False-Positive Diagnosis of Brain Death Following the Pediatric Guidelines: Case Report and Discussion

D. Alan Shewmon, MD1

Abstract

Journal of Child Neurology
2017, Vol. 32(14) 1104-1117

A 2-year-old boy with severe head trauma was diagnosed brain dead according to the 2011 Pediatric Guidelines. Computed tomographic (CT) scan showed massive cerebral edema with herniation. Intracranial pressures were extremely high, with cer- ebral perfusion pressures around 0 for several hours. An apnea test was initially contraindicated; later, one had to be terminated due to oxygen desaturation when the PCO2 had risen to 57.9 mm Hg. An electroencephalogram (EEG) was probably isoelectric but formally interpreted as equivocal. Tc-99m diethylene-triamine-pentaacetate (DTPA) scintigraphy showed no intracranial blood flow, so brain death was declared. Parents declined organ donation. A few minutes after withdrawal of support, the boy began to breathe spontaneously, so the ventilator was immediately reconnected and the death declaration rescinded. Two hours later, life support was again removed, this time for prognostic reasons; he did not breathe, and death was declared on circulatory- respiratory grounds. Implications regarding the specificity of the guidelines are discussed.

The criteria that physicians use in determining that death has occurred should:

1. (1)  Eliminate errors in classifying a living individual as dead,
2. (2)  Allow as few errors as possible in classifying a dead body as alive . . .

The 2011 Guidelines for the Determination of Brain Death in Infants and Children (hereinafter, “the Guidelines”),2-4 patterned closely after the American Academy of Neurology’s 2010 guideline update for determining brain death in adults,5 are considered the current diagnostic standard in the United States, having been endorsed by 9 specialty societies. The Guidelines are widely believed to possess the requisite speci- ficity for a criterion for death, and alleged counterexamples have been dismissed as misdiagnoses because of the failure to follow them completely. Citing a chapter by Ashwal,6 the Guidelines affirm that “there are no reports of children recovering neurologic function after meeting adult brain death criteria based on neurologic examination findings.”2(p.e727) In the words of the lead author, “Reports of people who have survived following a determination of neurologic death are the result of diagnostic error, and these patients did not meet initial criteria for neurologic death.”i Joffe and colleagues reported a false-positive diagnosis of brain death based on the criteria of the Canadian Neurological Determination of Death Forum, which do not require a second examination.7 Their patient breathed during the apnea test in conjunction with a second examination 2 days after the first. The drafters of the US Guidelines dismissed this case as not a true counterexample to the US criteria, which require fulfillment of a second examination and apnea test.2(p.e727) The authors considered it “likely that if a second brain death examination had been done within hours after the first, it would have confirmed brain death.” That is probably true, and if “within hours” meant “between 12 and 15 hours,” then the patient would probably have fulfilled the US Guidelines as well as the Canadian criteria; but that remains speculative.

1 David Geffen School of Medicine at UCLA, Los Angeles, CA, USA

Corresponding Author:

D. Alan Shewmon, MD, Pediatrics Department, Olive View-UCLA Medical Center, 14445 Olive View Drive, Sylmar, CA 91342-1495, USA.
Email: ashewmon@mednet.ucla.edu

Shewmon

1105

There are 4 reported cases of recovery of brain function in adults after allegedly fulfilling the adult criteria. One was contended by writers of the adult guideline update and defended again by the authors as a legitimate case.8-10 Another had incomplete documentation of 2 apnea tests that reportedly revealed no respiratory effort, but a 4-vessel angiogram showed no opacification of all intracranial ves- sels.11 Two others were considered to have fulfilled clinical criteria, but there were several unrecognized confounding factors; the authors advocated for more routine use of ancil- lary blood flow tests, which happened to reveal flow in both cases.12

I recently consulted on a case involving a false-positive diagnosis of brain death following the pediatric Guidelines. The case will be presented, followed by a discussion of the specificity of the Guidelines for diagnosing “irreversible cessa- tion of all functions of the entire brain” (the definition of brain death endorsed by the Guidelines, quoted from the Uniform Determination of Death Act1(p119)).

Case Report

A 2-year-old boy was tragically run over by a slowly mov- ing car in his own driveway, sustaining crush injuries to the head and thorax. Family members rushed him directly to a nearby emergency room, where he was in full cardiopul- monary arrest. There was bleeding from the right ear and both nares, swelling of the right side of the face, and mul- tiple abrasions, especially over the left thorax. He was immediately intubated, and an intraosseous line was inserted. He received 2 rounds of cardiopulmonary resusci- tation, and a spontaneous pulse was obtained after 5 min- utes. He remained motionless, not overbreathing the ventilator. Pupils were fixed and dilated, right larger than left, and the Glasgow Coma Score was 3. He was placed in a cervical collar and transferred promptly to a nearby pedia- tric trauma center.

An immediate computed tomography (CT) scan of the head (Figure 1) showed extensive pneumocephalus and loss of gray- white differentiation in the bilateral frontal and temporal lobes, concerning for edema. Maxillofacial CT revealed multiple, severe fractures of the skull base, midface, right orbit and right zygoma, and dislocation of the left temporomandibular joint with displacement of the left mandibular condyle into the left middle cranial fossa (Figure 2).

CT of the cervical spine showed no acute fracture and a small left apical pneumothorax. CT of the chest, abdomen, and pelvis revealed nondisplaced fractures of the left first through seventh posterior ribs, diffuse, mild fluid-filled distension of the small bowel with hyperenhancing mucosa, suggestive of shock bowel, and air within the right femoroacetabular joint without fracture. Chest radiograph showed severe, diffuse bilateral interstitial opacities, consistent with either edema or aspiration pneumonitis.

On admission to the pediatric intensive care unit at 2.2 hours, blood pressure was 86/41 and heart rate 180. (The timing

Figure 1. Initial head computed tomography (CT) at 1.6 hours, showing extensive pneumocephalus and early cerebral edema. (A) Axial cut, (B) reconstructed midsagittal plane.

Figure 2. Maxillofacial computed tomography (CT): (A) Axial image showing multiple fractures of the skull base, (B) coronal image showing displacement of the left mandibular condyle into the middle cranial fossa.

of events will be presented as hours from injury.) The boy received fluid resuscitation and required epinephrine and dobu- tamine infusions to maintain blood pressure. On examination by a neurosurgeon at 3 hours, there was no movement of upper or lower extremities. The left pupil was nonreactive and the right pupil was sluggish. Around the same time a pediatric surgeon described the right pupil as dilated and both pupils sluggishly reactive.

The neurosurgeon placed a Codman intracranial pressure– monitoring bolt (intraparenchymal), which showed an initial pressure of 9 mm Hg. The time courses of blood pressure, heart rate, and intracranial pressure are plotted together in Figure 3, along with the timing of the most important diagnostic tests. The patient’s bed was placed at a 30 angle, and ventilator parameters were adjusted to maintain arterial PCO2 in the low to the mid-30s range.

Over the next 12 hours, the cerebral perfusion pressure (CPP) fluctuated between 30 and 80 mm Hg, but after 16.2 hours, difficulty maintaining blood pressure resulted in CPPs consistently less than 40, and less than 30 after 18.2 hours (Figure 4). Therefore, a vasopressin infusion was added at 19.3 hours. Figure 5 summarizes the time courses of the 3 pressor medications. Around the same time, the boy’s intracra- nial pressure suddenly shot up, reaching a maximum of 110 mm Hg at 21.2 hours (Figures 3 and 5). The blood pressure rose in parallel, reaching a maximum of 166/120 despite drastic

1106 Journal of Child Neurology 32(14)

Figure 3. Time courses of heart rate (beats/min), blood pressure (mm Hg), and intracranial pressure (mm Hg), together with timing of major diagnostic tests. Abbreviations: BD, brain death; Bl Flow, blood flow; HR, heart rate; MAP (art), mean arterial pressure from arterial line monitor; MAP (cuff), mean arterial pressure from the upper arm cuff.

Figure 4. Time course of cerebral perfusion pressure.

reductions in the pressors, with CPPs in a low range around 25 mm Hg (Figure 4). The malignant intracranial hypertension was so precipitous that there was hardly an opportunity to initiate osmotic treatment before it became clear that the battle had already been lost.

A repeat head CT scan at 21.2 hours showed massive cere- bral edema with loss of gray-white differentiation, obliteration of the ventricles and basal cisterns, and uncal and tonsillar

herniation (Figure 6). A neurosurgery note around then docu- mented that the pupils had become dilated to 8 mm and fixed, and noxious stimuli elicited no reaction. Because vecuronium and fentanyl were no longer needed to assist ventilation and a brain death evaluation would soon be undertaken, those med- ications were discontinued at 22.6 and 22.7 hours, respectively.

The blood pressure began to drop, so the epinephrine and vasopressin infusions were increased starting around 40 hours

Shewmon 1107

Figure 5. Time courses of the 3 pressor medications, dobutamine and epinephrine in mg/kg/min and vasopressin in units/kg/min. Because the

scale for dobutamine is different from the other two, a separate axis

Figure 6. Head computed tomographic scan at 21.2 hours, showing massive cerebral edema and brain herniation. (A) Axial image, (B) reconstructed midsagittal plane, (C) reconstructed coronal plane through the foramen magnum.

(Figures 3 and 5). Because of the refractoriness of the intracra- nial pressure and the judgment of a lost cause, the neurosurgeon canceled the order for frequent measurement of intracranial pressure, to lighten the nursing burden; thereafter, intracranial pressure documentation became sporadic, but whenever it was checked, it was always above 50 mm Hg. During one 2-hour period (53.2-55.2 hours), it was between 95 and 114, with CPPs between –4 and 1 mm Hg.

On rounds the next morning, the neurosurgeon found no evidence of brain stem function; specifically, pupils were still dilated at 8 mm and nonreactive, there was no cough or gag, and no motor response to noxious stimuli. He recommended a formal brain death evaluation, at which point I was consulted.

An EEG had already been performed at 43.7 hours, when the CPP was documented at only 4 mm Hg (Figure 4) and the patient’s body temperature was 36.5 C. The EEG, performed according to the American Clinical Neurophysiology Society’s guidelines for suspected cerebral death,13 showed no definite electrocerebral activity at a sensitivity of 2 mV/mm, consistent

is given for it on the right.

with electrocerebral silence (ECS) (Figure 7). There was no behavioral or electrographic response to loud noise, flashing lights, or noxious tactile stimuli. Regularly repeating, very low- amplitude waves (around 4-8 mV) in the theta/alpha range max- imal at T5 and F3 were clearly ventilator-related artifacts. But on rare occasions similar waveforms occurred mainly at T5 between breaths (Figure 8). Although I suspected these to be artifactual as well, to err on the safe side I formally interpreted the EEG as “equivocal” for ECS: it was probably isoelectric, but one could not be absolutely sure that the waves at issue were not of cerebral origin.

During a brief period off the ventilator, as part of the hos- pital’s ECS protocol, despite an FiO2 of 100% and a PO2 around 100, the patient’s heart rate plummeted after only 20 seconds, so he was immediately reconnected to the ventilator; the heart rate quickly returned to its previous range. Such marked instability was considered a medical contraindication to formal apnea testing.

At 46.8 hours, the intensivist documented a Glasgow Coma Score of 3, pupils fixed and dilated, and absence of corneal, gag, and oculovestibular reflexes.

I examined the patient at 49.2 hours; his temperature was 36.5 , blood pressure 119/70, and heart rate 117. The ventilator rate was 22, and he did not breathe over it. He exhibited no response to name calling or loud noise and no response to supraorbital pressure or sternal rub. Cranial nerve examination revealed pupils 4 mm, equal and nonreactive. Corneal, oculo- cephalic, oculovestibular, and gag reflexes were absent. There was no cough or other response to deep tracheal suctioning. Motor examination showed diffusely flaccid tone and no spon- taneous movements. With fingernail pressure on either side, there was slight extension at the wrist. With toenail pressure, there was a weak triple flexion response. Tendon reflexes were

1108 Journal of Child Neurology 32(14)

Figure 7. ECS montage, sensitivity 2 mV/mm, patient temporarily off the ventilator. No discernible activity apart from a 20 Hz subharmonic of filtered 60-cycle artifact and ECG/pulse artifact. Vertical calibration bars to the left of each channel are 10 mV. ECG, electrocardiograph; ECS, electrocerebral silence.

Figure 8. Very low-voltage (4 mV) 7- to 8-Hz waves at T5 (rectangle) between respiration artifacts (ellipses). The waves were better seen on this “double banana” montage than the ECS montage. Sensitivity 1⁄4 2 mV/mm. Vertical calibration bars to the left of each channel are 10 mV. ECS, electrocerebral silence.

absent throughout, and plantar responses were mute bilaterally. Interpreting the responses to nailbed pressure as spinal reflexes, I concluded that he fulfilled the clinical components of the Guidelines, short of an apnea test.

Because of the contraindication to apnea testing, the equivocal formal interpretation of the EEG, and a slight concern whether brownish debris in the left external auditory canal might have lessened the ice-water caloric stimulus on that side, a technetium-99m diethylene-triamine-pentaacetate (DTPA) blood flow study was performed at 54.1 hours. The excellent quality images showed no discernible intracranial flow on the dynamic sequence, and the static images showed no visible uptake of isotope by the brain (Figure 9). The study was formally interpreted as positive for brain death, with “complete lack of perfusion and uptake throughout the brain parenchyma.”

A second formal brain death examination was conducted by the pediatric intensivist at 70.7 hours, 21.5 hours after my first

examination. Body temperature at the time was 36.9 . There were no spontaneous movements and no response to noxious stimuli except for triple flexion with toenail pressure. Tone was flaccid. Pupils were dilated and fixed. Corneal, oculocephalic, oculovestibular, and gag reflexes were absent. There was no cough or response to tracheal suctioning.

An apnea test was attempted. The patient had been preoxy- genated with 100% FiO2 all along, and the pretest arterial blood gas was as follows: PCO2 41.9, PO2 182.8, and pH 7.274. After 2.5 minutes off the ventilator, the oxygen saturation plum- meted, and he was immediately reconnected to the ventilator. During those 2.5 minutes, no respiratory effort was made. The immediate posttest arterial blood gas was as follows: PCO2 57.9, PO2 48.5, and pH 7.165. Although the apnea test was not quite complete (the final PCO2 of 57.9 almost met the Guidelines’ minimum requirement of 60, and the 16 mm Hg rise above baseline almost met the minimum requirement of 20), the blood

Shewmon

1109

Figure 9. Technetium-99m DTPA blood flow study showing no perfusion of the brain. (A) Selected images from the dynamic sequence; (b) static phase. DTPA, diethylene-triamine-pentaacetate.

flow study established the diagnosis according to the Guide- lines’ algorithm.

The intensivist therefore formally declared death at 70.7 hours. The regional organ procurement organization had pre- viously been notified, and a representative was present to dis- cuss donation with the parents, who declined and wished their deceased son to be disconnected from the lines and machines so they could say goodbye in a more dignified manner.

The patient was disconnected from the ventilator at 71.2 hours. After 2 or 3 minutes (possibly more—no one was keep- ing time), as he became markedly bradycardic and cyanotic, to everyone’s shock he began to breathe spontaneously. The father was the first to point this out, and it was witnessed by the nurse and the intensivist as well as other family members. The quality of the breaths was described by the intensivist as not agonal, but “slow” and “neither deep nor shallow.” After 3 or 4 such breaths, the intensivist immediately reconnected the patient to the ventilator and rescinded the death declaration in the medical record. Family members were dismayed that their son, who had just been pronounced dead, with organ donation requested and life support discontinued on that basis, was obvi- ously still alive. The medical and nursing staff were also greatly distressed.

After further deliberation among family members and with the support of one of the surgical consultants who happened to be of their ethnic group, parents elected to withdraw life sup- port, this time on the basis of the grim prognosis for survival and neurologic outcome. At 73.2 hours, he was therefore dis- connected from the ventilator a second time, and this time he remained apneic. He became cyanotic and turned pulseless at 73.5 hours, which was recorded as the official time of death by circulatory-respiratory criteria.

Discussion

Fulfillment of the Guidelines in this case is most easily sum- marized in the form of the completed Check List for Documen- tation of Brain Death (Guidelines, Appendix 1 for infants and children), presented in the Appendix of this paper. The false- positive diagnosis hinged on the role of ancillary tests in the Guidelines, in particular on the radionuclide blood flow test. But before focusing on that, let us consider a different potential source of false-positive error, discussed and dismissed in the Guidelines.

Minimum PCO2 Requirement

After an extensive literature review, the multi-society drafting Committee singled out for consideration 4 published cases that seemed to represent exceptions to the Guidelines. All were instances of “irregular breaths or minimal respiratory effort with a pCO2>60 mm Hg in children who otherwise met criteria for brain death.”2(p.e727) Two did not satisfy the Guidelines, because they had only one formal brain death examination and breathed during the apnea test.14,15 Nevertheless, the cases are still concerning, because the PCO2 that stimulated breathing was far above the Guidelines’ minimum requirement of 60 mm Hg, namely 91 and 112; had the apnea tests been of slightly shorter duration, with final PCO2 between 60 and 91, they could easily have fulfilled the Guidelines for the first examination and apnea test. Who knows what the second examination and apnea test might have shown?

The third case cited by the Committee, reported by Haun et al, had 2 examinations and a single apnea test, which fol- lowed the second examination.16 There was no respiratory effort as the PCO2 rose from 42 to 67 mm Hg. Brain death was declared, but after life support was withdrawn, the patient began to breathe regularly at a rate of 15 to 20/min. The PCO2 after several minutes of spontaneous respirations was 86, implying that it was even higher when he started to breathe. Although the case fulfilled institutional criteria for brain death, technically it did not fulfill the Guidelines, which require 2 apnea tests, one with each examination. It is likely, however, that had an apnea test been done along with the first examina- tion, the result would have been similar, and the patient would have fulfilled the Guidelines completely.

The fourth case was even more concerning. Okamoto and Sugimoto reported a 3-month-old who fulfilled all brain death criteria including 2 isoelectric EEGs and 2 apnea tests with

1110

final PCO2s of 69.3 and 62.1 mm Hg 2 days apart.17 The infant continued to fulfill clinical criteria (excluding repeat apnea tests) on days 7, 14, and 21. On day 43, she regained sponta- neous respiration, with an irregular rate of 2 to 3 per minute and tidal volume of 40 to 50 mL. Breathing continued until she died of pneumonia on day 71.

From these 4 cases, the Committee concluded:

Review of this case and others remind us to be cautious in applying brain death criteria in young infants. However, these cases should not be considered to represent reversible deficits or failure of cur- rent brain death criteria.2(p.e727)

But the “caution” applies not only to young infants. Two of the 4 cases involved a 4-year-old and a 3-year-old. Besides, if the criteria are 100% reliable, what can “cautious” mean beyond following them to the letter, which should be done anyway for any standard criteria? The only other way to be “cautious” is not to put faith in the validity of the criteria, at least as far as PCO2 threshold goes. It is also hard to understand the logic of the last quoted sentence, which the Committee took from a commentary by Fishman.18 Why should these cases “not be considered to represent reversible deficits,” when a deficit in fact reversed? And why was the reversibility of such deficit not a “failure of current brain death criteria,” which purportedly establish “irreversible cessation of all functions of the entire brain”?

Hansen and Joffe recently reported yet another case with breathing at a PCO2 above the Guidelines’ minimum require- ments (“Case 2,” a 15-year-old with final PCO2 of 61 mm Hg, 22 over baseline).19

Both Haun et al and Joffe et al pointed out another serious problem with the apnea test, namely that “the combination of hypercarbia and hypoxia is a greater ventilatory stimulus than hypercarbia alone”16(p183) and that medullary gasping is sti- mulated by hypoxia rather than hypercarbia.20(pp1438-1439) Therefore, “the hyperoxia required to perform an apnea test without causing dangerous hypoxia can actually be what causes apnea to manifest, because of the suppression of func- tion in the damaged preBo ̈tzinger complex.”20(p1438) So an apnea test can be falsely diagnostic on that basis, regardless how high the PCO2 is.

The present case and that of Haun et al seem to bear that out, and the current patient’s respirations were described as even more normal than agonal gasping. The resumption of breathing after discontinuation of support, but not during the apnea test, suggests indeed that hypoxia rather than hypercarbia, or per- haps the combination, was what stimulated the respiratory effort. At the rate of rise of PCO2 (16 mm Hg in 2.5 minutes), had the apnea test continued for only 38 more seconds, the final PCO2 would have reached 62 mm Hg and fulfilled the require- ments for a valid apnea test (>60 and 20 above baseline). We will never know whether the patient would have remained apneic during those 38 seconds, but it seems likely enough to reinforce the already expressed concerns about the Guidelines’ minimum requirements for declaring apnea.

Journal of Child Neurology 32(14) Confirmatory Value of Ancillary Tests

We turn now to the reliability of ancillary tests and their role in the Guidelines. These tests used to be called “confirmatory tests.” The reason for this relatively recent change in the brain-death lexicon is explained by the Committee: “The term ‘ancillary study’ is preferred to ‘confirmatory study’ since these tests assist the clinician in making the clinical diagnosis of brain death.”2(p.e728)

In some other countries ancillary testing is a mandatory component of the diagnostic protocol,21-24 but in the United States ancillary tests “are not required to establish brain death and should not be viewed as a substitute for the neurologic examination.”2(p.e728) Nevertheless, they can assist the clinician in making the diagnosis of brain death
(i) when components of the examination or apnea testing can- not be completed safely due to the underlying medical condition of the patient;
(ii) if there is uncertainty about the results of the neurologic examination;
(iii) if a medication effect may be present; or
(iv) to reduce the inter-examination observation period.2(p.e728)

Regarding (i), the Guidelines require an apnea test with

each of the 2 neurologic examinations, “unless a medical

contraindication exists. Contraindications may include con-

ditions that invalidate the apnea test (such as high cervical

spine injury) or raise safety concerns for the patient (high

oxygen requirement or ventilator settings). If apnea testing

cannot be completed safely, an ancillary study should be

performed to assist with the determination of brain death.”2(p.e727)

In the present case, the patient did not tolerate even 20 seconds off the ventilator during the EEG recording, so an apnea test was considered medically contraindicated at the time of the first brain death examination a few hours later. For this reason, in accordance with the Guidelines, I recommended a radionuclide blood flow study (because the EEG was formally read as equivocal for ECS).

After reviewing the literature on the diagnostic yield of

EEG and radionuclide cerebral blood flow (CBF) studies in

children, the Committee concluded that “both of these ancil-

lary studies remain accepted tests to assist with determination

of brain death in infants and children. The data suggest that

EEG and CBF studies are of similar confirmatory value.”2(p.e729)

But there are logical problems with the whole concept of ancillary tests. In situations that require an ancillary test according to the Guidelines’ algorithm, how can it “assist” with the diagnosis, if it is not in fact “confirmatory”? And if the test indeed has “confirmatory value,” why should we no longer call it a “confirmatory test”? And how is it not a “substitute” for a deficiency in the clinical examination or apnea test, if, without it, brain death cannot be declared because of the deficiency, but with it, brain death can be declared despite the deficiency?

Shewmon

1111

The Committee concluded that “other ancillary studies such as the Transcranial Doppler study and newer tests such as CT angiography . . . have not been studied sufficiently nor vali- dated in infants and children and cannot be recommended as ancillary studies to assist with the determination of brain death in children at this time.”2(p.e729) (Wijdicks et al came to the same conclusion as regards the adult criteria.5) But more importantly, neither EEG nor radionuclide blood flow scans have been validated as possessing an essentially infinitesimal false-positive rate, as required of a test for death.

Confirmatory value of EEG. It is widely known that EEGs can be falsely negative in brain death (EEG activity despite brain death),25-34 because electrocerebral activity is not considered a clinical brain function, and brain death is supposed to be a clinical diagnosis.35-38 More important, however, are the false positives (an isoelectric EEG despite clinical brain func- tion).8,39-43 The case of a 3-month-old reported by Kohrman and Spivack, not cited by the Guidelines, seems to contradict both the confirmatory value of EEG and the infallibility of the Guidelines.43 Although the baby had only one apnea test (to a PCO2 of 64 mm Hg), there were 2 isoelectric EEGs, overabun- dantly compensating for the missing second apnea test accord- ing to the Guideline’s algorithm.ii Around 4 hours after the second isoelectric EEG, the baby began making sucking move- ments. “Over the next 3 days, she continued sucking and devel- oped spontaneous eye-opening and blinking, spontaneous and elicited facial grimacing, a normal oculocephalic response, spontaneous eye movements with vertical nystagmus, and bilaterally intact corneal responses. She did not regain sponta- neous respirations. These brain stem responses continued unchanged for the next 30 days.”43(p47) Ashwal and Schneider reported a term newborn with clinical brain death, who had a nearly isoelectric first EEG (only a photic response) and a completely isoelectric second EEG, but then developed spon- taneous respirations and motor activity.31

In the present case, the EEG was probably isoelectric, but the interpretation was confounded by a few very low-amplitude waveforms that were probably, but not certainly, artifacts—a not uncommon problem in the electrically noisy intensive care setting.34,44,45 It is quite possible that another expert electro- encephalographer might have interpreted them as definite arti- facts, in which case the EEG would have sufficed for the patient to have fulfilled the Guidelines even without a blood flow study. In any event, since the Guidelines require only one ancillary study, and the radionuclide scan was definitive, the “equivocal” EEG did not contradict brain death; it merely failed to support that diagnosis sufficiently to serve as the only ancillary test.

Although much lower than the false-negative rate, the false- positive rate for EEGs is clearly not zero.34 This should not be surprising, since it has been known for decades that EEGs in hydranencephalic infants, for example, can be virtually isoelec- tric despite completely intact brain stem functions.46-48 Although this is a different clinical context from the kinds of brain insults that progress to brain death, the phenomenon

highlights the fact that EEG findings imply nothing about the

integrity of most of the brain, especially the brain stem.49 Nei-

ther does an isoelectric EEG per se establish irreversibility,

even excluding cases of medication effect or hypother- mia.43,50-55

Because an isoelectric EEG establishes neither totality nor irreversibility of brain nonfunction, it cannot possibly confirm “irreversible cessation of all functions of the entire brain, including the brainstem” in the face of clinical doubt. As Ber- nat succinctly summed up:

Showing electrocerebral silence by EEG is the oldest confirmatory test but generates too many false-positive determinations to be reliable. . . . The most reliable confirmatory tests are those demon- strating the cessation of intracranial blood flow.56(pp374-375)

That conclusion was echoed in a comprehensive review of ancillary tests by Young et al:

Thus, the EEG is far from an ideal or even a suitable test for brain death; it meets very few of the desired criteria for a suitable ancil- lary test. At best EEG, when applied to brain death, is mildly confirmatory; at worst it is misleading or irrelevant. Thus, despite official recommendations in some countries, EEG seems poorly suited as a necessary component of brain death declaration in children. When confirmation is needed, a test of blood flow to the brain is more appropriate.34(p622)

The Guidelines also acknowledge that “radionuclide CBF techniques are increasingly being used in many institutions replacing EEG as an ancillary study.”2(p.e729) It is therefore difficult to understand how the Committee, after thoroughly reviewing the literature, concluded that “the data suggest that EEG and CBF studies are of similar confirmatory value”2(p.e729) (emphasis added). An isoelectric EEG is cer- tainly supportive of the diagnosis, but in no way does it possess “confirmatory value” with the specificity requisite for a declaration of death.

Confirmatory value of radionuclide scintigraphy. This brings us to the confirmatory value of blood flow studies and the impor- tance of the current case. Although the “gold standard” test of brain blood flow is 4-vessel contrast angiography,2(p.e728),34,57-60 its invasiveness, risks, requirement of special procedural exper- tise and other impracticalities (especially in children) have led to its having given way to radionuclide studies as the most widely used test of brain blood flow, specifically endorsed by the Guidelines.2(p728)

The literature on planar radionuclide scans suggests that both sensitivity and specificity for clinical brain death (i.e., for the “cardinal” triad of coma, brain stem areflexia, and apnea) are high. In a detailed review, Joffe and colleagues determined the overall sensitivity to be 119/1531⁄477.8% (95% confidence interval 70.5%-83.7%) and the specificity to be 41/411⁄4100% (95% confidence interval 92.6%-100%).61 The statistics for single-photon emission computed tomography (SPECT) were similar, except for smaller N and wider confidence interval for

1112

Journal of Child Neurology 32(14)

specificity (12/121⁄4100%, confidence interval 78.4%-100%). The authors emphasized the importance of including lateral views, to optimally visualize cerebellar flow. They also pointed out the superiority of SPECT over planar imaging for detecting flow to the brain stem and the advantages of the newer diffu- sible tracers over the older nondiffusible ones.

The 100% specificity suggests that radionuclide imaging could be a reliable confirmatory test for clinical brain death. But as Joffe et al have pointed out, “the limited numbers of patients that allow a statement about the specificity of planar and SPECT imaging for brain death are concerning. . . . These numbers may be inadequate when the test is used to confirm a diagnosis of death, with a desirable specificity of 100%.”61(p58) They emphasized that “the tests for brain blood flow are being used not to give a prognosis of poor neurological outcome or death, rather, they are being used to diagnose the state of death. For this reason, the specificity of the test should be clar- ified.”61(p58) (emphases in original)

But more importantly, specificity for clinically diagnosed brain death is not even the kind of specificity that really mat- ters. Because radionuclide blood flow testing is supposed to have “confirmatory value,” its specificity should be measured against the “true condition” (i.e., brain death vs not brain death), not against some other test (i.e., the clinical criteria), which requires its own independent validation and which, by implication, may stand in need of confirmation by the “confirmatory” test.

The condition that both the clinical criteria and the ancillary test are intended to diagnose, and against which sensitivity and specificity should be measured, is “the irreversible cessation of all functions of the entire brain, including the brainstem.” Because the ultimate basis for irreversibility and totality of nonfunction is neuronal cell death throughout the brain, caused by the vicious cycle of increased intracranial pressure and her- niation, a reliable and sufficient way to know that all functions of the entire brain have irreversibly ceased is to demonstrate absence of blood flow to the entire brain.62,63

For the pediatric population, however, blood flow testing is particularly problematic in neonates. As Fishman expressed:

Another factor to consider is that the lower limits of blood flow necessary to support the immature brain have not been determined. Altman et al64 have demonstrated flows of 7 to 11 mL/100 g/ minute by positron emission tomography are sufficient to sustain the brains of premature infants. Minimal blood flow necessary to support the brainstem in the developing brain is uncertain. Abso- lute lack of flow to the cerebrum and the brainstem would need to be demonstrated if this information is to be used to support a diagnosis of brain death. We do not know what minimum flow to the brainstem can be demonstrated by angiography. Can we be assured that cerebral angiography can visualize low flows to the brainstem of infants?18(p515)

These concerns were borne out by Ashwal and Schneider’s series of 17 newborns with clinical brain death, one of whom (a term infant) had no flow on a radionuclide study and a nearly

isoelectric EEG (only a photic response).31 A second radio- nuclide study showed resumption of flow, and the baby devel- oped spontaneous respirations and activity. That patient survived for 3 months in a vegetative state. (This case is spe- cifically mentioned in the Guidelines.2(p.e729))

But even in older children and adults, there are reports indicating that the specificity of radionuclide studies for demonstrating no blood flow throughout the brain is less than 100%. Coimbra’s adult case had nonopacification of intracra- nial arteries on a 4-vessel contrast angiogram, which has even greater specificity than radionuclide studies, yet that patient had return of weak respiratory effort and developed reflex jaw closing and lip protrusion triggered by oral hygiene, clonic seizure-like episodes responsive to diazepam and lasting more than 10 minutes if untreated, and EEG showing a low- amplitude alpha coma pattern.11(pp315-316) In a large series of radionuclide angiographies between 1984 and 1996, Flow- ers and Patel reported that, of the 10 patients who were not clinically brain dead, 5 had no visualizable arterial flow (50% specificity).iii,65 Their technique, however, was only a dynamic scan with only an anterior view; they concluded by recommending perfusion scintigraphy as a more specific method, especially for assessing posterior fossa flow. Berlit and Wetzel reported a study of 27 clinically brain dead patients, 9 of whom had both hexamethylpropyleneamine- oxime (HMPAO) perfusion scintigraphy and 4-vessel angio- graphy; 2 of the 9 showed no flow on scintigraphy but flow on panangiography.iv,32 Drake et al reported that 5 out of 32 children with suspected brain death and no flow on radionu- clide scan at 24 hours had EEG activity, indicating viable portions of cortex with low blood flow not detected by the scan.66 Grigg et al described 2 patients who met all clinical criteria for brain death short of an apnea test, who had iso- electric EEGs and no apparent blood flow on radionuclide testing, yet breathed spontaneously during the apnea test.29 (It is worth noting, however, that both of the latter 2 studies utilized nondiffusible tracers that are not taken up by viable brain cells.)

The current case is the first report of a non-neonatal child with return of a clinical brain function (spontaneous breathing) after fulfilling the current Guidelines, specifically its algorith- mic branch that relies on radionuclide scintigraphy showing no intracranial blood flow. A limitation of the case is that a non- diffusible tracer, DTPA, was used, whereas diffusible tracers may have greater specificity for no flow.61 On the other hand, the Guidelines do not make any distinction among types of tracer. Be that as it may, the boy’s medullary respiratory center was evidently still viable, receiving undetected flow, despite MRI evidence of foramen magnum herniation and despite a CPP around 0 sustained for at least 2 hours. Medullary blood flow despite herniation is not unprecedented. Korein et al reported that of 20 cases of contrast angiography (both carotids and one or both vertebral arteries) with no clinical signs of brain function and isoelectric EEG, 2 had sluggish but present posterior fossa blood flow despite tonsillar herniation.67 In such cases, the lack of brain stem function could have been

Shewmon

1113

due to ischemic penumbra rather than necrosis.68 In any event, given that a test for death is supposed to have an essentially zero false-positive rate, even a single instance of false-positive error suffices to negate the confirmatory value of radionuclide scintigraphy in diagnosing brain death and therefore also that branch of the Guidelines’ algorithm that relies on ancillary tests to establish the diagnosis.

Wijdicks went a step further, offering cogent arguments against the value of ancillary testing in general, at least in adults, on account of an unacceptable rate of both false- positive and false-negative results.38 As regards radionuclide blood flow studies specifically, he cited studies that documen- ted discrepancies between radionuclide results and clinical examination, contrast angiography, transcranial Doppler, and CT angiography, concluding:

So, what are neurologists confirming? If documentation of a loss of all neuronal function is the ultimate goal for the definition of brain death, the goal is not attainable because no confirmatory test can provide such documentation with certainty. . . . The “confirmatory” tests do not confirm anything.38(pp79,81)

I and many others would take issue with the conditional clause suggesting that “the ultimate goal for the definition of brain death” is “documentation of a loss of all neuronal function.” The term neuronal function is generally under- stood to mean metabolic or electrical activity at the neuro- nal cellular level, and that is explicitly excluded by the 1981 President’s Commission as being of any interest for the diagnosis of brain death.1(pp6,28,33,75,112) The ultimate goal is rather documentation of the “irreversible cessation of all functions of the entire brain.” But Wijdick’s main point is well taken: the “confirmatory” tests do not really confirm anything.

Young et al laid out five requirements for “truly con- firmatory” ancillary tests:

1. There should be no “false positives”, i.e. when the test con- firms “brain death” there should be none who recover or who have the potential to recover.
2. The test should be sufficient on its own to establish that brain death is or is not present.
3. The test should not be susceptible to “confounders” such as drug effects or metabolic disturbances.
4. The test should be standardized in technology, technique and classification of results.
5. The test should be available, safe and readily applied. Testing should not be restricted to only a few research centres; ideally it could be applied within any intensive care unit (ICU) and the technique should be reliable and mastered without difficulty.

Care must be taken to avoid “self-fulfilling prophesies” in devel- oping tests for brain death; some type of confirmation is desirable. Verification methods include the continued support of the patient for a suitable time after the test has been applied, with observation and outcome determination and/or a neuropathological examina- tion of the brain.34(p621) (emphases added)

From a logical perspective, if requirement 2 is really ful- filled, why not just make the ancillary test the sole diagnostic criterion and be done with it? Be that as it may, no currently available ancillary test fulfills these requirements. No study has validated any of the blood flow tests as capable of distinguish- ing low flow that is barely enough to preserve viability of some parts of the brain, from no flow in all parts of the brain—not only in newborns but also in children and adults. The present case and the several others cited above29,31,32,65-67 demonstrate that, in fact, radionuclide scanning lacks sufficient specificity for confirming death.

But if ancillary tests are not required, as the Guidelines permit, or are to be completely done away with, as Wijdicks advocates38—on the grounds that brain death is essentially a “clinical” diagnosis (meaning by means of the bedside exam- ination)—then the set of clinical criteria should satisfy require- ments 1 to 3 of Young et al no less than ancillary tests should (requirements 4 and 5 being inapplicable), as well as hold up against the caveat about “self-fulfilling prophesies” and the recommendations for verification.

The contention that there are no reports of recovery of brain function after meeting the adult or pediatric Guidelines sounds impressive at first, but in the enormous majority of cases, either organs are harvested or support is withdrawn as soon as the diagnosis is made (“self-fulfilling prophecy”).7,69 One of the more disturbing aspects of the present case is the realization that if parents had agreed to organ donation, the discovery that the boy was not dead despite fulfillment of the Guidelines would never have come to light. Who knows how many organ donors over the decades have been in a similar condition?

If there is “insufficient evidence to determine the minimally acceptable observation period to ensure that neurologic func- tions have ceased irreversibly” as well as other aspects of the adult criteria,5 all the more so with the pediatric criteria. As the Guidelines point out,

No randomized control [sic] trials examining different strategies regarding the diagnosis of brain death exist. Standard evidence- based approaches for guidelines used by many organizations attempting to link the “strength of the evidence” to the “strength of the recommendations” therefore cannot be used in this instance. There is, however, considerable experiential consensus within observational studies in the pediatric population.2(p.e723)

In the end, the Guidelines are a consensus-based set of rec- ommendations, which inherently cannot live up to the main requirement of any diagnostic criterion of death, to “eliminate errors in classifying a living individual as dead.”1(p161) And no amount of iterative tweaking, taking into account this case or any future counterexamples, will produce the requisite 100% specificity, because, even if a multicenter prospective study were to be carried out to fulfill the verification requirement of Young et al,34(p621) the study size would have to be impos- sibly large by orders of magnitude in order to keep the lower 95% confidence limit for specificity acceptably close to 100%.

1114

Journal of Child Neurology 32(14)

A Pragmatic Objection

One might object that, although this patient was not strictly brain dead at the time of fulfillment of the Guidelines, he was “as good as dead”—or at least that the outcome was the same regardless whether he was truly brain dead or not. As a corollary, so the objection goes, it would not really matter if the only false-positive errors occasioned by the Guidelines were cases of this sort (setting aside the additional emotional trauma of seeing the child breathe after being declared dead). But who knows if 100% of the false positives would fall into this extreme category? And even if they did, a major problem with this line of reasoning, from the point of view of public policy, is that “dismal neurological prognosis” or the vague concept “as good as dead” does not correspond to the statutory definition of death in any state or country. Organ retrieval on this basis would place surgeons in a leg- ally precarious position at best, and it is not what organ procurement organizations advertise to the general public or explain to families in intensive care units. Although in the end it made no practical difference for this particular patient whether he was strictly brain dead or “as good as dead,” the distinction could make a difference for other patients in a similar condition, if organs were to be harvested or if parents insisted on continued support if the patient was not truly dead.

Appendix

Check List for Documentation of Brain Death (Guidelines, Appendix 1)

Brain Death Examination for Infants and Children
Two physicians must perform independent examinations separated by specified intervals.

Conclusion

Counterexamples to the standard brain death diagnostic criteria are extremely rare, which is why this one is reportable. But the rarity is not reassuring, because of the unknowable extent to which it could be due to self-fulfilling prophecy. Even a single false-positive diagnosis of death is one too many.

Notes

i. http://www.cnn.com/2016/05/12/health/california-israel-stinson- case/index.html? sr1⁄4twCNN051416california-israel-stinson- case0212AMVODtopPhoto&linkId1⁄424502814.

ii. Strictly speaking, the Guidelines allow an ancillary test to establish brain death in the context of a missing apnea test, if the reason for the omission was a medical contraindication. At the time of Kohr- man and Spivack’s case, the 1987 Pediatric Brain Death Guidelines were in effect, which left the required number of apnea tests ambig- uous (see Guidelines, Appendix 7). Regardless of the reason for not doing a second apnea test, the case proves that the EEG cannot reliably confirm irreversible absence of brainstem functions.

iii. The presentation of data in their article is conflicting as regards this detail. In some places it says that 10 cases did not meet clinical criteria for brain death, and in another place it says 9. Five scans showed flow, resulting in a specificity of either 50% (5/10) or 56% (5/9), and both versions of specificity were stated in different places.

iv. The text of the article states that 2 of 9 studies were discordant, but the data summary table shows 1 of 9 discordant. The reason for the discrepancy is unclear.

Age of patient

Term newborn 37 weeks gestational age and up to 30 days old

31 days to 18 years old

Section 1. PREREQUISITES for brain death examination and apnea test

A. IRREVERSIBLE AND IDENTIFIABLE Cause of Coma (Please check)

Timing of first exam

c First exam may be performed 24 hours after birth OR following cardiopulmonary resuscitation or other severe brain injury

ý First exam may be performed 24 hours following cardiopulmonary resuscitation or other severe brain injury

Interexam interval

c At least 24 hours c Interval shortened because ancillary

study (section 4) is consistent with brain death

ý At least 12 hours OR c Interval shortened

because ancillary study (section 4) is consistent with brain death

ExaminationTwo

ý Yesb c No ý Yesd c No

ý Traumatic brain injury c Anoxic brain injury c Known metabolic disorder c Other (specify)

B.Correctionofcontributingfactorsthatcaninterferewiththeneurologic ExaminationOne examination

a.CoreBodyTempisover95 F(35 C) ýYesa c No

b. Systolic blood pressure or MAP in acceptable range (Systolic BP not less than 2 ý Yesc c No standard deviations below age appropriate norm) based on age

(continued)

Shewmon

1115

Appendix (continued)
c. Sedative/analgesic drug effect excluded as a contributing factor d. Metabolic intoxication excluded as a contributing factor

e. Neuromuscular blockade excluded as a contributing factor

ý Yese ý Yesg

ý Yesi

c No c No

c No

ý Yesf ý Yesh

ý Yesj

c No c No

c No

ý If ALL prerequisites are marked YES, then proceed to section 2, OR
c ______confounding variable was present. Ancillary study was therefore performed to document brain death. (Section 4).

Section 2. Physical Examination (Please check)
NOTE: SPINAL CORD REFFLEXES ARE ACCEPTABLE

a. Flaccid tone, patient unresponsive to deep painful stimuli
b. Pupils are midposition or fully dilated and light reflexes are absent

c. Corneal, cough, gag reflexes are absent
d. Sucking and rooting reflexes are absent (in neonates and infants)

e. Oculovestibular reflexes are absent
f. Spontaneous respiratory effort while on mechanical ventilation is absent

Examination One

Examination Two Date/time: 70.7 hrs

ý Yes cNo ý Yes cNo

ý Yes cNo

ý Yes cNo ý Yes cNo

Date/time: 49.2

hrs

ý Yes ý Yesk

ý Yes

ý Yes ý Yes

cNo cNo

cNo

cNo cNo

ý The [oculovestibular reflex] (specify) element of the exam could not be performed because [It was performed, but its validity was considered uncertain due to debris in the external auditory canals]. Ancillary study (EEG or radionuclide CBF) was therefore performed to document brain death. (Section 4).

Section 3. APNEA Test

NospontaneousrespiratoryeffortswereobserveddespitefinalPaCO2 60mmHgand a 20 mm Hg increase above baseline. (Examination One)

NospontaneousrespiratoryeffortswereobserveddespitefinalPaCO2 60mmHgand a 20 mm Hg increase above baseline. (Examination Two)

Examination One Date/time: _____l

PretestPaCO2:_____ Apnea duration: ___ min PosttestPaCO2:____

Examination Two Date/time: 70.7 hours

PretestPaCO2:41.9 Apnea duration: 2.5 min PosttestPaCO2:57.9

Apnea test is contraindicated or could not be performed to completion because O2 saturation plummeted. Ancillary study (EEG or radionuclide CBF) was therefore performed to document brain death. (Section 4).

Section 4. ANCILLARY testing is required when (1) any components of the examination or apnea testing cannot be completed; (2) if there is uncertainty about the results of the neurologic examination; or (3) if a medication effect may be present.

Ancillary testing can be performed to reduce the inter-examination period; however a second neurologic examination is required. Components of the neurologic examination that can be performed safely should be completed in close proximity to the ancillary test.

Date/Time: _______

ý Electroencephalogram (EEG) report documents electrocerebral silence OR ý Cerebral Blood Flow (CBF) study report documents no cerebral perfusion

Probably Formally read as ECS equivocal

43.7 hrs 54.1 hrs

ý Yes c No ý Yes c No

aTemp 1⁄4 36.5 .
bTemp 1⁄4 36.9 .
cBP 1⁄4 119/70.
dBP 1⁄4 133/81.
eFentanyl administration was: 13.7 mcg IV push at 3.9 hours and a 17-hour continuous infusion at 0.5 mcg/kg/hr from 5.52 to 22.65 hours. It had been discontinued for 26.5 hours prior to the first examination. The elimination half-life of fentanyl, as stated in Appendix 2 of the Guidelines (in the two 2011 publications2,3 but not included in the 2012 publication4), is: “5 months to 4.5 yrs: 2.4 hrs (mean); 0.5-14 yrs: 21 hrs (range, 11-36 hrs for long-term infusions).” “Long-term” was not defined. Any residual sedative effect at the time of the first examination would have been minor, would have become negligible by the second examination, and would have been rendered irrelevant by the confirmative blood flow study.

fFentanyl had been discontinued for 48 hours prior to the second exam examination. See previous footnote regarding elimination half-life.
gThere was no reason to suspect recent ingestion, as his behavior was normal up to the time of the accident, so urine toxicology screening was not clinically indicated. Routine metabolic studies were normal, except for hyperglycemia, treated with insulin.
hIdem.
iVecuronium had been discontinued for 26.6 hours prior to the first exam. The presence of a triple flexion response ruled out any residual paralytic effect. jVecuronium had been discontinued for 48.1 hours prior to the second exam. The presence of a triple flexion response ruled out any residual paralytic effect. kPupils were 4 mm, both nonreactive.
lMedically contraindicated, due to precipitous drop in heart rate during 20 seconds off the ventilator during the EEG.

1116

Journal of Child Neurology 32(14)

Acknowledgment

The author thanks Michael Nair-Collins, PhD, for assistance in the preparation of this manuscript.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

References

1. President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research. Defining Death: Medical, Legal, and Ethical Issues in the Determination of Death. Washington, DC: U.S. Government Printing Office; 1981.
2. Nakagawa TA, Ashwal S, Mathur M, et al. Guidelines for the determination of brain death in infants and children: an update of the 1987 Task Force recommendations. Pediatrics. 2011;128: e720-e740.
3. Nakagawa TA, Ashwal S, Mathur M, et al. Guidelines for the determination of brain death in infants and children: an update of the 1987 Task Force recommendations. Crit Care Med. 2011; 39:2139-2155.
4. Nakagawa TA, Ashwal S, Mathur M, Mysore M; Committee for Determination of Brain Death in Infants and Children. Guidelines for the determination of brain death in infants and children: an update of the 1987 Task Force recommendations—executive summary. Ann Neurol. 2012;71:573-585.
5. Wijdicks EFM, Varelas PN, Gronseth GS, Greer DM and Amer- ican Academy of Neurology. Evidence-based guideline update: determining brain death in adults: report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurol- ogy. 2010;74:1911-1918.
6. Ashwal S. Clinical diagnosis and confirmatory testing of brain death in children. In: Wijdicks EFM, ed. Brain Death. Philadel- phia, PA: Lippincott Williams & Wilkins; 2001:91-114.
7. Joffe AR, Kolski H, Duff J, deCaen AR. A 10-month-old infant with reversible findings of brain death. Pediatr Neurol. 2009;41: 378-382.
8. Webb AC, Samuels OB. Reversible brain death after cardiopul- monary arrest and induced hypothermia. Crit Care Med. 2011;39: 1538-1542.
9. Wijdicks EFM, Varelas PN, Gronseth GS, Greer DM. There is no reversible brain death. Crit Care Med. 2011;39:2204-2205.
10. Webb AC, Samuels OB. There is no reversible brain death [reply]. Crit Care Med. 2011;39:2206.
11. Coimbra CG. Are “brain dead” (or “brain-stem dead”) patients neurologically recoverable? In: de Mattei R, Byrne PA, eds. Finis Vitae. ‘Brain Death’ Is Not True Death. Oregon, OH: Life Guard- ian Foundation; 2009:313-378.
12. Roberts DJ, MacCulloch KA, Versnick EJ, Hall RI. Should ancil- lary brain blood flow analyses play a larger role in the neurolo- gical determination of death? Can J Anaesth. 2010;57:927-935.

13. Stecker M, Sabau D, Sullivan L, et al. American Clinical Neuro- physiology Society Guideline 6: minimum technical standards for EEG recording in suspected cerebral death. J Clin Neurophysiol. 2016;33:324-327.

14. Vardis R, Pollack MM. Increased apnea threshold in a pediatric patient with suspected brain death. Crit Care Med. 1998;26: 1917-1919.

15. Brilli RJ, Bigos D. Altered apnea threshold in a child with sus- pected brain death. J Child Neurol. 1995;10:245-246.

16. Haun SE, Tobias JD, Deshpande JK. Apnoea testing in the deter- mination of brain death: is it reliable? Clin Intensive Care. 1991; 2:182-184.

17. Okamoto K, Sugimoto T. Return of spontaneous respiration in an infant who fulfilled current criteria to determine brain death. Pediatrics. 1995;96(3 pt 1):518-520.

18. Fishman MA. Validity of brain death criteria in infants. Pedia- trics. 1995;96(3 pt 1):513-515.

19. Hansen G, Joffe AR. Confounding brain stem function during pediatric brain death determination: two case reports. J Child Neurol. 2017;32:676-679.

20. Joffe AR, Anton NR, Duff JP. The apnea test: rationale, confoun- ders, and criticism. J Child Neurol. 2010;25:1435-1443.

21. Haupt WF, Rudolf J. European brain death codes: a comparison of national guidelines. J Neurol. 1999;246:432-437.

22. Wijdicks EF. The diagnosis of brain death. N Engl J Med. 2001; 344:1215-1221.

23. Wijdicks EFM. Brain death worldwide: accepted fact but no glo- bal consensus in diagnostic criteria. Neurology. 2002;58:20-25.

24. Wahlster S, Wijdicks EF, Patel PV, et al. Brain death declaration: practices and perceptions worldwide. Neurology. 2015;84:1870-1879. 25. Deliyannakis E, Ioannou F, Davaroukas A. Brain stem death with persistence of bioelectric activity of the cerebral hemispheres.

Clin Electroencephalogr. 1975;6:75-79.
26. Ashwal S, Schneider S. Failure of electroencephalography to diag-

nose brain death in comatose children. Ann Neurol. 1979;6:512-517. 27. Pallis C. ABC of brain stem death: the arguments about the EEG.

Br Med J (Clin Res Ed). 1983;286:284-287.
28. Furgiuele TL, Frank LM, Riegle C, Wirth F, Earley LC. Prediction

of cerebral death by cranial sector scan. Crit Care Med. 1984;12:1-3. 29. Grigg MM, Kelly MA, Celesia GG, Ghobrial MW, Ross ER. Electroencephalographic activity after brain death. Arch Neurol.

1987;44:948-954.
30. Darby J, Yonas H, Brenner RP. Brainstem death with persistent

EEG activity: evaluation by xenon-enhanced computed tomogra-

phy. Crit Care Med. 1987;15:519-521.
31. Ashwal S, Schneider S. Brain death in the newborn. Pediatrics.

1989;84:429-437.
32. Berlit P, Wetzel E. HM-PAO-Hirnblutflußszintigraphie in der

Manifestationsphase des Hirntodes [HM-PAO cerebral blood flow scintigraphy in the manifestation stage of brain death]. Ner- venarzt. 1992;63:101-104.

33. Paolin A, Manuali A, Di Paola F, et al. Reliability in diagnosis of brain death. Intensive Care Med. 1995;21:657-662.

34. Young GB, Shemie SD, Doig CJ, Teitelbaum J. Brief review: the role of ancillary tests in the neurological determination of death. Can J Anesth. 2006;53:620-627.

Shewmon 1117

35. Bernat JL. A defense of the whole-brain concept of death. Hastings Cent Rep. 1998;28:14-23.
36. Bernat JL. Refinements in the definition and criterion of death. In: Youngner SJ, Arnold RM, Schapiro R, eds. The Definition of Death: Contemporary Controversies. Baltimore, MD: Johns Hopkins University Press; 1999:83-92.
37. Bernat JL. The biophilosophical basis of whole-brain death. Soc Philos Policy. 2002;19:324-342.
38. Wijdicks EF. The case against confirmatory tests for determining brain death in adults. Neurology. 2010;75:77-83.
39. Brierley JB, Graham DI, Adams JH, Simpsom JA. Neocortical death after cardiac arrest. A clinical, neurophysiological, and neu- ropathological report of two cases. Lancet. 1971;2:560-565.
40. Pollack MA, Kellaway P. Cortical death with preservation of brain stem function: correlation of clinical, electrophysiologic, and CT scan findings in 3 infants and 2 adults with prolonged survival. Trans Am Neurol Assoc. 1978;103:36-38.
41. Mizrahi EM, Pollack MA, Kellaway P. Neocortical death in infants: behavioral, neurologic, and electroencephalographic characteristics. Pediatr Neurol. 1985;1:302-305.
42. Blend MJ, Pavel DG, Hughes JR, Tan WS, Lansky LL, Toffol GJ. Normal cerebral radionuclide angiogram in a child with electro- cerebral silence. Neuropediatrics. 1986;17:168-170.
43. Kohrman MH, Spivack BS. Brain death in infants: sensitivity and specificity of current criteria. Pediatr Neurol. 1990;6:47-50.
44. Hughes JR. Limitations of the EEG in coma and brain death. Ann

    N Y Acad Sci. 1978;315:121-136.

45. Pfurtscheller G, Schwarz G, List W. Brain death and bioelectrical

    brain activity. Intensive Care Med. 1985;11:149-153.

46. Sutton LN, Bruce DA, Schut L. Hydranencephaly versus maximal hydrocephalus: an important clinical distinction. Neurosurg.

    1980;6:35-38.

47. Iinuma K, Handa I, Kojima A, Hayamizu S, Karahashi M. Hydra-

    nencephaly and maximal hydrocephalus: usefulness of electro- physiological studies for their differentiation. J Child Neurol. 1989;4:114-117.

48. Shewmon DA, Holmes GL, Byrne PA. Consciousness in congeni- tally decorticate children: “developmental vegetative state” as self- fulfilling prophecy. Dev Med Child Neurol. 1999;41:364-374.
49. Boutros AR, Henry CE. Electrocerebral silence associated with adequate spontaneous ventilation in a case of fat embolism. A clinical and medicolegal dilemma. Arch Neurol. 1982;39: 314-316.
50. Green JB, Lauber A. Return of EEG activity after electrocerebral silence: two case reports. J Neurol Neurosurg Psychiatry. 1972; 35:103-107.
51. Juguilon AC, Reilly EL. Development of EEG activity after ten days of electrocerebral inactivity: a case report in a premature

neonate-hydranencephaly or massive ventricular enlargement.

Clin Electroencephalogr. 1982;13:233-240.
52. Holzman BH, Curless RG, Sfakianakis GN, Ajmone-Marsan C,

Montes JE. Radionuclide cerebral perfusion scintigraphy in deter-

mination of brain death in children. Neurology. 1983;33:1027-1031. 53. Toffol GJ, Lansky LL, Hughes JR, et al. Pitfalls in diagnosing

brain death in infancy. J Child Neurol. 1987;2:134-138.
54. Schmitt B, Simma B, Burger R, Dumermuth G. Resuscitation after severe hypoxia in a young child: temporary isoelectric EEG and loss of BAEP components. Intensive Care Med. 1993;19:

420-422.
55. Rimmele T, Malhiere S, Ben Cheikh A, et al. The electroence-

phalogram is not an adequate test to confirm the diagnosis of brain

death [in French]. Can J Anaesth. 2007;54:652-656.
56. Bernat JL. The concept and practice of brain death. Prog Brain

Res. 2005;150:369-379.
57. Fackler JC, Rogers MC. Is brain death really cessation of all

intracranial function? J Pediatr. 1987;110:84-86.
58. Link J, Schaefer M, Lang M. Concepts and diagnosis of brain

death. Forensic Sci Int. 1994;69:195-203.
59. Berenguer CM, Davis FE, Howington JU. Brain death confirma-

tion: comparison of computed tomographic angiography with

nuclear medicine perfusion scan. J Trauma. 2010;68:553-559. 60. Kramer AH. Ancillary testing in brain death. Semin Neurol. 2015;

35:125-138.
61. Joffe AR, Lequier L, Cave D. Specificity of radionuclide brain

blood flow testing in brain death: case report and review. J Inten-

sive Care Med. 2010;25:53-64.
62. Bernat JL. The whole-brain concept of death remains optimum

public policy. J Law Med Ethics. 2006;34:35-43.
63. Bernat JL. Contemporary controversies in the definition of death.

Prog Brain Res. 2009;177:21-31.
64. Altman DI, Powers WJ, Perlman JM, Herscovitch P, Volpe SL, Volpe

JJ. Cerebral blood flow requirement for brain viability in newborn

infants is lower than in adults. Ann Neurol. 1988;24:218-226.
65. Flowers WM Jr, Patel BR. Radionuclide angiography as a con- firmatory test for brain death: a review of 229 studies in 219

patients. South Med J. 1997;90:1091-1096.
66. Drake B, Ashwal S, Schneider S. Determination of cerebral death

in the pediatric intensive care unit. Pediatrics. 1986;78:107-112. 67. Korein J, Braunstein P, George A, et al. Brain death: I. Angio- graphic correlation with the radioisotopic bolus technique for evaluation of critical deficit of cerebral blood flow. Ann Neurol.

1977;2:195-205.
68. Coimbra CG. Implications of ischemic penumbra for the diagno-

sis of brain death. Braz J Med Biol Res. 1999;32:1479-1487.
69. Nair-Collins M. Clinical and ethical perspectives on brain death.

Medicoleg Bioeth. 2015;5:69-80.