17.5 Delirium

Delirium is the most common neuropsychiatric disorder experienced by patients with advanced illness. It is characterized by disturbances in level of alertness, attention, thinking, perception, cognition, psychomotor behaviour, mood, and sleep–wake cycle. Acute onset and fluctuation of symptoms are critical features. Although conceptualized as a reversible process, it may not be reversible in the context of advanced illness, and indeed, may be a harbinger of imminent death (Breitbart and Alici, 2008).

The occurrence of delirium indicates significant physiological disturbance, usually involving multiple aetiologies. It may be associated with significant immediate morbidity, increased costs due to prolonged hospitalization, and long-term cognitive decline (Breitbart and Alici, 2008). It often distresses both families and staff (Breitbart et al., 2002) and can interfere with the recognition and control of other physical symptoms, such as pain.

Unfortunately, delirium often is under-recognized and undertreated. The heterogeneity of neuropsychiatric signs and symptoms, and the fluctuating course, may delay diagnosis and treatment. Specialists in palliative care must be able to screen for delirium, diagnose it accurately, undertake appropriate assessment of aetiologies, and be familiar with risks and benefits of pharmacological and non-pharmacological treatment options.

Delirium is a complex syndrome with a multifactorial aetiology, characterized by disturbance of cognition, arousal, and attention (Lipowski, 1990a; Caraceni and Grassi, 2011). The term ‘delirium’ is derived from the Latin word delirare, which means to ‘go out of the furrow’ (Lipowski, 1990b).

Based on data from five studies, delirium prevalence on admission to palliative care units or hospices ranges from 13.3% to 42.3% and the incidence after admission ranges between 3% and 45% (Hosie et al., 2013). Two studies that explored delirium rates during the period prior to death observed a prevalence of 58.8–88% (Lawlor et al., 2000a; Hosie et al., 2013).

Recent studies indicate that delirium has a multifaceted pathophysiology. In addition to perturbations in dopamine-containing and cholinergic pathways, other neurotransmitters and neurobiological pathways have been implicated (Flacker and Lipsitz, 1999; Trzepacz, 2000; van der Mast and Fekkes, 2000; Wilson et al., 2005; MacLullich et al., 2008), Disturbances in serotonin, gamma-aminobutyric acid (GABA), cortisol, cytokines, and oxygen free radicals may be involved (Flacker and Lipsitz, 1999; Trzepacz, 2000; van der Mast and Fekkes, 2000; Wilson et al., 2005; MacLullich et al., 2008). Drug-induced delirium may be related to transient thalamic dysfunction caused by medication-induced interference with central glutamatergic, GABAergic, dopaminergic, and cholinergic pathways at critical sites of action (Gaudreau and Gagnon, 2005).

A recent model classifies the proposed aetiologies of delirium into (a) direct brain insults and (b) aberrant stress responses (MacLullich et al., 2008). Direct brain insults include general and regional energy deprivation (e.g. hypoxia, hypoglycaemia, and stroke), medication effects, and metabolic abnormalities (e.g. hyponatraemia, hypercalcaemia) (MacLullich et al., 2008). Aberrant stress responses may involve systemic inflammation, the sickness behaviour response, and activity of the limbic–hypothalamic–pituitary–adrenal axis (MacLullich et al., 2008).

The gold standard for clinical diagnosis is a clinician assessment utilizing standard diagnostic criteria, most commonly the Diagnostic and Statistical Manual of Mental Disorders (DSM), fourth edition—text revised (DSM IV-TR), or alternatively, the International Classification of Disease, version 10 (ICD-10) of the World Health Organization (WHO) (American Psychiatric Association, 2000; WHO, 2010). The major components of the DSM IV-TR classification are disturbance of consciousness, a change in cognition, short and fluctuating chronology, and presence of an underlying medical condition. The fifth edition of the DSM (DSM-5) was released in May 2013 (American Psychiatric Association, 2013).The term consciousness has been removed from the delirium diagnostic criteria in DSM-5 to avoid diagnostic confusion between delirium and coma states. There has been much controversy and currently no data to inform implications of the DSM-5 for clinical care and research. The ICD-10 classification has the main components of impairment of consciousness and attention, global disturbance of cognition, psychomotor disturbance, disturbance of sleep–wake cycle, and emotional disturbance.

Delirium includes both cognitive and non-cognitive neuropsychiatric symptoms, including changes in motor behaviour, sleep–wake cycle, affective expression, perception, and thinking (Box 17.5.1) (Gupta et al., 2008). Both the frequency and specificity of these symptoms have been part of the debate in developing a definition of delirium, and approaches to classify and measure it. The frequency of core symptoms varies: attentional deficits 97–100%, thought process abnormalities 54–79%, disorientation 76–96%, memory deficits 88–96%, sleep–wake disturbance 92–97%, motoric alterations 24–94%, language disturbance 57–67%; and the non-core symptoms such as perceptual disturbance 50–63%, delusions 21–31%, and affective changes 43–86% (Webster and Holroyd, 2000; Meagher et al., 2007; Gupta et al., 2008).

Delirium screening instruments can be helpful to characterize clinical symptoms and potentially improve delirium detection rates. Commonly used screening instruments include the Confusion Assessment Method (Ryan et al., 2009) and the Nursing Delirium Assessment Scale (Gaudreau et al., 2005a).

The most studied delirium assessment tools for populations with cancer or advanced illness are the Memorial Delirium Assessment Scale (Breitbart, 1997) and the Delirium Rating Scale—Revised 98 (Trzepacz et al., 2001; Leonard et al., 2008b). These instruments assist clinicians in operationalizing DSM-IV criteria and monitoring severity.

In particular, patients who have predisposing risk factors, such as visual and hearing impairment, age greater than 65, pre-existing cognitive impairment, dehydration, immobility, multiple medications, and comorbid illness, should be monitored closely for delirium symptoms (National Institute for Health and Clinical Excellence (NICE), 2010). The studies exploring risk factors in cancer patients have been small, but implicate psychoactive medication, low albumin, and bone, liver, and intracranial metastases as additional risk factors (Fann et al., 2002; Ljubisavljevic and Kelly, 2003; Gaudreau et al., 2005b).

Three clinical subtypes of delirium have been identified based on psychomotor features and arousal levels (Meagher, 2009). Patients with the hypoactive subtype appear lethargic and drowsy, respond slowly to questions, do not initiate movement, and have reduced awareness of surroundings (Breitbart and Alici, 2012). The hypoactive presentation often leads to misdiagnosis (fatigue or depression) (Meagher and Trzepacz, 2000.). Patients with the hyperactive subtype have restlessness, agitation, and psychomotor overactivity (Camus et al., 2000.). Perceptual disturbance and delusions are more common in this subtype than in hypoactive delirium (Boettger and Breitbart, 2011). More recently, longitudinal studies have demonstrated a group of delirious patients—the so-called no subtype—who have minimal motor activity change (Meagher et al., 2012b). Once established, the type of delirium seems to remain stable over time (Meagher et al., 2012b).

The DSM IV-TR criteria for delirium do not include categories to define psychomotor subtype Meagher and Trzepacz, (1998, 2000). Methods that have been used include the presence or absence of particular psychomotor behaviours, quantitative measurement of psychomotor activity, validated scales to rate agitated behaviours (not specific for delirium), and most recently, the development of the delirium motor subtype scale (O’Keeffe, 1999; Meagher et al., 2008a, 2008b). There is poor concordance between these different methods of subtyping (Meagher et al., 2008b). It has been suggested that focusing on purely motoric features (Meagher et al., 2008a), and using independent quantitative methods such as electronic motion analysis (accelerometry), may assist in determining the true relationships between clinical subtypes, aetiologies, and outcomes (Leonard et al., 2007; Godfrey et al., 2010).

Increasing subsyndromal presentations also are recognized. For example, some patients with delirium present with symptoms in all core domains of sleep–wake cycle, thought process, language, attention, orientation, and visuospatial ability, but less severe symptoms overall (Meagher et al., 2012a; Trzepacz et al., 2012). This presentation can be particularly difficult to diagnose.

Several studies have explored delirium subtypes in varied populations, including those receiving specialist palliative care, those with advanced cancer, and those with haematological malignancies. These studies suggest that 40–78% of patients with delirium have the hypoactive subtype (Lawlor et al., 2000b; Morita et al., 2001; Lam, 2003; Fann et al., 2005; Gaudreau et al., 2005c; Spiller and Keen, 2006).

There has been much effort to determine if the delirium subtype has implications for differential diagnosis, aetiology, treatment, or prognosis (Meagher and Trzepacz, 1998; O’Keeffe, 1999; Gupta et al., 2008). This work has led to some tentative conclusions.

Differential diagnosis should consider that hyperactive delirium may mimic psychosis or anxiety disorders, whereas hypoactive delirium may be confused with depression or uncooperative behaviour (O’Keeffe, 1999).

It has been postulated that different neurotransmitter abnormalities may associate with delirium subtype (O’Keeffe, 1999.). For example increased GABA and reduced glutamate activity have been demonstrated in hepatic encephalopathy, and this may relate to the high prevalence of hyperactive delirium with this condition (O’Keeffe, 1999.). Some neurotransmitter abnormalities can cause varied phenomenology, for example, anticholinergic medication is typically associated with the hyperactive subtype but also can cause hypoactive and mixed presentations (Itil and Fink, 1966). The concentration of the melatonin metabolite, urinary 6-sulphatoxymelatonin, is highest in hypoactive delirium, lower in delirium with mixed presentation, and lowest in hyperactive delirium (Balan et al., 2003). Interleukin 6 has been associated with the hyperactive and mixed subtypes (van Munster et al., 2008).

Localized neuroanatomical lesions are also associated with particular presentations. For example, hyperactivity has been linked with middle temporal gyrus damage and frontostriatal injury has been associated with hypoactive presentations (Gupta et al., 2008).

In the main, the evidence seems to point to the hypoactive subtype having poorer outcomes (Meagher et al., 2011). Differences in morbidity include more pressure sores and hospital-acquired infection in hypoactive delirium and more falls in hyperactive presentations (O’Keeffe and Lavan).

The presence of delirium implies the existence of an underlying medical disorder and a detailed history and examination should be undertaken to identify the aetiologies and determine whether they are reversible. Most episodes of delirium have an average of three precipitants (Leonard et al., 2008a). Common aetiologies include infection, substance withdrawal (alcohol and nicotine), metabolic abnormalities (e.g. hypercalcaemia), hypoxia, and psychoactive medications (corticosteroids, opioids, anticholinergics, and benzodiazepines) (Lawlor et al., 2000a; Breitbart et al., 2008; Leonard et al., 2008b). Predisposing delirium risk factors should be reviewed including older age, physical frailty, multiple medical comorbidities, dementia, admission to the hospital with infection or dehydration, visual impairment, deafness, polypharmacy, renal impairment, and malnutrition (Inouye, 2006).

The evaluation of delirium may require laboratory test, imaging, or an electroencephalogram. Laboratory tests can identify metabolic abnormalities (e.g. hypercalcaemia, hyponatraemia, and hypoglycaemia), hypoxia, or disseminated intravascular coagulation. Imaging can detect brain metastases or leptomeningeal disease, intracranial bleeding, or ischaemia. An electroencephalogram may rule out seizures. Occasionally, a lumbar puncture is needed to assess the possibility of leptomeningeal carcinomatosis or infectious meningitis (Breitbart and Alici, 2008).

When confronted with delirium in the terminally ill or dying patient, a differential diagnosis should always be formulated as to the likely aetiology or aetiologies. However, consideration of individual goals of care, and prior function and disease trajectory, is needed to determine the extent to which these aetiologies are investigated (Leonard et al., 2008a).

Many of the clinical features of delirium can be associated with other psychiatric disorders such as depression, mania, psychosis, and dementia (Breitbart and Alici, 2008). Delirium, particularly the hypoactive subtype, is often initially misdiagnosed as depression. In distinguishing delirium from depression, particularly in the context of advanced disease, an evaluation of the onset and temporal sequencing of depressive and cognitive symptoms is particularly helpful. Importantly, the degree of cognitive impairment in delirium is much more severe than in depression, and usually has a more abrupt onset. Additionally, disturbance in level of alertness, which is characteristic of delirium, is not a feature of depression (Breitbart and Alici, 2008).

A manic episode may share some features of a hyperactive or mixed subtype of delirium. The temporal onset and course of symptoms, the presence of a disturbance in level of alertness, as well as of cognition, and the identification of a presumed medical aetiology for delirium are helpful in diagnosis. A past psychiatric history or family history of mood disorders is usually evident in patients with depression or a manic episode (Breitbart and Alici, 2008).

Delirium that is characterized by vivid hallucinations and delusions must be distinguished from a variety of psychotic disorders. In delirium, such psychotic symptoms occur in the context of advanced medical illness with features of disturbance in level of alertness and impaired attention span, as well as memory impairment and disorientation, which is not the case in other psychotic disorders (Breitbart and Alici, 2008).

The most challenging differential diagnostic issue is whether the patient has delirium, or dementia, or a delirium superimposed upon a pre-existing dementia. Both delirium and dementia are cognitive impairment disorders and share such common clinical features as impaired memory, thinking, judgement, aphasia, apraxia, agnosia, executive dysfunction, and disorientation. Delusions and hallucinations can occur in both disorders. To distinguish delirium from dementia, specific differences should be sought. In a study of 100 cancer patients that compared patients with delirium superimposed on dementia to those with delirium in the absence of dementia, the former presentation was characterized by more severe cognitive symptoms, poorer response to treatment, and a lower rate of resolution (Boettger et al., 2011). Clinically, the patient with dementia is alert and does not have the disturbance of level of alertness that is characteristic of delirium. The temporal onset of symptoms in dementia is more subacute and chronically progressive. Delirium is an acute change from the patient’s baseline cognitive functioning, even if the patient has dementia or other cognitive disturbances at baseline (Breitbart and Alici, 2008). Treatment may lead to reversal of the symptoms and signs of delirium, even in the patient with advanced illness; this also distinguishes the syndrome from dementia. However, as noted previously, delirium may not be reversible in the last 24–48 hours of life. A number of instruments can aid clinicians in the diagnosis of delirium, dementia, or delirium superimposed on dementia (Wong et al., 2010).

The management of delirium should include both correction of underlying causes, if possible, and treatment of the symptoms and signs of the disorder (Breitbart and Alici, 2012). To minimize distress to patients, staff, and family members, treatment of the symptoms and signs should be initiated before, or in concert with, a diagnostic assessment of the aetiologies. In the terminally ill patient who develops delirium in the last days of life (‘terminal’ delirium), management may present a number of dilemmas, and the desired clinical outcome may be significantly altered by the dying process. The desired and often achievable outcome is a patient who is awake, alert, calm, comfortable, cognitively intact, not psychotic, not in pain, and communicating coherently with family and staff (Breitbart and Alici, 2008).

In addition to seeking out and potentially correcting underlying causes of delirium, non-pharmacological and supportive therapies are important (Box 17.5.2). Fluid and electrolyte balance, nutrition, measures to help reduce anxiety and disorientation; interactions with and education of family members may be useful. A quiet, well-lit room with familiar objects, a visible clock or calendar, and the presence of family may help reduce anxiety and disorientation. In non-palliative care settings, there is evidence that these non-pharmacological interventions result in faster improvement in delirium and slower deterioration in cognition. However, they were not found to have any beneficial effects on mortality or health-related quality of life when compared with usual care (Milisen et al., 2005; Pitkälä et al., 2006, 2008; Flaherty et al., 2010).

Physical restraints should be avoided in patients who are at risk for developing delirium and for those with delirium (Breitbart and Alici, 2008). Recent evidence suggests that restraint-free management of patients should be the standard of care for prevention and treatment of delirium (Flaherty and Little, 2011). One-to-one observation may be necessary while maintaining safety of the patient without use of any restraints.

There have been an increasing number of delirium prevention and treatment studies published within the last decade. Although there are no drugs specifically approved by the US Food and Drug Administration (FDA) for the treatment of delirium, antipsychotic or sedative medications often are required to control the symptoms or signs (Breitbart and Alici, 2012). Antipsychotics, cholinesterase inhibitors, and alpha-2 agonists are the three groups of medications studied in randomized controlled trials in different patient populations. In palliative care settings, the evidence is supportive of short-term, low-dose use of antipsychotics in the control of symptoms of delirium with close monitoring for possible side effects especially in older patients with multiple medical comorbidities (Breitbart and Alici, 2012).

There have been a number of case reports, case series, retrospective chart reviews, open-label trials, randomized controlled comparison trials, and most recently placebo-controlled trials with antipsychotics in the treatment of delirium (Breitbart et al., 1996; Han and Kim, 2004; Hu et al., 2004; Jackson and Lipman, 2004; Lonergan et al., 2007; Devlin et al., 2010; Girard et al., 2010; Tahir et al., 2010; Grover et al., 2011; Breitbart and Alici, 2012). Study populations mostly include medically ill patients, postoperative patients, and patients in intensive care unit settings. Only a few studies focus specifically on patients with delirium in palliative care settings.

The American Psychiatric Association practice guidelines published in 1999 recommended the use of antipsychotics as the first-line pharmacological option in the treatment of symptoms of delirium (American Psychiatric Association, 1999).The guidelines recommend low-dose haloperidol (i.e. 1–2 mg orally (PO) every 4 hours as required or 0.25–0.5 mg PO every 4 hours for the elderly) as the treatment of choice when medication is necessary (American Psychiatric Association, 1999).

A 2004 Cochrane review concluded that haloperidol was the most suitable medication for the treatment of patients with delirium near the end of life, with chlorpromazine being an acceptable alternative based on one randomized controlled trial with haloperidol, chlorpromazine, and lorazepam (Jackson and Lipman, 2004).

A 2007 Cochrane review that compared the efficacy and adverse effects of haloperidol and atypical antipsychotics concluded that, like haloperidol, selected atypical antipsychotics (risperidone, olanzapine) were effective in managing delirium (Lonergan et al., 2007). Haloperidol doses greater than 4.5 mg/day resulted in increased rates of extrapyramidal symptoms compared with the atypical antipsychotics, but low-dose haloperidol (i.e. < 3.5 mg/day) was not shown to result in a greater frequency of extrapyramidal adverse effects (Lonergan et al., 2007).

Based on these published data, low-dose haloperidol continues to be the first-line agent for treatment of symptoms of delirium (Breitbart and Alici, 2012). In those with advanced illness, haloperidol is preferred due to its efficacy, tolerability (due in part to few anticholinergic effects), and its lack of active metabolites (Breitbart and Alici, 2008). It also may be given by various routes of administration. When used intravenously, however, the US FDA has warned about the risk of QTc prolongation and torsades de pointes; therefore, monitoring QTc intervals closely during intravenous haloperidol has become standard clinical practice. If haloperidol is not tolerated, the atypical antipsychotics are effective alternatives (Lonergan et al., 2007; Breitbart and Alici, 2012).

Based on clinical experience, lorazepam often is added to haloperidol to manage an acute delirium, particularly when hyperactive. Lorazepam (0.5–1.0 mg every 1–2 hours PO or intravenously (IV)) may be more effective in rapidly sedating the agitated delirious patient and may minimize extrapyramidal side effects associated with haloperidol. An alternative strategy is to switch from haloperidol to a more sedating antipsychotic, such as chlorpromazine. The latter approach is most acceptable when the patient is in a monitored setting and blood pressure can be checked frequently. It is important to monitor for anticholinergic and hypotensive adverse effects of chlorpromazine, particularly in elderly patients (Breitbart and Alici, 2012).

The potential risks associated with antipsychotic drug treatment should be considered before the first dose is given. These risks include the possibility of extrapyramidal effects, sedation, anticholinergic side effects, cardiac arrhythmias, and possible drug–drug interactions. These concerns are particularly salient in older patients with dementia. The US FDA has issued a warning about the increased risk of mortality associated with the use of antipsychotics in elderly patients with dementia-related behavioural disturbances. This warning was based on a meta-analysis by Schneider et al of 17 placebo-controlled trials involving patients with dementia (Schneider et al., 2005). The risk of death in patients treated with atypical antipsychotic agents was 1.6–1.7 times greater than in those who received placebo. Most deaths were associated with cardiovascular disease or infection. A second, retrospective study of nearly 23 000 older patients found that mortality rates were higher with typical than with atypical antipsychotics—whether or not they had dementia (Wang et al., 2005). This finding led to an extension of the FDA warning to typical antipsychotics, including haloperidol.

Other studies that have evaluated the risk of excess mortality during antipsychotic drug treatment have yielded mixed results. A retrospective cohort study of Medicaid enrolees in Tennessee demonstrated an increased risk of serious ventricular arrhythmias and sudden cardiac death among users of both typical and atypical antipsychotics (Ray et al., 2009). In contrast, a retrospective, case–control analysis of 326 elderly hospitalized patients with delirium at an acute care community hospital included 111 patients who received an antipsychotic and found non-significant odds ratios of dying during treatment of 1.53 (95% confidence interval (CI) 0.83–2.80) in univariate and 1.61 (95% CI 0.88–2.96) in multivariate analyses (Elie et al., 2009). Additional prospective studies with larger sample sizes will be needed to clarify this association further.

In palliative care settings, the evidence is most clearly supportive of short-term, low-dose use of antipsychotics for the control of symptoms of delirium. Close monitoring for possible side effects is essential, especially in older patients with multiple medical comorbidities (Table 17.5.1) (Breitbart and Alici, 2012).

The use of psychostimulants in the treatment of hypoactive subtype of delirium has been suggested by case reports and one open-label study (Keen and Brown, 2004; Gagnon et al., 2005). These drugs can precipitate agitation and exacerbate psychotic symptoms, however, should be used very cautiously in palliative care settings.

Impaired cholinergic function has been implicated as one of the final common pathways in the pathogenesis of delirium (MacLullich et al., 2008). Despite case reports of beneficial effects of donepezil and rivastigmine, a 2008 Cochrane review concluded that there is currently no evidence from controlled trials supporting the use of cholinesterase inhibitors in the treatment of delirium (Overshott et al., 2008). On the basis of the existing evidence from general hospital and critical care settings, and the lack of studies in palliative care settings, cholinesterase inhibitors cannot be recommended in the treatment of delirium.

Dexmedetomidine is a selective alpha-2 adrenergic receptor agonist indicated in the United States for the sedation of mechanically ventilated adult patients in intensive care settings and in non-intubated adult patients prior to and/or during surgical and other procedures. It has analgesic effects and has been considered for the prevention and treatment of delirium (Riker et al., 2009; Pandharipande et al., 2010; Prommer, 2011). Clinical trials in the intensive care unit have shown mixed results for the prevention and treatment of delirium (Riker et al., 2009; Pandharipande et al., 2010) and there have been no studies in palliative care settings.

The development of effective strategies to prevent delirium should be a high priority in palliative care settings. Promising results in the elderly hospitalized population have been observed with various non-pharmacological interventions (Inouye et al., 1999; Siddiqi et al., 2007; NICE, 2011). In the latter population, the study effect sizes suggest that the incidence of delirium may be reduced by as much as one-third with multicomponent interventions (Siddiqi et al., 2007; NICE, 2012). In contrast, a simple multicomponent preventive intervention was found to be ineffective in reducing delirium incidence or severity among cancer patients (N = 1516) receiving end-of-life care; no difference was observed between the intervention and the usual-care groups in delirium incidence (odds ratio 0.94, p = 0.66), delirium severity (1.83 vs 1.92, p = 0.07), total days in delirium (4.57 vs 3.57 days, p = 0.63), or duration of first delirium episode (2.9 vs 2.1 days, p = 0.96) (Gagnon et al., 2010).

A variety of pharmacological interventions have also been considered for the prevention of delirium. Antipsychotics, cholinesterase inhibitors, melatonin, and dexmedetomidine have been evaluated in randomized controlled studies conducted in different settings (Kalisvaart et al., 2005; Liptzin et al., 2005; Sampson et al., 2007; Gamberini et al., 2009; Riker et al., 2009; Larsen et al., 2010; Pandharipande et al., 2010; Prommer 2011; NICE, 2012). A 2007 Cochrane review concluded that the evidence for the effectiveness of interventions to prevent delirium was sparse and no recommendations could be made regarding the use of drug therapy for delirium prevention (Siddiqi et al., 2007). Studies published since 2007 have shown mixed results with cholinesterase inhibitors, antipsychotics, melatonin, and dexmedetomidine (Breitbart and Alici, 2012), and there continues to be no established treatment to prevent delirium in palliative care settings.

Delirium that evolves to active dying is terminal delirium. The management of this syndrome is controversial. One study showed that physicians from different disciplines manage terminal delirium differently (Agar et al., 2012); medical oncologists were found to be more likely to manage terminal delirium with benzodiazepines or benzodiazepine and antipsychotic combinations, whereas palliative care physicians were more likely to use antipsychotics to manage delirium symptoms, including the hypoactive subtype of delirium.

Some have argued from a philosophical perspective that pharmacological interventions with antipsychotics or benzodiazepines are inappropriate in the dying patient because delirium is a natural part of the dying process and should not be altered. Clinical experience, however, indicates that the development of delirium can be highly distressing to the patient and family and that the use of antipsychotics to manage agitation, paranoia, hallucinations or altered sensorium is safe, effective and often quite appropriate (Breitbart and Alici, 2008). The use of drug therapy in those believed to be approaching death is best managed on a case-by-case basis, balancing a ‘wait and see’ approach without the use of pharmacological interventions against the level of overt distress and the knowledge that patients with hypoactive delirium—who may not appear distressed—may quickly and unexpectedly become agitated. It is important to remember that, by their nature, the symptoms of delirium are unstable and fluctuate over time.

Perhaps the most challenging clinical problem is the management of the dying patient with a ‘terminal’ delirium that is unresponsive to standard antipsychotic interventions, whose symptoms can only be controlled by sedation to the point of a significantly decreased level of consciousness. While antipsychotics are most effective in diminishing agitation, clearing the sensorium and improving cognition is not always possible in delirium which complicates the last days of life. Approximately 30% of dying patients with delirium do not have their symptoms adequately controlled with antipsychotic medications (Fainsinger et al., 2000; Rietjens et al., 2008). Processes causing delirium may be ongoing and irreversible during the active dying phase. In such cases, a reasonable choice is the use of sedative agents such as benzodiazepines (e.g. midazolam, lorazepam) or propofol (and sometimes an opioid, if the drug has proved to be demonstrably sedative) to achieve a state of quiet sedation (Fainsinger et al., 2000; Rietjens et al., 2008). Indeed, delirium has been identified as the main indication for the use of palliative sedation (Fainsinger et al., 2000; Sykes and Thorns, 2003; Lo and Rubenfeld, 2005; Rietjens et al., 2008). Clinicians who are concerned that the use of sedating medications can hasten death via respiratory depression, hypotension, or even starvation, should note that studies in hospice and palliative care settings have not confirmed this association (Bercovitch and Adunsky, 2004; Vitetta et al., 2005).

Before administering a sedative to achieve a calm and comfortable but sedated and unresponsive patient, the clinician must first take several steps. The clinician must have a discussion with the family (and the patient, if there are lucid moments when the patient appears to have capacity), eliciting their concerns and wishes for the type of care that can best honour their desire to provide comfort and symptom control during the dying process. The clinician should describe the optimal achievable goals of therapy as they currently exist. Family members should be informed that the goal of sedation is to provide comfort and symptom control, not to hasten death. Sedation in such patients is not always complete or irreversible; some patients have periods of wakefulness despite sedation, and many clinicians will periodically lighten sedation to reassess the patient’s condition. Ultimately, the clinician must always keep the goals of care in mind and communicate these goals to the staff, patient, and family members. The clinician must weigh each of the issues outlined above in making decisions on how to best manage the dying patient who presents with delirium that preserves and respects the dignity and values of that individual and family (Breitbart and Alici, 2008).

It is important to emphasize the prognostic value of delirium in terminally ill patients. Delirium is a relatively reliable predictor of approaching death in the coming days to weeks (Casarett and Inouye, 2001). In the palliative care setting, several studies provide support that delirium reliably predicts impending death in patients with advanced cancer (Morita et al., 2011). Given the prognostic significance of delirium, recognizing an episode of delirium in the late phases of an illness is critically important in treatment planning and in advising family members.

Palliative care clinicians commonly encounter delirium as a major neuropsychiatric complication of terminal illness. Screening, assessment, diagnosis, and management of delirium are essential in improving quality of life and minimizing morbidity in palliative care settings for patients, families, and health-care professionals.