Neurocritical Care in Children

Published on 26/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 3314 times

CHAPTER 174 Neurocritical Care in Children

Critically ill children with neurological disorders have unique needs that include specialized continuous functional monitoring of the nervous system, clinical strategies designed to detect secondary insults to the nervous system, and guided interventions crafted to reverse any evolving pathology while protecting uninjured nervous tissue.

Children suffering from traumatic injuries, recovering from complex surgical procedures such as brain tumor resection or cranial reconstruction, or suffering from respiratory disorders complicating neuromuscular disease require a high level of expertise that is best provided by a multidisciplinary team and specialized equipment.

The team is composed of physicians, nurses, and support staff with extensive experience and expertise in critical care, neurology, and neurosurgery. The multidisciplinary nature of such a team requires careful coordination of services and resources. The optimal level of neurocritical care can be delivered through the development and implementation of best clinical practice guidelines, a comprehensive educational and training program, and a specialized quality improvement program.

Neurological Assessment in the Pediatric Intensive Care Unit

A variety of tools are available for monitoring neurological function in critically ill children and adults (for review, see Tisdall and Smith ¹ and Wartenberg and colleagues²). Such tools range from electroencephalographic (EEG) monitoring, which is well established as a modality essential for the detection of nonconvulsive seizures (NCSs) in the adult and pediatric intensive care unit (ICU),³^,⁴ to intracranial pressure (ICP), transcranial Doppler,⁵ and tissue oxygen monitoring.¹^,⁶ Although thresholds for detection of cerebral ischemic injury have been proposed for brain oxygen tension⁷ and near-infrared spectroscopy,⁸ there is no consensus on normal values and age dependence of these end points, and there are only limited data from adult studies to show impact on outcome.⁹ Indeed, most of these studies focus on traumatic brain injury (TBI), and the application of these data to the quantification of primary neurological injury or detection of secondary insults in the context of critically ill children with stroke, refractory status epilepticus (RSE), or central nervous system (CNS) infections remains to be determined.

The importance of the bedside neurological examination for assessment of the initial neurological injury and progression of the insult, the need for serial examinations, and involvement of the entire medical team in this assessment cannot be overstated. It is clear that admission to a specialized neurosciences ICU,¹⁰ management by a neurocritical care team, or the presence of a neurointensivist¹¹^,¹² is associated with reductions in length of stay and mortality in adults. Accordingly, the Glasgow Coma Scale (GCS) score is simple to determine and yields reproducible results among observers,¹³ but the GCS is not a neurological examination and does not assess brainstem (particularly pupil reactivity) function or the presence of focal findings.

The neurological examination in the ICU should include many of the components (mental status, cranial nerves, muscle tone and strength, reflexes) of the complete neurological examination but with a focus on changes in mental status, loss of reflexes, development of pathologic reflexes, or asymmetries of tone or strength. The initial neurological examination serves to establish the baseline against which serial examinations are compared. Importantly, the baseline examination should be confirmed by the nurses and physicians caring for the patient. Involvement of the entire team is an essential step in ensuring that pathologic changes in neurological function are detected as early as possible. The neurological examination in a critically ill child may be limited, but this population has multiple risk factors for the development of new or progression of acute neurological insults. Therefore, attention to subtle changes in the results of the examination is essential. The resulting interventions (for example, to minimize secondary injury), additional diagnostic studies (CNS imaging, EEG monitoring, ICP monitoring), or therapeutic interventions will be contingent on the nature of the primary insult and mechanism of the injury.

The Critically Ill Child

Children with neurological disorders can become critically ill, and regardless of the primary diagnosis, quality of life is determined mostly by neurological outcome. Physiologic decompensation because of critical illness can expose the brain to systemic secondary insults such as hypotension, hypoxia, hypercapnia, and hyperthermia. These secondary insults may aggravate neurological injury, especially in the presence of intracranial hypertension or disrupted cerebral blood flow autoregulation.

Airway Management and Respiratory Failure

Respiratory decompensation in children is the most common cause of cardiorespiratory arrest. Careful monitoring of the patient’s respiratory status in the pediatric intensive care unit (PICU) aims at preventing respiratory decompensation and supporting patients with respiratory failure.

Respiratory decompensation creates additional challenges in critically ill children with neurological disorders. Given the impact that abnormalities in ventilation and oxygenation have on cerebral hemodynamics and metabolism, respiratory abnormalities can result in secondary insults to the brain. For example, hyperventilation can lead to hypocapnia and detrimental changes in cerebral blood flow—a 3% reduction in cerebral blood flow for each 1-mm Hg reduction in PCO₂. Although targeted hyperventilation is considered a treatment option during periods of severe intracranial hypertension and cerebral herniation, unintentional hyperventilation has been independently associated with increased mortality.¹⁴ Avoidance of unintentional hyperventilation is especially important in the immediate postinjury period after TBI, when cerebral blood flow may already be markedly reduced or mismatch between oxygen demand and supply may be present.¹⁵^,¹⁶ This principle should also be kept in mind when treating patients with other acute neurological disorders complicated by marginal brain perfusion.

Hypoventilation can lead to detrimental increases in PCO₂. Hypercapnia will increase cerebral blood volume and can lead to severe intracranial hypertension in patients with decreased cerebral compliance. Causes of hypoventilation in the PICU include deterioration of neurological status as a result of side effects of narcotic therapy. Children younger than 6 months are at higher risk for respiratory depression related to narcotics because of their immature liver metabolism. For example, the mean half-life of morphine has been reported to be 5.4 ± 3.4 hours in infants 1 week to 2 months of age and 2.6 ± 1.7 hours in infants aged 2 to 6 months.¹⁷ Developmental differences in pharmacokinetics should be taken into consideration when determining doses of narcotics in the PICU to minimize the risk for respiratory depression.

Although uncommon, severe respiratory failure may develop in critically ill children with primary neurological diagnoses. Current management strategies for critically ill children with severe respiratory failure include permissive hypoxia (saturation of ≥85%) and permissive hypercapnia. The potential benefits of these strategies should be balanced against the risk for secondary insults to the brain, especially in children without invasive brain monitoring, in whom the threshold for brain hypoperfusion is unknown.

Stroke in Children

There are a number of excellent reviews on stroke in children,¹⁸^–²⁷ including consensus statements that also address the controversies regarding treatment.²⁸^–³¹ This summary focuses on acute vascular insults, but it is important to consider that other disorders such as metabolic or mitochondrial disease, migraine, or seizures may mimic stroke.³² Timely and efficient evaluation for suspected stroke is therefore essential. The average interval between the onset of initial symptoms and medical evaluation is 34 hours.³³

The incidence of ischemic stroke in non-neonates is 2 to 6 per 100,000 children per year, and in neonates it is 1 in every 5000 live births.²⁶^,³⁴ The mortality and neurological morbidity associated with stroke in the young are substantial. Stroke is among the top 10 causes of death in children, and 60% of survivors have residual neurological deficits.^20,^25,^35,³⁶ Pooled data from patients with arterial ischemic stroke (AIS) in the United Kingdom and Canada who were not treated with antithrombotic drugs suggest a recurrence rate of nearly 50%.²⁹ This rate may be reduced (although no controlled trials have been performed) because studies in which most children were treated with aspirin or anticoagulation drugs report recurrence rates of 5% to 25%.³⁷^,³⁸ This question remains unresolved.

The causes of stroke in children differ substantially from those in adults, and large-artery atherosclerosis and small-vessel occlusive disease are not common pathologies in pediatric cerebrovascular disease (for review, see Lynch and colleagues ²⁶ and deVeber³⁹). For AIS, the most common risk factors are congenital or acquired heart disease and, less frequently, hypercoagulable or autoimmune disorders and arteriopathies (arterial dissection, moyamoya syndrome, and vasculitis).^21,^26,^40,⁴¹ In the United States, in the largest population-based study of AIS in children, the most common causes were cerebral arteriopathy (24%), infection (23%), and cardiac disease (12%), with sickle cell disease (SCD) accounting for just 3% of cases in this series.⁴² Among the arteriopathies, an association between antecedent varicella-zoster infection and increased risk for AIS has been reported in numerous studies,²¹ often involving the basal ganglia.⁴³ Similarly, the frequency of cervicocephalic arterial dissection is relatively higher in the young, in whom it accounts for 8%⁴⁴^,⁴⁵ to 20%³⁵ of cases in different series. The causes of moyamoya disease are not known, but there is a robust association with progressive neurological dysfunction, which occurs in up to 60% of cases⁴⁶ and may be reduced by revascularization surgery.⁴⁷ In most published studies, up to 25% of children have no identifiable risk factor.

The contribution of thrombophilias to stroke in the young has been addressed in multiple studies.^40,⁴⁸^–⁵⁰ Coagulation abnormalities implicated in the increased risk for stroke in children include deficiencies in protein C, protein S, or antithrombin III; activated protein C resistance; factor V Leiden mutation; prothrombin gene mutation (G20210A); methylenetetrahydrofolate reductase TT677 mutation; and antiphospholipid antibody syndrome. The contribution of a single biochemical or genetic risk factor to elevated risk for stroke is not precisely defined and is likely, given the relative rarity of childhood stroke, to be low.⁴⁸ In practice, this means that other factors should be investigated, even in the presence of a biochemical or genetically determined procoagulant state.²⁸

There are limited data and no controlled trials addressing the management of hemorrhagic stroke in children. Intracerebral hemorrhage (ICH) in the young has symptoms similar to those in adults, most commonly (59%) headache or vomiting, and is associated with seizures in 37%.⁵¹ For nontraumatic ICH, the most common risk factor in this study was an arteriovenous malformation (AVM) or arteriovenous fistula, which occurred in 34% (23 of 68) of cases. This frequency of AVM is comparable to that reported (47%) in a smaller study of 34 children with nontraumatic ICH.⁵²

Guidelines for the approach to diagnostic studies have been published.³⁰^,⁵³ A high index of suspicion is the essential first step for considering stroke in the differential diagnosis of new neurological deficits in children. Imaging studies should begin with computed tomography (CT) or magnetic resonance imaging (MRI) to first establish the diagnosis of stroke, identify the presence of hemorrhage, and delineate the extent of infarction. Computed tomography angiography (CTA) or magnetic resonance angiography (MRA) should be included if carotid or vertebral dissection is a possible mechanism. Venous sinus thrombosis is underdiagnosed, may be manifested by subtle signs, including seizures and subarachnoid hemorrhage or ICH,⁵⁴^,⁵⁵ and therefore merits a low threshold for the inclusion of computed tomography venography (CTV) or magnetic resonance venography (MRV) in the imaging studies. If the stroke is not in a vascular distribution or mitochondrial disorders are a consideration for any reason, magnetic resonance spectroscopy should by performed. If the results of MRA are normal and dissection, small-vessel vasculitis, AVM, aneurysm, or moyamoya disease are suspected, conventional angiography should be performed. Cardiac echocardiography is indicated for the evaluation of new stroke and should include bubble contrast enhancement. Based on the index of suspicion, a transesophageal study may be needed. Laboratory studies should include measures of coagulation, platelets, homocysteine, fasting cholesterol, triglycerides, and lipoprotein (a). Screening for infection may include antibodies to varicella-zoster²¹^,⁴³ and Mycoplasma and, if indicated, Chlamydia, Helicobacter, and Borrelia titers. Studies of genetic and biochemical risk factors that have been implicated in childhood stoke, together with von Willebrand factor antigen and plasminogen, may be performed along with other screening studies for autoimmune disorders, including lupus anticoagulant, as well as the erythrocyte sedimentation rate. Because anticardiolipin antibodies may be abnormal acutely, studies need to be repeated 8 to 12 weeks after the stroke, along with any abnormal biochemical measures of coagulation. If a metabolic disorder is suspected, lactate, pyruvate, serum amino acids, and urine organic acids should be obtained during the acute phase of the stroke, together with measurement of lactate levels in cerebrospinal fluid if available.

Stroke in Children with Sickle Cell Disease

Children with SCD pose a separate set of challenges for diagnosis and management of acute neurological deficits (for review, see Kirkham ⁵⁶ and Switzer and associates⁵⁷). By 45 years of age, 25% of patients with hemoglobin (Hb) SS disease and 10% of patients with HbSC disease have had a stroke.⁵⁸ In children with HbSS disease, up to 25% have had a “silent” infarction without detectable neurological deficits by adolescence.⁵⁹ The incidence of stroke in children with SCD (285 per 100,000 per year⁶⁰) is at least 20 times higher than the incidence (2.3 to 13.0 per 100,000 children¹⁸) in the general pediatric population. Overall, stroke is the primary cause of death in 10% of patients with SCD.⁶¹ Importantly, stroke in children with SCD may be accompanied by a diverse range of symptoms from difficulty with rapid alternating movements to headache, seizures, or coma.⁶² In practice, therefore, stroke should be the leading differential diagnosis in children with SCD and new neurological findings.

Strokes in children with SCD may occur as a result of multiple mechanisms, including arterial ischemia, intracerebral or subarachnoid hemorrhage,⁶³ aneurysm rupture, dissection, moyamoya syndrome,⁶⁴ or venous sinus thrombosis.⁵⁴ Thus, the imaging studies required for evaluation of acute neurological deficits in this population may include initial CT, MRI with diffusion weighting, MRA or CTA, and MRV or CTV. If imaging studies do not identify stroke as the mechanism, there should be a high index of suspicion for seizures, which occur 10 times more frequently in patients with SCD than in the general population.⁶²

Specific risk factors that increase the risk for stroke in patients with SCD have been identified and include previous transient ischemic attack, low hemoglobin levels, elevated white blood cell count, hypertension, and acute chest syndrome.⁵⁸ Indeed, children with acute chest syndrome are at significantly increased risk for stroke or stroke mimics, including posterior reversible leukoencephalopathy syndrome.⁶⁵ In SCD patients with increased flow velocity (identified by transcranial Doppler) in the middle cerebral artery or distal internal carotid artery, the annual risk for stroke increases from 1% to 10%.⁶⁶

The mainstay of stroke prevention in high-risk children with increased blood flow velocity and SCD is transfusion based on the 92% reduction in relative risk shown in the Stroke Prevention Trial in Sickle Cell Anemia (STOP) over a 2-year follow-up.⁶⁷ Such therapy has limitations, however. In this study stroke developed in only 15% of nontransfused subjects with high blood flow velocity. Furthermore, chronic transfusion therapy may result in complications such as infection, iron overload, and alloimmunization. Continued long-term transfusion therapy is required even when velocities return to a normal range.⁶⁸

For children with SCD and acute stroke, the initial focus of clinical management is exchange transfusion and hydration.²⁸^,⁵⁶ Exchange transfusion is undertaken to reduce the percentage of HbS to less than 30% of total hemoglobin, increase the percentage of HbA, and raise hemoglobin levels to 10 to 12.5 g/dL.³⁰ It should be noted that no controlled studies have proved the benefits of this approach, which is supported by class C evidence. Equally and again supported by class C evidence, children with acute ischemic stroke should receive adequate hydration and undergo correction of hypoxia and hypotension.²⁸

Data on pharmacologic prophylaxis for prevention of stroke in patients with SCD are limited. Antiplatelet therapy with aspirin did not reduce stroke frequency.⁶⁹ There is extensive evidence of the efficacy of statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) in the primary and secondary prevention of stroke in adults.⁷⁰^,⁷¹ However, there are no published data on their use in patients with SCD, and this merits further investigation.⁵⁷^,⁷²

Stroke—Anticoagulation and Supportive Therapy

In contrast to adults, in whom there is evidence-based recommendations for the treatment of AIS,⁷³ evidence-based data for children are more limited. There is universal consensus on the general goals of management: reestablishment of flow to the ischemic brain, minimization of secondary injury, and the principle that “time lost is brain lost.”²⁸^,⁷⁴ There are, however, no published data showing that such management improves outcome in children with stroke, although the authors regard this as standard of care. Guidelines for the management of children with AIS and ICH are available²⁸ and are summarized in Table 174-1. Notably, short-term anticoagulation may be considered for pediatric AIS, pending determination of the cause of the stroke. This is in contrast to the recommended practice in adults, for whom urgent anticoagulation is not recommended⁷³ given the absence of benefit⁷⁵ and increased risk for hemorrhage. In eligible adults within 3 hours of the onset of AIS, thrombolysis with tissue plasminogen activator (t-PA) improves outcome.⁷⁶ Intravenous t-PA for the treatment of AIS has been used safely in small pediatric series⁷⁷ and holds promise as therapy for pediatric stroke.

TABLE 174-1 Treatment Guidelines for Pediatric Stroke: Antithrombotic Treatments

As for TBI, management of AIS or ICH in the ICU should seek to minimize secondary injury. Strategies that aggressively treat fever, maintain euglycemia and normotension, and prevent hypoxia are adapted from the management of adults with AIS.⁷⁴ Although these recommendations for children are based on limited pediatric data (class I, level C evidence),²⁸ there is broad agreement that such management is indicated.^28,^30,⁵³

In the ICU, the challenges posed by children with stroke are threefold. First, there is the need to expedite diagnostic assessment of a child with suspected stroke so that therapy (including interventions to prevent secondary injury) can be started quickly if indicated. Second, there is the difficulty of identifying new cerebrovascular insults in critically ill children as a complication of other organ injury. This is often difficult in the sedated, intubated ICU patient population, so a high index of suspicion is warranted. Last, decisions about management of these patients will often need to be individualized because of limited or weak evidence-based guidelines for such management. Accordingly, decisions about anticoagulation, thrombolysis, blood pressure, temperature, and glucose management, as well as indications for ICP monitoring or hyperosmolar therapy, should be based on a consensus of the services involved. Criteria for escalation in therapy, up to and including thrombolysis and decompressive hemicraniectomy,⁷⁸ should be discussed early in the hospital course if needed. Serial neurological examination and frequent communication among members of the treating team are essential for early recognition of evolution of the neurological injury or a new insult and a rapid, coordinated approach to diagnostic studies and therapeutic intervention.

Fluids, Electrolytes, and Nutrition

Children with acute brain injury and those recovering from neurosurgical procedures commonly require intravenous fluids early in the ICU course. Maintenance intravenous fluids (1500 L/m² per day) are recommended to avoid dehydration and the consequent hemodynamic instability. The composition of maintenance intravenous fluids is tailored to avoid glucose and electrolyte abnormalities. Although hyperglycemia (blood glucose >180 mg/dL) is generally avoided because of its potential to worsen brain injury,⁷⁹^,⁸⁰ care should be taken to avoid hypoglycemia, especially in neonates and infants. Even though older children generally tolerate glucose-free intravenous fluids, infants and neonates frequently need glucose supplementation to avoid severe hypoglycemia and its related consequences, including seizures, alterations in mental status, and brain damage. Hypotonic fluids such as lactated Ringer’s solution and 0.25 normal saline solution are avoided because of the risk for hyponatremia, which can lead to worsening brain edema and even herniation.⁸¹^,⁸² In patients requiring fluid resuscitation to treat hypotension, normal saline solution is used as first-line therapy. The use of hypertonic saline (i.e., 3% saline solution) is considered a treatment option given its ability to restore hemodynamic stability and the added benefit of potentially reducing ICP.⁸³^,⁸⁴ Limited information is available on the safety and efficacy of the use of hypertonic saline solutions in children with brain injury.⁸⁵

Although the optimal time for initiation of enteral nutrition has not been established in children with brain injury, it is usually started as soon as possible and as tolerated by the patient’s neurological status. Guidelines for the management of severe TBI in children recommend replacement of 130% to 160% of resting metabolic expenditure, with nutritional support initiated by 72 hours after injury and full replacement by 7 days.⁸⁶ Although gastric feeding is commonly well tolerated, transpyloric (jejunal) feeding is an option in children with brain injury who have gastric intolerance because of gastroparesis. It is important to note that transpyloric feeding does not decrease the risk for vomiting and aspiration in critically ill children or adult patients with brain injury.⁸⁷^–⁸⁹ Additionally, vomiting can be a sign of neurological deterioration.

Fluid balance and electrolyte abnormalities can bee seen in children with acute brain injury and in the immediate postoperative period after neurosurgical procedures. Diabetes insipidus (DI) results from a deficiency of arginine vasopressin and can result in severe water and electrolyte imbalance.⁹⁰ It develops in approximately 75% of patients after transcranial resection of a pituitary tumor and in 10% to 44% after transsphenoidal pituitary surgery.⁹¹^,⁹² DI can also develop after other neurosurgical procedures such as ventricular fenestrations, as well as after TBI and cardiac arrest. In a few patients persistent DI will develop. Transient syndrome of inappropriate antidiuretic hormone secretion (SIADH) may develop in some patients as part of a triphasic manifestation (DI-SIADH-DI), or symptoms of cerebral salt wasting syndrome (CSWS) can develop.⁹²^–⁹⁸ Diagnostic criteria for postoperative DI are summarized in Table 174-2.

TABLE 174-2 Diagnostic Criteria for Postoperative Diabetes Insipidus

Polyuria (urine output >3 mL/kg/hr)

Serum sodium >145 mEq/L

Increased plasma osmolarity (>300 mOsm/kg)

Hypotonic urine (urine Osm <300 mOsm/kg; specific gravity <1.010)

Symptoms of DI can develop intraoperatively or in the immediate postoperative period (usually within the first 12 hours). The diagnostic criteria listed in Table 174-2 facilitate distinction between normal postoperative diuresis and DI. Management of children with DI can be complicated by overhydration or underhydration and electrolyte imbalance. These complications can be life-threatening when they result in hypotension and poor organ perfusion, severe hypernatremia, or hyponatremia. Hyponatremia can cause or worsen brain edema and seizures. These events can result in severe neurological injury.

Different treatment modalities for children with DI can be considered,⁹¹ including (1) aggressive free water replacement, (2) free water replacement with intermittent dosing of intranasal or oral vasopressin, and (3) free water replacement with continuous titration of intravenous vasopressin. The preferred approach is the one that would maximize patient safety and comfort. Although all three approaches have been used and can be considered, use of continuous, low-dose intravenous vasopressin with controlled fluid administration (1 L/m² per day plus additional fluid as needed to maintain fluid balance and hemodynamic stability) has been shown to produce better results in the perioperative period in children with postoperative DI who cannot regulate their own oral intake. This approach facilitates safe and effective management of children with perioperative DI while minimizing fluctuations in serum sodium concentrations and avoiding complications that can result from delays in effective therapy and inadequate fluid administration. This approach will also avoid the need for intense fluid replacement (intravenously or orally) for excessive urine output, especially during the immediate postoperative period. Satisfaction may be improved by minimizing excessive thirst postoperatively in patients who may not be able to maintain adequate oral intake because of their age or neurological status.

Other disorders of fluid and electrolyte balance seen in the postoperative period and in patients with acute brain injury include SIADH. Patients with SIADH have decreased plasma osmolality (<275 mOsm/kg H₂O), inappropriate urinary concentration (urine osmolality >100 mOsm/kg H₂O with normal renal function), urinary sodium loss (>40 mEq/L) with normal salt and water intake, and clinical euvolemia. A less common but important disorder, CSWS is characterized by hyponatremia, hypovolemia, natriuresis, and diuresis.⁹⁹ Severe hyponatremia can develop in patients with CSWS, and they require careful titration of fluid intake and sodium replacement. Given the detrimental effects that delayed diagnosis and treatment can have in patients with the disorders just described, our pediatric neurocritical care team has created a bedside document outlining diagnosis, monitoring, and treatment parameters for postoperative DI. This document is available in the Appendix.

Intracranial Hypertension

ICP monitoring techniques and treatment principles are described in Chapters 24 and 25. TBI and unique pediatric aspects are discussed in Chapter 214. The main indication for ICP monitoring in critically ill children is TBI. Children of all ages, including infants, are at risk for the development of intracranial hypertension and the consequent cerebral hypoperfusion and cerebral herniation. ICP monitoring in patients with medical conditions other than TBI is highly controversial. Examples of reversible brain pathology that results in increased ICP include meningitis and liver failure with encephalopathy. In patients with meningitis, ICP monitoring is used in an estimated 7% of patients.¹⁰⁰ Mortality is higher in patients with meningitis and mean cerebral perfusion pressure (CPP) less than 50 mm Hg despite CPP-directed therapies.¹⁰¹ In patients with liver failure and encephalopathy, the potential benefit of ICP-directed therapy should be balanced against the risk for complications.¹⁰²^,¹⁰³ Global ischemia from near drowning or cardiac arrest with consequent brain edema can result in increased ICP, but the incidence in children is not known, and because these patients are not routinely monitored, there are no data on the benefit of monitoring ICP and other parameters with the goal of optimizing cerebral perfusion and balancing cerebral metabolism. Limited reports suggest that ICP monitoring and interventions aimed at lowering ICP are not beneficial in patients after global ischemia.¹⁰⁴ Overall, although ICP monitoring is feasible in conditions other than TBI, no information on the safety and efficacy of ICP-directed therapies is available.

Seizures

The approach to seizures in the ICU is based on the need to detect their occurrence, terminate them, and minimize any secondary injury that they may produce. Seizures may be the initial symptom of new ischemic (venous or arterial) or hemorrhagic injury in critically ill patients. In parallel with acute management, investigation of common ICU causes of seizures (metabolic derangement, drug reaction, vascular injury, infection) should begin immediately. Here, we focus on the detection of seizures, including NCSs, and the management of status epilepticus (SE), including RSE.

As is the case for stroke, time is of the essence for the treatment and prevention of SE. The duration of seizures required for the definition of SE has been shortened to 30 minutes,¹⁰⁵ and a duration as short as 5 minutes has been proposed.¹⁰⁶ This variation in operational definition has come about in response to important observations from animal studies that within 15 minutes repetitive seizures become self-sustaining and pharmacoresistant.¹⁰⁷^–¹⁰⁹ A further refinement of the definition of SE has been proposed to include “impending” SE (continuous or intermittent seizures lasting more than 5 minutes without full recovery of consciousness between seizures) and “established” SE (clinical or electrographic continuous seizures lasting more than 30 minutes without full recovery of consciousness between seizures).¹¹⁰ Criteria for RSE include persistence of seizure activity despite appropriate medical and anticonvulsant drug (ACD) therapy, although the duration of seizures (1 to 2 hours)¹⁰⁵^,¹¹¹ and the number of drugs (two or three)¹¹² vary among studies.

SE and RSE are associated with a significant increase in mortality in adults and children.¹¹³^–¹¹⁶ In a series of 22 children with RSE, 33% died and none of the previously healthy children recovered normal neurological function.¹¹⁵ Overall, a meta-analysis of pediatric RSE indicates a mortality of 16%.¹¹²

Early recognition of clinical or electrographic seizures is an important factor in enhancing the efficacy of ACDs to terminate the seizures and thereby prevent the additional metabolic stress that seizures impose on the injured brain. However, detection of seizures in the ICU is challenging. NCSs or SE may be difficult to diagnose and are often unrecognized in comatose patients ¹¹⁷ in both adult and pediatric ICUs.^81,^84,⁸⁵ Convulsive seizures or NCSs are reported in up to 50% of patients undergoing treatment in an ICU when examined by continuous EEG (cEEG) monitoring¹¹⁸^–¹²⁰; such patients often have nonconvulsive SE.¹²¹ Importantly, routine EEG recording misses NCSs in approximately 50% of critically ill patients whose NCSs were detected with cEEG monitoring.¹²⁰^,¹²² Where there is a high index of suspicion for NCSs, cEEG monitoring for at least 12 hours is required to detect NCSs.^120,^123,¹²⁴

Protocols for the treatment of SE are well established (for review see Walker and Teach ¹²⁵ and Glauser and colleagues¹²⁶). The typical protocol involves intravenous, rectal, or buccal administration of a short-acting benzodiazepine, commonly followed by (depending on age, current ACD use, and seizure type) phenytoin, phenobarbital, or valproic acid. A protocol in use at one of our (M.W.) institutions is shown in Figure 174-1. Key features of this approach in the ICU are the need for early identification of seizures, possibly requiring cEEG monitoring, the ready availability of ACDs, and attention to causes of the seizures in high-risk critically ill children.

FIGURE 174-1 Inpatient guidelines for the management and evaluation of status epilepticus. ABCs, airway, breathing, circulation; BP, blood pressure; CBC, complete blood count; D10W, 10% dextrose in water; IV, intravenously; NICU, neonatal intensive care unit; PE, phenytoin equivalent; PHT, phenytoin; PICU, pediatric intensive care unit; PR, per rectum; RR, respiratory rate.

Many patients in the ICU will meet the criteria for RSE. In contrast to SE, the approach to the management of RSE is not as well defined. Treatment protocols for adult ¹²⁷ and pediatric¹²⁸ RSE have been proposed. After failure of first-line ACDs, treatment with midazolam followed by pentobarbital or other coma-inducing drugs is initiated. Other agents such as ketamine¹²⁹ or topiramate¹³⁰ have been shown in case series to be effective and may also be tried. Use of midazolam as the first-line agent for RSE in children is based on a number of studies showing efficacy¹³¹ (for review see Abend and Dlugos¹²⁸). A protocol for the treatment of RSE in use in the ICU at one of our (M.W.) institutions is shown in Figures 174-3 and 174-4. This approach combines rapid escalation of the primary treatment (midazolam) with defined targeted physiologic parameters and selected diagnostic studies. This approach seeks to treat RSE in the context of requirements for maintaining cardiovascular function.

FIGURE 174-3 Management of refractory status epilepticus. AED, antiepileptic drug; BP, blood pressure; cEEG, continuous EEG; EEG, electroencephalogram; ICP, intracranial pressure; IV, intravenously; IVF, intravenous fluid; NG, nasogastrically; PB, phenobarbital; PE, phenytoin equivalent; PHT, phenytoin; PO, orally; PTB, pentobarbital; RSE, refractory status epilepticus; VPA, valproic acid.

FIGURE 174-4 Secondary management of refractory status epilepticus. BP, blood pressure; CBC, complete blood count; CMV, cytomegalovirus; CPK, creatine phosphokinase; CSF, cerebrospinal fluid; CT, computed tomography; CTA, computed tomography angiography; CTV, computed tomography venography; EBV, Epstein-Barr virus; ECHO, echocardiography; EEG, electroencephalogram; GAD, glutamic acid decarboxylase; HHV-6, human herpesvirus 6; ICP, intracranial pressure; IVF, intravenous fluid; LFTs, liver function tests; MAP, mean arterial pressure; MRA, magnetic resonance angiography; MRI, magnetic resonance imaging; MRV, magnetic resonance venography; NG, nasogastically; NH₃, ammonia; PB, phenobarbital; PHT, phenytoin; PTB, pentobarbital; Pyr, pyruvate; RSE, refractory status epilepticus; VPA, valproic acid.

Infections

Encephalitis and bacterial meningitis are reviewed in detail in Chapter 45. Here, we focus on an overview of infectious issues most relevant to pediatric neurocritical care.

Mortality from bacterial meningitis in children has decreased to as low as 2%.¹³² In an Australian series of 122 children with pneumococcal meningitis, 15 died and 39 had permanent neurological deficits.¹³³ In a U.S. series of 559 pediatric patients, 7% died in the PICU and mortality was higher in those in coma or shock on admission to the ICU.¹³⁴ Treatment with dexamethasone improves outcome in children with all-cause bacterial meningitis, including both Haemophilus influenzae¹³⁵^,¹³⁶ and pneumococcal meningitis.¹³³ However, the use of adjunctive steroids for the treatment of pneumococcal meningitis is controversial, and practice varies in the United States.¹³²

The fundamental principles of prevention of secondary neurological injury also apply to the management of meningitis. Essential in optimizing neurological outcome are maintenance of normal oxygenation and adequate CPP, prevention of hypoglycemia and hyponatremia, and early detection of seizures, AIS, ICH, and cerebral sinus venous thrombosis.¹³⁷^,¹³⁸ The maximum elevation in ICP occurs in the first 24 to 48 hours after diagnosis.¹³⁹ A retrospective cohort study of U.S. meningitis cases reported ICP monitoring in 7% of patients.¹⁴⁰ There was no difference in mortality between groups, consistent with a previous series reported from 15 U.S. PICUs.¹³⁴ Taken together, variations in practice and lack of demonstrated effect on outcome render impossible any recommendations for or against ICP monitoring in children with bacterial meningitis.¹⁴⁰

Tuberculous meningitis (TBM) may occur as a reactivation of latent infection or as a new infection (for review, see Thwaites and Hien ¹⁴¹). Early recognition is important to enhance the chance of good neurological outcome but is difficult given the nonspecific prodrome and, commonly, lack of fever.¹⁴² In one adult series, 28% of subjects reported a headache, 13% were febrile, and only 2% reported meningitis symptoms.¹⁴³ Mortality and neurological morbidity are high, with a reported mortality of 13% and permanent neurological sequelae in 45% of 38 children in a British series.¹⁴⁴ Strikingly, there were no deaths or neurological sequelae in children vaccinated with bacille Calmette-Guérin (BCG). The neurological complications of TBM are myriad, but hydrocephalus, stroke, and seizures are common and TBM should be considered in the differential diagnosis of any neurological deficits.¹⁴⁵ Hydrocephalus is more frequent in children (80% to 90%) than adults (12%) with TBM.¹⁴² Accordingly, serial lumbar punctures or external ventricular drainage may be considered both to prevent the need for ventricular shunting and to increase the potential for benefit from shunt placement,¹⁴⁶^,¹⁴⁷ despite the lack of data from controlled trials. The use of adjunctive corticosteroid treatment remains controversial, but cumulative evidence¹⁴⁸ suggests that dexamethasone reduces morbidity but not mortality and should be administered to all patients regardless of age or disease severity.¹⁴¹

In the developing world, cysticercosis is the most common cause of adult-onset epilepsy.¹⁴⁹ For patients from regions with endemic disease, neurocysticercosis should be considered in the differential diagnosis of brain calcifications and seizures. Criteria for treatment with anthelmintic agents and the optimal agents for such use have not been established (for review see Nash and colleagues¹⁵⁰).

The neurological manifestations of viral encephalitis in children are protean and are most commonly manifested in the ICU setting as seizures. The most common agents are the nonpolio viruses, enteroviruses, respiratory viruses, and herpes group viruses, including Epstein-Barr virus, although a specific agent may be identified in just 40% of cases.¹⁵¹ In adults, the occurrence of seizures during the acute phase of a CNS infection significantly increases the risk for unprovoked seizures thereafter, with the risk being greatest in the first 5 years but remaining elevated (22%) over 20 years after recovery.¹⁵² In children, risk factors associated with a poor outcome after encephalitis include SE during the acute phase of encephalitis and herpes simplex as the etiologic agent.¹⁵³ Neurological complications of the common respiratory syncytial virus are rare but include seizures, cardiac arrest, and postulated brainstem dysfunction.¹⁵⁴ Finally, new CNS infection or reactivation of latent human herpesvirus 6 (HHV-6) should be considered in any immunosuppressed patient in the ICU setting with a change in neurological function.¹⁵⁵ Limbic encephalitis in these patients may be manifested as an inability to sleep and short-term memory loss.¹⁵⁶ A high index of suspicion is needed in any transplant recipient with these symptoms given the risk for long-term cognitive impairment associated with HHV-6 encephalitis.¹⁵⁷

For any CNS infection requiring ICU care, seizures should be suspected either as a complication of infection or in the setting of deterioration of neurological function.¹⁴¹ Frequently, seizures in this population will be nonconvulsive, and cEEG monitoring will typically be necessary for detection of electrographic seizures in the ICU.¹²⁰^,¹⁵⁸

Prevention of Secondary Insults to the Brain

The main objective of intensive care management of children with or at risk for brain injury is to prevent secondary insults to the brain. This goal may be achieved by direct and focused attention on variables associated with poor neurological outcome. Neurocritical care in children aims at preventing cerebral hypoxia and ischemia through avoidance of intracranial hypertension and systemic hypoxia, hypotension, and hypocapnia. Cerebral ischemia may occur as a direct result of inadequate cerebral perfusion, decreased oxygen or glucose supply, increased metabolic demand of the brain, or increased cerebral vascular resistance.

Inadequate perfusion may have a direct and deleterious effect on the pathophysiology of brain injury. Cerebral perfusion is generally described in terms of CPP, which is calculated as the difference between mean arterial pressure (MAP) and mean ICP: CPP = MAP − ICP. Current recommended guidelines suggest that CPP be maintained around 70 mm Hg in adult patients. CPP between 40 and 65 mm Hg has been recommended for children with severe TBI.¹⁵⁹ Based on data from adult patients, CPP of 60 to 70 mm Hg is recommended for adolescents. Children with CPP below these thresholds have worse outcomes. Optimal CPP levels for children younger than 2 years have not been established, but several studies have demonstrated that CPP of 40 mm Hg or less is associated with higher mortality and worse outcome in children of any age. We therefore consider 45 mm Hg a critical threshold for CPP in children younger than 2 years. Using normal MAP values for age and 10 mm Hg to estimate normal ICP, threshold levels of 50 mm Hg (2 to 6 years), 55 mm Hg (7 to 10 years), and 60 mm Hg (11 to 16 years) can be estimated. CPP values should be interpreted in the context of other determinants of brain tissue perfusion such as cerebral blood flow autoregulation.¹⁶⁰^–¹⁶³ Adjustments in MAP to maintain CPP above critical thresholds may be necessary in children with disrupted autoregulation. Because ICP and blood pressure are commonly monitored in children after TBI, CPP values are usually available and help guide treatment. In critically ill children with other acute neurological disorders, ICP data are much less commonly available, thus making estimations of critical thresholds for brain perfusion difficult.

FIGURE 174-2 Guidelines for the management of refractory status epilepticus. CMV, cytomegalovirus; CPK, creatine phosphokinase; CSF, cerebrospinal fluid; CT, computed tomography, CTV, computed tomography venography; EBV, Epstein-Barr virus; EEG, electroencephalogram; ER, emergency room; HHV-6, human herpesvirus 6; ICH, intracranial hemorrhage; ICU, intensive care unit; IV, intravenously; LFT, liver function tests; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; MRV, magnetic resonance venography; NG, nasogastrically; NH₃, ammonia; PE, phenytoin equivalent; PO, orally; PR, per rectum; Pyr, pyruvate; SE, status epilepticus; TBI, traumatic brain injury.

Even in the absence of increased ICP, low blood pressure has been shown to be a predictor of poor outcome in patients with brain injury.¹⁶⁴ One of the difficulties we face while caring for children with a brain injury is determining an appropriate systolic blood pressure and MAP for a specific disease state. Additionally, normal systolic blood pressure and MAP values change with age (Table 174-3).¹⁶⁵ The formula 70 mm Hg + (2 × age in years) allows easy calculation of the lower limit (fifth percentile) of systolic blood pressure for age.

TABLE 174-3 Normal Mean Arterial Blood Pressure in Children

AGE (yr)	NORMAL MAP (mm Hg)
1-2	53-59
3-4	61-65
5-6	67-69
7-8	70-72
9-10	73-75
11-12	75-77
13-14	77-79
15-17	80-84

Adapted from Haque IU, Zaritsky AL. Analysis of the evidence for the lower limit of systolic and mean arterial pressure in children. Pediatr Crit Care Med. 2007;8:138.

Different definitions of hypotension have been used. A recent study using the 75th percentile for age-appropriate systolic blood pressure as the threshold for hypotension in children with TBI described an association between early hypotension and outcome.¹⁶⁶ Maintenance of age-appropriate systolic blood pressure during the initial stabilization phase is important and may influence outcome in children with TBI and possibly other types of acute brain injury. As new technology becomes available, it may be possible to more effectively determine adequate brain perfusion thresholds in patients with nontraumatic brain injury. In the meantime, careful attention should be given to systemic hypotension.

Other factors that may result in secondary insults to the brain include hyperthermia, anemia, decreased cardiac output, seizures, pain and agitation, cerebral edema, and CNS thrombosis. Hyperthermia is known to correlate with worse outcome and longer ICU stay in patients with brain injury.¹⁶⁷^–¹⁶⁹ Avoidance of hyperthermia is important during periods of brain ischemia. Although data in children are limited, both medications and external cooling methods to treat hyperthermia in ICU patients appear to be ineffective in a significant proportion of patients.¹⁷⁰^–¹⁷² Intravascular cooling methods are more effective but unavailable for children. Promising new external cooling technology may become available for children in the near future.¹⁷²

Anemia and low cardiac output can aggravate brain injury and should be avoided. There is no clear threshold for transfusion in children, and although hemoglobin concentrations of 7 g/dL are well tolerated in stable children in the ICU,¹⁷³ no data are available for children with marginal brain perfusion. Recent information has also raised concern about the effects of transfusion of stored blood on brain tissue perfusion.¹⁷⁴

As for cardiac output, although coronary artery disease is much less common in children, pediatric patients can suffer from low cardiac output secondary to systemic hypoperfusion, severe brain injury, or the effect of medications such as pentobarbital, narcotics, and benzodiazepines. Propofol in particular has been reported to rarely result in refractory shock and metabolic acidosis. This medication is not recommended for continuous sedation in the PICU (http://www.fda.gov/medwatch/SAFETY/2001/safety01.htm#dipriv).

Pain and agitation resulting in hypermetabolism and increased ICP can also aggravate brain injury. The beneficial effects of sedatives and analgesics are balanced against the effects of these medications on the child’s neurological examination. In patients undergoing mechanical ventilation and ICP monitoring, continuous sedation and analgesia are more desirable than intermittent administration. In patients recovering from intracranial procedures such as resection of posterior fossa tumors, intermittent careful titration of analgesics is preferred, and consideration can be given to non-narcotic analgesics such as ketorolac in the absence of obvious contraindications.¹⁷⁵^–¹⁷⁷ Ketorolac appears to be well tolerated in children after significant surgical procedures without increasing the risk for bleeding and renal dysfunction, but its safety and efficacy in neurosurgical patients have not been established. Ketorolac is typically prescribed for 48 hours or less.

Finally, although less common, deep venous thrombosis is possible in children with brain pathology.¹⁷⁸ Risk factors include immobility, obesity, thoracic or abdominal surgery, need for central venous access, and procoagulant states.¹⁷⁹^,¹⁸⁰ As children recover and tolerate external stimulation, physical and occupational therapy interventions will increase mobility and decrease the risk for venous thrombosis. Anticoagulation prophylaxis is indicated in selected cases. Treatment of cerebral venous thrombosis, a life-threatening condition in critically ill children, is rather complex and more studies are needed.¹⁸¹ The risk of bleeding should be balanced against the risk of expanding thrombosis, which can lead to severe brain injury.

Diagnosis of Brain Death in Children

Guidelines for determining death, including brain death, have developed over time at many institutions based on the Harvard criteria.¹⁸² In 1981, both national policy and state law regarding the determination of death were established.¹⁸³ These directives indicate that as with irreversible failure of circulation and respiration, irreversible failure of brain function (including the brainstem) is death. Criteria for establishing brain death in children were published in 1987 in a statement issued by the American Academy of Pediatrics.¹⁸⁴ These principles are consistent with the report by the Quality Standards Subcommittee of the American Academy of Neurology and address additional aspects of the determination of brain death that are unique to children.¹⁸⁵

Before the determination of brain death, the patient must be known to have an irreversible disease that can cause brain death. Reversible disorders that may result in loss of brain function must be excluded. Examples of such recoverable disorders include drug poisoning, toxin, metabolic disorders, severe electrolyte disturbances, hypothermia, and shock. The degree of hypothermia that could interfere with a brain death examination is in general not well defined. While defining such a threshold at a given institution, consideration should be given to the fact that brainstem reflexes can be lost at 32°C.¹⁸⁶ Additionally, a temperature of 36.5°C or lower can create difficulty in the interpretation of target carbon dioxide partial pressure (PCO₂) values. Controversy surrounding the need to correct arterial blood gas values to the patient’s temperature is avoided at normothermia. Given the difficulty in achieving normothermia in children with loss of brain function and the paucity of data supporting a specific lower limit for temperature, a threshold of 35°C can be considered feasible and appropriate.

To establish brain death, there must be complete loss of brain function. The patient must have no spontaneous movement or vocalization and must not respond to noxious stimuli. There must be no decorticate or decerebrate posturing. Rudimentary spinal reflex responses may be present, such as deep tendon reflexes, plantar reflexes, triple flexion of the legs, and superficial abdominal reflexes. The pupils may be either widely dilated or midway between constriction and dilation, and stimulation with a bright light must fail to produce any pupillary constriction. The eyes must be motionless. Oculocephalic (doll’s eye) and oculovestibular (ice-water caloric) responses must be absent. The gag, cough, sucking, and rooting reflexes must be absent. Finally, to document the absence of spontaneous ventilatory efforts, an apnea test can be performed if there is no contraindication.¹⁸⁷ Before interrupting the ventilator, the patient should be ventilated with 100% oxygen for at least 5 minutes to minimize the risk for hypoxia. Blood gas analysis should be performed to demonstrate that the patient has adequate acid-base balance and acceptable ventilation (PCO₂ of 35 to 45 mm Hg) and oxygenation (PO₂ of 100 mm Hg or higher). If any of these criteria are not met, the patient may not tolerate the apnea test without hypoxia developing. Although different methods have been described, a technique involving transition from the mechanical ventilator to a self-inflating resuscitation bag may reduce the risk for barotrauma and hypoxia.¹⁸⁸ The bag is set for 100% oxygen and positive end-expiratory pressure (PEEP) that approximates the patient’s current mechanical ventilator settings. The patient is then observed constantly for 10 to 15 minutes. If no spontaneous effort at ventilation is observed at a PCO₂ of 60 mm Hg or 20 mm Hg above baseline, the test is interpreted as consistent with the absence of brainstem function. If circulatory instability occurs at any time, the patient should be reconnected to the ventilator. Patients with lung disease or chronic CO₂ retention may tolerate only a shorter period of observation, at least 2 to 3 minutes and preferably 5 minutes. Criteria for discontinuation of the apnea test include a decrease in oxygen saturation to less than 85% to 90%, hypotension, any arrhythmia associated with hemodynamic instability, and presence of respiratory effort.

Although a specific interval between the two required brain death evaluations is optional in patients 18 years and older, because of the difficulty in performing a brain death examination in younger patients, age-related observation periods between examinations have been recommended (Table 174-4). A single apnea test is typically conducted at the time of the second examination. The apnea test supports but is not required for the diagnosis of brain death.

TABLE 174-4 Age and Recommended Observation Periods for the Diagnosis of Brain Death

7 days to 2 months: Two examinations separated by 48 hr

2 mo to 1 yr: Two examinations separated by 24 hr

>1 yr: An observation period of 12 hr is recommended. In some circumstances, a shorter interval (4-6 hr) may be preferred

Ancillary tests such as a nuclear medicine perfusion scan, EEG recordings, cerebral angiography, and transcranial Doppler constitute options that can be used when deemed likely to be useful, but it is important to recognize that these ancillary tests have limitations.¹⁸⁹^,¹⁹⁰ These tests support the diagnosis but are not required to diagnose brain death. A nuclear medicine perfusion scan may be preferred when drugs that may suppress EEG activity are present (i.e., pentobarbital).

Once the determination of brain death is made, appropriate documentation in the patient’s medical record and communication with the patient’s family are important.¹⁹¹ The use of a standardized documentation form may facilitate accurate and complete documentation of the brain death examination. One such form, used at our institution (J.P.), is available in the Appendix. At this point it may also be appropriate to discuss the recovery of organs for humanitarian purposes. Permission from the family may be requested for evaluation of the patient by an organ and tissue procurement agency.