Clinical manifestations

Presentation and course

Perinatal hypoxic-ischemic encephalopathy represents a distinctive, recognizable clinical syndrome that evolves after acute intrapartum disruption of fetoplacental circulation and neonatal cardiorespiratory failure (24; 08; 131; 42; 93). Perinatal hypoxic-ischemic encephalopathy can result from events related to antepartum factors such as chronic hypoxia (uteroplacental insufficiency), superimposed on intrapartum factors such as infection (chorioamnionitis), and hypoxic-ischemic insult such as placental abruption. Studies using MRI, MRS, NIRS, and EEG suggest that many brain injuries in infants with hypoxic-ischemic encephalopathy occur in the immediate perinatal period, but some injuries follow insults days or weeks earlier. Other factors, such as infection, can play a role by sensitizing the brain to subsequent hypoxia-ischemia (58).

Although perinatal hypoxic-ischemic encephalopathy represents a distinct, recognizable clinical syndrome, none of the clinical manifestations are specific for this disorder. Several components of the clinical presentation that should be considered in determining if an infant has incurred a significant intrapartum asphyxial insult that has resulted in acute neonatal encephalopathy include: (1) presence of sentinel events such as uterine rupture or cord prolapse during labor and at delivery; (2) evolution of signs of encephalopathy such as seizures and EEG abnormalities in the early postnatal period; (3) associated systemic derangements noted acutely in the first few days of life. It used to be that the asphyxia severe enough to cause brain injury was defined by cord blood gases showing pH of less than 7.0 with base excess of -12 or greater and Apgar score of 5 or less, but the standards for asphyxia have changed (61). According to the American College of Obstetricians and Gynecologists’ Task Force on Neonatal Encephalopathy, a 5-minute APGAR score of 7 or higher makes asphyxia unlikely; an umbilical cord pH of 7.2 or higher is unlikely to be associated with neonatal encephalopathy and brain injury (06). A cord pH of less than 7 is associated with a higher risk of neonatal encephalopathy and brain injury, though not necessarily. An umbilical cord base excess of -12 or less is unlikely to be associated with neonatal encephalopathy and brain injury. Interestingly, Yeomans and colleagues investigated normal values for umbilical arterial and venous pH, pCO2, pO2, and bicarbonate in the cord blood of 146 infants after uncomplicated labor and vaginal deliveries at 37 to 42 weeks’ gestation (136). Mean umbilical arterial values were pH of 7.28 +/- 0.05 and bicarbonate of 22.3 +/- 2.5 mEq/L. Mean umbilical venous values were pH of 7.35 +/- 0.05 and bicarbonate of 20.4 +/- 4.1 mEq/L.

Asphyxial insult. An integral component of the history in babies with possible hypoxic-ischemic encephalopathy is identification of specific evidence of intrauterine events that would predispose to asphyxia, for example abruption of placenta, cord compression, fetal-maternal transfusion, maternal eclampsia, maternal hypotension, and cord prolapse. The value of retrospective analysis of fetal heart rate monitoring to identify infants who have incurred significant asphyxia is questionable because the false positive rate of heart rate abnormalities is high (94). Sustained marked reduction or loss of the fetal heart rate during labor is highly suggestive of intrauterine asphyxia, but similar changes may occur in fetuses that have sustained injuries weeks or months before labor. Graham and colleagues found that abnormalities during the last hour of fetal heart rate monitoring before delivery are not predictive of neonatal hypoxic-ischemic encephalopathy (55). Failure to breathe at birth is also not necessarily an indicator of intrauterine asphyxia; it can occur for many reasons, including congenital brainstem anomalies, CNS infection, weakness of respiratory musculature, or respiratory depressant drugs administered to the mother during labor. Although metabolic acidemia is significantly associated with hypotonia at the time of birth, the majority of neonates with hypotonia, depression, or seizures at birth do not have objective evidence of asphyxia as measured by a cord gas at the time of delivery (119).

It is important to keep in mind that asphyxia is the term used to describe severe reduction of oxygen delivery to the fetus or neonate, whereas potential hypoxic-ischemic encephalopathy describes the brain manifestations of this deprivation. Similar to other types of brain injury at other ages, the severity of asphyxia is associated with a wide range of possible outcomes that are influenced by numerous factors including gestational age, body temperature, presence of infection, sex, and preconditioning due to prior stress (71). For example, placental inflammatory villitis was associated with poor neurodevelopmental outcomes (90).

Clinical features of neonatal encephalopathy. In neonates who have incurred major asphyxial insults that result in cerebral hypoxia-ischemia, distinct patterns of clinical abnormalities have been recognized in the acute postnatal period. The major features of the clinical syndrome of hypoxic-ischemic encephalopathy in the first 48 hours of life include decreased responsiveness, respiratory suppression, seizures (which are often extremely difficult to treat for 24 to 48 hours), cerebral edema visualized on head ultrasound or CT, and loss of brainstem reflexes. Classification schemes have been developed in an attempt to link clinical manifestations (111).

Table 1. Sarnat Classification of Neonatal Encephalopathy

Commonly cited features of mild encephalopathy include hyperalertness and hyperexcitability with a normal or mildly normal EEG, and these features are unlikely to be associated with permanent brain injury. Moderate encephalopathy is characterized by lethargy, hypotonia, suppressed primitive reflexes, and seizures. Features of severe encephalopathy include stupor or coma, flaccidity, absent reflexes, increased intracranial pressure, and seizures. The EEG in babies with moderate hypoxic-ischemic encephalopathy can show signs of slowing or seizures, whereas babies with severe encephalopathy often show the burst suppression pattern or severe flattening. The prognostic value of these clinical manifestations is greatest when the most severe signs are present (44). Infants with severe hypoxic-ischemic encephalopathy often manifest intractable seizures, beginning 6 to 8 hours after the ischemic insult, persisting for several days, and then often remitting completely. These seizures are often difficult to treat. Yet, after seizures stop, antiepileptic drugs can generally be weaned. Using MR imaging, Cowan and colleagues reported that babies with hypoxic-ischemic encephalopathy have distinctive and relatively symmetric lesions, whereas babies with seizures alone without other signs of encephalopathy have strokes (32). This may be as useful clinical pearl.

Associated systemic abnormalities. Asphyxial insult of sufficient severity to cause brain injury can also lead to injury of other organ systems (103), and the most common target organs are the kidneys (acute tubular necrosis, oliguria, or anuria); the heart (elevated cardiac enzymes, cardiac dysfunction, tricuspid insufficiency, or myocardial ischemia on ECG); the lungs (persistent pulmonary hypertension) and the liver (increased liver enzymes). Furthermore, hypoxia and ischemia results in disbalance microvascular tone, which may have a supporting role in the pathogenesis of necrotizing enterocolitis. However, multiorgan involvement is not essential in the diagnosis of hypoxic-ischemic encephalopathy in some cases. For example, babies with profound or near total asphyxia can escape without kidney damage.

Prognosis and complications

It is clear that the interpretation of outcome research data is often complicated by lack of precision in the initial diagnosis of birth asphyxia in study patients (81). Currently, it is believed that if an infant sustains a significant hypoxic-ischemic insult intrapartum that is responsible for subsequent neurologic deficits, the infant will have clinical signs of encephalopathy in the acute postnatal period (46). In individual patients, assessment of prognosis is generally based on a combination of clinical features and severity of EEG and brain imaging abnormalities (44; 108; 06).

Burst suppression, low-voltage tracing, and flat tracing in the EEG of these neonates were found to be effective in predicting long-term neurodevelopmental outcomes (09). Many infants who exhibit significant intrapartum asphyxia recover postnatally and exhibit no further neonatal or later abnormalities. In term infants, there is consensus that the only children with adverse neurologic outcome are those who manifest early patterns of moderate or severe encephalopathy as newborns (06). In this group, approximately 20% have significant adverse neurologic outcomes including motor deficits, cognitive impairment, and epilepsy. Analysis of large cohorts, such as the NIH Perinatal Collaborative Study, showed that the developing motor system is the brain region that is most susceptible to perinatal ischemic injury. If perinatal asphyxia causes CNS injury, then a motor disorder (ie, cerebral palsy) will usually be an integral component of the resulting neurologic disorder.

Term infants with perinatal asphyxia who did not receive hypothermia were prospectively followed in neurodevelopmental clinics in Nepal; Denver Developmental Screening Tool was utilized at 3 months, 6 months, 9 months, 1 year, 18 months, and 2 years (02). One hundred and eighty-seven assessments were completed; language delay was found in about 19%, gross motor delay was found in 29%, 18% had a fine motor delay, 17% had social delay, and impaired hearing and vision was noted in 5%. Seizures were persistent in 15% of patients at 2 years of age (02).

The impact of common functional polymorphisms in the antioxidant genes (SOD2, GPX1, and CAT) associated with decreased capacity for depending against ROS was investigated; the CAT rs1001179 polymorphism was found to be useful in identifying children that have a higher susceptibility to cerebral palsy after perinatal hypoxic-ischemic encephalopathy (40).

The American College of Obstetrics and Gynecologists (ACOG), the American Academy of Pediatrics, and the International Cerebral Palsy Task Force have concluded that the following criteria should be comprised in order to establish that intrapartum hypoxic-ischemic insult has caused moderate to severe neonatal encephalopathy: (1) evidence of metabolic acidosis in fetal umbilical cord arterial blood obtained at delivery (pH < 7 and base deficit of -12 mmol/L or more); (2) early onset of severe or moderate neonatal encephalopathy in infants born at 34 or more weeks’ of gestation; (3) low Apgar scores at 5 and 10 minutes; (4) exclusion of other identifiable causes such as trauma, coagulation disorders, infectious conditions, or genetic disorders (06). The Task Force also emphasized spastic quadriplegic and dyskinetic subtypes of cerebral palsy that are more commonly associated with hypoxic-ischemic insult than other subtypes. Besides these criteria the following criteria provide additional suggestive evidence for an intrapartum event: (1) a sentinel hypoxic event occurring immediately before or during labor; (2) a sudden and sustained fetal bradycardia or absence of fetal heart rate variability in the presence of persistent, late, or variable decelerations, usually after a sentinel hypoxic event when the pattern was previously normal; (3) onset of multisystem involvement within 72 hours of birth; (4) early imaging studies showing evidence of acute nonfocal cerebral abnormality (61).

Clinical vignette

A 23-year-old woman was in labor with her first pregnancy and making good progress when at 8 centimeters of dilatation a vaginal exam revealed a knuckle of umbilical cord protruding between the baby’s head and cervix, and the fetal heart rate tracing revealed a sustained bradycardia to 40 beats per minute. The obstetrician pushed the head upward to try to decompress the cord, and the mother was taken into the operating room for an emergency C-section, which was accomplished within 20 minutes. The male baby weighed 3700 grams, and the baby’s 1-minute Apgar score was 2, 1 for heart rate of 70 and 1 for respiratory effort, and the 5-minute Apgar was 3. The baby was resuscitated and by 10 minutes the Apgar score was 7. The cord blood gases were arterial pH = 6.92, pO2 = 5 CO2 = 70, and base excess = -16.4. The baby remained floppy for the next 2 hours and did not cry. Because of the sentinel event of the prolapsed cord associated with severe bradycardia, as well as evidence of low Apgar score of 3 at 5 minutes, cord arterial gases of pH less than 7 and base excess greater than -12, and floppiness with little cry, the baby was diagnosed with asphyxia (Apgar score of 3 at 5 minutes + pH < 7.0, base excess > -12) and hypoxic-ischemic encephalopathy (low tone, depressed level of consciousness). In this case moderate total body hypothermia is indicated because, as described below, it has been shown in multiple randomized controlled trials to reduce the incidence of later cerebral palsy. The baby was screened for infection because hypothermia could’ve worsened sepsis and caused bleeding, but no evidence of this was found. A head ultrasound was done to rule out hemorrhage, which may worsen with hypothermia, but none was detected. At 4 hours of age the baby was put on a hypothermic blanket with rectal temperature monitor and the temperature was maintained at 33.5o C for 36 hours. The baby was monitored with an integrated EEG monitor that did not show any abnormality for 12 hours at which time seizures were suspected, and the baby was started on a continuous 21 lead EEG monitor that revealed electrographic seizures. The baby was then loaded with 20 mg/kg of levetiracetam after which the seizures stopped. A head ultrasound at 2 days indicated moderate cerebral with echogenicity in the basal ganglia bilaterally. The baby was rewarmed over 6 hours off after 3 days and began to improve with respect to activity, tone, and feeding and began to take oral feedings at 5 days. An EEG and MRI were done at 12 days and were both normal. Follow-up in a neonatal follow-up clinic at 2 years of age revealed no sign of cerebral palsy, and speech was normal. The baby continued in follow-up for neurodevelopmental disorders.

In this case, the sentinel event of cord prolapse gave ample warning that asphyxia was occurring and that the baby was at risk of developing hypoxic-ischemic encephalopathy. The baby was started on hypothermia because all data showed asphyxia severe enough to damage the brain, and the baby appeared encephalopathic. The integrated EEG indicated seizure activity, and the baby was started on levetiracetam, which is well tolerated in babies. Head ultrasound initially did not show edema but did show edema at 48 hours in the basal ganglia, the expected focus on injury in this case of near-total asphyxia due to cord compression. The baby recovered nicely and had normal EEG and MRI at 12 days, which has a reasonably good correlation with later outcome. Although the baby had definite asphyxia, encephalopathy, and evidence of seizures and echogenicity on ultrasound, the hypothermia was apparently effective in preventing permanent brain damage.

Biological basis

Etiology and pathogenesis

Interruption of gas exchange between the fetoplacental and maternal circulations results in hypoxemia, hypercarbia, and acidosis, and if prolonged, results in hypoperfusion leading to concomitant cerebral ischemia. It is clear that many infants who have respiratory or circulatory depression at birth do not develop evidence of encephalopathy, and it is unclear whether differences in vulnerability to hypoxic-ischemic encephalopathy simply reflect differences in the severity and duration of the ischemic insult, or whether other undefined genetic or environmental influences influence the neonate's susceptibility to cerebral ischemia.

With respect to identifiable etiologic factors, these are currently identifiable only in a minority of cases and include obstetric complications such as placental abruption, placenta previa, or fetal-maternal transfusion.

Cranial imaging (especially magnetic resonance imaging) and detailed neuropathologic assessment are enabling identification of a range of pathogenetically relevant developmental anomalies in children whose deficits would otherwise have been attributed to asphyxia. Also, there is increasing recognition that preexisting CNS anomalies (eg, of the brainstem) may predispose infants to neonatal apnea and, thus, to perinatal asphyxia at birth.

There is also increasing evidence that prenatal hypoxia-ischemia appears to be an important etiologic factor in the pathogenesis of brain injury in some children who manifest evidence of CNS dysfunction in the perinatal period (19).

Upregulation of markers of tissue hypoxia, namely adrenomedullin and vascular inducible growth factor, in placental tissues of pregnancies complicated by birth asphyxia are indicative of the role of antepartum events that can play a role in perinatal asphyxia (129; 128).

The clinical manifestations of perinatal hypoxic-ischemic encephalopathy are directly attributable to cerebral hypoxia-ischemia, which evolves because of interruption of the fetoplacental circulation or cardiorespiratory failure in the neonate. Immaturity of cerebral autoregulation increases the neonate's vulnerability to systemic hypotension. Oxidative stress, excitotoxicity, and inflammation contribute to brain injury following hypoxic-ischemic insult and result in accelerated cell death or apoptosis. Following brain injury, the variability in the occurrence of neurodevelopmental sequela such as cerebral palsy may be influenced by genetic polymorphisms (43; 70).

A complex cascade of biochemical events triggered by the accumulation of extracellular levels of the excitatory amino acid neurotransmitter glutamate secondary to impaired glutamate reuptake due to energy failure leads to irreversible neuronal injury, and that progression of injury can be blocked pharmacologically using a variety of potentially complementary strategies. The cellular targets of hypoxia ischemia differ depending on age; however, the cascade of injury occurs in a similar way regardless of age. This cascade has been classified according to an apoptosis-necrosis cell death continuum based on morphological and biochemical data. Programmed cell necrosis has prominent contribution to the neurodegeneration of hypoxic-ischemic encephalopathy seen in animal models (96).

Damage does not stop at the time of the acute injury. Cell death occurs from hypoxia and energy depletion. Reperfusion injury, increased free radical formation, excitotoxicity, and nitric oxide production with secondary energy failure and delayed death are considered phase 2 events in this cascade (73).

Neuropathologic and MR imaging studies have identified several patterns of ischemic injury in asphyxiated term neonates. Near total asphyxia lasting from 10 to 30 minutes due to cord compression as described above usually produces damage to the putamen, ventrolateral thalamus, and peri-Rolandic cortex (132; 70). These children usually have extrapyramidal or dyskinetic cerebral palsy and do not speak. Partial prolonged asphyxia over several hours usually produces multi-cystic encephalomalacia associated with spastic quadriplegia. Watershed infarcts are generally associated with severe hypotension. One study, using near-infrared spectroscopy, of cerebral hemodynamics and oxygenation in asphyxiated infants documented cerebral hypoperfusion in the first 12 hours of life in severely asphyxiated neonates, who subsequently developed neurologic abnormalities (130). Utilizing simultaneous EEG-diffuse optical tomography has allowed researchers to observe distinct hemodynamic changes that are temporally correlated with electrographic seizures. Diffuse optical tomography provides hemodynamic information in the form of changes in concentration of de/oxygenated hemoglobin, which can improve our understanding of the relationship between neural and vascular processes (121).

Defined as an inherent dynamic capacity of the central nervous system to undergo maturation and change in function and structure in response to experience and injury, patterns of neuroplasticity expressed by the developing brain were described: (1) developmental plasticity, (2) adaptive (experience-dependent) plasticity, (3) reactive plasticity to pre- and postnatal CNS injury, (4) excessive plasticity, and (5) brain vulnerability under certain conditions such as hypoxic-ischemic encephalopathy. Time-sensitive heightened plasticity responses in the developing brain are proposed as “windows of opportunity” for neuromodulatory intervention (68).

Epidemiology

Epidemiological studies have prompted a critical re-analysis of appropriate criteria for diagnosis of hypoxic-ischemic encephalopathy (46; Nelson and Ellenberg 1986; 81). An accurate estimate of the incidence of ischemic encephalopathy is too difficult to determine because of limitations in the precision of diagnostic criteria. Data from the Collaborative Perinatal project indicated that 5.4 per 1000 term infants had three or more neonatal neurologic abnormalities (decreased activity after the first day of life, feeding problems, poor suck, respiratory difficulty or seizures); these data do not comment as to etiology, and only a fraction of these cases is attributable to hypoxic-ischemic encephalopathy.

Current approximate estimates are that 3 and 1.5 per 1000 live births manifest neonatal encephalopathy and hypoxic-ischemic encephalopathy, respectively (77). It is estimated that 15% to 20% of cerebral palsy in the United States and similar countries is caused by intrapartum asphyxia, whereas in low-income countries such as rural Nepal and India, as many as 50% are due to asphyxia (70).

Prevention

It is unclear if altered obstetric practice can further reduce the incidence of perinatal hypoxic-ischemic encephalopathy. However, a study using magnetic resonance spectroscopy to assess neuroprotective treatments after perinatal hypoxic-ischemic brain injury suggests that any early intervention should focus on the following: preservation of energy metabolism, the reduction of glutamate excite-toxicity and oxidative stress, the maintenance of calcium homeostasis, and the prevention of apoptosis (16). Although it has been a major goal of fetal monitoring to identify infants at risk early enough to intervene, there are no convincing data that this approach has been effective.

Regarding the prevention of hypoxic-ischemic encephalopathy in neonates, Duran and colleagues from Turkey reported improved APGAR scores and a trend toward decreased hypoxic-ischemic encephalopathy following implementation of a neonatal resuscitation program (37). The other area of interest has been the use of room air oxygen versus 100% oxygen for neonatal resuscitation. The animal data show that resuscitation with room air oxygen is less harmful. Meta-analysis of previous clinical trials has shown that the resuscitation of term or near-term infants with room air as compared to 100% oxygen appears to be safe, reduces neonatal mortality, trends toward a decrease in severe hypoxic-ischemic encephalopathy (112). The 2019 European Consensus Statement advised using an initial inspired oxygen concentration of 30% for newborns less than 28 weeks of gestation, 21% to 30% for 28 to 31 weeks of gestation, and 21% for more than 32 weeks of gestation (123). To achieve adequate oxygenation, the oxygen concentration should be titrated using pulse oximetry.

Dietary creatinine supplementation during pregnancy is being explored as a prophylaxis to protect the fetus from multi-organ consequences of severe hypoxia at birth (38). Resveratrol, a natural polyphenol, was given before and immediately after a hypoxic-ischemic brain event to neonatal rats (07). Mitochondrial function and cognitive function were assessed; pretreatment with resveratrol was reportedly neuroprotective but did not prove to be helpful when given post hypoxic injury.

Differential diagnosis

Confusing conditions

There is increasing recognition that the clinical syndrome of neonatal encephalopathy cannot be attributed to an intrapartum hypoxic-ischemic insult unless specific diagnostic criteria are met and that in the majority of cases of full-term infants with neonatal encephalopathy etiologic factors often originate in the antepartum period (01). Although the importance of developing more precise diagnostic criteria for hypoxic-ischemic encephalopathy is evident, this task has proven difficult. The first step in approaching the differential diagnosis of neonatal encephalopathy more rigorously is not to equate low Apgar scores with antecedent asphyxia. Minimum criteria for diagnosis of perinatal asphyxia, as an etiologic factor for neonatal encephalopathy include: identification of a specific event that could lead to fetal hypoxemia and concomitant observation of intrauterine signs of fetal distress (eg, fetal bradycardia, low fetal scalp pH), severe acidemia (pH < 7) on umbilical artery blood sample; prolonged low Apgar scores (less than 3, greater than 5 min); neonatal seizures, hypotonia, and stupor or coma; and multiorgan dysfunction that could result from hypoxia and ischemia acutely.

In the clinical setting of neonatal encephalopathy and without a clear-cut etiology, a thorough diagnostic evaluation is essential; important diagnostic considerations include identification of any treatable causes of CNS dysfunction, such as infection (in particular, treatable disorders such as bacterial meningitis and herpes encephalitis) or correctable metabolic derangements, maternal drug abuse, in utero toxin exposure, or developmental anomalies. Lumbar puncture is essential to evaluate the possibility of underlying CNS infection; cerebrospinal fluid abnormalities may provide evidence of bacterial or viral CNS infection. In addition to routine measurements of calcium, glucose, magnesium, electrolytes, and liver function tests, measurements of ammonia, lactate and pyruvate, serum amino acids, and organic acids may rarely identify inherited metabolic diseases; extensive metabolic evaluation is particularly important if there is a family history of unexplained neonatal encephalopathy, consanguinity, or death in siblings. Familial neonatal seizures due to genetic channelopathies should be considered. Metabolic disorders such as nonketotic hyperglycemia (NKH), sulfite oxidase deficiency, molybdenum cofactor deficiency, urea cycle disorders, and mitochondrial disorders need to be kept in mind to avoid misdiagnosing infants with hypoxic-ischemic encephalopathy. A diagnosis of a potentially fatal cause of neonatal encephalopathy is pyridox(am)ine 5’-phosphate oxidase deficiency; this diagnosis should be considered in a neonate with presumed hypoxic-ischemic encephalopathy in whom the degree of encephalopathy is not expected from perinatal history, cord gases, or neuroimaging (63). NKH causes a flat EEG that resembles hypoxic-ischemic encephalopathy, and patients often have characteristic hiccupping and need ventilator support. Genetic disorders such as Angelman syndrome, Joubert syndrome, Rett syndrome in males, Aicardi syndrome, and Zellweger syndrome have neonatal encephalopathies that can be mistaken as hypoxic-ischemic encephalopathy (42). Congenital muscular dystrophies with labor problems and severe hypotonia may also lead to a misdiagnosis.

Diagnostic workup

Essential historical features include identification of specific antenatal or intrapartum factors that might predispose the infant to asphyxia or other cause of neonatal encephalopathy. In the physical examination, important clues as to etiology of encephalopathy may be provided by presence of congenital anomalies, or evidence of discrepant growth parameters (eg, relative microcephaly, which would suggest an antenatal insult and birth weight). With respect to factors that might predispose the infant to asphyxia, examination of the placenta may be helpful.

Routine laboratory tests. Routine laboratory tests essential for evaluation and treatment of any sick neonate include complete blood count, electrolytes, serum glucose, calcium, magnesium, blood urea nitrogen or creatinine, liver function tests, arterial blood gases, and blood culture. There have been attempts to develop reliable metabolic markers of brain injury, but currently none are of established clinical utility. In the setting of a clear-cut antecedent insult, eg, placental abruption, typical clinical syndrome of encephalopathy, and evidence of multisystem ischemic injury, little diagnostic workup is indicated.

Both conventional and amplitude integrated EEG (aEEG) have critical roles in catching seizures, measuring effectiveness of antiepileptic drugs in terminating seizures, which may be difficult to assess by clinical observation, and distinguishing nonepileptiform paroxysmal events from seizures. The American Clinical Neurophysiology Society recommended continuous conventional EEG monitoring for detection of especially electrographic seizures and monitoring encephalopathy degree of encephalopathy in neonates with hypoxic-ischemic encephalopathy, though they also pointed that there may be handicaps with regard to the availability of equipment and technical and interpretive personnel in many centers (118). Furthermore, the current neonatal seizure classification of the International League Against Epilepsy highlights the importance of EEG for seizure diagnosis. Clinical events that do not have an EEG/aEEG correlation are not considered as seizures. Seizures are divided into two types: electroclinical and electrographic only. Electroclinical seizures are further classified as motor, nonmotor, and sequential seizures based on the predominant clinical feature (106).

Although conventional EEG is more sensitive and specific, introduction of aEEG in neonatal intensive care units has facilitated the ability to continuously monitor brain function. The aEEG has been used in the selection of patients with hypoxic-ischemic encephalopathy for the randomized control trial of hypothermia as neuroprotective therapy. Features that have been useful clinically in management and predicting outcome include (1) initial background pattern on admission as well as the rate of recovery in the first 24 to 48 hours after birth, (2) presence or absence of sleep-wake cycles, and (3) presence of electrographic seizures, (4) presence of burst suppression and low voltage (35; 09). A continuous abnormal aEEG at 48 hours or more has been shown to be associated with a more adverse neurodevelopmental outcome (26).

Cranial imaging (ultrasounds, CT, MRI, MRS, NIRS). Cranial imaging studies may be used to attempt to confirm the diagnosis of hypoxic-ischemic encephalopathy, to exclude intracranial hemorrhage, and to assess the extent of irreversible brain injury. Cranial ultrasound studies enable bedside evaluation of severity of cerebral edema (evidence includes diffuse increases in echogenicity, loss of landmarks, and decreased ventricular size); this modality is less sensitive for acute identification of infarcted tissue. Ultrasound studies have been particularly valuable for delineating the periventricular ischemic white matter injury that evolves in premature infants. CT scans (noncontrast) are often helpful in the first few days of life, to assess the extent of ischemic injury and edema, and to exclude major developmental anomalies. Patterns of abnormalities detected include watershed infarcts, multifocal vascular territory infarcts, hemorrhages, and venous thrombosis. Prior studies showed extensive areas of decreased attenuation on noncontrast CT scans ("decreased density") acutely appears to have strong prognostic value for severity of ischemic brain injury (45).

MRI have the highest diagnostic and prognostic yields beyond other imaging tools, especially in moderate and severe hypoxic-ischemic encephalopathy. The American College of Obstetricians and Gynecologists’ Task Force on Neonatal Encephalopathy recommends MRI in first 24 to 96 hours of life and a further imaging in follow-up. The utility of different modalities of MRI scanning in the work-up of infants with hypoxic-ischemic encephalopathy has been studied by various investigators (74; 31; 14; 110; 67). In the first 10 days of life several distinct patterns of brain injury have been identified by MR imaging in asphyxiated neonates, including deep gray matter involvement (basal ganglia thalamus pattern) affecting central grey nuclei, perirolandic cortex, and, not uncommonly, hippocampus and brainstem; primarily cortical involvement; primarily periventricular white matter injury (watershed pattern) affecting watershed zones between anterior-middle and middle-posterior cerebral arteries; and mixed patterns (14; 34). Diffusion-weighted MR imaging may be particularly sensitive for early diagnosis and grading of the severity of hypoxic-ischemic brain injury (31). Abnormal magnetic resonance signal in the internal capsule, the first area to myelinate in the immature brain, as well as decreased apparent diffusion coefficient in the internal capsule by diffusion weighted imaging MRI, is associated with poor neurodevelopmental outcome (110; 67). Experience with MRI and EEG in the period after exposure to an asphyxial insult and evolution of encephalopathy suggests that brain MRI at 10 to 12 days is often predictive of later outcome, and a normal MRI at this time predicts good outcome. At this time, a normal EEG and a normal MRI provides even stronger evidence of good outcome. A depressed EEG activity during the first 72 hours of life and a diffused alteration of basal ganglia at MRI were correlated with a poor neurodevelopmental outcome at 18 months of follow-up (33). More follow-up studies are needed to refine this information (17; 70). A retrospective study was conducted to compare the prognostic value of early (less than or equal to 6 days) and late (greater than or equal to 7 days) MRIs in predicting adverse outcome at 2 years of age in asphyxiated term neonates treated with hypothermia; two radiologists analyzed each MRI independently. Utilizing a visual analysis, the imaging was classified as normal, subnormal, or abnormal. Apparent diffusion coefficient (ADC) values were measured within predefined areas, and the signal intensity of the posterior limb of the internal capsule was analyzed. Neurodevelopmental outcome was assessed at 18 to 24 months as adverse or favorable. Early MRI was found to be a good predictor of neurodevelopmental outcome at 2 years of age (27). Similarly, a retrospective review was performed to determine the optimal time window for MR imaging with quantitative ADC measurement in neonatal hypoxic-ischemic encephalopathy post-hypothermia. The time frame of 3 to 10 days was found to be the optimal window for MR imaging with ADC post-hypothermia (79).

Other newer modalities that may prove useful adjuncts to evaluation of neonatal encephalopathy include nuclear magnetic resonance spectroscopy (based on 31P or 1H) and near-infrared spectroscopy. Martin and colleagues (85) reported that 31P magnetic resonance spectroscopy measurements of phosphocreatine or ATP correlated strongly with the degree of hypoxic-ischemic encephalopathy in asphyxiated term neonates, and that there was a significant correlation between depressed levels of these metabolites and adverse neurologic outcome. Similarly, Penrice and colleagues reported that measurements of brain lactate relative to N-acetylaspartate using proton magnetic resonance spectroscopy in asphyxiated infants indicated that raised lactate and N-acetylaspartate ratios carried a poor long-term prognosis (1-year follow-up) (101). Proton-magnetic resonance spectroscopy findings of increased lactate-choline ratios in basal ganglia and thalami of infants (1 to 10 days) with asphyxia were found to be related to worse outcome (death or poor neurologic outcome at 5 to 19 months of age) (137). In a meta-analysis of studies addressing MRI performed in the neonatal period and neurodevelopmental outcome at greater than 1 year, Thayyil and colleagues found that proton magnetic resonance spectroscopy deep gray matter lactate/N-acetyl aspartate had better diagnostic accuracy than conventional MRI (126). They proposed the use of magnetic resonance biomarkers as surrogate end points for clinical trials. Gray matter novel magnetic resonance imaging subscore has also been shown to be an independent predictor of adverse outcomes at 2 years and school age whereas white matter and cerebellum added no predictive value (133).

Near-infrared spectroscopy is being developed as a bedside noninvasive method for assessment of cerebral blood flow and cerebral blood volume (135; 130). Meek and colleagues (88) using near infrared spectroscopy (NIRS) studied cerebral blood volume (CBV) and its response to changes in arterial PaCO2 (CBVR) in the first 24 hours of life in perinatally asphyxiated neonates and showed that an increase in CBV was a sensitive predictor of death or disability whereas a reduction of CBVR was not significantly correlated with the severity of adverse outcome. Nakamura and colleagues found that at 6 hours after birth, CBV was significantly higher in neonates with adverse outcomes and that combined CBV at 2 hours after birth plus cytochrome C oxidase assembly protein had the best predictive ability for neurodevelopmental outcome (92).

Pre- and post-hypothermia transfontanellar duplex brain sonography resistive indices (RI) were measured as a marker of cerebral hemodynamics in neonates who suffered from hypoxic-ischemic encephalopathy; values were found to be predictive of specific long-term neurodevelopmental outcomes (51). Neonates with RI values less than 0.60 prior to and following cooling were more likely to die or have severe neurodevelopmental disability by the age of 20 to 32 months than those with an RI greater than 0.60. Lower RI values were associated with specific neurodevelopmental deficits in motor skill attainment.

Diffusion tensor imaging of corpus callosum and corticospinal tract was measured to determine if there was an association of diffusion tensor imaging results with developmental outcomes after therapeutic hypothermia. Impaired microstructural organization of the corpus callosum and corticospinal tract was found to predict poorer cognitive and motor performance, respectively (86).

Biomarkers. In neonatal hypoxic-ischemic encephalopathy, the biomarkers may be able to determine the extent and the severity of injury, the timing of initial insult, and the evolution of brain damage. The biomarkers can help identify the infants at risk of adverse outcomes and, thus, identify those who would benefit the most from therapy aimed at postasphyxial brain injury. Studies have explored the utility of blood, urine, and CSF measurement of several biomarkers as indicators of asphyxial brain injury and tried to correlate them with neurologic outcome. Increased lactate to creatinine ratio in urine has been suggested as predictor of moderate to severe hypoxic-ischemic encephalopathy (66; 102). Cord blood cytokine levels, especially IL6, have been reported to be not only higher in infants with hypoxic encephalopathy but also in those with concomitant chorioamnionitis. Additionally, L-6, IL-8, IL-10, and IL-13 levels were shown to be inversely related to selected heart rate variability metrics, thus indicating that inflammation-modulating cytokines may be significant mediators in the autonomic dysfunction observed in newborns with hypoxic-ischemic encephalopathy (04). The magnitude of increase seems to correlate with degree of severity of encephalopathy, and neurodevelopmental outcome at 2 years (113; 28). Fetoplacental and cerebral hypoxia can stimulate the production of Actvin A, which then leads to increased production of erythropoietin with elevated numbers of nucleated red blood cells (NRBCs) in cord blood. Following perinatal asphyxia, the cord blood levels of nonprotein bound iron (NPBI), NRBCs as well as Activin A are reportedly increased (104). NPBI acts as a potent oxidant via production of OH groups.

Hagberg and colleagues reported high concentrations of excitatory amino acids in the CSF of asphyxiated infants (59); Garcia-Alix and colleagues reported that concentrations of neuron-specific enolase were elevated in the CSF of asphyxiated neonates between the ages of 12 and 72 hours, and the magnitude of increase correlated with adverse outcome at 1-year follow-up (49). Other investigators have shown that higher CSF neuron-specific enolase levels were seen in more severe encephalopathy and that the higher levels correlated with MRI and EEG finding (41). Increased serum level of astroglial protein S100 on the first day of life in infants with hypoxic-ischemic encephalopathy was reported to occur more frequently in those who died or developed cerebral palsy, than those with no impairment (127). Increased levels of nonprotein bound iron and ortho-tyrosine and meta-tyrosine, markers of protein oxidation induced by hydroxyl radicals, have been found in CSF of newborn infants with hypoxic-ischemic encephalopathy (97). Serum interleukin-1b and 6, cerebrospinal neuron-specific enolase, and interleukin-1b measured before 96 hours of age are also being examined as putative predictors of abnormal outcomes (107). The median values of serum anti–Hsp 70 antigen titers were found to be significantly higher in asphyxiated compared with non-asphyxiated neonates and proposed as a serum marker for early diagnosis of prenatal hypoxia (21). Blennow and colleagues found that CSF concentrations of the glial structural protein glial fibrillary acidic protein (GFAP) were increased 5-fold in full-term asphyxiated infants, in comparison with controls (20). Ennen and colleagues (2011) reported that serum GFAP is a biomarker for neonatal hypoxic-ischemic encephalopathy treated with whole body cooling and were predictive of brain injury on MRI (39). Massaro and colleagues reported that ubiquitin carboxyl-terminal esterase L1(UCH-L1) is also a biomarker for hypoxic-ischemic encephalopathy treated with whole body cooling with a time course that complements changes in GFAP (87). GFAP and UCH-L1 are proposed as biomarkers that predict long-term neurologic sequelae of hypoxic-ischemic encephalopathy (84). Salivary lactate dehydrogenase was found to be significantly higher in the hypoxic-ischemic encephalopathy group compared to controls and is a proposed biomarker (89). Microribonucleic acids (miRNA) are non-coding RNA molecules that have been demonstrated to change during hypoxic-ischemic encephalopathy, and assessing miRNAs could be another promising biomarker (98; 134).

Diffusion tensor imaging provides qualitative and quantitative information about the microstructure of the brain; a near-infrared spectroscopy index can assess cerebrovascular autoregulation. Tekes and colleagues hypothesized that lower apparent diffusion coefficient scalar values would correlate with worse autoregulatory function (124). They found that in neonates who had MRI on day of life 10 or greater, lower apparent diffusion coefficient scalars in the posterior centrum semiovale and the posterior limb of the internal capsule correlated with blood pressure deviation below the range with optimal autoregulation during hypothermia. Lower apparent diffusion coefficient scalars in the basal ganglia correlated with worse autoregulation during rewarming. They proposed that blood pressure deviation from the optimal autoregulatory range measured by apparent diffusion coefficient scalars may serve as a biomarker of injury in the posterior centrum semiovale, posterior limb of the internal capsule, and basal ganglia (124). However, there are no biomarkers that are established as reliable for clinical use at this time.

Management

Treatment of hypoxic-ischemic encephalopathy in term infants has taken a dramatic turn with the introduction of moderate total body hypothermia to maintain core (rectal or esophageal) temperature at approximately 33.5°C for 3 days and started within 6 hours of birth (120). Three large randomized controlled trials have demonstrated an approximately 40% improvement in survival without cerebral palsy and related neurodevelopmental disabilities in the cooled group compared to controls (53; 114; 10). A Cochrane Database systematic review of 11 studies found that the technique is beneficial in term and late preterm infants (69). A 2017 regional cohort trial has also demonstrated reduced infancy/childhood epilepsy following hypothermia treatment (83).

Eligibility criteria for hypothermia are gestational age of 36 weeks or more and 6 or fewer hours of age, moderate-severe encephalopathy, pH of 7.0 or less, or a base deficit of 16 or more mmol/L in a sample of umbilical cord blood or blood obtained during the first hour after birth (99). If blood gas is not available or pH is between 7.01 and 7.15, or base deficit is between 10 and 15.9 mmol/L on blood sample obtained within the first hour of birth, two additional criteria are needed: a history of an acute perinatal event (eg, cord prolapse, fetal heart rate decelerations) and either the need for assisted ventilation initiated at birth and continued for 10 minutes or an Apgar score of 5 or less at 10 minutes after birth. Birth weight is another binding criterion, yet there are no sufficient data on cooling the babies born under 1800 gr.

One study showed that therapeutic hypothermia in term newborns with hypoxic-ischemic encephalopathy provides better survival and less neurologic sequelae (139). A randomized clinical trial was performed to determine if longer duration cooling (120 hours), deeper cooling (32.0 degrees Celsius), or both are superior to cooling at 33.5 degrees Celsius for 72 hours in full-term neonates with moderate to severe hypoxic-ischemic encephalopathy/cytokine. Among neonates who were full-term with moderate or severe hypoxic-ischemic encephalopathy, longer cooling, deeper cooling, or both compared with hypothermia at 33.5 degrees Celsius for 72 hours did not reduce NICU death, so no changes were made to the standard protocol of 33.5 degrees for 3 days (115). A similar study performed in 2017 found the same results: that there was no reduction of death with longer and deeper cooling compared to the normal 72 hours and 33.5 degrees Celsius; however, this study found a statistical interaction between longer and deeper cooling, suggesting support for the current cooling time and degrees (116). Another randomized clinical trial from the United States focused on hypothermia initiated at more than 6 hours after birth and its influence on death and short-term disability. The researchers concluded that this strategy showed a probable benefit but had uncertainty in effectiveness (78).

The decision to begin hypothermia for babies with probable hypoxic-ischemic encephalopathy is made by neonatologists at specialist centers often in collaboration with pediatric neurologists. Each center has a set of guidelines that includes a matrix including possible sentinel events associated with delivery, Apgar scores, cord blood or early neonatal arterial blood gas values, early assessment of signs of encephalopathy, and contraindications such as evidence of sepsis or bleeding, especially into the brain. Probable sepsis and evidence of bleeding on head ultrasound or other excessive bleeding are contraindications to cooling. During cooling the baby may need medications such as morphine to stop shivering, and brain activity is monitored continuously by an EEG. If there is evidence of seizures, a prolonged conventional EEG is done. Both hyperglycemia and hypoglycemia in neonates with hypoxic-ischemic encephalopathy undergoing hypothermia were associated with poorer neurodevelopmental outcomes (29; 03; 15). Although the nutritional support for patients undergoing therapeutic hypothermia varies between centers, introducing enteral feeding and using maternal milk, if possible, during hypothermia appear to be safe in neonates (48).

Seizures are clearly a marker of neurologic disturbance, but it is uncertain to what extent uncontrolled seizures cause further brain injury; most clinicians rely primarily on phenobarbital for treatment of neonatal seizures. Some clinicians add a second antiepileptic drug such as phenytoin, benzodiazepine, or levetiracetam, if initial treatment with barbiturates is ineffective, but no systematic data are available comparing the efficacy of all these therapeutic approaches. A multicenter, randomized, blinded, controlled, phase IIb study investigating the efficacy and safety of levetiracetam compared with phenobarbital as a first-line drug treatment for neonatal seizures of any cause (54% had hypoxic-ischemic encephalopathy) has revealed greater efficacy of phenobarbital with more adverse events (117).

A pilot randomized, controlled, double-blind trial compared the effect of bumetanide, an inhibitor of Na-K-Cl cotransporter, to placebo (saline) as an add-on therapy to phenobarbital in neonatal seizures (122). The research showed that the bumetanide group experienced a statistically significant decrease in seizure load.

Furthermore, there are practical differences between physicians about the discontinuation of antiepileptic treatment on hospital discharge. A prospective, observational, multicenter study addressed this issue and evaluated 303 infants with acute symptomatic seizures, almost half of whom had hypoxic-ischemic encephalopathy (52). The study showed no difference in neurodevelopmental scores or epilepsy rates between patients who were discharged with or without antiepileptic medication, which supported discontinuation of antiepileptic drugs prior to hospital discharge in most of the infants.

In addition to hypothermia, standard supportive treatment includes ensuring adequate oxygenation and blood pressure as well as correction of acidemia and other metabolic derangements (specifically hypoglycemia, hypocalcemia, hypomagnesemia, or hyponatremia). In any neonate with seizures, metabolic causes must be excluded (ie, hypoglycemia, hypocalcemia, hypomagnesemia). No routine antiepileptic drug is recommended for seizure prophylaxis.

Various unresolved issues pertaining to hypothermia as neuroprotective therapy need further studies. Some of these include the following: the most effective technique of cooling (selective head cooling vs. whole body cooling); the most effective duration of cooling 48 or 72 hours or longer; the most appropriate target temperature; the most important CNS anatomic target (superficial cortex, deep cortex, or brainstem); the best way of identifying infants most likely to benefit (history, clinical criteria, or subjective measures such as EEG, amplitude integrated EEG, or diagnostic imaging); optimal age from birth to institution of therapy and how late is too late; the safest time and method of rewarming; and the long-term neurodevelopmental outcome (4 or 8 years or longer) (64).

Therapeutic moderate hypothermia is being applied for moderate to severe hypoxic-ischemic encephalopathy in the United States, Europe, Australia and many other countries, but it has not been widely applied in India or other moderate-income countries. In a pilot study, Horn and colleagues showed that moderate hypothermia could be induced using a servo-controlled fan, thereby offering an inexpensive method to provide neuroprotective therapy (65), and another group reported the use of convective cooling, which does not need electricity as a cost-effective treatment for hypoxic-ischemic encephalopathy in low-income countries (75).

A wide range of novel and potentially complementary pharmacological strategies are being developed to limit ischemic brain injury. The goal of therapy is to provide neuroprotection during the hypoperfusion and reperfusion periods following hypoxic injury. A period of 2 to 6 hours following termination of hypoxic-ischemic injury is the "therapeutic window," during which the therapy has to be initiated (131). Pharmacologic agents investigated include calcium channel blockers, oxygen free radical scavengers, excitatory amino acid antagonists, nitric oxide synthase inhibitors, and phenobarbital. Zhu and colleagues reported on the neuroprotective role of systemically administered erythropoietin for the treatment of neonatal hypoxic-ischemic encephalopathy (138). Besides erythropoietin, other pharmacological agents such as melatonin, topiramate, an antiepileptic drug, N-acetylcysteine, and inhaled argon, either alone or in conjunction with moderate hypothermia have been explored (30; 70; 05; 62; 22). In a neonatal cerebral hypoxia-ischemia rat model, early behavioral intervention combined with immunomodulatory inter-alpha inhibitor proteins has been shown to improve long-term working memory beyond that of each individual intervention alone (50). Hypothermia with concomitant mesenchymal stem cell therapy is considered another promising neuroprotective strategy, though preclinical data have not yet shown sufficient efficacy (125). Hypothermia plus therapies promises a potential benefit of additional neuroprotection in hypoxic-ischemic encephalopathy. Future studies with new agents about timing/dosing and early and late neuroprotective effects of them are needed.

Outcomes

The adverse events reported in the trials of therapeutic hypothermia were thrombocytopenia, leukopenia, sinus bradycardia, hypotension, elevated liver enzymes, pulmonary hypertension, and coagulopathy (18).

A retrospective cohort study investigated the decreased response to hypothermia in neonates with hypoxic-ischemic encephalopathy and infection (72). Chorioamnionitis diagnosed clinically and histologically was not found to affect the cord gas at delivery but was associated with an increased metabolic acidosis on the initial neonatal arterial gas. The investigators concluded that chorioamnionitis is associated with a persistent state of acidosis in neonates with hypoxic-ischemic encephalopathy that may contribute to worse neurologic outcomes.

A pilot study explored the relationship between acute autoregulatory vasoreactivity during treatment and neurodevelopmental outcomes at 24 months of age (23). The study found that motor and cognitive impairments at 21 to 32 months of age were associated with greater blood pressure deviation below the individual optimal mean arterial blood pressure at which autoregulatory vasoreactivity is greatest during rewarming following therapeutic hypothermia, but not with regional cerebral oximetry or blood pressure below gestational age + 5. Noninvasive assessment of cerebral pressure autoregulation utilizing near-infrared spectroscopy and systemic mean arterial blood pressure (NIRS-MAP) monitoring to evaluate whether impaired cerebral autoregulation measured by NIRS-MAP monitoring during therapeutic hypothermia and rewarming relates to outcome in newborns with hypoxic-ischemic encephalopathy was conducted (86). Higher pressure passivity index in both cerebral hemispheres and gain in the right hemisphere were associated with neonatal adverse outcomes.

A prospective study assessed the levels of specific cytokines and chemokines in 99 newborns with hypoxic-ischemic encephalopathy in the first week of life and explored the association of cytokines levels with neurocognitive outcome and death (100). They performed multiple regression analysis and compared neonates who survived without neurologic impairment to newborns who died or had neurologic impairment. Neonates who died or were impaired at 6 to 7 years following hypoxic-ischemic encephalopathy had lower RANTES and higher MCP-1 levels than those who survived without impairment.

Data of all neonates registered in the National Asphyxia and Cooling Register in Switzerland between 2011 and 2013 were analyzed for risk factors associated with subcutaneous fat necrosis, which is a potential adverse event of whole-body cooling (56). No association was found between method applied, proportion of temperature measurements outside target temperature range, and severity of hypoxic-ischemic encephalopathy.

The resistive index measured by cranial Doppler ultrasonography was measured in the anterior cerebral artery within 72 hours of hypothermia in term neonates to predict the risk of death/abnormal neurodevelopment. The presence of abnormal resistive index increased the risk of death/abnormal neurologic outcomes at 6 to 12 months (76).

Higher seizure burden (accumulated duration of seizures over a defined period) and excessive EEG discontinuity (diagnosed with a continuous two-channel amplitude-integrated EEG for a minimum of 48 hours) was found to be associated with increased cerebral tissue injury on MRI and was predictive of abnormal neurodevelopmental outcome in infants treated with hypothermia (36). However, the antiepileptic effects of hypothermia and its association with neurologic outcome in infants with moderate and severe hypoxic-ischemic encephalopathy were explored. Hypothermia was found to have an antiepileptic effect in moderate and severe neonatal hypoxic-ischemic encephalopathy. A lower seizure burden in cooled newborn infants with severe hypoxic-ischemic encephalopathy was more commonly associated with improved outcome at 24 months (57).

Most of the studies have focused on survival rates and neurologic outcome after hypoxic-ischemic encephalopathy. Rates of death, cerebral palsy, and major disability are lower in children who received therapeutic hypothermia compared to the ones who did not, but 29% of the treated babies still died, and 21% of them experienced cerebral palsy (11). Furthermore, newborns with hypoxic-ischemic encephalopathy, even the ones that underwent therapeutic hypothermia, are still under risk for intellectual disability, language deficits, and behavioral problems.

Both selective head and whole-body cooling modalities showed results in neurologic outcome. No randomized controlled trial determined the appropriate techniques of cooling, yet whole body cooling may have better neuroprotective effect (54).

Although most publications focus on moderate and severe hypoxic-ischemic encephalopathy, it is likely that mild cases also have long-term neurologic issues. A prospective multicenter study focused on the short- and midterm outcomes of infants with mild encephalopathy and revealed that a significant proportion of neonates showed impairment in MRIs and aEEGs during hospitalization and were discharged with an abnormal neurologic exam, with disability noted in 16% of the cohort at 18 to 22 months of age (25; 105). Despite the lack of evidence-based data on its usage, an increasing proportion of neonates with mild hypoxic-ischemic encephalopathy undergo hypothermia treatment (60). The effect of therapeutic hypothermia on MRIs was studied in a small nonrandomized cohort study of neonates with moderate encephalopathy (91). Cooled newborns showed lower white matter impairment scores and less metabolic injury than noncooled ones, but the difference in neurodevelopmental outcomes was not statistically significant. Researchers must also focus on these patients and formulate neuroprotective strategies (25).