Definition

Neonatal encephalopathy (NE) is an alteration in consciousness or neurologic exam in the neonate. The possible etiologies of NE are broad ( Table 47.1 ) but it is most commonly (approximately 50%) caused by hypoxic-ischemic encephalopathy (HIE). ^, The following sections are focused on HIE, although many details are applicable to other NE etiologies.

Diagnosis

The initial diagnosis of HIE relies on evidence of an acute or subacute (prolonged) perinatal event leading to brain injury and exam findings consistent with encephalopathy. Initial evaluation of an infant with suspected HIE should include perinatal history, neurologic exam, and in some cases, electroencephalogram (EEG) and imaging, when available.

Perinatal History

Evidence of intrauterine distress such as an abnormal biophysical profile, decreased fetal movements, fetal bradycardia, or acidosis on umbilical cord/initial infant blood gas (within 1 hour of birth), along with need for extensive resuscitation and/or low Apgar scores (<5) at 5 and 10 minutes, are needed for the diagnosis. ^,

Neurologic Examination

The clinical presentation depends on the duration, timing, and severity of the insult and may evolve over the subsequent hours to days. Although it is not the only scoring system, clinical staging of encephalopathy is often based on the Sarnat criteria ( Table 47.2 ). Using the modified Sarnat criteria, which do not include EEG, the current recommendations are to offer therapeutic hypothermia (TH) to infants with moderate to severe encephalopathy, although the efficacy of TH in severe encephalopathy is the subject of ongoing studies (see Chapter 32 ). The Sarnat score was used as an inclusion criterion in the first randomized controlled trials of TH ^{, ,} and is still used by most neonatal intensive care units for evaluation of the neonate needing TH. Longitudinal follow-up of Sarnat scoring suggests that worsening staging regardless of the initial staging is more predictive than the initial score alone.

Acute Phase

The fetal brain requires blood flow to deliver oxygen and glucose for cellular energy and metabolic homeostasis. Interruption of blood flow leads to hypoxemia and eventually decreases cardiac output. Depending on the severity and timing of the disruption of blood flow, cerebral injury may occur and lead to HIE. Abrupt disruption of cerebral blood flow is classically associated with deep grey matter (basal ganglia and thalamus) injury. Chronic and less severe (partial) disruption of cerebral blood flow is associated with cortical injury, particularly in the watershed regions.

The initial disruption of cerebral blood flow leads to mitochondria failure. The resulting adenosine triphosphate (ATP) depletion and lactic acid build-up impairs the function of excitatory amino acid transporters within the astrocytic membrane forming the synaptic cleft. As a result, excitatory amino acids such as glutamate accumulate and activate N-methyl-D-aspartate and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors, leading to an intracellular influx of sodium, calcium, and water. Cell death results from the progression of cytotoxic edema, protease activation, and free radical production ( Fig. 47.1 ).

After Neonatal Hypoxic-Ischemic Brain Insult, ATP Production Is Impaired Due to Decreased Oxygen and Glucose Delivery and Mitochondrial Failure.

The function of glutamate (glu) transporters with the astrocytic membrane is adenosine triphosphate (ATP)-dependent, and thus energy failure results in (A) impaired function of these transporters and (B) glutamate accumulation within the synaptic cleft. The resulting glutamate excitotoxicity activates NMDA and AMPA receptors. (C) NMDA receptor activation facilitates Ca ²⁺ influx followed by activation of synthases and production of free radicals from oxygen (i.e., superoxide) or nitrogen (i.e., nitric oxide). (D) Activation of AMPA receptors produces influx of Na ²⁺ and water, with it inducing cytotoxic edema. These events and the resulting cell death will initiate and perpetuate neuroinflammation. AMPA, α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; NMDA, N-methyl-D-aspartate.

Latent/Chronic Phase

Restoration of cerebral blood flow, or reperfusion, after disruption and injury begins the latent phase of injury. The latent phase is the therapeutic target of TH. Secondary energy failure, which signals the end of the latent phase, is marked by irreversible mitochondrial failure and involves oxidative stress, intracellular calcium accumulation, inflammation, and cell death. This neuronal death includes cell death along the entire cell death continuum, including apoptosis, autophagy, continuum cell death, necroptosis, and necrosis. Many of the therapeutic agents tested in the past decade have targeted many of these downstream mechanisms ( Figs. 47.2 and 47.3 ).

Downstream Mechanisms of Several Therapeutic Agents Being Tested Over the Last Decade for Hypoxic-Ischemic Encephalopathy.

(Robertson et al. Which neuroprotective agents are ready for bench to bedside translation in the newborn infant? J Pediatr . 2012;160:544–552.e4.)

Flow Chart Showing the Mechanisms Contributing to Each Phase of the Evolution of Neonatal Encephalopathy Over Time.

(Davidson et al. Seminars in Fetal and Neonatal Medicine , article 101267.)

Gestational Age/Prematurity

HIE is classically reported in full-term and late preterm infants and rarely in preterm infants. Although HIE occurs at all gestational ages, the assessment of encephalopathy in preterm infants is difficult and the clinical presentation of seizures may be subtle. There have been several small studies on preterm HIE, ^, which have found that due to differences in neurodevelopmental stage, injury mechanisms and patterns differ from those of the full-term neonate. Oligodendrocyte maturity, vascular and blood-brain barrier permeability, and selective vulnerability of certain cell lines such as developing oligodendroglia, astrocytes, and microglia likely determine the severity and type of injury. Preterm infants appear to exhibit greater susceptibility to white matter injury, most often associated with concurrent basal ganglia and thalamic injury, than do full-term infants (see Fig. 47.8 ahead).

Encephalopathy Score

Multiple scoring systems have been developed to assess severity of encephalopathy and predict long-term risk of neurodevelopmental disability (see Table 47.2 ). The first of these was the Sarnat scoring system. Clinically, in settings where other prognostic tools (magnetic resonance imaging [MRI], EEG) can be used, these scores are most commonly used to assess eligibility for TH. Regardless of the scoring or staging system used, trending the score/exam over time while in the acute period is important, because the examination will evolve over time and persistently high encephalopathy scores are generally predictive of worse outcomes. Of note, administration of sedation and/or antiepileptic drugs may alter the exam, which should considered.

Findings of moderate or severe encephalopathy on exam are inclusion criteria for treatment with TH; traditionally, children with mild encephalopathy were not thought to be at significant risk for neurodevelopmental disability, but there is growing evidence that neonates with mild encephalopathy may be at risk for long-term neurodevelopmental abnormalities. Neonates with mild HIE have been described to have abnormal short-term outcomes such as longer time to reach full feeds, seizures, need for a surgical feeding device, or abnormal MRI or neurologic exam findings at discharge. In a retrospective analysis of infants who had received TH, MRI abnormalities after TH were equally common among infants with mild HIE compared with those with moderate or severe HIE. Mild encephalopathy on early EEG is associated with lower IQ at age 5 years. The PRIME (Prospective Research in Infants with Mild Encephalopathy) study revealed that over half of infants with mild HIE had an abnormal short-term outcome. A recent meta-analysis showed that 20% of infants with mild encephalopathy have adverse neurologic outcomes at 18 months. Study of long-term outcomes in infants with mild HIE as well as the efficacy of TH for this group is ongoing.

Amplitude Integrated EEG/Continuous Video EEG

Encephalopathy on neonatal EEG has been standardized by the American Clinical Neurophysiology Society. Depending on available resources and setting, amplitude integrated EEG (aEEG; Fig. 47.4 ) or full montage EEG can be used ( Table 47.3 ). A normal or mildly abnormal early EEG background is highly predictive of normal neurodevelopmental outcomes at age 2 years. Recovery of EEG background within the first 24 hours is associated with better outcomes. Persistently abnormal EEG background activity 36 to 48 hours after injury has been associated with worse neurodevelopmental outcomes. aEEG is used by units particularly when continuous full montage neonatal EEG services are not available for monitoring throughout TH. A systematic review of aEEG use in HIE including 17 studies (both with and without TH) concluded that the aEEG background during the first 72 hours after HI injury has strong predictive value for long-term neurologic outcome.

Common Amplitude-Integrated Electroencephalography Patterns in Infants With Neonatal Encephalopathy.

(A) Normal sleep-wake cycling in a full-term infant with a Sarnat score of 0 at admission to the neonatal intensive care unit. Periods with voltage from 5–25 μV alternate with periods exceeding those limits. (B) Burst-suppression pattern in an infant with severe hypoxic-ischemic encephalopathy (Sarnat score of 3) cycling between periods of profound suppression of electrical activity and usually short bursts of increased activity. (C) Pattern of rhythmicity compatible with neonatal seizures in an infant with moderate encephalopathy (Sarnat score of 2), with typical elevation of baseline exceeding 10 μV. (D) Low voltage background pattern in an infant with a Sarnat score of 2. (Courtesy Charlamaine Henson, RN coordinator for the Johns Hopkins University—Neuroscience Intensive Care Program.)

Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIRS) is a bedside monitoring technique that continuously provides information on mixed arterial and venous saturations in the brain. In some cases cerebral saturations may be compared with control tissue such as the kidney ( Fig. 47.5 ). Mixed saturation (rSO ₂ ) is a measure of the differential between oxygen delivery to the brain tissue and oxygen extraction by the brain tissue. Small studies have examined the use of NIRS in infants with HIE undergoing TH, with varying results. Higher rSO ₂ values in the first 10 hours of TH were associated with injury on MRI ; however, other studies have not found a relationship between rSO ₂ values during TH and 18-month outcomes. Ongoing studies are examining the use of NIRS both in guiding management of blood pressure and cerebral autoregulation as an neuroprotective strategy and as a prognostic tool and in trending a number of NIRS parameters as a prognostic tool. ^,

Near-Infrared Spectroscopy Is a Noninvasive Technology for Assessing Cerebral Hemodynamics.

The probe emits light at a particular wavelength that passes through the scalp, skull, and brain tissue up to 1 to 2 cm. This light is absorbed by receiving optodes at two different wavelengths reflecting oxygenated and deoxygenated hemoglobin. Regional cerebral tissue oxygen saturation is calculated by the ratio between oxygenated and total hemoglobin. In some cases, cerebral saturations (labeled by the letter C ) may be compared with a control tissue (in this case renal, labeled by the letter R ). This is only appropriate when injury of the kidney or the control organ is minimal. Normal cerebral saturation ranges are between 65% and 75%. High cerebral oxygen saturation suggests decreased brain oxygen extraction but also occurs during therapeutic hypothermia because the brain metabolism is decreased. In contrast, low cerebral saturation suggests low brain perfusion.

Imaging

Head Ultrasound

Although much of the literature regarding imaging in HIE has focused on MRI, head ultrasound has been shown to be a useful tool for prognosis and evaluation of severity. Signs of HI injury that may be detectable on head ultrasound include small, effaced ventricles, which may be indicative of cerebral edema (although they can be a normal variant); echogenicity of the basal ganglia and thalamus can also be a sign of ischemia. Doppler measurements of the anterior cerebral artery resistive index have been associated with 2-year neurodevelopmental outcomes. ^, The ratio of white matter to gray matter echogenicity has also been described as a prognostic tool ( Figs. 47.6 and 47.8 ).

Full-Term Infant With Severe Hypoxic-Ischemic Encephalopathy (HIE) Due to Placental Abruption.

(A) Head ultrasound obtained at day-of-life (DOL) 1 demonstrates bilateral symmetric increased echogenicity of the white matter ( 1 ) with pronounced gray/white matter differentiation ( 2 ) and increased echogenicity of the basal ganglia and thalamus ( 3 ). At DOL 4, the above-described findings are more conspicuous and there is reduced resistive index ( b versus a ). (B) Brain magnetic resonance imaging at DOL 9 demonstrates diffuse increased T2 signal of the white matter ( 1 ), restricted diffusion and decreased T2 signal in the bilateral basal ganglia and thalami ( 2 ), and Rolandic gyri, mesial occipital cortices, insular and temporal polar cortices, and subcortical white matter. Prolonged pseudonormalization can be seen in neonates treated with therapeutic hypothermia with severe HIE (NICHD score of 3). (Courtesy Dr. Aylin Tekes, Director, Pediatric Radiology and Pediatric Neuroradiology, Johns Hopkins Hospital.)

Magnetic Resonance Imaging

When available, MRI is the gold standard imaging technique in neonates with HIE. A specific MRI protocol for neonates is important and should be discussed with pediatric neuroradiology. Depending on the timing of imaging after injury, findings will be present on different sequences. Restricted diffusion on diffusion weighted imaging (DWI) will be present for the first 3 to 5 days after injury, with the exact timing dependent on the severity of injury. Evolution of the MRI findings occurs over the first several weeks; nearing the end of the first week, pseudonormalization may occur. Pseudonormalization of the MRI is when the diffusion weighted changes are no longer visible and injury on T1/T2 sequences are not yet apparent. ^, There are two major patterns of HI injury, depending on the length and severity of the insult. Acute, more complete compromise results in basal ganglia-thalamic involvement, and more prolonged, less severe compromise is associated with a watershed/cortical injury pattern. In severe, prolonged hypoxic-ischemic events, extensive HI injury, which affects both the deep gray matter and the cortex, may be present. EEG and MRI may be useful in preterm infants to assess long-term risk for neurodevelopmental deficits.

Certain findings and scoring systems on MRI in HIE have been predictive of outcome. One highly predictive finding on MRI is abnormalities of the apparent diffusion coefficient in the posterior limb of the internal capsule, which is highly associated with survival and neuromotor outcome in infants with HIE. One MRI scoring system, specifically the grey matter injury score, is highly predictive of adverse motor and cognitive outcomes at age 2 and at school age. Conversely, absence of injury on MRI is not necessarily predictive of a favorable outcome, because infants with minimal or no injury on MRI may still have moderate to severe delays (see Figs. 47.6 , 47.7, and 47.8 ).

Hypoxic-Ischemic Brain Injury in Term Infant.

Infant with history of profound and brief anoxia A. An 8-day-old full-term infant with a classic central pattern of injury; note the bilateral symmetric high T2 and T1 signal in the posterior aspect of the putamen, posterior limb of the internal capsule, and ventrolateral thalami (arrows) . At day-of-life (DOL) 8, the diffusion tracer image (diffusion weighted imaging [DWI]) no longer demonstrates high signal due to pseudonormalization (no restricted diffusion). (Courtesy Dr. Aylin Tekes, Director, Pediatric Radiology and Pediatric Neuroradiology, Johns Hopkins Hospital.)

Patterns of Hypoxic-Ischemic Brain Injury in Preterm Infants.

A male infant of 26 4/7 weeks’ gestational age with OEIS complex (omphalocele-exstrophy-imperforate anus-spinal defects) and severe congenital anemia. (A) Head ultrasound at day-of-life (DOL) 0 demonstrates bilateral symmetric increased echogenicity of the white matter ( 1 ) and the basal ganglia and thalami ( 2 ). Transmastoid views demonstrate bilateral focal hyperechogenicities consistent with cerebellar hemorrhages ( 3 ). (B) Infant at 2 months of age: brain magnetic resonance imaging showed marked periventricular white matter volume loss ( 1 ), severe cystic encephalomalacia in the residual periventricular white matter, and cystic necrosis in basal ganglia. Note microhemorrhages in the periventricular white matter as demonstrated in the susceptibility weighted image (SWI) ( 2 ). Bilateral cerebellar hemorrhages are replaced by volume loss and cystic encephalomalacia and hemosiderin straining ( 3 ). (Courtesy Dr. Aylin Tekes, Director, Pediatric Radiology, Johns Hopkins Hospital.)

Biomarkers

The search for a reliable, testable biomarker for outcome in HIE is the topic of ongoing research. A number of serum biomarkers have undergone investigation, with mixed prognostic utility. Those with the greatest evidence are summarized in Table 47.4 . Ongoing trials are examining various biomarkers for feasibility of testing, specificity, and sensitivity for long-term neurodevelopmental outcomes (BiHiVE2 NCT 02019147; BANON study NCT03357250).

Seizure Control

Status epilepticus, seizures lasting >5 minutes, or more than two 30-second seizures per hour are accepted guidelines to start antiseizure drugs, particularly when encephalopathy is secondary to stroke. Phenobarbital is the most common agent used to control seizures in this setting, but evidence guiding the duration and the combination with other antiseizure drugs is limited. One of the few points of agreement among experts is that the duration of phenobarbital (and phenytoin) treatments should be as brief as possible after controlling the initial seizures, due to the suspected harmful effects on the developing brain. Between 45% and 70% of clinical seizures respond to phenobarbital (loading of 20 mg/kg), and its metabolism is grossly unchanged by TH. In rodent models of neonatal HI injury, the addition of phenobarbital to TH may provide additional protection concerning the early extent of the injury, although it does not prevent motor impairments. This lack of motor protection may be because phenobarbital, along with valproate and phenytoin, produces increased neuronal loss, even without a superimposed brain insult. Thus, other alternatives such as levetiracetam and topiramate, which do not have those effects, are being explored in the developing brain.

Shivering Control

Shivering exacerbates brain injury in large animal models of neonatal HI brain injury, likely due to increased glucose metabolism. Morphine is often used to control shivering in the setting of NE in many neonatal intensive care units, following the protocols used in large RCTs. However, opiates do not specifically target the mechanism of shivering and have significant adverse effects including respiratory depression, hypotension, and later withdrawal, effects that are potentiated due to impaired clearance during TH after an HI insult. α-2-Agonist agents such as clonidine or dexmedetomidine control shivering and thus may be potential adjuvant therapies to TH. ^, Although studies are limited, treatment with clonidine has been shown to decrease brain injury after neonatal HI in rodents. Similarly, neonatal clonidine treatment delays the development of epilepsy kindling in rodent models. One phase 1/2 RCT studying clonidine in NE has been completed supporting the safety of intravenous (IV) clonidine up to 1 μg/kg/dose every 8 hours in neonates receiving TH, with decreased need for morphine PRN (NCT01862250), and a second RCT is still enrolling (SANNI project, NCT03177980).

A second α-2-agonist agent under evaluation is dexmedetomidine. The results in different rodent models of perinatal brain injury and analysis of neurodegeneration suggest dexmedetomidine has neuroprotective properties, with an unknown link to the control of shivering. However, in a piglet model of neonatal HI, dexmedetomidine at clinically relevant doses (loading of 2 μg/kg at 10 minutes followed by 0.028 μg/kg/h for 48 hours) and plasma levels (within 1 μg/L) in combination with 48 hours of TH produced significant cardiovascular instability, increased mortality, and worsened neuronal death. Additionally, in a case report, dexmedetomidine (without TH) was temporally associated with development of seizures in an infant diagnosed with NE. An early phase 1 clinical study assessing the pharmacokinetics of dexmedetomidine (the Cool DEX study; NCT02529202) demonstrated slower rise in dexmedetomidine in patients receiving TH, with a good safety profile up to a dose of 0.4 μg/k/h. Until more safety data are available, caution must be exercised in the routine use of dexmedetomidine and clonidine.