Current Medical Diagnosis & Treatment 2015


Palliative Care & Pain Management

Michael W. Rabow, MD
Steven Z. Pantilat, MD


The focus of palliative care is to improve symptoms and quality of life at any stage of any serious illness, to support patients’ family and loved ones, and to help align patients’ care with their preferences and goals. At the end of life, palliative care often becomes the sole focus of care, but palliative care alongside cure-focused treatment is beneficial throughout the course of a serious illness, regardless of prognosis, whether the goal is to cure disease or manage it.

Palliative care includes management of pain, dyspnea, nausea and vomiting, constipation, and agitation; emotional distress, such as depression, anxiety, and interpersonal strain; and existential distress, such as spiritual crisis. While palliative care is a medical subspecialty recognized by the American Board of Medical Specialties, all clinicians should possess the basic skills to be able to manage pain; treat dyspnea; identify possible depression; communicate about important issues, such as prognosis and patient preferences for care; and help address spiritual distress. Advanced certification in palliative care is offered by the Joint Commission to hospitals providing high quality palliative services.

During any stage of illness, symptoms that cause significant suffering are a medical emergency that should be managed aggressively with frequent elicitation, continuous reassessment, and individualized treatment. While patients at the end of life may experience a host of distressing symptoms, pain, dyspnea, and delirium are among the most feared and burdensome. Management of these common symptoms is described later in this chapter. The principles of palliative care dictate that properly informed patients or their surrogates may decide to pursue aggressive symptom relief at the end of life even if, as a known but unintended consequence, the treatments preclude further unwanted curative interventions or even hasten death, although increasingly palliative care has been shown to prolong life.

Quill TE et al. Generalist plus specialist palliative care—creating a more sustainable model. N Engl J Med. 2013 Mar 28;368(13):1173–5. [PMID: 23465068]

Rabow M et al. Moving upstream: a review of the evidence of the impact of outpatient palliative care. J Palliat Med. 2013 Dec;16(12):1540–9. [PMID: 24225013]

Smith TJ et al. American Society of Clinical Oncology provisional clinical opinion: the integration of palliative care into standard oncology care. J Clin Oncol. 2012 Mar 10;30(8):880–7. [PMID: 22312101]

Temel JS et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med. 2010 Aug 19;363(8):733–42. [PMID: 20818875]



The experience of pain includes the patient’s emotional reaction to it and is influenced by many factors, including the patient’s prior experiences with pain, meaning given to the pain, emotional stresses, and family and cultural influences. Pain is a subjective phenomenon, and clinicians cannot reliably detect its existence or quantify its severity without asking the patient directly. A useful means of assessing pain and evaluating the effectiveness of analgesia is to ask the patient to rate the degree of pain along a numeric or visual pain scale (Table 5–1).

Table 5–1. Pain assessment scales.

General guidelines for management of pain are recommended for the treatment of all patients with pain. Clinicians should ask about the nature, severity, timing, location, quality, and aggravating and relieving factors of the pain. Distinguishing between neuropathic and nociceptive (somatic or visceral) pain is essential to proper tailoring of pain treatments. The goal of pain management is properly decided by the patient. Some patients may wish to be completely free of pain even at the cost of significant sedation, while others will wish to control pain to a level that still allows maximal functioning.

Chronic severe pain should be treated continuously. For ongoing pain, a long-acting analgesic can be given around the clock with a short-acting medication as needed for “breakthrough” pain. Whenever possible, the oral route of administration is preferred because it is easier to administer at home, is not painful, and imposes no risk from needle exposure. Patient-controlled analgesia (PCA) of intravenous medications can achieve better analgesia faster with less medication use and its principles have been adapted for use with oral administration.

The underlying cause of pain should be diagnosed and treated, balancing the burden of diagnostic tests or therapeutic interventions with the patient’s suffering. For example, radiation therapy for painful bone metastases or nerve blocks for neuropathic pain may obviate the need for ongoing treatment with analgesics and their side effects. Regardless of decisions about seeking and treating the underlying cause of pain, every patient should be offered prompt relief.


 Definition & Prevalence

Pain is a common problem for patients with serious illness. Up to 75% of patients dying of cancer, heart failure, COPD, or other diseases experience pain. Pain is what many people say they fear most about dying and is routinely undertreated. Joint Commission reviews of healthcare organizations now include pain management standards.

 Barriers to Good Care

Deficiencies in pain management in the seriously ill have been documented in many settings. Many clinicians have limited training and clinical experience with pain management and thus are understandably reluctant to attempt to manage severe pain. Lack of knowledge about the proper selection and dosing of analgesic medications carries with it attendant and typically exaggerated fears about the side effects of pain medications, including the possibility of respiratory depression from opioids. Most clinicians, however, can develop good pain management skills, and nearly all pain, even at the end of life, can be managed without hastening death through respiratory depression. In rare instances, palliative sedation may be necessary to control intractable suffering as an intervention of last resort.

A misunderstanding of the physiologic effects of opioids can lead to unfounded concerns on the part of clinicians, patients, or family members that patients will become addicted to opioids. While physiologic tolerance (requiring increasing dosage to achieve the same analgesic effect) and dependence (requiring continued dosing to prevent symptoms of medication withdrawal) are expected with regular opioid use, the use of opioids at the end of life for relief of pain and dyspnea is not generally associated with a risk of psychological addiction (misuse of a substance for purposes other than one for which it was prescribed and despite negative consequences in health, employment, or legal and social spheres). The risk for problematic use of pain medications is higher, however, in patients with a history of addiction or substance abuse. Yet even patients with such a history need pain relief, albeit with closer monitoring. Some patients who demonstrate behaviors associated with addiction (demand for specific medications and doses, anger and irritability, poor cooperation or disturbed interpersonal reactions) may have pseudo-addiction, defined as exhibiting behaviors associated with addiction but only because their pain is inadequately treated. Once pain is relieved, these behaviors cease. In all cases, clinicians must be prepared to use appropriate doses of opioids in order to relieve distressing symptoms for patients at the end of life.

Harms from the use of opioid analgesics, including medication diversion or death from accidental or intentional overdose, are known and significant risks. Some clinicians fear legal repercussions from prescribing the high doses of opioids sometimes necessary to control pain at the end of life. The US Food and Drug Administration (FDA) released a Risk Evaluation and Management Strategy for long-acting and extended-release opioids to help inform physicians about appropriate prescribing and reduce abuse []. Some states have special training, licensing and documentation requirements for opioid prescribing. However, governmental and professional medical groups, regulators (including the FDA), and the US Supreme Court have made it clear that appropriate treatment of pain is the right of the patient and a fundamental responsibility of the clinician. Although clinicians may feel trapped between consequences of over- or under-prescribing opioids, there remains a wide range of practice in which clinicians can appropriately treat pain. Referral to pain management or palliative care experts is appropriate whenever pain cannot be controlled expeditiously or safely by the primary clinician. In the field of chronic, nonmalignant pain management, many clinicians are using pain medication contracts and urine drug testing to help decrease the chance of abuse and diversion (see Box, Opioids for Chronic, Noncancer Pain). Clinicians who are caring for patients earlier in the course of life-threatening illness and are concerned that their patient may be misusing opioids (with serious negative consequences) can conduct periodic urine toxicology screening to confirm that the patient is taking the medication as prescribed and not using other medications.


In general, pain can be well controlled with opioid and nonopioid analgesic medications. For mild to moderate pain, acetaminophen, aspirin, and nonsteroidal anti-inflammatory drugs (NSAIDs) may be sufficient. For moderate to severe pain, opioids often are necessary. In all cases, the choice of analgesics must be guided by a careful consideration of the physiology of the pain and the benefits and risks of the particular analgesic being considered.

 Acetaminophen & NSAIDs

Appropriate doses of acetaminophen may be just as effective an analgesic and antipyretic as NSAIDs but without the risk of gastrointestinal bleeding or ulceration. Acetaminophen can be given at a dosage of 500–1000 mg orally every 6 hours, although it can be taken every 4 hours as long as the risk of hepatotoxicity is kept in mind. Hepatotoxicity is of particular concern because of how commonly acetaminophen is also an ingredient in various over-the-counter medications and because of failure to account for the acetaminophen dose included in combination acetaminophen-opioid medications such as Vicodin or Norco. With a recognition that total acetaminophen doses should not exceed 3000 mg/d long-term or 2000 mg/d for older patients and for those with liver disease, the FDA is now limiting the amount of acetaminophen available in combination analgesics.

Commonly used NSAIDs and their dosages are listed in Table 5–2. The NSAIDs are antipyretic, analgesic, and anti-inflammatory. NSAIDs increase the risk of gastrointestinal bleeding by 1.5 times normal. The risks of bleeding and nephrotoxicity from NSAIDs are both increased in elders. Gastrointestinal bleeding and ulceration may be prevented with the concurrent use of proton pump inhibitors (eg, omeprazole, 20–40 mg orally daily) or with the class of NSAIDs that inhibit only cyclooxygenase (COX)-2. Celecoxib (100 mg/d to 200 mg orally twice daily) is the only COX-2 inhibitor available and should be used with caution in patients with cardiac disease. The NSAIDs, including COX-2 inhibitors, can lead to fluid retention and exacerbations of heart failure and should be used with caution in patients with that condition. Unlike all other NSAIDs, naproxen has not been shown to increase the risk of major cardiovascular events and thus may be preferred in patients with coronary artery disease or at risk for cardiac disease. Topical formulations of NSAIDs (such as diclofenac 1.3% patch or 1% gel), placed over the painful body part for treatment of musculoskeletal pain, are associated with fewer side effects than oral administration.

Table 5–2. Acetaminophen, COX-2 inhibitors, and useful nonsteroidal anti-inflammatory drugs.

 Opioid Medications

  1. Formulations and Regimens

For many patients, opioids are the mainstay of pain management. Opioids are appropriate for severe pain due to any cause, including neuropathic pain. Opioid medications are listed in Table 5–3. Full opioid agonists such as morphine, hydromorphone, oxycodone, methadone, fentanyl, hydrocodone, and codeine are used most commonly. Hydrocodone and codeine are typically combined with acetaminophen or an NSAID, although acetaminophen in these combinations is restricted to 325 mg per unit dose due to the risk for toxicity. Extended-release hydrocodone without acetaminophen is FDA approved. Short-acting formulations of oral morphine sulfate (starting dosage 4–8 mg orally every 3–4 hours), hydromorphone (1–2 mg orally every 3–4 hours), or oxycodone (5 mg orally every 3–4 hours) are useful for acute pain not controlled with other analgesics. These same oral medications, or oral transmucosal fentanyl (200 mcg oralet dissolved in mouth) or buccal fentanyl (100 mcg dissolved in the mouth), can be used for “rescue” or “breakthrough” treatment for patients experiencing pain that breaks through long-acting medications.

Table 5–3. Useful opioid agonist analgesics.

Scott Steiger, MD

The potential harms and benefits of taking opioids daily to manage chronic, noncancer pain (CNCP) may be different than receiving opioids to manage the pain associated with a life-limiting illness in the setting of palliative and end-of-life care, as discussed elsewhere in this chapter.

CNCP is common in the United States. The Institute of Medicine estimates that CNCP afflicts 100 million adults and costs $635 billion annually in treatment and lost productivity. Efforts by medical societies, patient advocacy groups, and pharmaceutical companies led to an increased awareness of the magnitude of this problem and more aggressive treatment over the last two decades. Consequently, the prescription of opioids has undergone a dramatic increase. Hydrocodone has been the most commonly prescribed medication in the United States since at least 2008, and prescriptions for methadone for pain increased 1300% between 1997 and 2006. The total amount of opioid prescriptions filled in 2007 would have been enough for each adult in the United States to receive 700 mg of morphine.

The increased attention to treating CNCP has undoubtedly improved the lives of many patients, but the increase in prescribing opioids has had a deleterious effect on the health of the population as a whole. From 1991 to 2008, the number of deaths from opioid overdose quadrupled, and the number of patients admitted to treatment for addiction to opioids has quintupled. The Centers for Disease Control and Prevention has named prescription drug abuse an epidemic in the United States.


Research demonstrates that the beneficial effect of opioids for CNCP is modest at best and no measures have been identified to predict a good response. The improvements are generally measured in terms of a reduction in the analog pain score of 2–3 points on a 10-point scale or in improvements in the less precise outcome of function. Before considering a trial of prolonged opioid therapy, clinicians should discuss these modest possible benefits with patients in order to help set realistic goals of treatment (eg, moving from an average pain of 7 to a 4). Clinicians should also consider a deadline for reaching the patient’s goals. Since the published trials have generally lasted less than 16 weeks, it is reasonable to set a deadline before that, with some experts advocating a 90-day trial period. Limiting the time of a trial also helps prevent dose escalation to levels associated with increased risk of adverse effects, including overdose.

Many experts recommend developing a specific goal of improved function, and tracking the patient’s progress toward achieving this goal over time. But for the many patients who do not have specific, measurable goals—or who come to the clinician already taking daily opioid medication—monitoring response to treatment over time can be difficult. A useful tracking measure is the PEG, which directly patients to quantify on a scale of 0–10 the following three outcomes: Pain intensity on average over the last week; how the pain has affected their Energy level; and how much pain has impacted their General activity. Patients who do not progress toward their goal or whose PEG scores remain high over time may have pain that is unresponsive to opioids, and clinicians should reconsider the original diagnosis and use other modalities (both pharmacologic and nonpharmacologic) to provide analgesia. Without a clear analgesic benefit from opioids for CNCP, the risks may predominate, and the ineffective therapy should be discontinued.


The risks of opioids are different than most other medications. One of the most attractive features of opioids as a class is the absence specific organ toxicity attendant to most other kinds of medications. However, not every CNCP patient is a good candidate for prolonged opioid therapy. In addition to the grave risks of addiction and death to patients and the common side effects of sedation and constipation, prolonged opioid use leads to increased risk of many problems, including hypogonadism, fracture, hyperalgesia, psychosocial problems, and fraught interactions with the health care team. When considering whether to initiate or continue opioids for CNCP, clinicians should delineate these specific risks for patients to inform decision-making.

In addition to overdose, the biggest concern of many patients and clinicians is addiction. The level of risk to the individual CNCP patient is uncertain. While physical opioid dependence has been demonstrated with daily use in virtually all patients, the published risk of addiction in patients prescribed opioids for CNCP varies between less than 1% to more than 20%, with the lowest frequencies observed in studies that excluded patients with a history of substance abuse or addiction. Recent survey data found that nearly 80% of people in the United States who tried heroin for the first time in the prior year had already abused opioid medications.

Importantly, diversion of medication from patients to whom they are prescribed into other hands is another risk that must be considered when prescribing prolonged opioid therapy for CNCP. Diversion can represent opportunism, as when a patient sells medication in order to make money. Family members, acquaintances, or strangers may steal or extort medication for their own use or gain.


In an effort to limit the risks of opioids for patients with CNCP, nearly all medical society consensus panels and experts guidelines recommend using a risk assessment tool, patient-provider agreements, urine drug testing, dose limitations, and limits on the use of some medications. However, data demonstrating the effectiveness of such measures are limited.

Risk Assessment Tool

Although there are no highly predictive models for who will benefit from prolonged opioid therapy for CNCP, some models can identify patients most likely to exhibit aberrant or addictive behaviors. Most published guidelines recommend using one of these tools to determine how closely to monitor patients who are taking opioids daily, or whether to offer prolonged opioid therapy at all.

Patient-Provider Agreements

Also known as “pain contracts,” these agreements have a modest effect with 7–23% reduction in aberrant behaviors reported. They do represent an opportunity for the clinician to discuss explicitly the risks and benefits of opioids for CNCP, protocols and procedures for refills and monitoring, and consequences of worrisome behaviors.

Urine Drug Testing

Toxicology testing is a tool borrowed from addiction treatment with goals of limiting diversion and identifying risky secondary drug use. Guidelines recommend increasing frequency of testing with increased risk as determined by dose, risk assessment tool, or recent behavior. It is imperative that clinicians choose their tests appropriately and understand the limitations of toxicology testing when using this tool. Universal testing is recommended due to the inability of most clinicians to judge misuse of medication as well as documented racial differences in monitoring.

Dose Limitations

No published data support the use of daily doses of opioids above approximately 200 mg of morphine for CNCP, while risk of overdose increases approximately linearly with dose in observational studies. This has led some guideline authors—and one state, Washington—to impose limits on the amount of opioid a prescriber can offer.

Special Medication Limitations

The US Food and Drug Administration requires companies making extended-release opioid formulations to provide trainings for prescribers. Many guidelines recommend that the prescription of methadone and fentanyl be limited to specialists, and some recommend against concurrent prescription of opioids and benzodiazepines.


Prescribing opioids for patients with CNCP is fraught with challenges for clinicians, but taking the approach of carefully evaluating risks and benefits allows the opportunity for shared decision-making in individual cases. Clinical trials do not suggest that the majority of people with CNCP will benefit significantly from daily opioids, and the dramatic increase in morbidity and mortality witnessed with the increased availability of these medications warrants very careful patient selection. It is incumbent upon the clinician to provide frank advice to patients prescribed ongoing opioid therapy for CNCP and offer safer alternatives when the benefit is insufficient or the risks too high.

Bohnert AS et al. Association between opioid prescribing patterns and opioid overdose-related deaths. JAMA. 2011 Apr 6;305(13):1315–21. [PMID: 21467284]

Noble M et al. Long-term opioid management for chronic noncancer pain. Cochrane Database Syst Rev. 2010 Jan 20;(1):CD006605. [PMID: 20091598]

Nuckols TK et al. Opioid prescribing: a systematic review and critical appraisal of guidelines for chronic pain. Ann Intern Med. 2014 Jan 7;160(1):38–47. [PMID: 24217469]

Starrels JL et al. Systematic review: treatment agreements and urine drug testing to reduce opioid misuse in patients with chronic pain. Ann Intern Med. 2010 Jun 1;152(11):712–20. [PMID: 20513829]

Warner M et al. Drug poisoning deaths in the U.S., 1980–2008. NCHS Data Brief. 2011 Dec;(81):1–8. [PMID: 22617462]

For chronic stable pain, long-acting medications are preferred, such as oral sustained-release morphine (one to three times a day), oxycodone (two or three times a day), hydrocodone (two times a day), or methadone (three or four times a day) (Table 5–3). A useful technique for opioid management of chronic pain is equianalgesic dosing (Table 5–3). The dosages of any full opioid agonists used to control pain can be converted into an equivalent dose of any other opioid. This approach is helpful in estimating the appropriate dose of a long-acting opioid based on the amount of short-acting opioid required over the preceding days. For example, 24-hour opioid requirements established using short-acting opioid medications can be converted into equivalent dosages of long-acting medications or formulations. Cross-tolerance is often incomplete, however, so generally only two-thirds to three-quarters of the full, calculated equianalgesic dosage is administered initially when switching between opioid formulations.

Methadone deserves special consideration among the long-acting opioids because it is inexpensive, available in a liquid formulation, and may have added efficacy for neuropathic pain. However, equianalgesic dosing is complex because it varies with the patient’s opioid dose and caution must be used at higher methadone doses (generally > 100–150 mg/d) because of the risk of QT prolongation. Baseline ECG is recommended before starting methadone except at the very end of life where comfort is the only goal. Given the complexities of management and the increasing prevalence of methadone overdose in the United States, consultation with a palliative medicine or pain specialist may be appropriate.

Transdermal fentanyl is appropriate for patients already tolerant to other opioids for at least 1 week at a dose equivalent to at least 60 mg/d of oral morphine (equivalent to transdermal fentanyl 25 mcg/h every 72 hours) and therefore should not be used in the postoperative setting or be the first opioid used. Medications that inhibit cytochrome P450 3A4, such as ritonavir, ketoconazole, itraconazole, troleandomycin, clarithromycin, nelfinavir, nefazodone, amiodarone, amprenavir, aprepitant, diltiazem, erythromycin, fluconazole, fosamprenavir, and verapamil, and grapefruit juice can cause increased levels and duration of transdermal fentanyl. Since transdermal fentanyl can require 24–48 hours to achieve pharmacologic “steady state,” patients should be given short-acting opioids while awaiting the full analgesic effect of transdermal fentanyl, and changes in dose of transdermal fentanyl should be made no more frequently than every 6 days.

While some clinicians and patients inexperienced with the management of severe chronic pain may feel more comfortable with combined nonopioid-opioid agents, full agonist opioids are typically a better choice in patients with severe pain because the dose of opioid is not limited by the toxicities of the acetaminophen, aspirin, or NSAID component of combination preparations. There is no maximal allowable or effective dose for full opioid agonists. The dose should be increased to whatever is necessary to relieve pain, remembering that certain types of pain, such as neuropathic pain, may respond better to agents other than opioids, or to combinations of opioids with co-analgesics (see below).

While physiologic tolerance is possible with opioids, failure of a previously effective opioid dose to adequately relieve pain is usually due to worsening of the underlying condition causing pain, such as tumor growth or new metastasis in a patient with cancer. In this case, for moderate unrelieved pain, the dose of opioid can be increased by 25–50%. For severe unrelieved pain, a dose increase of 50–100% may be appropriate. The frequency of dosing should be adjusted so that pain control is continuous. Long-term dosing may then be adjusted by adding the average daily amount of short-acting opioid necessary for breakthrough pain over the preceding 72–96 hours to the long-acting medication dose. In establishing or reestablishing adequate dosing, frequent reassessment of the patient’s pain and medication side effects is necessary.

  1. Opioid Adverse Effects

As opioids are titrated upward, increasing difficulty with the side effects can be expected. Constipation is common at any dose of opioid, and tolerance to this side effect does not develop over time. Opioid-induced constipation should be anticipated and prevented in all patients (see below).

Sedation can be expected with opioids, although tolerance to this effect and to other nonconstipation side effects typically develops within 24–72 hours at a stable dose. Sedation typically appears well before significant respiratory depression. If treatment for sedation is desired, dextroamphetamine (2.5–7.5 mg orally at 8 AM and noon) or methylphenidate (2.5–10 mg orally at 8 AM and noon) may be helpful. Caffeinated beverages can also ameliorate minor opioid sedation.

Opioid-induced neurotoxicity—including myoclonus, hyperalgesia, delirium with hallucinosis, and seizures—may develop in patients who take high doses of opioids for a prolonged period. These symptoms may resolve after lowering the dose or switching opioids, especially to ones like fentanyl or methadone that do not have active metabolites. While waiting for the level of the offending opioid to fall, low doses of lorazepam, baclofen, or gabapentin may be helpful for treating myoclonus; haloperidol may be useful for treating delirium. Avoiding or correcting dehydration may be helpful for prevention and treatment of opioid-induced neurotoxicity.

Nausea due to opioids may occur with initiation of therapy and resolve after a few days. Notably, unrelieved constipation may be a more common cause of nausea in the setting of opioid use than opioid-induced nausea. Severe or persistent nausea despite treatment of constipation can be managed by switching opioids or by giving haloperidol, 0.5–4 mg orally, subcutaneously, or intravenously every 6 hours; prochlorperazine, 10 mg orally or intravenously or 25 mg rectally every 6 hours; or metoclopramide, 5–20 mg orally, subcutaneously, or intravenously before meals and bed. Ondansetron, 4–8 mg orally or intravenously every 6 hours, also relieves nausea but can contribute to constipation. Most antiemetic treatments can contribute to sedation.

Although clinicians may worry about respiratory depression with opioids, this side effect is uncommon when a low dose is given initially and titrated upward slowly. Patients at particular risk for respiratory depression include those with chronic obstructive pulmonary disease, obstructive sleep apnea, and baseline CO2 retention, those with liver or kidney or combined liver-kidney failure, and those with adrenal insufficiency or frank myxedema. Yet, even patients with severe pulmonary disease and obstructive sleep apnea can tolerate low-dose opioids, but they should be monitored carefully. Hospitalized patients with these conditions who require increased doses of opioids should be monitored with continuous pulse oximetry. Clinicians should not allow unfounded concerns about respiratory depression to prevent them from treating pain adequately.

 Medications for Neuropathic Pain

It is essential when taking a patient’s history to listen for descriptions such as burning, shooting, pins and needles, or electricity, and for pain associated with numbness. Such a history suggests neuropathic pain, which is treated with some medications not typically used for other types of pain. While opioids are effective for neuropathic pain, a number of nonopioid medications also have been found to be effective in randomized trials (Table 5–4). Successful management of neuropathic pain often requires the use of more than one effective medication.

The tricyclic antidepressants (TCAs) are good first-line therapy. Nortriptyline and desipramine are preferred because they cause less orthostatic hypotension and have fewer anticholinergic effects than amitriptyline. Start with a low dosage (10–25 mg orally daily) and titrate upward in 10 mg increments every 4 or 5 days to 50 mg. It may take several weeks for a TCA to have its full effect as a neuropathic pain analgesic.

The calcium channel alpha2-delta ligands gabapentin and pregabalin are also first-line therapies for neuropathic pain. Both medications can cause sedation, dizziness, ataxia, and gastrointestinal side effects but have no significant drug interactions. Both drugs require dose adjustments in patients with kidney dysfunction. Gabapentin should be started at low dosages of 100–300 mg orally three times a day and titrated upward by 300 mg/d every 4 or 5 days with a typical effective dose of 1800–3600 mg/d. Pregabalin should be started at 150 mg/d in two or three divided doses. If necessary, the dose of pregabalin can be titrated upward to 300–600 mg/d in two or three divided doses. Both drugs are relatively safe in accidental overdose and may be preferred over TCAs for a patient with a history of heart failure or arrhythmia or if there is a risk of suicide. Prescribing both gabapentin and morphine for neuropathic pain may provide better analgesia at lower doses than if each is used as a single agent.

The selective serotonin norepinephrine reuptake inhibitors (SSNRIs), duloxetine and venlafaxine, are also first-line treatments for neuropathic pain. Patients should be advised to take duloxetine on a full stomach because nausea is a common side effect. Duloxetine generally should not be combined with other serotonin or norepinephrine uptake inhibitors, but it can be combined with gabapentin or pregabalin. Because venlafaxine can cause hypertension and induce ECG changes, patients with cardiovascular risk factors should be carefully monitored when starting this drug.

Other medications effective for neuropathic pain include tramadol and the 5% lidocaine patch. The 5% lidocaine patch is effective in postherpetic neuralgia and may be effective in other types of localized neuropathic pain.

 Adjuvant Pain Medications & Treatments

If pain cannot be controlled without uncomfortable medication side effects, clinicians should consider using lower doses of multiple medications, as is done commonly for neuropathic pain, rather than larger doses of one or two medications. For metastatic bone pain, the anti-inflammatory effect of NSAIDs can be particularly helpful. Radiation therapy (including single-fraction treatments) and bisphosphonates may also relieve bone pain. For some patients, a nerve block can provide substantial relief, such as a celiac plexus block for pain from pancreatic cancer. Intrathecal pumps may be useful for patients with severe pain responsive to opioids but who require such large doses that systemic side effects (eg, sedation and constipation) become limiting. There is some evidence for the use of cannabinoids as analgesics.

Corticosteroids such as dexamethasone or prednisone can be helpful for patients with headache due to increased intracranial pressure, pain from spinal cord compression, metastatic bone pain, and neuropathic pain due to invasion or infiltration of nerves by tumor. Because of the side effects of long-term corticosteroid administration, they are most appropriate for short-term use and in patients with end-stage disease. Low-dose intravenous or oral ketamine has been used successfully for neuropathic and other pain syndromes refractory to opioids.


Nonpharmacologic therapies are valuable in treating pain. Hot or cold packs, massage, and physical therapy can be helpful for musculoskeletal pain. Similarly, biofeedback, acupuncture, chiropractic, meditation, music therapy, cognitive behavioral therapy, guided imagery, cognitive distraction, and framing may be of help in treating pain. Because mood and psychological issues play an important role in the patient’s perception of and response to pain, psychotherapy, support groups, prayer, and pastoral counseling can also help in the management of pain. Major depression, which may be instigated by chronic pain or may alter the response to pain, should be treated aggressively.

 When to Refer

  • Pain not responding to opioids at typical doses.
  • Neuropathic pain not responding to first-line treatments.
  • Complex methadone management issues.
  • Intolerable side effects from oral opioids.
  • Severe pain from bone metastases.
  • For a surgical or anesthesia-based procedure, intrathecal pump, or nerve block.

 When to Admit

  • Severe exacerbation of pain that is not responsive to previous stable oral opioid around-the-clock plus breakthrough doses.
  • Patients whose pain is so severe that they cannot be cared for at home.
  • Uncontrollable side effects from opioids, including nausea, vomiting, and altered mental status.

Bengoechea I et al. Opioid use at the end of life and survival in a Hospital at Home unit. J Palliat Med. 2010 Sep;13(9):1079–83. [PMID: 20799903]

Dworkin RH et al. Recommendations for the pharmacological management of neuropathic pain: an overview and literature update. Mayo Clin Proc. 2010 Mar;85(3 Suppl):S3–14. [PMID: 20194146]

Salpeter SR et al. The use of very-low-dose methadone for palliative pain control and the prevention of opioid hyperalgesia. J Palliat Med. 2013 Jun;16(6):616–22. [PMID: 23556990]

Smith EM et al; Alliance for Clinical Trials in Oncology. Effect of duloxetine on pain, function, and quality of life among patients with chemotherapy-induced painful peripheral neuropathy: a randomized clinical trial. JAMA. 2013 Apr 3;309(13):1359–67. [PMID: 23549581]



Dyspnea is the subjective experience of difficulty breathing and may be characterized by patients as tightness in the chest, shortness of breath, or a feeling of suffocation. Up to 50% of terminally ill patients may experience severe dyspnea.

Treatment of dyspnea is usually first directed at the cause (see Chapter 9). At the end of life, dyspnea is often treated nonspecifically with opioids. Immediate-release morphine given orally (2–4 mg every 4 hours) or intravenously (1–2 mg every 4 hours) treats dyspnea effectively. Doses are typically lower than would be necessary for the relief of moderate pain. Sustained-release morphine given orally at 10 mg daily appears to be safe and effective for most patients with dyspnea. Supplemental oxygen may be useful for the dyspneic patient who is hypoxic. However, a nasal cannula and face mask are sometimes not well tolerated, and fresh air from a window or fan may provide relief. Judicious use of noninvasive ventilation as well as nonpharmacologic relaxation techniques such as meditation and guided imagery may be beneficial for some patients. Benzodiazepines may be useful adjuncts for treatment of dyspnea-related agitation. Nonpharmacologic treatments, including acupuncture, may be quite effective.


Nausea and vomiting are common and distressing symptoms. As with pain, the management of nausea may be maximized by regular dosing and often requires multiple medications. An understanding of the four major inputs to the vomiting center may help direct treatment (see Chapter 15).

Vomiting associated with opioids is discussed above. When vomiting is due to stimulation of peripheral afferent nerves from the gut, offering patients small amounts of food only when they are hungry may be helpful. Nasogastric suction may provide rapid, short-term relief for vomiting associated with constipation (in addition to laxatives), gastroparesis, or gastric outlet or bowel obstruction. Prokinetic agents, such as metoclopramide (5–20 mg orally or intravenously four times a day), can be helpful in the setting of partial bowel obstruction. Transdermal scopolamine (1.5-mg patch every 3 days) can reduce peristalsis and reduce cramping pain, and ranitidine (50 mg intravenously every 6 hours) and octreotide (starting at 50–100 mcg subcutaneously every 8 hours) or as continuous intravenous or subcutaneous infusion, beginning at 10–20 mcg/h) can reduce gut secretions improving nausea and vomiting. High-dose corticosteroids (eg, dexamethasone, 20 mg orally or intravenously daily in divided doses) can be used in refractory cases of nausea or vomiting or when it is due to bowel obstruction or increased intracranial pressure.

Vomiting due to disturbance of the vestibular apparatus may be treated with anticholinergic and antihistaminic agents (including diphenhydramine, 25 mg orally or intravenously every 8 hours, or scopolamine, 1.5-mg patch every 3 days).

Benzodiazepines can be effective in preventing the anticipatory nausea associated with chemotherapy. Finally, many patients find medical cannabis (in particular strains with high cannabidiol [CBD] content) or dronabinol (2.5–20 mg orally every 4–6 hours) helpful in the management of nausea and vomiting.


Given the frequent use of opioids, poor dietary intake, and physical inactivity, constipation is a common problem in the seriously ill or dying. Clinicians must inquire about any difficulty with hard or infrequent stools. Constipation is an easily preventable and treatable cause of discomfort, distress, and nausea and vomiting (see above and Chapter 15).

Constipation may be prevented or relieved if patients can increase their activity and their intake of fluids. Simple considerations such as privacy, undisturbed toilet time, and a bedside commode rather than a bedpan may be important for some patients.

For patients taking opioids, anticipating and preventing constipation is key. A prophylactic bowel regimen with a stimulant laxative (senna or bisacodyl) should be started when opioid treatment is begun. Table 15–4 lists other agents (including polyethylene glycol) that can be added as needed. Docusate, a stool softener, adds little to senna in hospitalized patients and is not recommended. Methylnaltrexone, a subcutaneous medication, is a peripherally acting mu-receptor antagonist and is available for severe, unrelieved, opioid-induced constipation.

Table 5–4. Pharmacologic management of neuropathic pain.


Fatigue is a distressing symptom and is the most common complaint among cancer patients. Specific abnormalities that can contribute to fatigue, including anemia, hypothyroidism, hypogonadism, cognitive and functional impairment, and malnutrition, should be corrected. Because pain, depression, and fatigue often coexist in patients with cancer, pain and depression should be managed adequately in the setting of patients with fatigue. Fatigue from medication adverse effects and polypharmacy is common and should be addressed. For nonspecific fatigue, exercise and physical rehabilitation may be most effective. Low doses of psychostimulants, such as methylphenidate 5–10 mg orally in the morning and afternoon or modafinil 200 mg orally in the morning, can be effective. Corticosteroids may have a short-term benefit. American Ginseng has been shown to be effective for cancer-related fatigue.


Many patients die in a state of delirium—a disturbance of consciousness and a change in cognition that develops over a short time and is manifested by misinterpretations, illusions, hallucinations, disturbances in the sleep-wake cycle, psychomotor disturbances (eg, lethargy, restlessness), and mood disturbance (eg, fear, anxiety). Delirium may be hyperactive, hypoactive, or mixed. Agitated delirium at the end of life has been called terminal restlessness.

Careful attention to patient safety and nonpharmacologic strategies to help the patient remain oriented (clocks, calendars, a familiar environment, reassurance and redirection from caregivers) may be sufficient to prevent or manage mild delirium. Some delirious patients may be “pleasantly confused,” and a decision by the patient’s family and the clinician not to treat delirium may be considered.

More commonly, however, delirium at the end of life is distressing to patients and family and requires treatment. Delirium may interfere with the family’s ability to feel comforting to the patient and may prevent a patient from being able to recognize and report important symptoms. While there are many reversible causes of delirium (see Chapter 25), identifying and correcting the underlying cause at the end of life is often complex because a patient may have many possible causes. When the cause of delirium cannot be identified, treated, or corrected rapidly enough, delirium may be treated symptomatically with neuroleptic agents, such as haloperidol (1–10 mg orally, subcutaneously, intramuscularly, or intravenously twice or three times a day) or risperidone (1–3 mg orally twice a day). The benefits of neuroleptic agents in the treatment of agitation must be weighed carefully against potential harms, based on evidence showing an association between antipsychotic medications and increased mortality for older adults with dementia. The role of hydration as a treatment of delirium at the end of life is unclear but intravenous fluids do not appear to attenuate most instances of delirium. When delirium is refractory to treatment and remains intolerable, sedation may be required to provide relief and may be achieved rapidly with midazolam (0.5–5 mg/h subcutaneously or intravenously) or barbiturates (especially useful in the outpatient setting).

Table 5–5. Suggestions for the delivery of bad news.

Candy B et al. Laxatives or methylnaltrexone for the management of constipation in palliative care patients. Cochrane Database Syst Rev. 2011 Jan 19;(1):CD003448. [PMID: 21249653]

Currow DC et al. Once-daily opioids for chronic dyspnea: a dose increment and pharmacovigilance study. J Pain Symptom Manage. 2011 Sep;42(3):388–99. [PMID: 21458217]

Davidson PM et al. Update on the role of palliative oxygen. Curr Opin Support Palliat Care. 2011 Jun;5(2):87–91. [PMID: 21532348]

Kerr CW et al. Effects of methylphenidate on fatigue and depression: a randomized, double-blind, placebo-controlled trial. J Pain Symptom Manage. 2012 Jan;43(1):68–77. [PMID: 22208450]

Peuckmann V et al. Pharmacological treatments for fatigue associated with palliative care. Cochrane Database Syst Rev. 2010 Nov 10;(11):CD006788. [PMID: 21069692]

Suzuki M et al. A randomized, placebo-controlled trial of acupuncture in patients with chronic obstructive pulmonary disease (COPD): the COPD-acupuncture trial (CAT). Arch Intern Med. 2012 Jun 11;172(11):878–86. [PMID: 22905352]

Tarumi Y et al. Randomized, double-blind, placebo-controlled trial of oral docusate in the management of constipation in hospice patients. J Pain Symptom Manage. 2013 Jan;45(1):2–13. [PMID: 22889861]


In the United States, nearly 2.5 million people die each year. Caring for patients at the end of life is an important responsibility and a rewarding opportunity for clinicians. From the medical perspective, the end of life may be defined as that time when death—whether due to terminal illness or acute or chronic illness—is expected within weeks to months and can no longer be reasonably forestalled by medical intervention. Palliative care at the end of life focuses on relieving distressing symptoms and promoting quality of life (as with all other stages of illness). For patients at the end of life, palliative care may become the sole focus of care.

 Prognosis at the End of Life

Clinicians must help patients understand when they are approaching the end of life. Most patients want prognostic information. This information influences patients’ treatment decisions and may change how they spend their remaining time, but does not negatively impact patient survival. One-half or more of patients do not understand that many treatments they might be offered are palliative and not curative. Patients require support for distress that may accompany discussions of prognostic information.

While certain diseases such as cancer are more amenable to prognostic estimates regarding the time course to death, the other common causes of mortality in the United States—including heart disease, stroke, chronic lung disease, and dementia—have more variable trajectories and difficult to predict prognoses. Even for patients with cancer, clinician estimates of prognosis are often inaccurate and generally overly optimistic. Nonetheless, clinical experience, epidemiologic data, guidelines from professional organizations (eg, the National Hospice and Palliative Care Organization), and computer modeling and prediction tools (eg, the Palliative Performance Scale or may be used to help patients identify the end period of their lives. Clinicians can also ask themselves “Would I be surprised if this patient died in the next year?” to determine whether a discussion of prognosis would be appropriate. If the answer is “no,” then the clinician should initiate a discussion. Recognizing that patients may have different levels of comfort with prognostic information, clinicians can introduce the topic by simply saying, “I have information about the likely time course of your illness. Would you like to talk about it?”

 Expectations About the End of Life

Patients’ experiences at the end of life are influenced by their expectations about how they will die and the meaning of death. Many people fear how they will die more than death itself. Patients report fears of dying in pain or of suffocation, of loss of control, indignity, isolation, and of being a burden to their families. All of these anxieties may be ameliorated with good supportive care provided by an attentive group of caretakers.

Death is often regarded by clinicians, patients, and families as a failure of medical science. This attitude can create or heighten a sense of guilt about the failure to prevent dying. Both the general public and clinicians often are complicit in denying death, treating dying persons merely as patients and death as an enemy to be battled furiously in hospitals rather than as an inevitable outcome to be experienced as a part of life at home. As a result, approximately 75–80% of people in the United States die in hospitals or long-term care facilities.

Even when the clinician continues to pursue cure of potentially reversible disease, relieving suffering, providing support, and helping the patient prepare for death are foremost considerations. Patients at the end of life and their families identify a number of elements as important to quality end-of-life care: managing pain and other symptoms adequately, avoiding inappropriate prolongation of dying, preserving dignity, preparing for death, achieving a sense of control, relieving the burden on others, and strengthening relationships with loved ones.

 Communication & Care of the Patient

Caring for patients at the end of life requires the same skills clinicians use in other tasks of medical care: diagnosing treatable conditions, providing patient education, facilitating decision-making, and expressing understanding and caring. Higher-quality communication is associated with greater satisfaction and awareness of patient wishes. Clinicians must become proficient at delivering bad news and then dealing with its consequences (Table 5–5). When needed, the use of a professional interpreter can facilitate clear communication and help broker cultural issues.

Three further obligations are central to the clinician’s role at this time. First, he or she must work to identify, understand, and relieve physical, psychological, social, and spiritual distress or suffering. Second, clinicians can serve as facilitators or catalysts for hope. While hope for a particular outcome may fade (such as cure of advanced cancer following exhaustive conventional and experimental treatments), it can be redefined by the patient’s belief of what is still possible. Although expecting a “miraculous cure” may be simplistic, hope for relief of pain, for reconciliation with loved ones, for discovery of meaning, and for spiritual transformation is realistic at the end of life. With such questions as “What is still possible now for you?” “When you look to the future, what do you hope for?” “What good might come of this?” clinicians can help patients uncover hope, explore meaningful and realistic goals, and develop strategies to realize them.

Finally, dying patients’ feelings of isolation and fear demand that clinicians assert that they will care for the patient throughout the final stage of life. The promise of nonabandonment is perhaps the central principle of end-of-life care and is a clinician’s pledge to an individual patient to serve as a caring partner, a resource for creative problem-solving and relief of suffering, a guide during uncertain times, and a witness to the patient’s experiences—no matter what happens. Clinicians can say to a patient, “I will care for you whatever happens.” Dying patients need their clinicians to offer their presence—not necessarily the ability to solve all problems but rather a commitment to recognize and receive the patients’ difficulties and experiences with respect and empathy. At its best, the patient-clinician relationship can be a covenant of compassion and a recognition of common humanity.

 Caring for the Family

In caring for patients at the end of life, clinicians must appreciate the central role played by family, friends, and partners and often must deal with strong emotions of fear, anger, shame, sadness, and guilt experienced by those individuals. While significant others may support and comfort a patient at the end of life, the threatened loss of a loved one may also create or reveal dysfunctional or painful family dynamics. Furthermore, clinicians must be attuned to the potential impact of illness on the patient’s family: substantial physical caregiving responsibilities and financial burdens as well as increased rates of anxiety, depression, chronic illness, and even mortality. Family caregivers, typically women, commonly provide the bulk of care for patients at the end of life, yet their work is often not acknowledged or compensated.

Clinicians can help families confront the imminent loss of a loved one (Table 5–6) and often must negotiate amid complex and changing family needs. Identifying a spokesperson for the family, conducting family meetings, allowing all to be heard, and providing time for consensus may help the clinician work effectively with the family. Providing good palliative care to the patient can reduce the risk of depression and complicated grief in loved ones after the patient’s death.

Table 5–6. Clinician interventions helpful to families of dying patients.

 Clinician Self-Care

Many clinicians find caring for patients at the end of life to be one of the most rewarding aspects of practice. However, working with the dying requires tolerance of uncertainty, ambiguity, and existential challenges. Clinicians must recognize and respect their own limitations and attend to their own needs in order to avoid being overburdened, overly distressed, or emotionally depleted.

Candy B et al. Interventions for supporting informal caregivers of patients in the terminal phase of a disease. Cochrane Database Syst Rev. 2011 Jun 15;(6):CD007617. [PMID: 21678368]

El-Jawahri A et al. Associations among prognostic understanding, quality of life, and mood in patients with advanced cancer. Cancer. 2014 Jan 15;120(2):278–85. [PMID: 24122784]

Hudson PL et al. A systematic review of psychosocial interventions for family carers of palliative care patients. BMC Palliat Care. 2010 Aug 5;9:17. [PMID: 20687960]

Kearney MK et al. Self-care of physicians caring for patients at the end of life: “Being connected... a key to my survival”. JAMA. 2009 Mar 18;301(11):1155–64. [PMID: 19293416]

Lennes IT et al. Predictors of newly diagnosed cancer patients’ understanding of the goals of their care at initiation of chemotherapy. Cancer. 2013 Feb 1;119(3):691–9. [PMID: 23011902]

Moon JR et al. Short- and long-term associations between widowhood and mortality in the United States: longitudinal analyses. J Public Health (Oxf). 2013 Oct 28. [Epub ahead of print.] [PMID: 24167198]

Pantilat SZ. Communicating with seriously ill patients: better words to say. JAMA. 2009 Mar 25;301(12):1279–81. [PMID: 19318656]

Sinclair S. Impact of death and dying on the personal lives and practices of palliative and hospice care professionals. CMAJ. 2011 Feb 8;183(2):180–7. [PMID: 21135081]

Weeks JC et al. Patients’ expectations about effects of chemotherapy for advanced cancer. N Engl J Med. 2012 Oct 25;367(17):1616–25. [PMID: 23094723]

 Decision-Making, Advance Care Planning, & Advance Directives

Patients deserve to have their health care be consistent with their preferences and goals of care. Well-informed, competent adults have a right to refuse medical intervention even if this is likely to result in death. Many are willing to sacrifice some quantity of life to protect a certain quality of life. In order to promote patient autonomy, clinicians are obligated to inform patients about the risks, benefits, alternatives, and expected outcomes of medical interventions such as cardiopulmonary resuscitation (CPR), mechanical ventilation, hospitalization and ICU care, and artificial nutrition and hydration.Advance directives are oral or written statements made by patients when they are competent that are intended to guide care should they become incompetent. Advance directives allow patients to project their autonomy into the future and are an important part of advance care planning—a process whereby clinicians help patients match treatments and care to their goals and values. Advance directives take effect when the patient can no longer communicate his or her preferences directly. While oral statements about these matters are ethically binding, they are not legally binding in all states. State-specific advance directive forms are available from a number of sources, including

Clinicians should facilitate the process for all patients—ideally, well before the end of life—to consider their preferences, to appoint a surrogate and talk to that person about their preferences, and to formulate an advance directive. Most patients with a serious illness have already thought about end-of-life issues, want to discuss them with their clinician, want the clinician to bring up the subject, and feel better for having had the discussion. Patients who have such discussions with their clinicians are perceived by their family as having a better quality of life at the end of life, are less likely to die in the hospital, and more likely to utilize hospice care. With advance care planning discussions, patients’ loved ones are less likely to suffer from depression during bereavement.

One type of advance directive is the Durable Power of Attorney for Health Care (DPOA-HC) that, in addition to documenting patient preferences for care, allows the patient to designate a surrogate decision-maker. The DPOA-HC is particularly useful because it is often difficult to anticipate what decisions will need to be made. The responsibility of the surrogate is to provide “substituted judgment”—to decide as the patient would, not as the surrogate wants. Clinicians should encourage patients to talk with their surrogates about their preferences generally and about scenarios that are likely to arise, such as the need for mechanical ventilation in a patient with end-stage emphysema. Clear clinician communication is important to correct misunderstandings and address biases. In the absence of a designated surrogate, clinicians usually turn to family members or next of kin. Despite regulations requiring health care institutions to inform patients of their rights to formulate an advance directive, only a minority of people in the United States (including clinicians themselves as well as patients with advanced illness) have completed them. POLST (Physician Orders for Life-Sustaining Treatment) forms are physician orders that accompany patients wherever they are cared for—home, hospital, or nursing home—and are an increasingly widely used complement to advance directives for patients at the end of life.

 Do Not Attempt Resuscitation Orders

As part of advance care planning, clinicians can elicit patient preferences about CPR. Most patients and many clinicians overestimate the chances of success of CPR. Only about 15% of all patients who undergo CPR in the hospital survive to hospital discharge. Moreover, among certain populations—especially those with multisystem organ failure, metastatic cancer, and sepsis—the likelihood of survival to hospital discharge following CPR is virtually nil. Patients may ask their clinician to write an order that CPR not be attempted on them.

For some patients at the end of life, decisions about CPR may not be about whether they will live but about how they will die. Clinicians should correct the misconception that withholding CPR in appropriate circumstances is tantamount to “not doing everything” or “just letting someone die.” While respecting the patient’s right ultimately to make the decision—and keeping in mind their own biases and prejudices—clinicians should offer explicit recommendations about DNAR orders and protect dying patients and their families from feelings of guilt and from the sorrow associated with vain hopes. Clinicians should discuss what interventions will be continued and started to promote quality of life rather than focusing only on what is not to be done. For patients with implantable cardioverter defibrillators (ICDs), clinicians must also address issues of turning off these devices as death approaches to prevent them from discharging during the dying process.

 Hospice & Other Palliative Care Institutions

In the United States, hospice is a specific type of palliative care service focused on comprehensively addressing the needs of the dying. In the United States, 45% of people who die use hospice with about 66% of hospice patients die at home where they can be cared for by their family and visiting hospice staff. Hospice care can also be provided in hospitals and institutional residences. As is true of all types of palliative care, hospice emphasizes individualized attention, human contact, and an interdisciplinary team approach. Hospice care can include arranging for respite for family caregivers and assisting with referrals for legal, financial, and other services. Patients in hospice require a physician who is responsible for their care. Primary care clinicians are strongly encouraged to fulfill this role.

Hospice care is highly rated by families and has been shown to increase patient satisfaction, to reduce costs (depending on when patients are referred to hospice care), and to decrease family caregiver mortality. Despite evidence that suggests that hospice care does not increase mortality and may even extend life, hospice care tends to be used very near the end of life. The average length of stay in hospices in the United States is 69 days, with 36% of patients dying within 7 of beginning hospice care.

In the United States, most hospice organizations require clinicians to estimate the patient’s probability of survival to be < 6 months, since this is a criterion for eligibility to receive Medicare or other insurance coverage. Regrettably, the hospice benefit can be difficult to provide to people who are homeless or isolated or who have terminal prognoses that are difficult to quantify.

 Cultural Issues

The individual patient’s experience of dying occurs in the context of a complex interaction of personal, philosophic, and cultural influences. Various religious, ethnic, gender, class, and cultural traditions help determine patients’ styles of communication, comfort in discussing particular topics, expectations about dying and medical interventions, and attitudes about the appropriate disposition of dead bodies. While each patient is an individual, being sensitive to a person’s cultural beliefs and respecting ethnic traditions are important responsibilities of the clinician caring for a patient at the end of life, especially when the cultures of origin of the clinician and patient differ. A clinician may ask a patient, “What do I need to know about you and your beliefs that will help me take care of you?”

Bischoff KE et al. Advance care planning and the quality of end-of-life care in older adults. J Am Geriatr Soc. 2013 Feb;61(2):209–14. [PMID: 23350921]

Wendler D et al. Systematic review: the effect on surrogates of making treatment decisions for others. Ann Intern Med. 2011 Mar 1;154(5):336–46. [PMID: 21357911]

Zier LS et al. Surrogate decision makers’ interpretation of prognostic information: a mixed-methods study. Ann Intern Med. 2012 Mar 6;156(5):360–6. [PMID: 22393131]

 Nutrition & Hydration

People approaching the end of life often lose their appetite and most ultimately stop eating and drinking. The anorexia-cachexia syndrome frequently occurs in patients with advanced cancer, and cachexia is a common and poor prognostic sign in patients with heart failure. Ill people often have no hunger with total caloric deprivation, and the associated ketonemia produces a sense of well-being, analgesia, and mild euphoria. Although it is unclear to what extent withholding hydration at the end of life creates an uncomfortable sensation of thirst, any such sensation is usually relieved by simply moistening the dry mouth. Ice chips, hard candy, swabs, popsicles, or minted mouthwash may be effective. Although this normal process of diminishing oral intake and accompanying weight loss is very common, it can be distressing to patients and families who may associate the offering of food with compassion and love and lack of eating with distressing images of starvation. In response, patients and families often ask about supplemental enteral or parenteral nutrition.

Unfortunately supplemental, artificial nutrition and hydration offers little benefit to those at the end of life and rarely achieves patient and family goals. For example, although tube feedings are often considered in patients with advanced dementia who aspirate, they do not prevent aspiration pneumonia and there is debate about whether artificial nutrition prolongs life in the terminally ill. Furthermore, force feeding may cause nausea and vomiting in ill patients, and eating can lead to diarrhea in the setting of malabsorption. Artificial nutrition and hydration may increase oral and airway secretions as well as increase the risk of choking, aspiration, and dyspnea; ascites, edema, and effusions may be worsened. Nasogastric and gastrostomy tube feeding and parenteral nutrition impose risks of infection, epistaxis, pneumothorax, electrolyte imbalance, and aspiration—as well as the need to physically restrain the delirious patient to prevent dislodgment of catheters and tubes.

Individuals at the end of life have a right to refuse all nutrition and hydration. Because they may have deep social and cultural significance for patients, families, and clinicians themselves, decisions about artificial nutrition and hydration are not simply medical. Eliciting perceived goals of artificial nutrition and hydration and correcting misperceptions can help patients and families make clear decisions. Family and friends can be encouraged to express their love and caring in ways other than intrusive attempts at force feeding or hydration.

 Withdrawal of Curative Efforts

Requests from appropriately informed and competent patients or their surrogates for withdrawal of life-sustaining interventions must be respected. Limitation of life sustaining interventions prior to death is an increasingly common practice in intensive care units. The withdrawal of life-sustaining interventions such as mechanical ventilation must be approached carefully to avoid needless patient suffering and distress for those in attendance. Clinicians should educate the patient and family about the expected course of events and the difficulty of determining the precise timing of death after withdrawal of interventions. Sedative and analgesic agents should be administered to ensure patient comfort even at the risk of respiratory depression or hypotension. Scopolamine (10 mcg/h subcutaneously or intravenously, or a 1.5-mg patch every 3 days), glycopyrrolate (1 mg orally every 4 hours), or atropine (1% ophthalmic solution, 1 or 2 drops sublingually as often as every hour) can be used for controlling airway secretions and the resultant “death rattle.” A guideline for withdrawal of mechanical ventilation is provided in Table 5–7.

Table 5–7. Guidelines for withdrawal of mechanical ventilation.

 Psychological, Social, & Spiritual Issues

Dying is not exclusively or even primarily a biomedical event. It is an intimate personal experience with profound psychological, interpersonal, and existential meanings. For many people at the end of life, the prospect of impending death stimulates a deep and urgent assessment of their identity, the quality of their relationships, and the meaning and purpose of their existence.

  1. Psychological Challenges

In 1969, Elisabeth Kübler-Ross identified five psychological stages or patterns of emotions that patients at the end of life may experience: denial and isolation, anger, bargaining, depression, and acceptance. Not every patient will experience all these emotions, and typically not in an orderly progression. In addition to these five stages are the perpetual challenges of anxiety and fear of the unknown. Simple information, listening, assurance, and support may help patients with these psychological challenges. In fact, patients and families rank emotional support as one of the most important aspects of good end-of-life care. Psychotherapy and group support may be beneficial as well.

Despite the significant emotional stress of facing death, clinical depression is not normal at the end of life and should be treated. Cognitive and affective signs of depression (such as hopelessness or helplessness) may help distinguish depression from the low energy and other vegetative signs common with end-stage illness. Although traditional antidepressant treatments such as selective serotonin reuptake inhibitors are effective, more rapidly acting medications such as dextroamphetamine or methylphenidate (in doses used for sedation described earlier in this chapter) may be particularly useful when the end of life is near or while waiting for other antidepressant medication to take effect. Oral ketamine is emerging as a promising, rapid-onset treatment for anxiety and depression at the end of life. Some research suggests a mortality benefit from treating depression in the setting of serious illness.

  1. Social Challenges

At the end of life, patients should be encouraged to discharge personal, professional, and business obligations. This might include completing important work or personal projects, distributing possessions, writing a will, and making funeral and burial arrangements. The prospect of death often prompts patients to examine the quality of their interpersonal relationships and to begin the process of saying goodbye (Table 5–8). Dying may intensify a patient’s need to feel cared for by the clinician and the need for clinician empathy and compassion. Concern about estranged relationships or “unfinished business” with significant others and interest in reconciliation may become paramount at this time.

Table 5–8. Five statements often necessary for the completion of important interpersonal relationships.

  1. Spiritual Challenges

Spirituality is the attempt to understand or accept the underlying meaning of life, one’s relationships to oneself and other people, one’s place in the universe, one’s legacy, and the possibility of a “higher power” in the universe. Spirituality is distinguished from particular religious practices or beliefs and is generally considered a universal human concern.

Unlike physical ailments such as infections and fractures, which usually require a clinician’s intervention to be treated, the patient’s spiritual concerns often require only a clinician’s attention, listening, and witness. Clinicians might choose to inquire about the patient’s spiritual concerns and ask whether the patient wishes to discuss them. For example, asking, “How are you within yourself?” or “Are you at peace?” communicates that the clinician is interested in the patient’s whole experience and provides an opportunity for the patient to share perceptions about his or her inner life. Questions that might constitute an existential “review of systems” are presented in Table 5–9. Formal legacy work and dignity therapy have been shown to be effective in improving quality of life and spiritual well-being. Patients should be supported by clinicians, but engagement with religious communities and professional chaplains as part of comprehensive care are key.

Table 5–9. An existential review of systems.

While dying may be a period of inevitable loss of physical functioning, the end of life also offers an opportunity for psychological, interpersonal, and spiritual development. Individuals may grow—even achieve a heightened sense of well-being or transcendence—in the process of dying. Through listening, support, and presence, clinicians may help foster this learning and be a catalyst for this transformation. Rather than thinking of dying simply as the termination of life, clinicians and patients may be guided by a developmental model of dying that recognizes a series of lifelong developmental tasks and landmarks and allows for growth at the end of life.

Balboni TA et al. Provision of spiritual support to patients with advanced cancer by religious communities and associations with medical care at the end of life. JAMA Intern Med. 2013 Jun 24;173(12):1109–17. [PMID: 23649656]

Chochinov HM et al. Effect of dignity therapy on distress and end-of-life experience in terminally ill patients: a randomised controlled trial. Lancet Oncol. 2011 Aug;12(8):753–62. [PMID: 21741309]

Rayner L et al. Antidepressants for the treatment of depression in palliative care: systematic review and meta-analysis. Palliat Med. 2011 Jan;25(1):36–51. [PMID: 20935027]


After the death of a hospitalized patient, the clinician is called upon to perform a number of tasks, both required and recommended. The clinician must plainly and directly inform the family of the death, complete a death certificate, contact an organ procurement organization, and request an autopsy. Providing words of sympathy and reassurance, time for questions and initial grief, and a quiet private room for the family is appropriate and much appreciated.

 The Pronouncement & Death Certificate

In the United States, state policies direct clinicians to confirm the death of a patient in a formal process called “pronouncement.” The diagnosis of death is typically easy to make and the clinician need only verify the absence of spontaneous respirations and cardiac activity. Attempting to elicit pain in a patient who has died is unnecessary and disrespectful and should be avoided. A note describing these findings and the time of death is entered in the patient’s medical record. In many states, when a patient whose death is expected dies outside of the hospital (at home or in prisons, for example) nurses may be authorized to report the death over the telephone to a physician who assumes responsibility for signing the death certificate within 24 hours. For traumatic deaths, some states allow emergency medical technicians to pronounce a patient dead at the scene based on clearly defined criteria and with physician telephonic or radio supervision.

While the pronouncement may often seem like an awkward and unnecessary formality, clinicians may use this time to reassure the patient’s loved ones at the bedside that the patient died peacefully and that all appropriate care had been given. Both clinicians and families may use the ritual of the pronouncement as an opportunity to begin to process emotionally the death of the patient.

Physicians are legally required to accurately report the underlying cause of death on the death certificate. This reporting is important both for patients’ families (for insurance purposes and the need for an accurate family medical history) and for the epidemiologic study of disease and public health. The physician should be specific about the major cause of death being the condition without which the patient would not have died (eg, “decompensated cirrhosis”) and its contributory cause (eg, “hepatitis B and hepatitis C infections and chronic alcoholic hepatitis”) as well as any associated conditions (eg, “acute kidney injury”)—and not simply put down “cardiac arrest” as the cause of death.

 Autopsy & Organ Donation

Discussing the options and obtaining consent for autopsy and organ donation with patients prior to death is usually the best practice. This approach advances the principle of patient autonomy and lessens the responsibilities of distressed family members during the period immediately following the death. However, after a patient dies, or in the case of brain death, designated organ transplant personnel are more successful than the treating clinicians at obtaining consent for organ donation from surviving family members. Federal regulations require that a designated representative of an organ procurement organization approach the family about organ donation if the organs are appropriate for transplantation. Most people in the United States support the donation of organs for transplants. Currently, however, organ transplantation is severely limited by the availability of donor organs. Many potential donors and the families of actual donors experience a sense of reward in contributing, even through death, to the lives of others.

Clinicians must be sensitive to ethnic and cultural differences in attitudes about autopsy and organ donation. Patients or their families should be reminded of their right to limit autopsy or organ donation in any way they choose, although such restriction may limit the utility of autopsy. Pathologists can perform autopsies without interfering with funeral plans or the appearance of the deceased.

The results of an autopsy may help surviving family members and clinicians understand the exact cause of a patient’s death and foster a sense of closure. A clinician-family conference to review the results of the autopsy provides a good opportunity for clinicians to assess how well families are grieving and to answer questions. Despite the advantages of conducting postmortem examinations, autopsy rates are approximately 5%. Families report refusing autopsies out of fear of disfigurement of the body or delay of the funeral or say they were simply not asked. They allow autopsies in order to advance medical knowledge, to identify the exact cause of their loved one’s death, and to be reassured that appropriate care was given. Routinely addressing these issues when discussing autopsy may help increase the autopsy rate; the most important mistake is the failure to ask for permission to perform it.

Follow-up & Grieving

Proper care of patients at the end of life includes following up with surviving family members after the patient has died. Following up by telephone enables the clinician to assuage any guilt about decisions the family may have made, assess how families are grieving, reassure them about the nature of normal grieving, and identify complicated grief or depression. Clinicians can recommend support groups and counseling as needed. A card or telephone call from the clinician to the family days to weeks after the patient’s death (and perhaps on the anniversary of the death) allows the clinician to express concern for the family and the deceased.

After a patient dies, the clinician too may need to grieve. Although clinicians may be relatively unaffected by the deaths of some patients, other deaths may cause feelings of sadness, loss, and guilt. These emotions should be recognized as the first step toward processing and healing them. Each clinician may find personal or communal resources that help with the process of grieving. Shedding tears, the support of colleagues, time for reflection, and traditional or personal mourning rituals all may be effective. Attending the funeral of a patient who has died can be a satisfying personal experience that is almost universally appreciated by families and that may be the final element in caring well for people at the end of life.

Chau NG et al. Bereavement practices of physicians in oncology and palliative care. Arch Intern Med. 2009 May 25;169(10):963–71. [PMID: 19468090]

Simon NM. Treating complicated grief. JAMA. 2013 Jul 24;310(4):416–23. [PMID: 23917292]

Thornton JD et al. Effect of an iPod video intervention on consent to donate organs: a randomized trial. Ann Intern Med. 2012 Apr 3;156(7):483–90. [PMID: 22473435]