CHAPTER 12 PALLIATIVE CARE
Practice of Geriatrics
CHAPTER 12 PALLIATIVE CARE
Diane E. Meier, M.D., R. Sean Morrison, M.D., and Judith C. Ahronheim, M.D.
Use of Life-Sustaining Treatments in the Terminally Ill
The dramatic advances in medical care that haver taken place during the past century have had the untoward side effect of changing the definition of death from the natural culmination of life into an unwanted outcome of disease and a failure of medical intervention. A century ago the majority of Americans died in their homes, but of the 2 million deaths that now occur annually in the United States, more than 80% occur within a health care institution (65% in hospitals and 15% in nursing homes). The causes of death have also changed during the past 100 years. Whereas in 1900 five of the ten leading causes of death in the geriatric population were infectious diseases, in 1995 only pneumonia and influenza remained in the top ten. Our enhanced ability to manage successfully and significantly extend life for patients with chronic disease (diabetes, congestive heart failure, coronary artery disease, renal failure) and to cure previously fatal diseases (e.g., cancer, bacterial infections, tuberculosis) has blurred the boundary between curable diseases and those that inexorably result in death. It is not an understatement to suggest that many patients and health care providers regard the prolongation of life and the cure of disease as the fundamental and exclusive goals of modern medicine. The inability of our medical culture to regard the quest for a peaceful death as a legitimate medical enterprise has rendered a profound disservice to millions of dying persons in this country and has led to a demonstrable inability to meet the needs of such persons and their families. Caring for dying patients is not easy. It requires an ability to address and balance the needs of patients and families, to communicate about intimate and existential issues, to use one’s expertise in pain and symptom management, and to coordinate and function within an interdisciplinary team.
Discussions about end-of-life care should ideally begin well before patients and physicians are confronted with a life-threatening event. A routine office visit presents an opportune time to inquire about whether an advance directive has been completed and to explore patients’ attitudes, values, and beliefs about their life and their medical care. All too often, however, advance care planning does not occur, and physicians face the challenge of exploring these issues after the patient has lost decision-making capacity with family members who may not know the patient’s wishes about care.
Breaking Bad News
Informing a patient that he or she has a fatal disease, even when advance care planning has been undertaken, is one of the most stressful tasks a physician faces. Yet the manner in which such news is delivered and the plan of care that is developed ultimately have a significant impact on the quality of life remaining for a patient. Although there are many forms of “bad news” and an equal number of clinical situations in which such news is delivered, the following framework provides some useful general guidelines for how to share such news with patients.1,2
Establishing the proper physical context before initiating the patient encounter is critical. Ideally, bad news should be delivered in person in a private area in which there will be no interruptions. One should sit down, make eye contact with the patient, and then begin the discussion by exploring the patient’s knowledge of the illness. Grasping the patient’s comprehension of the disease and his or her understanding of its impact on the future provides the physician with a useful starting point for the remainder of the conversation. Patients with a thorough understanding of the complexities of their disease require a different approach than uninformed or less sophisticated patients. The next step is to ascertain how much information the patient wants to know. For example, “Would you like me to give you the full details of the situation or is there someone else with whom you would rather I discuss it?” Although the majority of patients desire to be fully informed about their condition and the options facing them, some may not, or they may prefer family members to be so informed and to make decisions for them. Recent research suggests that this may be particularly true for patients of certain ethnic groups in the United States.3
Once it has been established how much information the patient wants, it is time to share the information with the patient—a process that has been referred to as aligning and educating.1 The discussion should begin at the patient’s level of understanding and should make use of common words and phrases. Throughout the conversation, the physician should listen to the patient’s concerns, frequently reinforce and clarify information, assess the patient’s comprehension of what has been said, and, very importantly, avoid the temptation to present too much information all at once. Finally, the clinician should respond to the patient’s feelings and organize an immediate therapeutic plan that includes specific references to the patient’s concerns and incorporates the patient’s agenda.
The question about whether to inform a patient of a diagnosis of Alzheimer’s disease has received little attention in the literature and remains controversial. Arguments against such truth-telling cite the uncertainty of the diagnosis in the early stages of the disease, the fact that patients are often brought to medical attention by concerned family members who may ask that the diagnosis not be shared, and the fact that patients with dementia may not have the capacity to fully comprehend the nature of the diagnosis or to take appropriate action based on the information. Authors favoring full disclosure cite the principle of patient autonomy and the need for patients to be aware of their diagnosis to plan appropriately for their future medical care and indeed, for their remaining life.4 The few empirical studies that have examined this question are limited by small sample sizes but have found that a majority of patients want to be told of a diagnosis of dementia.5
Eliciting Patient Preferences
Once a terminal diagnosis has been made and the information has been shared, the discussion should shift to an attempt to elicit patient preferences and goals of care. On a global level, one wants to determine whether the patient prefers to direct treatment toward achieving maximum life prolongation or whether optimization of comfort is more important given the limited life expectancy. Outlining a realistic evaluation of the available treatment options (e.g., probability and extent of response to treatment, duration and quality of extended life, anticipated side effects), identifying patients’ short- and long-term goals and needs, discovering their expectations about therapy, and evaluating their coping strategies are critical components of this discussion. The goal of this conversation is to lead the patient to an understanding of the benefits and burdens of the available treatment options and to develop an individualized plan of care that is acceptable to them and is in keeping with their values and goals. This is also an appropriate time to elicit from patients their preferences about specific life-sustaining treatments such as cardiopulmonary resuscitation, artificial nutrition and hydration, and mechanical intubation. It is also appropriate to review with them the potential indications for such therapies and to offer a realistic appraisal of the outcome in their situation. A review and revision, if necessary, of a previously completed advance directive or the execution of a new directive are natural prologues to such a discussion.
As defined by the World Health Organization, palliative care is the active total care of a patient whose disease does not respond to curative treatment. Control of pain, of other symptoms, and of psychological, social, and spiritual problems is paramount. The goal of palliative care is the achievement of the best quality of life that is possible for the patient and the family.6 Whereas the model disease for palliative care has traditionally been cancer, palliation is also appropriate for patients with other chronic illnesses such as congestive heart failure, chronic obstructive pulmonary disease, renal failure, and degenerative neurologic conditions such as Alzheimer’s disease, amyotrophic lateral sclerosis, and Parkinson’s disease. Although many speak of switching from life prolongation to palliation, or from cure to care, it is rare for such a defining moment to be clearly evident in the course of most chronic illnesses. Rather, it is more reasonable to begin a dialogue with the patient at the time of diagnosis, emphasizing that the initial focus of care will be directed toward life prolongation, if indeed this is appropriate and desired by the patient. A continual reassessment of the benefits and burdens of treatment is necessary as the disease advances so that as the ratio of benefits to burdens shifts, there will be an accompanying increase in the proportion of palliative interventions.
Patients for whom the sole focus of care is palliation can and should be offered the option of hospice care. Hospice provides interdisciplinary team care (physicians; nurses; social workers; home health aides; clergy and spiritual advisors; speech, language, occupational, and physical therapists; and volunteers) to dying patients, primarily in the home, although some hospices serve patients in long-term care facilities and residences. Since 1982, hospice care has been available under the Medicare program for all part A beneficiaries who have a life expectancy of 6 months or less and are willing to forego standard Medicare-reimbursed services in favor of palliative care. Services covered under this program include physician and nursing care, medical equipment and supplies, outpatient medications for symptom management and pain relief, short-term inpatient care, up to 4 hours per day of a home health aid, and physical and occupational therapy. Hospice is also covered by the Civilian Health and Medical Program for Uniformed Services (CHAMPUS), many third-party insurers and managed care organizations, and Medicaid in 36 states.
Despite the existence of the Medicare benefit, a substantial number of dying patients are unable to utilize hospice services. Patients living in rural areas and inner cities often have very limited access to hospice services because of a scarcity of programs in these areas. Patients without family or significant others or patients with working family members may also be unable to avail themselves of hospice. The hospice benefit typically provides for only 4 hours of custodial home health aide services, thus requiring the family to provide the remaining 20 hours—a time commitment or financial obligation that is impossible for many families to meet. Although inpatient hospice care might be considered an alternative, the financial limitations of the hospice benefit (inpatient costs must be paid out of the hospice per-diem rate) preclude this option except in situations of brief intercurrent illnesses or terminal care in the last few days of life. Lack of knowledge also contributes to the underuse of hospice palliative care services. Despite the fact that hospice has existed in this country for more than 20 years, many physicians and patients are still not aware of the availability of hospice or fail to consider hospice as an option early enough in the course of the disease. Indeed, a substantial number of hospice referrals are made after the patient has become moribund and is not able to benefit from the interdisciplinary services that hospice provides. Finally, the difficulty of determining a 6-month prognosis in dying patients, particularly those with chronic illnesses (e.g., chronic obstructive pulmonary disease, congestive heart failure) or slowly progressive neurologic disorders such as Alzheimer’s disease often inhibits referral to hospice for palliative care.
The ability to predict the course of disease accurately becomes increasingly important in terminally ill patients. Patients and families desire a reasonable estimate of life expectancy to plan for their remaining time and to get their affairs in order. Physicians must be able to prognosticate accurately to time the referral to hospice appropriately. Despite the importance of this topic, relatively little attention has been paid to prognostication in the professional literature, and there exist very few instruments that can be used to estimate life expectancy accurately for individual patients. Life expectancy for patients with dementia, motor neuron disease, acquired immune deficiency syndrome (AIDS), and advanced cardiac or pulmonary disease is highly variable, making prognostication even more difficult. Most studies that report objective prognostic data do so for very select patient populations and often do not take into account such variables as comorbid conditions or health-related quality of life. A task force from the National Hospice Organization recently reviewed the available data on prognostication and has published guidelines to aid physicians in determining when hospice referral is appropriate for patients with noncancer diagnoses.7 Until future research further defines this issue, the following guidelines have been suggested as a reasonable approach for clinicians faced with estimating life expectancy for an individual patient: (1) Identify objective measures of prognosis of available; (2) assess the validity and reliability of the measures; (3) assess the relevance and utility of the measures; and (4) evaluate how the patient’s unique clinical or social attributes modify the prognosis.8
Relief of pain and other physical symptoms is one of the central goals of palliative care. A review of multiple surveys in cancer patients demonstrates that more than half of patients have pain that is unrelieved.9 The recent SUPPORT (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment) study10 found that 50% of conscious seriously ill, hospitalized patients suffered moderate to severe pain at least half the time during their last 3 days of life. In studies of institutionalized elderly, the prevalence of pain ranges from 45% to 80%11 and is associated with musculoskeletal and degenerative joint disorders, neuropathies, ischemic vascular disease, and cancer. Significant functional consequences follow from untreated pain in this population, including depression, social withdrawal, immobility, sleep disturbance, cognitive impairment, and malnutrition, among others.11 These data demonstrate the problem of underrecognition of pain in our medical culture and emphasize the need for enhanced identification of pain and suffering through formal assessment procedures.
Pain is a subjective sensation that is influenced by many physical, emotional, and social factors and requires comprehensive assessment of these domains to permit proper treatment. Formal pain assessment procedures permit rational choices of diagnostic and therapeutic interventions, ensure that pain and other distressing symptoms are actually recognized if they are present, and convey the message to the patient that the presence of pain is important and of interest to the health care professional.
Pain assessment is particularly complex in elderly patients, who typically underreport their symptoms and usually present with several medical comorbid conditions in addition to the main disease process; in addition, the prevalence of cognitive impairment is significantly higher in this group than in the general population. Published instruments that measure pain intensity have not been well validated in elderly populations. Brevity, use of a visual-analog or numerical scale, and ease of administration are critical to the usefulness of these instruments in the practice setting. Commonly used and validated tools include the Edmonton Symptom Assessment Scale,12 the Wisconsin Brief Pain Inventory,13 and the Memorial Pain Assessment Card (Fig. 12-1).14 Patients with visual, hearing, or motor impairments may be unable to complete the instrument without help, and no valid tools for assessment of pain in the moderately or severely cognitively impaired are currently available. In all patient populations it is critical to take seriously the patient’s expression of distress, particularly if it represents a change from baseline status. Grimacing or avoidance behaviors during movement or during a physical examination are obvious signs of pain or discomfort and should prompt a search for the source of the symptom, and appropriate therapy. A change in functional capacity is another marker for the presence of physical distress when patients are unable to self report. Agitation and screaming commonly attributed to the underlying dementia in cognitively impaired patients and managed with neuroleptics may be due to untreated physical pain, a possibility that should be assessed by thorough physical examination and appropriate diagnostic studies.
Figure 12-1 Pain intensity scales. (From Jacox A, Carr D, Payne R, et al: Management of Cancer Pain. Clinical Practice Guideline No. 9. AHCPR Publication No. 94-0592. Rockville, MD, U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, 1994).
Because pain research has focused almost exclusively on young and middle-aged patients, the applications of currently available data in older adults are limited. The difficulty of balancing therapeutic benefits in terms of symptom relief with the burden of drug side effects is particularly challenging in this age group, mandating aggressive and preventive approaches to predictable side effects such as constipation and implementation of a knowledgeable and rapid response to less common side effects such as hallucinations. The essential principles of pain management, based on recommendations of the United States Agency for Health Care Policy and Research, are listed in Table 12-1.14 All clinicians should be familiar with the World Health Organization (WHO) three-step analgesic ladder, which is a validated, simple, and effective method of ensuring pain relief.15 Step 1, for mild to moderate pain, recommends the use of a nonopioid analgesic with or without an adjuvant agent and followed in patients with worsening or unresponsive pain, by the addition of a so-called weak opioid such as codeine or oxycodone (step 2). Pain that is moderate to severe at baseline or that does not respond adequately to a step 2 regimen, should be treated with larger doses of a weak opioid or a more highly potent opioid such as morphine, hydromorphone, levorphanol, or fentanyl (step 3). An adjuvant analgesic agent may be added at any step on the ladder (see later under Adjuvant Analgesics for a discussion of adjuvant agents). Frequent upward dose titration may be necessary as the disease progresses.
TABLE 12-1 ESSENTIAL PAIN MANAGEMENT GUIDELINES*
There is no ceiling or maximum permitted dose for opioid agonists (although codeine doses above 65 mg are associated with dose-limiting side effects), and some patients require large doses for control of severe pain. Development of unmanageable side effects from one agent may be addressed by lowering the dose and increasing dose frequency (if adequate analgesia is present) or by switching to a different agent and taking advantage of incomplete cross-tolerance. The new agent should be started at approximately half the equianalgesic dose of the initial drug (Table 12-2). Appropriate initial dose reductions should be employed for elderly patients, particularly those with renal or hepatic disease or congestive heart failure. In some elderly patients the risks of renal injury or hepatotoxicity associated with acetaminophen and the risks of renal injury, gastrointestinal toxicity, or bleeding associated with nonsteroidal anti-inflammatory agents may support the initial use of step 2 weak opioids (codeine or oxycodone alone) rather than the nonopioid step 1 agents.11
TABLE 12-2 DOSE EQUIVALENTS FOR OPIOID ANALGESICS IN OPIOID-NAIVE ADULTS >50 kg BODY WEIGHTa
MANAGEMENT OF OPIOID SIDE EFFECTS
Because of the high prevalence of constipation, cognitive impairment, and polypharmacy in the elderly, skilled preventive efforts and management of the side effects of opioids are critical to the successful use of analgesics in this population. The most common side effects of opioids are constipation and sedation, but confusion, nausea, dry mouth, urinary retention, hallucinations, dysphoria, euphoria, altered cognition, and the syndrome of inappropriate secretion of antidiuretic hormone are also seen. Because of their high risk of toxicity, several opioids should not be used or should be used only with great caution in elderly patients. These include meperidine (accumulation of toxic metabolites); propoxyphene (no more effective than nonopioids as an analgesic but has all the side effects and toxicities of an opioid); pentazocine (associated with agitation in elderly patients); and long-acting opioids such as methadone and transdermal fentanyl (which may accumulate to dangerous levels in elderly persons with impaired drug metabolism).
Gastrointestinal and Urologic Side Effects. Constipation is a nearly universal side effect of opioid analgesics that can lead to serious and sometimes life-threatening complications if not identified and treated appropriately. Unlike the tolerance that develops to other side effects of opioids, constipation does not diminish over time and requires preventive and combined use of stool softeners, rectal suppositories, “irritant” laxatives (e.g., bisacodyl or senna), and hyperosmotic agents (milk of magnesia, lactulose). Fecal impaction may present with a change in mental status, diarrhea, or urinary retention and should always be ruled out when there is any change in clinical status in a patient receiving opioid agents.
Nausea and vomiting may occur as a drug side effect, as a consequence of intestinal obstruction due to fecal impaction, or as a result of another underlying disease process. Antiemetic agents (ondansetron, cisapride, metoclopramide, haloperidol, prochlorperazine, scopolamine, or hydroxyzine) given at regular dosage intervals are usually effective in treating opioid-induced nausea. The side effects of phenothiazine derivatives, antihistamines, and scopolamine as antiemetics include anticholinergic effects, hallucinations, and extrapyramidal disorders, and these side effects pose relative contraindications to the use of these agents in the elderly. Ondansetron and cisapride have fewer antidopaminergic effects but are costly. Dry mouth may be treated with oral lubricants (lemon glycerin swabs), ice chips, artificial saliva, and frequent sips of fluid.
Urinary retention may result from the anticholinergic effect of the opioid, drug interactions with other anticholinergic agents, fecal impaction, prostatic hypertrophy, or progression of tumor. It is common in patients who are receiving spinal opioid analgesics. If a change in opioid type or route of administration is ineffective, intermittent urethral catheterization or an indwelling catheter may be necessary to allow continued administration of the drug.
Central Nervous System Side Effects. Sedation is a common early side effect of opioids, but patients can be reassured that tolerance to this symptom usually develops over several days. Because of the risk of confusion, falls, and other central nervous system effects of multiple drug-drug interactions in older patients, close supervision at home is necessary early in the course of therapy for opioid-naive patients. Persistent sedation should be managed by ruling out other causes of delirium (including interaction with other sedating medications), reducing the dosage and increasing the frequency of administration of the lower dose, switching to a different opioid, or adding psychostimulants (e.g., methylphenidate, 2.5 to 5 mg morning and early afternoon) if the previous measures fail. Hallucinations and confusion should also be managed by ruling out other causes of delirium, reducing the opioid dose, increasing the dosage frequency, switching to an alternative opioid, or, if these measures fail, adding a low dose of a major tranquilizer such as haloperidol or resperidol.
Respiratory depression may occur in opioid-naive patients until tolerance develops to this side effect or, rarely, when a pain stimulus suddenly ceases and sedative effects predominate. A build-up of agents with a long half-life in the elderly (methadone, sustained-release morphine sulfate, transdermal fentanyl, levorphanol), especially in patients with comorbid cardiac, renal, or liver disease, typically results in progressive sedation that will lead to respiratory depression if it is not identified and corrected early. Withholding the next one or two doses and reducing the 24-hour dosage total by about 25% will usually correct the oversedation before it is necessary to resort to opioid antagonists. If opioid antagonists (naloxone) are needed to prevent respiratory arrest, they must be slowly titrated to avoid the recurrence of severe pain and opioid withdrawal symptoms.14 When adequate pain control cannot be achieved without risking respiratory depression, careful review of the goals of therapy is mandated. If the patient is near death and is suffering greatly, the benefits of pain relief and a peaceful death usually outweigh the risk of a slightly hastened death.
Adjuvant agents are used to supplement analgesia when opioids are insufficient, to manage concurrent symptoms (such as anxiety or depression contributing to pain) or side effects (such as sedation), and to treat symptoms that are relatively unresponsive to opioid agents (such as neuropathic pain syndromes). Tricyclic antidepressants are widely used for the treatment of neuropathic pain and as supplements to opioid analgesia through their direct analgesic effects, potentiation of opiates, and mood elevation. Amitriptyline is the best studied agent for these indications, but its strong anticholinergic profile with the associated side effects of constipation, urinary retention, orthostatic hypotension, risk of falls, dry mouth, and sedation make it a difficult drug to use safely in the elderly, especially in combination with opioids.
Although there are fewer data to support their use as adjuvant analgesics, nortriptyline and desipramine have fewer associated side effects and should be tried first. Maximum benefits of therapy may not be seen for 4 to 6 weeks after initiation of a tricyclic antidepressant agent. Other commonly used adjuvant agents include corticosteroids (for appetite stimulation and treatment of brain edema, cord compression, and plexopathies), anticonvulsants for neuropathic pain (e.g., gabapentin, carbamazepine, phenytoin), systemic administration of local anesthetic (e.g., oral mexiletene or tocainide, or intravenous lidocaine), neuroleptics (e.g., methotrimeprazine [a phenothiazine with analgesic properties], haloperidol, or resperidol) for delirium, restlessness, and agitation, and psychostimulants (methylphenidate) for treatment of sedation, lack of energy, and depressed mood.
NONPHARMACOLOGIC PAIN MANAGEMENT APPROACHES
Physical therapy, occupational therapy, massage, music, psychotherapy, support groups, biofeedback, hypnosis, meditation, and relaxation techniques may significantly reduce pain and enhance the quality of life in selected patients.
Management of Other Symptoms
Anorexia. Loss of appetite is a nearly universal accompaniment of terminal illness and may not require intervention if the patient is near death and does not express a desire to eat. Earlier in the course of a serious illness, improved appetite and nutrition may contribute to higher energy levels, well-being, and functional capacity. Elderly patients should be encouraged to eat whatever is most appealing without regard to fat, sugar, and salt restrictions if possible. Appetite stimulants including corticosteroids and megestrol acetate may help in some patients, but their potential for side effects and toxicity is substantial.
Nausea and Vomiting. Common causes of nausea and vomiting (ulcers, gastritis, obstipation, drug side effects) should be sought and treated if present. Antiemetic agents are usually effective in ameliorating nausea, but the extrapyramidal and anticholinergic side effects of metoclopramide and phenothiazine derivatives often limit the use of these drugs in the elderly. Cisapride or ondansetron, although expensive, are likely to be better tolerated.
Constipation. Constipation has been discussed earlier in the section on Management of Opioid Side Effects.
Diarrhea. Often due to fecal impaction or antibiotic-associated colitis, diarrhea is a particularly distressing and exhausting symptom in the terminally ill. Once these common causes have been ruled out, kaolin-pectin or psyllium may be used effectively to reduce diarrhea. Loperamide or tincture of opium may be used if these measures fail.
Mouth Care. Dry mouth and lips are a common consequence of diminished oral intake and anticholinergic medications. Lip lubricants, lemon glycerin swabs, and artificial saliva are helpful.
The differential diagnosis of urinary incontinence in the terminally ill includes functional incontinence, overflow incontinence due to fecal impaction or other source of obstruction, vaginal atrophy, infection, and spinal cord or neurologic disease. If the postvoid residual urine volume exceeds 200 cc after impaction is ruled out, an indwelling urethral catheter may be necessary for patient comfort and easing the burden of personal care tasks for the caregivers. In the absence of obstruction, detrusor instability may be treated with the bladder smooth muscle relaxant and anticholinergic agent oxybutynin or a low dose of a tricyclic antidepressant agent, observing carefully for the development of urinary retention and other anticholinergic side effects.
Depression. Depression is discussed in detail in Chapter 28 but several aspects of depression in the terminally ill deserve review. Because of the underlying physical illness, the standard DSM IV criteria for vegetative symptoms (sleep disturbances and appetite and weight changes) are not useful. The clinician should watch for evidence of withdrawn or depressed mood, loss of interest, and suicidal ideation. Studies have shown that coexisting and untreated major depression is present in a majority of terminally ill patients with a serious desire to die. Depression may be underdiagnosed in this population because of the clinician’s belief that depression is “appropriate” in the context of a terminal illness. In fact, only 25% of terminally ill persons have evidence of depression,16 which is often unrecognized and untreated. Expression by the patient of a desire to hasten death should be openly discussed in an effort to discover the possible alternative meanings of the request to the patient and to respond as directly as possible to the identified sources of the patient’s suffering. Escalation of efforts to control distressing pain and other symptoms, a trial of antidepressant therapy, and a psychiatric consultation are appropriate initial responses. The patient should be reassured that she can discontinue life-sustaining interventions (such as chemotherapy, antibiotics, corticosteroids, and artificial nutrition and hydration) when their burdens outweigh their benefits to the patient and that any otherwise uncontrollable suffering can be managed with terminal sedation. A continued dialogue with the patient about the wish to die often reveals a change of mind as time goes on.
Standard antidepressant agents are effective for treatment of depression in the terminally ill but may have a delayed (2- to 6-week) onset of action. Psychostimulants (methylphenidate) are well tolerated and effective alternatives that have a rapid onset and a beneficial effect on energy, mood, appetite, and mental alertness. Methylphenidate (starting with 2.5-mg doses) or dextroamphetamine should be given in the early morning and early afternoon to avoid night-time sleep disturbances, and they can be used while therapy with a standard antidepressant is being initiated.
Delirium. Delirium, agitation, and confusion are extremely distressing to patients and families alike and are common among terminally ill elderly patients. Initial efforts to identify reversible medical causes (especially drug side effects and fecal impaction or urinary retention in patients receiving opioids) should precede pharmacologic intervention. Major tranquilizers (haloperidol, resperidol) in low doses are effective for treatment of agitated delirium. Benzodiazepines may be useful in treating patients with agitated delirium who respond inadequately to a neuroleptic agent, but they should not be used as first-line agents.
Anxiety. Anxiety secondary to pain or dyspnea can be managed by treating the underlying symptoms, but persistent anxiety is often a manifestation of depression. Antidepressants and benzodiazepines with a short half-life are appropriate initial steps in management.
Dementia. End-of-life care in patients with dementia requires recognition of the terminal nature of the disease and a shift from life-prolonging therapies to palliative interventions. Factors associated with a poor prognosis in dementia include dependency in all activities of daily living, urinary and fecal incontinence, nonverbal and nonambulatory state, aversive feeding behavior or malnutrition, and recent medical comorbid conditions (such as pneumonia, sepsis, and pressure sores). Avoidance of burdensome and risky interventions that are of minimal benefit in the end stages of a dementing illness (e.g., tube feeding, restraints, urethral catheters, needle sticks, and intravenous lines) are major components of palliative care in this population. Treatment of agitation requires assessment for painful stimuli before major tranquilizers are initiated.
Dyspnea refers to the subjective symptom of breathlessness. Dyspnea is mediated variously through chemoreceptors in the respiratory tract and central nervous system, stretch receptors in the chest wall skeletal muscles that participate in respiration, and the pulmonary vasculature. The quality and intensity of dyspnea depend on the individual patient and are related to ill-defined psychological factors that determine specific reactions to physical stimuli. For these reasons, the extent of dyspnea experienced by the patient may be vastly different from “objective” signs, such as respiratory rate or auscultatory abnormalities, and measurements, such as hypoxia or hypercarbia. The pathophysiology of dyspnea is complex and has been reviewed in detail elsewhere.17
Dyspnea in people with a terminal illness can be due to multiple causes, including pulmonary edema, ascites, bronchospasm, respiratory infection, pleural effusion, acidosis, anemia, and other problems. An effort should be made to determine the underlying cause, and therapy should be directed toward reversing treatable contributing factors. Invasive or noninvasive testing is often inconsistent with the goals of care for the patient, but careful clinical diagnosis is generally revealing, and the standard medical treatments, such as diuretics for pulmonary edema or bronchodilators for wheezing, may provide significant relief. In patients with malignant pleural effusion, thoracentesis may be helpful, but these effusions often recur quickly. Recurrence can be prevented with chemical pleurodesis, but, unlike simple needle thoracentesis, pleurodesis can be very painful. Thus it is essential to provide adequate analgesia during and after this procedure. Nonrespiratory causes of dyspnea such as anemia and, less often, acidosis may also respond to specific therapies.
Oxygen is an important supplement to any active treatment for dyspnea. Although the concentration of oxygen must be carefully adjusted for patients who retain carbon dioxide, the symptomatic benefits generally outweigh the potential hazards even in these circumstances. In certain situations, oxygen may reduce dyspnea even if the oxygen saturation is 90% or more. Although specific requirements, such as an oxygen saturation of less than 90%, are required by most U.S. insurance plans to pay for the apparatus used at home, the hospice benefit covers home oxygen use independent of such stipulations.
A cool breeze on the face, through an open window or a fan, may reduce dyspnea. Under experimental conditions, cool air breathed through the nose reduces the ventilatory response to carbon dioxide, which in turn may improve dyspnea by reducing ventilatory effort. This temperature-sensitive or touch-sensitive response may be mediated by reduction in brain temperature or through cold or other receptors in trigeminal nerve afferent fibers.
Systemic opioids are effective in managing dyspnea caused by a variety of factors and need not be reserved for the last days of life. The initial doses for patients with dyspnea are subject to the geriatric prescribing principles outlined in Chapter 5. Opioids act by reducing central respiratory center responsiveness to carbon dioxide tension in a dose-dependent manner. This results in decreased respiratory drive and an increase in partial pressure of carbon dioxide, which sometimes limits this approach to management of dyspnea. However, the decreased perception of dyspnea may also be centrally mediated directly, or it may due to decreased sensitivity of carbon dioxide receptors at the level of the lungs, independent of reduced respiratory drive or rate.18 Likewise, the mood-altering effects of opioids may contribute indirectly to a reduced sensation of dyspnea.
Nebulized inhaled morphine may sometimes be useful as an alternative or adjunctive treatment. The theoretical basis for this result is that the opioid receptors in the afferent sensory nerves of the respiratory tract are involved in reflex responses that can be inhibited by opioid antagonists that act only at these peripheral sites. There are anecdotal reports of dyspneic patients with end-stage chronic obstructive pulmonary disease, heart failure, cancer, and pulmonary fibrosis who have obtained significant relief of dyspnea with nebulized morphine,19 although controlled trials in diverse groups, including normals, have not shown consistent benefit. Only 10% of nebulized morphine is systemically absorbed, and medicinal use as described in published reports appears to be well tolerated. The recommended starting dose is 5 mg of morphine in 2 mL of saline at approximately 4-hour intervals, with dose increases titrated to the patient’s symptoms. The potential dose range is wide, and high doses may be required in patients who are undergoing chronic systemic opioid treatment.
Nebulized morphine should not be relied on to the exclusion of systemic opioids, particularly in patients with severe shortness of breath. The dose of systemic or nebulized opioids should be titrated according to the patient’s report, not “objective” guidelines such as respiratory rate or breathing pattern.18 A dose as low as 2 to 5 mg of parenteral morphine sulfate is often effective.
Anxiety heightens dyspnea, and relief of anxiety is very important in management. Oxygen may help to relieve anxiety and may also relieve dyspnea directly. Sedatives, including neuroleptics, benzodiazepines, and barbiturates, can be given. Frequent reassurance and companionship are very important.
Morphine and virtually all sedatives depress respiration, and this can become clinically significant. However, there may be no alternative in patients with intractable dyspnea. This dilemma is particularly profound in ventilator-dependent patients who have decided to discontinue ventilator support knowing this will result in death. Patients undergoing such terminal weaning may experience intense dyspnea if ventilator withdrawal is abrupt and if sedation is not given. Suffering is heightened if the patient is conscious and has a feeling of suffocation until the moment of death. The natural coma of carbon dioxide narcosis can sometimes be induced quickly by adjusting the respirator settings. If not, adequate sedation must be given before respirator support is withdrawn. Most, although not all, alert ventilator-dependent patients have been receiving morphine or other sedating medications and will require more than a small dose of sedative for this withdrawal process; the dose should be strictly titrated according to the patient’s needs. Although doctors and nurses may be concerned that this sedation is hastening the patient’s death, it is impossible to know whether the actual moment of death was determined by the medication or the disease.20 The ethical principle of double effect that governs this clinical situation permits use of treatments intended to relieve terminal suffering even though an unintended side effect may be a slight hastening of death. When a patient is near death and will suffer without the sedation, the benefits of the treatment outweigh its risks.
USE OF LIFE-SUSTAINING TREATMENTS IN THE TERMINALLY ILL
An important component of palliative care is the avoidance of treatments and diagnostic tests that are painful or intrusive yet fail to contribute to the overall goal of preventing suffering. However, blanket policies, such as “no antibiotics,” should be avoided, and each situation should be evaluated on its own merits. Each test or treatment should be considered negotiable. For example, a patient who wishes to avoid resuscitation may still wish to receive antibiotics; one who wishes to avoid amputation may still wish to receive intravenous hydration. A patient who cannot direct care still has the right to make these choices through an authorized surrogate decision maker.
Artificial Nutrition and Hydration in Terminal Care
Artificial nutrition and hydration (ANH) is among the most controversial life-sustaining treatments in patients with incurable illness. Much of the controversy may stem from misconceptions about the medical aspects of this treatment as well as the notion that withholding ANH is “wrong” according to religious or personally held beliefs. However, foregoing ANH at the end of life is fully consistent with the principles of palliative care.21,22
Total parenteral nutrition (TPN) has been used to treat the “anorexia cachexia” syndrome of AIDS and cancer. This is a complex metabolic abnormality associated with wasting and weakness but not due primarily to inadequate caloric intake. Extensive study of parenteral and enteral calorie replacement in these patients has failed to show that it improves the quality of life. Enteral tube feeding is commonly employed in patients with end-stage dementia or other advanced neurologic impairment who have feeding disorders, but there is no evidence that tube feeding in most of these situations is palliative. Tube feeding has a number of side effects, including pain and discomfort, tube trauma with bleeding, dislodgement leading to reinsertion of the tube or the imposition of restraints, and aspiration. Since the advent of percutaneous endoscopic gastrostomy, gastrostomy tubes have increasingly replaced nasogastric tubes for long-term feeding. However, gastrostomy and jejunostomy tubes may produce serious side effects such as peritonitis or abdominal wall cellulitis, and unusual though devastating problems such as bowel perforation and obstruction and death. These tubes may require revision or replacement, which sometimes leads to hospitalization. Finally, gastrostomy and jejunostomy tubes are often inserted for the explicit purpose of preventing aspiration pneumonia, but there is no evidence that any form of enteral tube feeding reduces the risk of pneumonia in debilitated patients.21
Enteral tube feeding is unlikely to be palliative in patients with incurable neurologic impairments or terminal illness. Theoretically, a patient with an obstructing gastroesophageal, head, or neck tumor who is alert and otherwise well enough to experience hunger might be more comfortable with complete or partial gastrostomy or jejunostomy tube feeding. Likewise, a patient with neurologic disease who visibly and consistently chokes on food despite careful spoon feeding with food of appropriate consistency might be more comfortable with tube feeding; however, such a patient must be distinguished from one who develops pneumonia from subclinical aspiration or one whose cough during feeding may be protective. In patients who visibly choke, tube feeding should prevent discomfort, apnea, and hypoxia. There is no evidence that omitting tube feeding in patients with terminal illness leads to discomfort. Even in situations such as those just cited, in which tube feeding might be temporarily palliative, the patient or his authorized surrogate has the option of forgoing tube feeding and gaining relief of symptoms by other means, such as sedation.
Doctors, nurses, and family members sometimes believe that hydration should be provided even in patients who refuse enteral tube feeding, but this practice is illogical in a patient who wishes to forgo life-sustaining treatment. Although providing hydration may sometimes comfort the caregiver, the palliative effect of hydration in patients with terminal illness is limited. Weakness and confusion are common symptoms at the end of life, and although dehydration could be a factor, such symptoms are generally multifactorial in origin and are unlikely to be reversed for any length of time with rehydration. Furthermore, an intravenous line can be intrusive and painful, especially in very ill patients with limited venous access. Hypodermoclysis, or subcutaneous infusion, is sometimes used when venous access fails, but the palliative effect is no greater. Sips of water as tolerated, ice chips, and gentle mouth care are likely to be more palliative than parenteral fluids.22
The medical aspects of artificial nutrition and hydration and the misconceptions surrounding their use have recently been reviewed in detail.21
Antibiotic Use in Terminal Illness
The treatment of infections in patients with terminal illness is complex. When antibiotics reduce symptoms caused by an infection (such as dysuria, fever, cough, or cellulitis), they are potentially palliative. However, treatment of a life-threatening infection may not be consistent with the goals of a patient who wishes to avoid treatment that is expected to prolong the process of dying. The dilemma is illustrated by the example of pneumonia, which may produce distressing symptoms that could be relieved by antibiotics, but when cured, the patient is restored to the ravages of the underlying terminal illness. In many patients, especially those with severe dementia, a serious infection may lead rapidly to coma, in which case antibiotics would not only fail to contribute to symptom relief but would reverse the natural palliation of coma. In alert patients, symptomatic treatment such as antitussives, antipyretics, or sedation is appropriate if this is consistent with the patient’s goals.
The experience of loss and separation begins well before the actual death of the patient, and the health professional can contribute a great deal toward initiating an effective bereavement process. The opportunity to provide direct care to the dying person is an important means of expressing love and connection. Overzealous delivery of professional services can take this role away from loved ones, who may begin to worry about their own “lack of expertise.” Family members may need to be explicitly encouraged to say good-bye, to try to think of the things they want to be sure to have said and to try to say them. This process should occur even if the patient is unresponsive or is unable to speak; she or he may hear and understand the tone of the words or the meaning of a touch or embrace. Families should be given information about what the actual death may be like and what various signs and symptoms mean (e.g., the “death rattle”). The actual time immediately after death is a shock no matter how surely the event is anticipated, and families need privacy and quiet together without undue staff intrusion. Respectful treatment of the body with attention paid to cultural norms and rituals, and allowing family members to participate in washing and dressing the body may all contribute to initiation of a healthy grieving process.
The main tasks of mourning—to accept the reality of the loss, to experience the pain of grief, to adjust to life without the dead loved one, and to reinvest emotional energy in living—occur in highly individual modes without a set pattern.23 Efforts to avoid the grieving process may be associated with depression, somatization, and social withdrawal, and professional facilitation of the process of mourning in such cases may be crucial to the later well-being of the family. Spouse, friends, and older relatives of an elderly person who has recently died may feel a strong sense of abandonment and isolation as fewer and fewer of their peers remain. A clinician’s acknowledgment of this reality, a willingness to begin the process of remembering the deceased by describing memories and listening to those of others, validating the family’s role in the care-giving process, and establishing clear expectations about the mourner’s future relationship with the doctors and nurses are all helpful approaches. Risk factors for a complicated bereavement process include a difficult or painful death, a sudden death, a suicide, or a death that occurred when the patient was alone. Other major losses, medical illness in the mourner, prior psychiatric history, especially of depression, and an absent or unsupportive family are other indicators of risk and should prompt an offer of professional support.23
Buckman R: How to break bad news: A guide for health care professionals. Baltimore, Johns Hopkins University Press, 1992.
Quill TE, Townsend P: Bad news: Delivery, dialogue, and dilemmas. Arch Intern Med 1991;151:463–468.
Blackhall LJ, Murphy ST, Frank G, Michel V, Azen S: Ethnicity and attitudes toward patient autonomy. JAMA 1995;274:820–825.
Drickamer MA, Lachs MS: Should patients with Alzheimer’s disease be told their diagnosis? N Engl J Med 1992;326:947–951.
Halroyd S, Snustad DG, Chalifoux SL: Attitudes of older adults on being told the diagnosis of Alzheimer’s disease. J Am Geriatr Soc 1996;44:400–403.
World Health Organization: Cancer pain relief and palliative care. Geneva, WHO, 1990.
National Hospice Organization: Medical guidelines for determining prognosis in selected non-cancer diseases. Arlington, VA, National Hospice Organization, 1995.
Christakis NA, Sachs GA: The role of prognosis in clinical decision making. J Gen Intern Med 1996;11:422–425.
Bonica J, Loeser J, Chapman C, Fordyce W: The Management of Pain. Philadelphia, Lea & Febiger, 1990, pp. 563–580.
Principal Investigators for the SUPPORT Project: A controlled trial to improve care for seriously ill hospitalized patients: The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA 1995;274:1591–1598.
Ferrell B: Pain evaluation and management in the nursing home. Ann Intern Med 1995;123:681–687.
Bruera E, Kuehn N, Miller MJ, Selmser P, Macmillan K: The Edmonton Symptom Assessment System (ESAS): A simple method for the assessment of palliative care patients. Palliative Care 1991;7:6–9.
Daut R, Cleeland C, Flanery R: The development of the Wisconsin Brief Pain Questionnaire to assess pain in cancer and other diseases. Pain 1983;17:197–210.
Jacox A, Carr D, Payne R, et al: Management of cancer pain. Clinical Practice Guideline No. 9. AHCPR Publication No. 94-0592. Rockville, MD, U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, 1994.
Jacox A, Carr DB, Payne R: New clinical practice guidelines for the management of pain in patients with cancer. N Engl J Med 1994;330:651–655.
Massie M, Holland J: Depression and the cancer patient. J Clin Psychiatry 1990;51:12–17.
Manning H, Schwartzstein R: Pathophysiology of dyspnea. N Engl J Med 1995;333:1547–1553.
Bruera E, MacEachern T, Ripamonti C, Hanson J: Subcutaneous morphine for dyspnea in cancer patients. Ann Intern Med 1993;119:906–907.
Farncombe M, Chater S: Clinical application of nebulized opioids for treatment of dyspnoea in patients with malignant disease. Support Care Cancer 1994;2:184–187.
Wilson W, Smedira N, Fink C, et al: Ordering and administration of sedatives and analgesics during the withholding and withdrawal of life support from critically ill patients. JAMA 1992;267:949–953.
Ahronheim J: Artificial nutrition and hydration in the terminally ill patient. Clin Geriatr Med 1996;12:379–391.
McCann RM, Hall WJ, Groth Juncker A: Comfort care for terminally ill patients. The appropriate use of nutrition and hydration. JAMA 1994;272:1263–1266.
Saunders C, Sykes N: The Management of Terminal Disease, 3rd ed. London, Edward Arnold, 1993.