Chronic “brain death”

Meta-analysis and conceptual consequences

D. Alan Shewmon, MD

Article abstract-Objective: One rationale for equating “brain death” (BD) with death is that it reduces the body to a
mere collection of organs, as evidenced by purported imminence of asystole despite maximal therapy. To test this
hypothesis, cases of prolonged survival were collected and examined for factors influencing survival capacity. Methods:
Formal diagnosis of BD with survival of 1 week or longer. More than 12,200 sources yielded approximately 175 cases
meeting selection criteria; 56 had sufficient information for meta-analysis. Diagnosis was judged reliable if standard
criteria were described or physicians made formal declarations. Data were analyzed by means of Kaplan-Meier curves,
with treatment withdrawals as “censored” data, compared by log-rank test. Results: Survival probability over time
decreased exponentially in two phases, with initial half-life of 2 to 3 months, followed a t 1 year by slow decline to more
than 14 years. Survival capacity correlated inversely with age. Independently, primary brain pathology was associated
with longer survival than were multisystem etiologies. Initial hemodynamic instability tended to resolve gradually; some
patients were successfully discharged on ventilators to nursing facilities or even t o their homes. Conclusions: The
tendency to asystole in BD can be transient and is attributable more to systemic factors than to absence of brain function
per se. If BD is to be equated with death, it must be on some basis more plausible than loss of somatic integrative unity.
NEUROLOGY 1998;51:1538-1545

The equivalence of “brain death” (BD) with death is inexorably and imminently deteriorate to cardiovas-
one of the few bioethical issues of this decade consid- cular collapse despite the most aggressive therapy
ered relatively settled.l(P115) (Together with Veatch? I and resuscitative efforts. The BD literature, right up
prefer to place “brain death” in quotation marks on to the present, is replete with statements to this
account of its semantic ambig~ity.~ For purposes of effect,16-21
such as the following (emphases added):
this paper, the term will be taken to mean “whatever Even with extraordinary medical care, these [somatic]
most people understand by the term ‘brain death’ functions cannot be sustained indefinitely-typically,
[with whatever ambiguity and inconsistency that en- no longer than several days (President’s Commission)
tails] ,” or equivalently, “a clinical neuropathologic (p. 351.13
state fulfilling official diagnostic algorithms and le- Despite all efforts to maintain the donor’s circulation,
gally equated with death in most jurisdictions [re- irreversible cardiac arrest usually occurs within 48 to 72
gardless of the rationale for, or validity of, that hours of brain death in adults, although it may take as
equation].”) What has been settled, however, is long as 10 days in children. Indeed, general acceptance
merely statutory definition and diagnostic proto- of the concept of brain death depended on this close
C O ~ S . ’ , ~Beneath
~~ this superficial consensus there is temporal association between brain death and cardiac
tremendous confusion about the fundamental ratio- arrest (p. 816).22
nale for equating the death of one particular organ What was clearly established in the early 1980s was
with death of the entire o r g a n i ~ m .In ~ the
~ ~ United
~~-~~ that no patient in apneic coma declared brain dead ac-
States and most other countries where a quasi- cording to the very stringent criteria of the United King-
official rationale has been articulated, the rationale dom code . . . had ever failed to develop asystole within a
is that the brain is the “central integrator” or “criti- relatively short time. That fundamental insight remains
cal organ” of the body, and its destruction or irre- as valid today as it was 20 years ago-and not only in
versible nonfunction entails a loss of somatic the United Kingdom but throughout the world (preface
integrative unity, a thermodynamic “point of no re- to second edition).23
turn,” a literal “dis-integration”of the organism as a It is important t o distinguish between this line of
~hole.lJ~-~~ reasoning and a conflation of prognosis of eventual
One line of evidence usually cited is that BD bod- death with diagnosis of present death. As David
ies cannot be maintained indefinitely; rather, they Lamb eloquently explained:

See also page 1530

From the Department of Pediatrics, Division of Neurology, UCLA Medical School, Los Angeles, CA.
Received September 3, 1997.Accepted in final form July 17, 1998.
Address correspondence and reprint requests to D. Alan Shewmon, MD, Department of Pediatrics, Division of Neurology, UCLA Medical Center, MDCC
22-474,Box 951752,Los Angeles, CA 90095-1752.
1538 Copyright 0 1998 by the American Academy of Neurology
 When evidence is cited to show that, despite the most port, miscellaneous details (e.g., pregnancy outcome, trans-
aggressive support, the adult heart stops within a week fer from hospital), city, patient identification, and source(s)
of brainstem death and that of a child within two weeks, of information. If patient names were public knowledge,
one is not marshalling empirical support for a prediction they were included; otherwise, cases were identified as in
of death. What is being said is that a point has been their original reference or by code.
reached where the various subsystems lack neurological To study the extent to which associated systemic inju-
integration and their continued (artificial) functioning ries, rather than brain destruction per se, might limit sur-
only mimics integrated life. That structural disintegra- vival potential, etiologies were divided into two categories
tion follows brain death is not a contingent matter; it is depending on whether the primary insult was restricted to
a necessary consequence of the death of the critical sys- the brain or also affected multiple organs.
tem. The death of the brain is the point beyond which Data were analyzed with Kaplan-Meier survival curves.
other systems cannot survive with, or without, mechan- Terminal event (spontaneous asystole versus treatment
ical support (emphasis in original; pp. 36-37).24 withdrawal or ongoing survival) distinguished “deaths”
from “censored data” according to Kaplan-Meier methodol-
Recent literature and collective personal experi-
ogy. Survival curves of age and etiology subgroups were
ences, however, c a s t serious doubt on this long-
statistically compared using the Mantel-Haenszel (log-
standing doctrine.
rank) test. Some partitionings of data were also evaluated
by the chi-square test. There is no pretension that these
Methods. An attempt was made to compile all known cases constitute a random or representative sampling of
cases of BD with survival of 1 week or longer through the entire BD population; the statistical techniques were
personal experiences of the author and other professionals, strictly for exploratory purposes within the set of cases
collection of articles, and systematic database searches. collected.
Melvyl Medline (UCLA) was searched on the key word
“brain death” over the entire history of BD (1966 to 19971, Results. Combined sources yielded approximately 175
yielding 12,219 articles. An article was examined if either cases of BD with survival of at least 1 week (the approxi-
title or abstract suggested a possible case of prolonged mation stemming from two articles that reported survival
survival. News media were surveyed through the World durations grouped as mean and standard d e v i a t i ~ n ~ ~ , ~ ~ ) .
Wide Web and Lexis-Nexis on key words “brain death,” The data for these cases, which constitute an integral part
“pregnancy,” and “life support,” using a variety of search of this article, are available in two tables, along with refer-
engines and Lexis-Nexis libraries. ences and 270 footnotes, from the National Auxiliary Pub-
Cases were selected if BD was formally diagnosed ac- lications Service (NAPS) (see Note at end of text). The
cording t o customary criteria. In the medical articles and distribution of sources was as follows: author’s personal
personal sources the basis for diagnosis was usually thor- experience (2 cases), other professionals (6 cases), medical
oughly documented. From the news stories great care was literature (approximately 154 cases), nursing literature (2
taken to include only cases in which BD had been formally cases), bioethics literature (2 cases), court transcripts (2
declared by (presumably) competent physicians, not cases), and newspapers or news wire services (17 cases,
merely journalistic hearsay in which the reporter might about which 79 articles were obtained). The stated total of
have confused BD with vegetative state or coma. Articles approximately 175 is less than the sum of the foregoing
typically stated physicians’ names, specialties, and institu- numbers because of overlap among sources.
tions. Often the diagnosis was confirmed by multiple phy- For purposes of meta-analysis, three cases were ex-
sicians, including at least one neurologist or neurosurgeon. cluded due t o extreme paucity of information and one due
Many articles mentioned results of confirmatory tests or t o lack of certainty of diagnosis (Rosemarie Maniscalco,
described circumstances comprehensible only if the cases whose case occurred before diagnostic standardization by
truly involved formal declarations of BD (e.g., that organ the President’s Commission13and in whom some residual
donation was proposed to the family). EEG activity of uncertain significance reportedly re-
Despite the internationality of cases, diagnostic criteria turned). An additional approximately 115 cases with reli-
were fairly uniform. In most, a US-style “whole-brain’’ able diagnoses, though of considerable interest for survival
standard was employed. A few cases involving the British duration, were unsuitable for meta-analysis because of in-
“brainstem” standard were nevertheless included because sufficient individual data (most reported as grouped statis-
the intrinsic survival capacity of a BD body should not be tics). One table on file with NAPS itemizes these
affected by possible residual cortical function in the con- approximately 119 segregated cases, whereas the other
text of a dead brainstem. Moreover, the purported immi- contains the remaining 56 cases subjected to meta-analysis
nence of asystole is cited as much by British proponents of (see Note at end of text). The case of Tracy Bucher, known
“brainstem death” as by US proponents of “whole-brain through the news media, and that of “Sheila” (a pseud-
death.” onym) in a nursing article had so many features in com-
As much information as possible was collected about mon that they were assumed to be the same patient.
each case and compiled into an Excel (Microsoft; Seattle, Although itemized separately in the NAPS table, they
WA) spreadsheet for statistical analysis and graphical rep- were analyzed as a single case.
resentation. The following data categories were of particu- A few cases featured diagnostic controversy at the time
lar interest: survival duration, age at BD, etiology, clinical of occurrence but were nevertheless included for meta-
signs essential to the diagnosis, confirmatory tests, associ- analysis because, in the author’s opinion, the evidence
ated clinical findings (e.g., diabetes insipidus, nonpurpose- strongly favored BD. These controversies arose many days
ful movements), terminal event (spontaneous asystole into the clinical course and revolved around the interpreta-
versus treatment withdrawal), reason for prolonged sup- tion of return of muscle tone or certain spontaneous or
December 1998 NEUROLOGY 5 1 1539
 Total Group
1

0.9
0.8

-- 0.7

06
‘5 o’5
3
n ’
0.4
n.
0.3

0.2

0.1
1 dY 1wk 1mO 1 yr lOyr 0
Survival Duration 1 dY 1w k 1mO 1 yr lOyr
Survival Duration
Figure 1. Survival curve for the 56 cases subjected to
meta-analysis (“Total”) subdivided according to terminal Figure 2. Kaplan-Meier survival curve corresponding to
event-spontaneous cardiac arrest or treatment figure 1 (with the 19 treatment withdrawals and 1 still
withdrawal. surviving patient analyzed as “censored”data) represent-
ing probability of survival as a function of duration
of brain death.
reflex movements (Camp, Rader, “TK”), including inef-
fectual respiratory-like movements on rare occasions been shifted up and to the right by unknown extents. This
(“Baby A,” “TK”). Both spontaneous trunk and limb uncertainty can be taken into account statistically by re-
movement^^^-^^ and respiratory-like movement^^^.^^ have garding the treatment withdrawals as “censored” data in
been described as spinal cord-mediated phenomena in Kaplan-Meier methodology, analogous to patients lost to
well-documented cases of BD and are explicitly compatible follow-up or still alive at data collection in a typical sur-
with the diagnosis according to the American Academy of vival study. Accordingly, the three curves of figure 1 trans-
Neurology.* Therefore, the diagnostic disputes engendered formed into the single Kaplan-Meier curve of figure 2, in
by such movements were not considered in themselves suf- which the 36 spontaneous arrests constitute the vertical
ficient grounds to reject a case from the current study, steps and the 20 censored cases (19 withdrawals and 1still
especially because criteria for BD were incontrovertibly surviving) modify the probability level at each step. The
fulfilled early in the patients’ courses (when organs could resulting curve is a better indicator of intrinsic survival
have been legally removed or life-support terminated and capacity than those of figure 1. Note that it has shifted
the later controversies never have arisen). markedly to the right, with the first phase (still nearly
Survival durations. Of the approximate total of 175 linear on the semilog plot, but with shallower slope corre-
BD patients surviving a t least 1 week, approximately 80 sponding to a half-life of 2 to 3 months) extending as long
survived at least 2 weeks, approximately 44 a t least 4 as 1 to 2 years, before the second, more gradual, phase
weeks, approximately 20 at least 2 months, and 7 at least sets in.
6 months. Even excluding the 14 cases known only Age efect. Figure 3 shows a scatter plot of age a t BD
through news media or mentioned merely in passing in versus log(duration) for all 56 cases. The longest survivors
medical articles, there remained approximately 161 docu- (2.7, 5.1, and 14.5 years) were all young children, two of
mented survivals of a t least 1 week, approximately 67 at whom were newborns, and all nine survivors beyond 4
least 2 weeks, approximately 32 at least 4 weeks, approxi- months were younger than 18 years. Conversely, all 17
mately 15 a t least 2 months, and 7 at least 6 months. patients over age 30 survived less than 2% months. The
The 56 cases with sufficient individual information for inverse relationship between age and maximum survival
meta-analysis are shown in figure 1 as actuarial survival duration was nearly linear on the semilog plot.
curves for the whole group and for the two subgroups dis- Age also influenced the proportion of treatment withdraw-
tinguished by terminal event: those supported indefinitely als, most of which involved adult patients. An age-partition
until spontaneous cardiac arrest (36 cases plus 1 still sur-
viving) and those from whom treatment was withdrawn
70
(19 cases). The longest survivals were so great (up to 14.5 1
I . I I I
years) that a logarithmic scale was required to fit every- 50

thing meaningfully on a single chart. The drop-off of all 50

three curves was biphasic, with an initially rapid exponen-
tial decay followed by a very slow decline, the transition g 40

$33
occurring around 4 to 6 months. 8
Most treatment withdrawals (15/19; 79%)occurred after 220
4 weeks, whereas spontaneous arrests were widely distrib- 10
uted across survival durations, with slightly more than 0
one-half (21/37; 57%) before 4 weeks. Chi-square testing 1
0
revealed a window of statistical significance for placement
Survival Duration
of the survival-duration partition between 21.5 and 38.5
days, with maximum significance at 28.5 days ( p = 0.006). Figure 3. Scatter plot of age at brain death versus log-
If support had hypothetically been continued in the with- (survival duration) shows inverse relationship between
drawal subgroup, the overall survival curve would have maximum survival and age.
1540 NEUROLOGY 51 December 1998
 Age Effect
I
I
I
1
ETIOLOGY
0.9 100 yr
Primary Brain
0.8
-- 0.7
40 44%
Multisystem
x ?
$ 0.6 31
o x
8
8
6 0.5
CP x"*4 x
oo
g
L1 0.4 10 yr 0 0 r X
n
0.3
L
0.2- 1
0.1 .-
Age > 35 Years
I
L-I - . !
01 : ::::! : : ! : ::::::Id : :::::::! : '
1 dY 1w k Im, 1 yr 10 yr
Survival Duration

Figure 4. Kaplan-Meier curves show the effect of age at

brain death on survival capacity: patients at or below
age 35 years survived significantly longer than did those
over 35.

anywhere between 13 and 22 years yielded a statistically

significant chi-square test, with maximal significance at age
14; older than this, 47%of cases (16/34) ended by treatment
withdrawal compared with only 14% (3/22) for age 14 or
younger ( p = 0.01).
Upon defining treatment withdrawals as censored data,
1 dY lwk Irm 1 Yr 10 yr
Kaplan-Meier curves for "young" and "old" subgroups dif-
Survival Duration
fered significantly for age-partitions placed anywhere be-
tween 27 and 57 years. The higher the partition, the more Figure 5. Scatter plot of log(age) versus log(duration),
divergent the survival curves (that of the older subgroup broken down according to etiology category, illustrating
shifting more to the left). Figure 4 exemplifies this with a the interaction between age and etiology of brain death as
partition a t 35 years, where statistical significance was co-determinants of survival capacity. Overall, primary
greatest ( p = 0.00002). brain pathology was associated with longer survivals and
Etiology. The distribution of etiology category was as multisystem insult with shorter, but at the extremes of age,
follows: primary brain pathology (241, diffuse systemic in- the effect of etiology was overshadowed by that of age.
sult (241, and uncertain (8). Chi-square testing revealed a
region of statistical near-significance ( p = 0.07) for
survival-duration partitions placed between 40.5 and 48.5 The greatest significance was for an age group between a
days, below which the majority of cases with known etiol- lower bound of 5 to 9 months and a n upper bound of 43 to
ogy category (17/28; 61%) had multisystem insult and 45 years ( p = 0.005),as illustrated in figure 6.
above which the majority (12/18; 67%)had primary brain Within the category of primary brain pathology, the
pathology. This etiology effect was not accounted for by age independent effect of age was even more powerful than it
as a proxy variable because a separate plot of etiology was across combined etiologies. Comparisons of survivals
versus age revealed no relationship between the two. above and below an age-partition placed anywhere be-
Nevertheless, age and etiology interacted as determi- tween 22 and 49 years were statistically significant (maxi-
nants of survival probability. Figure 5 shows a scatter plot mally so between 34 and 45 years, p = 0.0000003), with
of log(age) versus log(duration) according to etiology cate-
gory. The association of primary brain pathology with
6 mo < Age <= 44 yr
longer survivals and diffuse pathology with shorter surviv-
als is evident; but the two extremes of age clearly consti-
tute notable exceptions to this general trend. The two p = 0.00519
newborns (lower right corner) had very long survivals de-
= 0.7
spite etiologies of severe hypoxia-ischemia. By contrast, .$ 06
older adults had shorter survivals regardless of etiology. E
.? 0.5
z .
'<
Because the age effect was statistically overpowering a t 0.4- 1 L, Etiology=
both ends of the age spectrum, the contribution of etiology Primaly Brain
0.3-
...-.-
was best appreciated in the large age group between these 0.2 .I I II
I

two extremes. Exploratory analysis was performed by sys-

tematically varying its upper and lower bounds and com- 04 : : : : : ! : : ! : : :::::d
! : ::::::! : I
paring the Kaplan-Meier curves for the two known etiology 1 dY 1wk Im, 1 Yr 10yr
Survival Duration
categories within it. There was a broad region approxi-
mately defined by a lower age limit between 1 month and 2 Figure 6. Etiology effect abstracted from age effect by ex-
years and an upper age limit anywhere from 13 to 48 years cluding extremes of age. Kaplan-Meier curves show that
within which primary brain pathology was significantly primary brain pathology is significantly associated with
associated with longer survival than multisystem insult. longer survival capacity than multisystem insult.
December 1998 NEUROLOGY 51 1641
the younger subgroup manifesting considerably longer sur- changes occur prior to 48 and even 24 hours after brain
vivals than the older subgroup. This striking effect re- death (p. 27).12
mained even after excluding extremes of age. Parallel
comparisons within the multisystem category could not be
That observation carries little weight, however,
meaningfully accomplished due to small numbers and ex- given the lack of systematic attempt to maintain BD
cessive inhomogeneity of data. patients aggressively to determine survival capacity.
The same disclaimer applies to the landmark review
article that same year by one section of
Discussion. Contrary t o popular belief, there are which equated the very essence of BD with “inevita-
many well-documented BD cases with survival be- ble bodily death.”
yond the “few days” typically cited as maximum pos- Another oft-cited study involved 609 BD cases
sible. An exhaustive search yielded approximately from three neurosurgical units during the 1960s and
175 cases surviving 1week or more. ~ ~ ~ it is difficult to draw conclusions re-
1 9 7 0 Again,
Validity of cases. Naturally, the amount and garding intrinsic survival capacity from this and
quality of information vaned tremendously, inviting from similar but smaller studiesZ3(p 30) because pa-
the criticism of possible misdiagnoses. But even in tients who did not succumb quickly to asystole (near-
cases with least information, formal diagnoses were ly one-half of the 609) were typically disconnected
unquestionably rendered by presumably competent from support. Unfortunately for BD research, ethical
physicians, usually including at least one neurologist patient management is incompatible with optimal
or neurosurgeon. If patients were “brain dead” scientific methodology. Studies permitting organ do-
enough to qualify as organ donors, they were surely nation are particularly unhelpful because the best
“brain dead” enough to qualify for this study. To donor candidates have the most intact organs and
dismiss the cases as presumptive misdiagnoses therefore also the greatest survival potential, which
would imply that organ donors are also often misdi- never becomes manifest; by contrast, patients with
agnosed and that BD declarations are inherently un- very unstable hemodynamics or multisystem failure
reliable. Even excluding news stories, many striking are typically rejected as donors, thereby biasing out-
examples remain of unequivocal BD confirmed by comes toward early asystole.
multiple clinical examinations, EEGs, intracranial Furthermore, neuro-intensive care has improved
blood flow, and necropsy findings. substantially since those pioneering studies. Some
Undoubtedly, more cases of prolonged survival aspects then would be considered substandard to-
have occurred than have been reported, and many day2fl,35;therefore, the apparent imminence of asys-
more potential cases have never been manifest be- tole is no evidence for limited intrinsic survival
cause BD is nearly always a self-fulfilling prophecy capacity.
of somatic demise through organ harvesting or dis- Given that BD pregnant women have been main-
continuation of upp port.^^,^^ (Thus, the small propor- tained for months and that Japanese teams have
tion of prolonged survivals among all BD cases in no supported BD patients extensively as far back as
way diminishes their conceptual importance. The 198436and virtually indefinitely with hormonal ther-
relevant denominator-the number supported maxi- it is difficult to interpret the dismal sur-
apy,25,26,37,38
mally until asystole-is unknowable but surely also vival data from a recent Taiwanese prospective
small; therefore, the meaningful ratio is not nearly study of “brain-stem dead” patients given “full car-
so tiny as it might initially seem.) diorespiratory ~ u p p o r t . ” 3Perhaps
~ more inter-
Enough information was available on 56 cases for institutional variation in “fullness of support” exists
meta-analysis of factors affecting survival capacity. than is generally recognized.
Although detailed inferences must be viewed cau- Finally, one cannot help wondering to what extent
tiously, the general conclusions are robust and ex- philosophical bias and sheer inertia of tradition may
tremely relevant to whether BD represents loss of have contributed t o perpetuating the anachronism
somatic integrative unity. despite increasingly abundant published counterevi-
Durations of suruiual. Of the meta-analyzed dence.
cases, one-half (28/56) survived more than 1 month,
Terminal event. The data collected here actually
nearly one-third (17/56) more than 2 months, seven
underestimate BD survival potential because in one-
(13%)more than 6 months, and four (7%) more than
third of cases support was withdrawn. Approxi-
1year, the record being 14% years (and still going).
mately 4 weeks into BD there was a statistically
If many of these cases have been in the medical
significant transition in the proportion of treatment
literature for some years, how did the “few days at
withdrawal versus spontaneous asystole, with 79%
most” dictum ever become so firmly entrenched? Fre-
quently cited is a 1978 multidisciplinary conference, of withdrawals after, and 57% of spontaneous arrests
in which: before, that time.
The most straightforward reason could be called
No investigator contributing to this volume has pre- “somatic plasticity.” The acute loss of all brain-based
sented evidence that irreversible cardiac arrest may be somatic regulation predisposes to cardiovascular col-
postponed more than a week (exclusiveof that in infants lapse. But those who survive gradually stabilize: ho-
and children), and most often these final irreversible meostasis adjusts, hemodynamic status improves,
1542 NEUROLOGY 51 December 1998
enteral nutrition can be resumed, and overall man- (Chamberlain, “Baby A ) or home (Hamilton, the
agement simplifies. This may be largely attributable case of Pinkus, “Baby Z”) similarly exemplify how
to recovery from spinal shock, with return of spinally chronic survival in BD does not necessarily require
mediated autonomic tone and reflexes. Such ten- “heroic,” “aggressive,” or “sophisticated technology.
dency to stabilization seems strong evidence for inte- Several Japanese studies teach a similar lesson.
grative unity. By merely adding vasopressin to epinephrine, mean
There also may be two subpopulations of BD pa- survival times in BD increased to 23 days.38With
tients: those absolutely unstable (possibly although pressor rather than antidiuretic doses of vasopres-
not necessarily because of lack of integrative unity) sin, stable hemodynamics were maintainable in all
and those relatively stable (implying some minimal patients seemingly indefinitely, and the epinephrine
degree of integrative unity). Supporting evidence, in- requirement gradually d e c r e a ~ e d .Similar
~ ~ , ~ ~results
ferred from the relationship between etiology and have been obtained with cortisol and triiodothyro-
survival duration, will be considered in the following. nine.26Because these studies were prospective, pro-
Perhaps the circumstances of treatment with- longed survivability seems representative of BD
drawal in many cases (especially cesarean delivery of patients in general, not merely rare anecdotal excep-
a fetus brought to viability) merely happened to en- tions. For such simple treatments to permit virtually
tail a several week latency. indefinite survival, the underlying somatic substrate
Regardless of the interpretation, the main point must be considerably integrated already.
remains incontrovertible: BD does not necessarily Even in the acute phase, the effort required to
lead to imminent asystole. At least some bodies with sustain most BD patients is not particularly extraor-
dead brains have survived chronically, and many dinary for contemporary ICU standards. That many
others must have an unrealized potential to do so. actually need much less sophisticated management
Age factor. Age and survival capacity were in- than many other ICU patients who are nevertheless
versely related. Adults treated indefinitely all suc- quite alive argues strongly that the former possess
cumbed to spontaneous arrest within 4 months. By integrative unity t o at least the same degree as the
contrast, children seemed capable of surviving virtu- latter.
ally indefinitely. The three most spectacular survi- Such relative simplicity of treatment contrasts
vors-with durations of more than 2 years-were all markedly with the technologic tour de force typically
young children, two being newborns. This age effect described with pregnant BD women. Plausible expla-
is not surprising. In general, children have more robust nations for the discrepancy are differences in clinical
health than do the elderly, and if the concept of somatic stage (before versus after recovery from spinal
plasticity is valid, children must have more of it than shock) and therapeutic goal. The treatment regimen
adults, just as they have more neuroplasticity. for a pregnant woman is not merely the minimum to
Complexity of care required. Seven very unusual sustain her own body, as in other BD cases (often
cases prove that complex technology and extraordi- with minimal enthusiasm of the health care team);
nary clinical effort are not always necessary for pro- rather, it is directed toward maintaining an optimal
longed survival. physiologic environment for the developing fetus
Of the two cases known personally to the author, (and with great enthusiasm). Thus, complexity of
“BES” was an almost-14-year-old head-trauma vic- management in the pregnancy cases does not indi-
tim who, after several weeks in an intensive care cate lack of unity in BD bodies in general.
unit (ICU), was transferred at the parents’ request t o Etiology factor. If some BD patients can survive
a skilled nursing facility. There he received nothing apparently indefinitely with relatively simple inter-
more than mechanical ventilation, desmopressin ventions, why do others deteriorate quickly to asys-
acetate, parenteral fluids, and basic nursing care. tole despite aggressive therapy? The heterogeneity
Hardly any laboratory tests were obtained. Survival suggests a difference in somatic substrate, contra-
was cut short at 65 days by untreated sepsis. dicting the brain-as-somatic-integrator hypothesis,
The other (‘‘TK”) is now an 18%-year-old boy who because the difference between two subgroups (im-
contracted Haemophilus influenzae meningitis at age minence versus nonimminence of asystole) cannot
4. Cerebral edema was so extreme that the cranial derive from what is common to both (lack of brain
sutures split. Multiple EEGs have been isoelectric, function).
and no spontaneous respirations or brainstem re- Apart from age, one possible somatic factor is di-
flexes have been observed over the past 14% years. rect multisystem injury. Listings of etiologies in the
Multimodality evoked potentials revealed no intra- BD literature often obscure this. For example, “head
cranial peaks, magnetic resonance angiography dis- trauma” is a commonly stated category, but motor
closed no intracranial blood flow, and neuroimaging vehicle accidents or severe falls are much more likely
showed the entire cranial cavity t o be filled with to damage internal viscera than are gunshots to the
disorganized membranes, proteinaceous fluids, and head. Similarly, one should not attribute to “loss of
ghost-like outlines of the former brain. He is fed by neural integration” a downward hemodynamic spiral
gastrostomy, and for the last 6 years has been thriv- actually caused by cardiac contusion or hypovolemic
ing sui generis on a ventilator at home. shock.
Five other cases transferred to nursing facilities For current purposes, therefore, etiologies were
December 1998 NEUROLOGY 51 1643
classified as either “primary brain” or “multisystem.” Conclusion. The phenomenon of chronic BD im-
Longer survivals (more than 6 weeks) were associ- plies that the body’s integrative unity derives from
ated with primary brain pathology, and shorter sur- mutual interaction among its parts, not from a top-
vivals with multisystem insults. Excluding age down imposition of one “critical organ” upon an oth-
extremes, etiology and age were independent and erwise mere bag of organs and tissues. If BD is to be
additive determinants of survival capacity. (At ex- equated with human death, therefore, it must be on
tremes the age effect predominated. Thus, babies “A” some basis more plausible than that the body is
and “ Z had very long survivals despite an etiology of dead. Whether other rationales, such as loss of “per-
perinatal asphyxia. Conversely, older adults had sonhood” from a biologically live body, might be con-
shorter survivals regardless of etiology.) ceptually more viable or desirable for societal
A second possible somatic factor is systemic pa- endorsement is beyond the scope of this physiologic
thology secondarily induced by sympathetic storm inquiry.
from the process of brain herniation before BD and
not attributable to mere absence of brain function Acknowledgment
afterward. Complications include transient severe The author thanks Ronald Cranford, MD, Rosa Lynn Pinkus,
hypertension, neurogenic pulmonary edema (mani- PhD, Arthur Allen, MD, and Tsu-pei Hung, MD, for sharing infor-
mation about their cases. He also thanks Charles Cannon, PhD,
fested by difficulty maintaining oxygenation), and and Frank Lopez for assistance in the Lexis-Nexis retrieval of
subendocardial infarctions (manifested by refractory news media references; Donald Guthrie, PhD, for statistical ad-
vice;
hypotension, cardiac arrhythmias, and a r r e ~ t ) . ~ O * ~ l and Mane1 Baucells for assistance in the advanced use of
Excel.
These are common with serious neurologic injuries
in general, so they are hardly surprising in cases Note. Readers can obtain 25 pages of supplementary material
culminating in BD. Neurogenic pulmonary edema from the National Auxiliary Publications Service, 248 Hempstead
Turnpike, West Hempstead, NY 11552. This is not a multiarticle
and subendocardial microinfarcts have been de- document. Request document no. 05467. Remit with your order,
scribed in experimental BD,42,43human BD autop- not under separate cover, in US funds only, $15.00 for photocopies
, ~ ~ICU management of organ donor^.^^,^^
~ i e s and or $5.00 for microfiche. Outside the United States and Canada,
Additionally, disseminated intravascular coagulation add postage of $4.50 for the first 20 pages and $1.00 for each 10
pages of material thereafter, or $1.75 for the first microfiche and
occurs commonly with severe head injury and in 25 $1.00 for each fiche thereafter. There is a $25.00 invoicing charge
to 65%of BD organ donor^.^^^^^ Gastrointestinal hem- on all orders filled before payment.
orrhage from neurologically induced stress ulcers
can also complicate management. References
The rapid, inexorable deterioration in a minority 1. Bernat JL. Ethical issues in neurology. Boston: Butterworth-
of BD patients (only approximately 10% of prospec- Heinemann, 1994:113-143.
2. Veatch RM. The impending collapse of the whole-brain defini-
tive organ donor^^^,^^) is therefore less attributable t o tion of death [published erratum in Hastings Cent Rep 1993;
absence of brain function per se than it is to su- 23(6):41.Hastings Cent Rep 1993;23(4):18-24.
pracritical damage to multiple organs, especially the 3. Shewmon DA. ‘Brain death’: a valid theme with invalid varia-
tions, blurred by semantic ambiguity. In: White RJ, Angst-
heart. Indeed, transplant specialists ironically con- wurm H, Carrasco de Paula I, eds. Working Group on the
sider cardiovascular stability a relative requirement Determination of Brain Death and its Relationship to Human
for heart d o n a t i ~ n .Conversely,
~~.~~ BD patients with Death. 10-14 December, 1989. (Scripta Varia 83). Vatican
City: Pontifical Academy of Sciences, 1992:23-51.
least hemodynamic stability have the least usable 4. American Academy of Neurology Quality Standards Subcom-
hearts due to intrinsic cardiac pathology. mittee. Practice parameters for determining brain death in
These considerations are elegantly summarized in adults (summary statement). Neurology 1995;45:1012-1014,
a book chapter entitled “Multisystem sequelae of se- 5. Wijdicks EF. Determining brain death in adults. Neurology
1995~45:1003-1011.
vere head injury” (significantly not about BD): 6. Halevy A, Brody B. Brain death: reconciling definitions, crite-
Head i n j u r y c a n n o t be considered an isolated, single- ria, and tests. Ann Intern Med 1993;119:519-525.
7. Taylor RM. Reexamining the definition and criteria of death.
system injury. Head-injured patients frequently sustain Semin Neurol 1997;17:265-270.
o t h e r o r g a n trauma. E q u a l l y important are the multi- 8. Tomlinson T. Misunderstanding death on a respirator. Bioeth-
system effects of trauma t o a n y p a r t of the body and the ics 1990;4:253-264.
unique effects of head trauma o n the hypothalamic- 9. Truog RD. Is it time to abandon brain death? Hastings Cent
p i t u i t a r y axis, which influences s u c h diverse pansys- Rep 1997;27(1):29-37.
10. Youngner SJ. Defining death: a superficial and fragile consen-
temic functions as blood coagulation, p u l m o n a r y venular sus. Arch Neurol 1992;49:570-572.
tone, gastric acid secretion, renal water conservation, 11. Conference of Medical Royal Colleges and their Faculties in
.
and glucose metabolism . . quality functional survival the United Kingdom. Diagnosis of death. Lancet 1979;1:261-
262.
obviously d e p e n d s o n a t t e n t i o n to these multisystem de-
12. Korein J. The problem of brain death: development and his-
rangements (p. 702).45 tory. Ann NY Acad Sci 1978;315:19-38.
13. President’s Commission for the Study of Ethical Problems in
Ironically, in a chapter on head injury these “mul- Medicine and Biomedical and Behavioral Research. Defining
tisystem derangements” are interpreted as therapeu- death: medical, legal, and ethical issues in the determination
tic challenges to keep a critically injured patient of death. Washington, DC: US Government Printing Office,
alive, whereas in a typical chapter on BD the same 1981.
14. Swedish Committee on Defining Death. The concept of death:
derangements are cited as evidence that the patient summary. Stockholm: Swedish Ministry of Health and Social
had already died. Affairs. 1984.
1544 NEUROLOGY 5 1 December 1998
15. White RJ,Angstwurm H, Carrasco de Paula I. Final consider- diagnosis of brain death: clinical and physiological observa-
ations formulated by the scientific participants. In: White RJ, tions. J Neurosurg 1981;55:942-946.
Angstwurm € I,
Carrasco de Paula I, eds. Working Group on 31. Turnbull J, Rutledge F. Spontaneous respiratory movements
the Determination of Brain Death and its Relationship t o with clinical brain death. Neurology 1985;35:1260. Letter.
Human Death. December 10-14, 1989. (Scripta Varia 83). 32. McCullagh P. Brain dead, brain absent, brain donors: human
Vatican City: Pontifical Academy of Sciences, 1992:81-82. subjects or human objects? Chichester: John Wiley and Sons,
16. Jennett B, Hessett C. Brain death in Britain as reflected in 1993:38.
renal donors. BMJ 1981;283:359-362. 33. Black PMcL. Brain death (first of two parts). N Engl J Med
17. Bernat JL. The definition, criterion, and statute of death. 1978;299:338-344.
Semin Neurol 1984;4:45-51. 34. Jennett B, Gleave J , Wilson P. Brain death in three neurosur-
18. Field DR, Gates EA, Creasy RK, Jonsen AR, Laros RK J r . gical units. Br Med J 1981;282:533-539.
Maternal brain death during pregnancy: medical and ethical 35. Darby JM, Stein K, Grenvik A, Stuart SA. Approach to man-
issues. JAMA 1988;260:816-822. agement of the heartbeating 'brain dead' organ donor. JAMA
19. Robertson KM, Cook DR. Perioperative management of the 1989;261:2222-2228,
multiorgan donor. Anesth Analg 1990;70:546-556. 36. Takeuchi K, Takeshita H, Takakura K, et al. Evolution of
20. Guerriero WG. Organ transplantation. In: Narayan RK, Wil- criteria for determination of brain death in Japan. Acta Neu-
berger J E J r , Povlishock JT, eds. Neurotrauma. New York: rochir (Wien) 1987;87:93-98.
McGraw-Hill, 1996:835-840. 37. Kinoshita Y, Yahata K, Yoshioka T, Onishi S, Sugimoto T.
21. Lew TWK, Grenvik A. Brain death, vegetative state, donor Long-term renal preservation after brain death maintained
management, and cessation of therapy. In: Albin MS, ed. with vasopressin and epinephrine. Transpl Int 1990;3:15-18.
Textbook of neuroanesthesia with neurosurgical and neuro- 38. Yoshioka T, Sugimoto H, Uenishi M, et al. Prolonged hemody-
science perspectives. New York: McGraw-Hill, 1997:1361- namic maintenance by the combined administration of vaso-
1381. pressin and epinephrine in brain death: a clinical study.
22. Soifer BE, Gelb AW. The multiple organ donor: identification Neurosurgery 1986;18565-567.
and management. Ann Intern Med 1989;110:814-823. 39. Hung TP, Chen ST. Prognosis of deeply comatose patients on
23. Pallis C, Harley DH. Al3C of brainstem death. London: BMJ ventilators. J Neurol Neurosurg Psychiatry 1995;58:75-80.
Publishing Group, 1996. 40. Samuels MA. Cardiopulmonary aspects of acute neurologic
24. Lamb D. Death, brain death and ethics. Albany, NY: State diseases. In: Ropper AH, ed. Neurological and neurosurgical
University of New York Press, 1985. intensive care. 3rd ed. New York: Raven Press, 1993:103-119.
25. Iwai A, Sakano T, Uenishi M, Sugimoto H, Yoshioka T, Sugi- 41. Yoshida K-I, Ogura Y, Wakasugi C. Myocardial lesions in-
mot0 T. Effects of vasopressin and catecholamines on the duced after trauma and treatment. Forensic Sci Int 1992;54:
maintenance of circulatory stability in brain-dead patients. 181-189.
Transplantation 1989;48:613-617. 42. Novitzky D, Wicomb WN, Rose AG, Cooper DK, Reichart B.
26. Taniguchi S, Kitamura S, Kawachi K, Doi Y, Aoyama N. Ef- Pathophysiology of pulmonary edema following experimental
fects of hormonal supplements on the maintenance of cardiac brain death in the chacma baboon. Ann Thorac Surg 1987;43:
function in potential donor patients after cerebral death. Eur 288-294.
J Cardiothorac Surg 1992;6:96-101; discussion 102. 43. Novitzky D, Rose AG, Cooper DK. Injury of myocardial con-
27. Heytens L, Verlooy J , Gheuens J, Bossaert L. Lazarus sign duction tissue and coronary artery smooth muscle following
and extensor posturing in a brain-dead patient: case report. brain death in the baboon. Transplantation 1988;45:964-966.
J Neurosurg 1989;71:449-451. 44. Antonini C, Alleva S, Campailla MT, et al. Morte cerebrale e
28. Ropper AH. Unusual spontaneous movements in brain-dead soprawivenza fetale prolungata [Brain death and prolonged
patients. Neurology 1984;34:1089-1092. fetal survival]. Minerva Anestesiol 1992;58:1247-1252.
29. Turmel A, Roux A, Bojanowski MW. Spinal man after declara- 45. Matjasko MJ. Multisystem sequelae of severe head injury. In:
tion of brain death. Neurosurgery 1991;28:298-302. Cottrell JE, Smith DS, eds. Anesthesia and neurosurgery. 3rd
30. Ropper AH, Kennedy SK, Russell L. Apnea testing in the ed. St. Louis: Mosby, 1994:685-712.

December 1998 NEUROLOGY 51 1545

 Chronic "brain death": Meta-analysis and conceptual consequences
D. Alan Shewmon
Neurology 1998;51;1538-1545
DOI 10.1212/WNL.51.6.1538

This information is current as of December 1, 1998

Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/51/6/1538.full.html

References This article cites 31 articles, 7 of which you can access for free at:
http://www.neurology.org/content/51/6/1538.full.html##ref-list-1
Citations This article has been cited by 35 HighWire-hosted articles:
http://www.neurology.org/content/51/6/1538.full.html##otherarticles
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright . All rights reserved. Print ISSN: 0028-3878.
Online ISSN: 1526-632X.