In the midst of an increasingly heated debate over physician-assisted suicide (PAS) another option available to patients who are determined to end their lives is receiving serious attention -- the conscious refusal of nutrition and hydration. Patient refusal of nutrition and hydration (PRNH) is hardly new, indeed, virtually all hospice clinicians remember people who came to a point in their illness when they could be described as having "lost their will to live" and who recognized that continued eating and drinking was having an undesired, life-prolonging effect. In the hospice context, death that follows the decision to refrain from food or drink is not usually considered a suicide, however, by choosing to do so these patients were conscious that their death would likely be hastened.
The general impression among hospice clinicians that starvation and dehydration
do not contribute to suffering among the dying and might actually contribute
to a comfortable passage from life. In contrast the general impression among
the public and non-hospice medical professionals is that starvation and dehydration
are terrible ways to die. Scientific support for either viewpoint has been
scanty, yet modern medical practice has reflected an aversion to allowing
a person to "starve to death."
Indeed, during the era in which most hospice providers have trained and practiced,
a patient unable to eat has been routinely treated with a feeding tube; the
option of declining such intervention never having been offered or fully considered.
The symbolic importance of offering food and fluids is well-recognized. While
it has been utilized by people throughout human history, in public discussion
and debates regarding physician-assisted suicide, hospice providers have wisely
avoided suggesting PRNH as an alternative. There has been concern that in
the political arena such a suggestion might appear as a self-serving way to
deny hospice providers' "ultimate responsibility" to the suffering
patient.
But the situation may now be changing. Several recent articles are serving
to dispel fears of suffering and are making it more acceptable to speak more
openly about this inherent ability of patients to influence the timing of
their demise. Late in 1993 an article entitled, Patient Refusal of Hydration
an Nutrition: An Alternative to Physician-Assisted Suicide or Voluntary Active
Euthanasia, by Bernat, et. al. in the Archives of Internal Medicine reviewed
the salient clinical literature and discussed the ethical implications of
this option.[Bernat] The authors include PRNH within the ethical category
of "voluntary passive euthanasia" since it involves not only the
refusal of oral food and fluid but the associated refusal of non-oral (enteral
or parenteral) alimentation and hydration. They assert that the critical moral
and legal distinction to be made regarding a life-ending decision is not whether
it involves an act of commission or omission on the part of the physician,
but whether or not it constitutes the refusal of a medical therapy by a competent
patient. Patient refusal of nutrition and hydration meets this criteria and,
thus, can be considered among the commonly accepted practices of patient-initiated
refusal (or withdrawal) of mechanical ventilation, renal dialysis, or antibiotic
use.
A more extensive review of the scientific literature relevant to starvation
and dehydration appears in an article by Sullivan entitled, Accepting Death
without Artificial Nutrition or Hydration.[Sullivan] Published studies of
healthy volunteers report that total fasting causes hunger for less than 24
hours. Ketonemia occurs and is associated with relief of hunger and an accompanying
mild euphoria. When ketonemia is prevented by small feedings hunger persists,
explaining the obsession with food commonly observed during semi-starvation
occurring in times of famine or war. Animal studies also suggest that ketonemia
may have a mild systemic analgesic effect. Experimentally induced dehydration
in normal volunteers may report thirst, yet this sensation is consistently
relieved by ad lib sips of fluid in cumulative volumes insufficient to restore
physiologic fluid balance. One study of healthy subjects suggests there is
a decrease in the severity of experienced thirst associated with older age.
Recently, two important clinical studies have been published which investigate
the effects of fasting and dehydration in the patient population relevant
to hospice. Burge surveyed patients at two inpatient palliative care units
in Canada. Visual analog scales (100-mm) were used to assess seven symptoms:
thirst, dry mouth, bad taste, nausea, pleasure in drinking, fatigue and pain.
[Burge] Reported symptoms were studied in relation to potential confounding
variables. Thirst was considered to be the principal outcome of interest.
The mean symptom rating for thirst was 53.8 mm. In multiple linear regressions
no statistically significant association between thirst and fluid intake,
serum sodium, urea or osmolality was found. The presence of oral disease yielded
the most significant association between thirst and examined variables.
In the October 26 issue of JAMA, McCann, et. al. report a prospective study
of patients in a ten-bed "comfort care unit" located within a long-term
care facility. [McCann] The physical care described closely resembles benchmark
hospice care. "Food was offered and if necessary fed to patients but
was never forced. All patients received meticulous mouth care that included
combinations of cleaning, various swabs, ice chips, hard candy, and lubricants.
Narcotics were used for most of the patients to treat symptoms of pain and
shortness of breath when present. The dose of narcotics was titrated to provide
pain relief while avoiding sedation. When the window of providing pain relief
and causing sedation was small, the patients' wishes were weighed regarding
the discomfort of pain vs. the discomfort of sedation in determining subsequent
doses and intervals of narcotic administration."
Of 32 patients assessed by this group during a 12 month period, 63% denied
hunger entirely, while 34% reported hunger during only the initial part (first
quarter) of their course in the unit. Thirst or dry mouth was reported by
66% of patients; 28% transiently and 38% throughout the terminal phase of
their illness. Thirty-four percent denied either symptom. The authors found
that in all patients reporting either hunger or thirst, these symptoms were
consistently and completely relieved by oral care or the ingestion of small
amounts of food and fluid. While patients could eat or drink ad lib, the amount
of food or fluid ingested -- and documented to relieve associated symptoms
-- was consistently less than that required to correct dehydration or to meet
obligate fluid and energy requirements.
Perhaps the most persuasive of recent articles is that entitled, A Conversation
with My Mother.[Eddy] It is a narrative written by Dr. David Eddy regarding
the progressive illness and dying of his mother. Initially published in the
Journal of the American Medical Association, it was subsequently reprinted
in the New York Times, eliciting substantial discussion and notably favorable
public response. Mrs. Eddy was suffering from progressive debilitation, chronic
depression, anemia, recent surgery and recurrent rectal prolapse. She expressed
a desire to die and, in the course of relentless decline, asked her son for
help. Dr. Eddy sought to provide his mother with the means to end her life
peacefully. However, prior to obtaining a lethal prescription, she developed
pneumonia and was hospitalized. Antibiotics were begun (we are not told why),
but quickly withdrawn at the patient's request. When she began to improve
despite the lack of life-prolonging intervention, Mrs. Eddy asked her son
about the option of refusing food and fluids. (It was her idea.) He assured
her that without nutrition and, especially without adequate fluid, the end
would come quickly. She was elated and, following the celebration of her 85th
birthday and with the support of her primary physician, she stopped eating
and drinking. (Her last morsel was chocolate.) She died, peacefully, six days
later. The description of her last few days is compelling. "Over the
next four days, my mother greeted her visitors with the first smiles she had
shown for months. She energetically reminisced about the great times she had
had and about things she was proud of... She also found a calming self-acceptance
in describing things of which she was not proud. She slept between visits
but woke up brightly whenever we touched her to share more memories and say
a few more things she wanted us to know. On the fifth day it was more difficult
to wake her. When we would take her hand she would open her eyes and smile,
but she was too drowsy and weak to talk very much. On the sixth day, we could
not wake her. Her face was relaxed in her natural smile, she was breathing
unevenly, but peacefully. We held her hands for another two hours, until she
died."
Sullivan's aforementioned review of the literature also begins with a clinical
narrative case presentation. The case is that of a 78 year old woman with
recurrent, complete small bowel obstruction, following treatment of endometrial
carcinoma. Although no evidence of recurrent or metastatic disease could be
found, the patient steadfastly refused further surgery. While accepting a
nasogastric tube for decompression, she declined any nutritional support and,
after an additional 14 days, also refused intravenous fluids. The patient
judged her symptoms, which included nausea and ascites, to be intolerable.
Sullivan states, "She repeatedly requested that her life be ended by
injection of a lethal dose of morphine. This request was respectfully declined
by her physician, who did offer to relieve any pain or discomfort." Eventually,
morphine was administered at the request of the patient "to relieve boredom
and help with sleep." The patient lived a surprisingly long time; a total
of 42 days of complete fasting and 29 days with minimal fluids. At no time
did she report pain. Her sensorium remained clear and affect normal until
the last day of life. "Her friends visiting during this period uniformly
confirmed that she was a woman who knew her own mind and had the courage of
her convictions."
Comment
Critical
examination of this latter case gives rise to a number of questions. Seasoned
hospice clinicians will wonder whether sophisticated palliative care would
have uncovered other options available to this patient. However, two points
of relevance to the current discussion stand out: a) neither starvation nor
dehydration seemed to contribute to this patient's physical distress, and
b) the patient lived for an unexpectedly long time. In addressing the prolonged
course of his patient, Sullivan emphasizes that she was not known to have
an underlying terminal illness or any ongoing catabolic process. Additionally,
she began her fast with apparently normal nutritional stores (she is described
as having been "always mildly obese") and her refusal of fluid followed
two weeks of intravenous hydration.
This patient's condition contrasts with the large majority of hospice patients
who are likely to undergo a shift from adipose to protein metabolism much
earlier in the course of fasting. This shift coincides with a decrease in
metabolic free water and renal function and a rapid progression to uremia,
electrolyte imbalance, somnolence and demise. A more prolonged course might
be expected for a person in the circumstance of a massive stroke or post-hypoxic
syndrome whose advanced directive prohibits non-oral nutritional support or
hydration.
In reviewing the literature on thirst associated with negative fluid balance
derived from studies of healthy volunteers as well as studies involving those
with far advanced illness, I have gained an impression that the word "thirst"
may well refer to two distinct sensations. In the patients whom hospices serve,
"thirst" most often refers to a sensation of dryness in the mouth
and throat rather than an experienced need to ingest a volume of fluid which
it normally conveys. This explains why the "thirst" of persons with
far-advanced illness is consistently relieved by amounts of fluid insufficient
to expand volume or osmotic receptors.
Risks and Potential Benefits of Starvation and Dehydration in Far-advanced
Illness
While the currently available research is not exhaustive, hospice clinicians
now possess respectable scientific data to supplement anecdotal experience
in forming their own opinions, and informing patients about nutrition and
hydration. This information has relevance beyond responding to a hospice patient
who is expressing suicidal intent. When diminished gag reflex, altered sensorium
or obstruction makes oral feeding impractical or risky, thoughtful clinicians
and their patients often struggle with decisions about whether to place nasogastric
or PEG feeding tubes.
Medical decision-making rest upon the pillars of informed consent and the
concept of proportionality. The publications cited collectively expand the
information base on which clinicians can make recommendations and patients
(and families) can make informed choices. Proportionality is commonly explained
as a weighing of the risks versus the potential benefits of a proposed intervention.
In the context of the present discussion, the principle of proportionality
can be understood more fully as requiring the weighing of the risks and potential
benefits of each available intervention against the risks and potential benefits
of non-intervention. In the process of informing and obtaining consent from
patients -- and, perhaps especially, in discussions with the legal surrogate(s)
of an incapacitated patient -- it must be remembered that the decision to
prevent malnutrition or dehydration is a de facto decision to have the person
die of something else. Stated another way, the euvolemic, nutritionally supported
patient may live longer, but is probably more likely to die from acute systemic
infection, from acute respiratory failure, from acute cerebral or myocardial
ischemia or from abrupt blood loss.
These clinical reviews, case reports and new research data lend credence to
the clinical impression that, among the terminally ill, the risks of uncorrected
malnutrition and dehydration are few. (In this circumstance, death is not
properly regarded as a risk, since it is a principal expected outcome.) Considered
together, they allow us to state that, at least within the context of adequate
palliative care, the refusal of food and fluids does not contribute to suffering
among the terminally ill. The literature is consistent on two points: a) rarely
does fasting cause any discomfort beyond occasional and transient hunger,
and b) symptoms referable to dehydration are few -- mostly dry oral and pharyngeal
mucous membranes -- and are readily relieved by simple measures.
Yet, while it may now be appropriate to present the risks of starvation and
dehydration within a hospice setting as being minimal, this is logically distinct
from being able to assert that there is tangible benefit from either condition.
From a patient's perspective, I suspect that the critical, intangible benefit
of PRNH will be an improved sense of confidence that death will occur peacefully.
As noted, this notion is supported (and probably fostered) by the commonly
expressed observations of hospice clinicians that such deaths are generally
preceded by a gentle, deepening somnolence with the person often described
as having "slipped away". What is known of the physiology of far-advanced,
progressive disease and of combined nutritional and fluid deprivation suggests
a synergism that, indeed, might explain a peaceful pattern of demise. This
may occur through gradual organ dysfunction (especially renal and hepatic)
with progressive metabolic derangements or via primary circulatory failure.
In either case, a clouding of sensorium can be expected with an arrhythmia
likely being the terminal event. While currently available studies do not
refute this clinical impression, neither are they sufficient to substantiate
it. Thus, while the heightened probability of a gentle passing by PRNH is
intriguing, at present, it remains speculative.
Another potential benefit merits discussion. At least for some persons, starvation
does correlate with reported euphoria. While there is a tendency for medical
professionals to dismiss euphoria as an aberration of consciousness -- a non-distressing
"symptom" -- it clearly has positive impact on a person's subjective
quality of life. Quality of life, it should be remembered, is a wholly subjective
construct. It has emerged as one of the "gold standards" of outcome
measures for hospice care. The question of whether euphoria is somehow "valid"
or "reality-based" may prove irrelevant to its implications for
hospice practice. Further prospective physiologic studies with clinical correlates
of affect and meaningful quality of life assessments should be actively encouraged.
Ethical and clinically effective response to requests for assistance in
suicide
The practice of a physician or nurse responding to a patient's request for
assistance in suicide by discussing the option of refusing to eat or drink
appears to stretch the "fine line" of ethical practice to the point
of disappearing. It will seem chilling to some, and self-absorbed hair-splitting
to others.
Perhaps all patients with essentially normal mental status who are aware that
they are dying engage in some degree of suicidal ideation. The request from
a patient to their doctor or nurse for assistance with suicide has many meanings,
at a minimum it must be heard as a call for help. It may represent an urgent
plea for symptoms to be better controlled or for the patient to be listened
to as a person -- for their suffering to be heard. In practice serious requests
for assistance with suicide commonly prove to be an opening to deeply important
clinical work. Thus, while suicidal ideation is common, suicide among hospice
patients is rare.
In a perfect world resources and expertise would be unlimited and this type
of therapeutic response -- combining effective symptom control with expert
counseling and an authentic healing presence -- would suffice. However, in
the imperfect world in which we practice, not every patient achieves an acceptable
quality of life, nor is every patient willing "to work through"
his or her suffering. Some people insist that they have "had enough"
and continue to plead for assistance with suicide. The information reviewed
has pragmatic application to a response that remains ethically consistent,
effective and genuinely caring.
Ethically, while a clinician may decline to actively assist a person's suicide,
we must never abandon our patient. Even a patient who is intent on suicide
continues to deserve our care. At a minimum the patient requires continued
attention to control of symptoms. Further, the patient remains entitled to
accurate medical information about the options available to them. Undeniably,
one such option is the refusal of further food and drink. But it is, indeed,
a fine line. Unlike physician-assisted suicide, refusing to eat or drink is
a purely personal act. While it may require information, the decision obviates
the need for physicians, nurses or other agents of society to participate.
After adequate discussion, and in the context of continued caring, at some
point the patient's choice becomes "none of our business".
Seeking a clear philosophical distinction, between the scenario of one patient
swallowing a prescribed lethal potion and the scenario another refusing to
swallow anything at all, may be an exercise in splitting ever-finer ethical
hairs. Yet, the interpersonal meaning of these two acts could not be more
different. In every circumstance -- legal or not -- acceding to a patient's
request for a lethal prescription entails a complicity on the part of the
clinician. In its meaning the act is a collusion in the patient's belief that
their situation is hopeless and that their existence is beyond conceivable
value.
In my own practice, while I steadfastly refuse to write a prescription with
lethal intent or otherwise help the patient commit suicide, I can share with
the patient information that he or she already has the ability to exert control
over the timing death. Virtually any patient with far-advanced illness can
be assured of dying -- comfortably, without any additional physical distress
-- within one or two weeks simply by refusing to eat or drink. This is less
time than would be legally imposed by waiting periods of assisted suicide
initiatives. The discussion and subsequent decision are wholly ethical and
legal, requiring no mandated psychiatric evaluations, attorneys, court decisions
or legislation. On the level of meaning there is wisdom in not discussing
the option of PRNH too soon. The impetus for the discussion should arise (as
it did for Mrs. Eddy) from the person dying.
Clinically, for a number of people at the very end of life, the decision to
refuse food and fluid may not arise from depression or emotional denial as
much as from a felt sense of "being done". Most such persons I have
encountered one way or another expressed a sense that eating or drinking were
no longer relevant to their situation. They were far along in a process of
withdrawal, having turned their attention inward or "beyond". Even
here the option of PRNH has important advantages over complying with a patient's
request to be killed, for it allows the clinician's attention to remain focused
on relief of suffering -- physical, psychosocial and spiritual. It requires
-- of frees -- the clinician to remain vigilant for treatable depression and
to remain, in humility, open to the possibility of unexpected opportunities
for the person to again discover value in the life that is waning.
The powerful symbolism of nurturing associated with feeding and the notion
of suffering associated with starvation and dehydration are deeply rooted.
This is perhaps especially true for those of our patients who lived through
world wars and the Great Depression of this century. While the topic of PRNH
may no longer be taboo, it must be approached with extraordinary sensitivity.
However, discussion of what we know can help reduce anxieties of what we fear.
The more we know and the more confidently we know it, the better able we are
to make sound decisions. My experience is that the information presented has
been well received and resulted in a noticeable allaying of fears. This has
been the case in private discussions, as well as in public and professional
forums. The recognition that patients with far-advanced illness have always
had control over the timing of their demise can enable the focus of discussion
and intervention to remain fixed on the goals of comfort and quality of life.
The debate over physician-assisted suicide will continue. The option of PRNH
deserves fuller attention and discussion. Years ago Cicely Saunders defined
a core principle for hospice, "You matter because you are you. You matter
to the last moment of your life, and we will do all we can not only to help
you die peacefully, but also to live until you die." In our statements
and our actions hospice clinicians would do well to reflect on those words.
References:
Bernat, JL, Gert, B, Mogielnicki, RP, Patient Refusal of Hydration an Nutrition: An Alternative to Physician-Assisted Suicide or Voluntary Active Euthanasia, Arch Intern Med Vol 153, Dec 27, 1993 pp2723-2728
Sullivan, RJ. Accepting Death without Artificial Nutrition or Hydration J Gen Intern Med, Vol. 8, April, 1993, pp 220-224.
Burge, F, Dehydration Symptoms of Palliative Care Cancer Patients. Journal of Pain and Symptom Management, 1993;8(7) pp454-464.
McCann, RM, Hall, WJ, Groth-Juncker, A, Comfort Care for Terminally Ill Patients: The Appropriate Une of Nutrition and Hydration, JAMA October 26, 1994 272(16) pp1263-1266
Eddy, D. A Conversation with My Mother, JAMA July 20, 1994 Vol. 272, No. 3 pp179-181