Anesthesiology Clin N Am

24 (2006) 129 144

Ethical Issues in Palliative Care

Richard A. Demme, MDa,T, Eric A. Singer, MD, MAb,
Jane Greenlaw, JDc, Timothy E. Quill, MDc
a
Nephrology Division, The Center for Palliative Care and Clinical Ethics,
University of Rochester Medical Center, 601 Elmwood Avenue, Box 675, Rochester, NY 14624, USA
b
Department of Urology, The Center for Palliative Care and Clinical Ethics,
University of Rochester Medical Center, 601 Elmwood Avenue, Box 656, Rochester, NY 14624, USA
c
Division of Medical Humanities, The Center for Palliative Care and Clinical Ethics,
University of Rochester Medical Center, 601 Elmwood Avenue, Box 601, Rochester, NY 14624, USA

Ethical concerns are common in palliative care settings. Rather than provide
an exhaustive list of possible ethical problems one may come upon, this article
describes areas of concern that are frequently encountered by perioperative health
care providers, especially anesthesiologists, in the palliative care arena.

Consent and capacity

Informed consent

Informed consent is a term familiar to all physicians, but it has particular

significance for those who regularly perform procedures that are risky or
invasive. It evolved as a legal doctrine that clarifies the obligations of physicians
toward their patients. Informed consent represents the recognition that physicians
who are proficient in diagnosis and treatment could nonetheless be found
negligent for failure to provide information that patients needed to decide whether
to accept recommended medical treatment. The adequacy of the disclosure of
information provided to the patient is measured against the standards of what a
reasonable physician with a similar background and training would tell the

T Corresponding author.
E-mail address: [email protected] (R.A. Demme).

0889-8537/06/$ see front matter D 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.atc.2005.11.004 anesthesiology.theclinics.com
130 demme et al

patient under the circumstances and what a reasonable patient would need to
know to decide whether to undergo the proposed procedure [1].
What information should be provided? In the early days of the informed
consent doctrine, physicians were expected to give a description of the proposed
treatment, inherent risks, expected benefits, and common complications. How-
ever, later legal developments have expanded the information that should be
provided to include a description of alternative treatments, an explanation of the
likely consequences of refusing treatment, an explanation of the diagnosis, and
the expected prognosis. A controversial question raised recently is whether the
physician should disclose personal complication rates, particularly if the rates
differ significantly from those generally experienced by other physicians per-
forming the same procedures. The trend is for courts and legislatures to require
disclosure of any information that, if known by the patient, would likely affect
the patients decision.
An important message about informed consent is that it involves a process,
not merely a form. The legal requirement is that there is a process of disclo-
sure and discussion giving information to the patient and answering the patients
questions. Consent forms are used in most institutions as a way of documenting
that the process has taken place. It is sound practice to document the process in
the progress or office notes, as well.
To be valid, an informed consent must be voluntary, that is, free of coercion.
Voluntariness can be difficult to assess in situations in which the patient ap-
pears to be influenced by others, such as family members. Some patients are
willing to undergo medical procedures that are risky and invasive, without the hope
of significant benefit, because they believe it is what their family wants. In a
coerced situation, the patient is not making a free choice. One helpful strategy for
evaluating these situations is to ask to speak with the patient alone and ask the
patient directly whether and how others are influencing his or her decision.
There are some exceptions to the general requirement that the patients in-
formed consent should be obtained. First is the emergency exception, in
which medical treatment should not be delayed to obtain consent if the delay will
cause harm to the patient. In such a situation, it is presumed that the patient, if
capable, would consent to the necessary treatment. The second exception is that
of waiver, that is, when the patient declines to be informed. This exception occurs
rarely and should be explored to understand the patients reasons. Generally,
people should be discouraged from agreeing to accept risks without having the
risks explained. In some of these situations, a patient with the capacity will
delegate someone else to make medical decisions for him or her. (This is not
the same thing as a health care agent or proxy, who is a person named by the
patient to make decisions if the patient becomes incapable of doing so.) A third
exception to the informed consent requirement is therapeutic exception, in which
information is not disclosed to the patient because the doctor believes that
having the discussion will cause the patients condition to worsen dangerously.
This situation is uncommon, and the risks of informing the patient should
be documented.
 ethical issues in palliative care 131

Generally, the person who will be performing the procedure is responsible

for ensuring adequate informed consent, even if that consent is obtained by a
resident or nurse practitioner. For nonemergency procedures, the process of dis-
closure and discussion has often taken place during an office visit or a series of
visits and the signing of the form in the hospital is merely a formality. None-
theless, it is sound practice to ascertain that the patient has been given the
necessary information and to give the patient the opportunity to ask any
additional questions.

Assessing the patients understanding

The general rule is that an adult who is capable can decide whether to accept
or refuse medical treatment, even treatment that is life saving. This rule has both
ethical and legal foundations. The legal doctrine of informed consent is sup-
ported by the ethical principle of autonomy (self-determination) and promotes
rational decision making. As a society, we agree that people should make
important decisions about what happens to their own bodies based on accurate
information. Accordingly, the informed consent doctrine places the obligation on
the physician to explain the relevant information in a way that is understandable
to the patient. Respect for autonomy means that the patients decision will be
honored, even if others do not agree with it.
This general rule applies when the decision maker is an adult who has the
capacity to make decisions. Adulthood, or the age of majority, is usually defined
by statute in most United States jurisdictions as a person who is 18 years old.
Many states give minors the legal authority to make their own medical decisions
when the minors are pregnant or seeking care for specialized services, such
as reproductive services, substance abuse treatment, or treatment for sexually
transmitted diseases. Two main elements define the decision-making capacity.
The first element is the ability to understand ones condition, prognosis, and the
treatment options that are available, including the risks and benefits of each, and
to make a choice, with the understanding of its consequences (the ability to par-
ticipate meaningfully in the informed consent process). Second, adults are
presumed to have decision-making capacity, so caregivers are not expected to
evaluate the decision-making capacity of adult patients, unless there is a reason
to question the persons capacity.
Decision-making capacity is best understood as an assessment of the person at
a particular time, like a photographic snapshot. The lack of capacity may be
temporary, long-term, or permanent. If it is determined that a patient lacks ca-
pacity, an important question is whether steps can be taken to bring the patient to
a level of capacity. Sometimes factors can be identified that interfere with the
persons capacity, such as the presence of alcohol or drugs or metabolic toxins. It
is important to attempt to correct for temporary negative influences on the
patients capacity to maximize the respect for self-determination. For example,
this measure may require stopping a medication or adding treatment to correct
an ongoing process.
132 demme et al

The fact that a person has a psychiatric history or ongoing condition would
not, itself, allow a conclusion of the lack of capacity to make treatment deci-
sions. Capacity is not determined by category; so, for example, it cannot be
concluded that a person lacks capacity because of mental retardation or
schizophrenia. The assessment of capacity must be based on the evaluation of
the persons ability to understand his or her particular medical condition and to
weigh the various therapeutic options, with the appreciation of the risks and
consequences of each option. Asking the patient whether he or she understands
the medical condition and the available options can be a helpful way to assess
capacity. Psychiatric consultation is not a required element of the assessment of
capacity, but when there is a suspicion that psychiatric or cognitive factors are
interfering with the patients decision-making ability, psychiatric or neurologic
consultation can be helpful.
Capacity is commonly questioned when the patient rejects a recommended
course of treatment. A decision that appears unreasonable can be a signal that the
patient needs more information or more time. Although a patients unreason-
able decision is often the triggering event for a capacity evaluation, it cannot
serve as proof of the patients incapacity; otherwise, patients would have no right
except to agree with the doctor.
What happens when the patient does not have capacity and it is not possible
to restore it? The first step is to determine whether the patient has named
someone to make decisions. Laws in the majority of states explicitly provide for
this type of decision-making authority. A formally designated health care agent or
proxy stands in the shoes of the patient and has the authority to make all
decisions that the patient would be able to make if he or she were capable. The
purpose is to best approximate the decision the patient would make if capable,
and someone previously chosen by the patient can best accomplish that goal.
When no one has been named by the patient, generally, family members serve
as surrogates and provide a decision when the patient cannot. This system is
based on an expectation that the family members will know the patients prior
decisions, values, and preferences and will make a decision that is consistent
with them. This is called the substituted judgment standard. If it is not possi-
ble for the family members to base a decision on the patients own values and
preferences (eg, in cases involving children or adults who have never been able
to make decisions, for example), the family can make a decision based on
a balancing of benefits and burdens to the patient. This is called the best
interests standard.
Usually, but not always, family members are best able to act on behalf of
an incapacitated patient. However, legal relationships that are created by blood,
marriage, or adoption are not always positive relationships, and there is
no guarantee that a family member would be prepared or suitable to best rep-
resent the patients values and interests. The goal of finding someone who
can make a decision on the incapacitated patients behalf should focus on find-
ing a person who knows the patient and will be able to act in the patients
best interests.
 ethical issues in palliative care 133

What if there is no one present who can act on the patients behalf? Ideally,
there may be time to make inquiries and discover someone who knows and can
act on behalf of the patient. When this is impossible, most institutions have
administrative policies for giving consent if treatment is needed urgently; if
time permits, legal proceedings can be instituted to appoint a guardian.

Perioperative do-not-resuscitate orders

Closed cardiac massage, or cardiopulmonary resuscitation (CPR), was intro-

duced in the 1960s as a technique to reverse anesthesia-associated complications
of pulmonary or cardiac arrest [2]. Subsequently, the application of CPR was
broadened to nearly every patient in the hospital who suffered cardiopulmonary
arrest, including those with irreversible terminal conditions. Lagging behind the
implementation of this new emergency intervention was a mechanism for ad-
vanced decision making, so that a patient or his or her surrogate could determine
if he or she were willing to undergo resuscitation and have this choice docu-
mented in the medical record [3].
The right for a patient or their surrogate decision maker to authorize a do-
not-resuscitate (DNR) order is now accepted commonly, but until recently the
same consideration was not applied to patients with DNR orders who were about
to undergo anesthesia. Two major obstacles to the acceptance of perioperative
DNR orders have been the initial incongruence of deciding to limit potentially
life-saving treatment, such as resuscitation, while wanting to undergo invasive
procedures that carry a significant risk of respiratory or cardiac complications,
and developing a cogent understanding of how general anesthesia actually
differs from resuscitation.
Patients who elect to complete DNR paperwork may do so because they have
a disease or comorbidities that put them at an elevated risk of cardiopulmonary
failure. Given that many such conditions are not reversible, receiving resus-
citation in the event of sudden cardiopulmonary collapse is unlikely to prolong
their lives significantly and, if successful, may even leave them more debilitated
than before the event. However, these same patients may very well benefit from
numerous other therapies that can enhance their comfort and mobility and palliate
troublesome symptoms. Operative interventions may be needed, with corre-
sponding levels of anesthesia, ranging from sedation to general endotracheal
anesthesia. Specific examples include open reduction with internal fixation of a
pathologic fracture, tumor debulking, obtaining indwelling central intravenous
access, bowel resection to relieve obstruction, managing ureteral obstruction
resulting from a kidney stone, tracheostomy, endoscopic treatment of a bleed-
ing ulcer, placement of a gastric feeding tube, or stent placement for pancreatic
duct obstruction. In each of these examples, the patient may choose palliation and
surgery without wanting to accept the burden of resuscitation and its possible
poor outcomes. The patient does not go to the operating room with the hope that
by doing so he or she will hasten death but with the understanding that, were a
134 demme et al

catastrophe to occur and resuscitation be implemented, it is feasible that he or

she could actually end up in a worse condition than before the procedure. If
health care providers are willing to accept patient self-determination in every
other venue of the hospital, it is reasonable to recognize the validity of DNR
orders in the perioperative period as well.

Resuscitation versus maintaining homeostasis

The challenge posed by the examples above is how to differentiate between

resuscitation in its unwanted form and the techniques and medications used
as part of routine anesthesia practice to maintain homeostasis. As with all
patients, preoperative discussions should take place among the patient, surgeon,
and anesthesiologist to determine the best anesthetic approach, given the case to
be performed. These discussions are even more critical when the patient has
decided to limit the scope of interventions available by completing a DNR form.
In most general anesthesia cases, the only constraint placed on the anesthesia and
surgical teams for a patient with a DNR would be the prohibition against closed
cardiac massage and defibrillation. All of the other modalities, medications,
and reversal agents would be available, as in any other case. These limitations
have caused many providers legitimate concern about inducing anesthesia when
they believe that all of the emergency protocols one usually relies on are not
available. Recognizing this conflict between the concern for patient safety from
the perspective of the health care provider, especially for iatrogenic problems,
and the importance of patient autonomy, several key professional societies have
addressed this dichotomy.

Required reconsideration

The American Society of Anesthesiologists (ASA) [4] was the first pro-
fessional organization to address the difficulties of perioperative DNR orders in a
1993 document Ethical Guidelines for the Anesthesia Care of Patients with Do-
Not-Resuscitate (DNR) Orders or Other Directives that Limit Treatment. This
report was amended in 1998 and 2001 to reflect the evolution in the standard of
care for this patient population [5]. The American College of Surgeons (ACS)
and Association of Operating Room Nurses (AORN) followed the ASA with
the development of their own position statements in 1994 and 1995, respec-
tively [6,7].
The guidelines put forth by the ASA, ACS, and AORN clearly state that it is
unacceptable to have policies that require the automatic suspension of a patients
DNR order before undergoing anesthesia because this undermines the principle
of patient autonomy. Additionally, they emphasize the need for discussion be-
tween the patient or surrogate and the anesthesia team before any procedure
requiring anesthetic care is performed. This discussion, or required reconsid-
eration, should review the patients treatment goals, the limitations of care he
or she wishes to keep in place, the risks associated with that plan, what anesthetic
 ethical issues in palliative care 135

and interventions seem to be indicated to the anesthesiologist, and whether any

additional clarifications to either the goals or limitations of treatment are needed.
It is advisable that a patients DNR status and other advance directives undergo
review whenever there is a significant change in condition. This review may
be prompted by the provision of a new diagnosis, the experiencing of new side
effects, or the need for a new level of care, such as a specialized treatment unit.
The need for a surgical intervention and the time spent in the operating theater
clearly represent such changes. The discussion and conclusions should be docu-
mented clearly in the patients medical record, and any necessary changes to
DNR orders, surgical permits, or anesthesia consents must be properly amended
and signed by the necessary parties.
The required reconsideration of a patients DNR status will result in one of
three possible scenarios. First, the patient may decide that he or she is willing
to temporarily suspend the DNR order and receive full resuscitation, if it is
needed. The patient and attending physician will need to determine in advance
when and under what circumstances the DNR will be reinstated postoperatively.
Second, conversely, a patient may decide that altering or suspending the existing
DNR order is incongruent with previously stated goals and values. Although
there may be some procedures that can be performed with this type of restriction
in place, physicians and other providers cannot be forced to violate their own
conscience or perform activities they believe to be wrong or in violation of
standard medical practice. Anesthesiologists and surgeons who believe that the
limitations that the patient is requesting make the operation fundamentally unsafe
can refuse to participate. Third, the patient and health care team can reach a
compromise with limited resuscitation, in which certain resuscitation tools will be
prohibited but others will be allowed.
Limited resuscitation can be either procedurally or goal-directed [8]. Procedure-
directed DNR orders list the treatments and interventions that are agreeable to
the patient; therapy that is not listed will not be provided. This allows a patient
to indicate that pharmacologic therapy would be acceptable, whereas chest com-
pressions and defibrillation would not. The greatest advantage of the procedural
approach is that it delineates clearly what care is desired and what is prohibited,
making it easily interpretable by providers who may have limited knowledge
of the patient. This approach can be especially useful in critical care settings,
emergency departments, and other areas where there are numerous staff mem-
bers responsible for any given patient during the course of a shift or day. The
greatest limitation of procedure-directed DNR order is that health care pro-
viders must strictly apply the chosen criteria to every emergency, which has the
potential to result in an easily reversible emergency going untreated.
To address the limitations of the procedure-directed approach, a goal-directed
method has evolved that places primary importance on the patients stated treat-
ment objectives and personal ideals. Goal-directed DNR orders provide the care
team with a values framework that is then applied to the circumstances of car-
diopulmonary arrest. This requires that the anesthesiologist and surgeon know
the wishes of the patient well enough to interpret what he or she would want,
136 demme et al

given the medical facts of the problem, the possible interventions, and the
likelihood of success; substituted judgment is used to guide any resuscitation
efforts undertaken. A patient may indicate that he or she would want easily
reversible problems to be treated but not want aggressive or prolonged
measures to be used. Such directives allow considerable medical flexibility but
also require significant trust between the patient and surgical team. A goal-
directed approach is best used in a setting such as the operating room, where the
clinicians who have the most detailed understanding of the patients wishes will
be present throughout the perioperative period, from induction to recovery.

Perioperative do-not-resuscitate synopsis

Perioperative DNR orders are legitimate expressions of patient self-determi-

nation, and attempts should be made to honor them. Numerous palliative
procedures can offer substantial benefit to patients despite the risk of cardio-
pulmonary arrest they inherently pose. The ASA, ACS, and AORN uniformly
recommend a process of required reconsideration that emphasizes open com-
munication between the patient or proxy, anesthesiologist, and surgeon to deter-
mine the best way to safely meet the treatment goals of the patient. To facilitate
this, DNR orders can be maintained, suspended, or revised so that they are either
procedure- or goal-directed. Documentation of the requisite preoperative dis-
cussions, proposed plan, and any changes or amendments to existing DNR orders
or consent forms should be recorded clearly in the medical record.

Rule of double effect, aggressive symptom management, and terminal

sedation

The trajectory of dying has changed in North America, with patients living
longer but also spending more time very ill and symptomatic before dying [9].
Not only does this evolution present complex end-of-life decision-making chal-
lenges, but it also presents increasing challenges in managing the symptoms of
complex, severely ill patients. Many patients with the most difficult symptom
management problems will be seen by anesthesiologists when usual measures
are insufficient. Although some patients may be helped by advanced technical
symptom-relieving procedures, many eventually will be too ill and too close to
death to benefit from such treatments. It is under these circumstances in which
aggressive symptom management or even terminal sedation may be considered,
and the anesthesiologist may be asked to both provide, and justify the practices.

Rule of double effect

Although it was derived originally from Catholic moral theology, the rule of
double effect (RDE) has been a cornerstone of end-of-life medical ethics in
the United States [10]. In evaluating a clinical act, the RDE puts strong emphasis
 ethical issues in palliative care 137

on clinicians intentions. There are four main components of RDE: (1) the act
must be good or at least morally neutral; (2) bad effects can be foreseen but not
intended; (3) the act cannot be a bad means to a good end; and (4) there must
be proportionality between the risk and benefit. These elements help clinicians
distinguish between permissible and impermissible acts. Aggressive pain and
symptom management would be permissible, even if they result in a hastened
death, provided that the clinicians intention is purely to relieve suffering and
that the level of pain is severe enough to warrant the risk of the dose of medi-
cation used. On the other hand, intentionally giving a patient an overdose to
relieve the same suffering would be impermissible, even though the same end is
achieved, because it would violate the first three components of RDE.
RDE is not helpful in assessing many end-of-life decisions, such as with-
holding or withdrawing life-sustaining therapy, which depends more on the right
to bodily integrity [11]. Thus, a patient may intentionally end his or her life by
discontinuing life support, even though it seems to violate the second and, pos-
sibly, the third component of RDE because the patients right to bodily integrity
gives the privilege of stopping an unwanted treatment. Such acts turn more on
patient voluntariness (ensuring the patient is fully informed about alternatives and
is mentally capable of making such decisions) and proportionality among the
degree of suffering, prognosis, and nature of the decision being made than they
do on the clinicians intention in stopping the life support. Although a central
part of the clinicians intention in aggressive pain management or stopping life
support should be to relieve patient suffering, this intent becomes more complex
when a patient is ready to die and intends to do so as a way of escaping suffering.
Clinicians intentions may, at times, be more textured and multilayered than
suggested by the RDE [12].

Aggressive pain and symptom management

RDE is not needed to justify usual pain management. In fact, invoking RDE
would overstate the risk of usual opioid treatment for pain, in which sedation
occurs long before respiratory depression in most patients. When treating the
sensation of dyspnea with opioids, there is a small risk of tipping the balance
toward respiratory failure. However, if one starts at a low dose and increases
the dose slowly, the risk generally is overstated and probably does not require
the invocation of RDE. Many treatments in other domains of medicine have
infrequent, lethal side effects about which patients must be informed (such as
chemotherapy), but RDE is not believed to be necessary to justify their usage.
The main clinical situation in which RDE is relevant is in the management of
accelerating pain or extreme shortness of breath in the imminently dying patient.
In these circumstances, all four components of RDE are relevant. First, aggres-
sive management of pain or shortness of breath is not only permissible but also
imperative because leaving such severe symptoms untreated when effective treat-
ment is available would be morally untenable. Second, although one can foresee
that death may be hastened, it is not the intent of the clinician. This is dem-
138 demme et al

onstrable because when the symptoms come under control, the dose of the opioid
is maintained but not increased (unless the symptoms increase in the future).
Third, opioid treatment is the medically indicated means to the desired end (relief
of the patients symptoms). Fourth, the patients symptoms in these circumstances
(severe pain or extreme dyspnea) are proportionately severe enough to warrant the
risk (which is more real in these circumstances) of hastening death from the
accelerating doses of opioids required. The anesthesiologist, often seeing such
patients in consultation, may need to provide guidance both in terms of dosing
(adjusting baseline rates and as needed supplementation) and in terms of
thinking through the justification of the process.

Terminal sedation

Because clinicians have struggled for ways to respond to severe physical

suffering when usual treatments are ineffective, the possibility of terminal se-
dation has been raised as one of the last-resort options [11,13]. Terminal sedation
and other last-resort options should be considered only in the presence of
intractable suffering, when all usual treatments have failed and the patient
(or surrogates) consent. Terminal sedation involves sedating patients to uncon-
sciousness so they can escape awareness of their intractable suffering; generally
life-sustaining treatment, including food and fluids, is withdrawn (although some
practitioners suggest that this should be a separate decision) [14]. The jus-
tification of the practice is twofold: (1) the sedation is supported by RDE because
the clinicians intent is to provide the patient with an escape from otherwise
severe and intractable suffering; and (2) the withdrawal of life support is justified
under the right to bodily integrity. However, virtually 100% of patients who are
terminally sedated die, and death in this circumstance does not depend on having
a terminal illness. Therefore, when one looks at the practice in aggregate, it is
possible that it goes beyond the confines of RDE in terms of clinician intent
(although it meets the other RDE components).
Some practitioners prefer the term palliative sedation to describe the prac-
tice because it emphasizes the goal of palliation and deemphasizes the terminal
nature of the practice. The present authors resist this terminology because it may
cause confusion about the gradual increase in the use of benzodiazepines or
antipsychotics in the anxious or agitated patient toward the end of life. With
this type of palliative sedation, sedation to unconsciousness is not the goal but
rather the lowest level of sedation that can relieve the patients suffering. This is a
circumstance in which RDE is genuinely helpful because the clinician will take
the risk of increasing sedation, but the intended endpoint is not unconsciousness
or death. To minimize the controversy, the present authors policy guiding ter-
minal sedation is descriptively titled sedation to unconsciousness for relief of
otherwise intractable physical symptoms in the terminally ill.
Terminal sedation usually is performed with intravenous infusions of benzo-
diazepines or barbiturates [15], most commonly in acute care settings. Because
anesthesiologists are available in this setting and they have experience provid-
 ethical issues in palliative care 139

ing sedation in other settings, they may be asked to provide guidance. Terminal
sedation allows practitioners to respond to a wide range of intractable suffering,
and, although it remains controversial in some settings [14], the practice is
generally acceptable to palliative care and ethics groups that otherwise are op-
posed to other last-resort practices, such as voluntarily stopping eating and drink-
ing or physician-assisted suicide [11]. In the present authors opinion, terminal
sedation should be reserved for the most difficult cases, when all other palliative
care interventions have been found to be ineffective and suffering is severe.
Consultations with palliative care specialists and ethicists are recommended if
there is any additional uncertainty about the appropriateness for a given case.
Within the present authors inpatient palliative care consultation program, terminal
sedation is needed in approximately 1% of cases, although in other programs it is
used in as many as 50% of deaths [16].

Critical role for anesthesiologists

Anesthesiologists have an important role in relieving severe pain, shortness

of breath, and other intractable symptoms in terminally ill patients. In addition to
their knowledge of special interventional techniques and basic opioid pain
management, anesthesiologists need to be familiar with approaches to and justi-
fications for aggressive symptom management of pain and shortness of breath,
as well as with terminal sedation as a last resort. Knowledge of RDE certainly is
helpful in justifying and explaining aggressive symptom management to patients,
family, and staff, whereas concepts such as proportionality, voluntariness, and
our obligation not to abandon patients [17] are more central to the justification of
terminal sedation. Although the anesthesiologist may be able to help the team
develop an effective approach to managing patients with symptoms that are dif-
ficult to control, consultation with palliative care and ethicists services may be
useful in seeking consensus about the best path in the most difficult cases.

Withdrawal of life support and ventilators

Over the past century, advances in nutrition, sanitation, and medicine have
led to an increased lifespan for a majority of Americans. Rather than dying of
acute infectious illnesses or trauma, many now die of chronic degenerative
diseases. There has been a shift in the occurrence of death at home to death in the
hospital or nursing home setting, where resuscitative equipment and advanced
technological apparatuses and medications can be used. A problem pointed out
by Thomasma [17] is that people dont die all at once anymore. They die in
pieces. . . and adds that youre never sure which piece to stop at. Presently,
the majority of Americans are likely to face either the withholding or withdrawal
of certain medical treatments in the dying process. In the 1990 Cruzan case
[18], Supreme Court Justice Brennan noted Of the approximately 2 million
people who die each year, 80% die in hospitals and long-term care institu-
140 demme et al

tions. . .and perhaps 70% of those after a decision to forgo life-sustaining treat-
ment has been made [18].
Before the 1970s, the removal or withdrawal of life-sustaining treatment
was commonly considered euthanasia. Because withdrawal of therapy was be-
lieved to be unallowable, it was easier in practice for physicians to withhold
treatment; yet, some ethicists argued, a more defensible course would be to offer
a therapy to see whether it would be useful for a particular individual in a specific
situation. However, if a treatment did not result in a desired outcome for a patient,
some patients and their physicians feared the possible result of a patient being
kept alive long-term, tethered to various machines.
Most ethicists in North America believe that withholding and withdrawal
of potentially life-sustaining medical treatment are both morally permissible and
that one practice is not morally better than the other. It may be more defensible
to try a therapy to see if it is useful for an individual patient but to discontinue the
therapy if it does not fulfill desired goals.
The Quinlan [19] case set a landmark legal precedent. In 1975, 21-year-old
Karen Quinlan suffered anoxic brain damage, leading to a persistent vegetative
state (PVS). She was kept alive with a ventilator and a feeding tube. Her father
petitioned the New Jersey Supreme Court to have the ventilator discontinued.
Although declaring that the State had an interest in the preservation of life,
the judge opined, [T]he individuals right to privacy grows as the degree of
bodily invasion increases and the prognosis dims. Ultimately there comes a point
at which the individuals rights overcome the State interest [19]. The court
admitted We are aware that such termination of treatment would accelerate
Karens death. We believe that the ensuing death would not be homicide but
rather expiration from existing natural causes. The Quinlan court recommended
that multidisciplinary hospital ethics committees be consulted in similar cases.
Although the Quinlan case was decided by a state (New Jersey) court, it had
widespread influence and led to a greater acceptance of the removal of ventilators
in certain medical circumstances. Of interest, Karen Quinlan was weaned from
the ventilator but continued to receive tube feeding and so did not die until
10 years later, when antibiotics were withheld for an infection.

Withdrawal of invasive nutrition

In 1983, 24-year-old Nancy Cruzan [20] was in a car accident and was thrown
face down in a water-filled ditch. Like Quinlan, she developed PVS and was kept
alive by the use of a feeding tube, which had been authorized by her husband.
Three years after the accident, her parents, at that time her court-appointed
guardians, requested that her tube feedings be discontinued. The Missouri Su-
preme Court held that artificial nutrition could not be withdrawn unless there was
proof that she had authorized such termination before losing capacity. The United
States Supreme Court reviewed the case and ruled that a competent person has a
liberty interest under the due process clause in refusing unwanted medical
treatment. The Court stated . . . it is assumed that a competent person would
 ethical issues in palliative care 141

have a constitutionally protected right to refuse lifesaving hydration and nu-

trition. [20]. Because Cruzan had lost decision-making capacity, the Court
agreed that Missouri could require clear and convincing evidence of her
previously expressed wishes regarding the withdrawal of life-sustaining treat-
ment. Convincing evidence of Cruzans wishes was presented to the Missouri
Court, and her tube feeding was discontinued.

Advance directives

The Cruzan case was decided in 1990. The publicity of the Cruzan case fueled
the passage of the Patient Self-Determination Act (PSDA), a federal law that
took effect in late 1991. The PSDA requires, as a condition of participation in
Medicare or Medicaid programs, that hospitals, skilled nursing facilities, home
health agencies, hospice programs, and prepaid health care organizations furnish
patients with information about advance directives. For example, on entering a
hospital, a patient must be given written information about the right to make
decisions concerning medical care, the right to refuse or accept treatment, and
the right to formulate advance directives.
All states now have at least one form of advance directive [21]. An advance
directive is the general term for a document designed to enable competent adults
to make health care plans in advance of future incapacity. Advance directives are
prescribed by individual state law and are generally of two types: instructional or
proxy [21]. An instructional directive, such as a living will, allows a competent
adult to specify treatment wishes in advance of a terminal illness or condition
during which the person is not capable of making sound decisions. A proxy
directive, or durable attorney for health care, allows a competent adult to choose
another adult to make health care decisions if he or she becomes unable to do
so. Some states allow both types of directives. Some states put limits on the
ability of the health care agent to make certain decisions. For instance, in New
York, a designated proxy cannot make decisions about the removal of artificial
hydration and nutrition, unless he has reasonable knowledge of whether the
patient would choose to forgo that treatment. Despite the availability of advance
directive forms, the majority of adults have not completed one.
As demonstrated by the recently publicized Schiavo case [22] in Florida,
decisions about the discontinuation of invasive nutrition have tended to be the
most controversial with regard to withdrawal of therapy. However, a recent poll
indicates that 74% of Americans say they would want their health care guardian
to remove a feeding tube if they were in a persistent vegetative state [22].
The withdrawal of life support and refusal of medical and surgical procedures
are now protected by law for competent patients. For patients without capacity,
each state may set its own standards of evidence required for such actions to
safeguard patients interests. Although the refusal of treatment or withdrawal of
treatment is legally allowed for competent patients, it is the clinicians obliga-
tion to ensure that there is informed refusal. Just as we aim for informed
consent when providing a treatment, we must ensure that patients who refuse
142 demme et al

therapy understand what they are refusing, that is, patients should demonstrate
that they understand the risks, benefits, and alternatives of not accepting
recommended therapy. For instance, a competent adult Jehovahs Witness can
refuse a life-saving blood transfusion but must be apprised of the risk of refusal,
and he or she should be able to explain the refusal and the likely outcome with or
without transfusion.
Anesthesiologists may be involved in cases in which a beneficial operation
is refused (eg, the removal of a diabetic patients gangrenous foot). According to
the Guidelines for the Ethical Practice of Anesthesiology [23], Anesthesiolo-
gists respect the right of every patient to self-determination. Anesthesiolo-
gists should include patients, including minors, in medical decision making
. . .Anesthesiologists should not use their medical skills to restrain or coerce
patients who have adequate decision-making capacity.
In critical care settings, specialists, including anesthesiologists, sometimes may
be involved with the discontinuation of dialysis, pressor medications, antibiotics,
and even ventilators. When removing a ventilator from a dying patient, the critical
care specialist may be involved in aggressive management of associated dyspnea
with opioids, as justified by the rule of double effect discussed earlier.

Brain death

Although Quinlan, Cruzan, and Schiavo had severe neurologic damage, none
of them was brain dead when their treatments were discontinued. Brain death
was described by Beecher and the Ad Hoc Committee of the Harvard Medical
School to Examine the Definition of Brain Death in 1968 [24]. The primary
purpose of the committee was to define irreversible coma as a new criterion for
death. Among the reasons given for a new definition was that the burden is great
on patients who suffer permanent loss of intellect, on their families, on the
hospitals, and on those in need of hospital beds already occupied by these comatose
patients. In addition, it was noted that [O]bsolete criteria for the definition of
death can lead to controversy in obtaining organs for transplantation [24].
All 50 states accept brain death as a legal definition of death. The Uniform
Determination of Death Act describes death as either the irreversible cessation
of the circulatory and respiratory functions or irreversible cessation of all func-
tions of the entire brain, including the brain stem [25].
Although it is not necessary to provide medical care to one who is dead, the
anesthesiologist might provide continued circulatory support to one who has
been declared dead by whole brain criteria to allow for the retrieval of organs for
transplantation. Because these brain-dead donors are unreceptive and unrespon-
sive, there is a debate among anesthesiologists about whether these cases should
receive anesthesia for organ procurement procedures [26].
Recently, it has become possible to retrieve organs from those who have
died as a result of cardiopulmonary criteria. Typically, this retrieval occurs when
a patient does not meet the standard of brain death but is severely ill and from
 ethical issues in palliative care 143

whom ventilatory support has been withdrawn after proper consent by advance
directive or from appropriate surrogate decision makers. This has been termed
controlled nonheart-beating cadaveric donation or donation after cardiac
death. In such an instance, it is critical that the determination to withdraw life
support must come before the discussion about organ retrieval, to avoid aiming
for death to obtain organs. To avoid conflict of interest, the transplantation
surgical team should not be involved with the determination of death. The
anesthesiologist from the transplant team should not be the one to provide ad-
ditional sedation during the withdrawal process. A separate team should provide
for the patients comfort care and determination of death. The Institute of Medi-
cine (IOM) [27] concluded that premortem cannulation of controlled nonheart-
beating donors is acceptable but requires informed family consent. To ensure
that the donor is dead and to avoid the perception of organ removal from a living
person, the IOM recommends that organ removal not occur until 5 minutes after
the donor heart has stopped and the pulse is lost.

Summary

It is not unusual for perioperative health care providers to confront ethical

concerns; the role of the anesthesiologist is particularly important. Obtaining
patient consent is necessary for each surgical procedure, so clear understanding
of the consent process is essential. A patient who has declined resuscitative
measures may still request a palliative operation, and wishes for a DNR order
must be addressed. The anesthesiologist may be the team member best-equipped
and most able to deliver high doses of sedatives and potent pain analgesia. An
appreciation of the rule of double effect may be instrumental for those providing
relief of suffering. At some point, withdrawal of technologic support may be
requested by the patient or the surrogate. In addition, the family may wish to offer
a dying patients organs to benefit others. In these instances, among others, the
anesthesiologist may play a key role in providing good palliative care.

References

[1] Canterbury v Spence, 464 F2d 772 (DC, 1969).

[2] Kouwenhoven WB, Jude JR, Knickerbocker GG. Closed-chest cardiac massage. JAMA 1960;
173:1064 7.
[3] Burns JP, Edwards J, Johnson J, et al. Do-not-resuscitate orders after 25 years. Crit Care Med
2003;31(5):1543 50.
[4] American Society of Anesthesiologists. Ethical guidelines for the anesthesia care of patients with
do-not-resuscitate (DNR) orders or other directives that limit treatment: approved by the House
of Delegates, October 10, 1993. Available at: http://www.asahq.org/publicationsAndServices/
standards/09.html. Accessed January 10, 2006.
[5] American Society of Anesthesiologists. Ethical guidelines for the anesthesia care of patients with
do-not-resuscitate (DNR) orders or other directives that limit treatment: approved by the House
144 demme et al

of Delegates, October 17, 2001. Available at: http://www.asahq.org/publicationsAndServices/

standards/09.html. Accessed January 10, 2006.
[6] American College of Surgeons. Statement of the American College of Surgeons on advance
directives by patients: do not resuscitate in the operating room. American College of Surgeons
Bulletin 1994;79(9):29.
[7] Association of Operating Room Nurses. Perioperative care of patients with do-not-resuscitate
(DNR) orders. 1994. Available at: http://aorn.org/about/positions/dnr.htm. Accessed January
10, 2006.
[8] Truog RD, Waisel DB, Burns JP. DNR in the OR: a goal directed approach. Anesthesiology
1999;90(1):289 95.
[9] Field MJ, Cassel CK. Approaching death: improving care at the end of life. Institute of Medi-
cine Report. Washington (DC)7 National Academy Press; 1997.
[10] Quill TE, Dresser R, Brock DW. Rule of double effect: a critique of its role in end-of-life deci-
sion making. N Engl J Med 1997;337:1768 71.
[11] Quill TE, Lo B, Brock DW. Palliative options of last resort: a comparison of voluntarily
stopping eating and drinking, terminal sedation, physician-assisted suicide, and voluntary active
euthanasia. JAMA 1997;278:2099 104.
[12] Quill TE. The ambiguity of clinical intentions. N Engl J Med 1993;329:1039 40.
[13] Quill TE, Byock I for the ACP-ASIM End-of-Life Care Consensus Panel. Responding to
intractable terminal suffering: the role of terminal sedation and voluntary refusal of food and
fluids. Ann Intern Med 2000;132:408 14.
[14] Jansen LA, Sulmasy DP. Sedation, alimentation, hydration, and equivocation: careful con-
versation about care at the end of life. Ann Intern Med 2002;136:845 9.
[15] Ventafridda V, Ripamonti C, DeConno F, et al. Symptom prevalence and control during cancer
patients last days of life. J Palliat Care 1990;6:7 11.
[16] Quill TE, Cassel CK. Nonabandonment: a central obligation for physicians. Ann Intern Med
1995;122:368 74.
[17] Siegel J. Deciding when to let life end. Chicago Tribune February 10, 1992;5, 2.
[18] Meisel A. What is the right to die?. In: The right to die. 2nd edition. New York7 John Wiley &
Sons, Inc; 1995. p. 3 36.
[19] In re Quinlan, 355 A 2d 647 (NJ 1976).
[20] Cruzan v Director, 497 US 261, 302 (1990) (Brennan J, dissenting).
[21] Clarke D. The patient self-determination act. In: Monagle J, Thomasma D, editors. Health
care ethics: critical issues for the 21st century. Gaithersburg (MD)7 Aspen Publishers; 1998.
p. 92 116.
[22] McGirt E. Who will speak for you when you cant? Money Magazine 2005;5:25 6.
[23] American Society of Anesthesiologists. Guidelines for the ethical practice of anesthesiology:
approved by the House of Delegates on October 3, 1967, and last amended on October 15, 2003.
Available at: http://www.asahq.org/publicationsAndServices/standards/10.pdf. Accessed January
10, 2006.
[24] Beecher H, Adams R, Barger A, et al. A definition of irreversible coma: report of the ad hoc
committee of the Harvard Medical School to examine the definition of brain death. JAMA 1968;
205(6):337 40.
[25] Paola F, Anderson J. The process of dying. In: Sanbar SS, Gibofsky A, Firestone M, et al,
editors. American College of Legal Medicine. Legal medicine. 4th edition. St. Louis (MO):
Mosby; 1998. p. 35264.
[26] Truog R, Robinson W. Role of brain death and the dead: donor rule in the ethics of
transplantation. Crit Care Med 2003;31(9):2391 6.
[27] Institute of Medicine. Non-heart-beating organ transplantation: medical and ethical issues in
procurement. executive summary. 2003 National Academy of Sciences. Available at: http://
www.nap.edu. Accessed June 29, 2005.