JOURNAL OF PALLIATIVE MEDICINE

Volume 3, Number 1, 2000

Mary Ann Liebert, Inc.

Palliative Care/Hospice and the Withdrawal of Dialysis

KATHY JOHNSON NEELY, M.D.1 and DAVID M. ROXE, M.D.2

ABSTRACT
While the majority of end-stage renal disease (ESRD) patients on dialysis lead satisfying lives,
an increasing number are choosing to withdraw from dialysis before death. A partnership between nephrology and palliative care/hospice healthcare teams would seem likely in the care
of ESRD patients, yet this is often not the case. In anticipation of increasing participation by
palliative care/hospice teams in the care of such patients, this article reviews the decisionmaking process of withdrawal and the medical care of the patient who withdraws. While
withdrawal can be an acceptable choice from a medical, legal, psychiatric, and ethical point
of view, it can nonetheless be complex. Profound decisions are often characterized by the
need for time to process, and by ambivalence among patient, family and healthcare providers.
In addition to caring for the patient and family, the palliative care/hospice team will want to
consider the needs of the referring nephrology team as well. A "uremic death" is characterized as painless; however, other symptoms related to the accumulation of toxins and fluid
can be anticipated and managed. Pharmacological intervention of uremic symptoms, as well
as the pain attendant to other, nonrenal comorbid disease is accomplished with awareness of
the impact of renal failure on the excretion of various drugs and their metabolites.

INTRODUCTION

Data System, which records mortality statistics,

reports that from 1988 to 1990, the percent of
NLY A FEW DECADES AGO, dialysis was a rare, deaths attributable to withdrawal was 8.4%. Beexperimental resource available to a fortu- tween 1990 and 1995, this percentage had innate minority of candidates who would other- creased to 17.8%.2 The dramatic increase is parwise die of end-stage renal disease (ESRD). Since tially explained by changes in the reporting
the entitlement act of 1972, near universal fund- system; however, patients with multiple lifeing for dialysis has been offered to all residents threatening diseases and their nephrologists are
of the United States with ESRD.1 The population increasingly willing to consider withdrawal from
receiving dialysis has evolved over time. Origi- dialysis as an option of care. Nephrologists are
nally young and acutely ill, most patients now not trained in the treatment of uremic death,
are older and have chronic, progressive comor- because dialysis averts that. Accordingly, a colbid illnesses. While the vast majority of dialysis laborative effort of palliative care/hospice carepatients are gratified with their lives, every year givers with nephrologists would seem an approan appreciable and apparently growing minority priate response to the needs of the ESRD patient
elects to discontinue dialysis. The U. S. Renal foregoing dialysis. Thoughtful consideration of

Northwestern University Medical School, Division of General Internal Medicine, Chicago, Illinois.
Northwestern University Medical School, Division of Nephrology/Hypertension, Chicago, Illinois.

57

58

NEELY AND ROXE

such is reflected in the joint statement of the Renal Physicians Association and the American Society of Nephrology regarding end-of-life treatment for ESRD patients who forego dialysis. They
recommend "a hospice or hospice-like approach"
as "the optimal method of patient care."3 However affirmed in theory, hospice enrollment of
ESRD patients is rare. For example, in North Carolina in 1996, of the 14,253 patients served by hospice, only 71 were ESRD patients.4 And as recently as 1999, while affirming that withdrawal
of dialysis is "an opportunity to participate and
assist in a major life event of both the patient and
family," Leggat and Port5 fail to mention palliative care or hospice as a means to facilitate that
end.
However, there is potential for change. Palliative care/hospice teams are becoming increasingly available in hospitals as inpatient con-

sultative teams, many offering acute inpatient

palliative services in addition to traditional home
hospice enrollment. Palliative care/hospice is
modeled on and expert in end-stage cancer care;
however, in recent decades it has intentionally included those often underserved in end-of-life
care, among which are patients with renal failure.6 In the hope of increasing palliative
care/hospice involvement with ESRD patients
and families, this review article addresses decision making and ethical issues leading to the decision to withdraw from dialysis. It also discusses
anticipated physical symptoms inherent in an
uremic death, as well as pain frequently encountered with comorbid diseases. Illustrative cases
from the experience of the Acute Palliative Care
Unit of the Northwestern Memorial Palliative
Care/Hospice Program will be included
(Table 1).

TABLE 1. PROFILES OF SIX PATIENTS W H O WITHDREW FROM DIALYSIS WHILE INPATIENTS ON THE ACUTE INPATIENT
PALLIATIVE CARE UNIT OF NORTHWESTERN MEMORIAL HOSPITAL FROM JANUARY 1, 1998 TO AUGUST 30, 1998

.7?.

G.Z.

D.W.

LR.

J.B.

Age
Gender
Renal
disease

71
male
ESRD

66

66
female
ESRD

93

77
female
ESRD

Comorbid
disease(s)

Endocarditis
and
perivascular

Calciphylaxis

CHF; PVD;
dry
gangrene

CHF;
aspiration
pneumonia

74
male
ATN
imposed on
CRI
CHF; PVD

Duration of
dialysis
Lifespan
after final
cuaiysis
Who
requested
dialysis
cessation

years

10 months

13 months

3 years

1 month

years

7 days

8 days

3 days

9 days

28 days

Competent
patient and
son

Competent
patient and
wife, in
conjunction

Family
acting as
healthcare
proxy

Patient

Patient and
sons

Multiorgan
failure;
imminent
death

Burden of
multiple
interventions

Multiorgan
failure;
sepsis?

Uremia

Futility of
Futility of
goals of
goals of
care;
treatment
burden of
multiple
interventions
Uremia

male
ESRD

male
ESRD

L.B.

Multi-infarct
dementia

Family
acting as
healthcare
proxy

"Wltrl

Rationale

Futility of
goals of
care; multiorgan
failure

Presumed
immediate
cause of
death

Ruptured
aortic
aneurysm

advance
directives
Futility of
goals of
treatment

Uremia

While J.B. did not have ESRD, we include him in this discussion as his decision-making process is illustrative.
ESRD, end-stage renal disease; CHF, congestive heart failure; ATN, acute tubular necrosis; CRI, chronic renal insufficiency; PVD, peripheral vascular disease.

WITHDRAWAL OF DIALYSIS

THE DECISION-MAKING PROCESS OF

DIALYSIS WITHDRAWAL
The right to withdraw

Within recent memory, a patient's request to

withdraw from dialysis or any life-sustaining
technology was regarded as a self-destructive
wish and thus tantamount to an attempt at suicide. The majority of literature from the 1960s
and 1970s viewed withdrawal with considerable
alarm. A desire for death was regarded as prima
facie evidence of depression; if sufficiently depressed to make a request for withdrawal, a patient was clearly incapable of making decisions.
However, in 1971, McKegeny and Lange7 presciently argued that withdrawal of dialysis was a
"legitimate topic" of discussion between ESRD
patients, their families, and physicians. Healthcare teams, they said, should individualize responses to such requests, acknowledging that
"death by the natural course of a disease is not
equivalent to self-destruction." Further, individual professionals should be wary of interpreting
personal uneasiness as ethical scruple when dissension in decision-making arises.
Legal, psychiatric, and ethical opinion has subsequently evolved along these lines. Currently,
the law recognizes an individual's right to refuse
excessively burdensome interventions. Death is
understood as an option sometimes preferable to
life with disease and with onerous medical interventions. In the face of a request to terminate or
limit therapy, legal competence is presumed until proven otherwise.
From a psychiatric view, the prevalence of major depression among seriously ill patients justifies an evaluation when a request for treatment
termination is made. Often, depression in the context of illness and loss is "reactive," and supportive therapy alone assists the patient to clearly
contemplate the future. Major depression, on the
other hand, often renders a patient incapable of
decision making. However, there are clearly
cases in which the presence of even "a major
depressive disorder is neither necessary nor sufficient for the impairment of competence," and
the requests of a depressed patient are honored.8 It may be that a more sophisticated understanding of depression in the seriously ill is
needed. Research indicates that among the
many characteristics of depression rendering an
individual self-destructive, it may be hopeless-

59

ness that singularly eclipses other aspects.

Hopelessness, not depression, is most correlated with suicidal intent.9
Ethically, a decision made by a willing patient
to refrain from active treatment such as dialysis in the context of imminently terminal disease is the closest to being a clearly acceptable
choice.10 However, many withdrawal decisions
are pondered against the backdrop of daunting
complexities, and ethicists advocate that clinicians be willing to engage in such vexing
cases.11'12'13 Nephrologists, or indeed any
physician focused on a curative paradigm of
care may find it difficult to recognize that there
is still a need for their guidance and support
when the end point is death. Paradoxically, the
end of life is often when physicians become
most indispensable.
The role of end-of-life planning
and advance directives

End-of-life decision making encompasses

many possible determinations. Withdrawal of
dialysis and other life-sustaining technology is
of greater magnitude than many other limitations placed by patients on healthcare
providers.14'15 Recent encouragement of advanced life planning may be changing the character of discussions of end-of-life decision making in general and options for patients with
ESRD in particular.
In 1997, Cohen et al.16 found that only 6% of
144 ESRD patients on dialysis had advanced directives; all of these were male, and tended to
be better educated than their cohorts. They
speculated that the better-educated, confident
of the power of choice, seize an opportunity for
self-determination.16 However, Sekkarie and
Moss17 found that advance directives were
available for more than half the patients in his
study who withdrew from dialysis. They concluded these larger numbers are attributable to
an enhanced understanding of the utility of advance planning within the nephrology community they studied, and called for continuing efforts to educate physicians.17
Sekkarie and Moss17 further noted in their
study that the majority of patients who lacked
decision-making capacity at the time of withdrawal had advance directives. These directives
were extremely helpful to the nephrologists as
they participated in decision making.17 This

60

finding resonates with the work of Singer et

al.,18 who in his study of nephrologists found
them reluctant in the absence of written advance directives to comply with family requests
and substitutive judgment to withdraw dialysis. However, caution is indicated in overreliance on written advance directives; while 39%
of ESRD patients on dialysis said that they
wished "no leeway" regarding their advance directives, a nearly equal number (31%) specified
on further clarification they wanted "complete
leeway" to be exercised by their healthcare
teams.19 Latitude in the specific directives of a
document, as well as the degree to which the
patient hopes it is interpreted further emphasizes the value of a healthcare proxy or surrogate. A trusted, informed loved one speaking
for the patient can be extremely helpful in endof-life decision making.20
Among those who withdraw from dialysis, a
special circumstance is notable, that of the cessation of the time-limited, therapeutic trial of an intervention that eventually proves excessively burdensome. While some studies indicate that
withholding dialysis from unstable patients is an
option more commonly exercised by nephrologists
than withdrawal of dialysis,18 others suggest that
a trial of dialysis is increasingly common for patients with multiple nonrenal illnesses.2 Optimally,
such a trial is initiated with a discussion of the limits of dialysis: that it may prove unhelpful and thus
may be subject to termination, and also, by what
criteria a trial might be judged a success.
What might be the impact of intentional,
broadly encouraged end-of-life planning for patients with ESRD on dialysis? A large majority
state that they would welcome such a discussion with their physicians, and that dialysis
units should ask all patients as a matter of routine about advance directives.21 Schwartz and
Perry22 came to the same conclusion after a 6year retrospective study. Patients who withdrew from dialysis in "reconciled fashion" were
those most likely to have a verbal or written advance directive. In their recommendations for
the nephrology team, they describe an approach
strongly reminiscent of palliative care/hospice.
Specifically, they describe an interdisciplinary
team providing a range of perspectives to the
patient and family, such that no one of the team
is engaged in a difficult discussion, isolated and
overburdened. End-of-life discussions, they
say, are best done over time, allowing the full

NEELY A N D ROXE

range of the patient's feelings and opinions to

be explored.22
Ambivalence

The human capacity to vacillate when making

weighty decisions is often dramatically encountered in end-of-life determinations. Of 200 terminally ill cancer patients interviewed by Chochinov et al.,23 6 who strongly desired death were
studied over time. Of these, 4 had changed their
minds within 2 weeks, demonstrating the "inherent transience" of the wish of many patients
to die.23 Similarly, ambivalence is noted among
ESRD patients considering withdrawal from dialysis.24 While ambivalence is anticipated amongst
any patients making difficult decisions, it is perhaps enhanced in patients with ESRD who live
day-to-day partnered to a technological intervention that is both life-sustaining and optional.
Many ESRD patients have acclimated to dialysis
over many years; others, however, have more recently initiated dialysis as a therapeutic trial, and
from its onset, have been weighing benefit versus burden. The decision-making process is
fraught with ambivalence. Even after an established decision to withdraw, generally more than
a week passes before death.24 This interlude before death is disproportionately generous in comparison to the withdrawal of other modalities
such as ventilator support, a pause that permits
review and sometimes reversal of previous
wishes.24-25 Furthermore, the decision to withdraw is often made with an uncertainty of prognosis among patients with multiple life-limiting
diseases. This uncertainly only heightens the potential for ambivalence among any of the participants in the decision, and the opportunity to rescind an earlier decision. In addition, the interval
before death may prompt reticent family members to step forward with needs or requests that
alter the dynamic of the end-of-life process.
The following cases from the Palliative
Care /Hospice Program of Northwestern Memorial Hospital illustrate complexities encountered
in the making decisions about withdrawal from
dialysis; specifically, uncertain prognosis, ambivalence amongst the many participants in the
process, and the specter of suicidal intent behind
a request to withdraw.
L.B. was a 77-year-old woman with multiinfarct dementia and ESRD, successfully

WITHDRAWAL OF DIALYSIS

maintained on outpatient hemodialysis for

11/2 years. At baseline, she recognized and
conversed with her family and required
their full assistance in her activities of daily
living. She was hospitalized with an abrupt
hypertensive crisis and new onset of
seizures. Although her seizures were controlled, she remained unresponsive. After 2
weeks of counseling with their primary care
physician, the family chose to withdraw
dialysis from the persistently unresponsive
patient. She was admitted to the Inpatient
Palliative Care Unit of Northwestern Memorial Hospital for symptom management and
family support. After 7 days without dialysis, the patient inexplicably regained consciousness, recognized her family and asked
to be fed. The delighted family concurred
with the astounded physician that she
should resume her previous dialysis regime.
She was discharged to home and outpatient
hemodialysis.
J.B. was a 74-year-old man whose claudication from peripheral vascular disease led
to hospitalization for a femoral-tibial bypass.
Postoperatively, his course was complicated
by volume overload, worsening heart failure, melena while being anticoagulated for
atrial fibrillation, and acute tubular necrosis
imposed on chronic renal insufficiency. He
underwent dialysis for 3 weeks. However, as
the burden of his therapies weighed heavily
against his uncertain future, ongoing dialysis became untenable to him. He discussed
with his primary care physician, nephrologist, and children his decision to withdraw
from dialysis. Each of them responded with
varying degrees of support, although one
physician remained convinced to the end
that the patient was making a terrible mistake. The patient continued alert and unperturbed by arguments to the contrary
throughout the next 28 days, and died comfortably in the Palliative Care Unit of Northwestern Memorial Hospital.
D.R. was a 78-year-old woman with severe degenerative joint disease (DJD), congestive heart failure, diabetes, recurrent
atrial fibrillation, and hypertension; she had
initiated outpatient hemodialysis as a therapeutic trial 3 months prior for treatment of

61

ESRD complicated by volume overload. She

was considerably more comfortable on dialysis, yet was distressed by the degree of debility imposed by DJD. She repeatedly
sought support from her physician, who affirmed her right to refuse treatment but encouraged her to continue. On a Friday, she
called to say that she'd had her last dialysis;
further discussion with her family confirmed
that she'd made this decision, and although
sad, they would support her. On the following Tuesday, she called to say that the day
prior, she'd consumed "about 70" of her
warfarin 5-mg pills in hopes that it would
make her "drift away." She was asymptomatic, but hospitalized in the Palliative Care
Unit of Northwestern Memorial Hospital in
anticipation of imminent death. After intensive discussion, the patient's wishes to
forego dialysis and phlebotomy were honored, but she was treated aggressively with
vitamin K subcutaneously to avoid hemorrhage. After 2 days, when she manifested no
evidence of volume overload, uremia, or coagulopathy; laboratory values were obtained
to assist with prognosis. Her blood urea nitrogen (BUN) was 44 mg/dL; creatinine was
4 mg/dL; and PT was 41.8 with INR of 11.
When told that her kidney function was
much more stable than had been anticipated,
and that for the immediate future she had no
need for life-sustaining dialysis, she became
angry and overtly suicidal. Threatening to
kill herself if left alone even for a moment,
she was transferred to the Older Adult Psychiatric service, where her diagnosis included mild dementia and depression. She
was supported in the therapeutic milieu of
the inpatient unit and started on antidepressant medication. After a week, she was no
longer suicidal, and worked with her family
on some adaptive problem solving regarding her DJD. She was discharged home. A
few months later, she experienced symptoms of volume overload, and was restarted,
with her approval, on outpatient hemodialysis.

Making the decision: conclusion

A patient and family considering withdrawal

from dialysis often find themselves in a complex
decision-making process. A palliative care/hos-

62

pice team requested to enter into that process

must be prepared for these complexities. To be
sure, there are sometimes ethically clear choices
and reconciled patients, families and physicians.
Sometimes, advance directives are both available
and useful. However, frequently vexing questions arise regarding ethics, depression, capability for decision making, and even suicidal intent.
The decision-making process is often characterized by ambivalence. Palliative care/hospice has
acquired expertise in careful inquiry and persistent listening, allowing decisions to take shape
over time. We have learned to encourage patients
to arrive at determinations that do not necessarily please us, but rather decisions that they will
"feel good about."26 In the practice of this expertise, there would seem to be an important role for
us as patients consider withdrawal from dialysis
and end-of-life care.
Furthermore, the palliative care/hospice team
would do well to consider the role of the healthcare professionals caring for the patient. Often,
over months and years, the nephrology team and
the primary care physician have been assisting
the patient to construct a fulfilling life while simultaneously coping with the impositions of
dialysis. Often this entails policing the patient's
diet and scrupulous attention to the fine-tuning
of laboratory values, activities somewhat foreign,
even perhaps abhorrent to palliative care/hospice.
A little imaginative empathy for the nephrology team is often in order. First, the palliative
care/hospice team needs to understand that it
might take the nephrology team time to resolve
their focus from maintenance into that of palliation. Furthermore, we must look to our nephrology colleagues for insights they have to offer us.
And, finally, we need to recognize the full implications of our care toward the entire "family."
"Family" says Ira Byock, is anyone "for whom it
matters."27 If this is true, then the circle of "family" often includes those colleagues who ask us
to help care for their patients.
On our unit, we watched as one nephrologist
repeatedly visited his delirious patient, assiduously attempting to find a window of lucidity
through which to assure himself that the patient
was indeed competently requesting withdrawal.
Another, convinced that the patient's withdrawal
was a blunder, urged the palliative care team to
persuade the patient to change his mind. A third,
baffled by the unexpected recovery of his patient

NEELY AND ROXE

who had been unresponsive for over two weeks,

exclaimed to the team with a beaming smile, "I
cannot NOT dialyze a miracle!"
We assist physicians in their care for their patient; also, as they struggle with difficult decisions, limitations of medicine to cure, and perhaps with personal failure, we care for them as
well. Most of this occurs "casually," simply in attentive listening and empathy offered in encounters in the hospital, on our unit and over the
phone. It is a role largely unassumed elsewhere
in the medical system and it is our privilege as
well as to our mutual benefit to meet this need.
PALLIATION OF SYMPTOMS
AFTER WITHDRAWAL

After dialysis is terminated, profound physiological aberrations predictably occur. Accumulation of toxic metabolites, imbalance in electrolytes, and inadequate fluid control may
present adverse effects. The first part of this discussion will address anticipated complications of
the "uremic death" and their management. While
uremia is a painless state, many patients experience significant pain from other comorbid diseases. The second part of this section will address
some of these common palliative care concerns,
and their treatment, in the patient with untreated
ESRD. Illustrative cases from the Northwestern
Memorial Hospital Palliative Care Unit are included.
Judicious use of pharmacology is a mainstay of
palliative medicine. Principles of drug administration change in the context of renal insufficiency. Alteration in the excretion of drugs and
their metabolites is the most important pharmacological change; however, modification of drug
distribution due to reduced albumin or pH-dependent drug binding may contribute to unexpected drug effects as well. Furthermore, a patient with longstanding renal failure often
experiences delayed gastric emptying and increased gastric pH, and with diabetes, autonomic
neuropathy, and gastroparesis. The following
discussion of symptom control suggests guidelines regarding alteration of drug dosing and/or
interval. These guidelines are just that: guidelines. Any drug, especially those with a low therapeutic index and significant renal clearance are
best approached cautiously.28 Palliative medicine
is adept at attentive titration; this practice will

63

WITHDRAWAL OF DIALYSIS

stand the provider in good stead. Table 2 provides a summary of the prescribing information
discussed in the text.
Symptoms attributable to ESRD
A uremic death is described as painless and
peaceful. As nitrogenous and other molecular
waste products accumulate, the final common
pathway for most patients is uremic encephalopathy, characterized by confusion and somnolence. There may be variable periods of lucidity,
but finally coma and death ensue over a period
of about a week. We prepared patients and families for this likelihood; they responded with family gatherings, previously forbidden foods, funeral planning, laughter and tears at the bedside.
Most of our patients died quietly and comfortably, in a manner typifying uremic death.
G.Z. was 66-year-old dentist with ESRD and
diabetes, complicated by calciphylaxis. He
was admitted to the palliative care unit for
pain control. Intensive opioid therapy was
initiated, and multiple consultants involved.
TABLE 2.

One of his treatment options included wide

excision and skin graft of his many open
wounds, with limited hope for prolonged
palliation. After intensive life review, he and
his wife declined surgery, and chose to discontinue dialysis. Subsequently, his family
transformed his room with pictures and personal objects. Always a gregarious man, he
invited friends to his bedside to laugh with
him as well as to personally request their
participation in his funeral. He startled an
administrator by personally calling to arrange for his own brain autopsy. He and his
wife subsequently enjoyed the telling of this
anecdote of forthrightness, so typical of G.Z.
Increasingly somnolent and confused, he
died quietly 7 days after his last dialysis.
The accumulation of BUN and creatinine and
other molecular waste products responsible for
coma and death may contribute to the "high
road"29 of agitated delirium before coma predominates. Delirium occurs in 70%-90% of all terminally ill patients in the last week of life30; it
usually creates great distress for patient, family,

SUMMARY OF DRUGS DISCUSSED IN THE TEXT

Indication

Drug

Alteration in ESRD

Delirium

Haloperidol

Unchanged: 0.5-1.0 mg
PO, SC, rv q 1 hr;
titrate to symptoms

EPS

Diphenhydramine

Generalized tonic-clonic
seizures

Phenytoin

Unchanged; 25-50 mg
IV q 4-6 hours
Unchanged; load with
15-20 mg/kg, maintain
with 30(M00/d
Unchanged; 0.1 mg/kg
IV
Reduce by 50%; 5-10 mg
PO, IV, QID
Unchanged; 5-10 mg PO
TID-QID; 25 mg PR
Unchanged; 0.5 mg TTD
Unchanged for acute,
short-lived events: MS
IR 2.5-5.0 mg PO;
0.5-1.0 mg SQ/IV; titrate
to relief
Details of use beyond
the scope of this chart;
see references 55,56.

Lorazepam
Emesis

Metoclopramide
Prochlorperazine

Myoclonus
Acute dyspnea

Clonazepam
Morphine

Pain

Hydrocodone
Hydromorphine
Fentanyl
Methadone

Data from references 33, 34, and 38.

ESRD, end-stage renal disease; EPS, extrapyramidal symptoms.

Notes

EPS is adverse effect;

akathisia, dystonia,
Parkinsonism within
days; treat with
diphenhydramine

For treatment of
status epilepricus
EPS is adverse effect
EPS is adverse effect
Long-term use risks
accumulation of active
morphine metabolites.
Active metabolites
suspected of causing
adverse effects.
Methadone not know to
have active metabolites

64

and staff. Terminal delirium is usually multifactoral; even a delirium with the "obvious" cause
of uremia mandates consideration of other, reversible factors such as volume depletion, adverse effects of medications, or fever.30"32
Haloperidol remains the mainstay of therapy for
delirium in the dying. It is hepatically oxidated
and its metabolites are inactive, and with the
usual dose and frequency, it may be safely used
in renal failure.33-34 As in any patient, extrapyramidal adverse effects (EPS) are a possibility.
Within a matter of days, any patient is at risk for
acute dystonia, akathisia, and Parkinsonism;
should they occur, diphenhydramine can be
given in usual doses.34
Perhaps the most distressing adverse effect of
accumulating toxic metabolites is seizures.35 Most
often these are generalized tonic-clonic, but metabolic aberrations such as uremia are known to elicit
focal seizures as well.36 Once occurring in up to
59% of ESRD, a study in the early 1980s described
generalized seizures as a late manifestation of uremia in only 10% of patients.35 It is likely that
seizures occur even less frequently now. In ESRD,
neither benzodiazepines nor phenytoin requires
adjustment in dose or frequency.35'37 Lorazepam
followed by phenytoin is now the first choice for
status epilepticus, and could accordingly be safely
used in the end-stage renal patient.38 None of our
patients experienced seizures.
Nitrogenous metabolites are the putative
agents of the nausea and vomiting of ESRD; other
contributing factors may include gastroparesis,
and often in palliative care, adverse gastrointestinal effects of opioids. A centrally acting
blocker at the level of the chemotrigger receptor
zone such as haloperidol is helpful. Metoclopramide would act similarly and also as a peripheral prokinetic. Most of metoclopramide is
conjugated in the liver, but up to 30% is excreted
in the urine unchanged. A recommended adjustment for renal failure is half the usual dose with
a 4-6-hour frequency.34 Alternatively, prochlorperazine, without adjustment, may be used.34
Again, diphenhydramine may be safely given in
event of EPS. Consideration of other causes of
nausea, particularly constipation in any dying patient, and peptic ulcer or gastrointestinal bleeding particularly in the renal patient, is imperative.
Asterixis and myoclonus may emerge in the
context of a metabolic encephalopathy such as
uremia.39 Often these are barely annoying, and
reassurance therapy suffices for patient and fam-

NEELY A N D ROXE

ily. At other times, myoclonus can be frankly distressful, physically jerking patients out of sleep.
Clonazepam effectively reduces or eliminates
myoclonus. It is hepatically metabolized, and, as
other benzodiazepines, needs no dosage adjustment for use in end-stage renal patients.34
The failing kidney can no longer manage electrolyte balance. Hyperkalemia is an anticipated,
life-threatening biochemical derangement of
ESRD. Although some authorities propose that
hyperkalemia is tolerated in renal patients,40
there are no data to support this.
Volume overload and pulmonary edema are a
potential concern for patients who have withdrawn from dialysis. Two of our patients, J.B. and
E.R. in particular, had prepalliative care courses
characterized by intensive attempts to reverse
their heart failure. However, as others have described elsewhere,25 we found neither signs nor
symptoms of volume overload or respiratory distress in any of our patients. Undoubtedly the selfimposed "fluid restriction" inherent in lethargy
helps to maintain euvolemia. Even after fluid and
sodium restrictions were lifted, our patients were
interested in consuming previously prohibited
foods only in small quantities. Should acute severe dyspnea occur, it would be appropriately
treated as a medical emergency with aggressive
use of opioids titrated to distress.41 Ultrafiltration
might be indicated in the unusual situation in
which the patient became symptomatically volume overloaded and was expected to live for
more than hours.42
PAIN CONTROL
Although end-stage renal failure might portend a peaceful, painless death, comorbid conditions are often attendant and present significant
challenges.
G.Z., mentioned previously, had severely
painful, widespread calciphylaxis, involving his trunk, lower extremities, and penis.
To avoid the accumulation of morphine
metabolites, continuous hydromorphone
was prescribed. Achieving a balance between acceptable pain relief and sedation
was difficult. On the third day after dialysis
cessation, G.Z. experienced bladder spasms;
on examination, penile necrosis from calciphylaxis had obstructed outflow. A supra-

WITHDRAWAL OF DIALYSIS

pubic catheter successfully provided palliation. As he became increasingly somnolent,

his reports of pain decreased, and he requested less hydromorphone for breakthrough pain. However, he required hydromorphone for analgesia until he died.
Calciphylaxis is the painful, spontaneous, progressive formation of cutaneous necrosis. It occurs most frequently, but not universally, in the
context of dialysis-dependent renal failure or failing kidney transplant.43 A variety of interventions, including parathyroidectomy, hyperbaric
oxygen, and early, aggressive wide excision of the
wounds have been attempted with varying reports of success. It is usually associated with
rapidly progressive decline and death, and may
be more common than previously reported.43"45
Calciphylaxis, like many cancers, is characterized by significant nociceptive soft tissue pain.
Distressing pain in end-stage disease usually
mandates opioid therapy. The elimination of
morphine is not compromised in renal insufficiency, but its active metabolites may rapidly accumulate. Morphine-6-glucuronide, a /x-receptor
agonist, clearly augments analgesia, and is implicated in respiratory depression and nausea.
Morphine-3-glucuronide, without antinociceptive affect at the /Lt-receptor, is associated with
hyperalgesia, allodynia, and myoclonus.46"48 The
use of morphine in the end-stage renal disease
patient must be balanced against the anticipated
longevity of the patient.
E.R. had ESRD complicated by endocarditis
and pericardial abscess. On the fifth day after dialysis cessation, he experienced abrupt
onset of "awful" back pain, from his right
shoulder distally down his back, consistent
with aortic dissection. He died within minutes.
In the context of catastrophic pain, and with
the anticipation of imminent death, the best opioid is the handiest opioid. Most often, that would
be liquid morphine concentratecertainly indicated, metabolites notwithstanding, in this scenario.
For use over a period of days to weeks, however, other opioids may be preferable. All opioids are hepatically metabolized. The perfect
opioid would be the one with no active metabolites. Hydromorphone is metabolized predomi-

65

nantly into
hydromorphone-3-glucuroide,
which is the suspected agent of neuroexcitation
in case reports of patients with renal failure.49"51
In rats, the antinociceptive activity of noroxycodonethe metabolite of oxycodonehas
been identified, as well as dose-dependent neuroexcitatory effects.52
Fentanyl is structurally related to meperidine,
notorious for its neuroexcitatory metabolite
normeperidine. Similarly, fentanyl is N-demethylated to become norfentanyl, excreted by the kidney and detectable in the urine a full 72 hours after a single intravenous dose.53 While the actions
of norfentanyl remain conjectural, delirium has
been associated with use of transdermal fentanyl.54
Accumulation of potentially noxious metabolites suggests cautious use of standard opioids in
ESRD. Methadone, long a second choice opioid,
has the unique characteristic of no known active
metabolite. Furthermore, urinary excretion is a
minor pathway of elimination; most occurs
through fecal excretion. Far more research is
needed on variations between patients in
bioavailability, equianalgesic dosing, and indepth study of metabolite activity. However,
methadone may emerge as the opioid of choice
for patients with renal insufficiency.55
In the meantime, the principles of palliative
opioid therapy still serve. First, that the goal of
therapy is balance of pain relief against desired
level of function. Attentive reassessment allows
careful titration. Adjuvant medication, as well as
nonpharmacological intervention may reduce the
total opioid used. Opioid rotation may be helpful if adverse effects emerge.55 And for the patient who has withdrawn from dialysis, the need
for opioids will likely be short term. Further,
there is the possibility that with it, uremic coma
will bring increasing levels of analgesia.
CONCLUSION
Daniel Callahan writes, "Of each serious illness
. . . a question should be asked and a possibility
entertained: could it be that this illness is the one
that either will be fatal, or, since some disease
must be fatal, should be allowed to be fatal?"57
For the majority of dialysis patients, the burden of intervention is far outweighed by benefit.
But since their therapy is elective, dialysis patients, their families and their healthcare teams

66

NEELY AND ROXE

may find themselves in a position to consider

Callahan's question in a particularly poignant
way. Palliative care /hospice has potentially 19.
much to offer in companioning such patients
through the decision-making process and caring
for them in the last days of life.
20.
REFERENCES
21.
1. Schreiner GE: How end-stage renal disease (ESRD)Medicare developed. Semin Nephrol 1997;17:152-159.
2. Leggat JE, Bloembergen WE, Levine G, HulbertShearon TE, Port FK: An analysis of risk factors for
withdrawal from dialysis before death. J Am Soc
Nephrol 1997;8:1755-1763.
3. Renal Physicians Association and the American Society of Nephrology: RPA/ASN Positions on the Quality
of Care at the End of Life. 1997: Dialysis and Transplantation.
4. Soltys FG, Brookins M, Seney J: Why hospice? The
case for ESRD patients and their families. Am
Nephrol Nurses Assoc 1998;25:619-622.
5. Leggat JE, Port FK. Journal club: Author's reply. Am
J Kidney Dis 1999;33:601-606.
6. Byock I: Hospice and palliative care: A parting of the
ways or a path to the future? J Palliative Med
1998;1:165-176.
7. McKegney FP, Lange P: The decision to no longer live
on chronic hemodialysis. Am J Psychiatry 128:
267-274.
8. Sullivan MD, Youngner SJ: Depression, competence,
and the right to refuse lifesaving medical treatment.
Am J Psychiatry 1994;151:971-978.
9. Chochinov HM, Wilson KG, Enns M, Lander S: Depression, hopelessness, and suicidal ideation in the
terminally ill. Psychosomatics 1998;39:366-370.
10. Kilner JF: Ethical issues in the initiation and termination of treatment. Am J Kidney Dis 1990;15:218-227.
11. Levine DZ. Managing conflict with families over dialysis discontinuation. Am J Kidney Dis 1998;5:868-883.
12. Tobe SW, Senn JS: Foregoing renal dialysis: A case
study and review of ethical issues. Am J Kidney Dis
1996;28:147-153.
13. Roy DJ: Dignity, dialysis, and dying. Dialysis Transplant 1990;19:19-21.
14. Cohen LM: Suicide, hastening death, and psychiatry.
Arch Intern Med 1998;158:1973-1976.
15. Singer PA, Siegler, M: Elective use of life-sustaining
treatments in internal medicine. Adv Intern Med
1991;36:57-79.
16. Cohen LM, McCue JD, Germain M: Denying the dying: Advance directives and dialysis discontinuation.
Psychosomatics 1997;38:27-34.
17. Sekkarie MA, Moss AH: Withholding and withdrawing dialysis: The role of physician specialty and education and patient functional status. Am J Kidney Dis
1998;31:464-472.
18. Singer PA: Nephrologists' experience with and atti-

22.
23.

24.
25.
26.
27.

28.
29.
30.
31.
32.
33.

34.

35.

tudes towards decisions to forego dialysis. J Am Soc

Nephrol 1992;2:1235-1240.
Seghal A, Galbraith A, Chesney M, Schoenfeld P,
Charles G, Lo B: How strictly do dialysis patients
want their advance directives followed? JAMA
1992;267:59-63.
Cohen LM, Germain M, Woods A, Gilman ED, McCue JD: Patient attitudes and psychological considerations in dialysis discontinuation. Psychosomatics
1993;34:395-401.
Holley JL, Nespor S, Rault R: Chronic in-center hemodialysis patients' attitudes, knowledge, and behavior towards advance directives. J Am Soc Nephrol
1993,3:1405-1408.
Schwartz RD, Perry E: Advance directives are associated with "good deaths" in chronic dialysis patients.
J Am Soc Nephrol 1993;3:1623-1630.
Chochinov HM, Wilson KG, Enns M, Mowchun N,
Lander S, Levitt M, Clinch JJ: Desire for death in the
terminally ill. Am J Psychiatry 1995;152:11851191.
Cohen LM, Fischel S, Germain M, Woods A, Braden
GL, McCue J: Ambivalence and dialysis discontinuation. Gen Hospital Psychiatry 1996;18:431^35.
Cohen LM, McCue JD, Germain M, Kjellstrand, CM:
Dialysis discontinuation: A good death? Arch Intern
Med 1995;155:42^t7.
Byock I: Dying Well: The Prospect for Growth at the End
of Life. New York: Riverhead Books, 1997.
Byock I: Reclaiming life's end: Models and strategies
for the 21 st century. Plenary speech at the 12th International Congress of Care of the Terminally 111. Montreal, 1998.
Lam YWF, Banerji S, Hatfield C, Talbert RL: Principles of drug administration in renal insufficiency.
Clin Pharmacokinetics 1997;32:30-57.
Freemont FR: Delirium and organic psychosis. In:
Freemon FR (ed): Organic Mental Disease. Jamaica, NY:
SP Medical and Scientific Books, 1981, pp. 81-94.
Massie MJ, Holland J, Glass E: Delirium in terminally
ill cancer patients. Am J Psychiatry 1983;140:
1048-1050.
De Stoutz ND, Tapper M, Fainsinger RL: Reversible
delirium in terminally ill patients. J Pain Symptom
Manage 1995;10:249-253.
Fainsinger RL, Tapper M, Bruera E: A perspective on
the management of delirium in terminally ill patients
on a palliative care unit. J Palliative Care 1993;9:4-8.
Baldessarini RJ: Drugs and the treatment of psychiatric disorders. In: Hardman JG, Limbird LE (eds): The
Pharmacological Basis of Therapeutics. New York: McGraw-Hill, 1996, p. 413.
Swan SK, Bennett WM: Use of drugs in patients with
renal failure. In: Schier RW, Gottshalk CW (eds): Diseases of the Kidney, Volume III. Boston: Little, Brown
and Co, 1996, pp. 2963-3007.
Tyler HR, Tyler KL: Neurologic Complications. In:
Eknoyan G, Knochel J (eds): The Systemic Consequences
of Renal Failure. Orlando: Grune and Stratton, 1984,
pp. 311-315.

WITHDRAWAL OF DIALYSIS

67

36. Delanty N, Vaughan CJ, French JA: Medical causes of 49. MacDonald N, Der L, Allan S, Champion P: Opioid
seizures. Lancet 1998; 352:383-390.
hyperexcitability: The application of alternate opioid
therapy. Pain 1993;53:353-355.
37. Boggs JG: Seizures in medically complex patients.
Epilepsia 1997;38(Suppl 4): S55-S59.
50. Hagan N, Thirlwell MP, Dhalwal HS, Babul N, Ha38. Lowenstein DH, Alldredge BK: Status epilepticus. N
ranyi Z, Darke AC: Steady-state pharmacokinetics of
Engl J Med 1998;338:970-976.
hydromorphone and hydromorphone-3-glucuronide
in cancer patients after immediate and controlled-re39. Caviness JN: Myoclonus. Mayo Clin Proc 1996;71:
lease hydromorphone. J Clin Pharmacol 1995;35:
679-688.
37-44.
40. Ifudu O: Care of patients undergoing hemodialysis.
N Engl J Med 1998; 339:1054-1062.
51. Babul N, Darke AC: Putative role of hydromorphone
metabolites in myoclonus [letter]. Pain 1992;51:
41. Hansen-Flaschen J: Advanced lung disease. Clin
260-261.
Chest Med 1997;18:645-655.
42. L'Abbate A, Emdin M, Piacenti M, Panichi V, Biagini 52. Leow KP, Smith MT: The antinociceptive potencies of
A, Clerico A, Del Chicca MG, Contini C, Palla R, Dooxycodone, noroxycodone and morphine after innato L: Ultrafiltration: A rational treatment for heart
tracerebroventricular administration to rats. Life Sci
failure. Cardiology 1989;76:382-390.
1994;54:1229-1236.
43. Coates T, Kirkland GS, Dymock RB, Murphy BF, 53. Peng PWH, Sandier AN: A review of the use of fentanyl analgesia in the management of acute pain in
Brealey JK, Mathew TH, Disney APS: Cutaneous
adults. Anesthesiology 1999;90:576-599.
necrosis from calcific uremic arteriolopathy. Am J
Kidney Dis 1998;32:384-391.
54. Steinberg RB, Gilman DE, Johnson F: Acute toxic
delirium in a patient using transdermal fentanyl.
44. Llach F: Cutaneous uremic arteriolopathy (calciphyAnesrh Analg 1992;75:1014-1016.
laxis): An evolving entity? Am J Kidney Dis 1998;
55. Ripamonti C, Zecca E, Bruera E: An update on the
32:514-518.
clinical use of methadone for cancer pain. Pain
45. Kane WJ, Petty PM, Sterioff S, McCarthy JT, Crotty
1997;70:109-115.
TB: The uremic gangrene syndrome: Improved healing in spontaneously forming wound following 56. Portenoy RK: Management of cancer pain. Lancet
1999;353:1695-1700.
subtotal parathyroidectomy. Plast Reconstr Surg
57. Callahan, D: Pursuing a peaceful death. Hasting Cent
1996;98:671-678.
Rep 1993;July-August 33-38.
46. Hagen NA, Foley KM, Cerbone DJ, Portenoy RK, Inturrisi CE: Chronic nausea and morphine-6-glucuronide. J Pain Symptom Manage 1991;6:125-128.
47. D'Honneur G, Gilton A, Sandouk P, Scherrmann JM,
Address reprint requests to:
Duvaldestin P: Plasma and cerebrospinal fluid conKathy Johnson Neely, M.D.
centrations of morphine and morphine glucuronides
Division of General Internal Medicine
after oral morphine: The influence of renal failure.
675 N. St. Claire
Anesthesiology 1994;81:87-93.
th
18
Floor, Suite 200
48. Sjogren P: Clinical implication of morphine metaboChicago,
IL 60611
lites. In: Bruera E, Portenoy RK (eds): Topics in Palliative Care, Volume III. New York: Oxford University
Press, 1998, pp. 163-175.
E-mail: k-neely@nwu.edu