secondary energy failure.53 Doppler studies of major intracranial arteries demonstrated loss of normal CO2 reactivity with high diastolic blood flow, first seen 12 to 24 hours after birth. In the most severe cases, overall cerebral blood flow will be reduced.84 Hypoxic-ischemic insults before 20 weeks’ gestational age, such as may occur as the result of severe maternal illness, may lead to neuronal heterotopia or polymicrogyria because the insult to the fetal brain occurs during the stage of neuronal migration, which is not complete until 21 weeks of gestation. Insults affecting the brain during midgestation (26 to 36 weeks) predominantly damage white matter, leading to cystic periventricular leukomalacia, and may have secondary negative effects on growth of the deep gray matter or may increase the risk of intracranial hemorrhage. Primary hypoxic-ischemic injury to the basal ganglia and thalamus in preterm infants has been reported and resulted in a poor outcome.27a,58a Insults near or at term (35 weeks and beyond) result predominantly in damage to deep gray matter or watershed areas of the brain.100 In the term brain, a cortical watershed area (the parasagittal region) is present between the three main arteries supplying each hemisphere. Volpe and colleagues used positron emission tomography to measure regional cerebral blood flow in 17 full-term asphyxiated newborns, and found a consistent decrease in blood flow to the parasagittal region of both cerebral hemispheres in most of the infants.118 Modern techniques such as diffusion-weighted MRI have demonstrated these watershed-type lesions in the full-term neonate (see Neuroimaging). The depths of the sulci are also sensitive to hypoxic-ischemic insult as the result of being watershed areas at term. Reduction in perfusion in small vessels at the base of sulci during hypoxia-ischemia leads to columnar necrosis and may lead to ulegyria in the older child’s brain.100 Ischemic infarction (stroke) has been recognized more commonly with the increased use of MRI in neonates with or even without encephalopathy. Although infarction of a major cerebral artery has been found in neonates with a hypoxic-ischemic insult, thrombosis and embolus, vasospasm, maternal smoking, or hypoglycemia are more common etiologic factors.40,92 The location of the lesions and neurodevelopmental outcome is dependent on the cerebral artery involved. Animal studies on primates have modeled two different patterns of hypoxic-ischemic insult and have shown different patterns of neuronal injury depending on the type of insult.74 Acute total asphyxia was produced in fetal monkeys by clamping the umbilical cord and preventing the animal from breathing. This produced injury to the thalamus, brainstem, and spinal cord structures. The longer the duration of the acute insult, the more extensive the damage was in these regions. Little damage was reported in higher structures. The second model attempted to mimic a prolonged, partial asphyxial insult lasting 1 to 5 hours. This produced damage predominantly in the cerebral hemispheres and particularly in the watershed distribution (see Vascular Territories), often with sparing of brainstem, hippocampus, and temporal and occipital lobes. Damage was also seen not uncommonly in the basal ganglia and the cerebellum. There has been much debate as to the role of pre-existing antenatal factors in exacerbating intrapartum HIE. Recently, placental changes were seen in asphyxiated full-term neonates with encephalopathy. Chronic villitis was associated with the basal ganglia/thalamus pattern of injury, whereas decreased placental maturation was linked to white matter and watershed injury.42 Polymorphisms of the MTHFR gene were demonstrated more often in neonates with HIE and cerebral lesions of the watershed type.39 An MRI study of term infants with hypoxic-ischemic encephalopathy has suggested that the majority of the brain damage present in these babies occurred in the immediate perinatal period and was not the result of long-standing damage,19 although this finding remains controversial. The neuropathologic features of hypoxic-ischemic injury vary depending on the type and duration of the insult and the gestational age of the child at the time of the insult. The pathologic features evident on examination also depend on the interval between the insult and death. The neuropathologic features become more apparent as they develop in a temporal sequence. The following represent some of the more consistent features of hypoxic-ischemic injury. Of course, findings on neuropathology are biased by the fact that these neonates have died. Thus, these findings will show different patterns when compared to imaging studies, for example MRI findings, in a cohort of neonates who survived after hypoxic-ischemic encephalopathy. In addition, MRI is less sensitive in detecting cerebral hypoxic-ischemic changes compared to histology.2 Different cell lines in the brain react in varying ways to an acute hypoxic-ischemic insult. Neurons may undergo either necrosis or apoptosis (see Pathophysiology). The nature of these two mutually exclusive processes can be distinguished to some extent by histologic staining. In necrosis, sections stained with hematoxylin and eosin show changes from 5 to 6 hours after the injury. Nuclear membranes degenerate, with release of nuclear chromatin and a secondary inflammatory reaction.80 Activated microglia express a number of cytotoxic cytokines. Although microglial cells respond rapidly after a hypoxic-ischemic insult and undertake the function of ingesting and lysing dead tissue, CD68-positive cells, indicating activated microglia and macrophages, were seen not earlier than at least 24 hours after the hypoxic-ischemic insult in a cohort of 22 full-term neonates who died after severe perinatal asphyxia.35 In this paper, caspase activity could be seen in the brain examined within 24 hours of the hypoxic-ischemic insult. In addition, nitrotyrosine staining, considered the product of NO toxicity, was present in brain and spinal cord tissue.37 Glial cells respond to hypoxic-ischemic insult by enlarging, proliferating, and later developing fibrillary processes with the expression of glial fibrillary acidic protein, which can be recognized by specific staining. A glial response occurs from 17 weeks of gestational age.100 Ulegyria refers to a particular form of pathology seen in the depths of the cortical sulci and is probably caused by particular watershed vulnerability (see Vascular Territories). The chronic phase of this process produces the appearance of mushroom-like gyri because of loss of deep gray matter in the sulci.100 Previous studies have described the clinical changes in the encephalopathic term neonate in detail.93,104 The Sarnat classification is based on clinical and electroencephalography (EEG) findings 24 hours after the insult. It has been modified to be used shortly after the insult for selecting neonates to determine eligibility for therapeutic hypothermia. The Thompson score is based on several clinical items and results in a quantitative score (see Table 61-2). The classification systems have been used not only for clinical purposes but also for inclusion of patients in clinical trials, as well as for stratification of patients within those trials.47 Although the predictive value for long-term neurodevelopment of mild and severe encephalopathy is good (no handicaps following mild encephalopathy, almost invariably poor outcome after severe encephalopathy), the predictive value of moderate encephalopathy is poor.111 Neonatal EEG is a well-recognized method to assess brain integrity after hypoxia-ischemia that results in hypoxic-ischemic encephalopathy. In newborn infants, the EEG commonly uses 16 channels and is in most centers only performed during the day; it requires skilled technicians and highly trained and experienced neurophysiologists to interpret the recordings. The 16-channel EEG provides detailed information and when performed with simultaneous video recording, helps to make a distinction between clinical and subclinical seizures. One study showed that less than 10% of neonatal seizures were correctly identified by the neonatal staff, compared with simultaneous video-EEG recordings (see Chapter 62).73 Normal and severely abnormal results on EEG are of important predictive value. Normal traces almost invariably predict a normal outcome, whereas persistent, severely abnormal traces predict an adverse outcome (Figure 61-5). Prediction in children with mild to moderate EEG abnormalities is less reliable and requires sequential recordings. EEG activity is fully differentiated in full-term infants. During quiet sleep, a discontinuous pattern (tracé alternant) alternates with a high-voltage continuous delta pattern. In wakefulness, the EEG is characterized by low-voltage mixed theta-delta activity. In active sleep, a mixed medium-voltage delta-theta activity is seen. Abnormalities on the EEG can be classified as background abnormalities, ictal abnormalities, and abnormalities in the organization of states and maturation. Background pattern abnormalities are highly predictive of outcome (Figure 61-6). It should be taken into account that antiepileptic drugs can (transiently) affect sleep states and the background pattern, especially in children with severe encephalopathy. The following background patterns can be recognized: isoelectric and extremely low voltage (less than 5 µV); burst suppression pattern, with long periods of inactivity (usually longer than 10 seconds) mixed with bursts of abnormal activity. Outcome is especially poor, with long periods of inactivity and brief periods of bursts (less than 6 seconds) with small-amplitude bursts and interhemispheric asymmetry and asynchrony.8,69
Hypoxic-Ischemic Encephalopathy
Definitions
Hypoxic-Ischemic Encephalopathy
Variable
Stage I
Stage II
Stage III
Level of consciousness
Alert
Lethargy
Coma
Muscle tone
Normal or hypertonia
Hypotonia
Flaccidity
Tendon reflexes
Increased
Increased
Depressed or absent
Myoclonus
Present
Present
Absent
Seizures
Absent
Frequent
Frequent
Complex Reflexes
Suck
Active
Weak
Absent
Moro
Exaggerated
Incomplete
Absent
Grasp
Normal or exaggerated
Exaggerated
Absent
Doll’s eye
Normal
Overactive
Reduced or absent
Autonomic Function
Pupils
Dilated, reactive
Constrictive, reactive
Variable or fixed
Respirations
Regular
Variations in rate and depth, periodic
Ataxic, apneic
Heart rate
Normal or tachycardia
Bradycardia
Bradycardia
Electroencephalogram
Normal
Low voltage, periodic paroxysmal
Periodic or isoelectric
Sign
Score
Day 1
Day 2
Day 3
0
1
2
3
Total score per day
Pathophysiology
Systemic Adaptation to Hypoxic-Ischemic Insult
Preconditioning
Perinatal Hypoxic-Ischemic Brain Damage
Glutamate Injury (Excitotoxicity)
Free Radical Formation
Apoptosis
Human Studies
Selective Vulnerability
Maturity
Vascular Territories
Types of Hypoxic-Ischemic Insult
Others
Neuropathology
Cellular Responses
Chronic Lesions
Assessment Tools
Clinical Examination
Neurophysiology
EEG
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree