Taking the pulse of brain death: A meta‐analysis of the natural history of brain death with somatic support

Abstract

Background and purpose

The conceptualization of brain death (BD) was pivotal in the shaping of judicial and medical practices. Nonetheless, media reports of alleged recovery from BD reinforced the criticism that this construct is a self‐fulfilling prophecy (by treatment withdrawal or organ donation). We meta‐analyzed the natural history of BD when somatic support (SS) is maintained.

Methods

Publications on BD were eligible if the following were reported: aggregated data on its natural history with SS; and patient‐level data that allowed censoring at the time of treatment withdrawal or organ donation. Endpoints were as follows: rate of somatic expiration after BD with SS; BD misdiagnosis, including “functionally brain‐dead” patients (FBD; i.e. after the pronouncement of brain‐death, ≥1 findings were incongruent with guidelines for its diagnosis, albeit the lethal prognosis was not altered); and length and predictors of somatic survival.

Results

Forty‐seven articles were selected (1610 patients, years: 1969–2021). In BD patients with SS, median age was 32.9 years (range = newborn–85 years). Somatic expiration followed BD in 99.9% (95% confidence interval = 89.8–100). Mean somatic survival was 8.0 days (range = 1.6 h–19.5 years). Only age at BD diagnosis was an independent predictor of somatic survival length (coefficient = −11.8, SE = 4, p < 0.01). Nine BD misdiagnoses were detected; eight were FBD, and one newborn fully recovered. No patient ever recovered from chronic BD (≥1 week somatic survival).

Conclusions

BD diagnosis is reliable. Diagnostic criteria should be fine‐tuned to avoid the small incidence of misdiagnosis, which nonetheless does not alter the prognosis of FBD patients. Age at BD diagnosis is inversely proportional to somatic survival.

Scatterplot depicting the negative linear relationship between age at brain death diagnosis (x‐axis) and somatic survival length (y‐axis). FBD, functionally brain dead; NNFR, normal neurologic function recovered. The NNFR patient was censored at the time of latest follow‐up (1 year). For every 1‐year increase in age, there was an almost 12‐year decrease in somatic survival length. Patients with a somatic survival ≥ 1 year were all within the pediatric range (0–14 years).

INTRODUCTION

Brain death (BD) is the conceptualization of a pathophysiologic entity that has affected the human species for only a tiny portion of its history. The dramatic impact of resuscitation technology (e.g., first cardiac defibrillation in 1947 [1] and first mechanical ventilation in 1950 [2]) created this clinical entity. In 1954, Schwab was the first to report withdrawing treatment from a patient without reflexes and spontaneous breathing, in the presence of electroencephalographic silence [3]. In 1959, French neurologists Wertheimer and Jouvet concurred with Schwab on diagnosis and treatment withdrawal for what they defined “death of the nervous system” [4, 5]. During the same year, at the 23rd International Neurological Meeting, Mollaret and Goulon presented 23 patients diagnosed with the then‐neologism “coma dépassé,” which was described as the state when “total and definitive abolition of vegetative functions is added to the abolition of the functions of relation” [6]. The idea of BD was gradually conceptualized over the following years on both sides of the Atlantic Ocean [4, 7], and later codified by the Harvard Ad Hoc Committee in 1968 [8] and endorsed by the World Medical Association [9]. In 1976, the UK Conference of Medical Royal Colleges and their Faculties published the guidance on the diagnosis of BD, which adopted the stance of Mohandas and Chou [10] on indicating the loss of brainstem function as pivotal to the state of BD described in the Harvard expert opinion [11]. Three years later, a follow‐up memorandum proposed brainstem death as equivalent to death of the whole person [12]. In the USA, the National Institutes of Neurological Disease and Stroke suggested adding an assessment of cerebral blood flow to the Harvard Criteria, giving additional validation to BD as a clinical entity [13]. National legal validation was provided in 1981 by the President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research [14]. The Commission adopted the conceptualization of BD as proposed by the senior author [15], and legally sanctioned it with the model statute (i.e., Uniform Determination of Death Act) [14]. The American Academy of Neurology standardized the diagnostic criteria for BD in 1995 [16] and again in 2010 [17].

The practice of BD has been endorsed by contemporary judicial and medical practices. But, despite its wide scientific and legal acceptance, the BD concept has been persistently challenged since its inception. The principal epistemological objection is that the concept of BD was introduced as a definition tout court, without the support of empirical data. By this perspective, BD was established as a self‐fulfilling prophecy because, upon its declaration, either mechanical ventilation is disconnected or vital organs are transplanted [18]. Such criticism has been reinforced in the general population by sensational news of patients “recovering” from BD, reported with layman jargon in nonscientific media (Table S1).

To scientifically rebut such criticism, we believed that evidence could be provided empirically. First, we elucidated the natural history of BD, to determine whether somatic expiration always ensues when somatic support is maintained. Second, the accuracy in diagnosing BD is so obviously consequential that its degree ought to be set extremely high. Therefore, we evaluated not only the rate, but also the natural history of patients with a BD misdiagnosis. Third, we investigated the predictors of prolonged somatic survival in somatically supported BD patients. By answering such questions meta‐analytically, we aimed at eliminating the stigma of “self‐fulfilling prophecy” and granting BD the dignity of a clinical entity based on scientific evidence.

METHODS

Registration

The study protocol for this meta‐analysis was registered with the number CRD42022347705 at PROSPERO (International Prospective Register of Systematic Reviews in Health and Social Care), developed and maintained by the Centre for Reviews and Dissemination of the University of York, United Kingdom [19]. The systematic literature review was undertaken according to Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines [20] and summarized in Table S2. Participants, interventions, comparisons, outcomes, and study design (PICOS), along with data items and eligibility criteria are detailed in Table S3.

Search strategy, data processing, and statistical software

Please refer to Supplemental Methods (Material S1).

Endpoints and eligibility criteria

The primary endpoint was the percentage of somatic expiration (i.e., death of the whole body) after a diagnosis of BD, when critical care support is maintained. Secondary endpoints were BD misdiagnosis, length of somatic survival after BD, and its predictors. Inclusion criteria were (i) aggregate data on the natural history of BD, when full somatic support is maintained (aggregate meta‐analysis); and (ii) patient‐level data on BD patients, with censoring in correspondence of treatment withdrawal or organ donation. Clinical exclusion criteria were as follows: aggregated data including BD patients in whom treatment was withdrawn or organ donation occurred (because it would not be possible to censor individual patients if only aggregated data were provided); unclear diagnostic algorithm of BD; severe brain damage other than BD; and unspecified modality of somatic death. Nonclinical exclusion criteria were as follows: necessary endpoint not specified; evidence predating the formulation of BD diagnosis; and overlapping series/cases. The vast majority of authors classified the etiology of BD as either primary cranial pathology (e.g., trauma, neoplasm, ischemia, hemorrhage, often coalesced together) or cardiopulmonary arrest. Additionally, in a sizable number of publications the various types of primary cranial pathology were coalesced together. Therefore, we are not able to discern, out of the hypoxic–ischemic brain injury group, which ones were due to cardiopulmonary arrest (extracranial etiology) and which ones due to an intracranial etiology (e.g., stroke, vascular compression by intracranial mass). Hence, although not ideal, we adopted this binary classification for statistical analysis purposes with the caveat that “cardiopulmonary arrest” refers to “hypoxic–ischemic injury of extracranial etiology.”

Measures of treatment effect, heterogeneity, and data distribution

Pooled estimates and their 95% confidence intervals (CIs) were calculated with the Mantel–Haenszel fixed and random effect models, depending on whether there was respective absence or presence of significant publication heterogeneity. In‐between study heterogeneity was examined with the Cochrane Q (χ ²) test, and we further quantified inconsistency by calculating I ², interpreted using the following guide: 0%–40% might not be important; 30%–60% may represent moderate heterogeneity; 50%–90% may represent substantial heterogeneity; and 75%–100% may represent considerable heterogeneity [21, 22]. Shapiro–Wilk test assessed continuous variables for Gaussian distribution before being analyzed as mean ± SD.

Assessment of methodological quality, reporting bias, and sensitivity analysis

The methodological quality of the selected studies was assessed with the JBI critical appraisal checklists for case series and case reports [23, 24]. The risk of reporting bias was evaluated quantitatively with the Egger regression intercept. Sensitivity analysis was performed for the primary endpoint, by determining whether consequent removal of one study at a time led to influential observations.

Meta‐regression

In the patient‐level analysis, predictors of somatic survival length were evaluated first with univariate linear regression, and then retained for the multivariable linear model if they presented a p‐value < 0.2 or otherwise were deemed clinically relevant by a priori selection. Multicollinearity among independent variables was evaluated with variance inflation factor analysis and excluded by a value ≤ 3.

RESULTS

Selection of studies

A total of 1906 studies were identified according to the search strategy. This initial pool was screened according to the prespecified eligibility criteria, and consequently many articles were excluded as detailed by the flowchart in Figure S1. The methodological quality of the included studies is detailed in Figure S2. Table S4 lists series that were cited by other authors [25] as aggregated evidence of the natural history of BD with somatic support, but did not pass our eligibility criteria partially [10, 14, 26, 27, 28] or completely [29, 30].

Overall sample

The 47 articles finally selected were published over a 5‐decade span (1969–2021) and encompassed 1610 BD patients [14, 18, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72]. The length of somatic survival after a diagnosis of BD ranged from 1.6 h to 19.5 years. Nine cases of BD misdiagnosis were detected (Table S5). Influential analysis revealed that all selected publications contributed homogeneously to the aggregated evidence; there were no influential observations for the primary endpoint (i.e., sequential removal of each publication did not alter the significance of the aggregated result; Figure S3).

Natural history of BD

Somatic support was maintained in 1564 patients. The mean age at the time of BD diagnosis was 32.9 years (95% CI = 21.3–44.6), and 54.9% of patients were of male sex (95% CI = 37.6–71.2). Primary cranial pathology (e.g., trauma, neoplasm, ischemia, hemorrhage) was largely prevalent (91.4%, 95% CI = 73.9–97.6), and cardiopulmonary arrest was the second most common cause of BD (7.7%, 95% CI = 2.7–20.3). With somatic support, somatic expiration ensued in 99.9% of BD patients (95% CI = 89.8–100). The mean length of somatic survival was 8.0 days (95% CI = 4.5–11.6). These results, along with an evaluation of heterogeneity and reporting bias, are detailed in Table 1.

TABLE 1.

Natural history of brain death with somatic support.

Variable	Estimate	Heterogeneity		Reporting bias
	Estimate (95% CI)	p	I 2, %	Intercept	SE	p
Demographics
Age, years	32.9 (21.3–44.6)	<0.01	99.4	14.2	3.1	<0.01
Male sex, %	54.9 (37.6–71.2)	<0.01	71	0.1	2.4	0.97
Etiology, %
Cranial	91.4 (73.9–97.6)	<0.01	78.4	3.2	0.9	<0.01
Cardiopulmonary	7.7 (2.7–20.3)	<0.01	69.3	−2.8	0.7	<0.01
Natural history after BD
Somatic expiration, %	99.9 (89.8–100)	0.97	0	1.2	2.3	0.62
Somatic survival, days	8.0 (4.5–11.6)	<0.01	91.5	2.9	1.1	0.03

Note: Natural history of patients after diagnosis of brain death, when somatic support was maintained (i.e., no withdrawal of treatment or organ donation). Etiology: cranial = traumatic or nontraumatic (e.g., tumor, hemorrhage) cranial pathology leading to brain death; cardiopulmonary = cardiac or respiratory arrest leading to brain death; the remaining portion of brain deaths had alternative etiologies. Somatic expiration: expiration of the soma by cardiac arrest, following the diagnosis of brain death; of 1564 patients who were somatically supported after a diagnosis of brain death, only one patient recovered to normal neurological function and survived. Somatic survival: survival of the soma after diagnosis of brain death.

Abbreviations: BD, brain death; CI, confidence interval.

Predictors of somatic survival length

The granular data necessary for a patient‐level risk analysis were available for 124 BD patients. Their median age was 30 years (range = newborn to 85 years). Female sex was prevalent (n = 55, 52.4%), with a high pregnancy rate (n = 23, 41.8%). BD was due to primary cranial pathology in the majority of cases (n = 88, 72.7%), and cardiopulmonary arrest was the second most common etiology (n = 24, 19.8%; Figure 1). Multivariable regression indicated that only age (at the time of BD diagnosis) was an independent predictor of somatic survival length (estimate = −11.8, SE = 4, p < 0.01); there was no multicollinearity in the model (Figure 2). The negative linear relationship between age and somatic survival, along with color‐coded censoring of treatment withdrawal or organ donation (n = 30) and of misdiagnosis (n = 9) cases, is depicted in Figure 3.

FIGURE 1.

Patient‐level analysis: demographics, etiology, and misdiagnoses. Patient‐level analysis is shown of 124 brain death cases that were published with granular data. The top graphical row depicts age at brain death diagnosis and somatic survival length, two continuous variables expressed in terms of quartiles (color‐coded quartile balls) and of range (minimum–maximum). The second graphical row from the top depicts how female sex was preponderant in publications providing granular data, likely because almost half of women were pregnant and kept on somatic support until delivery at term. The third graphical row from the top depicts the etiological partition of such cases. The bottom graphical row depicts the rate of misdiagnosis in publications providing granular data, distinguishing “functionally brain‐dead" (FBD; i.e., subsequent detection of ≥1 findings that are incongruent with a previous diagnosis of brain death, albeit of no clinical consequence in altering the lethal prognosis) patients from the single case of “normal neurological function recovered” (NNFR).

FIGURE 2.

Multivariable linear regression. At multivariable linear regression, age at the time of brain death diagnosis was the only independent predictor of somatic survival length. The "zero effect" line is depicted in red. Coefficient estimates (dots) and their SE (width) are depicted by the blue lines; **p < 0.05. Multicollinearity was excluded by variance inflation factor (VIF) analysis (green line estimates), which did not reach the threshold of 3 (purple line).

FIGURE 3.

Age and somatic survival. Scatterplot depicts the negative linear relationship between age at brain death diagnosis (x‐axis) and somatic survival length (y‐axis). The specified color code was applied to distinguish the modality of somatic expiration (spontaneous cardiac arrest vs. withdrawal of treatment organ donation) and the type of misdiagnosis (“functionally brain‐dead” [FBD] vs. “normal neurological function recovery” [NNFR]). The NNFR patient was censored at the time of latest follow‐up (1 year). For every 1‐year increase in age, there was an almost 12‐year decrease in somatic survival length. Patients with a somatic survival ≥1 year were all within the pediatric range (0–14 years).

Misdiagnoses

The nine patients with BD misdiagnosis included eight termed “functionally brain‐dead” (FBD; i.e. after pronouncement of BD according to properly applied diagnostic criteria, there was subsequent detection of ≥1 findings that were incongruent with guidelines for its diagnosis, albeit the lethal prognosis was not altered) cases and one case of neurological recovery and survival [48, 52, 57, 59, 63, 64, 66, 71]. The age at BD misdiagnosis ranged from 2 days to 55 years. In the eight FBD cases, the type of neurological finding that was incongruent with a previous diagnosis of BD varied from resurfacing of spontaneous breathing to movements on command. The duration of such incongruent findings ranged from 30 min to 4.5 years. We specifically refer to the lapsed time between the detection and the disappearance of a clinical finding incongruent with the previous declaration of BD. The incongruent finding could not be present at the declaration of BD; otherwise, such declaration would not have been made. Additionally, the incongruent finding could disappear before somatic death occurred. Possible confounding factors for certain diagnoses of BD were hypothermia and neurotropic drugs, although their relevance was contested by the authors [48, 52, 71]. Treatment was withdrawn in five of the eight FBD patients, whose somatic survival ranged from 2 h to 4.5 years. The sole case of neurological recovery occurred in a premature (35 weeks gestational age) female newborn, who was declared BD (with absence of confounding factors) following cerebral hemorrhage. She was deemed BD from 40 to 120 h after birth, after which a steady neurological recovery ensued. The patient was found in good health at 1‐year follow up [53]. Details of the BD misdiagnoses are available in Figure 4.

FIGURE 4.

Brain death misdiagnoses. Reports of incongruent findings with brain death after its diagnosis are shown. Etiology = cause of brain death. Incongruent Finding = type of finding inconsistent with a diagnosis of brain death. Incongruence Duration = interval of time during which findings incongruent with brain death were found after its diagnosis. Confounding Factors = factors that were present but should be excluded before diagnosis of brain death. "Somatic" pertains to the body after a diagnosis of brain death. ¹Eye movements, blinking, nystagmus, grimacing, oculocephalic and corneal reflexes, disorganized EEG pattern reappearance, delayed brainstem conduction at auditory evoked potential. 2Minimal deficit at last follow‐up (1 year). ³Debatable and contested by authors. ⁴The timing of initial resurfacing of movements on command is unknown. EEG, electroencephalogram; EP, auditory evoked potentials; F/U, follow‐up.

DISCUSSION

A declaration of BD is so consequential that the degree of certainty ought to be extremely high. Our meta‐analysis confirmed that the diagnosis of BD is one of the most reliable in medicine, with the caveat that all diagnostic criteria are rigorously and correctly applied as the World Brain Death Project advocates [73].

Selected evidence

We conducted a comprehensive meta‐analysis of the natural history of BD and of the reliability of its diagnosis, aggregating the relevant evidence beginning with the conceptualization of this clinical entity in 1968 [8]. In the overall sample, the average BD patient was a man in his 30s, likely a consequence of cranial trauma being the most common etiology of brain injury in this demographic. The publications presenting aggregated evidence served to provide empirical evidence on the natural history of BD with somatic support. Publications with patient‐level data were usually prompted by unusual cases that authors deemed worthy of special attention, illustrating extended somatic survival or misdiagnosis. In this subset of publications with granular data, female sex predominated, because nearly half were pregnant and maintained on somatic support to deliver the offspring at term [18, 32, 37, 38, 39, 40, 42, 45, 49, 54, 55, 56, 68].

Natural history of BD

Our results provide further empirical evidence to refute the criticism that BD is not a reliable diagnosis but rather is a self‐fulfilling prophecy because it is followed by either treatment withdrawal or organ donation. Our meta‐analysis showed that somatic expiration invariably followed a BD diagnosis when somatic support was maintained, with the single exception published by Pasternak [59] more than a half‐century ago. The average somatic survival length after a diagnosis of BD was of 8 days with maintained somatic support, although there have been a few sensational outliers.

Somatic survival and chronic BD

In the early BD era, the belief that somatic expiration invariably followed BD in a matter of days was generally accepted by the medical community [14]. Subsequently, the dramatic improvement of critical care technology allowed some young patients to overcome the acute phase of BD (characterized by multiorgan failure, severe hypotension, spinal shock with related hemodynamic instability, disseminated intravascular coagulation, and electrolyte imbalances), allowing patients to transition into a chronic phase. Alan Shewmon classified BD as “chronic” when somatic survival was ≥1 week [18]. First, the validity of BD diagnosis should not be based on the duration of somatic survival, because that commits the fallacy of confusing prognosis and diagnosis. The conceptual validity of a BD diagnosis should be based on its condition per se, and not its duration. Furthermore, our meta‐analysis indicated that no patient had ever recovered from chronic BD. Irrespective of the length of the temporal gap between the two, somatic expiration always followed chronic BD. Because the probability of BD misdiagnosis is inversely proportional to the somatic survival length, chronic BD cases paradoxically testify to the irreversibility of a properly diagnosed BD. Therefore, medical providers should dedicate a sufficient amount of time to counsel proxies and next of kin to selflessly consider that artificially extending somatic existence has never yielded any desired outcome.

The longest somatic survival of a BD patient is the case presented by Repertinger and colleagues. This 4‐year old boy, diagnosed with BD following meningitis, had a somatic survival of 2 decades. He spent most of his BD existence in the maternal home, only requiring mechanical ventilation, corticosteroids, and occasional antibiotics. He underwent somatic growth, although with minimal secondary sexual development. At age 24 years, his body succumbed to cardiac arrest after the mother decided that no resuscitative efforts should be undertaken [62]. Our meta‐regression indicated that the only independent predictor of somatic survival length was age; for every 1‐year increase in age at the time of BD diagnosis, there was an almost 12‐year decrease in somatic survival length. All the BD patients with a somatic survival ≥ 1 year were in the pediatric age range (i.e., 0–14 years; Figure 4). Nonetheless, we ought to acknowledge that besides the expected higher physiological resilience typical of younger bodies, age may well be a surrogate of the level of care provided. It is likely that the willingness to provide somatic support in terms of its degree and duration is inversely proportional to the patient's age. The only way to address this selection bias would be to perform a subanalysis relating the duration of somatic survival to the level of somatic support. Unfortunately, this subanalysis was not feasible due to lack of specific data on the degree of somatic support in a considerable number of publications.

Furthermore, the reader should beware that, although the term “somatic support” indicates a level of care able to sustain at minimum basic physiological functions, the degree of physiological optimization (e.g., type of cardiovascular, respiratory, and renal support, hormonal replacement therapy, route of nutrition) varied on a spectrum that was likely influenced by the patient's age.

Misdiagnoses

Only nine cases of BD misdiagnosis have been recorded in peer‐reviewed evidence [48, 52, 57, 59, 63, 65, 66, 71], of which two were exceptionally noteworthy. The case presented by Pasternak et al. was sensational because it represents the only human being scientifically recorded as “resurrected” from BD. This premature newborn met all the criteria of BD diagnosis with no confounding factors and was thriving at the 1‐year follow‐up [59]. It is extremely important to point out that this case was published in 1979. Current societal guidelines of neurology do not endorse the declaration of BD in premature infants less than 37 weeks of gestational age, because of the lack of validation due to insufficient evidence in the literature [74].

The second case reported by Shewmon was the only published case of a person reliably declared BD who later moved on command. This 13‐year‐old girl, Jahi McMath, was declared BD following a cardiac arrest from a hemorrhage complicating oropharyngeal surgery. Because her family contested the diagnosis, a court‐appointed external expert (i.e., the chief of pediatric neurology at Stanford University Medical Center) confirmed her to be BD, with unequivocal evidence also by ancillary tests. Expired by Californian law, the patient was transported by the family to New Jersey, where religious objections to the neurological determination of death are permitted. She spent the next 4 years between hospitals and home care, undergoing somatic growth and a degree of secondary sexual maturation. The extraordinary feature, which brought wide media attention to the case, was that she allegedly was able to move parts of her body on command. This unexpected finding was witnessed by nurses, lawyers, her internal medicine physician (i.e., Dr. Alieta Eck), and an experienced neurologist (i.e., Dr. Alan Shewmon) [66]. McMath represents the most exceptional outlying case well documented both in the mass media and the scientific literature.

Conversely, other sensational cases in the mass media have not been submitted to the scrutiny of scientific peer review, for several reasons. First, sensational news can be generated by individuals (family members, nonmedical health care staff, physicians without neurological understanding, mass media personnel) who lack the knowledge to differentiate states of altered consciousness (e.g., vegetative state, minimally conscious state, coma, locked‐in syndrome, BD). Second, these news reports of truly misdiagnosed cases of BD are unlikely to be submitted to the scrutiny of the scientific community, because the diagnosing physicians are aware that official diagnostic criteria were not properly executed (Table S1).

Therefore, a global publication bias is plausible and prevents the medical community from precisely discerning, quantifying, and learning from diagnostic fallacies. Hence, a “super partes” governance platform, to which to address anonymous and confidential enquiries bypassing the reluctance of directly involved providers, is warranted. Such governance platforms are, for instance, already available in the UK to inquiry about maternal and perinatal deaths [75].

Relationship with previous meta‐analyses

A quarter‐century ago, Shewmon published the only other relevant meta‐analysis, which was limited to the cases of chronic BD available at that time [18]. Compared to Shewmon's meta‐analysis, ours adds considerable information, because we added 1535 subjects, we tabulated the data on somatic survival length for some patients who were BD but still surviving at the time of the previous meta‐analysis, and our analysis benefited from advanced statistical analysis.

Limitations

One intrinsic limitation of our meta‐analysis lies in its material: retrospective data from institutional databases and suprainstitutional registries. Reciprocal strengths were (i) correspondence in meta‐analytic results of granular (patient‐level) and aggregated data; and (ii) rigorous methodology, ensured by a prospectively registered protocol and PRISMA standards, and by a high methodological quality of the selected studies. Another limitation could be the potential reporting bias, which was detected for some secondary endpoints. Nonetheless, the primary endpoint was homogeneous among studies with no reporting bias or heterogeneity. Consequently, this meta‐analysis represents the best evidence possible, as no prospective randomized studies are foreseeable due to the nature of the condition investigated.

CONCLUSIONS

Somatic expiration followed a diagnosis of BD in all adult patients, and in all cases of chronic BD irrespective of age. The only case of survival and neurological recovery recorded in the medical literature was BD in a premature newborn. The remaining cases of misdiagnosis regarded FBD patients, and therefore the fatal prognosis was not altered. The only independent predictor of somatic survival length is age at the time of BD diagnosis. Thus, the current diagnostic criteria for BD are highly reliable, though there is a duty for doctors and institutions involved in determining BD to investigate any cases of misdiagnosis and always be seeking to improve diagnostic accuracy. Medical providers are advised to counsel patients' proxies that a diagnosis of BD is one of the most accurate in medicine according to the best level of evidence possible, provided that all the diagnostic criteria are properly applied.

AUTHOR CONTRIBUTIONS

Ivancarmine Gambardella: Conceptualization; investigation; writing – original draft; methodology; validation; visualization; writing – review and editing; software; formal analysis; project administration; data curation; supervision. Francesco Nappi: Conceptualization; supervision. Berhane Worku: Data curation; supervision. Robert F. Tranbaugh: Supervision; data curation. Aminat M. Ibrahim: Data curation; investigation. Sandhya K. Balaram: Validation; supervision. James L. Bernat: Conceptualization; writing – review and editing; supervision.

CONFLICT OF INTEREST STATEMENT

None of the authors has any conflict of interest to disclose.

Supporting information

Appendix S1.

Gambardella I, Nappi F, Worku B, et al. Taking the pulse of brain death: A meta‐analysis of the natural history of brain death with somatic support. Eur J Neurol. 2024;31:e16243. doi: 10.1111/ene.16243

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.