Part B: Brain

INDICATION

Randomized clinical trials performed around the world have shown that therapeutic hypothermia reduces death and major neurodevelopmental disability for infants with moderate-to-severe hypoxic ischemic encephalopathy (HIE).^{1, 2, 3, 4, 5, 6, and 7}

Hypothermia is initiated within 6 hours after birth and continued for 72 hours, with a target temperature of 33°C–34°C for whole-body hypothermia and 34°C–35°C for selective head cooling. Eligible infants were 35 weeks or greater gestational age in 2 trials and 36 weeks or greater or 37 weeks or greater in the remaining trials. A recent systemic review and meta-analysis that included 7 large randomized clinical ­trials and 1214 newborns showed a reduction in the risk of death or major neurodevelopmental disability (risk ratio [RR], 0.76; 95% confidence interval [CI], 0.69–0.84) and an increase in the rate of survival with normal neurological function (RR, 1.63; CI, 1.36–1.95) at age 18 months.⁸ The number needed to treat is 7 to prevent 1 case of neonatal death or major disability. Newborns with moderate HIE had a greater reduction in the risk of death or major neurodevelopmental disability at age 18 months (RR, 0.67; 95% CI, 0.56–0.81) when compared to those with severe HIE (RR, 0.83; CI, 0.74–0.92). There was no difference in head vs whole-body cooling. Follow-up at 6 to 7 years of age of the National Institute of Child Health and Human Development (NICHD) Neonatal Network trial participants found that death or an IQ score below 70 occurred in 47% of the hypothermia group compared to 62% of the control group (p = .06).⁹ This result was not statistically significant; however, hypothermia was associated with a lower death rate without an increase in the rate of severe disability in survivors.

Clinical Findings

History

The recognition of moderate or severe HIE requires a detailed obstetric history, including any acute perinatal event, such as a placental abruption, cord events such as prolapse or complete knot, maternal hemorrhage, uterine rupture, prolonged fetal bradycardia, or nonreassuring fetal heart tracings. Other essential criteria include Apgar score of 5 or less at 10 minutes or the continued need for ventilator support initiated at birth and continued for more than 10 minutes.

Physical

A modified Sarnat neurologic examination is performed to determine eligibility for therapeutic hypothermia unless seizures have been observed. Abnormalities must be identified in at least 3 of the following 6 categories: level of consciousness, spontaneous activity, posture, tone, primitive reflexes, and autonomic nervous system (Table 76-1).

Baseline Tests

The laboratory eligibility criteria for therapeutic hypothermia include a pH of 7.0 or less or base deficit of less than 16 mmol/L in an umbilical cord gas or any gas during the first hour after birth. If the pH is between 7.01 and 7.15, base deficit is between 10 and 15.9 mmol/L, or no blood gas is available, additional criteria are required, including an acute perinatal event and 10-minute Apgar score of 5 or less or assisted ventilation for 10 or more minutes.

Several of the randomized clinical trials required amplitude-integrated electroencephalography (aEEG) with abnormal background activity for study entry; many centers continue to use this eligibility criterion for therapeutic hypothermia.^1,4,5

Other baseline laboratory studies performed in this population include a complete blood cell count (CBC), blood culture, C-reactive protein value, coagulation panel, complete metabolic panel, and creatinine kinase, troponin, and lactate values. Conventional EEG, when performed with video recording, helps with identification of infants with clinical and subclinical seizures.

Differential Diagnosis

The following diagnostic categories must be considered when evaluating an encephalopathic infant, particularly in the absence of an acute perinatal event:

• Metabolic disorders

• Sepsis

• Neuromuscular disorders

• Chromosomal disorders and genetic syndromes

• Central nervous system malformations or injury (eg, subdural or subgaleal hemorrhage)

Exclusions

The following are exclusions for treatment:

1. Infants with lethal chromosomal or congenital anomalies are usually excluded.

2. Infants presenting after 6 hours of age are often not cooled as there are few data suggesting benefit. The window for therapeutic hypothermia is not specifically known; however, it is known that neuroprotection is influenced by the timing of therapy.^10,11 The NICHD Neonatal Research Network is currently conducting a trial to evaluate the effect of cooling initiated between 6 and 24 hours of age (http://ClinicalTrials.gov, identifier: NCT00614744).

3. Infants with profound asphyxia are not likely to benefit; however, early identification of these infants is challenging. At this time, no reliable early clinical markers of nonresponse have been identified, and in the absence of this information, therapeutic hypothermia is initiated and later discontinued if other information, such as neurologic examination, EEG, and magnetic resonance imaging (MRI), confirm that the prognosis is poor.

4. The use of therapeutic hypothermia to treat mild HIE is controversial. Two trials unintentionally enrolled newborns with mild HIE.^6,7 Zhou et al found no death or severe disability in the newborns with mild HIE; Jacobs et al reported a sizable rate of death and disability associated with mild HIE.^6,7 The latter study did not use a standardized neurologic assessment tool or formally certify the transport personnel who assessed newborns for eligibility. Misclassification of the level of encephalopathy is potentially responsible for the higher-than-anticipated rates of death and ­disability.

5. Parental refusal is a cause for exclusion.

PREPARATION

Ethical Considerations and Consent

Neonatal therapeutic hypothermia has been studied in a specific patient population: near-term infants who are less than 6 hours of age and meet strict criteria for moderate-to-severe HIE. Multiple prospective, randomized, controlled trials have shown improved neurodevelopmental outcomes in infants with HIE who are cooled. Based on the results of these studies, it would be unethical to withhold this treatment from near-term infants with moderate-to-severe HIE.¹² Parents must be informed of the availability of this therapy in a timely fashion and offered the option of transfer to another facility if hypothermia is not available at the birth hospital.

The situation becomes less clear if the patient is identified as a candidate for cooling outside the 6-hour window postbirth, if the infant is less than 36 weeks’ gestational age, or if the neurologic status does not meet the specific criteria used in randomized clinical trials. Because the complications of hypothermia are few, it could be argued that ethically any patient who might benefit from the therapy should be offered treatment, for instance, the infant who qualifies for treatment but does not arrive at a tertiary facility until 7 hours of age or the infant who qualifies for treatment but is 35 weeks’ gestation. Is it ethical to withhold treatment from such patients because they do not meet criteria used in studies? Conversely, is it ethical to treat patients who may not benefit from the therapy? Some centers have chosen only to treat infants who meet the strict criteria followed in the clinical trials; others decide whether to treat on an individual basis. If treatment is offered to infants who do not meet the criteria used in randomized clinical trials, informed parental consent is essential.¹²

Another ethical dilemma encountered by centers that offer hypothermia treatment is the potential for delay in outcome prediction. The early sequelae of HIE present during the 72-hour period in which cooling has been shown to be helpful. Cooling can suppress seizures and alter the EEG pattern, confounding the prediction of poor neurologic outcome based on electrographic evidence. Imaging studies such as computed tomography (CT) or MRI are difficult, if not impossible, to obtain in the cooled infant. Cooled patients are often treated with sedatives or narcotics that alter neurologic status. Discussions about prognosis and decisions about continuation of intensive care are often delayed until after the 72-hour treatment window, by which time the cardiorespiratory status of these patients may have improved to such an extent that discontinuation of support is no longer an option. The result can be prolongation of life in a severely damaged infant, greatly altering the options available to parents.¹³ If during the 72-hour treatment period the prognosis for a severely ill infant becomes hopeless, parents should be informed that they may choose to end hypothermia treatment and consider withdrawal of intensive support.¹²

Multiple studies have shown that cooling should begin as soon as possible following a neurologic insult, hence the 6-hour postbirth window used in the clinical trials completed to date. Informed parental consent is mandatory when offering a treatment within the confines of a clinical trial. Obtaining consent within the 6-hour window can be difficult for many reasons and may result in nontreatment because of lack of ­consent. Because hypothermia is considered by many to be standard of care for moderate-to-severe HIE in the first 6 hours of life, consent is not routinely required. In fact, failure to treat based on lack of consent could be viewed as unethical. Outside a study protocol, treatment should be initiated within the 6-hour window, with the intent of informing the parents and reaching agreement about treatment as soon as possible.¹²

Monitoring for Treatment

Intensive Care Monitoring

Infants with neonatal encephalopathy are at risk for multiorgan failure. These infants should be cared for in a level III neonatal intensive care unit (NICU) that can provide intensive monitoring and cardiorespiratory support, correction of metabolic disturbances, treatment of seizures, access to neuroimaging, and consultation with pediatric neurology.

Laboratory Workup

The CBC with platelet count; coagulation panel; chemistries, including liver enzymes; and lactate should be monitored.

Although no increased risk for bleeding has been seen in the randomized controlled trials of hypothermia therapy, changes in hematologic parameters are evident with cooling. Lower platelet counts, abnormalities of platelet activation and aggregation, and delayed activation of the fibrinolytic system have been reported.

Blood Gas Monitoring

Temperature Correction

Blood gas equipment is calibrated to analyze a sample at a “typical” patient body temperature, generally 37°C. Two of the early trials of hypothermia recommended that blood gases run on cooled patients should be performed on an instrument calibrated to the core temperature.^1,3 In an evaluation of the iStat^® instrument (Abbott Point of Care, Princeton, NJ), blood gases from patients cooled to 33.5°C were 15% lower on the corrected instrument than on the instrument calibrated to 37°C, and pH was 15% higher (Sheehan, personal communication). Failure to monitor ventilation with an instrument calibrated to the appropriate temperature in hypothermic patients can result in nontreatment of hypocarbia and alterations in cerebral blood flow.

Avoidance of Hypocarbia

Pappas et al explored the association of hypocarbia and outcome and concluded that death and disability was increased with longer cumulative exposure to carbon dioxide levels below 35 mm Hg.¹⁴

Electroencephalography

It is unknown whether seizures result in brain injury or if their presence signifies an underlying brain disorder. It also remains unclear whether treatment of seizures provides neurodevelopmental benefit. Studies have shown an association between neonatal seizures and neurologic deficits.^15,16 Glass et al compared the neurodevelopmental outcomes of infants who had HIE and seizures with those of infants who had HIE alone and found infants with HIE and seizures had more adverse neurologic outcomes.¹⁷ They concluded that clinical neonatal seizures and their treatment are associated with adverse cognitive and neuromotor outcome and speculated that seizures impair the developing brain. The available evidence supports routine EEG or aEEG monitoring of infants who qualify for cooling, as well as treatment with antiepileptics when seizures are detected.

Drug Monitoring

Therapeutic hypothermia may alter the pharmacokinetics and pharmacodynamics of drugs used in critically ill newborns with HIE. Some drugs have reduced clearance, longer elimination half-life, and a smaller volume of distribution when used during cooling, related to the decline in function of the cytochrome P450 enzyme system and other physiologic adaptations to low temperature. Drugs known to be affected by hypothermia include morphine, midazolam, fentanyl, vecuronium, phenytoin, phenobarbital, and gentamicin.¹⁸

Sedation and Pain Control Considerations

Pain and discomfort can be challenging to assess in cooled newborns as the majority are encephalopathic and do not exhibit the usual indicators of pain. Some cooled infants will demonstrate signs of stress with grimacing, crying, and shivering. Because most patients develop sinus bradycardia while cooled, a heart rate in the normal range may be an indicator of pain. Elevated cortisol levels were reported in cooled animals as compared to normothermic controls.¹⁹ Furthermore, in the same animal model, hypothermia without sedation failed to provide neuroprotection.¹⁹ The authors speculated that inadequate sedation of infants undergoing hypothermia may block the protective effects of hypothermia.

PROCEDURE

Therapeutic hypothermia can be achieved using passive or active cooling methods. The ideal method achieves the desired temperature quickly, maintains that temperature without fluctuation, and avoids over- and undercooling during the initiation and rewarming phases.

Passive cooling is defined as withholding of external heat sources. The isolette or warmer is turned off, and no blankets or clothes are used. The temperature is monitored manually, and heat is provided if the patient becomes too cold. This method of cooling is frequently associated with profound overcooling and great fluctuations in temperature.^{20, 21, 22, and 23}

Active cooling employs devices such as water- or gel-filled mattresses, fans, and gel or water pillows. Active cooling methods may be manual, with adjustments to devices made by the caregiver, or servo controlled, with the patient temperature regulated by an electronic temperature probe with a feedback mechanism to the cooling device. Compared to adults, infant body temperature is more responsive to conductive heating and cooling because of a higher ratio of skin contact area to body mass. Servo-controlled cooling is superior to manually controlled cooling with respect to avoidance of over- and undercooling and fluctuations in core temperature.²⁴

Active servo-controlled cooling has been studied in newborns with HIE using several different methods, including selective head cooling, whole-body cooling, and servo-regulated fan. Selective head cooling utilized a servo-regulated, water-filled, fitted cooling cap attached to a heating/cooling unit (Olympic Medical Cool Care System, Olympic Medical, Seattle, WA). It was used in conjunction with a radiant warmer to maintain a core rectal body temperature of 34°C–35°C. Studies of newborns with moderate-to-severe encephalopathy demonstrated a neuroprotective benefit from mild hypothermia achieved with selective head cooling.^1,6 A servo-controlled fan was designed to provide an inexpensive and safe method of delivering whole-body hypothermia to infants with HIE. This technique, which uses a fan directed cephalocaudally over the infant and a radiant warmer to control core temperature, was successful in cooling patients without overshoot or significant temperature variability.²⁵

Whole-body cooling with a water-filled blanket system is the most extensively used and studied of the techniques currently available for hypothermia in newborns. The radiant warmer is turned off, and cooling to 33°C–34°C is achieved using a water-filled blanket, esophageal or rectal temperature probe, and servo-controlled heating/cooling unit. The devices available in the United States suitable for use in newborns are the CritiCool^® (Mennen Medical, Southampton, PA) and the Blanketrol^® system (Cincinnati Sub-Zero, Cincinnati, OH).

Initiation

Whole-Body Cooling Using the Blanketrol System

The Blanketrol II and III are used to control patient temperature through conductive heat transfer. Both devices contain a heater, compressor, circulating pump, and microprocessor board. For initiation and maintenance of hypothermia and for rewarming, the servo-regulated mode is used with the Blanketrol II. The Blanketrol III can be used like the Blanketrol II for maintenance and rewarming in standard servo mode or by utilizing a software program called Gradient Control to minimize fluctuations in water temperature. With either system, an infant-size blanket is placed under the patient and water is continuously circulated between the Blanketrol heater/cooler unit and the blanket. A probe is placed in the esophagus or rectum for servo-regulation of patient temperature. When using the Blanketrol in standard servo mode, a second larger blanket added to the cooling circuit will act as a heat sink and damp patient temperature fluctuations that occur when only the small neonatal-size blanket is used. The current standard of care is to cool to a temperature of 33°C–34°C and maintain this temperature for 72 hours. Rewarming is done slowly, increasing the temperature by 0.5°C per hour. Care should be taken to avoid rapid rewarming and hyperthermia. Patients should be monitored closely as seizures and hypotension can occur during, and in the hours following, rewarming.

Infants Who Are Slow to Rewarm

Many infants will not reach normothermia in the initial 7-hour rewarming period, and some may take up to 24 hours. The Blanketrol system should remain in place with the temperature set to 37°C degrees. The radiant warmer should remain off. The system should be checked to make sure it is functioning properly, and the environment should be checked for drafts. If after 24 hours of rewarming, normothermia has not been achieved, the Blanketrol system can be discontinued and the final warming done slowly using the radiant warmer, being careful not to exceed a warming rate of 0.5°C per hour.

Hypothermia on Extracorporeal Membrane Oxygenation

Selected patients on extracorporeal membrane oxygenation (ECMO) may also meet the criteria for therapeutic hypothermia. All ECMO systems include a heating/cooling device that is used to control blood temperature in the extracorporeal circuit. This can be used to provide therapeutic hypothermia by setting the device to cool the patient to the desired temperature and manually adjusting based on a core temperature reading. The ECMO heater displays fluid temperature (circuit blood temperature), but this reading does not accurately reflect core body temperature and should not be used to servo control cooling during hypothermia treatment. Esophageal or rectal temperatures can be used to monitor the hypothermia patient on ECMO, but care must be taken with use of these probes because of systemic heparinization during extracorporeal bypass.

Hypothermia During Neonatal Transport

Initiation of hypothermia must occur within 6 hours of birth for optimal neuroprotection. The 6-hour window is easily achieved when an infant is born at a level III NICU; however, the majority of the infants in the large randomized trials were outborn and required transport to a setting where therapeutic hypothermia could be provided. In the 3 randomized trials that performed cooling in transport, there were limited data or discussion of the specific implementation, temperature profiles, or outcomes of infants cooled in transport.^2,4,7

Many cooling centers have chosen to offer cooling during transport by withholding external heat sources (passive cooling) or by using ice/gel packs (active cooling).²⁶ Despite careful attention to temperature management in transport and the development of clinical protocols, the temperatures achieved during transport are not consistently in the target range (33°C–34°C).^{20, 21, and 22} Until recently there were no portable servo-regulated cooling devices approved by the Food and Drug Administration (FDA) for use during transport in the United States. The Tecotherm Neo (Inspiration Healthcare, LTD, Leicester, UK), a small, portable, servo-regulated cooling device with continuous core temperature monitoring, is now approved for use.

Monitoring

Temperature Monitoring

During cooling, the minimal standard for temperature monitoring is a core temperature measurement every 15 minutes for the first 4 hours, every hour for the next 8 hours, and every 4 hours during the remaining period of hypothermia treatment.²⁷ The goal of hypothermia treatment is to cool the brain temperature to 33°C–34°C. Because brain temperature cannot be directly monitored at the bedside, esophageal or rectal temperatures are routinely used as surrogates. In a study of patients undergoing deep hypothermia during cardiopulmonary bypass, with brain temperature measured directly with a thermocouple embedded in the cerebral cortex, esophageal temperature was found to be a better approximation of brain temperature than nasopharyngeal, rectal, pulmonary artery, bladder, or axillary measurements.²⁸ Skin temperature as a proxy for core temperature was studied; a wide discrepancy between skin and rectal temperatures was seen.²² Landry et al compared simultaneous axillary and rectal temperature measurements and found a wide variability at all stages of cooling.²⁹

Overcooling is seen with non–servo-regulated cooling, particularly during transport. It can lead to “cold-injury syndrome,” which is associated with body temperature below 34°C and is characterized by increased mortality, sclerema, renal failure, pulmonary hemorrhage, disseminated intravascular coagulation (DIC), hypoglycemia, electrolyte and acid-base disturbance, increased risk of infection, and cardiac complications.³⁰

Elevated body temperature may occur during rewarming or after hypothermia is discontinued. It is important to be aware that elevated body temperature has been associated with exacerbation of the brain injury caused by HIE.^31,32 For this reason, it is imperative that patients be rewarmed slowly, no more than 0.5°C hourly, and that body temperature be monitored frequently in the postcooling period.

Positioning

Infants must remain supine. Rolls should be placed under the blanket for nesting, bringing it up around the sides of the infant and under the legs, increasing the surface area in contact with the blanket.

Vital Signs

Heart Rate

Most cooled infants develop benign sinus bradycardia, with heart rates between 60 and 100 beats per minute. Heart rates above 120 beats per minute may be a sign of pain or stress in the cooled infant.

Blood Pressure

Therapeutic hypothermia has not been associated with hypotension. A multicenter, randomized, controlled study showed that mild systemic hypothermia did not affect arterial blood pressure or the need for inotropes or volume in infants with moderate-to-severe encephalopathy.³³

Skin Integrity

In the cooled infant, perfusion of the skin may be decreased by perinatal asphyxia and further decreased by hypothermia. Cooled patients may exhibit skin changes, including cyanosis, erythema, and sclerema. More rarely, patients may develop subcutaneous fat necrosis, which usually presents several days or even weeks after discontinuation of cooling therapy.^34,35 Skin complications may be lessened by patient positioning, such as tilting side to side to avoid constant pressure on the back during cooling. Avoidance of overcooling may be protective of skin integrity. Filippi et al studied 39 infants cooled with the Blanketrol III system and compared the automatic temperature control mode to the gradient control mode.³⁶ Two of the 11 infants cooled in the automatic mode developed skin complications consistent with subcutaneous fat necrosis; none of the 28 infants treated with the gradient control mode developed skin complications. The gradient control mode may be superior because of fewer temperature fluctuations.

Hypercalcemia is a potentially lethal complication of subcutaneous fat necrosis.³⁷ The etiology is uncertain but may involve production of nonrenal 1, 25(OH)₂D₃ from the skin lesions. It may present at the same time as the skin lesions or up to 6 months later. Treatment includes limiting calcium and vitamin D ingestion, hydration, and diuretics. Any infant with the diagnosis of subcutaneous fat necrosis posthypothermia should have serum calcium levels followed regularly in the first 6 months of life.

Exit Strategy

Infants who qualify for hypothermia are often critically ill. Despite intensive care, their condition may deteriorate, and a decision may be made to discontinue support. The timing of this decision should not be affected by a desire to complete the therapeutic hypothermia treatment regimen.

REFERENCES

DIAGNOSIS/INDICATION

History

1. Seizure-predisposing conditions in the neonate are, for example,

   • Asphyxia

   • Prolonged low Apgar scores/resuscitation

   • Birth trauma

   • Sepsis or central nervous system (CNS) infection

2. Maternal factors

   • Diabetes (hypoglycemic infant)

   • Fever/chorioamnionitis (sepsis)

   • Viral infection (eg, herpes)

   • Group B streptococcus carriage

3. Pharmacologically restrained and sedated neonates with “at-risk” conditions

4. Family history of siblings or parents with recognized or suspected genetic conditions predisposing to seizures (eg, benign familial neonatal seizure [BFNS], tuberous sclerosis)

Physical

1. Dysmorphic features or multiple congenital anomalies (suggesting possible cerebral malformation)

2. Seizures most commonly occur as repeated, stereotyped episodes of

   • Focal clonic limb jerking

   • Tonic limb posturing with or without head and eye deviation

   • Subtle phenomena, such as stereotyped chewing, tongue thrusting, bicycling, or swimming movements

   • Vital sign changes (eg, apnea, tachycardia, hypo- or hypertension)

Confirmatory/Baseline Tests

1. Initial blood tests

   • Glucose

   • Electrolytes

   • Calcium/magnesium

   • Culture

   • Complete blood cell count and differential

2. Urine analysis and culture

3. Cerebrospinal fluid

   • Cell count and differential

   • Glucose (simultaneous serum glucose recommended) and protein

   • Bacterial and viral cultures

   • Herpes polymerase chain reaction (PCR)

4. Electroencephalogram (EEG) to look for evidence of frequent or continuous seizures, correlate suspicious movements with ictal EEG rhythm, features suggesting focal cortical dysfunction or underlying inborn error of metabolism

5. Neuroimaging

   • Ultrasound (intraparenchymal hemorrhage, malformation with ventriculomegaly)

   • Computerized tomography (intraparenchymal and superficial hemorrhage, cerebral malformation, intraparenchymal calcification [congenital infection])

   • Magnetic resonance imaging (MRI) (cerebral malformation, stroke, hemorrhage, sinovenous thrombosis, vascular malformation)

Further Tests

1. Further blood tests

   • Amino acids

   • Arterial lactate, pyruvate, and ammonia levels

   • Carnitine level and acyl-carnitine profile

   • Karyotype

   • Hepatic enzymes

   • Biotinidase

2. Urine

   • Organic acids

   • Purine and pyrimidine metabolites

   • Sulfites

3. Many other possible diagnostic and screening tests for genetic conditions/inborn errors of metabolism depending on circumstances

Differential Diagnosis/Diagnostic Algorithm

• Physical examination: Vigorous stimulation and repositioning are often useful means to try to interrupt suspicious, seizure-like behavior that is, in fact, nonepileptic (eg, benign sleep myoclonus, jitteriness).

• EEG testing (as previously described) should help differentiate subtle seizures from nonepileptic behaviors (eg, apnea, tachycardia, oxygen desaturation).

• Consider long-term EEG monitoring for less-frequent, stereotyped, seizure-like behavior.

TREATMENT/PROCEDURE

Initiation

• Monitor respiratory and cardiac function, oral secretions.

• Secure vascular access.

Tests to Follow

• Correct deranged metabolic conditions: hypoxia, hypoglycemia, hypocalcemia, hypomagnesaemia, hyponatremia.

• Once treatment with an antiseizure medication is initiated, follow serum levels of readily measurable drugs (eg, phenobarbital, phenytoin).

Management Algorithm/Decision Tree

• Acute treatment with antiseizure medication: Initially consider a single dose of a benzodiazepine:

  • Lorazepam: 0.05 mg/kg/dose IV (preferred; longer duration of action)

  • Diazepam: 0.2 mg/kg/dose IV (hard to give, incompatible with most intravenous solutions, may precipitate or adhere to intravenous tubing)

  • Midazolam: 0.05 mg/kg/dose (very short duration of action)

• Persistent clinical or electrograph seizures require longer-term treatment, classically with phenobarbital:

  • Load: 15–20 mg/kg initially

    • Initial “target” serum levels are 15–40 μg/mL, with level achieved usually approximately the same as the total loading dose.

    • Seizure control may require additional doses to achieve serum drug levels of 40–50 μg/mL.

  • Maintenance: Give 3–4 mg/kg daily, divided every 12 hours; start once the baby demonstrates the ability to metabolize the drug.

  • Follow levels closely, as once fully “loaded,” neonates in the first few days have minimal phenobarbital metabolism and can reach markedly elevated levels; subsequently, the dose may need to increase to 5 mg/kg daily or higher as liver function improves and matures.

• For seizures that persist after phenobarbital loading to high therapeutic serum levels, consider loading with additional medications, such as

  • Fosphenytoin/phenytoin: Load with 15–20 mg/kg (dosing for phenytoin and fosphenytoin is the same, but fosphenytoin is measured as “mgPE,” indicating “phenytoin equivalents”). Maintenance is 5–8 mg/kg daily, usually divided every 12 hours.

  • Levetiracetam: Give 10–15 mg/kg initial dose, titrating to 50 mg/kg daily over 5 days.

• If initial antiseizure medications fail, particularly in the absence of definite seizure risk factors, consider

  • Pyridoxine challenge: Give 100 mg intravenous pyridoxine during EEG monitoring. Caution: This may cause acute respiratory depression.

  • Folinic acid: Give 2.5 mg twice daily.

Exit Strategy/Weaning/Indications for Discontinuation

• For acute seizures following asphyxia, a transient electrolyte derangement, or hypoglycemia, there may be no need for maintenance therapy after a good response to initial loading dose.

• For initially difficult-to-control seizures following an acute asphyxia insult, consider early medication weaning if no seizures recur after the first week. There is no “set time” for weaning, but between a week and 3 months may be considered.

• For seizures in the setting of a traumatic or hemorrhagic insult, it is reasonable to treat for 3 months and wean medication if there has been no seizure recurrence.

• For seizures secondary to congenital structural abnormalities, assess a trial of weaning medication at 3 months.

CONCLUSION/FOLLOW-UP

• Aftercare monitoring: Follow-up care should focus on monitoring for recurrent seizures, screening for medication toxicity, and following developmental progress.

• Follow-up tests: For children who remain on seizure medications, following serum concentrations of phenobarbital, carbamazepine, and phenytoin is reasonable.

  • In most cases, a “routine” follow-up EEG will be of little benefit. If there are paroxysmal spells suspicious for seizures, in that case consider an extended monitoring study, such as inpatient or outpatient video EEG.

• Clinical follow-up

  • Obtain a history to look for stereotyped paroxysmal spells that may be seizures, particularly if associated with altered level of consciousness or motor phenomena that cannot be physically interrupted.

  • Watch for signs of delayed attainment of developmental milestones.

• Long-term outcome/implications for patient/family

  • Neonatal seizures mean there is an increased risk of epilepsy (see chapter).

  • Parents should be instructed in basic seizure first aid (eg, turn child on his or her side and do not put anything in the mouth; activate emergency services for seizures over 3–5 minutes).

  • Parents should be instructed to observe seizure precautions (eg, no bathing unattended by an adult).

SUGGESTED READING

AMPLITUDE-INTEGRATED ELECTROENCEPHALOGRAPHY

What Is Amplitude-Integrated Electroencephalography?

Amplitude-integrated electroencephalography (aEEG) is a simplified, bedside EEG device used to continuously monitor brain function. Brain wave activity is recorded from scalp electrodes, and the raw EEG data are processed using a filter to attenuate frequencies less than 2 Hz and greater than 15 Hz, transform the amplitudes into a rectified and time-compressed signal, and display on a semilogarithmic scale in microvolts (μV). This results in patterns that are easily recognizable without formal training in EEG analysis. Interpretation of the aEEG recording includes assessment of the minimum amplitude, maximum amplitude, bandwidth, and the presence of sleep-wake cycles (SWCs) (Figure 78-1).

Indications for aEEG Monitoring

The following are indications for aEEG monitoring:

1. Confirmation of suspected clinical seizures, especially when formal EEG is not available

2. Assessment for subclinical seizures in high-risk infants

3. Continuous monitoring of therapeutic effects of antiepileptic medication

4. Prediction of outcomes in hypoxic ischemic encephalopathy (HIE)

5. Other potential, but not well-studied, uses for aEEG include metabolic encephalopathy, patients on extracorporeal membrane oxygenation (ECMO), assessment of neuromaturation in preterm infants, assessment of posthemorrhagic hydrocephalus, surgical anesthesia, and postoperative recovery

Placement of aEEG

1. Types of electrodes

   1. Hydrogel: disposable electrodes similar to regular electrocardiographic (ECG) electrodes; used in most commercial aEEG devices

   2. Cup-disk: reusable; requires EEG conductive paste

   3. Needle: minimal skin preparation required and achieves low impedances; invasive, with theoretical risk for infection

2. Recording montages: location of electrodes based on the 10–20 international classification system modified for neonates (Figure 78-2)

   1. Single channel: Traditionally, electrodes are positioned at P3 and P4, which is the area overlying the cortical watershed zone in infants with HIE; alternatively, if using the CoolCap^® System, electrodes may be modified to Fp1 and Fp2.

   2. Three channel: Electrodes positioned at C3, P3, C4, P4 create two biparietal channels (C3-P3, C4-P4) and a cross-cerebral channel (P3-P4).

   3. All devices require an additional ground electrode, usually placed on the back or forehead. Some devices may also require a reference electrode on the scalp (typically Cz).

   4. Other EEG recording systems may utilize additional channels.

3. Skin preparation: When cup-disk or hydrogel electrodes are used, skin should be cleaned and exfoliated prior to application.

4. Data quality: Impedance is continuously measured during the aEEG recording. Impedances greater than 10 kΩ indicate poor electrode contact, insufficient skin preparation, or electrical interference, and the electrodes should be checked.

Interpretation of aEEG Recordings

1. Background pattern interpretation

   1. Basic background pattern descriptions (Figure 78-3)

      • (1) Continuous: minimum amplitude greater than 5 μV and maximum amplitude greater than 10 to 25 μV

      • (2) Discontinuous: minimum amplitude less than 5 μV but demonstrating variability; maximum amplitude greater than 10 μV

      • (3) Burst suppression: minimum amplitude less than 3–5 μV with minimal variability; bursts greater than 25 μV of variable ­density

      • (4) Low voltage: continuous background with both minimum and maximum amplitudes less than 10 μV

      • (5) Inactive/flat: primarily isoelectric tracing with amplitudes less than 5 μV

   2. Term infants with hypoxic-ischemic encephalopathy and prediction of outcome

      • (1) Background pattern: Recovery of a normal background pattern (either continuous or discontinuous) within 36–48 hours has been associated with a good outcome.

      • (2) Sleep-wake cycles: Shorter time to recovery of SWCs (before 36 hours) is associated with a good outcome. Never developing SWCs is strongly predictive of a poor outcome.

      • (3) Recovery of background pattern and SWCs may be delayed for infants treated with hypothermia.

   3. Preterm infants: In healthy preterm infants, the aEEG matures from a discontinuous pattern to a continuous pattern with advancing gestational and postmenstrual age (Figure 78-4). In other words, the minimum amplitude increases, the bandwidth and maximum amplitude decrease, and there is interval development of SWCs. Isoelectric and low-voltage patterns have been associated with the development of intraventricular hemorrhage.

2. Seizures

   1. An isolated seizure is reflected by an increase in the minimum amplitude; repetitive seizures appear as a “sawtooth” pattern (see Figure 78-5).

   2. It is important to confirm seizure activity on the raw EEG tracing (see the section on limitations).

3. Sleep-wake cycling

   1. Immature SWCs may begin to appear as early as 25 to 27 weeks’ gestational age, as evidenced by sinusoidal fluctuations in the minimum amplitude.

   2. SWCs that are more mature in healthy infants appear between 30 and 34 weeks’ gestational age but are often shorter in duration compared to mature SWCs, which emerge at 36–37 weeks’ gestational age.

   3. In healthy preterm infants, SWCs may appear at earlier postmenstrual ages than would be expected of a newborn at an equivalent gestational age.

Limitations of aEEG Monitoring

1. Because of the nature of the time-compressed recording, brief seizures (<30 seconds) may be missed.

2. As only 1 to 3 channels are recorded, seizures or abnormal background activity in the frontal or occipital regions will not be recognized.

3. Artifacts

   1. Mimics of seizures include any repetitive movement, such as burping, patting, swinging.

   2. High-frequency ventilation, external electrical interference, ECG, and muscle movement can falsely elevate amplitudes.

   3. Contact between electrodes (either by physical contact or by a bridge created by paste/moisture) can result in falsely low amplitudes.

   4. Sedative medications may depress the background pattern or inhibit SWCs.

NEAR-INFRARED SPECTROSCOPY

What Is Near-Infrared Spectroscopy?

Near-infrared spectroscopy (NIRS) is a noninvasive technique to continuously monitor regional tissue oxygenation. It has most commonly been used for cerebral oximetry, although monitoring of renal and mesenteric oxygenation are other potential applications. A skin sensor transmits near-infrared light, which penetrates underlying tissues to a depth of 2.5 to 3 cm (depending on device manufacturer). Light is differentially absorbed by oxyhemoglobin and deoxyhemoglobin from the tissue vasculature, and remaining light is then detected by the sensor (see Figure 78-6). A regional tissue oxygen saturation (rSO₂) or a hemoglobin difference (HbD) signal is then calculated based on a weighted average of 20% arterial blood, 75% venous blood, and 5% capillary blood in the underlying tissue. The rSO₂ value reflects a regional balance between tissue oxygen supply and demand.

Indications for NIRS Monitoring

1. Conditions with altered hemodynamic states causing cerebral hypoxemia

   1. Hypoxic ischemic encephalopathy

   2. Congenital heart disease, especially with potentially decreased systemic outflow for perioperative monitoring

   3. Patent ductus arteriosus, particularly after interventions to close the ductus

   4. Prematurity with altered cerebral autoregulation

2. Conditions with altered hemodynamics affecting mesenteric or renal oxygenation

   1. Shock. A drop in somatic oxygenation may be an early potential indicator of low cardiac output before other indicators such as blood pressure, acidosis, and decreased urine output.

   2. Congenital heart disease.

   3. Renal dysfunction.

   4. Necrotizing enterocolitis.

   5. ECMO, to evaluate interventions while on ECMO or to assist with prediction of successful weaning.

3. Other potential applications for NIRS include assessment of intraventricular hemorrhage (IVH), intraoperative management, anesthetic management and with conditions such as anemia, hypoglycemia, seizures, or metabolic disorders.

Placement of NIRS

1. Locations (Figure 78-7)

   1. Cerebral sensor on lateral forehead

   2. Renal sensor on flank

   3. Mesenteric (splanchnic) sensor on abdomen, typically inferior to umbilicus in lower left quadrant

2. Duration is dependent on manufacturer’s recommendations. Evaluate for skin sensitivity, especially if monitoring for longer than 24 hours.

3. Data quality

   1. Right-to-left differences in regional cerebral oxygenation can be up to 10% during periods of unstable systemic arterial oxygenation.

   2. Excessive light may interfere with signal detection.

   3. Adhesion of sensor may interfere with signal detection.

NIRS Measures (Direct and Indirect)

1. rSO₂ (%).

2. HbD signal.

3. Tissue oxygenation index (TOI) (%).

4. Cerebral blood volume as calculated from HbD.

5. Fractional tissue oxygen extraction (FTOE) = (SaO₂ — rSO₂)/SaO₂. Elevated FTOE reflects higher oxygen consumption in relation to oxygen delivery; decreased FTOE suggests less utilization of oxygen by the tissues.

6. There are autoregulation measures such as the pressure passivity index (PPI), which is calculated by percentage time with concordance between mean arterial blood pressure and rSO₂ to indicate a pressure passive circulation.

Interpretation of NIRS Data

1. Values are validated using jugular venous saturation measurements (Figure 78-8).

2. Normal values are controversial because of ­variability between patients.

   1. Cerebral oxygenation

      • (1) Term infants: 60%–80% or approximately 30% less than arterial saturation

      • (2) Preterm infants

        • (a) The cerebral oxygen saturation decreases over the first 1–2 weeks of life in preterm infants less than 34 weeks’ gestation.

        • (b) Range is typically 70% ± 8% in the first 72 hours of life.

        • (c) There is a mean PPI of 20%.

   2. Somatic oxygenation

      • (1) Term infants: typically 5%–15% above cerebral saturations

      • (2) Preterm infants

        • (a) Renal oxygen saturations also typically decrease over the first 1–2 weeks of life in preterm infants less than 34 weeks’ gestation.

        • (b) Renal rSO₂ range is typically 80% ± 10% in first week of life.

        • (c) Mesenteric (abdominal) oxygen saturation values show significant variability and increase with gestational age, ­ranging from 32% to 66% in the preterm population.

3. Abnormal values and associations with outcomes

   1. Cerebral oxygenation

      • (1) Cardiac surgery absolute levels: Values less than 35%–40% for more than 10 minutes are associated with early postoperative neuropsychological dysfunction, lower developmental quotient scores, EEG changes, intracellular anaerobic metabolism, and histopathologic neuronal injury.

      • (2) Cardiac surgery changes in rSO2: Greater than a 20% decrease in rSO₂ is associated with development of seizures, hemiparesis, and adverse neurodevelopmental outcome at 18 months of age.

      • (3) Hypoxic ischemic encephalopathy: Greater than 50% decrease in FTOE and increased rSO₂ levels are associated with adverse neurodevelopmental outcomes at 1 to 2 years of age. Increased cerebral blood volume predicts death or disability.

   2. Somatic oxygenation

      • (1) A ratio less than 0.75 for splanchnic:cerebral oxygenation has been associated with intestinal ischemia.

      • (2) Splanchnic (abdominal) saturations are lower in infants with feeding intolerance.

      • (3) Renal saturations less than 50% for more than 2 hours are associated with higher incidence of acute kidney injury after cardiac surgery.

   3. Autoregulation: Loss of autoregulation (Figure 78-9) has been associated with increased mortality and severe IVH and PVL in preterm infants.

Interventions to Improve rSO₂

1. Improve blood pressure

2. Avoid hypocarbia or hypercarbia

3. Correct anemia with a red blood cell transfusion

4. Increase the fraction of inspired oxygen (FiO₂)

5. Minimize metabolic demand (ie, temperature regulation, sedation)

Limitations of NIRS Monitoring

1. The depth of signal penetration limits measurement of deeper tissue structures: For cerebral monitoring, regional oxygenation of cortex and not deeper brain structures is measured.

2. It is better to trend rSO₂ values over time or with interventions in an individual patient rather than compare absolute values of rSO₂ between patients because of the potential for significant variability between patients.

3. There is a risk of skin sensitivity caused by the sensor adhesive.

4. Interpretation must be made in the context of all factors that influence oxygen extraction and delivery, such as anemia, hypotension, hypocarbia, and metabolic rate.

REFERENCES

SUGGESTED READING

DIAGNOSIS/INDICATION

Definition of Neonatal Stroke

A stroke, in general, is defined as an event that leads to poor blood flow to a localized area of the brain. One of the difficulties in studying strokes in the neonate is that, until recently, there has not been full agreement on the definition of neonatal stroke. Recently, however, the National Institute of Child Health and Human Development and the National Institute of Neurological Disorders and Stroke provided a consensus definition of neonatal stroke as “a group of heterogeneous conditions in which there is a focal disruption of cerebral blood flow secondary to arterial or cerebral venous thrombosis or embolization between 20 weeks of fetal life through 28th post-natal day, and confirmed by neuroimaging or neuropathological studies.” Neuroimaging was defined as T2-weighted magnetic resonance imaging (MRI), magnetic resonance angiography (MRA), or diffusion-weighted imaging, as head ultrasounds can miss anterior and posterior lesions and computed tomography (CT) may not detect small or early lesions.¹ However, ultrasounds and CT scans remain commonly used techniques in the evaluation of neonatal stroke. Neuropathological studies included autopsy and detailed evaluation of the placenta. Thus, the definition of neonatal stroke is currently defined by neuroimaging or neuropathologic findings, is either venous or arterial, and is either thrombotic or hemorrhagic. The location and type of stroke, as well as the neonate’s clinical status, help guide the workup.

Neonatal stroke is a common clinical event; with an estimated incidence between 1/2300 and 1/5000 live births.¹ The incidence in this age category is second only to the incidence seen in elderly adults. The majority of neonatal strokes are caused by thrombosis of the arteries (70%), with a lower rate of hemorrhagic strokes (20%) and sinus venous thrombosis (10%). There is a slightly increased frequency in males compared to females and in African American infants when compared with Caucasian babies. A majority of these events occur in term infants because of their physiological hypercoagulable state near the time of delivery. As neuroimaging techniques become more sophisticated and the technology becomes more available, neonatal stroke may be more easily diagnosed, and this reported incidence may begin to rise.

Maternal Risk Factors

1. Thrombotic Stroke

   1. Pregnancy results in a relative hypercoagulable state with several plasma proteins affected and increased thrombin generation, putting both mother and baby at increased risk of thrombus.² Several publications reported that about half (50%) of mothers of infants diagnosed with a thrombotic stroke were found to have prothrombotic abnormalities.

   2. Chorioamnionitis has also been shown to be associated with increased risk of thrombotic strokes in neonates. A multivariate analysis of a retrospective case control study involving a total of 208 infants found that maternal fever, even in the absence of an identified infection, had a statistically significant association with perinatal arterial ischemia.³

      Maternal autoimmune disorders, in particular lupus, can lead to neonatal stroke. The maternally derived autoantibodies can cross the placenta and cause myriad clinical findings in the neonate. In particular, these antibodies can also cause a vasculitis in the neonate, leading to cerebral arterial or venous thrombosis.

   3. Other maternal factors that compromise the placenta, including the use of cocaine, increase the risk of stroke in the prenatal or perinatal periods.

2. Hemorrhagic Stroke

   1. Quantitative platelet disorders can lead to hemorrhagic neonatal stroke and may be related to the mother. The main maternal risk factor for hemorrhagic stroke is a maternal-paternal platelet antigen mismatch leading to neonatal alloimmune thrombocytopenia (NAIT).Because of the risk of subsequent neonatal hemorrhagic stroke in future infants, this is an important diagnosis to make correctly. Prophylactic measures, such as maternal intravenous immune globulin (IVIG) administration, may be effective in preventing thrombocytopenia, and thereby stroke, caused by NAIT.

      The mother may also have maternal immune-mediated thrombocytopenia (ITP), which carries less risk than NAIT. In this scenario, the mother’s autoantibodies, which are directed at antigens on her own platelets, cross the placenta and then increase the clearance of the infant’s platelets, leading to thrombocytopenia and the potential for increased hemorrhagic stroke.Thus, obtaining the mother’s platelet count, if available, is often helpful in guiding the workup in neonatal hemorrhagic stroke caused by neonatal thrombocytopenia.

Infant Risk Factors

1. Thrombotic Strokes

   1. The physiology of newborns places them at increased risk for thrombotic strokes. Their physiologically elevated hematocrit, which can be made worse in the setting of twin-twin transfusion, increases the risk of thrombosis. Neonates also have a relatively decreased rate of cerebral blood flow. The patent foramen ovale, always present at birth with variable rate of closure, allows a thrombus entering the right side of the heart to pass to the brain.

   2. Congenital cardiac abnormalities are at risk for developing intracardiac thrombi, and these may embolize to the central nervous system (CNS), leading to arterial stroke. Testing for congenital thrombophilic disorders remains controversial in neonates. Testing for congenital thrombophilias, such as factor V Leiden, prothrombin mutations, hyperhomocysteinemia, and protein C/S and antithrombin deficiencies is often performed, but the utility for screening for these disorders in the neonate with thrombotic strokes, both arterial and sinus venous thrombosis, is controversial. ­Testing may not influence treatment or ultimate length of treatment. Thus, with the limitations of blood available for diagnostic testing, the clinician should be judicious in ordering these tests and when to begin testing. One recommended test in infants is to look for antiphospholipid antibodies, which may be placentally transferred from the mother, possibly contributing to development of thrombotic stroke.

   3. Any foreign bodies, such as arterial or venous catheters, particularly in the setting of extracorporeal membrane oxygenation (ECMO), increase turbidity of the blood and therefore increase the risk of thrombosis.

   4. Similar to the contribution of infections to hemorrhagic stroke, infections leading to disseminated intravascular coagulation (DIC) may thus also contribute to the development of a thrombotic stroke. Hemorrhagic stroke may also occur because of DIC.

2. Hemorrhagic Strokes

   1. Gestational age is important in the evaluation of hemorrhagic strokes. Typically, there is an increased risk for hemorrhagic stroke with increasing prematurity.¹ Most neonatal intensive care units (NICUs) have standardized screening policies that utilize head ultrasounds to screen for hemorrhagic stroke based on ­gestational age. The pathophysiology is likely caused by the increased vascular fragility of blood vessels in the premature infant.

   2. Newborns are born physiologically deficient in vitamin K and have a high rate of hemorrhagic strokes (hemorrhagic disease of the newborn) before the use of vitamin K right after delivery. Congenital bleeding diatheses, most commonly such deficiencies in factors 8 or 9 and 13, may lead to hemorrhages as well. Abnormal fibrinogen, both quantitatively and qualitatively, and severe homozygote factor 7 deficiency may also lead to hemorrhagic stroke in the neonate.

Clinical Symptoms and Presenting Features

Neonates who have suffered a stroke often have minimal detectable symptoms. Because of this, the symptoms are often found incidentally in an asymptomatic infant. Subtle clinical findings may include unexplained apnea, hypotonia, or irritability.¹ Many of the deficits caused by stroke, although perhaps easily identifiable in adults, are frequently missed in infants. This may be because developmental milestones, which would reflect asymmetry or loss of function, have not yet been reached. If not detected during the perinatal period, the presenting sign to the clinician may be early preference for a hand (ie, early right handedness) or may be a child who has failed to reach developmental milestones. In combination with a concern for a perinatal neurological event, these children are often given the diagnosis of “presumed perinatal stroke.”

Clinical symptoms that are more overt include seizures, which are often the first sign of neonatal stroke. All seizures in the neonatal period should be worked up with neuroimaging. Hemiplegia may also be an overt finding. For unclear reasons, the left middle cerebral artery is the most commonly affected vessel; thus, right hemiplegia is more often observed than left hemiplegia.

Differential Diagnosis

Many other possible diagnoses may mimic the symptoms of a stroke in the neonate. Table 79-1 lists a few diagnostic considerations to take into account and how to differentiate these from either hemorrhagic or thrombotic stroke in the neonate.

CONFIRMATORY/DIAGNOSTIC TESTING

Imaging Studies

Ultrasound

As cranial ultrasounds are easily obtained and do not require sedation, they are often the imaging modality of choice when initially assessing an infant with neurologic symptoms. Numerous studies have assessed the role of ultrasound in diagnosing stroke. Unfortunately, the sensitivity of cranial ultrasound in the acute phase is poor, with studies reporting 30%–68% of strokes being identified by ultrasound.^{4, 5, and 6} These studies also clearly demonstrated that the skill of the ultrasonographer can greatly contribute to their sensitivity. In addition, this technique does not expose the infants to any radiation.

Computed Tomography

Computed tomography may not detect small or early lesions.⁷ CT is most likely to miss strokes during the first 24 hours after they occur. However, the scans are generally easy to obtain and fast, but they do expose the infant to radiation. A CT scan is generally not used for diagnosis of neonatal stroke but may be ordered when considering other diagnoses as the cause for workup.

Magnetic Resonance Imaging

If the patient is stable enough to tolerate the duration of the imaging, and the suspicion for stroke is high, then the recommended neuroimaging is T2-weighted MRI, MRA, or diffusion-weighted imaging.¹ These techniques also spare the infant from radiation exposure. However, cost, availability, and expertise in reading neuroimages in the neonate may limit these techniques. MRA, may also be obtained at the same time, which permits visualization of the blood vessels of the CNS, allowing for determination of anatomical abnormalities that may have led to the stroke. This technique is the recommended diagnostic examination when suspicion for neonatal stroke is high.

Laboratory Testing

Maternal Risk Factors

1. Sepsis, chorioamnionitis: Blood culture of the mother and full sepsis workup of the infant once imaging confirms a lumbar puncture is safe to perform.

2. Antiphospholipid syndrome: Antiphospholipid antibody, anticardiolipin antibody, and lupus anticoagulant screens may be indicated in a woman who has exhibited the clinical triad suggestive of antiphospholipid syndrome. These symptoms include recurrent miscarriages, thrombotic events, and thrombocytopenia. Any history of maternal autoimmune disease should trigger consideration of testing for these antibodies.

3. Maternal ITP: Maternal thrombocytopenia or history of previous maternal ITP, especially with a history of maternal splenectomy, should raise concern for maternal antiplatelet antibodies having crossed the placenta.

4. Urine toxicology

Congenital Thrombophilia in the Neonate

See Chapter 32 for a discussion of congenital thrombophilia in the newborn.

1. Antithrombin III level

2. Protein C and S activity

3. Factor V Leiden mutation analysis

4. Prothrombin G20210A mutation analysis

5. Homocysteine level or methylene tetrahydrofolate reductase (MTHFR) C677 mutation analysis

TREATMENT CONSIDERATIONS

Thrombotic Stroke

1. Thrombolysis: If a thrombus is felt to be life threatening and there is no significant hemorrhage and no recent or planned invasive surgery, fibrinolytic therapy can be considered.⁸ Although likely with a limited availability for this age group, interventional radiology for mechanical removal and ­catheter-directed thrombolysis may be considered for serious thromboses. Prior to initiation of thrombolytic therapy, the platelet count should be at least 100,000/μL and the fibrinogen at least 1. Neonates may require supplementation of plasminogen (which is naturally lower in neonates) via FFP. Tissue plasminogen activator (tPA) is generally the drug of choice because it has been most effective toward neonatal thrombosis, with the least risk for hypersensitivity.⁸ There have not been any clinical trials to define ideal dosing, but bleeding complications are not infrequent, so the blood bank should be notified and prepared to provide the appropriate products.

2. Anticoagulation and antiplatelet agents: Numerous groups, including the American Heart Association, recommend anticoagulation for neonates only if the following criteria are met: The neonate does not have significant intracranial hemorrhage, there is evidence of thrombus propagation or emboli, or the neonate is found to have a severe prothrombotic state. Based on guidance from the American College of Chest Physicians, the use of anticoagulants depends on the type of stroke and comorbidities.⁹ If surgical intervention is a potential, then often standard, unfractionated heparin may be used because this is easily reversible. It is important when administering heparin that antithrombin III levels be checked; they may be physiologically low and require replacement for effective anticoagulation. Antiplatelet agents have not been well studied in this age group for treatment of thrombotic events.

3. Cerebral venous sinus thrombosis (CVST). In neonates with CVST, many may be associated with concomitant hemorrhage around the thrombus, so it is critical to assess for the degree of hemorrhage prior to initiating anticoagulation. In the absence of a significant hemorrhage, low molecular weight heparin (LMWH) is a recommended treatment modality because of its stable delivery, but an optimal dose has not been determined in a controlled study for neonates because renal clearance is likely higher in this age group. Checking for a therapeutic level is important in this age group. CVSTs tend to recanulate quickly in neonates, so if anticoagulation is initiated, reimaging should be performed at 6 weeks; if full recanulation has occurred, anticoagulation can be discontinued. If at 6 weeks thrombosis is still present, the neonate should complete a 3-month course of LMWH. If a neonate with CVST presents with significant hemorrhage, the infant should be reimaged at 5–7 days to assess for propagation of the thrombus, at which point the degree and direction of propagation will have to be weighed against the degree of hemorrhage in determining whether to begin anticoagulation.

4. Arterial ischemic stroke (AIS). In the absence of evidence of an ongoing source, such as a cardioembolic source, no anticoagulation or aspirin therapy is recommended. If no cause is found, the likelihood of recurrence is low, whereas hemorrhagic conversion is seen.⁹

5. Remove any noncritical indwelling line.

Hemorrhagic Stroke

1. Intracerebral hemorrhage: Platelets should be corrected and maintained at least at 100,000/μL. Prothrombin time (PT) and partial thromboplastin time (PTT) should be checked at least twice daily and corrected with FFP as needed. Vitamin K should be administered if not already given or if there is a defect in vitamin K-dependent coagulation factors. Urgent neurosurgical intervention may be required for drainage or shunting.

2. Neonates who are found to have homozygous protein C deficiency may benefit from protein C replacement. Provide either FFP 10–20 mL/kg every 12 hours or protein C concentrate at 20–60 units/kg until the lesion has resolved.

Supportive Care

Regardless of the cause, brain injury requires standard supportive care. Patients should receive adequate oxygen to maintain normal PaO₂, sodium should be closely followed to minimize cerebral edema, blood pressure should not be allowed to be elevated above or fall below age-appropriate values, and cooling may be useful in decreasing CNS injury during the acute phase.

On the Horizon

Many studies are underway to assess the role of stem cells (cord blood and other sources) in the attenuation of brain injury following neonatal stroke, but significant evidence supporting this therapy has not yet been shown.

REFERENCES