Evaluation of Coma and Brain Death

David J. Michelson and Stephen Ashwal

Coma is a nonspecific sign of widespread central nervous system impairment resulting from various metabolic and
structural etiologies. The rapid recognition of this neurologic emergency and results from the history, physical
examination, and early investigative studies are key to the identification and treatment of its underlying cause. The
prognosis for recovery depends greatly on the underlying etiology as well as on its optimal treatment, which seeks
to preserve neurologic function and maximize the potential for recovery by reversing the primary cause of brain
injury, if known, and preventing secondary brain injury from anoxia, ischemia, hypoglycemia, cerebral edema,
seizures, infections, and electrolyte and temperature disturbances. Brain death must be diagnosed with similar
care and precision, and families approached compassionately about the diagnosis and their decisions regarding
organ donation.
© 2004 Elsevier Inc. All rights reserved.

N ORMAL CONSCIOUSNESS is recognized

with little effort, but is considerably more
difficult to describe. Traditionally, consciousness
state, the minimally conscious state (MCS), and
akinetic mutism (Table 1). Patients may progress
through these states as they recover from or dete-
has two components: arousal or wakefulness, and riorate toward coma.
awareness of the self and the environment. Arousal In the vegetative state, environmental awareness
is dependent on ascending sensory stimulation of is absent, although hypothalamic and brainstem
brainstem, hypothalamus, thalamus, and, ulti- (vegetative) functions are sufficient to allow for
mately, the cerebral cortex. Conscious awareness, prolonged survival with supportive care.3 Such
although difficult to localize with certainty, arises patients often have some preservation of both re-
from the cerebral cortex and its interactions with spiratory function and sleep-wake cycles, in which
the major subcortical nuclei.1 Impaired conscious- periods of apparent sleep alternate with periods of
ness can result from alterations in either arousal or spontaneous eye opening. Movements in response
awareness (although awareness is dependent on at to external stimuli can occur; these may be inter-
least partial arousal), or from loss of both. Coma is preted by hopeful observers as being purposeful
defined as a pathological state of deep and sus- and voluntary, but in reality are reflexive, unsus-
tained (⬎1 hour) unconsciousness with the eyes tained, and unreproduceable.4 A vegetative state is
closed, distinguishable from normal sleep by the considered permanent if it lasts longer than 12
inability to be aroused.2 months after traumatic brain injury or longer than
3 months after non-traumatic brain injury.
IMPAIRMENT OF CONSCIOUSNESS In contrast, patients in MCS can reproduceably
demonstrate purposeful (even if severely limited)
Numerous descriptive terms for states of acute nonreflexive movements or affective behaviors,
altered awareness, spanning a continuum between such as simple command following, gestural or
normal consciousness and coma, are frequently verbal responses to questions, intelligible verbal-
used, although this is often done without precision. izations, smiling or crying in response to the emo-
One principal distinction between these various tional content of stimuli, reaching accurately to-
states is whether alertness is heightened or re- ward the location of an object, or visual pursuit or
duced. A heightened mental state, occasionally fixation in response to visual stimuli.5 A person in
with hypervigilance, is seen in patients with delir- a MCS can easily be misdiagnosed as being in a
ium, delusions, and hallucinations. Mental states
with reduced alertness range from lethargy, in
which patients appear sleepy without stimulation, From the Division of Child Neurology, Department of Pedi-
to obtundation, in which patients appear abnor- atrics, Loma Linda University School of Medicine, Loma Linda,
mally drowsy even when stimulated, to stupor, in CA.
which even less responsiveness is elicited only Address reprint requests to David J. Michelson, MD, Loma
with continuous and vigorous stimulation. Linda University School of Medicine, Coleman Pavilion, Room
A1120G, 11175 Campus Street, Loma Linda, CA 92350.
Chronic states of severely impaired conscious- © 2004 Elsevier Inc. All rights reserved.
ness from which coma can be clinically differen- 1071-9091/04/1102-0000$30.00/0
tiated have been defined, including the vegetative doi:10.1016/j.spen.2004.03.010

Seminars in Pediatric Neurology, Vol 11, No 2 (June), 2004: pp 105-118 105

106 MICHELSON AND ASHWAL

Table 1. Severe Disorders of Consciousness and Related Conditions

Self- Pain and Sleep-Wake Respiratory

Condition Awareness Suffering Cycles Motor Function Function Outcome

Brain death Absent Absent Absent None or only spinal Absent Death
reflexes
Coma Absent Absent Absent No purposeful Variable Variable
movements
Vegetative state Absent Absent Present No purposeful Present Variable
movements
Minimally conscious Minimal Present Present Limited purposeful Present Unknown
state movements
Akinetic mutism Present Present Present Limited Variable Limited
recovery
Locked-in syndrome Present Present Present Absent; may spare Variable Variable
vertical eye
movements

PVS without the aid of repeated assessments by a both cerebral hemispheres or disrupt the ascending
multidisciplinary team with experience in the man- brainstem and diencephalon activating systems.
agement of complex disabilities.6 The functional impairment itself may be caused by
Patients with locked-in syndrome have a largely destructive or compressive lesions, which most
intact consciousness but a severely limited ability commonly result from severe head trauma, or met-
to communicate their awareness due to paralysis of abolic disturbances, which typically result from
voluntary muscles. This condition is rare, espe- nontraumatic causes and have diffuse effects.2
cially in children, but can occur with brainstem Traumatic and nontraumatic causes of coma
injuries sparing the midbrain resulting from have roughly equal incidences of about 30 per
trauma,7 pontine glioma,8 or basilar artery occlu- 100,000 children, with nontraumatic causes seen
sion9 or with profound neuromuscular dysfunction more frequently in younger children, particularly
from Guillain-Barré syndrome,10 spinal muscular in infants, in whom congenital heart malformations
atrophy,11 botulism, or organophosphate toxic- and inborn errors of metabolism are most likely to
ity.12 become symptomatic. In a prospective, population-
Akinetic mutism is a state of profound apathy based study of nontraumatic coma in children,
with preserved awareness, revealed by attentive infection was quite common, accounting for 38%
visual pursuit, but a paucity and slowness of vol- of cases, followed by accidental and deliberate
untary movements not due to paralysis. It is seen intoxication (exogenous toxins), prolonged sei-
with bilateral injuries within the paramedian mid- zures, congenital malformations of the brain or
brain, basal diencephalon, or inferior frontal lobes, heart, accidental hypoxic-ischemic injury, and
occurring with traumatic brain injury,13 hydro- metabolic diseases (endogenous toxins), including
cephalus,14 central nervous system (CNS) infec- diabetic ketoacidosis.19 A list of the common eti-
tion,15 tumors,16 and tumor resection.17 ologies is presented in Table 2.
Profound coma due to an irreversible destructive
brain injury and the absence of brainstem function History and Physical Examination
are essential components shared by all definitions Although the initial evaluation must be rapid
of brain death, although variation exists in the and the history and physical exam abbreviated,
requirement for destruction of both the cortex and identification of the underlying cause of coma is
brainstem and in the confirmatory testing needed.18 crucial for optimal management and must be
sought out in the most complete manner possible.
CAUSES OF COMA Coma may result from the progression of a known
Coma is a nonspecific consequence of various illness or injury or have no obvious precedent.20 A
CNS insults, but generally is seen only in condi- sudden loss of responsiveness is suggestive of
tions that either cause widespread dysfunction of intracranial hemorrhage, seizure, cardiac arrhyth-
COMA AND BRAIN DEATH 107

Table 2. Potential Etiologies for Coma intracranial involvement or cardiorespiratory fail-

Metabolic ure. However, fever, immune system activation, or
Electrolyte, acid-base, glucose, water, or thermal fasting associated with infection may secondarily
disturbances precipitate impaired consciousness due to seizures,
Inborn errors of metabolism
acute disseminated encephalomyelitis (ADEM),
Organic acidemias and amino acidemias
Urea cycle defects diabetic ketoacidosis, Reye’s syndrome, or an in-
Carbohydrate disorders born error of metabolism. Nonaccidental head
Mitochondrial and carnitine disorders trauma is rarely witnessed or confessed to, but may
Fatty acid oxidation defects be accompanied by various external signs of abuse
Leukodystrophies: CACH, MLSC
or suspected on the basis of a vague or inconsistent
Nutritional: thiamine, pyridoxine, or niacin deficiency
Toxins: accidental or intentional ingestion, overdose, or history.21
inhalation The evaluation begins with vital sign measure-
Medications: overdose or adverse reaction ment with continuous monitoring. Hyperthermia is
Organ dysfunction suggestive of systemic and/or intracranial infec-
Endocrine: pituitary, thyroid, pancreas, adrenal
tion, but is also seen with inflammatory causes of
Nonendocrine organ
Renal failure: uremia coma, such as ADEM, as well as with heat stroke,
Hepatic failure: encephalopathy, Reye’s syndrome neuroleptic malignant syndrome, status epilepti-
Cardiac failure: shock cus, and anticholinergic poisoning. Hypothermia
Respiratory failure: hypoxia, CO2 narcosis occurs with infection, especially in infancy, but is
Infectious
more often due to drug intoxication or environ-
Systemic
Cerebral mental exposure. Tachycardia occurs with fever,
Meningitis or encephalitis pain, hypovolemia, cardiomyopathy, or tachyar-
Cerebritis, empyema or abscess rhythmia. Bradycardia occurs with hypoxic-isch-
Paroxysmal emic myocardial injury and as part of the Cushing
Seizure: convulsive, nonconvulsive, or postictal
triad of increased intracranial pressure (ICP), along
Migraine: “basilar”
Traumatic with hypertension and irregular respirations.
Intracranial hemorrhage or contusion Tachypnea occurs with respiratory compromise, as
Diffuse axonal injury compensation for metabolic acidosis, or with pon-
Neoplastic tine injury. Slow, irregular, or periodic respirations
Infiltration or edema: herniation, hydrocephalus, seizures
occur with metabolic encephalopathies, such as
Chemotherapy or radiation therapy toxicity
Vascular barbiturate or benzodiazepine intoxication, or with
Ischemia: thrombosis, embolism, strangulation, or arterial brainstem injury. Hypotension can suggest hypo-
dissection volemic, septic, or cardiogenic shock, toxic inges-
Hemorrhage: AVM, aneurysm, coagulopathy, atrophy, or tion, or adrenal insufficiency. Hypertension may be
trauma
due to pain, toxic ingestion, renal failure, or in-
Venous sinus thrombosis
Structural creased ICP.
Hydrocephalus: communicating or noncommunicating Further examination should include a complete
Inflammatory inspection of the head, scalp, and skin (removing
Vasculitis all clothing) and turning to visualize the back. Skin
Acute disseminated encephalomyelitis
color can suggest hypoxia, jaundice, anemia, or
Psychiatric: conversion, malingering, or Munchausen by
proxy carbon monoxide poisoning. Bruised or swollen
areas of the face and head suggest a traumatic brain
CACH ⫽ Childhood Ataxia with Central Nervous System
Hypomyelination; MLSC ⫽ Megalencephalic Leukodystrophy injury, as do signs of a basilar skull fracture, such
with Subcortical Cysts. as bleeding or cerebrospinal fluid (CSF) rhinorrhea
or otorrhea. Bruises or burns of varying ages or
of characteristic shapes suggest nonaccidental
mia, or acute intoxication. A more slowly progres- trauma. Rashes may be characteristic of particular
sive deterioration in consciousness has a large infections, such as those due to meningococcal or
differential diagnosis, which the history, physical rickettsial organisms. Pigmentation abnormalities,
exam, vital signs, and investigative tests may nar- such as the macular lesions of tuberous sclerosis
row. A history of infection raises the possibility of and neurofibromatosis, or generalized hyperpig-
108 MICHELSON AND ASHWAL

mentation, as seen with adrenocortical insuffi- upper and lower eyelids, and the sweat glands of
ciency, can also be helpful clues. the face. Lesions anywhere along this long path-
way can result in ipsilateral pupillary constriction
Neurologic Examination (miosis), partial ptosis, and anhydrosis (Horner’s
The neurologic exam is crucial for determining syndrome).
etiology. Funduscopic examination may reveal Unequal pupil size (anisocoria) may be due to
papilledema in the setting of subacute to chronic direct injury to one iris or to structural disruption
elevations in ICP or retinal hemorrhages in the of unilateral sympathetic or parasympathetic inner-
setting of trauma. Evidence of meningeal irritation, vation. Symmetrically small but minimally reac-
including passive resistance to neck flexion, knee tive pupils are seen with metabolic disorders, drug
flexion when the hip is flexed (Kernig’s sign), or intoxication, and loss of sympathetic stimulation
hip and knee flexion when the neck is flexed resulting from structural lesions of the hypothala-
(Brudzinski’s sign), can be seen with meningitis, mus, pons, medulla, or cervical spinal cord. Le-
subarachnoid hemorrhage, tonsillar herniation, or a sions involving the midbrain are more likely to
parameningeal focus of inflammation. cause midposition and unresponsive pupils due to
Without patient cooperation, as long as medica- interruption of both sympathetic and parasympa-
tions have not induced deep sedation or paralysis, thetic innervation. Symmetrically dilated and un-
important information can be quickly obtained responsive pupils are seen with severe hypoxic-
about the function of brainstem and spinal cord ischemic injuries.
motor and sensory pathways. The pupillary reflex Extraocular movements derive from the oculo-
depends on the passage of light through eye struc- motor (cranial nerve [CN] III) and trochlear (CN
tures, activation of the photoreceptors and retinal IV) nuclei in the midbrain and the abducens (CN
nerve fibers, and transmission along the optic VI) nuclei in the pons. Conjugate horizontal eye
nerve, which hemidecussates at the optic chiasm, movements are made possible by the connections
to bilateral nuclei in pretectal areas of the rostral between the ipsilateral abducens (lateral rectus
midbrain. Signals are then distributed to the bilat- muscle) and contralateral oculomotor (medial rec-
eral parasympathetic Edinger-Westphal nuclei, tus muscle) nuclei along the medial longitudinal
which effect pupillary constriction, and to the bi- fasciculus. Rapid eye movements (saccades) initi-
lateral sympathetic nuclei in the hypothalamus, ate from the frontal eye fields, whereas the smooth
which effect pupillary dilation. Decreased visual pursuit of a moving stimulus arises from the oc-
input from one eye, due to lesions within the eye, cipitoparietal eye fields. Eye movements are mod-
retina, or optic nerve, result in an afferent pupillary ulated by proprioceptive (spinocerebellar tract) and
defect in which the ipsilateral eye has little to no vestibular (CN VIII) inputs from the medulla and
direct response to light but continues to have a coordinated by the cerebellum. The normal ability
normal consensual response to light. Disruption of of these influences to maintain visual fixation de-
a parasympathetic nucleus or of its fibers running spite movement of the head is the source of the
along the surface of the ipsilateral third cranial oculocephalic, or “doll’s eyes,” reflex.
nerve results in ipsilateral mydriasis (pupillary di- An oculocephalic reflex is abnormal in a coma-
lation) and loss of both the direct and consensual tose patient when one or both eyes do not move in
responses to light, but does not affect the direct or the opposite direction of head rotation or vertical
consensual responses of the contralateral eye. movement. Because neck movement is contraindi-
Sympathetic efferents to the pupil originate in cated with cervical spine injuries, the caloric re-
the hypothalamus, descend uncrossed through the sponse (oculovestibular reflex) is used to test the
brainstem and cervical spinal cord to synapse in same pathways. Irrigation of one external ear canal
the nuclei of the intermediolateral columns, exit with warm or cold water induces convection cur-
the spinal cord along with the ventral rami to rents in the endolymph of the semicircular canals.
synapse again in the paraspinal sympathetic chain, Warm water stimulation leads to slow horizontal
and run along the surface of the internal carotid deviation of the eyes away from the irrigated ear,
artery and the ophthalmic division of the trigemi- whereas cold water will effect slow deviation to-
nal nerve to reach the ipsilateral radial pupillodi- ward the irrigated ear. In conscious patients, sac-
lator muscles of the eye, Muller’s muscles of the cades back toward the initial point of fixation
COMA AND BRAIN DEATH 109

Table 3. Glasgow Coma Scale (GCS) Scoring and Modification for Children

Sign GCS GCS Modified for Children Score

Eye opening Spontaneous Spontaneous 4

To command To sound 3
To pain To pain 2
None None 1
Verbal response Oriented Age-appropriate vocalization, smile, 5
or orientation to sound
Confused, disoriented Irritable, consolable, uncooperative, 4
aware of the environment
Inappropriate words Irritable, inconsistently consolable 3
Incomprehensible sounds Inconsolable, unaware of the 2
environment, restless, agitated
None None 1
Motor response Obeys commands Obeys commands, spontaneous 6
movements
Localizes pain Localizes pain 5
Withdraws Withdraws 4
Abnormal flexion to pain Abnormal flexion to pain 3
Abnormal extension to pain Abnormal extension to pain 2
None None 1
Best total score 15

attempt to counter these slow deviations. It is from ing, with flexion of the arms and extension of the
the direction of these rapid eye movements that the legs. Motor responses, along with verbal and ocu-
familiar mnemonic COWS (Cold water Opposite, lar responses, are important cortically determined
Warm water Same) is derived. In the unconscious aspects of consciousness assessed by rating scale
patient, no such saccades are seen. scores, such as the Glasgow Coma Scale (GCS)
Corneal reflexes test the transmission of sensory and its modification for use in children (Table 3).
input along the ophthalmic division of the trigem- These measures of gross neurologic function omit
inal nerve (CN V) to its nucleus in the medulla to many physical findings that are important for the
bilateral facial nerve (CN VII) nuclei in the pons to evaluation of a comatose patient, such as brainstem
effect direct and consensual eyelid closure. An reflexes, but have proven descriptive and prognos-
asymmetric response is suggestive of a structural tic value.
lesion interrupting the reflex arc. Bilateral absence
of the corneal reflex can be seen with metabolic Herniation Syndromes
disorders, but can also occur with sedation and Localized mass lesions or increased ICP can
paralysis. cause several important cerebral herniation clinical
Nasopharyngeal and gag reflexes can be elicited syndromes. Uncal herniation is most often due to a
if there is intact functioning of the glossopharyn- local mass lesion or swelling, which pushes the
geal (CN IX) and vagal (CN X) afferents. Endo- uncus through the tentorial notch and progressively
tracheal tube manipulation is a less-specific test, compresses the ipsilateral CN III, initially affecting
because it stimulates additional spinal cord-medi- the parasympathetic pupillomotor fibers running
ated reflexes. along its surface and resulting in a sluggishly
Asymmetrical movements or postures suggest reactive then fixed and dilated pupil. Further com-
unilateral structural lesions. Flaccid weakness can pression of CN III causes partial ophthalmoplegia
occur with metabolic insults, spinal cord injuries, with the affected eye turned outward and down-
and brainstem lesions below the pons. Lesions of ward due to unopposed functioning of the lateral
the brainstem between the pons and midbrain are rectus and superior oblique muscles. Ipsilateral
thought to be responsible for decorticate posturing, hemiparesis can also occur due to compression of
characterized by extension and internal rotation of the contralateral cerebral peduncle by the tentorial
the upper and lower extremities. Bilateral lesions notch.
above the midbrain may cause decorticate postur- Central herniation is most often due to diffuse
110 MICHELSON AND ASHWAL

brain edema or obstructive hydrocephalus, which (commonly barbiturates, benzodiazepines, opioids,

causes increased ICP, with movement of the thal- amphetamines, cocaine metabolites, phencyclid-
amus and hypothalamus downward through the ine, and marijuana metabolites) and specific serum
tentorial notch and of the cerebellar tonsils through tests for acetaminophen, salicylate, ethanol, sym-
the foramen magnum. Compression and ischemia pathomimetic amines, or tricyclic antidepressants.
of the diencephalon and brainstem can occur in a Drug screens broad enough to detect less common
slow rostrocaudal progression or can advance causes of intoxication, such as thin-layer chroma-
abruptly to give signs of lower brainstem injury. tography and ultraviolet spectroscopy, are not rap-
With a diencephalic-level injury, comatose patients idly available or reliable (often requiring a repeat
have decorticate posturing, Cheyne-Stokes respira- confirmatory test).24
tions, and small pupils but have preserved brain- Neuroimaging by computed tomography (CT),
stem reflexes and remain able to localize noxious which by its rapid acquisition is most suited to the
stimuli. With involvement of the midbrain and critically ill patient, is sufficiently sensitive for
upper pons, posturing becomes decerebrate, pupils detecting structural pathology that needs immedi-
become midposition in size and sometimes irreg- ate intervention, such as cerebral edema, tumor,
ular in shape, and there is a loss of pupillary, hemorrhage, and herniation. Magnetic resonance
oculocephalic, and oculovestibular reflexes. Respi- imaging (MRI) should be obtained when the child
rations may become regular and rapid. With in- can safely withstand the longer acquisition times
volvement of the lower pons and medulla, all and provides greater structural detail as well as
brainstem reflexes are lost, posturing is replaced better early evidence of infection, infarction, dif-
with a flaccid paralysis, and respirations become fuse axonal injury, and demyelination.25,26 Patients
ataxic, irregular, and slow before eventually ceas- without a surgically remediable lesion on initial,
ing. 24-, and 48-hour CT scans do not appear to benefit
from continued serial scanning on a routine ba-
DIAGNOSTIC TESTING OF COMA sis.27
Routine laboratory testing of a patient with An electroencephalogram (EEG) is critical to
coma of unknown etiology should include an im- the evaluation of clinical seizures and clinically
mediate bedside test for serum glucose level, a nonapparent seizures (nonconvulsive status epilep-
complete blood count, and a complete metabolic ticus [NCSE]), especially in paralyzed or heavily
profile, including serum calcium, ammonia, and sedated patients.28 Nonepileptiform EEG features,
lactate levels. The blood count may show evidence such as the degree and distribution of slowing, may
of infection, anemia, or leukemia. The complete suggest a diagnosis or contribute prognostic infor-
metabolic profile may disclose the presence of an mation.29,30 Periodic epileptiform discharges
electrolyte imbalance, a disorder of osmolality, (PEDS) have been associated with NCSE.31 In
renal or hepatic dysfunction, or an inborn error of patients without epilepsy, PEDS are suggestive of
metabolism. Additional tests for inborn errors of underlying brain injury and, when lateralized
metabolism can be performed, such as those for (PLEDS), are suggestive of herpes encephalitis.
serum amino acids, urine organic acids, carnitine Multifocal or generalized periodic discharges can
and acylcarnitines, thyroid function, cortisol, and also be seen with metabolic and infectious causes
urine porphyrins. Respiratory management will of- of widespread cerebral dysfunction and are char-
ten require an arterial blood gas determination. acteristic of subacute sclerosing panencephalitis
Any suspicion for infection should lead to the (SSPE).32
collection of blood, urine, and CSF cultures. Most
children will undergo neuroimaging before a lum- MANAGEMENT OF COMA
bar puncture is performed. The primary brain insult, regardless of etiology,
Toxic ingestions are best identified by history may cause irreversible cellular energy failure, lead-
and the recognition of specific clinical toxidromes, ing to early and delayed cell death along necrotic
but children are more likely to present with unfa- and apoptotic pathways. The mechanisms for in-
miliar or atypical clinical features.22,23 Some lab- jury include excessive release of excitatory neuro-
oratory tests can provide rapid results, such as transmitters, loss of ion homeostasis, elevated in-
standard serum or urine screens for drugs of abuse tracellular calcium, and mitochondrial and
COMA AND BRAIN DEATH 111

inflammatory free-radical production, DNA cleav- Table 4. Steps in the Management of Coma

age, proteolysis, and lipid peroxidation.33 Labora- 1. Assure a stable airway, adequate oxygenation and
tory research into experimental therapies to ad- ventilation, and circulation
2. Recognize and treat hypoglycemia
dress these specific mechanisms has been
3. Consider specific antidotes for medication overdoses
promising, but has not yet led to effective clinical 4. Monitor for, prevent, and treat increased intracranial
applications. The goal of stabilization and treat- pressure
ment of coma, apart from reversing its underlying 5. Evaluate for and treat seizures
cause, if possible, is to prevent secondary brain 6. Treat suspected infection
7. Correct electrolyte and acid-base imbalances
injury from systemic hypotension, hypoxia, hypo-
8. Normalize body temperature
glycemia, seizures, infection, or cerebral herniation
due to an expanding intracranial hemorrhage or
worsening cerebral edema.34 Unfortunately, de-
spite improvements in our ability to recognize and tained in adult stroke patients.39 Similar recom-
treat these secondary factors, there has not yet been mendations have not yet been made in children.
a parallel improvement in outcomes.35 In coma resulting from the effects of opiates,
As with any neurologic emergency, coma man- anticholinergics, or benzodiazepines, reversal can
agement begins with stabilization of the airway, be attempted (with appropriate caution) by using
breathing, and circulation. Patients with altered the respective specific antagonists naloxone, phy-
consciousness nearly always require airway man- sostigmine, and flumazenil. Flumazenil should be
agement to prevent aspiration. Unless the coma is used with caution to avoid precipitating with-
drawal seizures.
clearly atraumatic, cervical spine precautions, with
All children with either traumatic or nontrau-
care in moving the neck (use of a rigid collar), is
matic etiologies should be considered at risk for
required. Intubation must be done using a jaw-
herniation from increased ICP, possibly due to a
lifting technique, which avoids neck hyperexten-
rapidly expanding mass lesion such as hemorrhage
sion. Respiratory failure may be readily apparent
or tumor, or from diffuse expansion of brain con-
or may require monitoring of peripheral oxygen
tents, such as with global cerebral edema or hy-
saturation or measurement of blood gases. The
drocephalus. In such cases, an urgent head CT is
circulatory system is maintained with the early
performed and neurosurgical evaluation requested.
placement of large-bore intravenous catheters or, if
Surgical and nonsurgical management of increased
lack of venous access might delay therapy, an ICP is addressed elsewhere in this issue, but tradi-
intraosseous needle. Maintenance of a cerebral tional measures include decompressive craniot-
perfusion pressure (CPP) at or above 50 mm Hg is omy, ventriculostomy, midline head positioning at
strongly correlated with survival after traumatic 30 degrees of elevation, hyperosmolar infusions,
brain injury.36 Individualized use of fluids and corticosteroids, hyperventilation, and sedation to
medications to maintain an adequate mean arterial avoid agitation. In severe hypoxic-ischemic injury,
pressure (MAP) and CPP leads to better outcome elevated ICP has been correlated with poor out-
than routine dehydration in patients at risk for come; however, interventions that lower ICP have
cerebral edema.37 decreased mortality but have led to increased sur-
Additional steps in the management of coma are vival of patients that remain in a vegetative state.40
outlined in Table 4. A serum glucose level should There are concerns about the adverse effects of
be measured at the bedside to immediately recog- prolonged hyperventilation, with arteriolar con-
nize hypoglycemia and treat it as soon as possible. striction and decreased cerebral perfusion.41 The
Glucose-containing fluids should be used cau- use of osmolar diuretics may lead to dehydration,
tiously, if at all, outside of the setting of hypogly- with decreased cerebral perfusion, and disruption
cemia, due to the risk of iatrogenic hyperglycemia, of the blood-brain barrier, with rebound cerebral
hyperosmolarity, and cerebral dehydration.38 The edema. Hypertonic saline may be a more effective
association between hyperglycemia and increased and safer method of reducing ICP.42
morbidity and mortality in patients and experimen- Status epilepticus necessitates urgent manage-
tal animals with ischemic brain injury has led to ment, but it bears emphasizing that clinically subtle
the recommendation that normoglycemia be main- or nonconvulsive seizures should be considered
112 MICHELSON AND ASHWAL

Table 5. Glasgow Outcome Scale Nutritional support tailored for the increased
1. Death energy and nitrogen demands of severe injury may
2. Persistent vegetative state minimize infectious complications and accelerate
3. Severe disability: conscious but disabled neurologic recovery.46 The incidence of deep ve-
4. Moderate disability: disabled but independent
nous thrombosis and pulmonary embolism can be
5. Good recovery
decreased with the use of low molecular weight
heparin (in patients not at an increased risk for
intracranial hemorrhage), sequential compression
stockings (in head trauma patients without associ-
when no other etiology for coma is apparent. An
ated leg injuries), or foot pumps.47 For children
urgent EEG should be requested in all such cases.
with severe traumatic brain injury, transfer to hos-
Infections are treated with empiric broad-spec-
pitals with pediatric trauma centers should be con-
trum antibiotics when suspected, and it may be
sidered, because these facilities have demonstrated
impossible to rule out an intracranial infection
improved outcomes.48
without a lumbar puncture. If an intracranial mass
lesion is present or suspected, and the risk of OUTCOME OF COMA
herniation outweighs the benefits of an immediate
The prognosis for neurologic recovery depends
bacteriologic diagnosis, then lumbar puncture may
greatly on the etiology and severity of the primary
be deferred. The fever seen in the first 48 hours
brain injury and on the success of treatments to
after submersion in up to 50% of near-drowning
minimize secondary brain injury. There is an high
victims may not be infectious and may resolve
overall mortality rate, with traumatic brain injury
spontaneously in 80% of victims. The use of em- consistently one of the most common causes of
piric antibiotics should not depend on a single childhood mortality and neurologic morbidity.49
laboratory test result, but rather should be guided Outcomes from coma are commonly reported us-
by the overall clinical presentation. ing one of several scales. The five-point Glasgow
Electrolyte and acid-base imbalances should be Outcome Scale (GOS) emphasizes functional in-
corrected, whether they are the direct cause or are dependence, as outlined in Table 5.50 The six-point
secondary. The syndrome of inappropriate antidi- Pediatric Cerebral Performance Category Scale51
uretic hormone frequently causes hyponatremia is more easily applied to younger children because
and decreased urine output, while central diabetes it evaluates function with respect to its age-appro-
insipidus may lead to hypernatremia and increased priateness, as outlined in Table 6. These scales are
urine output. The syndrome of cerebral salt wast- easily applied and have good interrater reliabilities,
ing can also cause hyponatremia and dehydration. but place little emphasis on the emotional and
These conditions can be differentiated by measure- social impairments that are common among survi-
ment of urine volume, sodium concentration, and vors of brain injury and greatly influence their
osmolarity.43 Wide swings in electrolyte levels and quality of life. There are substantially more com-
serum osmolarity put patients at risk for cerebral prehensive scales in use, such as the Functional
edema, cardiac dysfunction, and osmotic demyeli- Independence Measure (FIM) and the pediatric
nation. WeeFIM, which assess 18 items in six categories,
Hyperthermia and hypothermia should be cor-
rected. Although some children rescued from sub-
mersion in icy water survive with surprisingly Table 6. Pediatric Cerebral Performance Category Scale
good neurologic outcomes and hypothermia is rou- 1. Normal: able to perform all age-appropriate activities
tinely used for neuroprotection during cardiac by- 2. Mild disability: conscious, alert, and able to interact at an
pass surgery, clinical trials of therapeutic hypother- age-appropriate level but may have a mild neurologic
deficit
mia in traumatic or hypoxic brain injury have had
3. Moderate disability: conscious, sufficient cerebral
mixed results to date.44 However, it has been function for most age-appropriate independent activities
shown that fever is detrimental to the injured brain 4. Severe disability: conscious, dependent on others for
and that cooling measures should be used as daily support because of impaired brain function
needed to bring core temperatures into the normal 5. Persistent vegetative state or coma
6. Death
range.45
COMA AND BRAIN DEATH 113

including self-care, sphincter control, mobility, lo- was worse for patients who were younger, had a
comotion, communication, and social/cognition.52,53 lower GCS score on presentation, or had absent
Detailed neuropsychologic testing remains the brainstem reflexes, worse motor responses, hypo-
most complete measure of neurologic functioning thermia, or hypotension.59
for those able to cooperate.54 Intact somatosensory evoked potentials (SSEPs)
Children with severe traumatic brain injury usu- within 48 hours of resuscitation have ⬎90% pre-
ally reach their highest GOS level by the sixth dictive value for good neurologic outcome,
month of recovery, although further improvements whereas bilaterally absent SSEPs make a poor
within this gross outcome measure can continue neurologic outcome (severe disability or worse)
for months and years afterward. The vast majority likely.60 Children and adolescents are more likely
of children who survive traumatic coma (up to than adults to awaken from coma despite absent
90% of those with traumatic coma lasting less than SSEPs. Unlike adults, children who do awaken
6 weeks) can be expected to eventually have suf- usually recover some degree of functional indepen-
ficient independent mobility to rate as moderately dence, making the test even less attractive for use
disabled or better on the GOS, although significant in making life-support decisions.61 Other forms of
cognitive and behavioral problems may persist. electrophysiologic testing, including EEG, brain-
The emphasis on motor skills of many outcome stem auditory evoked responses (BAERs), and
measures has contributed to the sense that young visual evoked potentials (VEPs), have a similar
children do better than adults after brain injury, but accuracy in outcome prediction.30 Burst-suppres-
persistent deficits in other areas may have a more sion and isoelectric EEG patterns indicate a very
profound influence on future learning ability and poor prognosis and were seen only in children with
later cognitive outcomes of younger subjects.55 In fatal outcomes in one series of near-drownings.62
one large prospective, population-based study of For children who are in a coma after resuscita-
nontraumatic coma, 46% of the 278 children died, tion from cardiopulmonary arrest, especially those
but 64% of the survivors were considered neuro- with an out-of-hospital arrest who remain pulseless
logically intact, 16% had mild to moderate disabil- or asystolic on arrival in the emergency room, the
ity, and only 14% were severely disabled or in a combined mortality and severe morbidity rates
vegetative state.56 approach 100%. Children with the worst progno-
sis—those who are found pulseless after an unwit-
Outcome Prediction nessed arrest, require more than 25 minutes of
Our ability to predict outcomes for comatose cardiac resuscitation, or suffer warm-water near-
patients is far from complete, but does provide drowning and do not have a perfusing cardiac
some basis for the expectations and treatment de- rhythm by the time they arrive in the emergency
cisions of family members and caregivers. Predic- room— have such a low likelihood of meaningful
tive accuracy using various clinical, serologic, ra- recovery that reasonable arguments can be made
diologic, and electrophysiologic measures ranges against providing them with continued aggressive
from 70% to 90% in various studies, but early care.63-65 Intact recovery rates of 40% to 75% have
assessment methods have yet to exclude all false- been reported with witnessed arrest, isolated respi-
positives— children who are predicted to have se- ratory arrest, and ice-water submersion. For sub-
vere disability but nevertheless make a good re- mersion victims, factors found to be associated
covery. Children with severe traumatic brain injury with poor recovery after arrival to the emergency
who present with a GCS of 4 to 8 are very unlikely room in various studies include loss of pupillary
to die, whereas those with an initial score of 3 (no reflexes, a GCS below 5, male gender, and an
response) have a mortality rate of 50% to 60%, initial blood glucose level ⬎250 mg/dL.66
often dying within a few days of hospitalization.57 Among widely available imaging studies, proton
Other than GCS score, outcome from severe trau- magnetic resonance spectroscopy has shown prom-
matic brain injury has also shown independent ise in predicting poor outcome in those near-
correlation with multiple trauma, hypoxia, hypo- drowning victims with lower metabolite ratios and
tension, disseminated intravascular coagulation, higher serum lactate levels.67,68 Even conventional
hyperglycemia, and early posttraumatic seizures.58 MRI has a significant advantage over CT scanning
Among patients with nontraumatic coma, outcome in the prognosis of patients with traumatic brain
114 MICHELSON AND ASHWAL

injury; one series found that trauma involving the Table 7. Guidelines for the Determination
of Brain Death in Children
brainstem, seen in 60% of cases, was 100% pre-
dictive of mortality.69 After hypoxic-ischemic in- History:
jury, scoring systems rating the extent of injury on 1. A cause of coma sufficient to cause irreversible loss of all
brain function
MRI, especially in watershed and basal ganglia
2. No surgically remediable intracranial condition
areas, demonstrate a strong correlation with neu- 3. Absence of confounding sedatives or paralytics
rologic outcome.70 Other, less accessible tests have Physical exam:
also proven to be highly predictive of poor neuro- 1. Coma and apnea requiring mechanical ventilation
logic outcome after brain injury, as is the case with 2. Absent spontaneous movements, postures, seizures, or
responses (other than spinal reflexes)
elevated serum levels of neuron-specific enolase
3. Absent brainstem function
after cardiac arrest.71 However, no combination of a. Mid-position or fully dilated pupils without light reflexes
early clinical variables or ancillary tests has yet b. Absent oculocephalic, oculovestibular (caloric), corneal,
been shown to have absolute predictive power. It nasopharyngeal, gag, cough, sucking, and rooting
thus has been recommended that even victims of reflexes
c. Absent facial movement with central noxious
warm water near-drowning who appear to have
stimulation
little chance of recovery be given at least 48 hours d. Absent respiratory effort with standardized apnea
of intensive support.72 testing
4. Exclusion of hypothermia (below 32°C) or uncontrolled
BRAIN DEATH hypotension
5. Exclusion of severe metabolic or endocrine disturbance
The concept of brain death and guidelines for its
Diagnosis is independently confirmed after a period of
clinical determination were originally proposed by observation determined by age:
an ad hoc committee of the Harvard Medical 1. Under 2 months: 48 hours with EEG confirmation at each
School faculty in 1968.73 The need for an alterna- exam
tive definition of death arose from the technologi- 2. 2 months to 1 year: 24 hours with EEG confirmation (or
less time with both EEG and CBF confirmation)
cal advances that allowed preservation of the car-
3. Over 1 year: 12-24 hours with confirmatory tests optional
diorespiratory functions that had previously been
used to define life and death. Two important mo-
tivations were the desire to avoid prolonged suf- common mode of death in one study.78 Although
fering for families of patients in a coma who had diagnosed less often in neonates, brain death rep-
no hope for recovery, and a desire to promote resented 6% of neonatal intensive care unit deaths
organ transplantation.74 Techniques at our ready in one study.79 Most children diagnosed as brain
disposal may be unable to prove the absolute dead are either removed from life support or have
cessation of all cellular activity within the brain but their organs removed for donation within a short
are accurate in diagnosing the loss of function. No time. When supportive care is continued, the vast
patient with a diagnosis of brain death, when di- majority suffer cardiac arrest within 2 to 3 weeks;
agnosed with care, has been known to have recov- few cases of prolonged maintenance have been
ered any neurologic function.75 described.80
Although some aspects of the diagnosis of brain
death have evolved over time and continue to show Clinical Guidelines
regional differences, the general concept is legally Guidelines for the diagnosis of brain death in
and socially established in most developed coun- older children and adults were first established by
tries. The emphasis in all cases is on clinical testing a Presidential commission in 1981.81 Subsequent
for profound coma, loss of brainstem reflexes, and discussion by a combined neurologic and pediatric
apnea, but there is significant variability in the task force82 and studies in younger children, in-
requirements for confirmatory observation periods fants, and newborns83 helped provide support for
and ancillary tests.76 Even in countries with well- recommendations for the diagnosis and confirma-
established and publicized guidelines, there may be tion of brain death in children down to age 34
wide variability in how closely they are followed weeks from conception (Table 7). Due to the im-
or documented.77 maturity of brainstem reflexes, the clinical exami-
Brain death occurs in up to 2% of all admissions nation for brain death is unreliable in younger
to pediatric intensive care units and was the most infants.84
COMA AND BRAIN DEATH 115

Apnea testing should be the last element of a meet criteria for brain death, especially young
brain death examination, preferably following a infants, may exhibit minimal amounts of electrical
period of ventilation that establishes normocap- brain activity at a first or repeat EEG examination.
nea.85 During the test, mechanical ventilation is Infants are also more likely to have preservation of
interrupted and the child is observed for respiratory CBF despite findings consistent with brain death,
effort until hypercarbia of at least 60 mmHg is perhaps because their open fontanels and cranial
achieved, as usually occurs within 5 to 10 minutes. sutures prevent ICP increases from exceeding
Rare examples of children showing respiratory mean arterial pressures. In one study of clinically
effort after reaching a PaCO2 of ⱖ90 mmHg have brain dead neonates, only 13 of 18 demonstrated
led some institutions to use prolonged apnea test- absence of CBF.89 In situations in which a patient
ing.86 Pretreatment with 100% oxygen and contin- is clinically brain dead but has a negative confir-
ued delivery of oxygen via a catheter inserted into matory test, a determination of brain death can still
the endotracheal tube is recommended to avoid be made after a more extended period of clinical
potential hypoxic injury to the myocardium and observation.
cardiac arrhythmia. If parents refuse consent
for this form of apnea testing, CO2 gas augmenta- Withdrawal of Care and Organ Donation
tion and brief ventilator interruption can be substi-
tuted.87 When a diagnosis of brain death is made, the
physician must make the difficult transition from
Confirmatory Testing dogged pursuit of patient survival to withdrawal of
There are two established tests for confirming a supportive care. Parents may be unprepared for the
suspected diagnosis of brain death: EEG and cere- diagnosis but should be informed of it as soon as it
bral blood flow (CBF) measurement. These tests is suspected. Physicians struggling with this task
are recommended for children under age 1 year often mistakenly use language that adds ambiguity
and for any patient in whom a complete clinical to the patient’s status. It is essential to convey that
examination cannot be performed. An isoelectric the diagnosis is equivalent to death and that further
EEG must be recorded for at least 30 minutes with, treatment cannot avoid death and will only delay
according to certain parameters, a sensitivity ⬎2 the mourning process for the family. Parents may
mV and filter settings at 70 Hz. Despite the wide benefit from being present during the performance
availability of EEG and its bedside portability, its of a confirmatory test or clinical examination for
interpretation is frequently hampered by the tech- brainstem reflexes, but deep tendon reflexes and
nical problems encountered in the typical critical other spinal cord-mediated movements must be
care environment.88 described and differentiated from higher brain
CBF can be evaluated with invasive four-vessel functions so that they are not interpreted as signs of
contrast angiography, scintigraphy using a radio- neurologic recovery. Physicians should be pre-
nuclide such as technetium-99m HmPAO, xenon pared to remain supportive of parents despite par-
CT, or transcranial Doppler ultrasonography. Scin- ents’ quite natural expressions of blame, hostility,
tigraphy is favored due to its noninvasiveness, and anger toward them and the hospital staff.
portability, and lack of dependence on operator Once the parents have been given time to un-
skill. The absence of discernable cerebral perfusion derstand the diagnosis of brain death, their prefer-
is considered indicative of severely decreased car- ences regarding organ donation should be ad-
diac output or increased ICP. CBF studies are dressed. The care and sensitivity with which this
generally not confounded by hypothermia or by question is raised may have a direct influence on
endogenous or exogenous intoxications, unlike the decision made. Donor and nondonor families in
EEG studies. However, as absent perfusion reflects one survey had significantly different perceptions
severely increased ICP, perfusion tests may be less of their physicians as either supportive or insensi-
helpful in patients with a ventriculostomy or crani- tive.90 Organ donation is also strongly influenced
ectomy. by the family’s understanding of the meaning and
Electrocerebral inactivity may be recorded when prognosis of brain death. Several studies have
there is persistent evidence of brain function on found that donor families are much more likely
examination. Conversely, patients who otherwise than nondonor families to agree that people cannot
116 MICHELSON AND ASHWAL

recover from brain death.91,92 Better donation rates tor. Maintenance of organs before transplantation
have been associated with the use of a team ap- requires continuation of supportive care, including
proach, quiet surroundings, and a decoupling of the mechanical ventilation and stabilization of temper-
discussion about donation from the announcement ature and blood pressure. Hemodynamic instability
of brain death. often worsens around the time of brain death, with
There are very few remaining medical condi- a loss of pituitary function and development of
tions (such as malignant neoplasia) that are still central diabetes insipidus. Desmopressin replace-
considered absolute contraindications for organ ment may be necessary for or may greatly facilitate
donation.18 Every potential organ donor should be blood pressure support,93 along with routine fluid
discussed with the local organ transplant coordina- and electrolyte compensation for urinary losses.

REFERENCES
1. Zeman A: Consciousness. Brain 124:1263-1289, 2001 18. Lutz-Dettinger N, de Jaeger A, Kerremans I: Care of the
2. Plum F, Posner JB: The Diagnosis of Stupor and Coma potential organ donor. Pediatr Clin North Am 48:715-749, 2001
(ed 3 rev). Philadelphia, PA, FA Davis, 1982 19. Wong CP, Forsyth RJ, Kelly TP, et al: Incidence, aeti-
3. Zeman A: Persistent vegetative state. Lancet 350:795- ology, and outcome of non-traumatic coma: A population based
799, 1997 study. Arch Dis Child 84:193-199, 2001
4. The Multi-Society Task Force on PVS: Medical aspects of 20. Bates D: The management of medical coma. J Neurol
the persistent vegetative state. N Engl J Med 330:1499-1508, Neurosurg Psychiatr 56:589-598, 1993
1994 21. Carty H, Pierce A: Non-accidental injury: A retrospec-
5. Ashwal S: Medical aspects of the minimally conscious tive analysis of a large cohort. Eur Radiol 12:2919-2925, 2002
state in children. Brain Dev 25:535-545, 2003 22. Abbruzzi G, Stork CM: Pediatric toxicologic concerns.
6. Andrews K, Murphy L, Munday R, et al: Misdiagnosis of Emerg Med Clin North Am 20:223-247, 2002
the vegetative state: Retrospective study in a rehabilitation unit. 23. Lifshitz M, Shahak E, Sofer S: Carbamate and organo-
BMJ 313:13-16, 1996 phosphate poisoning in young children. Pediatr Emerg Care
7. Landrieu P, Fromentin C, Tardieu M, et al: Locked-in 15:102-103, 1999
syndrome with a favourable outcome. Eur J Pediatr 142:144- 24. Hammett-Stabler CA, Pesce AJ, Cannon DJ: Urine drug
145, 1984 screening in the medical setting. Clin Chim Acta 315:125-135,
8. Masuzawa H, Sato J, Kamitani H, et al: Pontine gliomas 2002
causing locked-in syndrome. Childs Nerv Syst 9:256-259, 1993 25. Tong KA, Ashwal S, Holshouser BA, et al: Hemorrhagic
9. Rosman NP, Adhami S, Mannheim GB, et al: Basilar shearing lesions in children and adolescents with posttraumatic
artery occlusion in children: Misleading presentations, “locked- diffuse axonal injury: Improved detection and initial results.
in” state, and diagnostic importance of accompanying vertebral Radiology 227:332-339, 2003
artery occlusion. J Child Neurol 18:450-462, 2003 26. Sundgren PC, Reinstrup P, Romner B, et al: Value of
10. Bakshi N, Maselli RA, Gospe SM Jr, et al: Fulminant conventional and diffusion- and perfusion-weighted MRI in the
demyelinating neuropathy mimicking cerebral death. Muscle management of patients with unclear cerebral pathology, ad-
Nerve 20:1595-1597, 1997 mitted to the intensive care unit. Neuroradiology 44:674-680,
11. Echenberg RJ: Permanently locked-in syndrome in the 2002
neurologically impaired neonate: Report of a case of Werdnig- 27. Figg RE, Burry TS, Vander Kolk WE: Clinical efficacy
Hoffmann disease. J Clin Ethics 3:206-208, 1992 of serial computed tomographic scanning in severe closed-head
12. Golden GS, Leeds N, Kremenitzer MW, et al: The injury patients. J Trauma 55:1061-1064, 2003
“locked-in” syndrome in children. J Pediatr 89:596-598, 1976 28. Markand ON: Pearls, perils, and pitfalls in the use of the
13. van Mourik M, van Dongen HR, Catsman-Berrevoets electroencephalogram. Semin Neurol 23:7-46, 2003
CE: The many faces of acquired neurologic mutism in child- 29. Young GB: The EEG in coma. J Clin Neurophysiol
hood. Pediatr Neurol 15:352-357, 1996 17:473-485, 2000
14. Abekura M: Akinetic mutism and magnetic resonance 30. Mewasingh LD, Christophe C, Fonteyne C, et al: Pre-
imaging in obstructive hydrocephalus. Case illustration. J Neu- dictive value of electrophysiology in children with hypoxic
rosurg 88:161, 1998 coma. Pediatr Neurol 28:178-183, 2003
15. Mellon AF, Appleton RE, Gardner-Medwin D, et al: 31. Brenner RP: Is it status? Epilepsia 43:S103-113, 2002
Encephalitis lethargica-like illness in a five-year-old. Dev Med 32. Yemisci M, Gurer G, Saygi S, et al: Generalised periodic
Child Neurol 33:158-161, 1991 epileptiform discharges: Clinical features, neuroradiological
16. Shinoda M, Tsugu A, Oda S, et al: Development of evaluation and prognosis in 37 adult patients. Seizure 12:465-
akinetic mutism and hyperphagia after left thalamic and right 472, 2003
hypothalamic lesions. Childs Nerv Syst 9:243-245, 1993 33. Bramlett HM, Dietrich WD: Pathophysiology of cerebral
17. Kadota Y, Kondo T, Sato K: Akinetic mutism and ischemia and brain trauma: Similarities and differences. J Cereb
involuntary movements following radical resection of hypotha- Blood Flow Metab 24:133-150, 2004
lamic glioma: Case report. Neurol Med Chir (Tokyo) 36:447- 34. Kochanek PM, Clark RS, Ruppel RA, et al: Biochemi-
450, 1996 cal, cellular, and molecular mechanisms in the evolution of
COMA AND BRAIN DEATH 117

secondary damage after severe traumatic brain injury in infants Measure. Buffalo, NY, State University of New York at Buf-
and children: Lessons learned from the bedside. Pediatr Crit falo, 1990
Care Med 1:4-19, 2000 53. Data Management Service of the Uniform Data System
35. Pfenninger J, Santi A: Severe traumatic brain injury in of Medical Rehabilitation and the Center for Functional Assess-
children: Are the results improving? Swiss Med Wkly 132:116- ment Research: Guide for Use of the Functional Independence
120, 2002 Measure for Children (WeeFIM) of the Uniform Data Set for
36. Hackbarth RM, Rzeszutko KM, Sturm G, et al: Survival Medical Rehabilitation. Buffalo, NY, State University of New
and functional outcome in pediatric traumatic brain injury: A York at Buffalo, 1989
retrospective review and analysis of predictive factors. Crit 54. Hannay HJ, Sherer M: Assessment of outcome from
Care Med 30:1630-1635, 2002 head injury, in Narayan RK, Wilberger JEJr, Povlishock JT
37. Yu PL, Jin LM, Seaman H, et al: Fluid therapy of acute (eds): Neurotrauma. New York, McGraw-Hill, 1996
brain edema in children. Pediatr Neurol 22:298-301, 2000 55. Fletcher JM, Ewing-Cobbs L, Francis DJ, et al: Variabil-
38. Jackson J, Bolte RG: Risks of intravenous administration ity in outcomes after traumatic brain injury in children, in
of hypotonic fluids for pediatric patients in ED and prehospital Broman SH, Michel ME (eds): Traumatic Head Injury in
settings: Let’s remove the handle from the pump. Am J Emerg Children: a developmental perspective. New York, Oxford
Med 18:269-270, 2000 University Press, 1995, pp 22-39
39. Wass CT, Lanier WL: Glucose modulation of ischemic 56. Forsythe RJ, Wong CP, Kelly TP, et al: Cognitive and
brain injury: Review and clinical recommendations. Mayo Clin adaptive outcomes and age at insult effects after non-traumatic
Proc 71:801-812, 1996 coma. Arch Dis Child 84:200-204, 2001
40. Bohn DJ, Biggar WD, Smith CR, et al: Influence of 57. Bruce DA: Outcome, does it work?, in Harris BH (ed):
hypothermia, barbiturate therapy, and intracranial pressure Progress in Pediatric Trauma. Boston, MA, Nob Hill Press,
monitoring on morbidity and mortality after near-drowning. 1985
Crit Care Med 14:529-534, 1986 58. Chiaretti A, Piastra M, Pulitano S, et al: Prognostic
41. Skippen PA, Seear M, Poskitt K, et al: Effect of hyper- factors and outcome of children with severe head injury: An
ventilation on regional cerebral blood flow in head-injured 8-year experience. Childs Nerv Syst 18:129-136, 2002
children. Crit Care Med 25:1402-1409, 1997 59. Johnston B, Seshia SS: Prediction of outcome in non-
42. Georgiadis AL, Suarez JI: Hypertonic saline for cerebral traumatic coma in childhood. Acta Neurol Scand 69:417-427,
edema. Curr Neurol Neurosci Rep 3:524-530, 2003 1984
43. Berkenbosch JW, Lentz CW, Jimenez DF, et al: Cerebral 60. Beca J, Cox PN, Taylor MJ, et al: Somatosensory evoked
salt wasting syndrome following brain injury in three pediatric potentials for prediction of outcome in acute severe brain injury.
patients: Suggestions for rapid diagnosis and therapy. Pediatr J Pediatr 126:44-49, 1995
Neurosurg 36:75-79, 2002 61. Robinson LR, Mickelsen PJ, Tirschwell DL, et al: Pre-
44. Bernard SA, Buist M: Induced hypothermia in critical dictive value of somatosensory evoked potentials for awakening
care medicine: A review. Crit Care Med 31:2041-2051, 2003 from coma. Crit Care Med 31:960-967, 2003
45. Bernardo LM, Henker R, O’Connor J: Treatment of 62. Kruus S, Bergstrom L, Suutarinen T, et al: The prognosis
trauma-associated hypothermia in children: Evidence-based of near-drowned children. Acta Paediatr Scand 68:315-322,
practice. Am J Crit Care 9:227-234, 2000 1979
46. Taylor SJ, Fettes SB, Jewkes C, et al: Prospective, 63. Goyco PG, Beckerman RC: Sudden infant death syn-
randomized, controlled trial to determine the effect of early drome. Curr Probl Pediatr 6:297-346, 1990
enhanced enteral nutrition on clinical outcome in mechanically 64. Quan L, Wentz KR, Gore EJ, et al: Outcome and pre-
ventilated patients suffering head injury. Crit Care Med 27: dictors of outcome in pediatric submersion victims receiving
2525-2531, 1999 prehospital care in King County, Washington. Pediatrics 86:
47. Spain DA, Bergamini TM, Hoffmann JF, et al: Compar- 586-593, 1990
ison of sequential compression devices and foot pumps for 65. Spack L, Gedeit R, Splaingard M, et al: Failure of
prophylaxis of deep venous thrombosis in high-risk trauma aggressive therapy to alter outcome in pediatric near-drowning.
patients. Am Surg 64:522-525, 1998 Pediatr Emerg Care 13:98-102, 1977
48. Potoka DA, Schall LC, Ford HR: Improved functional 66. Ibsen LM, Koch T: Submersion and asphyxial injury.
outcome for severely injured children treated at pediatric Crit Care Med 30:S402-408, 2002
trauma centers. J Trauma 51:824-832, 2001 67. Holshouser BA, Ashwal S, Luh GY, et al: Proton MR
49. Arias E, MacDorman MF, Strobino DM, et al: Annual spectroscopy after acute central nervous system injury: Out-
summary of vital statistics: 2002. Pediatrics 112:1215-1230, come prediction in neonates, infants and children. Radiology
2003 202:487-496, 1997
50. Jennett B, Bond M: Assessment of outcome after severe 68. Dubowitz DJ, Bluml S, Arcinue E, et al: MR of hypoxic
brain damage. Lancet 1:480-484, 1975 encephalopathy in children after near drowning: Correlation
51. Fiser DH: Assessing the outcome of pediatric intensive with quantitative proton MR spectroscopy and clinical out-
care. J Pediatr 121:68-74, 1992 come. Am J Neuroradiol 19:1617-1627, 1998
52. Data Management Service of the Uniform Data System 69. Woischneck D, Klein S, Reissberg S, et al: Prognosis of
of Medical Rehabilitation and the Center for Functional Assess- brain stem lesion in children with head injury. Childs Nerv Syst
ment Research: Guide for Use of the Uniform Data Set for 19:174-178, 2003
Medical Rehabilitation Including the Functional Independence 70. Christophe C, Fonteyne C, Ziereisen F, et al: Value of
118 MICHELSON AND ASHWAL

MR imaging of the brain in children with hypoxic coma. Am J 82. Guidelines for the determination of brain death in chil-
Neuroradiol 23:716-723, 2002 dren. Pediatrics 80:298-300, 1987
71. Fogel W, Krieger D, Veith M, et al: Serum neuron- 83. Ashwal S, Schneider S: Brain death in the newborn.
specific enolase as early predictor of outcome after cardiac Pediatrics 84:429-437, 1989
arrest. Crit Care Med 25:1133-1138, 1997 84. Swaiman KF: Neurological examination of the term and
72. Christensen DW, Jansen P, Perkin RM: Outcome and preterm infant, in Swaiman KF, Ashwal S (eds): Pediatric
acute care hospital costs after warm water near-drowning in Neurology: Principles and Practice. (ed 3). St Louis, MO,
children. Pediatrics 99:715-721, 1997 Mosby, 1999
73. Beecher HK: A definition of irreversible coma. Report of 85. Goudreau JL, Wijdicks EF, Emery SF: Complications
the ad hoc committee of Harvard Medical School to examine during apnea testing in the determination of brain death: Pre-
the definition of brain death. JAMA 205:337-340, 1968 disposing factors. Neurology 55:1045-1048, 2000
74. Wijdicks EF: The neurologist and Harvard criteria for 86. Vardis R, Pollack MM: Increased apnea threshold in a
brain death. Neurology 61:970-976, 2003 pediatric patient with suspected brain death. Crit Care Med
75. Jennett B, Gleave J, Wilson P: Brain death in three
26:1917-1919, 1998
neurosurgical units. BMJ 282:533-539, 1981
87. Lang CJ: Apnea testing by artificial CO2 augmentation.
76. Wijdicks EF: Brain death worldwide: Accepted fact but
Neurology 45:966-969, 1995
no global consensus on diagnostic criteria. Neurology 58:20-25,
88. Schneider S: Usefulness of EEG in the evaluation of
2002
brain death in children: The cons. Electroencephalogr Clin
77. Chang MY, McBride LA, Ferguson MA: Variability in
Neurophysiol 73:276-278, 1989
brain death declaration practices in pediatric head trauma pa-
tients. Pediatr Neurosurg 39:7-9, 2003 89. Ashwal S: Brain death in the newborn: Current perspec-
78. Martinot A, Lejeune C, Hue V, et al: Modality and tives. Clin Perinatol 24:859-882, 1997
causes of 259 deaths in a pediatric intensive care unit. Arch 90. Weiss AH, Fortinsky RH, Laughlin J, et al: Parental
Pediatr 2:735-741, 1995 consent for pediatric cadaveric organ donation. Transplant Proc
79. Parker BL, Frewen TC, Levin SD, et al: Declaring 29:1896-l901, 1997
pediatric brain death: Current practice in a Canadian pediatric 91. Siminoff LA, Mercer MB, Arnold R: Families’ under-
critical care unit. Can Med Assoc J 153:909-916, 1995 standing of brain death. Prog Transplant 13:218-224, 2003
80. Shewmon DA: Chronic “brain death”: Meta-analysis and 92. Franz HG, DeJong W, Wolfe SM, et al: Explaining brain
conceptual consequences. Neurology 51:1538-1545, 1998 death: A critical feature of the donation process. J Transplant
81. President’s Commission for the Study of Ethical Prob- Coord 7:14-21, 1997
lems in Medicine and Biomedical and Behavioral Research: 93. Katz K, Lawler J, Wax J, et al: Vasopressin pressor
Guidelines for the determination of death. JAMA 246:2184- effects in critically ill children during evaluation for brain death
2186, 1981 and organ recovery. Resuscitation 47:33-40, 2000