Determination of Brain Death in Children

Updated: Dec 26, 2018
  • Author: Samuel Koszer, MD; Chief Editor: Stephen L Nelson, Jr, MD, PhD, FAACPDM, FAAN, FAAP, FANA  more...
  • Print


The medical and legal aspects of determining brain death have evolved over the past 3 decades. The need for special criteria for the determination of brain death in infants and children was clear as early as 1967, when the first US committee met to form a consensus opinion. The pediatric CNS may be more resilient to certain forms of injury; this should be considered when interpreting diagnosis and confirming brain death in infants and children.

See also Special Issues in Medicolegal Neurology and Evaluation of Fetal Death.


Definition of Cerebral Death

The definition of cerebral death has evolved out of necessity through consensus opinions and information based on retrospective studies and case reports. An early age-independent definition of death in Black's Law Dictionary (1950) was "the cessation of life; the ceasing to exist; defined by physicians as a total stoppage of the circulation of the blood, and a cessation of the animal and vital functions consequent there upon such as respiration, pulsation, etc." [1]

In the latter half of the 20th century, technological advancements created situations in which cardiopulmonary resuscitation (CPR) and prolonged artificial support maintain vital functions even in the presence of an irreversible CNS insult. The continued use of expensive medical resources on patients without a reasonable prognosis for recovery expends a substantial portion of our limited resources. This has prompted medical and legal professionals to define a state of cerebral or brain death. The identification of brain death without somatic death (ie, death of the entire body) also allows the harvesting of organs for transplantation from patients who have no possibility of recovery. [2]


American Electroencephalographic Society

In 1967, the American Electroencephalographic Society retrospectively surveyed 1665 patients with electrocerebral silence (ECS)—that is, no evidence of brain electrical activity greater than 2 µV between electrode pairs placed at a distance of 10 cm or more—who were in various levels of coma. Only 3 of the 1665 patients recovered cerebral function. Two of the patients who recovered were in barbiturate coma and the third, a child, had experienced a meprobamate overdose. No details were given regarding the ages of the remaining patients; therefore, generalizations concerning age could not be made. The report concluded that ECS together with complete unresponsiveness, apnea, absent cephalic reflexes, and the inability to maintain circulation without artificial means is strongly presumptive of irreversible coma (ie, cerebral death). [3, 4] The term electrocerebral inactivity (ECI) is often used instead of ECS.


Harvard Medical School Ad Hoc Committee

In 1968, a consensus report of the Ad Hoc Committee of the Harvard Medical School was the first effort to define cerebral death as a new criterion for death. No specific recommendations with respect to age were made in these landmark definitions. However, irreversible coma was defined as unresponsiveness to external stimuli, absent movements or breathing, absent reflexes, and a flat EEG. [5]


American Neurological Association

The American Neurological Association (ANA) met in 1975 and reviewed the Harvard criteria. With regard to children, the consensus was that "these criteria may be inapplicable for children under 5 years of age since there are indications that the immature nervous system can survive significant periods of electrocerebral silence."

This conclusion was based on a case report in 1972 by Green and Lauber describing a 5-year-old child who became comatose after severe liver failure. He was unresponsive to any stimulation, had dilated and fixed pupils, and was on a ventilator. A 30-minute EEG revealed no electrical activity attributable to the cerebrum. Twenty-four hours later, the patient's clinical status improved. Spontaneous movements and withdrawal from noxious stimuli were noted. The EEG at this time demonstrated fast activity at 2-4 µV intermittently and independently in the right and left hemispheres and occasionally bilaterally. The patient died 7 hours after the second EEG. No apnea or oculocephalic reflexes were documented in this report; therefore, conclusions concerning the applicability of the Harvard criteria for children younger than 5 years were questionable. [6]


Retrospective Collaborative Study

In 1977, a retrospective Collaborative Study on brain death was performed with strict criteria for entry. The prerequisite for entry was that all appropriate diagnostic and therapeutic procedures were performed to determine the cause of brain death (including a drug toxicology screen). At least 6 hours after ictus, patients had documented lack of response to external stimuli (ie, cerebral unresponsiveness), apnea, dilated pupils, absent cephalic reflexes, and electrocerebral silence (ECS) for at least 30 minutes. If 1 of the criteria was not met precisely (eg, pupils small, intoxication with a small amount of a sedative), a cerebral blood flow (CBF) study was performed. [7]

Of 503 patients with suspected brain death, 43 were aged 1-9 years and 58 were aged 10-19 years. Of the 503 patients, 187 met their criteria for brain death, and 185 of the 187 subsequently died. The 2 that survived were likely from the group of patients that did not meet the criteria precisely, but this was not indicated clearly in the report. Unfortunately, an age breakdown of those who died was not provided. No conclusions could be drawn concerning the age specificity of their criteria and its relationship to eventual death. [8]


Uniform Determination of Death Act

As more states legally recognized brain death, the American Bar Association, the American Medical Association, and the National Conference of Commissioners on Uniform State Laws worked together to develop a national standard for this determination. The Uniform Determination of Death Act was adopted as part of the President's Commission report in 1981. [9] This document stated that "an individual who has sustained either (1) irreversible cessation of circulatory and respiratory functions, or (2) irreversible cessation of all functions of the entire brain, including the brain stem, is dead. A determination of death must be made in accordance with accepted medical standards." This consensus was based on the fact that in no cases was return of brain function after a 6-hour absence documented by clinical examination and then confirmatory EEG. In the absence of confirmatory tests, a period of observation of at least 12 hours is recommended. For anoxic injury, the observation period may be increased to 24hours.

The criteria for brain death of the President's Commission are listed below. The commission's recommendations for children were similar to the American Neurological Association’s (ANA's) conclusion based on the case study by Green and Lauber: "Caution should be used applying neurological criteria to determine death in children younger than 5 years." Age-specific guidelines were not provided. [6]

President's Commission - Determination of Cerebral Death

An individual with irreversible cessation of all functions of the entire brain, including the brain stem, is dead if the following are true:

  • Cessation of all brain function is recognized

  • Cerebral functions are absent (ie, unresponsiveness)

  • The following brainstem functions are absent: pupillary light reflex, corneal reflex, oculocephalic/oculovestibular reflex, oropharyngeal reflex, and respiratory (apnea using an accepted apnea testing procedure) [10]

  • Irreversibility of brain function cessation is recognized

  • The cause of coma is established and is sufficient to account for the loss of brain function

  • The possibility of recovery of any brain function is excluded

  • Cessation of brain function persists for an appropriate period of observation or trial of therapy

  • Complicating conditions are excluded, such as the following: drug and metabolic intoxication, hypothermia, age younger than 5 years, and circulatory shock

  • The patient has been monitored for an appropriate observation period

  • Without confirmatory tests: (1) 12 hours when the etiology of the irreversible condition is well established or (2) 24 hours for anoxic injury to the brain

  • With confirmatory tests (may reduce the observation period): (1) EEG: irreversible loss of cortical functions with electrocerebral silence (ECS), together with the clinical findings of absent brainstem functions, confirms the diagnosis of brain death; (2) cerebral blood flow (CBF): absent CBF demonstrated by radionuclide scanning or intracranial 4-vessel cerebral angiography in conjunction with clinical determination of absence of all brain function for at least 6 hours is diagnostic of brain death


Task Force for Determination of Brain Death in Children

The Task Force for the Determination of Brain Death in Children was assembled in 1987 to recommend guidelines specific for children. These guidelines were a consensus opinion regarding necessary clinical history, physical examination criteria, observation periods, and confirmatory laboratory tests required to determine brain death in children. [11]

In 1995, the American Academy published a revised guideline for brain death determination in adults. No new recommendations for children were established. Updates in 2011-2012 do, however, include changes. [12]

Guidelines of the Task Force for the Determination of Brain Death in Children

History: Determination of cause of death is necessary to ensure the absence of treatable or reversible conditions (ie, toxic or metabolic disorders, hypothermia, hypotension, or surgically remediable conditions).

Physical examination findings include the following:

  • Coexistent coma and apnea

  • Loss of consciousness and volitional activity

  • Absent brainstem function - Fixed and dilated or midposition pupils; absent spontaneous and oculocaloric/oculovestibular eye movements; absent movement of facial and oropharyngeal muscles; and bsent corneal, gag, cough, sucking, and rooting reflexes

  • Spinal cord reflex withdrawal not included

  • Consistent examination throughout the observation period (see Table 1, below)

Table 1. Age-Dependent Observation Period (Open Table in a new window)


Hours Between 2 Examinations

Recommended Number of EEGs

7 days-2 months



2 months-1 year



>1 year


Not needed

Observation period: If hypoxic encephalopathy is present, observation for 24 hours is recommended. This may be reduced if an EEG shows electrocerebral silence (ECS) or a radionuclide study is negative for cerebral blood flow (CBF). One report suggests that a second EEG is not necessary at all; however, the number of patients in this study, aged 2 months to 1 year, was small.

As discussed earlier, however, in unresponsive neonates with absent brainstem reflexes, a single EEG is not sufficient to predict lack of potential survival. In infants younger than 2 months (especially when premature), the Task Force recommends that 2 EEGs 48 hours apart, both with ECS, as well as a clinical determination of brain death, be documented in order to declare death. All other age groups should be treated as adults. No data or guidelines are available for newborns younger than 7 days. Especially in premature newborns, ECS alone is not sufficient to determine brain death.


Studies on Determination of Brain Death in Children

Ashwal and Schneider reported findings in 5 children aged 28 weeks gestational age to 30 months. All 5 children were unresponsive to external stimuli and lacked cephalic reflexes and spontaneous respiration. Despite meeting these criteria for brain death, they had persistent low-voltage delta EEG activity between 2 and 4 µV. Radionuclide isotope bolus studies demonstrated no cerebral blood flow (CBF) in any of the patients, and cerebral angiography revealed no cerebral blood flow (CBF) in 4 of 5. [13, 14]

The authors concluded that this demonstrated the limitation of EEG in confirming brain death in infants and young children. However, a lack of even low-voltage ECG artifact in the cerebral channels and the presence of numerous electrode artifacts suggest technical difficulties in the acquisition of these EEGs. For these reasons, conclusions regarding the limitations of EEG in confirming brain death in children could not be made.

Drake et al studied the usefulness of confirmatory tests in 61 children who met specific criteria for brain death (age range, newborn to 18 y). These subjects were apneic. In addition, they showed no response to external stimulation, had no brainstem reflexes, and had no history of hypothermia, hypotension, drug toxicity, or metabolic disorders. Of these patients, 27 had electrocerebral silence (ECS) after 24 hours without CBF, whereas 6 had ECS with persistent CBF after 24 hours of unresponsiveness. Follow-up EEG at 48 hours showed that only 1 patient with ECS still had persistent CBF; at 72 hours, no patients with ECS had persistent CBF. This study demonstrated that children with the stated findings on examination and ECS after 24 hours do not have a reasonable prognosis for recovery. [15]

Nonaccidental trauma (eg, child abuse) is a frequent factor in childhood deaths, and courts are reluctant to accept clinical criteria alone for the determination of cerebral death prior to the removal of life support systems. Suggestions have been made that confirmatory tests be used when determining cerebral death in children.

In the algorithm recommended, if an EEG is isoelectric at 24 hours, then a CBF study should be performed. No CBF at 24 hours is a confirmation of cerebral death. If flow is present, then a repeat EEG is performed at 48 hours, at which time no activity would confirm cerebral death. Whenever an EEG activity of cerebral origin greater than 2 µV is recorded, a repeat examination 24 hours later is recommended if diagnosis of cerebral death is still being considered.

A retrospective comparison of children who met the President's Commission criteria for brain death and those who met the criteria of the Task Force was done subsequent to publication of the Task Force criteria. The study separated 47 children into 3 age groups, as follows: 2 months to 1 year, 1-5 years, and 5-17 years. All had an EEG consistent with electrocerebral silence (ECS). The observation period needed to fulfill brain death criteria was not different among the age groups. These data are contrary to the Task Force's recommendation that different age groups should be evaluated with separate criteria. Other groups have confirmed this finding.

Alvarez et al addressed the role of EEG in determining brain death. They studied the cases of 52 children younger than 5 years who met the President's Commission criteria for brain death and either died spontaneously or were disconnected from life support after being declared brain dead. Twenty-eight had 2 EEGs showing ECS 1 day apart and 24 had 1 EEG showing ECS. Autopsies on all these subjects showed pathology consistent with brain death. This strongly suggests that, in children, a single EEG is a reliable confirmatory test and a second EEG after 24 hours does not improve diagnostic accuracy. [16]


Case Reports

In 1972, Green and Lauber described 2 children, aged 5 years and 6 weeks, who each were in a coma with electrocerebral silence (ECS). After 24 hours, each showed low-voltage beta activity on EEG. As discussed previously, these cases are not well supported and no apnea or oculocephalic testing was documented in these reports. [6]

Blend et al reported a case of a 3-year-old boy with trisomy 18, mental retardation, hydrocephalus ex vacuo, and cerebral atrophy who suffered a cardiopulmonary arrest. He was described as deeply comatose, but a full neurological examination was not documented. ECS was observed; however, the radionuclide scan showed preserved cerebral blood flow (CBF). [17]

Alvarez et al explained these findings by drawing information from another patient with a skull defect. They postulated that the cerebral atrophy and hydrocephalus provided enough space during the period of anoxic encephalopathy to prevent any increase in intracranial pressure (ICP). Since CBF is proportional to mean arterial pressure (MAP) minus ICP, hydrocephalus and cerebral atrophy kept ICP low and CBF was maintained. Although EEG showed no evidence of neuronal activity, CBF continued. This finding suggested that in these children, CBF measurements may be unreliable and that EEG may be the most sensitive test when all other criteria for brain death are met. [18]

Massive ventricular enlargement and hydranencephaly in a single comatose premature neonate was reported to yield flat EEGs that improved over time. Three weeks after the initial flat EEG, low-voltage beta and delta activity appeared and responsiveness improved. The baby died 4 months later. This case suggests that a single EEG is not sufficient to predict the potential for survival in neonates.

Montes and Conn reported ECS after near-drowning in a child aged two and a half years. The child had CPR for 30 minutes, was placed on a mechanical ventilator, had initial decerebration, and was then paralyzed with pancuronium. Phenobarbital was loaded intravenously, and a core temperature of 35°C was documented. Two hours later, an EEG was performed and described as flat in all leads. Ten hours later, EEG activity reappeared and the patient went on to recover fully. The authors suggested allowing 24 hours of observation after near-drowning before declaring brain death in children. [19]

This case clearly shows that caution should be used when the clinical examination or EEG is done soon after an acute insult. Moreover, clinicians should recognize that the existence of posturing, the use of pancuronium and phenobarbital, and hypothermia preclude the clinical diagnosis of brain death as described by the President's Commission report and that a flat EEG alone does not diagnose brain death. Cases have been reported of successful CPR after prolonged near-drowning in children, but no documented cases are known of recovery after determination of clinical brain death. [20]

A review of 23 patients by Grattin-Smith et al suggested that loss of brainstem reflexes in patients with barbiturate levels of less than 40 µg/mL is not due to the action of barbiturates alone. Other intoxicants or other causes must be ruled out. However, even moderate barbiturate levels may contribute to an altered clinical picture, and the patient may appear clinically dead. [21]

This may have been the case in a patient reported by Kohrman and Spivack. This 3-month-old infant suffered from hypoxic-ischemic encephalopathy. He was found to be unresponsive, with absent brainstem reflexes, apnea, and ECS. The phenobarbital level was 21.4 µg/mL, but 4 hours later the child regained sucking movements as well as a disorganized EEG pattern. [22, 23]

Caution should be applied in diagnosing brain death when even a modest degree of intoxication has occurred, especially when the intoxication is acute.

In contrast to clinically assessing brainstem function, clinical assessment of cortical function in a comatose or brain dead patient is limited to determining an absence of spontaneous movements and reactivity to external stimuli, except for spinal reflexes. Two cases of acute inflammatory demyelinating polyradiculoneuropathy have been reported that satisfied the clinical criteria for brain death but had preserved EEG activity. Thus, EEG has an important role in the confirmation of brain death in such cases.


Studies of Other Clinical Parameters and Tests

The effect of hypothermia and total circulatory arrest during open-heart surgery was examined by Reilly et al. They studied intraoperative monitoring records and found that 6 of 14 children younger than 2 years undergoing circulatory bypass, with body temperatures from 17-23°C, had periods of electrocerebral silence (ECS) lasting 6-32 minutes. The other 8 children had some low-voltage EEG activity despite hypothermia and sustained circulatory arrest lasting 11-46 minutes. Those who were hypothermic with ECS had no subsequent differences in clinical course. [24]

This study demonstrates that hypothermia may produce a situation resembling brain death by clinical criteria and laboratory confirmation. These observations support the Task Force's recommendation that hypothermia should be eliminated before declaring brain death.

Quantification of autonomic variability has shown that heart rate variation is decreased in brain dead children and that a cold pressor test (immersion of the hand in ice water) yields no variations in heart rate.

Apnea testing for determination of brain death in adults, as specified in the President's Commission report, requires that 100% oxygen be administered for 10 minutes, followed by removal of ventilatory support and oxygen for 10 minutes. [25, 26] Arterial blood gases (ABG) at 10 or 15 minutes should show a PaCO2 greater than 60 mm Hg, and withdrawal of support should not induce spontaneous respiration. When children were evaluated, their PaCO2 rose faster due to a higher metabolic rate; ABG testing for apnea at 5 minutes was considered adequate. However, use of therapeutic hyperventilation may prolong the time needed for the PaCO2 to reach 60 mm Hg. In Canada, and in the 2011-2012 guidelines, two brain death examinations are specified. [27]

Several groups studied the effects of barbiturate levels on the determination of a diagnosis of brain death in children aged 1 day to 18 years. They found that barbiturate levels of less than 40 mg/mL when allowed to fall to 0 µg/mL had no effect on the eventual outcome of children with a diagnosis of brain death. [28]

Another important clinical observation in children is that central diabetes insipidus occurs in 38-88% of children with a diagnosis of brain death. This results from destruction of hypothalamic neurons. [29, 30]

Additional confirmatory tests

As demonstrated by 4-vessel angiography, the criterion standard of CBF studies, absence of CBF by radionuclide imaging also occurs in brain dead children older than 13 months. [14, 31] Similar findings have been reported using xenon CT scanning and in a case report in which positron emission tomography (PET) scan was used. [32]

In children aged 2-8 years with clinical brain death, Tc-99m-hexamethylpropyleneamine oxime (HMPAO) single photon emission computed tomography (SPECT) demonstrated nonvisualization of the cerebrum and cerebellum in cases where the EEG was flat. In 2 newborns at 7 days of life with similar clinical and EEG findings, the first SPECT scan showed activity, but a follow-up 5 days later demonstrated no activity. No clinical change was noted during the intervening 5 days. This suggests that 2 SPECT scans separated in time are needed to confirm brain death in newborns. [33, 34]

Ashwal et al found that any evidence of EEG activity (ie, burst suppression, occipital or temporal activity) or clinical seizure activity in infants younger than 1 year correlated with positive CBF by radionuclide scanning. They also found lack of CBF in 4 infants younger than 5 months who had ECS and absent cephalic and brainstem functions and were unresponsive and apneic. [32]

A few pitfalls always should be kept in mind when interpreting CBF studies. As pointed out by Holzman et al, children who are hypothermic or on barbiturates may continue to have normal CBF in the presence of attenuated EEG activity and a clinical examination consistent with brain death. [35] In addition, as discussed earlier, hydrocephalus and cerebral atrophy alter intracerebral pressure dynamics and may allow CBF despite EEGs consistent with ECS. [35]

Evoked potentials may be useful as a confirmatory test. In brainstem auditory evoked potentials (BAEPs), conduction through cranial nerve VIII (wave I) must be demonstrated in order to evaluate other brainstem responses (waves II - VII). In the absence of a wave I, no comment can be made about the absence of the more proximal responses. Wave II may be present in brain death, since it may originate from the proximal portion of the auditory nerve. Similarly, peripheral nerve responses must be demonstrated to evaluate the function of cortically generated somatosensory evoked potentials (SEPs).

In 10 clinically brain dead children aged 3 months to 17 years studied by Steinhart and Weiss, BAEPs were absent except wave I, which is due to cranial nerve VIII function. All 13 comatose but non–brain dead patients had preserved BAEPs. This suggests that BAEPs may be useful in confirming brain death; however, this confirmatory test is not specific and can have a significant number of false-positive results. [36]

Initially absent BAEPs returned to normal 10 days later in anoxic infants who were unresponsive and had apnea, loss of corneal reflexes, and burst suppression on EEG. Recovery of BAEPs has also been reported in a single comatose 33-month-old child who was not brain dead by clinical examination. [37]

Newer diagnostic tests have been explored for potential usefulness in children. [38] Common carotid pulsed Doppler ultrasound revealed a characteristic pattern of a large positive systolic spike and a second lower amplitude positive peak in middiastole in 19 of 23 brain dead children older than 4 months. A transorbital approach to transcranial Doppler ultrasound (TCD) has also been found useful in adults but has not been examined in children. [39] Phosphorus-31 MRI also may have prognostic value for children younger than 5 years who are brain dead by clinical examination. However, only 3 children were evaluated in this study. [40, 41, 42]

A single patient with clinical brain death and an isoelectric EEG was found to have normal brain glucose metabolism by PET scanning. Autopsy revealed severe white matter necrosis and intense mononuclear cell infiltrates in the cortical layers. The glucose metabolism was secondary to an inflammatory reaction and not due to neuronal activity. This suggests that the presence of glucose metabolism in PET scanning should not be used to alter a clinical diagnosis of brain death that has been confirmed by EEG. [43]


Conclusions, Ethical Considerations, and Controversies

The Task Force's criteria can be applied with a number of caveats. For example, confirmatory testing with EEG or cerebral blood flow (CBF) studies is necessary in fulminant cases of acute peripheral neuropathies, such as acute inflammatory demyelinating polyradiculopathy or infant botulism.

A single EEG with a uniform observation period of 12 hours (24 h if the etiology is anoxia) is likely to be sufficient for children older than 2 months. Infants younger than 2 months require 2 EEGs; in those younger than 7 days, a diagnosis of brain death should be made with caution.

"Do not resuscitate" orders are written for children in hospitals in the United States on the basis of these criteria; however, a more controversial issue has been the application of these criteria to anencephalic newborns. [44] The use of anencephalic newborns as organ donors is dependent on declaring them legally dead. [45] The lack of higher cortical functions in anencephaly with preservation of brainstem reflexes does not meet present day criteria for brain death. To use these subjects for transplantation raises an ethical dilemma and would mean radically altering our concepts of cerebral death.

Fackler and Rogers proposed the controversial idea that brain death should be declared despite minimal EEG and CBF measurements. They reported a case of an 8-year-old child with an intracranial abscess and high intracranial pressure (ICP) who was unresponsive, without brainstem reflexes, and apneic. This patient had a "nearly flat" EEG, and a 4-vessel angiogram demonstrated some right and left carotid and middle cerebral artery filling. The authors raised the question "... whether brain death should be declared even while the brain is found to have ineffective cerebral perfusion or to have traces of feeble and ineffective electrical activity?" [46]

Other situations have been reported in which "fragments" of low-voltage EEG activity were measured when the neurological examination was consistent with brain death. Plum and Posner found that 37 of 449 patients described as clinically brain dead had "fragments" of low-voltage EEG activity and that generation of EEG activity by islands of perfused neurons was not uncommon. [47] To date, ineffective but present EEG activity or CBF has not been quantified, making application of this controversial notion of brain death difficult.

Over the last decade and a half, organ donation after removing life support, but before brain death is determined, has become medically and legally controversial. Expanding the pool of potential pediatric heart donors beyond those who meet brain death criteria can help meet a pressing need. [48, 49]

About 100 infants younger than 1 year receive life-saving heart transplants every year. But as many as 50 infants in need of heart transplants die each year while waiting on the United Network for Organ Sharing list, according to a New England Journal of Medicine editorial. [50] The timing of transplantation has also proven contentious. The so-called Pittsburgh protocol, published in 1993, called for a 2-minute wait after cardiopulmonary arrest before declaring death and retrieving organs. The Institute of Medicine in 1997 said transplant teams should wait 5 minutes after cardiac functioning ceases before retrieving organs. [51]