Hypoxic Ischemic Encephalopathy—What Can We Learn from Humans?

Abstract

Hypoxic ischemic encephalopathy (HIE) is a condition that occurs in both human newborns and foals. The condition is the subject of extensive current research in human infants, but there have been no direct studies of HIE in foals, and hence, knowledge of the condition has been extrapolated from studies in humans and other animal models. The purpose of this review article is to highlight the most up-to-date and relevant research in the human field, and discuss how this potentially might have an impact in the management of foals with HIE.

Abbreviations

ADC

apparent diffusion coefficient

aEEG

amplitude-integrated encephalography

ASL

arterial spin-labeling

cEEG

conventional electroencephalography

DWI

diffusion-weighted imaging

EPO

erythropoietin

GABA

gamma amino butyric acid

HIE

hypoxic ischemic encephalopathy

MDI

mental developmental index

MRI

magnetic resonance imaging

MRS

magnetic resonance spectroscopy

NE

neonatal encephalopathy

NMS

neonatal maladjustment syndrome

PAS

perinatal asphyxia syndrome

PLIC

posterior limb of the internal capsule

RNS

reactive nitrogen species

ROS

reactive oxygen species

VEGF

vascular endothelial growth factor

Hypoxic ischemic encephalopathy (HIE) is a condition in foals that is known by many names, including neonatal maladjustment syndrome (NMS) and perinatal asphyxia syndrome (PAS). The incidence in foals is reported to be 1–2% of all births.[1] There have been no direct studies of HIE in foals, and hence, knowledge of the condition has been extrapolated from studies in humans and other animal models with attendant limitations. Because the basic pathophysiology of brain injury in many species shares common features, the purpose of this review is to summarize the pathophysiology, diagnosis, and treatment of HIE in human neonates. Current protocol in the human neonatal intensive care unit (NICU) and ongoing research in this area have raised important points for discussion regarding our approach to and management of the condition in foals.

Incidence and Consequences

HIE affects 3–5 per 1000 human live births, with moderate or severe hypoxia affecting 0.5–1.0 per 1000 live births.[2] HIE has very important sequelae with 10–60% of affected infants dying in the newborn period and with at least 25% of survivors having important long-term neurodevelopmental impairment.[3] The most important long-term complication is development of cerebral palsy, which is defined as a group of permanent motor disorders that are attributed to nonprogressive disturbances which occur in the developing fetal or infant brain.[4]

Etiology

The origin of hypoxic ischemic brain injury results from a reduction in cerebral blood flow and oxygenation during the antepartum, peripartum, or postnatal period. Evidence suggests that approximately 70–80% of cerebral palsy cases are the result of antepartum injury, with birth asphyxia responsible for only 20%.[5] Hypoxia can develop during labor because of compression of the umbilical cord, insufficient uteroplacental circulation, cord prolapse, uterine rupture, shoulder dystocia, or vaginal breech delivery.[6] Most preventative measures and therapies target the birth asphyxia group, as there is a limited time period after insult to prevent or reduce injury, and often with antepartum injury that period has passed by the time the infant is born.[7] Risk factors for the development of hypoxic brain injury in foals include prolonged dystocia, premature placental separation (red bag delivery), and need for resuscitation after cesarean section.[8] Recently, the term neonatal encephalopathy (NE) has been used to encompass all neonatal foals that exhibit neurological abnormalities. It is important to note that HIE is a specific type of NE, and not all NE cases are caused by hypoxia ischemia. An example of NE without true hypoxic brain injury is evident in human infants born to mothers experiencing pyrexia as a result of infection. This is thought to lead to cerebral damage through the production of inflammatory mediators and can also make the infant potentially more vulnerable to an adverse intrapartum incident leading to HIE.[5] This is an important point to consider in equine medicine, as often foals born to mares with severe placentitis exhibit neurological abnormalities, without experiencing an obvious hypoxic incident. For this reason, it is important not to label all neurologically abnormal foals as having HIE, and to emphasize the importance of a thorough history and complete diagnostic workup for each case.

Clinical Findings

Clinical findings in a newborn infant suggesting hypoxic ischemic injury include evidence of a sentinel event during labor, such as fetal heart rate abnormality, a severely depressed infant with a low extended Apgar score, the need for resuscitation in the delivery room, and evidence of severe fetal acidemia. The definition of severe acidemia in human infants is a cord umbilical artery pH < 7, base deficit >/= 16 mEq/L or both.[9]

Because of a high proportion of unattended foalings and less fetal monitoring performed on horses, it is difficult to accurately define a hypoxic ischemic event. This then makes the distinction between true HIE and NE because of another cause unclear, but it does highlight the importance of considering the global picture when presented with a neurologically abnormal foal. This includes obtaining as detailed history of the pregnancy as possible, and close examination of the mare's placenta. To fully assess the risk of hypoxic ischemic brain injury in the foal, we must consider the above parameters described for humans. A modified Apgar score has been developed for foals[8] (see Table 1) and although it is performed as standard in the human NICU, it is often not carried out in the equine clinic.

Table 1. Apgar score: 7–8 = normal foal, 4–6 = requires intervention, 0–3 = life threatening.[8]

Score	0	1	2
Heart rate	Absent	<60 or irregular	>60 and regular
Respiratory rate	Absent	Irregular	Regular
Muscle tone, and postural response	Laterally recumbent, no muscle tone	Some flexion of limbs and muscle tone	Sternal recumbency and attempts to stand
Response to stimuli	No response	Weak ear movement when stimulated, facial grimace	Head shake, sneeze, or cough

In a referral hospital situation, it is standard practice to monitor arterial blood gas and this should be performed as soon after birth as possible to enable accurate prediction of those foals that are likely to have experienced a hypoxic ischemic insult. Normal blood gas values for term foals have been published.[10] Umbilical arterial blood gas values immediately after birth are pH 7.323 ± 0.014, base excess 0.2 ± 1.2 mmol/L, PaO₂ 43.2 ± 3.9 mmHg, and PaCO₂ 53.0 ± 1.8 mmHg.

Pathophysiology

At the most simplistic level, a reduction in cerebral blood flow and oxygen delivery initiates a cascade of deleterious biochemical events. The lack of oxygen causes a switch to energy inefficient anaerobic metabolism, leading to depletion of high energy phosphate reserves, lactate accumulation, and an inability to maintain cellular homeostasis. This results in failure of critical transcellular ion pumps, which causes intracellular accumulation of water (cytotoxic edema) and calcium. The increase in calcium stimulates the release of, and also inhibits the reuptake of, excitatory amino acids such as glutamate. Glutamate is a potent, rapidly acting neurotransmitter which under normal physiological conditions is rapidly removed from the extracellular space by powerful neuronal and glial uptake mechanisms.[11] In the immature brain, glutamate is an important trophic factor that mediates normal brain development and plasticity.[12] These properties of the immature brain contribute significantly to its increased vulnerability to excitotoxic cell death, and this excitotoxicity is central to brain injury.[13]

Free radicals are highly reactive compounds with an uneven number of electrons in the outermost orbital. They can react with certain normal cellular components, such as the unsaturated fatty acids of membrane lipids, to generate further free radicals leading to a chain reaction and irreversible biochemical injury. In less intense insults, free radicals can lead to apoptotic cell death by activation of specific cell death genes. There are various sources of free radicals after hypoxic ischemia. Most importantly, oxygen deprivation prevents the complete passage of electrons to cytochrome c oxidase within the electron transport system, leading to the generation of free radicals. Other sources are related to increased cytosolic calcium concentrations, which include phospholipase A acting on arachidonic acids, xanthine oxidase, and nitric oxide (NO) synthase. Microglia are also potent sources of free radicals.[14] The most commonly generated free radicals are the reactive oxygen species (ROS) that include the superoxide anion, which is converted to hydrogen peroxide by superoxide dismutase. Hydrogen peroxide is detoxified by antioxidant systems catalase and glutathione peroxidase, unless free iron is available. Under hypoxic conditions, the nontoxic ferric state of free iron is converted to the dangerous ferrous form, and in this environment, superoxide radicals react with hydrogen peroxide producing toxic hydroxyl radicals.[15]

As well as ROS, reactive nitrogen species (RNS) incite free radical-mediated damage as a result of HI injury. There are 3 forms of nitric oxide synthase (NOS), a constitutive neuronal form (nNOS), a constitutive endothelial form (eNOS), and an inducible form (iNOS) found in astrocytes and microglia. Both constitutive forms are activated by calcium, whereas iNOS is stimulated by products of inflammatory processes, such as cytokines and lipopolysaccharide (LPS).[16] Under pathologic conditions of hypoxia-ischemia-reperfusion, an increase in cytosolic calcium via NMDA receptors results in increased activity of eNOS leading to vasodilatation. Increases in nNOS result in generation of NO, which can diffuse to adjacent neurons, and under conditions of oxidative stress with abundant superoxide anion, react to form the particularly toxic RNS, peroxynitrite.[15]

Microglia are the resident immune cells of the brain and are believed to originate from hematogenous monocytes, which invade the fetal brain during development.[17] These cells are abundant within the brain accounting for up to 10% of all cell types in both the gray and white matter. The microglia exist in 2 states, first, a resting or ramified state in which the cells are characterized by a small cell body and multiple processes. Under pathologic conditions, the microglia are rapidly activated and expand in response to an injurious stimulus. Activated microglia, also known as amoeboid-globoid microglia, develop a more rounded appearance and the processes are retracted.[18] Activated microglia are the predominant cell type producing inflammation-mediated neurodegeneration.[19] Activation of microglia is believed to play a central role in neonatal hypoxic ischemic injury.[13] Evidence for the central role of activated microglia in neonatal hypoxic ischemic damage is supported by the finding that minocycline, a drug that specifically inhibits microglia, and significantly reduces brain damage in response to injury in animal models.[20]

Based on the multiple pathways involved in the pathogenesis of hypoxic ischemic injury, it is clear that there are many intervention points for possible therapies. Unfortunately, however, there is often failure of promising treatments in animal models of disease to translate to human patients. Reasons for this include differing dose rates, routes of administration, and the disease model not truly reflecting what happens in the patient. As a result, there is a lack of scientifically proven treatments available for the human neonate suffering from this devastating condition. The situation in equine medicine is similar, with a range of treatments on offer in the management of HIE, but all without proven benefit.

Diagnosis

Imaging

Magnetic resonance imaging (MRI) is the imaging modality that provides most information in the diagnosis of HIE in the newborn.[14] Conventional MRI includes T1- and T2-weighted sequences, and these provide the most accurate anatomical detail currently available in neuroimaging. However, when injury occurs, these conventional sequences may remain normal for several days. In the very acute phase, it can be very difficult to detect abnormalities on either conventional T1-weighted or T2-weighted images. Other sequences such as diffusion-weighted imaging (DWI) are more useful in the early detection of perinatal brain ischemia. DWI is based on the molecular diffusion of water, and injury is often demonstrable in the first 24–48 hours after birth as decreased diffusion (increased signal) in injured areas. In conventional MRI, contrast derives from differences in water MR relaxation properties, whereas in DWI, the contrast is based on differences in the translational motion of water molecules. This motion is described quantitatively as a water apparent diffusion coefficient (ADC). With acute cellular injury, cytotoxic edema develops and the cell membrane restricts water movement when a diffusion gradient is applied, resulting in a reduced ADC value.[21] The ADC of brain water declines within minutes after the onset of cerebral ischemia, giving rise to DWI hyperintensity with little or no change in the corresponding T2-weighted MRI. By 5–10 days after injury, the ADC value returns to normal (pseudo normalization) and can then increase because of ongoing cell death. After this, the ADC value returns to normal values at 5–10 days (pseudo normalization) before increasing further with cell death associated with a highly diffusing extracellular space around cellular loss. ADC mapping with ADC values calculated from the DWI sequence shows restricted diffusion as dark areas of diminished signal intensity.

The pattern of injury seen on MRI in human infants is used to help predict long-term neurodevelopmental outcome.[22] It has been demonstrated that abnormalities within the basal ganglia and thalamus lead to cerebral palsy, with the severity of the lesion being proportional to the degree of motor impairment. Infants suffering with HIE that have lesions within the brainstem are at increased risk of mortality.[22] It has been demonstrated that the ADC value in the posterior limb of the internal capsule (PLIC) is useful as an objective prognostic marker for infants with HIE and is correlated with clinical outcome (Fig 1).[23]

Figure 1. The image shows (a) diffusion-weighted imaging performed on day 2 of life in a child showing restricted diffusion affecting all regions of the brain, (b) T2-weighted imaging performed on day 9 of life showing evolving injury in areas originally identified in (a) including the basal ganglia with global loss of gray–white differentiation at cortical margin, and (c) T2-weighted imaging performed at 9 months of age, showing end stage, encephaloclastic dissolution of the brain associated with severe motor and mental disability.

To date, there are no published reports of MRI findings in foals with HIE. One study reported the MRI findings of 12 horses exhibiting neurological signs, one of which was a 2-day-old foal with a suspected diagnosis of HIE. The scan of that foal, however, showed the presence of severe hydrocephalus in emphasizing the importance of not using HIE as an all encompassing term for all foals with neurological disease.[24] There is a technique for successfully imaging the brain of normal neonatal foals and determined normal brain anatomy.[25] MRI is becoming much more common in referral hospitals, and imaging of foals with neurological signs caused by presumed HIE will be more accessible; our ability to diagnose the disease and monitor its progression will increase markedly.

Two other magnetic resonance sequences are utilized in the investigation of HIE in people. First, magnetic resonance spectroscopy (MRS) provides a biochemical snapshot of a selected region of the brain. The 2 types of MRS are phosphorous (31P) and proton (1H) MRS. The 31P nucleus is used for the study of cerebral oxidative phosphorylation and adenosine triphosphate (ATP) concentrations, thus enabling studies of cerebral metabolism. Normally, the neonatal phosphorus MR spectrum shows the major phosphorus-containing constituents important for energy metabolism (the alpha, beta, and gamma forms of ATP, phosphocreatine (PCr), and inorganic phosphate (Pi)).[14] The 1H MRS can provide information on neuronal/axonal viability, cellular energetic, and cellular membrane status. It provides information about the concentrations of many neurochemicals, such as N-acetyl aspartate (NAA), creatine (Cr) and PCr, choline (Cho), and lactate (Lac). NAA resonance is widely regarded as a marker of neuronal integrity, with reduced NAA indicating a reduction in neuronal/axonal number.[26]

Second, arterial spin-labeling (ASL) provides an imaging tool for signal within blood vessels, and so determines perfusion. This sequence is useful to determine if perfusion of some regions of the brain is differentially affected in certain clinical circumstances.

Electroencephalography

MRI provides information about the appearance of structures of the injured brain; however, it is also necessary to have an understanding of cerebral function when diagnosing and prognosticating in HIE. Traditionally, conventional electroencephalography (cEEG) has been used for intermittent evaluation of functional brain status. However, cEEG is recorded on a brief and intermittent basis, and experienced neurophysiologists are required for interpretation. Amplitude-integrated encephalography (aEEG) has become popular as a tool in the human NICU that is constantly available, readily applied, and easy to interpret for the continuous assessment of brain function.

All EEGs measure the difference in electrical potential between 2 scalp locations.[14, 27] aEEG is an increasingly used method for the continuous monitoring of cerebral electrical activity in critically ill newborns.[14] It usually involves a single channel recording obtained from 1 pair of biparietal electrodes. The EEG signal is amplified and is then passed through a filter that attenuates activity lower than 2 Hz and higher than 15 Hz to minimize artifacts. The signal is processed further by amplitude and time compression, rectifying and smoothing before recording on a semilogarithmic scale at relatively slow speed, usually 6 cm/hour. The bandwidth reflects variations in minimum and maximum EEG amplitude. Advantages include ease of application, ability to monitor continuously, and capacity to detect seizures, relatively severe encephalopathy, effects of drugs and outcome. The recordings are evaluated especially for background and seizure activity. The major background patterns identified are termed continuous normal voltage, discontinuous normal voltage, burst suppression, continuous low voltage, and flat line. The latter three are associated with severe brain pathology.[14] Seizure activity is characterized as a rapid rise in both the lower and upper margins of the trace.[14] The raw EEG trace should show simultaneous seizure activity with a gradual build up followed by reduction in both frequency and amplitude in a repetitive spike pattern lasting at least 10 seconds.

The aEEG has proven to be sensitive for early prediction of outcome in HIE in human newborn infants. A continuous or slightly discontinuous aEEG pattern during the first 6 hours is associated with a high chance of cerebral recovery and normal outcome.[28] Because aEEG provides continual assessment of cerebral function, one of its main uses is to highlight the presence of clinically silent electrographic seizures, which despite an absence of clinical signs of seizure activity can induce further brain injury.

Although aEEG has not been utilized in foals, it has been applied to a piglet model of HIE and is increasingly being used in the management of HIE in newborn infants (Fig 2).[29] The tool is easily applied with 5 small needle electrodes placed subcutaneously (SC) around the animal's head. The tracing provides useful information regarding cerebral activity and can detect clinically silent electrographic seizures. This is very important as moderate to severe HIE is the most common underlying cause of neonatal seizures, and both clinical and experimental studies indicate that they can lead to long-term impairments in brain development.[30, 31]

Figure 2. An example of an amplitude-integrated encephalography (aEEG) tracing in a piglet with experimentally induced HI brain injury. The upper 2 traces show the raw EEG trace in the left (top) and right (lower) hemispheres, and the lower 2 traces display the aEEG tracing. Electrographic seizure activity is prominent in all traces; however, no clinical seizures were noted during this time. (1) Repetitive sawtooth pattern on raw EEG trace indicating seizure activity. (2) Arrow indicates abrupt rise in the amplitude of the trace in association with seizure activity.

Clinical seizure activity in foals can range from very subtle signs such as lip smacking to tonic clonic seizure activity. An important question that needs to be answered is whether foals, like babies, have suspected clinical seizures or abnormal movements which do not have electrographic correlates, and conversely have electrographic seizures with no clinical correlate. Application of the aEEG to foal medicine could potentially help to answer this question and increase our understanding of seizure pattern and response to treatment in these patients. It is important to be able to accurately determine seizure activity, because clinical neonatal seizures in the setting of birth asphyxia are associated with a worse neurodevelopmental outcome independent of the severity of the hypoxic ischemic brain injury.[32]

Neuropathology

The neuropathological features of neonatal HIE vary with the gestational age of the infant, the nature of the insult, the types of interventions, and other factors. There are 4 major categories of neuropathology seen in cases of HIE: selective neuronal necrosis, parasagittal cerebral injury, periventricular leukomalacia, and focal ischemic brain necrosis (stroke). In term infants with HIE, the major features are selective neuronal necrosis and parasagittal cerebral injury, whereas periventricular leukomalacia is the most predominant lesion in premature infants.

The major sites of neuronal necrosis are the cerebral neocortex, hippocampus, deep nuclear gray structures (thalamus and basal ganglia, especially the putamen), brainstem, cerebellum, and spinal cord. This combination of putaminal and thalamic neuronal injury is typical of neonatal hypoxic ischemic disease in the term infant. Within the cerebral cortex, the neurons in the depths of the sulci are most susceptible to injury, and similarly, the Purkinje cells of the cerebellum. There are several factors that confer region-specific vulnerability to neuronal damage. Vascular factors play a role because neuronal injury is more marked in vascular border zones (depths of sulci and parasagittal cerebral cortex).[14] Regional metabolic factors also play a role, but the most important underlying cause has been shown to be the result of distribution of glutamate receptors on neurons.[14]

There are few reports of the neuropathologic change in foals that have died as a result of HIE, and there is a widely held belief that there are no or minimal pathologic changes associated with this condition. Pathologic findings from 18 foals diagnosed with NMS report necrosis of the cerebral cortex and local hemorrhage in 9/18 foals.[33] A recent article investigating biomarkers of brain injury in foals with HIE described the pathologic findings of 3 foals that had performed histopathology. There was neuronal necrosis of the gray matter of the cerebral cortex, caudate nuclei, thalamus, hippocampus, cerebellar cortex, and medulla oblongata, which was consistent with a diagnosis of hypoxic ischemic brain injury.[34]

Treatment

Hypothermia

Therapeutic hypothermia is recognized as an effective cytoprotectant and has been used in many medical situations, such as organ transplantation, cardiopulmonary bypass, spinal cord injury, and neonatal hypoxia ischemia.[35] Animals that hibernate during the winter provide an example of how a reduction in body temperature and metabolism can aid survival during adverse conditions. Hypothermia can be classified based on the depth of cooling from a normal body temperature of 37–38°C. Mild hypothermia describes a body temperature in the range 32–35°C, moderate hypothermia (28–32°C), and deep hypothermia (<28°C). In general, mild to moderate degrees of hypothermia are used clinically because of a similar neuroprotective effect alongside reduced risks of medical complications, such as infection, arrhythmia, hypokalemia, coagulopathies, and bradycardia. Recent research has shown that hypothermia does not just act to slow metabolism, but rather can have an effect on a wide range of cell death and cell survival pathways leading to inhibition of apoptosis and inflammation, and stimulation of pro-survival pathways. There are many proposed mechanisms of action by which hypothermia offers neuroprotection in HIE. It has been proposed that cooling suppresses many of the pathways leading to delayed cell death, and may modify pathways in cells programmed for apoptosis, resulting in their survival. A reduction in body temperature leads to a reduction in cellular metabolic demands, and it has been reported that for every degree Celsius of temperature reduction, the cerebral metabolism is reduced by 5%.[36] Hypothermia might also protect the neurons by attenuating the release of excitatory amino acids, ameliorating the ischemia-impaired uptake of glutamate, and by lowering production of toxic nitric oxide and free radicals.[37]

Hypothermia is now the major treatment modality in the management of neonatal HIE in humans and involves reduction in temperature to 33°C initiated within 6 hours of delivery, which is continued for a total of 72 hours, followed by a gradual rewarming to normothermia over 12 hours. Therapeutic hypothermia aims to lower the temperature of the vulnerable deep brain structures, the basal ganglia, to 32–34°C. There are 2 methods for achieving this, whole body cooling and selective head cooling with mild systemic hypothermia. The rationale for selective head cooling is that the newborn infant's brain produces 70% of total body heat and that systemic hypothermia may be physiologically harmful to the sick neonate. However, a theoretical modeling of cooling, investigating temperature distribution within the neonatal head, found that the only situation that resulted in a significant reduction in deep brain temperature was when the core temperature was lowered to 34°.[38]

There have been a number of randomized controlled trials that report improved outcomes for cooled infants with HIE compared to patients maintained at normothermia.[39] The most recent meta-analysis to date[40] evaluated a total of 13 clinical trials of therapeutic hypothermia, consisting of a total of 1,440 patients. The results showed a significant reduction in the risk of mortality or of moderate to severe neurodevelopmental disability in infants who received hypothermia compared to control which was used as the primary outcome. The secondary outcomes assessed were efficacy and safety. There was a significant reduction in the risk of cerebral palsy, severe visual deficit, cognitive delay or Mental Developmental Index (MDI) <70, and psychomotor developmental index <70 in the hypothermia groups compared to the control group. There was no difference in the risk of epilepsy or severe hearing deficit, and there was no difference in the relative risk of withdrawal of life sustaining medical treatment between the hypothermia and control group. There was an increased risk of arrhythmia and thrombocytopenia reported in the hypothermia group, but neither resulted in any clinically significant impact. No other significant differences between the groups were observed for other adverse outcomes, including organ dysfunction. The overall conclusions from this meta-analysis of 13 trials confirm that therapeutic hypothermia is associated with a reduced risk of the combined outcome of mortality or moderate to severe neurodevelopmental disability in infancy or childhood. Hypothermia was also effective in reducing individual outcomes or mortality, moderate to severe neurodevelopmental disability, severe cerebral palsy, cognitive delay, and psychomotor delay.

Therapeutic hypothermia is the only treatment that has been shown to improve neurological outcome in HIE in children, but it does not provide complete protection and there is evidence that it is not as successful in the treatment of the most severely affected neonates.[31] Additional treatments can be administered during or after hypothermia that can extend the therapeutic window or provide an additive or synergistic effect are desperately needed, and several research groups are currently working on animal models to develop further therapies, some of which are described later in this review.

To the authors’ knowledge, there is only 1 published report of the use of therapeutic hypothermia in veterinary species.[41] In this article, the author describes successful management of intractable seizures in a dog suffering from traumatic brain injury with a combination of high dose barbiturate therapy and controlled hypothermia. Several studies have shown that critically ill foals which are hypothermic have a worse prognosis for survival.[42] In 1 study,[43] the odds ratio for survival for foals with a subnormal temperature was 0.19, with a mean temperature of survivors of 38.3 ± 1.8°C versus a mean of 37.5 ± 1.7°C in the nonsurvivors. Based on this information, it would seem counterintuitive to suggest the use of therapeutic hypothermia in the equine NICU. However, it is important to distinguish therapeutic hypothermia and its benefits for brain injury, and hypothermia as a result of critical illness. One study investigated the risk of death or disability in term infants with HIE, which had increased temperature between 6 and 78 hours after birth. In this report, the controls were those that received “usual care” and were not cooled, and this group was compared to cooled infants. The authors reported that the odds of disability or death were increased by 3.6- to 4-fold for each 1°C increase in skin or esophageal temperature.[44] Once we start to consider utilizing this treatment modality in foals, it becomes very clear how important our assessment and diagnostic workup is, as there are specific indications for the use of therapeutic hypothermia and it is certainly not to be used as a blanket treatment for any foal which is admitted to the NICU. We cannot ignore the results in both experimental animal models and human clinical trials, and we have to seriously contemplate that by keeping foals that have HI injury warm, we may be inadvertently worsening their prognosis.

Other Therapies

Animal studies are critical to improving the care of the human fetus and newborn. The goals of animal models are to increase our understanding and knowledge of mechanisms of injury, evolution of injury, and to provide a clinically relevant template on which to develop and test new therapies. To study HIE in the term infant, the models used should, first, closely mimic the etiology of injury, express the functional outcomes seen in human infants, and accurately reflect histopathologic injury. There is, however, quite frequent failure of seemingly successful treatments in animal models of disease, to translate to human patients. Reasons for this include differing dose rates, routes of administration, and the disease model not truly reflecting what happens in the patient. As a result of this, despite all of the advances, adaptation of neuroprotective drugs to the clinical setting has been largely unsuccessful. At present, the question of how these models relate to foals remains largely unanswered, and is an area that needs further research. There are numerous therapies that have been used in the past to treat foals with HIE, and many are still used today, including dimethylsulfoxide (DMSO), thiamine, allopurinol, magnesium sulfate, mannitol, theophylline, ascorbic acid, naloxone, and alpha-tocopherol, despite a lack of supporting scientific evidence. There are several research groups across the world whose focus of attention lies with developing new therapies for the treatment of HIE in human infants to add to cooling, and the most relevant of these treatments are summarized below.

Desferrioxamine

As described earlier, during hypoxic ischemia, free radicals are generated and the neonatal brain is particularly vulnerable to this damage because of its high concentration of lipids, high rate of oxygen consumption, altered balance of antioxidants, and increased availability of nonprotein bound free iron. A potential treatment strategy for protection against HI injury is to inhibit free radical production. The high affinity iron chelator desferrioxamine forms a complex with iron so that it cannot be reduced to the ferrous form, thus inhibiting free radical production.[45] Experimental models have documented a reduction in brain injury in response to HI injury with desferrioxamine treatment. Desferrioxamine administered after induction of cerebral hypoxia ischemia, reduced injury in 7-day-old rats,[46] and was found to significantly reduce hypoxia-ischemia-induced nonprotein bound iron formation in newborn lambs.[47] More recently, Papazisis et al[48] found that in newborn rats, desferrioxamine administered SC immediately postinsult had a neuroprotective effect on the hippocampus, and it decreased asphyxia-induced brain tissue concentrations of the excitotoxins glutamate and aspartate.

Erythropoietin

Erythropoietin (EPO) has also been investigated as a potential neuroprotectant for use in neonatal HI injury.[49] The neuroprotective actions of EPO have been known for approximately 15 years, initially based on the observation that recombinant EPO (rEPO) protected neurons against hypoxia in vitro and that endogenous EPO was produced in brain astrocytes.[50, 51] The proposed mechanisms of neuroprotection include direct neurotrophic effects, decreased susceptibility to glutamate toxicity, induction of anti-apoptotic factors, decreased inflammation, increased blood flow to injured tissue, decreased nitric oxide mediated injury, and antioxidant effects. Recombinant human EPO administration to newborn rats has been shown to promote tissue protection, revascularization, and neurogenesis in neonatal HI injured brain, leading to improved neurobehavioral outcomes.[52] The results from a pilot clinical trial by means of EPO in neonates with HIE have recently been published.[53] A prospective case control study was performed with 45 neonates enrolled. The investigators split the neonates into 3 groups of 15, a normal healthy group, neonates with HIE treated with human rEPO (2500 IU/kg q24h for 5 days), and neonates with HIE who received no EPO treatment. They reported that by 2 weeks of age, EEG tracings had improved and nitric oxide (NO) concentrations had decreased in the EPO group. However, no changes in MRI findings were seen. By 6 months of age, the neonates treated with EPO had fewer neurologic and developmental abnormalities. There are now, however, concerns about the use of high doses of EPO to treat neonates, based on findings in adults where phase II/III trials suggest a higher death rate in patients receiving EPO compared to a placebo, as well as increased risk of serious complications, such as intracerebral hemorrhage and thrombotic events.[54] Additional concerns exist with the use of rEPO in horses because of the development of anti-recombinant human erythropoietin (anti-rhEPO) antibodies resulting in erythroid hypoplasia and severe anemia.[55]

Anticonvulsant Therapy

The major common goal of the pharmacological treatment by means of antiepileptic drugs is to counteract abnormal brain excitability by either decreasing excitatory transmission or enhancing neuronal inhibition.[56] An excessive release of excitatory amino acids and a reduced neuronal inhibition also occur in brain ischemia, and therefore the use of anticonvulsants may also contribute to neuroprotection.[31, 57] There is concern, however, that anticonvulsants which suppress synaptic activity in the brain have been associated with widespread apoptotic neurodegeneration throughout the brain when administered to normal immature rodents during the period of the brain growth spurt.[31, 58] Compounds that may cause neuronal apoptosis in the developing brain include NMDA antagonists (ketamine), agonists of gamma amino butyric acid A (GABA A) receptors (barbiturates and benzodiazepines), and sodium channel blockers (phenytoin).[58] Barks et al investigated the combined effect of hypothermia with phenobarbital in 7-day-old rats with cerebral hypoxia ischemia. The authors reported that early posthypoxia ischemia, prophylactic administration of phenobarbital may augment the neuroprotective efficacy of therapeutic hypothermia, and that hypothermia may limit potential adverse effects of phenobarbital.[57] In 2010, Meyn[59] investigated the use of prophylactic administration of 40 mg/kg phenobarbital to infants with HIE at the commencement of whole body cooling. They reported that although a reduction in clinically detectable seizures was noted, there was no significant improvement in neurodevelopmental outcome.

Melatonin

Melatonin is produced from 1-tryptophan in the pineal gland, retina, and gastrointestinal tract.[60] It is considered to be a natural neuroprotectant and a potent free radical scavenger itself, and also induces the production of other antioxidant enzymes. It has been demonstrated in fetal sheep that melatonin administration after umbilical cord occlusion reduced the number of activated microglial cells and apoptotic cells[61] Melatonin given to mice during the last 7 days of pregnancy before hypoxia ischemia and delivery resulted in a reduction in markers of inflammation and apoptosis.[62]

Xenon

One of the most recent adjunctive therapies currently being investigated is xenon.[31] Xenon is approved for use as a general anesthetic in Europe, and there are encouraging results supporting its role as a potential neuroprotective agent. The major disadvantage of xenon in the clinical setting is the fact that it requires special respirators for administration, and is very expensive. The gas exhibits noncompetitive antagonism at the NMDA subtype of glutamate receptor.[63] It is also thought to work via the activation of antiapoptotic effectors, Bcl-xL and Bcl-2, as well as induced expression of hypoxia inducible factor 1α and its downstream effectors, EPO and vascular endothelial growth factor (VEGF), which can interrupt the apoptotic pathway.[60, 64] Ma et al reported the use of combined xenon and therapeutic hypothermia in neonatal asphyxia. They reported that cultured neurons injured by glucose-oxygen deprivation were protected by combinations of interventions of xenon and hypothermia that, when administered alone, were not efficacious.[63] Further studies have provided more evidence that xenon and hypothermia combine additively to provide neuroprotection in animal models of HIE[64, 65]^.

Conclusions

We have provided a detailed description of the pathogenesis, diagnosis, and treatment of human neonatal HIE and discussed the most recent advances in neuroprotection for this disorder. There are interesting questions that arise in relation to treating neonatal foals with neurological signs. The most pertinent of these include: (1) Can we more accurately define hypoxic ischemic brain injury? (2) Is it practical and financially feasible to start to integrate MRI into the diagnostic workup of foals that have severe, more prolonged neurological abnormalities? (3) Could we obtain pilot aEEG data to help increase our understanding of seizure activity? (4) Should we be considering the use of therapeutic hypothermia to treat HIE foals? Despite this intervention, the outcome for some foals may be fixed, but given the value of these animals, any advance in management may prove to be cost-effective.