Hypoxic-Ischemic Injury in the Term Infant: Clinical-Neurological Features, Diagnosis, Imaging, Management, and Prognosis

The clinical aspects of neonatal hypoxic-ischemic encephalopathy (HIE) are appropriately discussed after the neuropathology (see Chapter 22 ) and pathophysiology (see Chapter 23 ), because understanding of the clinical phenomena is facilitated greatly by an awareness of the underlying pathological substrates. Moreover, choice of appropriate diagnostic modalities, formulation of rational prognostic statements, and development of appropriate plans of management are based, in many ways, on awareness of the probable neuropathologies. In this chapter, we discuss the clinical settings for neonatal HIE, the clinical syndrome, diagnostic studies, clinical correlations, management, and prognosis.

CLINICAL SETTINGS

The importance of the early recognition of the clinical risk factors for hypoxic-ischemic cerebral injury in the term-born infant has escalated significantly in the last two decades with the implementation of successful neuroprotection with therapeutic hypothermia (TH) (see later). To initiate such therapy requires a recognition of the infant who may have suffered hypoxic-ischemic cerebral injury—predominantly in the peripartum period. The peripartum period is defined as the period shortly before, during, and immediately after birth . The clinical settings for neonatal HIE are dominated by the ultimate occurrence of ischemia (i.e., diminished blood supply to brain), usually, but not necessarily, preceded or accompanied by hypoxemia (i.e., a diminished amount of oxygen in the blood supply). Hypoxemia leads to brain injury principally by causing myocardial disturbance and loss of cerebrovascular autoregulation, with ischemia the major consequence. The temporal characteristics and the severity of the hypoxemia and ischemia, as well as the gestational age (GA) of the infant, are the principal determinants of the type of resulting neuropathology (see Chapters 18 and 22 ).

The major causes of serious hypoxemia in the peripartum period are (1) hypoxia-ischemia with intrauterine disturbance of gas exchange across the placenta (i.e., asphyxia) or with failure to establish independent respiration at the time of birth or both; (2) postnatal respiratory insufficiency secondary to severe respiratory disease; and (3) severe right-to-left shunt secondary to persistent fetal circulation or cardiac disease. The major causes of serious ischemia are (1) intrauterine asphyxia (i.e., hypoxemia, hypercarbia, and acidosis) with cardiac insufficiency and loss of cerebrovascular autoregulation both in utero and at the time of birth; (2) postnatal cardiac insufficiency secondary to severe hypoxemia or congenital heart disease; and (3) postnatal (postcardiac) circulatory insufficiency secondary to patent ductus arteriosus (with “ductal steal”) or vascular collapse (e.g., with sepsis).

In this chapter, we will focus on the encephalopathic syndrome in the term infant and the presence of evidence for peripartum hypoxic-ischemic injury. We prefer the term “peripartum” hypoxic-ischemic injury as it acknowledges the potential presence of (1) fetal or maternal prepartum conditions that may accentuate propensity to intrapartum hypoxic-ischemic injury; (2) intrapartum hypoxic-ischemic injury per se; and (3) the often-associated protracted postpartum resuscitative efforts for such infants, with no or low heart rate for several minutes. Although the postnatal cardiac compromise is not the primary etiology of the poor birth transition, it may contribute to the extent of the ultimate hypoxic-ischemic cerebral injury in the infant. However, there is also evidence that even with prolonged resuscitation, in the era of TH there can be better outcomes than might be expected (see ‘Prognosis’ section).

It is important to recognize that not all neonatal encephalopathies are related to hypoxic-ischemic disease. Antepartum and postpartum disorders (e.g., infectious, metabolic, dysgenetic) may lead to neonatal encephalopathies, as discussed throughout this book. Reassuringly, a number of studies have reported recent reductions of neonatal encephalopathy, including those related to hypoxic-ischemic injury at 0.5 to 1.5/1000 live births. However, the latter still remains the principal cause with fully 56% of all cases of newborn encephalopathy relating to hypoxic-ischemic injury that occurs during the intrapartum period. The findings just noted are consistent with a large cohort study of 4165 singleton term infants with any one of the following: seizures, stupor, coma, Apgar score at 5 minutes less than 3, and/or receiving hypothermia therapy. In this study, 15% of the infants experienced a clinically recognized sentinel event, such as antenatal hemorrhage (presumably, often placental abruption), uterine rupture, or cord prolapse, all of which are capable of compromising oxygen supply. Although intrapartum sentinel events do provide clear evidence of a hypoxic-ischemic insult, four additional studies found that sentinel intrapartum events were only identified in 8% to 36% of infants. Thus it can be challenging to confirm a hypoxic-ischemic etiology for the infant with neonatal encephalopathy and/or the need for resuscitation as only 10% to 30% of such infants may have a clinical history of an overt intrapartum risk factor, whereas approximately 50% or more may have a constellation of risk factors, including maternal history, cord acidemia, and the need for resuscitation that support hypoxia-ischemia as the most likely etiology for their neurological syndrome. Of note, signs of inflammation were also not uncommon, with 27% of mothers displaying elevated maternal temperature in labor and 11% clinical chorioamnionitis. However, the contributing role of chorioamnionitis is not consistently supported. A combination of both acute and chronic placental abnormalities may be more common in infants with HIE, as seen in a multicenter cohort study, underscoring the complex causal pathways. Finally, in relation to acute sentinel events it has been noted that there is an association of uterine rupture with previous cesarean deliveries, that may provide a potential opportunity for obstetrical management in reducing the incidence of HIE.

In addition, although obvious, hypoxic-ischemic injury may affect the infant’s brain during the antepartum and postnatal periods, albeit much less commonly than the intrapartum period . On the basis of earlier work, approximately 20% of hypoxic-ischemic injury recognized in the newborn period was said to be related primarily to antepartum insults. These data should be interpreted with the awareness that assessment of timing of insults to the fetus in these reports generally was based on imprecise methods, and the variability of findings is considerable. Moreover, more recent studies with consistent use of magnetic resonance imaging (MRI) suggest that the large majority of hypoxic-ischemic injury occurs during the intrapartum period.

The best data indicate that most infants with neonatal HIE and intrapartum evidence of hypoxic-ischemic insult exhibit, on MRI, evidence only of injury from the immediate peripartum period with no clear evidence of long-standing antenatal hypoxic-ischemic disease ( Table 24.1 ).

Table 24.1

Timing of Insults Leading to Hypoxic-Ischemic Encephalopathy

Of 245 infants who had an MRI scan after neonatal neurological signs (“neonatal encephalopathy”), and evidence of intrapartum perinatal asphyxia:
197 (80%) had MRI evidence of acute peripartum lesions consistent with hypoxic-ischemic insult; only 8 (4%) also had MRI evidence of antenatal injury
40 (16%) had normal MRI scans
8 (4%) had other disorders (e.g., neuromuscular or metabolic diseases)

MRI , Magnetic resonance imaging.

For example, in one study of 245 term infants with neonatal encephalopathy and evidence of intrauterine asphyxia, fully 80% had evidence of acute lesions (within the period immediately before or during labor and delivery) consistent with hypoxic-ischemic disease, 16% had normal MRI scans, and only 4% had concomitant evidence of chronic antenatal injury (see Table 24.1 ). In another MRI study of 173 term newborns with encephalopathy and signs of intrauterine asphyxia, only acute injury was observed. Related clinical and epidemiological data also support a marked preponderance of intrapartum events in the origin of neonatal HIE, especially in the term infant.

The principal intrapartum events leading to hypoxic-ischemic fetal insults include acute placental or umbilical cord disturbances, such as abruptio placentae or cord prolapse, prolonged labor with transverse arrest, difficult forceps extractions, or rotational maneuvers (see Chapter 21, Chapter 22, Chapter 23 ). Postpartum events alone (e.g., severe persistent fetal circulation, severe recurrent apneic spells, cardiac failure secondary to large patent ductus arteriosus or other congenital heart disease, severe pulmonary disease) may lead to HIE and may account for approximately 5% to 10% of cases. Most of these and related postnatal factors are much more important in the pathogenesis of hypoxic-ischemic brain injury in the premature infant than in the term infant (see Chapters 18 and 20 ). Although hypoxic-ischemic injury certainly can occur in the antepartum period (e.g., secondary to maternal trauma, maternal hypotension, uterine hemorrhage), this injury cumulatively may account for less than 20% of neonatal HIE. However, antepartum factors appear to be of some contributory importance in the risk for neonatal encephalopathy related to peripartum events. Such factors may indeed predispose to intrapartum hypoxia-ischemia during the stresses of labor and delivery, especially through threats to placental flow. Such factors include maternal diabetes, preeclampsia, placental vasculopathy, intrauterine growth restriction, and twin gestation that may compromise fetal cerebral perfusion ( Table 24.2 ; see Chapter 16 ). In one series, such factors were present in approximately one-third of cases of intrapartum asphyxia. Indeed, “perinatal asphyxia” was identified in 27% of infants of diabetic mothers, and its occurrence correlated closely with diabetic vasculopathy (nephropathy) and presumed placental vascular insufficiency. In a more recent cohort of infants that received TH ( n = 98), the frequency of pregestational diabetes and preeclampsia was significantly higher (threefold to fivefold) in women with infants requiring TH.

Table 24.2

Antepartum/Maternal Clinical Factors Associated With Neonatal Encephalopathy

		FREQUENCY IN GENERAL POPULATION (%)	FREQUENCY IN HIE POPULATION (%)
Antepartum/Maternal	Hypothyroidism	0.5	3
	Obesity	10–25	15–50 ^a
	Diabetes (particularly pregestational)	0.5–2	5–20 ^a
	Fetal growth restriction <5%	5	10–15
	Hypertension	3–5	5–15
	Clinical chorioamnionitis	1–4	5–10

HIE , Hypoxic-ischemic encephalopathy.

a Dependent on geographical cohort.

Regarding intrauterine growth restriction , in the largest North American series of neonatal encephalopathy, collected by the Vermont Oxford Registry, 16% of infants were defined as less than the 10th percentile for weight. In one of the larger regional population studies of neonatal encephalopathy, 16% of infants with neonatal encephalopathy were growth-restricted compared with only 1.2% of term infants without encephalopathy. Growth restriction was the strongest predictor of neonatal encephalopathy examined, associated with a 30-fold increase in risk. The additional stress of labor may be expected to compromise placental blood flow. Further, in a fetus that already has impaired placental function, impaired placental function and an increased risk of perinatal asphyxia in the infant with intrauterine growth restriction are recognized and appear to account for some of the increased risk of subsequent neurological disability in such infants. In a recent series of infants receiving hypothermia, infant birthweight below the 5th percentile for GA was significantly associated with the need for TH. Other factors (e.g., dysmorphic syndromes, severe undernutrition, infection) may also lead to increased risk of neurological disability in intrauterine growth restriction. Moreover, studies in fetal and neonatal animals suggest that the mechanisms for the increased vulnerability of the growth-restricted fetus relate not only to placental insufficiency but also to diminished glucose reserves in the heart, liver, and brain, and to impaired capability to increase substrate supply to the brain with the hypoxic stress of vaginal delivery.

Other less characterized maternal factors have been recognized as important risk factors for neonatal encephalopathy, although pathogenetic mechanisms remain unclear. One such factor is maternal hypothyroidism. In four prospective studies, an elevated risk of up to 10-fold was found for neonatal encephalopathy among infants born to mothers with hypothyroidism (see Table 24.2 ). Further, maternal drug use can impair the transition of infants after delivery, and infants can display abnormal neurological signs that can mimic neonatal encephalopathy. The details of these agents are outlined in Chapter 42 .

Although the particular importance of intrauterine hypoxic-ischemic injury, especially intrapartum asphyxia with or without antepartum predisposing factors, in the genesis of the clinical syndrome of neonatal HIE is apparent, most infants who experience intrapartum hypoxic-ischemic insults do not exhibit overt neonatal neurological features or subsequent neurological evidence of brain injury. The severity and duration of the hypoxic-ischemic insult is obviously critical. The elegant studies of Low and others demonstrate a striking relationship among the severity and duration of intrapartum hypoxia, assessed by the use of fetal acid-base studies (see Chapter 21 ), the subsequent occurrence of a neonatal neurological syndrome, and later neurological deficits. Current data suggest that approximately 10% of all term deliveries require some resuscitation with 1% requiring extensive resuscitation. Of the latter, only 1 to 3 per 1000 will develop signs of evolving encephalopathy consistent with HIE.

NEUROLOGICAL SYNDROME

The neurological syndrome that accompanies serious peripartum hypoxic-ischemic cerebral injury is the prototype for neonatal HIE. In considering the nature and timing of hypoxia-ischemia as the etiology of neonatal HIE , we consider three features to be important: (1) evidence of fetal distress and/or fetal risk for hypoxia-ischemia (e.g., fetal heart rate [FHR] abnormalities, sentinel event, fetal acidemia); (2) the need for resuscitation and/or low Apgar scores; and (3) an overt neonatal neurological syndrome in the first hours or day of life.

Defining the neurological syndrome of HIE by clinical evaluation is important. Central to this definition is awareness of the characteristics of the normal neurological examination (see Chapter 12 ). The abnormal features of the examination in infants with HIE were discussed in previous editions of this book and in Chapter 12 . To improve interobserver reliability, standardized scores have been developed, and have proven useful in large-scale clinical research studies (see later) ( Table 24.3 ).

Table 24.3

Standardized Scoring Systems for Neonatal Encephalopathy

SCORING SYSTEM	PURPOSE/UTILITY	NUMBER OF ELEMENTS	ELEMENTS	EEG NECESSARY
Sarnat	Prognosis applied in first 7 days	14	Alertness, tone, posture, reflexes, myoclonus, suck, Moro, oculovestibular, tonic neck, pupils, heart rate, secretions, GI motility, seizures, EEG	Yes
Modified Sarnat	Prognosis applied in first 7 days	5	Alertness, tone, suck, Moro, seizures	No
Thompson	Prognosis applied in first 7 days	6	Alertness, tone, respiratory status, reflexes, seizure, feeding method	No
NICHD	Selection in first 6 h of life of moderate-severe NE for hypothermia	9	Alertness, spontaneous activity, posture, tone, suck, Moro, pupils, heart rate, respirations	No
SIBEN	Defining mild, moderate, and severe NE in first 6 h of life	10	Alertness, spontaneous activity, posture, tone, suck, Moro, pupils, heart rate, respirations, seizures	No

EEG , Electroencephalography; GI , gastrointestinal; NE , neonatal encephalopathy; NICHD , National Institute of Child Health and Human Development; SIBEN , Score of the Iberoamerican Society of Neonatology.

The initial neurological examination classification systems developed evaluated infants over the first week of life to define the severity of their encephalopathy for prognostication. The first of these scoring systems was that developed by Sarnat, which was based on serial examinations of 21 term-born infants over the first few weeks of life. Three clinical stages of “postanoxic encephalopathy” were described. Stage 1 lasted less than 24 hours and was characterized by hyperalertness, uninhibited Moro and stretch reflexes, sympathetic effects, and a normal electroencephalogram. Stage 2 was marked by obtundation, hypotonia, strong distal flexion, and multifocal seizures. The electroencephalography (EEG) showed a periodic pattern sometimes preceded by continuous delta activity. Infants in stage 3 were stuporous, flaccid, and their brainstem and autonomic functions were suppressed. The EEG was isopotential or had infrequent periodic discharges. Infants who did not enter stage 3 and who had signs of stage 2 for less than 5 days appeared normal in later infancy. Persistence of stage 2 for more than 7 days, or failure of the EEG to revert to normal, was associated with later neurological impairment or death. This classification system was then further simplified, to be known as the Modified Sarnat score (see Table 24.3 ).

The next scoring system developed for prognostication in neonatal encephalopathy was the Thompson Encephalopathy Score, developed in 1997 (see Table 24.3 ). This scoring system was simpler to apply and did not require EEG to increase its widespread applicability. The initial evaluation showed a good correlation between the maximal score in the first 7 days of life and neurodevelopmental outcome at 18 months in 44 infants with neonatal HIE.

It is important to note that both the Sarnat and the Thompson scoring systems aimed to define neonatal neurological signs during the first week of life to improve the prediction of subsequent neurological deficits. However, as the era of neuroprotection emerged, it became apparent that a standardized neonatal neurological examination tool to be applied in the first few hours of life would be necessary to define eligibility for randomized controlled trials (RCTs), such as TH. To define encephalopathy eligibility for clinical trials, a modified examination (Modified Sarnat Staging) was developed by the National Institute of Child Health and Human Development (NICHD) trial to select those with moderate to severe encephalopathy at less than 6 hours. This modified staging system clustered nine neurological signs into six categories: consciousness, spontaneous activity, posture, tone, primitive reflexes, and autonomic system. To qualify as moderate or severe, an infant needed to have moderate or severe abnormalities in at least three of the exam items or have seizures. In a similar fashion, the CoolCap trial adopted another “modified” Sarnat staging that relied on abnormal consciousness, with one or more of hypotonia, abnormal reflexes, abnormal suck, or clinical evidence of seizures (in addition to use of amplitude-integrated EEG [aEEG]). After a recent study of infants with milder forms of neonatal encephalopathy demonstrated increased risk of disability up to 16% in those identified as mild encephalopathy, a Total Sarnat Score was developed, which adds up numeric values for these six categories to give a single score ranging from 0 to 18. In this study of mild encephalopathy (PRIME study), all infants diagnosed with disability at 18 to 22 months of age had a Total Sarnat Score of 5 or more at less than or equal to 6 hours of life. A score of greater than or equal to 5 predicted disabilities with good accuracy (area under the curve = 0.83, P = 0.004, sensitivity 100%, and specificity 67%).

The HIE Score of the Iberoamerican Society of Neonatology (SIBEN) was developed in 2016 and involved the assessment of 10 clinical aspects that could be undertaken immediately after delivery room resuscitation (see Tables 24.3 and 24.4 ). This was the first scoring system to allow for the recognition of mild encephalopathy, which was recognized to be of great relevance because at least 40% of hypoxic-ischemic cerebral injury presents as mild disease (see later). To classify HIE as mild, moderate, or severe, each item evaluated varies according to the degree of severity (see Table 24.4 ). With this scoring system, a point is given to every item that corresponds to a level in the SIBEN score, with the diagnosis of HIE considered in the presence of 3 points or more. In a recent study, when the Sarnat Total Score was compared with the SIBEN score in 145 infants with mild, moderate, and severe encephalopathy, both scores were comparable and superior to a more simplified grading system in predicting brain injury on MRI.

Table 24.4

Score of the Iberoamerican Society of Neonatology Neurological Score

	MILD	MODERATE	SEVERE
Level of consciousness	Hyperalert	Lethargy	Stupor/coma
Spontaneous activity	Normal	Decreased	Not present
Posture	Mild distal flexion	Marked distal flexion	Decerebrate
Tone	Normal	Hypotonia	Flaccidity
Suck	Weak	Weak or absent	Not present
Moro	Strong	Weak	Not present
Pupils	Mydriasis	Miosis	Diverted/nonreactive
Heart rate (HR)	Tachycardia	Bradycardia	Lack of HR variability
Breathing	Spontaneous	Periodic	Apnea
Seizures	Absent	Present—frequent	Present—infrequent

Temporal Evolution of the Neurological Syndrome

The temporal evolution of the neurological syndrome in the era before TH has been outlined in detail in previous editions of this book. Although it is more complex in the infant undergoing TH because of sedation and response to hypothermia, the principles remain unchanged.

With less severe forms of HIE , often characterized as mild or moderate encephalopathy, changes in the clinical syndrome may be less stereotyped, and thus careful serial clinical evaluation every hour over the first 6 to 12 hours of life is more sensitive to the evolution. In the first hours after insult , these infants may be hyperalert or irritable (e.g., difficult to console, jittery) with normal to slightly abnormal posture and decreased spontaneous activity ( Table 24.5 ). Tone is often preserved, although earlier increased tone may be found distally and evident as cortical thumb flexion. On the contrary, appendicular tone during this time tends to be lower and evolve into increased tone over the coming days to weeks. Autonomic function is typically undisturbed, although respiratory function may be impaired. Primitive reflexes may be weak or incomplete in moderately affected infants.

Table 24.5

Clinical Features of Hypoxic-Ischemic Encephalopathy: Birth to 12 Hours

Mild to moderate HIE	Hyperalert or irritable Decreased spontaneous activity Some disturbance of respiratory function Hypotonia with distal hypertonia (i.e., cortical fisting) Seizures
Severe HIE	Depressed level of consciousness: usually deep stupor or coma Ventilatory disturbance: “periodic” breathing or respiratory failure Intact pupillary responses Intact oculomotor responses Hypotonia, minimal movement > hypertonia Seizures

HIE , Hypoxic-ischemic encephalopathy.

The emergence of seizure activity is an important marker of the evolution of encephalopathy related to HIE, occurring in those with less severe forms of HIE as well as those with severe HIE. In the modern era of TH for HIE, seizure onset is usually during the first 48 hours but sometimes occurs during rewarming or even after return to normothermia. An early study of 26 consecutive newborn infants with HIE undergoing TH, continuous EEG (CEEG) identified electrographic seizures in 65% of infants with onset in the first 48 hours for 76% of patients. In another recent retrospective study of 114 term neonates with moderate to severe HIE undergoing TH and CEEG, seizures were identified in 49% of infants with seizure onset being in the initial 24 hours, 24 to 48 hours, or greater than 72 hours in 88%, 7%, and 5% of the infants, respectively. Infants with normal or mildly abnormal EEG background constituted 40% of the cohort, and those infants who experienced seizures always had their seizure onset within the initial 24 hours of CEEG. All of the infants with first seizures beyond 24 hours had moderately or markedly abnormal early EEG backgrounds. A larger study of 331 term infants with HIE undergoing TH included 216 infants with normal or mildly abnormal EEG background. Only 6% of the infants with a normal or mildly abnormal EEG experienced seizures, and all occurred in the initial 12 hours of CEEG, which corresponded to the initial 24 hours of life. Finally, a three-center retrospective study of 90 infants with HIE managed with TH indicated that 48% experienced electrographic seizures, including 10% with electrographic status epilepticus. Seizures occurred at a median of 20 hours from CEEG initiation. An abnormal initial EEG background classification (excessively discontinuous, depressed and undifferentiated, burst suppression, or extremely low voltage) was associated with seizure occurrence. Compared with a normal initial EEG background, the risk ratios (RRs) of seizures were 18-fold and 13-fold higher for infants with excessively discontinuous and severely abnormal initial EEG backgrounds, respectively. These models indicate that the initial 24 hours and the period during TH carry the highest risk for seizures . Based on these data, the Newborn Brain Society Guideline Committee recommended that infants with encephalopathy undergoing hypothermia with a normal background in the initial 24 hours have a very low risk of developing seizures. However, it is important to note that there can be some variability in the timing of onset, including some with seizure onset late during hypothermia or during rewarming, particularly in infants with an abnormal background in their CEEG as noted above.

As noted in Chapter 15 , there can be significant challenges in the correct clinical recognition of seizures. In one recent study of staff observing high-risk newborns, only 9% of 526 electrographic seizures were identified by clinical observation, indicating an underdiagnosis of seizures . In addition, 78% of 177 nonictal events were incorrectly identified as seizures, indicating an overdiagnosis of seizures.

From approximately 12 to 24 hours , the infant’s level of consciousness changes in a variable manner ( Table 24.6 ). Infants with less severe disease may often begin to exhibit some degree of improvement in alertness, and others may become “hyperalert.” A report from Biselele and colleagues that serially examined 21 infants with evidence of a hypoxic-ischemic insult documented that in the first hour of life, approximately 70% of the infants displayed significantly abnormal neurological signs, with a score of greater than 7 on the Thompson scale (see earlier), thereby permitting entry into TH, whereas at 6 hours only 20% of infants scored at this level. The authors warned that this “apparent” improvement may prevent infants that need to receive TH from being eligible, if their examination was delayed. Thus it is regarded that in many of these infants, this improvement is more apparent than real , because the appearance of alertness may not be accompanied by visual fixation or following, habituation to sensory stimulation, or other signs of intact cerebral function. The notion of apparent rather than real improvement in such cases is further supported by the occurrence at this time of seizures, apneic spells, jitteriness, and weakness. Apneic spells appear in approximately 50% to 65% of infants. Jitteriness develops in about one-fourth of infants and may be so marked that the movements are mistaken for seizures. Distinction can usually be made at the bedside (see Chapter 12 ).

Table 24.6

Clinical Features of Hypoxic-Ischemic Encephalopathy: 12 to 24 Hours

Mild to moderate HIE	Hyperalert (irritable or jittery) More seizures Apneic spells Evolving hypertonia
Severe HIE	Variable change in level of alertness More seizures Apneic spells Jitteriness Weakness Proximal limbs, upper > lower (full term) Hemiparesis (full term) Lower limbs (premature)

HIE , Hypoxic-ischemic encephalopathy.

Between approximately 24 and 72 hours , the level of consciousness continues to change and is influenced by degree of sedation, presence of seizures, and any antiseizure medications administered ( Table 24.7 ).

Table 24.7

Clinical Features of Hypoxic-Ischemic Encephalopathy: 24 to 72 Hours

Mild to moderate HIE	Variable change in level of alertness Hypertonia
Severe HIE	Stupor or coma Respiratory arrest Brainstem oculomotor and pupillary disturbances

HIE , Hypoxic-ischemic encephalopathy.

Beyond 72 hours , the level of consciousness improves, often dramatically. However, disturbances of feeding are extremely common and relate to abnormalities of sucking, swallowing, and tongue movements ( Table 24.8 ). The power and coordination of the muscles involved (innervated by cranial nerves V, VII, IX, X, and XII) are deranged. In the large series studied by Brown and coworkers, 80% of infants required early tube feedings because of feeding difficulty.

Table 24.8

Clinical Features of Hypoxic-Ischemic Encephalopathy: After 72 Hours

Mild to moderate HIE	Improving level of alertness Increased spontaneous activity Disturbed suck, swallow, and tongue movements Tone normalization with some residual truncal hypotonia
Severe HIE	Persistent, yet diminishing stupor Disturbed sucking, swallowing, gag, and tongue movements Hypotonia > hypertonia Weakness Proximal limbs, upper > lower Hemiparesis

HIE , Hypoxic-ischemic encephalopathy.

With regard to the more severe form of neonatal encephalopathy, occurring in 10% to 20% of infants with HIE, a clear evolution occurs. In the first hours after insult , the severely affected infant is either deeply stuporous or in coma (i.e., not arousable and minimal, or no response to sensory input). Periodic breathing, or respiratory irregularity akin to this pattern, is prominent, which may be considered as a form of respiratory disturbance as a neonatal counterpart of Cheyne-Stokes respiration, which is observed in older children and adults with bilateral hemispheric disease. In one series, approximately 80% of infants with severe neonatal encephalopathy had abnormal breathing patterns, particularly periodic breathing. Those most severely affected may exhibit marked hypoventilation or apnea. Pupillary responses to light are intact, spontaneous eye movements are present, and eye movements with the oculocephalic response (doll’s eyes maneuver) are usually full. (Pupillary size is variable, although dilated reactive pupils tend to predominate in the less affected infants, and constricted reactive pupils are common in the more severely affected infants. ) Commonly, dysconjugate eye movements are apparent. However, only in a few babies are eye signs of major brainstem disturbance seen. Fixed, midposition, or dilated pupils and eye movements fixed to the doll’s eyes maneuver and to cold caloric stimulation are unusual at this stage. If either of these signs is evident at this time, especially in the full-term infant, injury to the brainstem is likely. Most infants at this stage are markedly and diffusely hypotonic with minimal spontaneous or elicited movement.

From approximately 12 to 24 hours , these severely affected infants remain deeply stuporous or in a coma ( Table 24.6 ). Infants with involvement of basal ganglia may exhibit an increase in their tone, especially in response to handling. Many infants manifest definite, albeit not marked, weakness (see Table 24.6 ). Although precise correlation is often difficult, these infants appear from the clinical circumstances surrounding their insult to have sustained particularly marked ischemic insults. The weakness is often most prominent in the hip–shoulder distribution, with more impressive involvement usually of the proximal extremities. Distinct asymmetry of these latter motor findings is unusual to elicit early in the neonatal period, although a few full-term infants do exhibit weakness that is confined to or is clearly more severe on one side than on the other, especially in the setting of stroke.

Between approximately 24 and 72 hours , the severely affected infant’s level of consciousness often deteriorates further, and deep stupor or coma may ensue ( Table 24.7 ). Respiratory arrest may occur, often after a period of irregularly irregular (“ataxic”) respirations. Brainstem oculomotor disturbances are now more common. These usually consist of skew deviation and loss of responsiveness of the eyes to the doll’s eyes maneuver and to cold caloric stimulation. Rarely, ocular bobbing may appear. Pupils may become fixed to light in the mid or dilated position. Reactive but constricted pupils are more common in less severely affected infants, especially if they are on opiates. Babies who die with HIE most often do so at this time, particularly if the criterion is “brain death.” In one large series of infants who died after perinatal asphyxia and HIE, the median age of death was 2 days. The cause for the apparent delay in progression to brain death until this period is not known definitely, but delayed cell death has been documented in in vivo models and in neurons in culture (see Chapters 22 and 23 ). The importance of excitatory amino acids, calcium-mediated deleterious metabolic events, and free radical production has been detailed in Chapters 16 and 23 . Indeed, studies by magnetic resonance spectroscopy (MRS) in the asphyxiated human infant (see later) have documented a delayed deterioration of cerebral energy state. However, although delayed cell death most probably accounts for this clinical deterioration, consideration should also be given to the occurrence of frequent subclinical electrical seizures as the reason for the deterioration. Although EEG is required for this determination, the potential effectiveness of anticonvulsant therapy warrants the procedure.

Infants who survive to greater than 72 hours will at this point usually improve over the next several days to weeks; however, certain neurological features persist ( Table 24.8 ). Many infants require tube feedings for weeks to months, particularly those with involvement of deep nuclear gray matter and brainstem (see ‘Prognosis’ section). Generalized hypotonia of limbs is common, although hypertonia, particularly with the passive manipulation of limbs, is frequent on careful examination, especially among infants with prominent involvement of basal ganglia. The patterns of weakness discussed in the previous section become more readily elicited, although the weakness is rarely marked. Feeding and suck-swallow difficulties are common with impaired hyporesponsive or excessive gag reflexes and risk for aspiration. There may be pooling of saliva with inability to swallow saliva.

Diagnosis

The recognition of neonatal HIE depends principally on information gained from a careful history and a thorough neurological examination. The contributing role of certain metabolic derangements requires evaluation. Determination of the site or sites and extent of the injury is made by the history, neurological examination, EEG, and neuroimaging studies (ultrasonography, MRI).

History

Recognition of neonatal HIE requires awareness of those intrauterine situations that account for most cases. Thus information should be sought regarding maternal disorders with an increased risk for peripartum HIE (see Table 24.2 ), such as those that could lead to uteroplacental insufficiency and disturbances of labor or delivery that could impair placental respiratory gas exchange or fetal blood flow. The value of fetal evaluation including electronic fetal monitoring, particularly when supplemented by fetal blood sampling to determine acid-base status, is discussed in Chapter 21 .

Neurological Examination

Recognition of the neurological signs outlined previously provides critical information concerning the presence, site, and extent of hypoxic-ischemic injury in the newborn infant. The neurological examination plays two critical roles. First, the systematic neurological examination in the first 6 hours of life may allow recognition by the clinician of the presence and severity of any neonatal HIE (see Chapter 12 ). This recognition is essential to allow that infant to be considered for eligibility for potentially neuroprotective strategies, such as TH. Second, the regular and systematic neurological examination of the infant with encephalopathy over the first week of life carries very important information for establishing a prognosis (see later).

Metabolic Parameters

Certain metabolic derangements may contribute significantly to the severity and qualitative aspects of the neurological syndrome, and the diagnostic evaluation should include evaluation of such derangements. Hypoglycemia, hyperammonemia, hypocalcemia, hyponatremia (inappropriate secretion of antidiuretic hormone [ADH]), hypoxemia, and acidosis are among the metabolic complications that may occur, often because of associated disorders, and that may exacerbate certain neurological features or add new ones.

Metabolic Acidemia

In 1958, James and colleagues recognized that umbilical cord blood gas analysis could reflect fetal hypoxia. Since that time, cord blood gas analysis has been widely performed as an objective measure of fetal condition at the time of delivery and when combined with other neonatal factors, can help identify infants at risk for neonatal encephalopathy. The American College of Obstetricians and Gynecologists and the American Academy of Pediatrics (AAP) now recommend umbilical cord blood gas analysis be performed in all high-risk deliveries.

Sampling umbilical artery blood is preferable to sampling umbilical vein blood as the arterial pH and base deficit provide the most accurate information on fetal acid-base status and correlate best with newborn morbidity. Ideally, the test is performed as soon as possible after delivery; however, pH, P o ₂ , and P co ₂ measured from clamped portions of the umbilical cord remain reliable for up to 60 minutes after birth. The arterial umbilical cord blood pH and base excess (or deficit) have been shown to be particularly useful values for interpreting fetal-neonatal condition and prognosis. Though cost-effectiveness for this practice has not been established, some clinicians prefer to obtain two samples, one from the artery and the other from the vein (considering the high sampling error rate of umbilical arterial gases).

If both vessels are sampled, the median arteriovenous pH difference is about 0.09 (range 0.02 to 0.49) with the mean umbilical cord arterial pH of 7.24 to 7.27, and the mean umbilical cord venous pH of 7.32 to 7.34.

A practical pH threshold commonly used for defining pathological fetal acidemia is umbilical artery pH less than 7.00 . However, this threshold has high variability as a predictor of neurological problems after birth and even when these infants are followed for 6.5 years. In a 2008 systematic review of studies of nonanomalous term births with pH less than 7.0, 51 of 297 (17.2%) survived with neonatal neurological morbidity, 45 of 276 (16.3%) had seizures, and 24 of 407 (5.9%) died during the neonatal period. In a 2010 meta-analysis of 51 cohort and case-control studies, including over 480,000 newborns, the strongest association between neonatal morbidity and pH was at a pH threshold of 7.0 (odds ratio [OR] 12.5, 95% confidence interval [CI] 6.1 to 25.6). For neonatal mortality, the strongest association was at pH 7.1 (OR 7.1, 95% CI, 3.3 to 15.3). Base deficit (negative base excess) and lactate levels are often used as more accurate markers of metabolic acidosis, as they may better reflect hypoxic-ischemic insults due to the lack of impact of secondary hypocarbia that may falsely reassure clinicians due to respiratory compensation to normalize the pH. An umbilical artery base deficit greater than or equal to 12 mmol/L (>2 standard deviations [SDs] above the mean), is commonly accepted as a reasonable threshold for predicting an increased risk of newborn complications . However, controversy also remains regarding the precise lactate cutoff value to be used for the prediction of adverse neurological outcome. Studies by Tuuli and colleagues reported a normal range for umbilical artery lactate of 2.55 to 4.63 mmol/L and umbilical venous lactate level of 3.4 mmol/L to be the optimal value in predicting for both arterial lactic acidemia and an adverse neonatal outcome. Despite ongoing uncertainty concerning the metabolic markers individually, when a pH less than 7.00 and/or a base deficit of more than 12 mmol/L are associated with other clinical indicators of impaired fetal/neonatal well-being (abnormal fetal heart tracings, 5-minute Apgar scores ≤5, a requirement for intubation), they become strong predictors of poor neurological outcome.

In 2006, Ambalavanan and colleagues suggested that the first postnatal blood gas base deficit was a better predictor of neurological outcomes than arterial umbilical cord gas (aUCG) measures. To further examine the strength of association between blood gas measures from the aUCG and the initial infant gas, Sakpichaisakul and colleagues studied 68 infants with clinical neurological examination and validated MRI scoring system used as a measure of injury severity. The authors concluded that the initial infant blood gases best predicted the severity of abnormality in the infant’s neurological examination and should be used to evaluate risk of HI cerebral injury. Additional research is needed to study the generalizability of these results and determine the value of the infant blood gas combined with other clinical or neurodiagnostic measures as a better predictor of both short- and long-term neurological sequalae.

Multiorgan Dysfunction From Hypoxic-Ischemic Insult

Although not discussed here in depth, important systemic abnormalities , presumably related to ischemia, often accompany the neonatal neurological syndrome. The relative frequencies of manifestations of organ injury in term infants with evidence of asphyxia have been investigated in several studies. The findings vary as a function of the severity of asphyxial insult and the definitions of organ dysfunction. In combined data from two reports, approximately 20% of infants with apparent perinatal hypoxic-ischemic insult had no evidence of organ injury ( Table 24.9 ). Evidence of involvement of the central nervous system occurred in 62% of infants. Indeed, in 16% of infants, involvement of only the nervous system was apparent. The order of frequency of systemic organ involvement overall has been hepatic > pulmonary > renal > cardiac .

Table 24.9

Manifestations of Organ Injury in Term Asphyxiated Infants ^a

ORGAN	PERCENTAGE OF TOTAL
None	0–36
CNS only	0–36
CNS and one or more other organsRenal and cardiac ~65%: Pulmonary and liver ~85%	46–100
Other organ(s), no CNS	10–20

CNS , Central nervous system.

a Cumulative total of 107 term infants; definition of asphyxia in both series included umbilical cord arterial pH < 7.2.

The relative frequency of organ involvement may vary somewhat with the manner of investigation. In one autopsy series, cardiac involvement was the most commonly affected systemic organ. With careful electrocardiographic and enzymatic studies of living infants after perinatal asphyxia, evidence of myocardial ischemia has also been commonly observed. Representative data from a well-studied series of 144 infants with moderate to severe encephalopathy found that all infants displayed some form of organ dysfunction, with pulmonary and hepatic approximately 85%, renal 70%, and cardiac 60%. These frequencies may relate, in part, to the nature of the diagnostic categories for these abnormalities, but they confirm that multiorgan dysfunction is very common in the setting of moderate to severe peripartum HIE and should be sought by appropriate diagnostic studies. Furthermore and related to this recognition, a multiorgan dysfunction in neonatal encephalopathy scoring (MODE) system was recently created, including the cardiovascular, respiratory, gastrointestinal, hematological, and neurological systems, and applied (with a maximum score of 15) in 85 infants to compare with the grade of encephalopathy as well as outcomes at 2 years of age. Infants with higher MODE scores were more likely to have moderate to severe encephalopathy, die, and have abnormal neurological outcomes. The authors concluded that consideration of multiorgan injury assisted in identifying risk of adverse short- and long-term outcomes in neonatal encephalopathy.

Hypoglycemia

Of additional interest in neonatal management in the setting of a hypoxic-ischemic insult is the occurrence of hypoglycemia and its potential role in the accentuation of brain injury. In a detailed study of 185 infants with evidence of intrauterine asphyxia (cord pH < 7.00), 15% exhibited blood glucose concentrations lower than 40 mg/dL in the first 30 minutes of life. The hypoglycemia may relate, in large part, to enhanced anaerobic glycolysis and, therefore, glucose use, in an attempt to preserve cellular energy levels (see Chapter 16 ). By multivariate analysis, the OR for an abnormal neurological outcome was 18.5 when infants with blood glucose levels lower than 40 mg/dL were compared with those with levels higher than 40 mg/dL. The accumulating evidence regarding maintenance of adequate glucose levels has important implications for management (see later).

Other Biomarkers of Ischemic Injury

Hyperammonemia may occur in newborns with severe perinatal asphyxia. Although very uncommon, levels of approximately 300 to 900 µg/mL have been detected in the first 24 hours of life and are usually accompanied by elevated serum glutamic oxaloacetic transaminase levels. Clinical correlates may be difficult to distinguish from those secondary to HIE, although hyperthermia and hypertension have been frequent additions in patients with hyperammonemia. Clinical improvement is coincident with falling blood ammonia levels. The pathogenesis of the hyperammonemia is unclear, although a combination of increased protein catabolism, secondary to hypoxic “stress,” and impaired liver function, and therefore hepatic urea synthesis, is a good possibility (see Chapter 31 ). Recall that hepatic disturbance is a common feature of the systemic multiorgan dysfunction observed with intrauterine asphyxia (see earlier).

Other metabolic parameters have been studied, and some may hold promise as measures of severity of the hypoxic-ischemic insult ( Table 24.10 ), although currently the precise sensitivity and specificity of these determinations require further study before general use is warranted. Furthermore, the type and timing of biological samples measured are of great importance as it rapidly changes with the evolving pathophysiologic changes seen during the hypoxic-ischemic insult.

Table 24.10

Potential Adjunctive Determinations in Blood, Urine, or Cerebrospinal Fluid in Assessment of Perinatal Asphyxia ^a

DETERMINATION	BODY FLUID
Energy metabolism
Glucose	Blood
Lactate	Blood, CSF
Lactate/creatinine ratio	Urine
Lactate dehydrogenase	CSF
Excitatory amino acids
Glutamate	CSF
Aspartate	CSF
Glycine	CSF
Free radical metabolism
Hypoxanthine	Blood, urine
Uric acid	Blood, urine
Non–protein-bound iron	Blood
Protein carbonyls	CSF
Isoprostanes	CSF
Ascorbic acid	CSF
Arachidonate metabolites	CSF
Nitric oxide	Blood, CSF
Antioxidant enzymes	CSF
Inflammatory markers
Interleukin-6	Blood, CSF
Interleukin-10	CSF
Interleukin-1beta	Blood
Tumor necrosis factor-alpha	Blood, CSF
Brain-specific proteins
Neuron-specific enolase	Blood, CSF
Neurofilament protein	CSF
Protein S-100	Blood, urine, CSF
Glial fibrillary acidic protein	Blood, CSF
Creatine kinase-BB	Blood, CSF
Other
Erythropoietin	Blood
Nerve growth factor	CSF
Cyclic adenosine monophosphate	CSF

CSF , Cerebrospinal fluid.

a See text for references.

The metabolites and markers are best considered in terms of their relevance to energy metabolism, excitatory amino acids, free radical metabolism, inflammation, brain-specific proteins, and compounds from other organ systems that may have sustained hypoxic-ischemic injury (see Table 24.10 ). The simultaneous evaluation of a panel of biomarkers for acute brain damage might provide a number of advantages over the measure of individual markers. Information about neuronal injury combined with free radical and cell injury markers would be very useful to a neonatologist understanding the etiology (see Chapters 16 , 22 , and 23 ). It is important to note that cerebrospinal fluid (CSF) biomarkers are generally found to have greater sensitivity and specificity than blood biomarkers and to consider that the early clinical detection of biomarkers might allow an earlier diagnosis. This identification would allow the earlier initiation of intervention measures to improve neonatal survival and reduce the degree of brain injury. In summary, such biomarkers could be important for diagnosis of neonatal HIE, selection of intervention, determination of efficacy, and assessment of the severity of illness and the estimation of prognosis .

More recent techniques have focused on the use of metabolomics to study the whole metabolic profile of the patient simultaneously . Metabolomics has been employed for biomarker discovery in neonatal HIE by taking full advantage of information-rich data sets obtained from multiple different analytical platforms. For details on the use of metabolomics in clinical and experimental studies on perinatal asphyxia, the reader is referred to recent reviews.

Concerning energy metabolism , perinatal asphyxia has been associated with hypoglycemia, elevated lactate in blood and CSF, elevated lactate/creatinine (L/C) ratio in urine, and elevated lactate, pyruvate, and hydroxybutyrate dehydrogenases in CSF. Of these, the value of detection of early hypoglycemia was discussed earlier. Of particular interest is the ratio of L/C in urine. In a study of 40 infants with evidence of intrapartum asphyxia, the mean (±SD) ratio within 6 hours of life was 16.8 ± 27.4 in the asphyxiated infants who subsequently developed the clinical features of HIE versus 0.2 ± 0.1 in those who did not develop encephalopathy and 0.09 ± 0.02 in normal infants. Moreover, the ratio was significantly higher in the infants who had neurological sequelae at 1 year (25.4 ± 32.0) than in those with favorable outcomes (0.6 ± 1.5). The degree of elevation of lactate in blood at 30 minutes of life also may be a useful predictor of the severity of perinatal asphyxia. A study of L/C in urine by proton nuclear MRS within 6 and 24 hours after birth in 50 normal infants and 50 infants with asphyxia who developed HIE showed that the L/C ratio was elevated among asphyxiated neonates in the first 6 hours after birth to 11-fold greater than in normal neonates ( P = 0.0001). This ratio decreased to 1.5 ± 0.55 for asphyxiated cases over the first 24 hours after birth, fivefold greater than in the control group ( P = 0.0001). The severity of asphyxia correlated with the greater L/C ratio among cases ( P = 0.0007). The sensitivity and specificity of the L/C ratio were 96.1% and 100%, respectively. This measure, or other biomarkers, is not used in routine clinical practice for the detection or confirmation of HIE. However, research is ongoing with a particular emphasis on energy-related metabolites.

Concerning excitatory amino acids , elevations of the excitotoxic amino acids glutamate, aspartate, and glycine, each of which acts through the N -methyl- d -aspartate (NMDA) receptor, have been observed in CSF in the first day of life (see Table 24.10 ). Correlations with severity of HIE have been shown. Clinical observation showed increased levels of serum glutamate after neonatal HIE insult. Within 24 hours, the increase of glutamate was significant and reached a peak at day 3 of postnatal life. At day 7, the levels returned to normal, and serum glutamic acid concentrations were closely related to the severity of HIE.

Concerning free radical metabolism , many studies support the involvement of reactive oxygen and nitrogen species in the final common pathway to cell death with neonatal HIE (see Table 24.10 ). These studies have shown elevations in sources of free radicals (e.g., hypoxanthine, nonprotein-bound iron [NPBI], arachidonate metabolites), indicators of lipid peroxidation (e.g., isoprostanes) or oxidized proteins (e.g., protein carbonyls), and markers of free radical use (e.g., ascorbic acid, antioxidant enzymes). Of note in clinical studies is superoxide dismutase (SOD), an antioxidant enzyme that removes the oxygen free radical, superoxide, to protect cells from free radical damage (see Chapter 16 ). Its activity level reflects the oxygen free radical scavenging capacity. Glutathione peroxidase, a second key antioxidant enzyme, detoxifies hydrogen peroxide, the product of action of SOD. In addition, also frequently studied is the lipid peroxidation product of free radical activity, malondialdehyde (MDA), which reflects the extent of oxidative damage to cells. Thus excess free radicals consume SOD, produce a large amount of MDA, promote the release of inflammatory factors in brain tissue, induce nerve cell apoptosis, and increase permeability of the blood-brain barrier in neonatal HIE . A study of 50 cases of asphyxiated full-term newborns found that serious asphyxia resulting in the death of newborns with HIE was associated with concentrations of MDA and glutathione peroxidase in plasma and CSF that were significantly higher than in infants who survived. A recent multicenter study evaluated a new panel of oxidative stress biomarkers (advanced oxidation protein products [AOPP], NPBI, and F2-isoprostanes [F2-IsoPs]) in blood samples of 81 term infants at risk for HIE. The authors concluded that newborns with severe asphyxia showed higher levels of oxidative stress biomarkers than those with mild asphyxia at birth . Furthermore, there was a significant association with AOPP and the severity of brain injury assessed by MRI, especially in males. Finally, a systemic review of literature evaluating the noninflammatory CSF biomarkers for clinical outcome in newborn infants with perinatal hypoxic brain injury reported SOD and MDA as promising CSF biomarkers for prognostication (among creatine kinase, xanthine oxidase, vascular endothelial growth factor, neuron-specific enolase [NSE]).

Concerning inflammatory markers , related potentially to hypoxic-ischemic or intrauterine infection or both, elevations of certain cytokines (interleukin [IL]-6, IL-10, IL-1beta, IL-16, IL-18, ICAM-1, P-selectin, granulocyte-macrophage-colony-stimulating-factor, and tumor necrosis factor-alpha [TNF-α]) have been documented in blood and CSF in both term and preterm infants (see Table 24.10 ). The degree to which the elevations in cytokines are primary or secondary is unclear (see Chapters 16 and 17 ). Recent studies have examined the role of inflammation in the exacerbation and recovery from hypoxic-ischemic injury with key mediators. It was found that the inflammatory process may persist into childhood and a longer therapeutic window may be available for neuroprotection therapies. Moreover, early changes in proinflammatory cytokine levels (specifically, IL-6 and TNF-α within the IL-1β pathway) in newborn infants with neonatal encephalopathy may be associated with remote epilepsy.

Concerning brain-specific proteins , specific components of neurons (NSE, neurofilament protein, creatine kinase-BB [CK-BB]) and astrocytes (S-100, glial fibrillary acidic protein, CK-BB) have been studied in blood and CSF to detect evidence of neuronal and glial injury. In general, elevations of these markers in blood or CSF in the first hours of life after a perinatal asphyxial insult have correlated approximately with the severity of clinical and brain imaging findings. However, the value of studies of blood is tempered somewhat by the finding of S-100 and NSE in placenta, suggesting that at least some of these molecules are not entirely brain specific. Recent studies of infants receiving hypothermia treatment for neonatal HIE demonstrated abnormal changes in blood or CSF NSE that correlated with brain injury on neuroimaging. However, findings in relation to the severity of the injury remained variable.

Available data suggest that the determination of CK-BB is a very sensitive indicator of brain disturbance. However, the extreme sensitivity of the indicator in blood impairs the specificity of the measure because variable but appreciable proportions of infants with elevated concentrations of CK-BB in cord blood or neonatal blood samples have no evidence of irreversible brain injury and have a normal neurological outcome. However, two studies of the concentrations of CK-BB in CSF suggested greater specificity and sensitivity concerning identification of hypoxic-ischemic brain injury than with determination of blood CK-BB concentrations (see Table 24.10 ).

Concerning other markers , elevations of erythropoietin in blood and nerve growth factor and cyclic adenosine monophosphate in CSF have been documented after perinatal asphyxia (see Table 24.10 ). The value of these markers and the significance of their elevations remain to be established.

Currently, none of the markers has been established to be of sufficiently high sensitivity and specificity to be appropriate for general clinical use. However, several appear to be promising (see references for reviews.).

Lumbar Puncture

A lumbar puncture should be performed on any infant with HIE in whom the diagnosis is unclear. It is particularly important to rule out other potentially treatable intracranial disorders (e.g., early-onset meningitis) that may mimic the clinical features of HIE.

Electroencephalogram

The EEG monitoring of the newborn with encephalopathy serves three key purposes: (1) selection of infants who may benefit from TH in the first 6 hours of life (almost exclusively by aEEG technology); (2) monitoring for neonatal seizures (both EEG and aEEG with and without seizure algorithms); and (3) monitoring of the evolution of the electrographic encephalopathy for prognostication of subsequent neurodevelopmental outcomes (aEEG and EEG) (see Chapter 15 ).

A full description of the technologies of EEG and aEEG is outlined in Chapter 13, Chapter 15 . However, briefly, aEEG provides one or two channels of both time-compressed rectified amplitude tracings alongside the raw signal from the centroparietal cerebral regions. This approach contrasts with 16 channels of conventional EEG signal, providing wider coverage of the cerebrum with expert monitoring and evaluation including concurrent video. However, aEEG can often be quickly implemented and applied widely in less resourced clinical settings. The utility of these two technologies in HIE in the term infant is outlined below.

Electroencephalographic Definition of Hypoxic-Ischemic Encephalopathy

Evaluation by EEG in HIE provides valuable information concerning the severity of the injury. A recommended approach has been to utilize EEG monitoring throughout the period of TH and rewarming in infants treated with hypothermia for HIE. However, recent studies have suggested a more tailored duration of monitoring based on early continuous video EEG background categorization and to optimize utilization of continuous video EEG resources. In addition, a study by Mahfooz and colleagues reported that the highest diagnostic yield of EEG monitoring was within the first 24 hours and during the rewarming phase and advised to limit the study to these periods in resource-limited settings.

EEG has also been studied as a means by which to improve the identification of infants presenting with clinical characteristics of mild HIE , who may be at risk for brain injury and benefit from inclusion in neuroprotective clinical trials. In a recent study by Garvey and colleagues, 72% of infants with mild HIE had at least one abnormal EEG feature in the first 6 hours on qualitative assessment. Alterations in EEG background, particularly in sleep–wake cycling and excessive slow wave activity were notable, and previously described as important indicators of the severity of neonatal HIE and prognosis. As early alterations in sleep–wake cycling can also be observed with aEEG, alterations in cyclicity on aEEG may also assist in identifying infants with mild HIE who may be considered for inclusion in future clinical trials. Though EEG/aEEG data alone may not be predictive of neurological outcome in mild HIE, based on the meta-analysis by Falsaperla and colleagues, additional research is needed to determine its potential value in combination with other clinical and diagnostic tools.

The most common evolution of EEG changes in moderate to severe HIE is initially voltage suppression and a decrease in the frequency (i.e., slowing) into the delta and low theta ranges. Within approximately 1 day, and often less, an excessively discontinuous pattern appears, characterized by periods of greater voltage suppression interspersed with bursts, usually asynchronous, of sharp and slow waves. Some infants exhibit multifocal or focal sharp waves or spikes at this time, often with a degree of periodicity. Over the next day or so, the excessively discontinuous pattern may become very prominent, with more severe voltage suppression and fewer bursts, now characterized by spikes and slow waves. This burst-suppression pattern is of ominous significance, especially in the full-term infant (see Chapter 13 ). However, it is critical to recognize that excessively discontinuous patterns with prolonged interburst intervals (IBIs), which are not as severe as classic burst-suppression patterns, nevertheless also are associated with an unfavorable outcome (see ‘Prognosis’ section and Chapter 13 ). Indeed, in one large series of infants, only 16% of excessively discontinuous tracings (in patients with a generally unfavorable outcome) exhibited burst-suppression patterns by classic definition. Notably, however, as many as 50% of asphyxiated term infants with a burst-suppression pattern identified by aEEG in the first hours of life develop normal or nearly normal tracings within 24 hours (see later). In a recent meta-analysis of 18 studies with 940 neonates who received TH, the pooled sensitivity for a burst suppression pattern for death and neurodevelopmental impairment was 0.87 (95% CI, 0.79 to 0.93), and specificity was 0.60 (95% CI, 0.44 to 0.74). Sensitivities were higher when assessed beyond the first 24 hours (0.89 [95% CI, 0.82 to 0.96], whereas specificities were lower (0.47 [95% CI, 0.40 to 0.55]). In general, the most useful tracings for detection of severe encephalopathy have been continuous low-voltage, flat, and burst-suppression tracings . Positive predictive values (PPVs) for an unfavorable outcome with such tracings in the first hours of life are 80% to 90% (see ‘Prognosis’ section). Of infants with these marked background abnormalities, 10% to 50% may normalize within 24 hours. Rapid recovery is associated with a favorable outcome in 60% of cases.

In the severely affected infant , the excessively discontinuous EEG may then evolve into an isoelectric tracing with a very poor prognosis ( Fig. 24.1 ). A recent study of 486 newborn infants who underwent TH confirmed the presence of burst suppression, isoelectric tracing, or flat tracing at 24 hours of life to be associated with a significant increase in the risk for death or neurodevelopmental impairment (OR 10.91, 95% CI, 6.00 to 19.86). Caution in the interpretation of apparent isoelectric tracings in the newborn not undergoing hypothermia therapy, especially in the first 12 hours of life , is indicated by the findings of Pezzani and coworkers, which showed that of 17 asphyxiated newborns with isoelectric or “minimal” background activity in the first 10 hours, one was normal and one exhibited only epilepsy on follow-up (15 of the 17 died in the neonatal period). In general from the earliest studies of EEG, those asphyxiated infants whose EEG tracings revert to normal within approximately 1 week have a more favorable outcome. More recently, Nash and colleagues continuously EEG monitored babies with video, from the beginning of the cooling procedure throughout the three cooling days and after the rewarming. They found that none of the newborns with a normal EEG background at the beginning of cooling had moderate to severe injury. Moreover, they found that all infants whose EEG background showed excessive discontinuity or extremely low voltage at the end of cooling had moderate to severe MRI lesions. Further data are quite similar; in another more recent study, none of a group of 20 newborns treated with TH for neonatal HIE with mild EEG abnormalities at 6 and 24 hours showed any brain injury on MRI, whereas all infants with low-voltage background at 72 hours EEG exhibited moderate to severe deep nuclear brain injuries.

Fig. 24.1, Evolution of the electroencephalographic changes in severe hypoxic-ischemic encephalopathy.

It is worthy of note that there has been a correlation between specific forms of brain injury and EEG patterns ( Table 24.11 ). Diffuse and severe abnormalities (excessive discontinuity with prolonged IBI, burst suppression, marked voltage suppression, isoelectric EEG) are observed most commonly with diffuse cortical neuronal necrosis. Involvement of the thalamus may also be important (see Chapter 13 ). Focal periodic epileptiform discharges are characteristic of focal cerebral infarction; in one series, approximately 90% of infants with such discharges had infarctions (see Chapter 25 ).

Table 24.11

Most Frequent Correlations of Electroencephalographic Patterns and Table: Topography of Neonatal Hypoxic-Ischemic Brain Injury ^a

EEG PATTERN	TYPE OF HYPOXIC-ISCHEMIC BRAIN INJURY
Excessive discontinuity, burst suppression, persistent marked voltage suppression, isoelectric EEG pattern	Diffuse cortical and thalamic neuronal necrosis
Excessive sharp waves: positive vertex or rolandic, positive frontal, and negative occipital sharp waves	Periventricular leukomalacia (also periventricular hemorrhagic infarction; see Chapters 20 and 28 )
Focal periodic lateralized epileptiform discharges	Focal cerebral ischemic necrosis (infarction)

EEG , Electroencephalographic.

a See text for references.

The role of the EEG in the assessment of brain death in the asphyxiated newborn has not been delineated decisively. Thus, an isoelectric EEG can be observed in infants with cerebral neuronal necrosis but not death of the entire brain (i.e., brain death). Conversely, persistent EEG activity for many days has been documented in infants with clinical and radionuclide evidence of brain death. Currently, the guidelines of the Task Force for the Determination of Brain Death in infants between the ages of 7 days and 2 months requires two clinical examinations indicative of loss of all cerebral and brainstem function and two isoelectric EEG tracings carried out according to standardized techniques separated by 48 hours. Although data are limited, a 72-hour observation period for term infants less than 7 days of age appears warranted in most cases and only when the cause of the coma is unequivocally established.

aEEG is a commonly applied method for continuous monitoring of electrical activity in the newborn ( Fig. 24.2 ; see also Chapter 13 ) and has considerable value in the assessment of the encephalopathic term newborn. aEEG background tracings have been most useful, particularly the burst-suppression, continuous low-voltage, and flat trace patterns. In one large study, the PPV for unfavorable outcome for aEEG detection of severe abnormalities at 3 hours of life was 78%, and at 6 hours it was 86%. Notably, approximately 10% to 40% of infants with marked background abnormalities may normalize within 24 hours, and more than 50% of this minority group will have a favorable outcome.

Although the aEEG acquired within the first 6 hours of age has been considered one of the best predictors of neurological outcome at 18 months in infants with neonatal HIE who did not receive hypothermia therapy, since the widespread use of hypothermia therapy, the predictive value of early aEEG has changed, and infants have been shown to have a normal neurological outcome if the aEEG background voltage activity recovers by 48 to 78 hours. In a meta-analysis of nine studies with 520 infants treated with TH for moderate or severe HIE, the predictive value of an abnormal tracing on aEEG, acquired at 6, 24, 48, and 72 hours of age, was examined. The authors found that (1) a persistent, severely abnormal aEEG background at 48 hours of age or beyond predicted an adverse outcome (PPV 85% and diagnostic OR 67 at 48 hours); and (2) at 6 hours of age, the aEEG background in hypothermia-treated infants had a good sensitivity at 96% (95% CI, 89% to 97%) but low specificity at 39% (95% CI, 32% to 45%). Another recent meta-analysis further reinforced the significance of timing when using the aEEG as a predictive tool, and though it may initially be false positive, it becomes more reliable after 24 hours.

Seizure

Continuous monitoring of conventional EEG with portable equipment has been found to be particularly useful in the identification of seizure activity (see Chapter 13, Chapter 15 ). Early detection of the seizures and the evaluation of response to anticonvulsant therapy are facilitated by modern portable monitoring systems. EEG data can assist in the determination of whether clinical events are correlated with electrical seizures requiring anticonvulsant medication or with nonepileptic events in which anticonvulsant medication administration can be avoided. As discussed earlier, some seizures have readily identifiable clinical manifestations (i.e., clonic or tonic components), whereas many seizures have more subtle manifestations (i.e., orolingual, ocular, or autonomic). The most comprehensive guideline on CEEG monitoring in the newborn was produced in 2011 by the American Clinical Neurophysiology Society. The guideline was created to standardize care and define the best neuromonitoring practices in the neonatal population, while recognizing that not all recommendations would be feasible or applicable across institutions. The guidelines recommend that (1) electrodes be placed using the International 10 to 20 system with additional electrocardiogram, respiratory, eye, and electromyography leads; (2) at least 1 hour of recording be assessed for cycling through wakefulness and sleep; (3) high-risk newborns be monitored for at least 24 hours to screen for the presence of electrographic seizures; and (4) in newborns with seizures, monitoring occurs during seizure management and for an additional 24 hours after the last electrographic seizure. Video EEG recording was recommended for 24 hours rather than a briefer EEG recording as many newborns will not have seizures in the first hour of recording but will experience electrographic seizures within the first day. A statement from the AAP recommends that centers performing TH in newborns with HIE have either aEEG or conventional EEG available for seizure identification. This approach provides insight not only into potentially treatable conditions (frequent, clinically silent seizures) but also into the neurophysiological state of the infant who may also be heavily sedated or therapeutically paralyzed.

Neuroimaging Evaluation

Neuroimaging is used to identify the key neuropathologies, as outlined in Chapter 22 ( Table 24.12 ). These include: (1) selective neuronal necrosis, including three basic patterns, that is, diffuse injury, cerebral cortex-deep nuclear injury, and deep nuclear-brainstem injury; (2) parasagittal cerebral injury; (3) white matter injury (WMI) (see Chapter 20 ); and (4) focal ischemic necrosis and stroke (see Chapter 25 ). The accurate diagnostic application of neuroimaging modalities in the newborn infant with HIE is related to the level of radiological expertise for the acquisition and interpretation of the studies, regardless of the neuroimaging method. The methods of acquisition and level of experience in the interpretation of neuroimaging studies in the newborn can vary greatly between institutions. Integration of neonatal, neuroradiological, and radiological physicians by joint conference and review of case materials on a regular basis (e.g., weekly) will assist in improved communication and the application of modern neuroimaging methods. In addition, expert interpretation at a center of excellence in perinatal and neonatal neurology should be requested. Recommendations have varied slightly over the last decade but can be summarized as early cranial ultrasound (US) and definitive evaluation by MRI ( Table 24.13 ).

Table 24.12

Major Techniques for Diagnosis of Specific Neuropathological Types of Neonatal Hypoxic-Ischemic Encephalopathy

	DIAGNOSTIC TECHNIQUE
NEUROPATHOLOGICAL TYPE	MRI		CT	ULTRASOUND
Selective neuronal necrosis: cerebral cortical	++		+	–
Selective neuronal necrosis: basal ganglia and thalamus	++		+	+
Selective neuronal necrosis: brainstem	++		±	–
Parasagittal cerebral injury	++		+	–
Focal and multifocal ischemic brain injury	++	++		+
Periventricular leukomalacia	++	+		++ ^a

++, Very useful; +, useful; ±, questionably useful; –, not useful.

CT , Computed tomography; MRI , magnetic resonance imaging.

a Very useful for detection of focal component; not useful for detection of diffuse component or “noncystic periventricular leukomalacia” (see text).

Table 24.13

Summary of Recommendations in Term-Born Infants With Neonatal Encephalopathy

Adapted from Wisnowski JL, Wintermark P, Bonifacio SL, et al. Newborn Brain Society Guidelines and Publications Committee. Neuroimaging in the term newborn with neonatal encephalopathy. Semin Fetal Neonatal Med . 2021;26(5):101304 - Inder coauthor

	Neuroimaging of the Neonate (2002)	Neonatal Encephalopathy and Neurological Outcome 2nd Edition (2019)	Imaging and the Term Neonatal Brain	Neuroimaging in the Term Newborn With Neonatal Encephalopathy
Organization	Practice Parameter of the American Academy of Neurology and Child Neurology Society	American College of Obstetricians and Gynecologists and American Academy of Pediatrics	Position Statement of the Canadian Pediatric Society	Newborn Brain Society Guidelines and Publication Committee
Cranial Ultrasound (CUS)	No specific recommendations	CUS on admission and in first 2–3 days to exclude dysgenesis	Useful if trained technologists Not recommended as sole imaging modality	May be useful if trained acquisition and interpretation for antenatal/dysgenesis Day 1 Not as only neuroimaging
CT scan	Noncontrast if trauma	Sensitive to hemorrhage but not brain injury	May be useful in urgent settings At 72 ± 12 h	Only in urgent situations Only when MRI or US not available
MR imaging	For all infants between day 2–8 Include MRS/DWI if available	Acquired in all infants with NE 24–96 hours DWI/MRS/MRI for timing Day 7–21 (day 10 optimal) for extent of injury	Day 3–5 in all NE infants Repeat MRI at day 10–14 if discordance between imaging and clinical features	T1/T2/DWI/MRS Day 3–5 in all NE infants Repeat MRI at day 10–14 if discordance between imaging and clinical features
Additional notes			Note if no experienced neuroradiologist, then send for external opinion	Note if no experienced neuroradiologist, then send for external opinion

CT , Computed tomography; DWI , diffusion-weighted imaging; NE , neonatal encephalopathy; MR , magnetic resonance; MRI , magnetic resonance imaging; MRS , magnetic resonance spectroscopy; T1W and T2W , T1- and T2-weighted images; US , ultrasound.

Cranial Ultrasound

Cranial US remains the most commonly used neuroimaging modality in the neonatal intensive care unit (NICU) setting and is commonly applied to the term infant with neonatal encephalopathy. In the Vermont Oxford Neonatal Encephalopathy Registry and more recent survey conducted by the Newborn Brain Society, cranial US was acquired in nearly 40% to 50% of all infants with neonatal encephalopathy undergoing TH. Cranial US may be the only imaging modality possible if an infant is too clinically unstable to transport from the NICU. Cranial US is sensitive for parenchymal hemorrhage, ventricular size, gross brain malformations, and cystic changes in the brain parenchyma (see Chapter 13 ). It is less sensitive for smaller and more subtle abnormalities within the brain, including cerebral cortical or brainstem neuronal disease, noncystic WMI, and minor cerebral dysgenesis. It is a very useful screening evaluation in the term infant with encephalopathy; cerebral dysgenesis has been identified in approximately 2% to 4% of infants who had been diagnosed with hypoxic-ischemic injury ( Table 24.13 ). Cranial US is often used to assess the presence of slit-like ventricles or sulcal effacement related to cerebral edema, and hemispheric or basal ganglia echodensity. Cranial US also can detect severe deep nuclear gray matter injury ( Figs. 24.3 and 24.4 ). However, cranial US lacks sensitivity in defining the full extent of the cerebral lesions, even in severe encephalopathy, and particularly in the first 24 hours of life. Thus, detection of pronounced abnormalities on the first day of life may be of antenatal origin. Recently, a validated cranial US scoring system developed for ease of use was shown to be associated with outcome in neonates with neonatal encephalopathy when performed on days 3 to 7 of life ( Fig. 24.5 ).

Fig. 24.3, Hypoxic-ischemic injury to basal ganglia in a 24-hour-old full-term infant who experienced severe perinatal asphyxia.

Fig. 24.5, Cerebral ultrasound, outcome and HIE: example images.

This scoring system was applied in two cohorts of 83 and 35 term-born infants with neonatal encephalopathy and revealed that the best predictive value was with the day 3 to 7 cranial US where most items predicted adverse neurological outcome.

COMPUTED TOMOGRAPHY

Computed tomography (CT), in medical centers without ready access to MRI, has limited value in the evaluation of the infant with HIE and as we discuss later, MRI is far preferable. Concerns for the potential risks from radiation exposure have persisted. This neuroimaging technique should be restricted to select settings in which the information obtained from the imaging study is clearly of benefit to the patient and cannot be readily obtained with some other modality, like cranial US and MRI. Examples would include infants with severe head trauma at risk for major epidural bleeding and those who require more definitive imaging in a very brief period because of severe clinical instability.

MAGNETIC RESONANCE IMAGING

MRI provides the highest sensitivity for both anatomical and functional detail and also offers an array of imaging options that can be tailored to the specific clinical question (see later). MRI, however, does have some drawbacks compared with other imaging modalities, particularly in the neonatal period. Unlike cranial US, patients must typically be transported to a radiology suite from the NICU for MRI, which may pose some risk to those infants who are unstable. Safely imaging encephalopathic neonates is a unique challenge. Studies have shown that at least 20% of term-born infants with severe HIE cannot be safely transported to the MRI scanning suite because of the severity of their illness. Further information on MRI in infants and the techniques applied are detailed in Chapter 13 .

MRI has been used in a large number of studies of neonatal HIE. The entire spectrum of hypoxic-ischemic brain injury has been demonstrated ( Table 24.12 ). The major findings by MRI are outlined in Table 24.14 . The relative frequency of these findings is summarized in Table 24.15 .

Table 24.14

Major Aspects of Magnetic Resonance Imaging in the Diagnosis of Hypoxic-Ischemic Encephalopathy in the Term Infant ^a

Major conventional MRI findings in first week
Cerebral cortical gray-white differentiation lost (on T1W or T2W) Cerebral cortical high signal (T1W and FLAIR), especially in parasagittal perirolandic cortex Basal ganglia/thalamus, high signal (T1W and FLAIR, usually associated with the cerebral cortical changes but possibly alone with increased signal in brainstem tegmentum in cases of acute severe insults; see Chapter 22 ) Parasagittal cerebral cortex, subcortical white matter, high signal (T1W and FLAIR) Periventricular white matter, decreased signal (T1W) or increased signal (T2W) Posterior limb of internal capsule, decreased signal (T1W or FLAIR) Cerebrum in a vascular distribution, decreased signal (T1W), but much better visualized as decreased diffusion (increased signal) on diffusion-weighted MRI Diffusion-weighted MRI more sensitive than conventional MRI, especially in first days after birth, when former shows decreased diffusion (increased signal) in injured areas

Major conventional MRI findings in first week

Cerebral cortical gray-white differentiation lost (on T1W or T2W)
Cerebral cortical high signal (T1W and FLAIR), especially in parasagittal perirolandic cortex
Basal ganglia/thalamus, high signal (T1W and FLAIR, usually associated with the cerebral cortical changes but possibly alone with increased signal in brainstem tegmentum in cases of acute severe insults; see Chapter 22 )
Parasagittal cerebral cortex, subcortical white matter, high signal (T1W and FLAIR)
Periventricular white matter, decreased signal (T1W) or increased signal (T2W)
Posterior limb of internal capsule, decreased signal (T1W or FLAIR)
Cerebrum in a vascular distribution, decreased signal (T1W), but much better visualized as decreased diffusion (increased signal) on diffusion-weighted MRI
Diffusion-weighted MRI more sensitive than conventional MRI, especially in first days after birth, when former shows decreased diffusion (increased signal) in injured areas

FLAIR , Fluid-attenuated inversion recovery; MRI , magnetic resonance imaging; T1W and T2W , T1- and T2-weighted images.

a See text for references and more details concerning timing of findings.

Table 24.15

Neuroimaging Characteristics on Magnetic Resonance Imaging in Term Encephalopathy

From Inder T. Role of neuroimaging. Chapter 10. In: Neonatal Encephalopathy and Neurologic Outcome . 2nd ed. American College of Obstetrics & Gynecology; 2014.

STUDY		COWAN 2003	MILLER 2005	OKEREAFOR 2008	CHAU 2009	STEINMAN 2009	RUTHERFORD 2010	VERMONT OXFORD 2014	CHEONG 2012	MARTINEZ-BIARGE 2010 2011
Inclusion criteria		Either fetal heart rate monitoring abnormalities or cord pH < 7.1 or delayed respiration or 5-min Apgar < 7 or multiorgan failure	One of: Apgar score ≤ 5 at 5 min, pH < 7.1, BD > 10	Acute sentinel event: umbilical cord prolapsed (19%), uterine rupture (23%), placental abruption (46%)	Neonatal encephalopathy and one of “fetal distress” cord acidemia, 5-min Apgar ≤ 5, or both	One of: Apgar score ≤ 5 at 5 min, pH < 7.0, BD > 10	Neonatal encephalopathy and abnormal EEG and cord pH < 7.0 or delayed respiration or 10-min Apgar < 5, or both	Seizures or stupor Coma or hypothermia or Apgar at 5 min < 3	Two of: Apgar score ≤ 5 Need for mechanical ventilation at 10 min pH < 7.0, BD > 12	Term neonatal encephalopathy and MRI in first 6 weeks Apgar score ≤ 5 at 5 min or pH < 7.1, or both “fetal distress”
Number of infants		245	173	48	48	64	67 no hypothermia	1074	61 no hypothermia	425
Day of MRI scan (range)		14 days of birth	6 days (1–24)	10 days	72 h	14 days	8 days (2–30)	7 days	6 days (3–8)	10 days (2–42)
	No abnormality	16%	30%		33%	40%	20%	18%	13%	See cortical injury
	Basal ganglia/thalamus	80%	25%	75% any 65% moderate to severe	31%	33%	80% any 58% moderate-severe	30%	36% moderate-severe	41%
	Cortical injury	10%	45%	58%	36%	45%	63%	15%	28%	44% ( n = 186) normal or white matter/cortical
	White-matter injury	5%	Not recorded	Not recorded	21%	15%	84% any 13% severe	25%	23%	See cortical injury
	Other diagnosis	4%	None noted	None noted	Arterial infarction in 10%		4%	2%		11% from total of 555

BD , Base deficit; EEG, electroencephalographic; MRI , magnetic resonance imaging.

Conventional MRI shows the abnormalities in the first 3 to 4 days, but generally not on the first day. However, diffusion-weighted imaging (DWI), based on the molecular diffusion of water, is not only more sensitive than conventional MRI but also shows abnormalities earlier, often in the first 24 to 48 hours after birth (see later discussion and Fig. 24.6 ). The correlates of the MRI findings with the neuropathological states described in Chapter 22 are apparent ( Figs. 24.6 to 24.13 ). Thus selective cerebral cortical neuronal injury is manifested by loss of the cerebral gray-white matter differentiation and by cortical high signal (highlighting) on T1-weighted (T1W) or fluid-attenuated inversion recovery (FLAIR) images at the sites of particular predilection, the parasagittal perirolandic cortex, and the depths of sulci (see Fig. 24.7 ).

Fig. 24.6, Deep nuclear-cortical selective neuronal injury.

Fig. 24.7, Magnetic resonance imaging scan of cortical neuronal injury.

Fig. 24.8, Magnetic resonance imaging (MRI) scans of selective neuronal injury.

Fig. 24.9, Magnetic resonance imaging scans of hypoxic-ischemic injury to basal ganglia and thalamus.

Fig. 24.10, Diffusion-weighted magnetic resonance imaging (DWI) of deep nuclear and brainstem injury.

Fig. 24.11, Severe hypoxic-ischemic injury on magnetic resonance imaging (MRI) without (A–D) and with (E and F) deep nuclear gray matter and profound cortical injury. Brain MRI: axial diffusion-weighted imaging (A and C), axial T2 (B, D, and F), and axial apparent diffusion coefficient (E) examination. (A–D) MRI images at 4 days of life of a female born at 36 5/7 weeks gestational age (GA) with neonatal encephalopathy showing multifocal regions of restricted diffusion and abnormal T2 signal involving the cerebral hemispheres, bilateral basal ganglia and thalami, corpus callosum, brainstem, and cerebellar hemispheres consistent with severe hypoxic-ischemic injury (A and B). MRI images at 2 weeks of life showing interval gliosis in the thalami and perisylvian cortices. The thalamic site of injury associated with moderate volume loss and some obscuration of the adjacent posterior limb of the internal capsule. Also, there is gliosis involving the right hypothalamus. Findings are consistent with expected evolution of profound-mechanism hypoxic-ischemic injury (C and D). MRI at 8 days of life of a female born at 40 1/7 weeks GA with neonatal encephalopathy showing diffuse symmetrical signal abnormality involving the deep gray nuclei, white matter, and cortex consistent with severe hypoxic-ischemic injury (E and F).

Fig. 24.12, Magnetic resonance imaging scan of multifocal ischemic brain necrosis.

Decreased diffusion (increased signal) is seen on DWI (see Fig. 24.8 ).

Selective cerebral cortical neuronal injury is usually accompanied by involvement of basal ganglia (especially dorsal putamen) and thalamus (especially lateral thalamus; see Fig. 24.9 ). In the unusual cases of principally deep nuclear and brainstem involvement, as with severe, acute asphyxial insults, high signal (T1W or FLAIR) is seen in the brainstem tegmentum as well as in the basal ganglia. DWI is more sensitive for detection of cerebral cortical and deep nuclear involvement (see Fig. 24.10 ). This pattern can be more global and severe (see Fig. 24.11 ).

As time progresses, with encephaloclastic dissolution of brain tissue there can be major cystic changes seen after several weeks ( Fig. 24.12 ).

In one series of 173 encephalopathic term newborns, predominant involvement of perirolandic cortex and basal ganglia/thalamus was observed in 44 (25%) and in an additional 24 (14%) in association with predominant involvement of parasagittal regions.

Parasagittal cerebral injury is seen readily as areas of increased signal (T1W and FLAIR) in the parasagittal cerebral cortex and subcortical white matter (see Figs. 24.13 to 24.15 ).

Fig. 24.14, Magnetic resonance imaging (MRI) scan of parasagittal cerebral injury.

Fig. 24.15, Magnetic resonance imaging (MRI) scan of parasagittal cerebral injury.

The relative distribution of the abnormalities among many large-scale MRI studies has varied somewhat because as different systematic scoring systems for neuroimaging assessment have been developed (see Table 24.16 ). A recent comparison of these systems in 135 term-born encephalopathic infants suggested that the Weeke scoring system had the greatest reproducibility, sensitivity, and predictive value for all neurodevelopmental outcomes at 2 years of age.

Table 24.16

Comparative Summary of the Parameters of MRI Scoring Systems

Barkovich Score	Mathur Score ^, ^a	Weeke Score ^, ^b
Total score 4	Total score 134	Total score 57
0 Normal	Subcortical region (caudate, GP/putamen, T, PLIC) 4 items Points possible: 72	Gray matter (BG, T, PLIC, brainstem, perirolandic cortex, hippocampus ^c ) 6 items Points possible: 23
1 Abnormal signal in BG orT	White matter Points possible: 18	¹ H-MRS (BG NAA, BG lactate) 2 items Points possible: 2
2 Abnormal signal in cortex	Cortex Points possible: 18	White matter (Cortex excluding perirolandic, WM abnormal SI, PMWL, hemorrhage, optic radiation, corpus callosum ^d ) 6 items Points possible: 21
3 Abnormal signal in cortex and BG or T	Brainstem Points possible: 12	Cerebellum (Cerebellum abnormal SI or diffusion restriction, Cerebellar hemorrhage) 2 items Points possible: 8
4 Abnormal signal in entire cortex and BGT	Cerebellum Points possible: 18	Additional (IVH, ^d SDH, ^d CVST ^d ) 3 items Points possible: 3

BG , Basal ganglia; BGT , basal ganglia and thalamus; CVST , cerebral venous sinus thrombosis; GP , globus pallidus; ¹ H-MRS , proton magnetic resonance spectroscopy; IVH , intraventricular hemorrhage; NAA, N -acetylaspartate; PLIC , posterior limb of the internal capsule; PWML , punctate white matter lesions; SDH , subdural hemorrhage; SI , signal intensity; T , thalamus; WM , white matter.

a Mathur score: Regional subscore: summation of independent T1 + T2 + DWI images for the region, each for the left and right side (except the brainstem scored from 0 to 2 in all three sequences due to its smaller size). MRI injury grade: summation of the five regional subscores, total of 138 points possible for this score (0 = No injury; 1–11 = mild injury; 12–32 = Moderate injury; and 33–138 = Severe injury).

b Weeke score: Each item receives 1 point if focal injury (<50%) or 2 points if injury is extensive (≥50%), and each item also receives 1 point if there is unilateral injury or 2 points if the injury is bilateral. There are 4 maximum points per individual item. There are 57 total points possible for this score.

c This item has a maximum point possible of 3.

d This item has a maximum point possible of 1.

In general, approximately 15% to 30% of MRI scans have been normal. Lesions in basal ganglia/thalamus, either predominantly or more commonly accompanying other areas of involvement, are present in approximately 40% to 80% of cases. Because the lesions in basal ganglia and thalamus often are microscopic in size, some instances of involvement likely may not be detected by MRI. Abnormalities of parasagittal (watershed) white matter and cortex are present in approximately 40% to 60%. The involvement of cortex in the watershed lesions also is likely an underestimate, because the cortical involvement typically is laminar (especially layers 3 and 5), and the entire cortical thickness in the human newborn is only approximately 2 mm. In more severe cases the classic watershed parasagittal cerebral injury involving cortex and subcortical/central white matter is readily apparent. Involvement of basal ganglia and brainstem preferentially occurs in approximately 10% to 20% of cases, usually after a catastrophic sentinel event (see later). Preferential involvement of periventricular/central white matter, similar to WMI of premature infants, is noted as a dominant feature in only approximately 15% of cases, and occurs especially in infants of somewhat lower GA (late preterms), or in the context of hypoglycemia or prolonged cardiovascular instability (e.g., congenital heart disease). About 13% of full-term infants with HIE have also been shown to have abnormal high mammillary body signal on T2-weighted images during the acute phase, though this finding did not relate to the severity of encephalopathy, MRI patterns, or pathology elsewhere in the brain. Although many of the lesions just discussed are visualized well by conventional MRI, they are visualized better and, importantly, earlier by DWI (see Figs. 24.11 and 24.13 ). Many studies of newborns with hypoxic-ischemic disease have demonstrated the superior sensitivity of DWI versus conventional MRI in delineating the site and extent of tissue injury early in the neonatal period. The DWI signal in neonatal HIE is influenced greatly by the timing of the scan and the region studied.

The timing of DWI abnormality in asphyxiated term infants with presumed selective neuronal necrosis or parasagittal cerebral injury, or both, is shown in Fig. 24.16 . Thus, although some infants exhibit abnormality on the first day, injury is generally underestimated at that time. The nadir for diffusion occurs between the second and third days. By 7 to 8 days, pseudonormalization is apparent and is probably related to recovery processes that ultimately lead to angiogenesis and other factors causing increased diffusion. Thus the optimal time for detection of DWI abnormality in the most common varieties of hypoxic-ischemic disease in the term newborn is approximately 2 to 3 days (see Fig. 24.16 ). The evolution of these diffusion changes appears to be altered by the commencement of TH , which is associated with a more protracted pseudonormalization of the diffusion coefficient, with full normalization requiring greater than 10 days ( Fig. 24.16B ). This more protracted course may reflect a slower evolution to cell death and thereby provide a longer window for a second neuroprotective agent. Further studies are needed to confirm this evolution of the diffusion findings in the setting of TH.

Fig. 24.16, (A) The time course of apparent diffusion coefficient (ADC) change after brain injury. Blue represents a 30-minute occlusion. Black represents a 90-minute occlusion. The data are a composite from animal studies. (B) The time course of ADC change after brain injury in term-born human infants. The red line represents normothermic infants. The blue line shows infants treated with therapeutic hypothermia. ADC ratios rather than absolute ADC values are used in the ordinate of both panels because ADC values vary regionally in infants and the areas of injury vary infant to infant. The ratio represents injured tissue over normal tissue, so values <1 indicate a reduction in ADC values.

The importance of the region injured in the evolution of changes in diffusion is illustrated by the scans shown in Fig. 24.17 . Thus in this unusual example of precise knowledge of the timing of the insult (postnatal cardiac arrest), decreased diffusion in basal ganglia and thalamus was apparent at 6 hours of age, but decreased diffusion did not appear in the cerebral cortex until 32 hours. Other investigators showed that although severe WMI is associated with early decreased diffusion, with more moderate WMI diffusion is normal or slightly increased early and increases in the ensuing days ( Fig. 24.18 ). A similar increase in white matter diffusion was observed in cerebral hypoxia–ischemia in the neonatal rat.

Fig. 24.17, Differential regional decreases in diffusion after neonatal hypoxia-ischemia.

Fig. 24.18, Diffusion abnormalities in cerebral white matter after apparent neonatal hypoxia-ischemia

Most recently, studies have attempted to evaluate the MRI scoring systems with the aim of assessing interobserver variability and developing subcategories for the severity of brain injury. These findings suggested that infants with mild HIE and subtle MRI abnormalities may benefit from detailed scoring systems, which would be important for investigating the benefit of hypothermia in mild HIE (see Table 24.16 ). In addition, a broader range of lesions with a greater focal and white matter hemorrhagic component has been recognized by MRI posthypothermia (see new Fig. 24.19 ). The neuropathological correlates of these MRI findings are unknown.

Fig. 24.19, Atypical magnetic resonance imaging (MRI) findings in infants receiving hypothermia for neonatal encephalopathy: axial diffusion-weighted imaging (A, C, and E) and axial T2 (B, D, and F) sequences. (A and B) MRI images at 5 days of life of a female born at 39 2/7 weeks gestational age (GA) with neonatal encephalopathy showing a small focus of cortical injury in the right parietal lobe, deep to the diastatic lambdoid suture, that could represent cortical contusion or embolic insult. (C and D) MRI images at 5 days of life of a male born at 37 weeks GA with neonatal encephalopathy showing a wedgelike area of cytotoxic edema in the left temporooccipital lobe within the larger left temporooccipital focus. This pattern of injury is suggestive of either embolic infarction involving the posterior left middle cerebral artery territory or left cortical vein thrombosis/congestion with subsequent parenchymal infarction. (E and F) MRI images at 5 days of life of a male born at 40 1/7 weeks GA with neonatal encephalopathy showing foci of T2 prolongation suggestive of subpial hemorrhage (F) associated with decreased diffusivity (E) along the anterior and lateral aspect of the anterolateral left temporal lobe and smaller foci involving the right temporooccipital region; injury might be venous in nature.

To summarize, MRI clearly provides superior imaging resolution for delineation of all hypoxic-ischemic lesions, both in the neonatal period and on follow-up (see Tables 24.13 and 24.14 ). DWI provides the capability for identification of injury by 24 to 48 hours after asphyxia in the term infant and thereby assists in the early delineation of the nature and severity of cerebral injury.

Cerebral Metabolic-Hemodynamic Neurodiagnostic Studies

Neurodiagnostic studies that address changes in metabolism and physiology after perinatal hypoxic-ischemic insults include MRS, positron emission tomography (PET), near-infrared spectroscopy, and other measures of the cerebral circulation (see Chapter 13 ). Of these, MRS has proven most useful for diagnostic assessment and is emphasized here.

Magnetic Resonance Spectroscopy

MRS has proven to be a diagnostic modality of particular importance in the evaluation of the infant with perinatal hypoxic-ischemic brain injury. Both phosphorus and proton MRS are useful, although currently the more readily available proton MRS is used most widely. Indeed, over the past few years at many institutions, proton MRS has joined DWI as part of the standard evaluation of infants evaluated by magnetic resonance (MR) techniques for hypoxic-ischemic disease. The basic principles of phosphorus and proton MRS and the normative data obtainable are described in Chapter 13 . The value of these techniques in the assessment of HIE in the term infant is summarized in Table 24.17 .

Table 24.17

Value of Magnetic Resonance Spectroscopy in Assessment of Hypoxic-Ischemic Encephalopathy in the Term Infant ^a

Phosphorus magnetic resonance spectroscopy
Detects high-energy phosphates (PCr, ATP), inorganic phosphate, and pHi; in first few hours after insult PCr, ATP, and pHi are often normal. After approximately 8 h, with development of secondary energy failure, PCr (and later ATP) declines, and pHi increases. In first 1–2 weeks, the severity of decline in PCr and the increase in pHi correlate with the severity of brain injury.
Proton magnetic resonance spectroscopy
Detects multiple compounds, especially lactate, N -acetylaspartate, choline, creatine, and glutamate. Lactate is elevated as early as a few hours after the insult, and it appears to be an earlier indicator of brain injury than is diffusion-weighted magnetic resonance imaging. In the first days after insult, elevations of lactate (and perhaps glutamate) and declines in N -acetylaspartate are identified, and the severity of changes correlates with the severity of brain injury.

ATP , Adenosine triphosphate; HIE , hypoxic-ischemic encephalopathy; PCr , phosphocreatine; pHi , intracellular pH.

a See text for references.

Phosphorus Magnetic Resonance Spectroscopy

Multiple studies of infants who sustained perinatal asphyxia, especially intrapartum, have focused on phosphorus-31 ( ³¹ P) spectra. The sequence of findings has been initially normal spectra (concentrations of phosphocreatine [PCr], inorganic phosphate [P _i ], and adenosine triphosphate [ATP]) in the first hours after birth, followed by a decline in concentration of PCr and a rise in that of P _i (and thus a decline in the PCr/P _i ratio) over approximately the next 24 to 72 hours (see Table 24.17 ). In the most severely affected infants, ATP concentrations also decline at this time. Subsequently, spectra return to normal over the ensuing weeks, although the total ³¹ P signal may be reduced when marked loss of brain tissue has occurred. This sequence of events is directly reminiscent of the progression of the “delayed energy failure” described in Chapter 16 . This secondary energy failure correlates directly with the ultimate degree of cell death. Consistent with the experimental data, the severity of this delayed energy failure in human infants correlates closely with the severity of the neonatal neurological syndrome ( Fig. 24.20 ) and with the subsequent occurrence of neurological deficits (see later).

Fig. 24.20, Relationship between cerebral levels of high-energy metabolites (mean ± standard deviation), determined by phosphorus magnetic resonance spectroscopy, and the severity of the neonatal neurological syndrome (NNS) in 23 asphyxiated term newborns.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here

CLINICAL SETTINGS

NEUROLOGICAL SYNDROME

Temporal Evolution of the Neurological Syndrome

Diagnosis

History

Neurological Examination

Metabolic Parameters

Metabolic Acidemia

Multiorgan Dysfunction From Hypoxic-Ischemic Insult

Hypoglycemia

Other Biomarkers of Ischemic Injury

Lumbar Puncture

Electroencephalogram

Electroencephalographic Definition of Hypoxic-Ischemic Encephalopathy

Seizure

Neuroimaging Evaluation

Cranial Ultrasound

COMPUTED TOMOGRAPHY

MAGNETIC RESONANCE IMAGING

Cerebral Metabolic-Hemodynamic Neurodiagnostic Studies

Magnetic Resonance Spectroscopy

Phosphorus Magnetic Resonance Spectroscopy

You're Reading a Preview

Related Posts

Passive Addiction and Teratogenic Effects

Brain Tumors and Vein of Galen Malformations

Injuries of Extracranial, Cranial, Intracranial, Spinal Cord, and Peripheral Nervous System Structures

Bacterial and Fungal Intracranial Infections