Eleven-year retrospective study characterizing patients with severe brain damage and poor neurological prognosis -role of physicians’ attitude toward life-sustaining treatment

Background Severe brain hemorrhage/infarction and cardiac arrest constitute the most critical situations leading to poor neurological prognosis. Characterization of these patients is required to offer successful end-of-life care, but actual practice is affected by multiple confounding factors, including ethicolegal issues, particular in Japan and Asia. The aim of this study is to evaluate the clinical courses of patients with severe brain damage and to assess the preference of end-of-life care for these patients in Japanese hospitals. Methods A retrospective observational study was conducted between 2008 and 2018. All intracranial hemorrhage/infarction and cardiac arrest out-patients (n = 510) who were admitted to our two affiliated hospitals and survived but with poor neurologic outcomes were included. Demographic characteristics as well as prognosis and treatment policies were also assessed. Results Patients were divided into two categories; cases with absent brainstem reflex (BSR) (BSR[-]) and those with preserved BSR (BSR[ +]). The survival rate was higher and the length of hospitalization was longer in patients with BSR[ +] than in those with BSR[-]. Among three life-sustaining policies (i.e., aggressive treatment, withdrawal of treatment, and withholding of treatment), withholding of treatment was adopted to most patients. In BSR[-], the proportion of three treatment policies performed at the final decision did not differ from that at the initial diagnosis on neurological status (p = 0.432). In contrast, this proportion tended to be altered in BSR[ +] (p = 0.072), with a decreasing tendency of aggressive treatment and a modest increasing tendency of withdrawal of treatment. Furthermore, the requests from patients’ families to withdraw life-sustaining treatment, including discontinuation of mechanical ventilation, increased, but actual implementation of withdrawal by physicians was less than half of the requests. Conclusions BSR constitutes a crucial determinant of mortality and length of hospitalization in comatose patients with severe brain damage. Although the number of withdrawal of life-sustaining treatment tends to increase over time in BSR[ +] patients, there are many more requests from patients’ families for withdrawal. Since physicians has a tendency to desist from withdrawing life-sustaining treatment, more in-depth communication between medical staff and patients’ families will facilitate mutual understanding over ethicolegal and religious issues and may thus improve end-of-life care.


Background
Severe brain damage with poor neurological prognosis is a critical state that reflects devastated conditions, including brain death, brainstem dysfunction and irreversible hypoxic encephalopathy. Cardiopulmonary arrest (CPA) is one of the most crucial events and constitutes a major leading cause of death, accounting for more than 10,000 cases annually in Japan [1]. Only 20% of out-of-hospital cardiac arrest (OHCA) patients are able to restore spontaneous systemic circulation, among whom, however, 80% have severe brain damage with poor neurological prognosis [2]. Furthermore, severe brain damage is also precipitated by intracranial disease such as cerebral/subarachnoid hemorrhage and cerebral infarction, which occur in approximately 300,000 subjects annually in Japan [3]. Despite vigorous implementation of medical care, current advances in clinical practice fail to halt the ravaging process or the progression to brain death and brainstem dysfunction in many patients resuscitated from CPA or with critical brain hemorrhage/infarction.
Although advanced brain damage frequently involves ethical considerations or arguments on end-of-life care, the strategy for the medical practice in patients with poor neurological prognosis may vary depending on personal attitudes of physicians or specific regions of the hospitals [4][5][6][7][8][9][10]. Most Western physicians recognize that withholding and withdrawal of life-sustaining treatment are acceptable [4][5][6][7]10]. In contrast, Asian countries are facing different views regarding physicians' attitudes toward end-of-life care [4,5,8,10]. In Japan, physicians used to be reluctant to withdraw life-sustaining treatment for patients who were comatose and had poor neurological prognosis [4,5]. Meanwhile, the Government [11] and several academic societies [12][13][14] published the guidelines for end-of-life care, which qualified the physicians' attitudes toward withdrawal of life-sustaining treatment. Under the milieu where medical care becomes more complicated and multifaceted, however, much concern arises regarding availability and necessity of life-sustaining measures, post-resuscitation care and prolonged hospitalization, all of which entail important issues to be settled with utmost requirement. Thus, there still remains a controversy as to the decision-making on the end-oflife care and management of the comatose patients with severe brain damage. Furthermore, although the presence of brainstem dysfunction is generally recognized as more severe brain injury [15], it remains undetermined whether the disappearance of brainstem reflex (BSR) modifies the attitudes of physicians toward end-of-life care.
Our medical facilities, i.e., St. Marianna University Hospital and St. Marianna University Yokohama Seibu Hospital, play a central role in the emergency care and management as tertiary medical centers over the Yokohama and Kawasaki area and accept a large number of critically ill patients. During these eleven years, we have experienced many patients with severe brain hemorrhage/infarction or comatose patients who survive from CPA. Among these patients, approximately 500 patients were afflicted with neurologically critical sequelae presenting variable degrees of brain/brainstem dysfunction but not reaching a level of brain death. The purpose of this study is to evaluate the consequence of these patients with special reference to BSR and prognosis. Furthermore, serial changes in the preference of end-of-life treatment, i.e., withdrawal/withholding of treatment, were assessed in Japanese patients with severe brain damage. time in BSR[ +] patients, there are many more requests from patients' families for withdrawal. Since physicians has a tendency to desist from withdrawing life-sustaining treatment, more in-depth communication between medical staff and patients' families will facilitate mutual understanding over ethicolegal and religious issues and may thus improve end-of-life care.

Study population
Keywords: Brain hemorrhage, Cerebral infarction, Cardiac arrest, Brain death, Withdrawal, Attitude toward treatment, Ethics The patients devoid of BSR who suffered brain hemorrhage/infarction and severe neurological prognosis or the OHCA patients who survived for more than 72 h but with poor neurologic outcomes were admitted to our hospitals (n = 569, Fig. 1). Patients who experienced CPA in our hospitals (i.e., in-hospital CPA) or who had suffered CPA in other hospitals and then transferred to our facilities were not enrolled in this study. In addition, the patients who met the following criteria were excluded; (1) patients in a potentially reversible coma (e.g., acute alcohol/drug addiction, hypothermia, sepsis, uremia, encephalopathy due to metabolic and endocrine diseases, use of neuromuscular blocking or high-dose sedative agents), (2) children under the age of 16, and (3) patients with a history of medical errors or social problems.
Among the potential cases (n = 530), the patients who were legally determined as brain death were excluded from the study (Fig. 1). The diagnosis of legally determined brain death was made based on the Japanese guidelines for organ transplantation [https:// www. mhlw. go. jp/ stf/ seisa kunit suite/ bunya/ 00000 40046. html] (i.e., flat EEG, absent BSR and positive apnea test). The patients with non-convulsive status epilepticus (NCSE) were excluded from the group of CPA patients. Accordingly, the patients finally enrolled in this study were arbitrarily categorized into two groups; (1) cases devoid of BSR (BSR [-], n = 275) and (2) those with preserved BSR (BSR[ +], n = 235, Fig. 1).
Demographic data, including age, gender and etiology, were evaluated. Vital signs and neurological status (e.g., BSR, consciousness level, apnea test and EEG data) along with clinical outcomes were also assessed. Prognosis of the patients (e.g., survival to hospital discharge or transfer) and length of stay (LOS) in ICU or hospital were also evaluated.
Upon admission, patients received multipronged life-sustaining therapies, including administration of catecholamines/antibiotics, enteral feeding and endotracheal intubation. The evaluation of the severity of brain damage and the prognostication was performed among the patients who survived more than 72 h and manifested consciousness level of M3 or lower under no influence of sedatives/hypnotics/analgesics [16,17]. Brain CT was performed in all patients and approximately half of the patients received EEG.
Following the assessment of patients' neurological status and prognostication, treatment policies were repeatedly discussed with their families/surrogates, i.e., whether they should receive aggressive life-sustaining support therapy or whether ongoing life-sustaining therapy should be withdrawn/withheld. We also obtained from the families the information regarding the perception about end-of-life care that the patients had embraced. Based on the results of the in-depth discussion, final therapeutic policies were determined by multiple physicians, including neurologists.

Statistical analysis
Results of continuous variables are expressed as the median (interquartile range; IQR). Categorical variables are presented as the number. The Mann-Whitney U test or the Kruskal-Wallis test was used for comparison between two groups or among three groups, respectively. The chi-square test or Fisher's exact test was applied for analysis of categorical valuables. Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 25 (IBM Japan Ltd, Japan). P values less than 0.05 were considered statistically significant.

A. Patients' characteristics
During the period between April 2008 and December 2018, we enrolled 510 patients with BSR[-] (n = 275) and BSR[ +] (n = 235) in our two affiliated hospitals ( At the time of admission, brain CT scan was also performed in all cases, showing that most of the CPA patients manifested diffuse anoxic brain injury.
Among the cases with equivocal CT findings, we confirmed the findings of diffuse anoxic brain injury by a second series of CT scan, which was apparent even in patients with no EEG evaluation or indeterminate EEG findings.
In patients with BSR[-], the survival rate was not different between the subgroup with flat EEG and Temporal changes in the ratio of each treatment policy were assessed. In patients with BSR[-], the proportion of three treatment practices conducted after the final decision did not differ from that of policies decided at the time of diagnosis (χ 2 test; p = 0.432) (Fig. 3). In patients with BSR[ +], the proportion of the treatment policies tended to be altered (χ 2 test; p = 0.072), with a decreasing tendency of the aggressive life-sustaining support subset (from 15.3% to 9.8%, p = 0.070) and a modest increasing tendency of the withdrawal subgroup (from 2.1% to 4.7%, p = 0.127). Figure 4A illustrates serial changes in the number of the requests of patients' families/surrogates to withdraw life-sustaining treatment during hospitalization. Most of the requests were made within 7 days from the initial diagnosis on patients' neurological status although Withdrawal/withholding of life-sustaining treatment included discontinuation of drugs (catecholamines, antibiotics), enteral feeding and hemodialysis (  Table 4A). When each medical practice was evaluated, physicians recommended organ transplantation following the withdrawal of life-sustaining treatment, particularly in BSR[-] patients (p = 0.022) and preferred the discontinuation of hemodialysis in BSR[ +] patients (p = 0.020, Table 4B). In contrast, patients' families requested the discontinuation of mechanical ventilation in both BSR[-] (p = 0.029) and BSR[ +] patients (p = 0.022).

Discussion
Intracranial hemorrhage/cerebral infarction and CPA are serious events that cause ominous sequelae, including severe brain damage with poor neurological prognosis and/or brainstem dysfunction. There have been reported many studies and reviews showing a substantial number of patients who develop severe brain damage following intracranial hemorrhage/infarction or resuscitation from CPA [18,19]. Additionally, in some resuscitated patients, irreversible hypoxic encephalopathy ensues which requires differential diagnosis for brain death. Such devastating status always worries physicians with their attitudes toward treatment policies because interrupting ongoing therapies might evoke ethicolegal and religious issues, particularly in Japan and other Asian countries [4,5,8,10]. To the extent that withdrawal of treatment and execution of organ donation depend on the brain status of patients, it appears extremely critical to recognize the Table 3 Withdrawal or withholding of life-sustaining treatment Table 4 Impacts of physicians' suggestions and families' requests on withdrawal policy characteristics and the temporal profiles of patients following brain hemorrhage/infarction or resuscitated from CPA.

A. Characterization of patients with BSR[-]/[ +]
The present study shows that during the last 11 years, we experienced 510 patients who suffered severe sequelae of brain damage resulting from intracranial hemorrhage/infarction and CPA in our two affiliated hospitals (Fig. 1). These unconscious patients were classified as cases with absent brainstem reflex (i.e., BSR [-]) and those with the preserved reflex (i.e., BSR[ +]). Thus, 53.9% of the patients developed severe brain damage with BSR[-], a status conventionally deemed as a category of 'brainstem death' [20][21][22] and resulted in markedly less survival; the patients of this subgroup had survival rate of 0.7% and LOS of 3 days (Table 2), as compared with the patients with BSR[ +] showing 31.9% of survival rate and 16 days of LOS. These findings are consistent with the premise that brainstem dysfunction reflects more severe brain injury [15] and may further lend support to the presumption that impaired BSR affects not only the physicians' attitudes toward treatment practices but also the decision-making process of patients' families; they might request to withdraw life-sustaining treatment and propose organ donation. Indeed, organ transplantation was performed in six cases ( Table 3).
The impact of brainstem function and EEG activity on mortality remains a matter of controversy. The present study showed that among patients with BSR[-], only 29.8% of the patients underwent EEG examination, and most of this subgroup manifested a flat EEG (Fig. 2). Furthermore, because of a paucity of survivors in the BSR[-] group, the survival rate did not depend on whether an EEG was implemented (p = 0.088) nor on the types of the EEG wave pattern (p = 1.0). In this regard, several statements and reports indicate that EEG is not a requirement for the diagnosis of brain death unless mandated by regional laws [23,24]. It turns out therefore that implementation of the EEG plays a permissive role in patients with BSR [-].
In contrast, a substantial number of the patients with BSR[ +] had a chance to undergo EEG examination (i.e., 69.4%, Fig. 2 Of note, the present study shows that there are some differences in the suggestions and requests for withdrawal between physicians and patients' families. As illustrated in Table 4, physicians suggested withdrawal of life-sustaining treatment, resulting in implementation of organ transplantation, particularly in patients with BSR [-]. In striking contrast, 20 families/surrogates of the patients requested the discontinuation of mechanical ventilation whereas none of the physicians had suggestion for this withdrawal. Unlike European physicians, Japanese doctors are less likely to withdraw life-sustaining treatment [4,5], which attitude may be reflected by no suggestion regarding the cessation of mechanical ventilation. Alternatively, the physicians essentially recognize that the recommendation of organ transplantation is rendered contributory to transplant therapy and may offer an opportunity for functional restoration to many potential recipients with end-stage organs. The attitudes of Japanese physicians may thus constitute a determinant of end-of-life care policies that cannot be modified so easily probably because of our religious belief or societal culture.

C. Limitations
The present study has been conducted in two affiliated institutions. Thus, the critical care management and ethical responses to CPA in these two institutions may not be fully identical. Nevertheless, communication between the staff of these two facilities is well maintained through regular assembly and medical and other technical information is shared, which would minimize the gap between two hospitals. Caveat is in order, however, since the attitudes of our medical staff toward treatment policies might be deviated from a national consensus on end-oflife care. Finally, life-sustaining treatment policy for patients with severe brain damage may be affected by personal attitudes of physicians toward treatment of which physicians' religious belief could be a determinant. Reinforced decision-making frameworks involving not only physicians but also co-medical staff and lawyers would hence unravel this problem.

Conclusions
The present study demonstrated the characteristics of the patients who suffered severe brain hemorrhage/ infarction or were resuscitated from CPA but with poor neurological outcomes. Preserved BSR constitutes a determinant of survival and length of hospitalization. More importantly, although requests of patients' families to withdraw life-sustaining treatment increased over time, actual interruption does not correspond to the request due to multifaceted confounding factors, including medicolegal frameworks and physicians' attitudes to the treatment. This discrepancy appears to be more conspicuous in Japan, and sociolegal, ethical and religious issues surrounding our medical circumstance should be taken into consideration when discussing end-of-life care for patients with severe brain damage with poor neurological prognosis.