Care of Patients with Delirium at the End of Life

One in four Americans dies in a nursing home¹; 30% of all nursing home residents die within 1 year of admission.² In the last 30 years, hospice and palliative care practitioners have advanced a model of comprehensive, interdisciplinary care that promotes comfort, symptom management, emotional and spiritual support, and advance care planning for patients.³ Long-term care facilities increasingly apply palliative care principles, and as they become more sophisticated at recognizing patients with shortened life expectancy, refer them for concurrent hospice enrollment.

Screening, assessment, and treatment of mental disorders are central to good terminal care in nursing home residents because psychological distress is the source of substantial suffering at the end of life. When faced with the knowledge of foreshortened life, most patients experience anxiety, fear, sadness, or anger. These symptoms meet criteria for a mental disorder at the point they become pervasive or impair an individual’s ability to function or make decisions.^4,5 The most common mental disorders in terminal care are delirium, depression, and anxiety. Among nursing home residents, these disorders often develop in those who already have dementia. Delirium is arguably the most important mental disorder at the end of life because of its high prevalence and deleterious impact on quality of life, behavior, and communication. Delirium undermines several important goals of care at the end of life, including comfort and meaningful interaction with family.⁶
__________________________________________________________________________________________________________________________
RELATED CONTENT
Caring for the Patient With End-Stage Dementia
Hip Fracture Rehabilitation in Persons with Dementia: How Much Should We Invest?
__________________________________________________________________________________________________________________________

There are very few studies of delirium in residents receiving palliative care in nursing homes. Most studies of delirium in persons at the end of life recruit from inpatient palliative care settings, where the most common terminal diagnosis is cancer. While it seems reasonable to extrapolate to some degree from patients dying of cancer in palliative care units to patients dying in nursing facilities, differences must be acknowledged. Cancer is the second most common cause of death in nursing home residents, but only 17% of all cancer deaths occur in nursing facilities.⁷ Cancer patients’ lengths of stay in nursing homes are likely brief, therefore they rarely make up a large proportion of the facilities’ population. For example, in a randomized controlled trial to increase hospice enrollment in the nursing home, 62% of study participants were diagnosed with dementia, 39% had cardiopulmonary disease, and only 4% had cancer.¹ Delirium in the nursing home often occurs in the context of dementia, a diagnosis that is underrepresented in inpatient palliative care and hospice patients. Because of the paucity of studies of delirium in patients dying in long-term care, I will draw on studies of delirium in other ill populations including patients in medical and surgical inpatient settings, intensive care units, and inpatient palliative care settings, acknowledging the limitations in generalizing to the nursing home setting.

CLINICAL FEATURES OF DELIRIUM
Delirium is a mental disorder that comes on over hours to weeks, and results most often from medications or physical illnesses, usually outside of the brain. Symptoms of delirium include confusion, inattention, diminished awareness, impaired memory, perceptual disturbances, and sleep disruption.⁸ Delirium is the most common mental disorder among dying patients, occurring in up to 90% of cancer patients in the final weeks of life.⁹ The prevalence of delirium in the final weeks of life in nursing home residents is unknown and complicated by the high prevalence of dementia. Among dying patients, delirium progresses to coma preceding death.¹⁰

Delirium has been subtyped based on whether the patient is restless (agitated or hypermotoric delirium) versus lethargic (hypomotoric or quiet delirium). Within hospice and terminal care, the major concern has been with agitated delirium, which is referred to as terminal restless or terminal agitation. Hypermotoric patients appear alert; are agitated, irritable and restless; may develop paranoia, delusions and hallucinations; and manifest verbal and physical aggressiveness. These patients are often perceived as having substantial suffering, especially when delirium occurs with pain. Agitated delirium may precipitate a crisis resulting in hospital or inpatient palliative care admission. When agitated delirium occurs in the final days and hours of life, it is considered a hallmark of a “bad death.” Controversy exists regarding the causes of agitated delirium in terminal care; some clinicians believe that patients may develop this syndrome purely from psychosocial and spiritual distress, whereas others assert that medical/physical causes are primary, as in other deliria.¹¹

In contrast, hypomotoric delirium, in which patients characteristically have apathy and lethargy, is so common among dying hospice patients that it is often not considered clinically important. Several research findings counter this laissez faire attitude. Among elderly patients, hypoactive delirium is the predominant form of delirium in all settings.¹² Spiller and Keen¹³ reported that among admissions to an inpatient palliative care unit, 29% of patients had delirium; in 86% it was the hypoactive subtype. Surprisingly, levels of distress are similar between the two types. Breitbart et al,¹⁴ who surveyed cancer patients after resolution of an episode of delirium, reported that most patients recalled and were upset when remembering their delirium experience. Although delusions were a significant predictor of patients’ distress, those with hypoactive delirium were as distressed as those with hyperactive delirium. Physical disability predicted spouse distress, and perceptual disturbances and delirium severity predicted nurse distress. Morita and coauthors¹⁵ surveyed family members of cancer decedents to determine the level of distress about delirium-related symptoms. Two-thirds of family members rated all delirium-related symptoms other than somnolence as distressing or very distressing when they occurred often, with restlessness and mood lability as the most upsetting for family members. Although delirium may be an inevitable part of dying, efforts to minimize its impact are important, and are as important for hypoactive as agitated deliria. Educating family and care providers about delirium is key to reducing distress.

Clinicians frequently fail to recognize deliria, overlooking about half of cases,^16,19 even in settings where it is common, such as inpatient palliative care units. The symptoms may be attributed to other psychiatric disorders such as depression or anxiety. Farrell and Ganzini¹⁷ reported that delirium was diagnosed in 42% of elderly medical inpatients referred to mental health for evaluation of depression. Similarly, Spiller and Keen¹³ noted a correlation between high ratings on depression screening tools and delirium severity among inpatient palliative care admissions—these authors voiced concerns that without systematic attempts to diagnose delirium, it might be mischaracterized. This point is very important—other psychiatric diagnoses cannot be validly made in the context of delirium, and most treatments for depression and anxiety will be ineffective or even harmful if, in fact, the patient has delirium.

In the nursing home, delirium often occurs in the context of dementia, and no studies describe the clinical manifestations of patients dying with both disorders. Dementia is a risk factor for the development of delirium. Studies support that medical and nursing personnel are less likely to recognize delirium in patients with dementia.¹⁸

Delirium has a complex and poorly understood association with pain. Pain may worsen delirium. Patients who use mental effort to cope stoically with pain may lose this mental ability if they have delirium. Delirium-associated agitation and restlessness are often misinterpreted as worsening pain. For example, Bruera and coauthors¹⁹ documented that both physicians and nurses misinterpret agitation as an expression of pain in patients with hyperactive delirium, even though before delirium development and after delirium resolution, pain was well controlled. This interpretation may lead to excessive opioid use, which may worsen delirium unnecessarily. Furthermore, many cognitively-impaired patients cannot use pain-rating instruments reliably, confounding investigation of the relationship between these two syndromes.²⁰ This area clearly requires further investigation.

DIAGNOSIS OF DELIRIUM
Experienced clinicians diagnose delirium based on history and mental status examination, knowing that the setting and “pretest probability” of the disorder add to diagnostic acumen. When evaluating any mood or behavioral change among very ill elderly patients in terminal care settings, intensive care units, and medical and surgical inpatient services, delirium is the “horse” and all other psychiatric disorders, even depression, are relative “zebras.” The Confusion Assessment Method (CAM) has been extensively researched, is brief, and demonstrates high sensitivity and specificity for delirium diagnosis. The criteria used in the CAM are: (1) acute onset and fluctuation, (2) inattention, (3) disorganized thinking, and (4) altered level of consciousness, with a delirium diagnosis resulting when patients have both criteria 1 and 2, and either criteria 3 or 4.²¹ Studies of bedside nurses suggest difficulty in identifying criteria 1 and 4, although they still reached 90% diagnostic accuracy.^22,23

One of the most commonly used measures of cognition in the elderly is the Mini-Mental State Examination (MMSE). The MMSE cannot reliably distinguish delirium and dementia at a single point in time. O’Keeffe and coauthors²⁴ in 2005, however, reported that a decline of two or more points was 93% sensitive and 90% specific for delirium diagnosis in patients admitted to an acute geriatric service. Clinical use of the MMSE has recently become more expensive since the copyright was purchased by Psychological Assessment Resources, and photocopying and Internet download of the form is no longer allowed.

PREDISPOSING FACTORS, PRECIPITANTS,AND CAUSES OF DELIRIUM
Most deliria in terminal care are multifactorial. For example, in a study of patients with advanced cancer, Lawlor and coauthors⁹ found a median of three precipitating factors per episode of delirium. Among patients with cancer in terminal care, delirium often starts weeks before death, with spontaneous remissions, but is a recognized harbinger of death. In half of deliria in this setting, evaluation will not reveal any other cause other than the patient’s primary terminal illness.¹⁶

Within long-term care, terminally-ill patients, their providers, and families must consider how aggressive and invasive they will be in evaluating the causes of delirium, as burdens will accrue. Often, patients will have lost decision-making capacity and will be unable to participate in balancing benefits and burdens of evaluation. Lawlor et al⁹ in a study of residents of an inpatient palliative care unit with advanced cancer, reported that delirium was more likely to be reversed if associated with dehydration and psychoactive medications, but less likely to be reversed if resulting from hypoxia and metabolic abnormalities. Gentle hydration is being reevaluated as palliative among patients in the final weeks of life. Bruera et al²⁵ randomized mildly dehydrated patients with cancer in palliative care to parenteral hydration with 1000 mL/day versus “placebo” hydration over 2 days. Myoclonus and sedation were improved in the hydration group, although there were no differences in hallucinations or fatigue. On the other hand, Morita et al¹⁵ compared patients with cancer who received fluids (1 liter or more per day) to those who did not. Both groups of patients had similar levels of agitation and hyperactive delirium, and hydration patients had more edema and ascites. Use of low doses of fluids remains controversial, and a standard of care among palliative care patients in last weeks of life has not emerged.

With terminally-ill patients, decreasing or stopping drugs that contribute to delirium may be easily facilitated. For opioid-induced deliria, experts recommend rotating to an equipotent or slightly less than equipotent dose of another opioid.¹⁰ Anticholinergic medications such as tricyclic antidepressants used in pain treatment; antisecretory agents such as scopolamine, or anti-nausea drugs; benzodiazepines (see below); and corticosteroids are among the most frequently implicated as deliriogenic in a variety of studies.^5,26,27

PHARMACOLOGIC MANAGEMENT
Hyperactive or agitated deliria require aggressive treatment because the experience is frightening to patients and families, with increased risk of wandering, falls, and self-harm. Neuroleptic medications (antipsychotics) are the first-line agents in management of patients with delirium with agitation, paranoia, delusions, or physical aggression, although none are U.S. Food and Drug Administration–approved for this purpose. Haloperidol is most often recommended by experts and consensus panels.^4,10 The advantages of haloperidol are its wide therapeutic margin of safety, ease of administration by a variety of routes, and relative lack of cardiopulmonary and anticholinergic effects at low doses. In a randomized, double-blind trial comparing haloperidol 1.5 mg per day versus placebo among elderly patients pretreated before hip surgery, there was no difference between groups in the incidence of delirium; however, delirium duration, severity, and duration of hospital stay were dramatically lower in the active treatment arm.²⁸ The major side effects of haloperidol that are relevant to elderly patients at the end of life are drug-induced parkinsonism (DIP) and akathisia. DIP is associated with discomfort, choking, falls in ambulatory patients, and decreased bed mobility in nonambulatory patients. It appears 3-5 days after neuroleptic drug treatment initiation, and, in elderly patients, can develop even at low doses. Akathisia is a syndrome of motor restlessness that can cause substantial discomfort and may be mistaken as worsened agitation. Akathisia can occur even with the first dose of antipsychotic medication. Acute dystonias are exceedingly rare in older, ill patients, and tardive dyskinesia, which develops after several months of chronic neuroleptic treatment, is not clinically relevant in patients with shortened life expectancy. Although some antipsychotics block alpha receptors, which can result in orthostatic hypotension, this is not common with haloperidol. More studies of the indications for and efficacy of neuroleptic use are needed in these populations.

Among patients with serious and persistent mental disorders such as schizophrenia, second-generation (atypical) antipsychotics, such as quetiapine, risperidone, and olanzapine, have generally supplanted older drugs. These newer neuroleptics are less likely to cause tardive dyskinesia (at about half the rate),²⁹ and may be associated with fewer “negative symptoms” of schizophrenia, such as apathy. These benefits are not relevant to end-of-life management of delirium. In general, olanzapine and quetiapine will cause less DIP, are much more expensive, and are more liable to cause orthostatic hypotension than haloperidol. In a small, blinded trial comparing risperidone with haloperidol in treatment of delirium in medically ill patients, the investigators found no differences in efficacy or adverse effects.³⁰ A comparison of olanzapine versus haloperidol for treating delirium in critical care showed no difference in efficacy, although the development of DIP may have been higher in the haloperidol group (this was not statistically analyzed).³¹ Second-generation antipsychotic drugs have been linked to hyperglycemia, hyperlipidemia, and increased risk of stroke. Among patients with dementia, antipsychotic use increases risk of death from 2% to 4% in the first 10 weeks of treatment.³² How relevant these risks are in treating patients with a shortened life expectancy is unknown.

Interesting case series suggested a possible role for cholinesterase inhibitors in delirium prevention.³³ In the only double-blind, placebo-controlled study, which was conducted in elderly patients undergoing elective joint replacement, investigators found no benefit of donepezil in reducing either incidence or duration of delirium.³⁴

Outside of terminal care, benzodiazepines universally worsen delirium, and should not be considered a treatment for this disorder unless the cause is alcohol and benzodiazepine withdrawal.^35,36 Breitbart et al,³⁷ in a randomized trial comparing neuroleptics to lorazepam for the treatment of delirium in patients with AIDS, stopped the lorazepam arm early when the patients’ cognition worsened and many developed adverse effects. Lorazepam was found to be an independent predictor for transition to delirium among patients in intensive care units.³⁸ Among hospitalized patients with cancer, those exposed to benzodiazepines (more than 2 mg of lorazepam daily or equivalent) had twice the risk of developing delirium.²⁶ Whether or not to use benzodiazepines in elderly, terminally-ill patients with delirium in the nursing home will depend on the goals of care. For bedbound, moribund patients with agitation, paranoia, dyspnea, or myoclonus, benzodiazepines can be used to facilitate calm, sedation, and anxiolysis, but at the expense of worsening confusion. In patients, however, who are not in the final days of life, who may have a reversible cause of delirium, and where the goals of care may still include maintaining function and communication with others, benzodiazepines should be avoided. Goals of care should be discussed with caregivers, since most delirious patients will lack decision-making capacity.

OTHER INTERVENTIONS
Multicomponent interventions among inpatient geriatric patients can result in close to 40% reduction in delirium, but they may be less easily applied to terminally-ill patients in nursing homes.³⁹ Assuring that eyeglasses and hearing aids are available may help patients correct misperceptions. Minimization of noxious stimuli and medical testing may reduce agitation. In a study of hospitalized patients randomized to home versus inpatient rehabilitation, home patients had lower rates of delirium, supporting that efforts to keep patients in familiar environments may minimize confusion.⁴⁰

SUMMARY
During the last decade, researchers in geriatrics and palliative care have published many studies about delirium, yet substantial new information is still lacking and unanswered questions remain. In terminal care, delirium is common, has substantial impact on quality of life, and is underrecognized. The approach to evaluating and treating delirium will depend on the goals of care. Delirium resulting from drugs or dehydration are among the most treatable. The most effective and least burdensome approaches include switching to a different opioid and administering antipsychotics such as haloperidol, though well-done, randomized, controlled trials have not been completed. Best practices in care of dying patients with delirium should be a focus of research in both palliative and long-term care.

The author reports no relevant financial relationships.