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 [16,17]. 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 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 (Figure 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’ [18-20] 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  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 (Figure 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 [21,22]. 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%, Figure 2), probably due to a longer period of hospitalization. Among this subgroup, there was no difference in survival rate between patients with flat EEGs (29.2% [=19/65]) and those with diffuse slow waves (36.0% [=31/86], p=0.378). Because a number of the patients with BSR[+] had no chance to undergo EEG examination, it awaits further evaluation whether EEG findings affect survival rate in patients with BSR[+].
B. Treatment policy
The treatment policies for patients with BSR[-] and BSR[+] and the temporal changes in attitudes of physicians merit comment. At the time of initial diagnosis on neurological status, most families of the patients with BSR[-] and BSR[+] had willingness to withhold life-sustaining treatment (Figure 3). In patients with BSR[-], however, the ratio of the patients receiving aggressive life-sustaining treatment was less than in those with BSR[+] (6.9% vs. 15.3%, p=0.002). This difference could be ascribed to the impact of impaired brainstem function . When temporal factors (i.e., interval between the diagnosis and actual implementation of treatment policies) were taken into consideration, there existed no serial changes in the proportion of each treatment policy (i.e., aggressive, withhold and withdraw) in patients with BSR[-] (p=0.432). In patients with BSR[+], by contrast, the changes in the proportion nearly attained significance (p=0.072), probably due to an additive impact of a decrease in cases with aggressive life-sustaining support (from 15.3–9.8%, p=0.070) and a modest increasing tendency in the patients who had withdrawal of life-sustaining treatment (from 2.1–4.7%, p=0.127). These distinct effects could be associated with longer hospitalization in BSR[+] cases (i.e., 16 days vs. 3 days, for BSR[+] and BSR[-], respectively, p<0.001).
The implications of withdrawal or withholding of treatment in end-of-life care may be affected by physicians’ attitudes toward life-sustaining treatment. Vincent  previously showed in a questionnaire survey conducted in 16 European countries that 93% of the respondents withheld treatment from patients with no hope of a meaningful life. A recent questionnaire survey on the treatment attitude of physicians in Asia demonstrated that they often withheld but seldom withdrew life-sustaining treatments in ICU patients . This study included 1465 respondents, among whom 224 physicians were enrolled from Japan. The study showed that 70.2% of respondents withheld and 20.7% withdrew life-sustaining treatment. When restricted to Japanese physicians, approximately 90% respondents withheld but only 10% withdrew life-sustaining treatment. The trends observed in these findings are hence compatible with the results of our current study. Collectively, these findings observed in Asia lend support to the contention that there is an unacceptable social background to interrupting medical practice. Close communication between physicians and families and in-depth discussion would offer more fruitful end-of-life care.
The present study shows a striking discrepancy in number between the requests from patients’ families to withdraw treatment and the actual withdrawal by the physicians. Thus, the cumulative number of requests to withdraw life-sustaining treatment increased during a couple of week period (BSR[-]; from 1 to 22 cases, BSR[+]; from 5 to 28 cases, Figure 4). Nevertheless, actual implementation of the withdrawal was much less in both groups (8 and 11 cases for BSR[-] and BSR[+], respectively). The discord between these results may be accounted for not only by personal attitudes of physicians toward treatment policies but also by religious belief of them. Indeed, Asian physicians, including Japanese, tend to desist from discussing with families the withdrawal of ongoing life-sustaining therapy [4,5], which might discourage the families from requesting withdrawal of treatment. Furthermore, ethicolegal circumstance in Japan may deter physicians from interrupting life-supporting treatment. Although the guidelines by the Government and several academic societies qualified the withdrawal/withholding of treatment as end-of-life care practices [11-14], our physicians might hesitate to implement justifiable withdrawal to avoid lawsuits. This important issue should be more thoroughly recognized, and successful end-of-life care needs to be established with nationwide consensus endorsed by ethicolegal frameworks.
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.
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-of-life 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.