This editorial appears in the January Issue of the American Journal of Bioethics
Ariane Lewis et al. have produced a comprehensive and insightful analysis of the tragic case of Adriana Smith, addressing its many medical, ethical, and legal issues, and particularly focusing on the medicolegal controversies arising in managing brain death (death by neurologic criteria) in pregnancy. Their analysis is necessarily constrained by the limited publicly available medical details of the case that are a prerequisite for optimal medical and ethical decision-making. Here, I add context to these issues by presenting a historical perspective that traces the evolution of medical and ethical thinking about brain death in pregnancy beginning with the first case reported in 1983. Then, I briefly comment on several aspects of the bioethical challenges of this vexing situation and highlight several essential scholarly contributions. I leave to others to comment on the complex medicolegal issues that Lewis and colleagues examined.
Historical Perspective
A longstanding obstetrical literature describes and analyzes the use of cesarean delivery to save a viable fetus in the heartbreaking event of maternal death. Understandably, most of this discourse concerns maternal death determination by circulatory criteria in the pre-modern era. The first application of this lifesaving fetal rescue to the emerging state of brain death was mentioned in a 1978 review of postmortem cesarean delivery. The author commented that, while theoretically possible, such an opportunity would be unlikely to occur in practice. A probable reason for this conclusion was that in the 1970s, ICU technology and medications that were available to maintain physiologic support in brain death permitted circulation to continue for only several days after which nearly all patients suffered cardiac arrest irrespective of treatment.
With subsequent developments in ICU technology and medications during the 1980s, cases began to be reported describing “prolonged somatic survival” of some aggressively treated brain-dead patients. One report described a 49-year-old man resuscitated after cardiopulmonary arrest who, over several days, became brain dead, had an isoelectric EEG and diabetes insipidus, but for whom mechanical ventilation and aggressive ICU care were continued. His mechanical ventilatory support was finally discontinued on hospital day 74 after his 4th isoelectric EEG. Postmortem examination showed widespread brain and spinal cord necrosis.
In the first case of successfully managing a pregnancy complicated by maternal brain death, Dillon et al. reported a 30-year-old pregnant woman who became brain dead from meningoencephalitis and who underwent successful cesarean delivery after 24 days of aggressive physiologic support. They proposed basic guidelines to be followed in such cases, recommending the consideration of maternal physiologic support in the presence of a 24–28 week gestation and a healthy fetus who was unaffected by illness, or by the condition that led to maternal death.
In 1985, Heikkinen and colleagues published a case of a pregnant woman with severe brain damage from a ruptured intracranial aneurysm producing coma, brain stem areflexia, ventilator dependence, and EEGs that were either isoelectric or showed only slight cortical activity. Although an apnea test was not performed and, therefore, brain death could not be confirmed, it seems likely that brain death was present given the clinical details. The patient was physiologically supported from fetal week 21 to cesarean delivery of a healthy baby at 31 weeks.
The first detailed ethical analysis of such a case was conducted by the Viennese American physician-bioethicist Erich Loewy in 1987. His thorough and fascinating account of the history of postmortem cesarean delivery includes its practice in Greek and Roman antiquity, the requirement in the Babylonian Talmud to attempt to save the child, and the debate over the etymology of the topic’s name. He pointed out that the term “cesarean” (or “Caesarean”) delivery derived from the Latin word caedere, meaning to cut, and that popular myths claiming that the term instead derived from the method in which Julius Caesar was born are apocryphal, particularly because Caesar’s mother was known still to be alive when he was 40 years old.
Loewy conducted the first comprehensive ethical analysis of whether a brain-dead mother should be physiologically maintained solely to permit the fetal development and live birth of her infant. He concluded that treating the pregnant woman as an incubator to permit a live birth is fundamentally a moral question of whether the harms inflicted to the mother could be justified by the benefits to others, particularly to the infant. He concluded that the benefit of delivering a viable infant justified certain harms to the mother but the greater the duration from the moment of fetal viability that maternal death occurred, the more that maintaining the mother solely as an incubator resembled experimental therapy that required explicit informed consent. He argued further that the emotional and financial costs of long-term maternal incubator treatment demanded community support to the family and that all participants must be sensitive to the ethical issues involved.
In 1988, Field and colleagues provided the second detailed ethical analysis in their description of a brain-dead pregnant woman whose physiologic support was continued for 9 weeks until cesarean delivery at 31 weeks’ gestation. Their article did not cite Loewy’s analysis, possibly because they were unaware of it as it had been published only 9 months earlier. They argued that maternal autonomy should have medical and legal priority over what is done to the fetus but that these “rights” are irrelevant to the decision to use the brain-dead mother as a fetal incubator for a prolonged period. They contended that the long tradition of attempting fetal rescue had been established since antiquity but when applied to chronic physiologic support of the dead mother, the most relevant issues are weighing the harms of treatment against its benefits. This utilitarian balancing required an assessment of the risk of intrauterine death, the risk of premature delivery, and the benefit of delivering a viable and healthy child. Monitoring the health and development of the fetus during treatment could lead to a dynamic shift in the balance of these risks. The risks and benefits to the family also must be weighed, particularly the great financial and emotional costs involved .
In early reports, the longest case of prolonged, successful maternal support in the setting of brain death was for 107 days prior to cesarean delivery. In that case, the pregnant woman’s brain death resulted from a traumatic brain injury and the infant was unaffected (Bernstein et al. . Over the next 15 years, additional cases with ethical and legal analyses were published, most notably those of Frader, Spike, Sperling, and Farragher and Laffey.
Bioethical Considerations
A fundamental question, first raised by Loewy, and most recently addressed by Caplan in his analysis of the Adriana Smith case, concerns whether medical interventions to provide continuous physiologic support to the pregnant woman that allow the fetus to mature and be born alive should be classified as medical treatment or as medical experimentation. Loewy was concerned that using a pregnant woman as a human incubator solely for the benefit of the fetus was ethically analogous to performing experiments on a physiologically maintained brain-dead patient. He argued that physicians in both cases committed the Kantian ethical fallacy of treating a person as an object rather than as an end itself. Loewy concluded that while short-term treatment of the brain-dead mother as an incubator might be defensible as a medical intervention, long-term treatment should require more formal informed consent as would be appropriate for an experiment on a human subject.
Caplan justified his conclusion that this intervention should be considered experimentation by citing four related lines of evidence: (1) there are extremely limited data from the very few cases published over the decades; (2) we have acquired only minimal knowledge to show that such treatment can result in a healthy baby; (3) there is no agreed-upon, standardized protocol for maternal medical management in such cases; and (4) we have acquired only minimal medical knowledge of the impact of prolonged postmortem maternal physiological support on fetal development and outcome. Caplan concluded that this intervention is therefore experimentation and, consequently, there is no unequivocal medical mandate to conduct it. Moreover, it lacks appropriate experimental consent.
Many writers have analyzed the ethical issues raised by these cases, usually invoking familiar ethical principles to evaluate them. But, in these cases, the ethical analysis must first be accompanied by a medical analysis for two reasons. First, the medical condition causing brain death in the pregnant woman (for example, prolonged cardiopulmonary arrest producing diffuse hypoxic-ischemic brain injury) may have produced concurrent profound damage to the fetal brain. Second, the mere performance of the full brain death determination, particularly the apnea test, may harm the fetus as has been recently explained and quantified.
Respecting the autonomy of the brain-dead mother is an obvious but difficult to achieve goal. While many commentators may assume that the pregnant woman would do anything possible to save the life and protect the health of her infant, it does not necessarily follow that she would want to undergo long-term invasive treatment required for the child to be born, particularly if she were not present to raise it. Especially troublesome is the ethical challenge for a lawful surrogate decision-maker to achieve the standard of substituted judgment by accurately discerning the wishes of a young, previously healthy woman who most likely had neither completed written directives nor discussed her treatment wishes in such an unanticipated situation. In most cases, appointing the fetus’s biological father as the primary surrogate decision-maker seems justified because he will be the child’s only surviving parent, he is most likely to accurately represent the pregnant woman’s wishes, and is the person whose wishes the mother would have taken into greatest consideration. But there are obvious limits to the certainty that the prospective father satisfies each of these criteria.
Because the utilitarian ethical analysis, conducted by Lewis et al. and by many earlier commentators, weighed the benefits of maternal physiologic support against its harms to all relevant parties, the length of anticipated treatment—which is directly dependent on the stage of fetal development when maternal brain death occurs—becomes an essential deciding factor. Experts have opined on gestational age limits for such treatment, with some declaring that maternal physiologic support during the first trimester or extending up to 16–17 weeks’ gestation is not justified.
Taking these factors into account, Spike offered operational medical-ethical conditions for proceeding with continuous maternal physiologic support in such cases. Continued support would be offered based on the satisfaction of two conditions: (1) there is clear evidence the brain-dead woman would approve of her posthumous physiologic support until successful delivery, and, if that evidence were absent, that the biological father approves it and it is not contrary to the mother’s known wishes; (2) the pregnancy stage is sufficiently advanced and there is no evidence of fetal injury, thus predicting a good fetal outcome and thereby satisfying the best interest standard of decision-making. Spike added that, if either the birth of a healthy baby was unlikely or if family members whom the mother trusted could not care for the baby, then long-term physiologic maternal support would be difficult to justify.
Despite the rarity and complexity of each case of brain death occurring during pregnancy, and the complications generated by the marked variability in the intrusiveness of applicable state laws, the development of broadly accepted clinical practice guidelines is a prerequisite to the successful management of future cases. Thoughtful interdisciplinary analyses of individual cases, such as of the tragic case of Adriana Smith provided here by Lewis and colleagues, will catalyze the eventual development of such guidelines.
James L. Bernat, MD