Article Text

Download PDFPDF

Withholding artificial feeding from the severely demented: merciful or immoral? Contrasts between secular and Jewish perspectives
  1. J Kunin
  1. PO Box 487, Zichron Yaakov 30900, Israel; jkunin{at}

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    According to Jewish law, to make a judgment that a life has no purpose and is not worth saving is contrary to the concept of justice

    Traditional medical practice dictates that when patients are unable to eat or drink enough to sustain their basic nutritional requirements, artificial feeding and hydration is indicated. Common clinical examples of this problem are patients with senile dementia and those in a persistent vegetative state (PVS). In recent decades, however, the practice of mandating artificial feeding has been increasingly questioned. A combination of legal, ethical, and clinical considerations has resulted in broad support for withholding and withdrawing artificial nutrition. The guiding ethical principle in the current clinical standards is that patient autonomy must be honoured. In the context of an incompetent adult (such as a patient with advanced dementia or in a PVS), advance directives or surrogate decision making are legally binding. Such requests to withhold artificial nutrition are considered appropriate and even encouraged.1 Such a view, however, is not unanimous. For example, Catholic writers have questioned the current consensus. The New Jersey Catholic Conference has written that withholding nutrition and hydration from such patients “ultimately results in starvation, dehydration, and death,” and that withdrawing such basic care from patients who are not dying but in a PVS “is a clear statement that the patient’s life has no moral value”.2 This question is also of great concern in Jewish law (halachah). In contrast to secular medical ethics, halachah requires that artificial feeding be given to patients with dementia or in a PVS. The following discussion gives an overview of the pertinent arguments that have led to the consensus in secular ethics that artificial feeding may be withheld from the severely demented and the halachic considerations that argue against this practice.


    The most obvious clinical objective in artificially feeding demented patients is to sustain their lives by preventing starvation. In recent years, however, some authors have argued that severe dementia is a terminal condition and that artificial nutrition does not extend life. As Gillick writes, “There is a pervasive failure to view advanced dementia as a terminal illness”. She argues that eating and swallowing disorders in a demented patient are “independently associated with mortality”.3 This declaration, however, is not supported by available data. In one study of demented patients who were totally dependent on others for eating, the one year survival rate was 80%.4 In another study, Mitchell et al, reported that in demented patients with eating and swallowing disorders 86% were alive at one year.5 These statistics do not support the notion that demented patients with eating and swallowing disorders are terminally ill. Indeed, Meyers and Grodin write that severely demented patients “are neither terminal nor suffering from a painful condition”. They continue: “Severe dementia is not predictive of eating dependency once abnormal oral motor behavior was considered. This suggests that the neurologic changes in swallowing that are seen in dementia may be variable in timing and presentation.” Meyers and Grodin conclude: “Elderly patients with dementia are similar to persistent vegetative state (PVS) patients in that neither group are terminal. Both these groups can be expected to live years if tube feeds are supplied.”6

    It has been argued3,7 that there is no evidence that tube feeding prolongs life. This unexpected conclusion is based largely on two retrospective studies. In a 1997 report by Mitchell et al,4 patients were studied who were totally dependent on others for eating. Patients who were tube fed and those who were hand fed had similar one year survival rates of 80%. Their conclusion was that “tube feedings did not affect survival in this frail population”. A serious weakness of this study was that they specifically excluded those who were in the most severe category of dementia (the patients who were unable to eat by mouth, even by spoon feeding). In other words, the investigators excluded those most likely to benefit from tube feeds and most likely to die from starvation if not artificially fed. In another widely quoted paper5 a large population of demented patient with eating and swallowing disorders was studied. The results showed that those patients without tube feeds had an 86% one year survival rate and that those patients who were tube fed actually had a 10% higher mortality rate. This study is cited by some authors to support the conclusion that there is no proof that artificial feeding prolongs life. This study suffers from being retrospective and cannot control for confounding factors that may make the two study groups not comparable. Moreover, the 86% one year survival rate in those who did not receive tube feedings implies that these patients, in fact, were able to eat despite their swallowing disorders and were likely not to have been comparable to the patients requiring tube feedings. In summary, no conclusion can be drawn from these studies regarding the effect of tube feedings on longevity. The cited studies do not address the group of patients that cannot eat by mouth, even by spoon feeding, and would starve if it were not for artificial feeding.

    In fact, there is widespread acknowledgment that withholding nutrition shortens life. A consensus statement of the American Academy of Neurology asserts that “withholding life-sustaining therapy may shorten life”.8 Peck et al, write that “the terminal phase of dementia may be prolonged for months or even years” if supportive care such as artificial nutrition is administered.9

    It has been argued that PEG (percutaneous endoscopic gastrostomy) tubes cause suffering manifested by high complication rates, that they require restraints, and that they lead to a more uncomfortable death. In a review of the literature, Larson et al, found that in 314 PEG placements, three deaths were attributable to the procedure itself—one from laryngospasm and two from aspiration.8 While any operative mortality is of great concern, a 1% mortality for a procedure that may indefinitely prolong life is halachically considered an acceptable risk (Rabbi Yaakov Weiner, written communication, 2002). One can think of palliative surgical procedures such as gastric and biliary bypasses in terminal cancer patients that are associated with higher operative mortality and a lower long term survival but which are well accepted medical procedures. Moreover, the high 30 day complication rates in all series of PEG placements are overwhelmingly related to the patient’s underlying medical conditions and not to the PEG placement itself. The most commonly cited PEG related complications are infections (which were reduced from 30% to 7% with the use of prophylactic antibiotics) and tube displacement.10 This latter complication is minor and only requires replacement of the gastrostomy tube—a very simple, painless procedure.

    A widely quoted concern is that tube feedings result in greater use of restraints, and that this, in turn, adds mental suffering to these fragile patients. The source for this assertion is a report from Peck et al.11 A careful reading of this report demonstrates, however, that there was no statistically significant difference in use of restraints between tube fed and non-tube fed patients. Moreover, 78% of those patients who were given artificial nutrition were fed by nasogastric (NG) tube. The PEG technique is a technological advance that makes artificial feeding more comfortable for patients. One would expect restraints would be less necessary with PEG tubes than with NG tubes.

    In conclusion, the available literature demonstrates that severe dementia is not a terminal disease, that there is no data to argue against the assumption that artificial nutrition will prolong life in patients who are unable to eat orally, that complications rates are not excessive with PEG tubes, and that there is no added suffering from tube feedings.


    The argument that withholding tube feedings in demented patients is permissible is based on the idea that withholding medical technology and allowing nature to take its course is morally acceptable. The assumption is that these patients can be kept alive, but there is nothing wrong with allowing them to die by starvation, particularly when advance directives have been established.

    Court decisions have led the way in not mandating tube feedings in PVS and demented patients. The principle of autonomy weighs heavily in these legal decisions. As Steinbrook and Lo have stated: “[The] right to refuse artificial feeding outweigh[ed all the state’s] interests”.12 One court summed up this principle by writing: “The right to refuse medical therapy is virtually absolute and patient motives were not a matter of debate”. The Council of Ethics of the AMA has stated: “It is not unethical to discontinue life prolonging therapy in patients with PVS”.13 In such circumstances (severe dementia being similar), the council argues, artificial feeding is extraordinary care and similar to cardiopulmonary resuscitation (CPR) and mechanical ventilation. It is important to emphasise that no distinction is made in these statements between terminal and non-terminal disease. In Jewish law the distinction is important. The question of whether to initiate endotracheal intubation and mechanical ventilation in a terminally ill cancer patient and whether to insert a PEG tube in a demented patient without any terminal illness are not considered comparable cases.

    All ethical arguments in favour of withholding nutrition assume that demented patients are qualitatively different from other patients. Lo and Dornbrand argue that severe dementia is the end of a progressive downhill course and “stopping to eat is a natural letting go of life”.14 Hanson et al, emphasise the concept of medical futility. They declare that when a favorable outcome “offers the opportunity to achieve life goals” it is not futile. Their conclusion is that such states as PVS and dementia exclude the possibility of achieving “life goals” and so any intervention to lengthen the lives of such patients is by definition futile.15 The core of their argument is that these patients are not sentient beings, and merit less care. So, too, Rabenick et al, have stated:

    If the intervention is only expected to improve the patient’s physiological effects, but does not have any effect on quality of life such as in a patient with PVS, then the physician should recommend against such interventions. The minimal function of PVS patients eliminates the possibility of experiences and activities that make human life valuable—that is, they cannot experience quality of life. Medical interventions that only sustain minimal function do not benefit the patient overall.16

    While autonomy has great legal weight, several authors argue that the case against nourishing such patients is so strong that it should override the principle of autonomy. Gillick has stated: “If family members are unable to make a decision, the physician should assume that a person with advanced dementia would not want a feeding tube”.3 Schneiderman et al, write:

    We believe that qualitatively poor results fall outside the range of patient autonomy and need not be offered as options. The clearest of these qualitatively poor results is continued biologic life without conscious autonomy. The patient has no right to be sustained in a state in which he has no purpose other than mere vegetative survival.17


    The current standard that allows for the withholding and withdrawing of artificial nutrition is not universally accepted. Rosin and Sonneblick, geriatricians reporting from a Jerusalem hospital that is guided by halachic principles, state that in their practice artificial feeding is given to patients with senile dementia and is considered proper, moral care.18 In an article surveying general attitudes among health care providers towards artificially feeding demented patients, Israelis were more in favour of force feeding than a similar survey in Sweden. The authors concluded that the Israeli attitude was based on the Jewish tradition that gives a very high priority to preservation of life.19

    A halachic ruling on the advisability of placing a PEG tube in a severely demented patient depends on several practical considerations. These include the prognosis of the disease, the physical and mental suffering the patient must withstand, and the efficacy of the PEG tube in achieving its goal of prolonging life. What is specifically not a consideration in the calculus is the perceived quality of life of the patient.

    Rabbi Yaakov Kanevsky, a leading rabbinic figure of the 20th century, has written20 that the first halachic principle of medical intervention is that whenever it is possible to increase the longevity of a patient, it should be done. A cardinal principle of Judaism is that every moment of life is considered precious and of infinite worth. He continues by warning that while there are exceptions to this general rule, they must be applied with great caution and only on a case by case basis. In other words, the halachic premise is always in favour of treating and promoting life. The exceptions to this rule usually involve persons known to be terminally ill and who are suffering. Both of these conditions must be met. Neither alone is sufficient to withhold treatment. Terminal illness in halachic terms means there is a near certainty that the patient will die soon. The exact time frame is a subject of dispute. The most influential living halachic authority, Rabbi Yosef Shalom Elyashiv has stated this to be a life expectancy of no greater than three months (Rabbi Yaakov Weiner, written communication, 2002). It is important to underscore that in rendering a halachic decision, the definition of terminal illness is not related to its natural history, but includes the prognosis, taking into account all the resources of modern medicine. A natural history of a disease independent of currently existing treatment is of no interest to the halachic judge. That a severely demented patient who cannot eat will “naturally” die from lack of nutrition in a “natural letting go of life,” is an artificial construct, which ignores the reality that artificial feeding will allow him to live on indefinitely.

    When a person is determined by medical opinion to be terminally ill despite all medical treatment and he is suffering, initiating new treatment is not required if it only minimally increases life expectancy. Rabbi Moshe Feinstein, one of the great halachic authorities of the last half century, has written that added suffering outweighs the small benefit of a minimally increased life expectancy, and the treatment should not be given.21 Given the data as outlined above, however, the severely demented patient who does not eat cannot qualify for this important exception to the general principle that life must be supported and promoted. As mentioned earlier, there is no medical evidence that such patients are terminally ill. To the contrary, most authorities specifically state that the severely demented are, in fact, not terminal and are not suffering from a painful condition.

    Regarding the efficacy of tube feeds, the halachic judge must weigh the complications versus the benefits of the procedure. The benefits, while not well quantitated clinically, are simply to prevent starvation in a patient who is unable to eat. In other words, there is a presumption that feeding a patient who is getting insufficient nutrition will certainly increase their life expectancy and convert them from what would be a terminal state of starvation to a life that is now indefinitely prolonged. Rabbi Feinstein specifically stated: “When a patient is currently terminal, but can be given medicine to make them non- terminal and live indefinitely, the therapy is required”.22 Moreover, in analysing the risk benefit analysis of tube feeds, the complication and mortality rates of PEG tubes are very low considering the likelihood that artificial nutrition will increase the patient’s longevity.

    It is important to note that Rabbi Feinstein adds that physical coercion should not be used to force patients to accept therapy. He limits his discussion to the competent adult. One may extrapolate from his concern about coercion that if a severely demented patient refuses tube feedings by repeatedly pulling out tubes, physical restraint is a most undesirable solution. No generalisations, however, can be offered here, and, as always in Jewish law, such problematic cases can only be judged by individual rulings. It would appear, though, that such a refusal in an incompetent patient does not necessarily mandate cessation of artificial nutrition. Such a situation may require adjustment of medications to control agitation or depression. It is clear, however, that physical restraints should be avoided.

    For advocates of withholding artificial nutrition the outstanding question is the value of the life of a severely demented patient. This concept is sharply opposed by halachic authorities. Professor Abraham B Abraham, a pupil of the great halachic authority, Rabbi Shlomo Zalman Auerbach, writes:

    Every portion of a life of a man has infinite worth—70 years of a man has exactly the same value as 30 years, as one year or one hour. The infinite value placed on human life [means] that if the value of the end of the life of a man is reduced, the absolute value of his life has been lost and one has made the value of his life relative—relative to his usefulness, the condition of his health, his benefit to fellow human beings, or by any other measure. Then, there will be a need to grade men and two men are never equal in their value. The moment that the life of a man becomes defective and he may soon die, his value deteriorates . . .. There is no limit to what becomes possible in limiting the value of men.23

    The Jewish argument, though, is not just out of concern for the slippery slope. Judging any human life as more or less valuable is categorically rejected. As Rabbi Auerbach writes:

    Even though it is clear that [such a] life is not life as we understand life to be, and it is extremely difficult for both the patient and family to withstand such suffering, we are none the less commanded to make every effort to preserve and save his life . . . . We are commanded [to save the life of] a sick old man, a deaf mute, one completely retarded and incapable of performing [any useful function]. Even in such circumstances it is of great [merit] to make every effort to pursue saving his life.24

    Rabbi Feinstein also addresses the issue of quality of life explicitly:

    There is no place to consider quality of life. Considering a retarded person or one in a vegetative state there is no place to think there is not the same mandate to treat such patients. When there is no suffering and they are able to live a long time it is clear and simple that they require the same care as anyone and no regard is paid to their sentience.21

    In measuring a man’s life by its quality or considering it worthwhile only when he is able to pursue certain goals, ethicists are making sweeping judgments. If ethicists, judges, physicians, or family members have the authority to determine which life is worth living, they are able to decide which life is worth saving and which is not. This is extraordinary power. To make the moral dilemma more difficult, there may be unconscious bias in allowing the severely demented to die from starvation. Rango has stated:

    As a group, the impaired elderly are a potent source of disagreeable and even hateful mental images capable of generating fear, anger, and guilt in the minds of those who are chronically exposed to the ravages of geriatric disease . . . . Why do these shadows of former selves refuse to die? Why do these vacant minds and misshapen bodies incapable of living alone continue to burden the lives of others? By appearance and behavior, the demented elderly sorely challenge one’s sense of compassion and benevolence. Senility itself poses a terrible psychic threat reminding one of the dark human potential to lose personal identity, awareness, and self control in old age. An insidious mental process can thus unfold if one begins to project these feelings of personal anguish to the demented person who is then regarded as having a diminished quality of life. This form of psychological projection can lead to a willingness to end the demented person’s life for his own good even though observable evidence of great pain and distress is lacking.25

    One exception to the judicial trend to allow a severely demented patient to die was the New Jersey Supreme Court rejection of a request to remove the feeding tube in the case of Conroy. In their summary they stated that it was not proper to terminate care on “assessments of the personal worth or social utility of another’s life or the value of that life to others”.26 Thus there is a concern that the manner in which “policy choices and clinical decisions are now being framed cuts against the best interests of the disabled, the sick, and the dying”. Rango concludes: “Recognising their mental incapacity, we ought to devise stringent ethical standards that protect persons with dementing illnesses from discrimination in its various forms”. Patients with severe dementia as well as with PVS are arguably the most vulnerable members of society, with few advocates, and no political power.

    It has become increasingly acceptable today to judge the life of some of our fellow human beings, such as those with severe dementia and PVS, as less worthy of preserving. Rabbi Shlomo Zalman Auerbach, who typified the ideal rabbinic personality of unparalleled intellectual power and scholarship, combined with renowned humility, wrote: “We have no ability to determine the preciousness and the importance of any life, even when that person is incapable of [performing any useful function]”.24 To think otherwise, to presume that we have the authority to decide who we will sustain and who we will allow to die, is a kind of arrogance that is an anathema to the benevolent human society that is at the core of what Jewish law and ethics promotes.


    I am grateful to Rabbi Yaakov Weiner of the Jerusalem Center for Research in Medicine and Halachah who reviewed the manuscript. It is with his counsel that I have attempted to present the accepted halachic consensus on this subject. Any error or misrepresentation in this regard is, however, because of my own misunderstanding.

    According to Jewish law, to make a judgment that a life has no purpose and is not worth saving is contrary to the concept of justice