National Cancer Institute
Last Modified: July 25, 2011
Cognitive disorders and delirium are neuropsychiatric syndromes that occur frequently in patients with cancer, particularly in those with advanced disease. Occurrence rates range from 28% to 48% in patients with advanced cancer on admission to hospital or hospice, 1 2 3 and approximately 90% of these patients will experience delirium in the hours to days before death. 3 4 5 The term acute confusional state has also been used to describe this syndrome; in the last days of life, the condition sometimes referred to as terminal restlessness probably represents a terminal delirium. 6 Although delirium clearly has a recognized association with the dying phase, many episodes of delirium are reversible; therapeutic intervention can result in delirium reversal, or at least improvement, in 30% to 75% of episodes. 3 5 7 8 9 10 Variability in reported occurrence rates and clinical outcomes most likely reflects sampling from different clinical settings or different stages in the clinical trajectory of cancer, in addition to inconsistency in diagnostic terminology. 11 The prevention of delirium in the patient with cancer has not been systematically examined, but studies in elderly medical patients suggest that early identification of risk factors reduces the occurrence rate of delirium and the duration of episodes. 12
Delirium has been defined as a disorder of global cerebral dysfunction characterized by disordered awareness, attention, and cognition. 13 In addition, delirium is associated with behavioral manifestations. The fourth edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-IV) cites the core clinical criteria for diagnosis as follows: 14
Other associated noncore clinical criteria features include sleep-wake cycle disturbance, delusions, emotional lability, and disturbance of psychomotor activity. The latter forms the basis of classifying delirium into three different subtypes: 15 16
In this summary, unless otherwise stated, evidence and practice issues as they relate to adults are discussed. The evidence and application to practice related to children may differ significantly from information related to adults. When specific information about the care of children is available, it is summarized under its own heading.
Despite the very limited systematic study of risk factors for delirium in patients with cancer, risk factors have been identified in general medical patients (some of them with cancer) and include severe illness, level of comorbidity, advanced age, prior dementia, hypoalbuminemia, infection, azotemia, and psychoactive medications. 20 21 22 Studies in elderly medical patients suggest that the level of risk is proportionate to the number of risk factors present. 23 Cancer is particularly prevalent in the elderly population. Many patients with cancer, particularly those with advanced disease, are likely to have a high level of baseline vulnerability. Such vulnerability leaves them predisposed to precipitants such as psychoactive medications. 24 It is also likely that the predictors of poor pain control in cancer patients (neuropathic pain, incidental pain, opioid tolerance, somatization, and a history of drug or alcohol abuse) result in higher opioid doses and thereby increase the risk of delirium. 25 Distinct from delirium, older (65 years or older), long-term (>5 years) cancer survivors are also at increased risk of cognitive deficits and possibly dementia, as noted in a co-twin control design study of 702 Swedish cancer survivors. 10
The clinical presentation of cognitive disorders and delirium is associated with a high level of distress in patients, family members, and health care personnel. 1 2 3 Delirium is associated with a high burden of symptom distress, particularly in relation to delusions, perceptual disturbances, and psychomotor agitation. Incontinence, falls, failure to maintain adequate hydration, a prolonged hospital stay, and death are more likely to occur in the patient with delirium. 4 5 6 7 8 9 10
Dysfunctional cognition in the delirious patient hinders communication between patient and family and between patient and health care personnel. As a result, reliable symptom assessment, counseling, and active patient participation in the therapeutic decision-making process are all compromised. 11 Psychomotor agitation and emotional lability in the delirious patient may be misinterpreted as a presentation of increased pain expression. 12 Consequently, conflict can arise among the patient, family, and staff concerning the level of analgesia needed. A potentially destructive triangle can develop when the patient's family misinterprets agitation as increased pain and advocates for inappropriate escalation of opioid dosing. 13
A psychosocial intervention for family caregivers of patients with advanced cancer may be beneficial in providing knowledge of delirium and detection rates and in increasing family caregiver self-confidence in decision making. 14
The diagnosis of cognitive disorders and delirium should be considered in any patient with cancer demonstrating an acute onset of agitation or uncooperative behavior, personality change, impaired cognitive functioning, altered attention span, fluctuating level of consciousness, or uncharacteristic anxiety or depression. The diagnoses of delirium and cognitive impairment are frequently missed and poorly documented, however. 1 2 3 4 5 Medical and nursing staff, as well as family members, may attribute a functional cause to some of the early, prodromal, and more subtle signs of delirium such as increased anxiety, restlessness, and emotional lability. 6 Failure to recognize delirium is particularly likely if the patient is encountered in a transient lucid phase, which can commonly occur as part of the fluctuating nature of delirium. 7 Delirium is most frequently misdiagnosed as depression or dementia. 7 8 9 10 The hypoactive subtype is considered especially likely to be misdiagnosed as depression. 7
Differentiating delirium from dementia or recognizing delirium superimposed on dementia can be difficult because of some shared clinical features such as disorientation and impairment of memory, thinking, and judgment. 11 12 13 Dementia, however, typically appears in relatively alert individuals; disturbance of consciousness is not a common feature. The temporal onset of symptoms of delirium is acute (hours to days), not insidious (months to years) as in dementia. 14 In elderly patients with cancer, delirium is often superimposed on dementia, giving rise to a particularly difficult diagnostic challenge. In this situation, the diagnosis may become more apparent when delirium fails to reverse or when some features of delirium, especially cognitive impairment, persist. Dementia is often then the most likely explanation for a persistent or residual cognitive deficit. 14
Vigilance on the part of nursing staff and a systematic approach to recording serial observations assist in the detection of delirium. Regular cognitive screening facilitates the diagnosis of delirium in cancer patients. 15 Instruments such as the Minimental State Examination (MMSE), Blessed Orientation Memory and Concentration Test (BOMC), and Confusion Assessment Method (CAM) have favorable psychometric properties and are brief enough to allow repeated administration in cancer patients. 16 17 18 The BOMC and MMSE screen for cognitive impairment and require active patient participation in assessment. The Bedside Confusion Scale (BCS) also requires active patient participation; however, it is remarkably brief, and its psychometric potential as a screening instrument compares favorably with the CAM. 19 The CAM does not require formal patient participation. The Memorial Delirium Assessment Scale (MDAS) and Delirium Rating Scale-Revised-98 have been validated as having diagnostic and severity rating potential. 20 21 The MDAS allows prorating of scores when a patient cannot actively participate in testing for reasons such as dyspnea or fatigue.
An integrated approach to management involves educating family members about the nature of the delirium syndrome and its potential treatment. 1 2 3 Family concerns, particularly the misinterpretation of symptoms such as agitation, emotional lability, and disinhibition, must be addressed. Depending on the clinical circumstances, guarded optimism regarding reversibility can be expressed. On the basis of discussion with family members, a consensus is then reached on the goals of care; this, in turn, will determine the desired and appropriate level of assessment and therapeutic intervention, which could be directed at identifying and treating underlying precipitants to reverse or improve delirium. 2 The extent of assessment will likely be influenced by the clinical setting, disease variables, and level of distress. It may therefore be appropriate in some situations to forego further assessment and focus solely on symptomatic treatment. Regardless of the level of investigational or therapeutic aggression, symptomatic treatment is usually required in most patients. Ongoing monitoring and reassessment should also occur, particularly when pharmacological sedation is required to initially control symptoms. 4
Various environmental strategies have been proposed to reduce the symptomatic distress associated with delirium. These strategies include discrete efforts at reorientation such as a well-lit room with familiar objects, a visible clock or calendar, limited staff changes (and possibly one-on-one nursing care), reduced noise stimulation, and the presence of family. 5 6 7 Although some controversy surrounds the use of physical restraints, their judicious use may sometimes be necessary to prevent self-harm or physical aggression directed at caregivers.
Delirium reversal is consistent with the goals of care; therefore, the standard management approach in patients with cancer is to search for and treat the reversible precipitants of delirium. 1 2 Although patients with cancer generally have a high level of baseline vulnerability to the development of delirium (owing to factors such as cachexia, hypoalbuminemia, advanced age, and prior dementia), the greatest therapeutic benefit is more likely to be derived from identifying and treating superimposed precipitants with relatively low-burden interventions such as discontinuation or dose reduction of psychoactive medications, 8 subcutaneous fluid administration via hypodermoclysis to treat dehydration, intravenous or subcutaneous bisphosphonate treatment of hypercalcemia, and possibly oral or intravenous antibiotics to treat infection. 9 10 This process typically involves a careful history and physical examination in addition to basic laboratory tests and imaging. 7 If no obvious precipitant is identified on preliminary searching, the decision to proceed with more invasive or elaborate tests is mainly determined by the goals of care.
Opioid analgesics, required in most patients with advanced disease, are among the psychoactive agents that precipitate delirium most frequently. 9 11 A prospective cohort study in an oncology inpatient population (n = 114) demonstrated that patients exposed to daily oral opioid doses of more than 90 mg of morphine or morphine equivalents were at significantly higher risk of developing delirium. These results are still significant when controlling for the effects of corticosteroids, benzodiazepines, and other psychotropic medications. However, this study did not provide enough data (on comorbidities, laboratory values, etc.) to determine which patients specifically were at risk for delirium when exposed to more than 90 mg of opioids per day. 12 Patients with delirium should be assessed for other symptoms that suggest opioid neurotoxicity, for example, tactile hallucinations, agitation, myoclonus, allodynia, hyperalgesia, and possibly seizures. It is postulated that this toxic state relates to the accumulation of the parent opioid compound or its metabolites. 11 13 Intervention in the form of dose reduction or opioid switching in association with assisted hydration typically allows for clearing of the offending opioid or its metabolites. On identifying opioid toxicity, therefore, it is important to search for other precipitants such as dehydration or infection. 14 A common clinical scenario consists of a patient with infection who may become drowsy, take less fluid, become dehydrated, and then exhibit signs of opioid toxicity, including delirium. Therapeutic intervention should target the triad of precipitants in this scenario. (Refer to the Opioid switching section in the PDQ® summary on Pain for more information.)
The pathogenesis of delirium is complex and is poorly understood. Abnormalities of various neurotransmitter and other endogenous agents have been postulated, including reduced cerebral cholinergic transmission (or excess dopaminergic transmission relative to cholinergic transmission), altered gamma aminobutyric acid transmission, altered serotonin transmission, cytokine production, and altered cortisol levels. 15 The mainstay of pharmacological management has been the use of powerful antidopaminergic neuroleptics such as haloperidol 7 and the newer atypical antipsychotic agents such as olanzapine, risperidone, and quetiapine. 16 17; 18[Level of evidence: I]
Some authors have tentatively suggested that the different subtypes of delirium (hypoactive, hyperactive, and mixed) reflect different underlying pathophysiologies and therefore have differing treatment requirements, treatment responsiveness, and outcomes. 19[Level of evidence: II] One study suggests that the hypoactive subtype is as distressing for patients as the hyperactive or mixed subtypes. 20[Level of evidence: II] Some preliminary evidence suggests that the hypoactive subtype is less responsive to neuroleptic treatment. 18 Evidence would therefore suggest that all three subtypes warrant symptomatic treatment but that the hyperactive or agitated subtype requires more sedating agents. Although psychostimulants have been proposed for the treatment of hypoactive delirium, 21[Level of evidence: III]; 22[Level of evidence: II] little empirical evidence attests to their benefit. In a prospective clinical study of 14 patients with advanced cancer and hypoactive delirium, patients demonstrated improvement in cognitive function after receiving 20 to 50 mg of methylphenidate hydrochloride per day. 23[Level of evidence: III] Relatively higher doses of stimulants (>10 mg of methylphenidate) should be used with caution in delirious patients because such doses can contribute to the unmasking of paranoia and confusion and can lead to agitation. Clinical experience suggests that psychostimulants should be avoided in the presence of hallucinations or delusions. 2
Haloperidol, a neuroleptic agent with potent antidopaminergic properties, is still considered the drug of choice for the treatment of delirium in the patient with cancer. 1 7 Haloperidol has a low incidence of cardiovascular and anticholinergic effects. Although there is no high-level evidence to substantiate the use of haloperidol in patients with cancer, a double-blind trial of haloperidol, chlorpromazine, and lorazepam in the treatment of hospitalized patients with delirium and acquired immunodeficiency syndrome suggested that haloperidol and chlorpromazine were equivalent in efficacy, and both were associated with a very low prevalence of extrapyramidal side effects. Lorazepam, however, was ineffective and associated with adverse effects, resulting in early closure of this arm of the protocol. 24[Level of evidence: I] The optimal dose range of haloperidol for patients with delirium has not been determined. Consensus guidelines recommended initial doses in the range of 1 to 2 mg every 2 to 4 hours as needed and lower starting doses, such as 0.5 mg every 4 hours as needed, in elderly patients. 7
Haloperidol can be administered orally, intravenously, subcutaneously, or intramuscularly. Parenteral doses are roughly twice as potent as oral doses. Peak plasma concentrations are achieved 2 to 4 hours after an oral dose, and measurable plasma concentrations occur 15 to 30 minutes after intramuscular administration. Haloperidol appears to cause fewer extrapyramidal side effects (akathisia, cogwheeling, rigidity, dystonias, and akinesias) when administered intravenously. The extrapyramidal side effects can be treated with agents such as benztropine in doses of 1 to 2 mg once or twice a day. Neuroleptic malignant syndrome, a rare complication with haloperidol use, is characterized by hyperthermia, increased mental confusion, leukocytosis, muscular rigidity, myoglobinuria, and high serum creatinine phosphokinase. Injectable haloperidol is approved by the U.S. Food and Drug Administration (FDA) only for intramuscular administration. However, it is frequently administered intravenously to treat agitated delirium. Intravenous administration and higher doses of haloperidol have been associated with risk of sudden death due to Torsades de Pointes (TdP) and QTc prolongation. In this context, an FDA alert has raised concerns about intravenous use of haloperidol. 25 The updated labeling of haloperidol includes a strong warning about the TdP and QTc prolongation issues, especially in patients with specific cardiovascular risk factors, and recommends consideration of alternative agents. The FDA alert also recommends extreme caution and close electrocardiogram monitoring if haloperidol is administered intravenously. The FDA alert is based on case reports and small case-control studies. There are no randomized studies investigating this issue, and the data from case reports and case-control studies are confounded by multiple factors (e.g., comorbid conditions and other agents known to cause QTc prolongation). 26 In select cases, the risk-benefit equation might favor intravenous use of haloperidol, especially in patients with established intravenous access; in addition, no other currently available antipsychotics can be delivered intravenously, and many of the currently available typical and atypical antipsychotics are also associated with QTc prolongation. Before the newer atypical antipsychotics became available, chlorpromazine was considered an alternative to haloperidol, although it is associated with orthostatic hypotension and a greater level of sedation.
Risperidone is an atypical antipsychotic with fewer extrapyramidal side effects than haloperidol. Risperidone has been used with increasing frequency in the treatment of delirium, although only case reports attest to its efficacy for the treatment of this condition. 27 Risperidone is available in oral tablet and liquid formulations; dosing begins at 1 to 2 mg per day in two divided daily doses that are titrated, if necessary, to a total daily dose of 4 to 6 mg per day. Olanzapine, another atypical antipsychotic, has been studied more extensively. In an open trial of 79 hospitalized cancer patients with delirium, 18 an oral formulation was used with an initial dose range of 2.5 to 10 mg and a mean of 3 mg per dose in two daily doses. Seventy-six percent of patients had complete resolution of their delirium while taking olanzapine. No patients experienced extrapyramidal side effects, but 30% experienced sedation, which was usually not severe enough to discontinue treatment. Predictors of a poor response included age over 70 years, history of dementia, central nervous system involvement with cancer, hypoxia, hypoactive subtype, and delirium of severe intensity. Olanzapine is also reported to have antiemetic and possibly analgesic properties, although it is not used primarily for these indications. 28[Level of evidence: II] The lack of an available parenteral formulation for both olanzapine and risperidone is a disadvantage, especially in the context of agitated delirium.
Except for lorazepam and midazolam in selected situations, benzodiazepines are generally not recommended for the treatment of delirium. 7 Lorazepam is a short-acting agent, and its use is largely reserved for the treatment of alcohol or benzodiazepine withdrawal. Lorazepam (0.51 mg orally or parenterally, every 12 hours) has also been used along with haloperidol in patients with delirium who are particularly sensitive to extrapyramidal side effects. Another exception is midazolam, a very short-acting benzodiazepine, which is given by continuous subcutaneous or intravenous infusion in doses ranging from 30 to 100 mg over 24 hours. Midazolam is used to achieve deep sedation, especially in a terminal hyperactive or mixed delirium when agitation is refractory to other treatments, for example, doses of haloperidol in the region of 20 mg per day. Similarly, methotrimeprazine (a very sedating neuroleptic) is also used to achieve deep sedation in doses of 6.25 to 25 mg subcutaneously or intravenously every 8 hours.
The decision to use a deep level of pharmacologically induced sedation in the treatment of agitated delirium often raises ethical concerns, fueled by the marked variability in the reported frequency (ranging from 10% to 52%) for this practice in patients dying from advanced cancer. Consistent with the goals of care, it is important that appropriate efforts are made to assess the reversibility of delirium, clarify the intent of sedation (the relief of refractory symptoms), and maintain clear communication with family members and health care team members regarding rationale and process. 2 4 (Refer to the Delirium and Sedation for Refractory Delirium and Other Intractable Symptoms section of this summary for more information.)
Delirium at the end of life often requires a pharmacological sedative approach. This issue cannot be considered in isolation from the ethical dilemma that it evokes. The need to sedate terminally ill patients for poorly controlled symptoms that include delirium, pain, dyspnea, and psychological effects has been reported frequently. 29 30 31 32; 33 34[Level of evidence: III] Although clinical experience suggests that good palliative care can effectively manage the symptoms of most cancer patients, patients may experience symptoms that can be termed refractory. 35 Although sedative drugs are a therapeutic option, the incidence of these refractory situations in advanced cancer patients is controversial. This highlights the need to distinguish between difficult and refractory symptoms. A clear understanding of the terminology describing sedation and sedative medications is necessary; however, the extent to which sedation has been used for managing agitated delirium is difficult to clarify because of inconsistent definitions and confusing terminology. 36 37 Nevertheless, agitated behavior requiring sedation that is variously described as delirium, terminal restlessness, mental anguish, and agitation is a recurring theme in the literature. 38
A systematic review of the definitions of sedation for symptom relief noted a marked variation in the literature. 37 A sedation definition was proposed to include two core factors:
This review defined palliative sedation as the use of sedative medications to relieve intolerable and refractory distress by the reduction in patient consciousness. The identified inconsistencies in the definition of sedation (i.e., primary versus secondary, light versus deep, and intermittent versus continuous sedation) should be subcategories of palliative sedation. 37
The use of palliative sedation for psychosocial and existential symptoms can be particularly controversial. Many ethical and clinical questions can arise for the clinicianquestions that are more easily resolved in the case of palliative sedation for pain and physical symptoms.
For example, the ethical basis for the use of terminal sedation (double effect) is less clearly applicable in the case of psychiatric symptoms. Under this principle, the intended effect (relieving psychological suffering) would be considered allowable as long as any risks or negative effects (i.e., shortened survival) are unintended by the professional. Difficulty arises here because the principle discusses only the professional's intention, when it is the patient's intention that can be unclear and potentially problematic. Is the depressed patient who no longer wants to suffer depressive symptoms asking only for that relief, or is it also the patient's intent to ask the professional to shorten his or her life? Clinicians who feel uncomfortable in such situations might want to seek guidance from their ethics committees.
Other difficult questions can arise from the potentially negative value that is culturally assigned to zoning out as a lower form of coping. Should the anxious patient who no longer wants to face the anxiety associated with the end of life and desires sedation be encouraged to work through such issues? Or is it allowable for the anxiety of such patients to be handled with sedation? How many alternatives should be tried before anxiety is considered unacceptable? When dealing with such requests, professionals should consider their own cultural and religious biases and the cultural and/or religious backgrounds of patients and their families.
Few studies detail the use of terminal sedation for psychiatric symptoms. Four palliative care programs in Israel, South Africa, and Spain participated in one survey. 39[Level of evidence: III] One unique study has described the Japanese palliative care experience around these issues. 40 41[Level of evidence: II]
Noting the limitations of surveys and retrospective chart reviews, 42 43 researchers have completed prospective studies to determine the use of sedation for uncontrolled symptoms in terminally ill patients. Four palliative care programs with inpatient units in Israel, South Africa, and Spain reported that 97 out of 387 patients (25%) required sedation. 39 In 59 of the 97 patients (60%), sedation was used for refractory delirium, with midazolam being the most common medication prescribed. A study of similar design in Canada reported that 80% of 150 patients developed delirium prior to death. 44 Of the 150 patients, 9 required sedation for refractory delirium. A retrospective study at the MD Anderson Cancer Center in Houston included 1,207 patients admitted to the palliative care unit. Palliative sedation was used in 15% of admissions. The most common indications were delirium (82%) and dyspnea (6%). Sedation in these circumstances is often used on a temporary basis and was reversible in 23% of this group of patients. 45
The relatively short period between onset of sedation and death has been consistently reported as 1 to 6 days. 38 It has been noted, however, that palliative care patients who have delirium and appear extremely ill may still have treatable, reversible complications. 46 47 48 49
Various medications have been utilized for palliative sedation. 50 51 These include benzodiazepines (midazolam), phenothiazines (methotrimeprazine, chlorpromazine), butyrophenones (haloperidol), anaesthetic agents (propofol), and barbiturates. Midazolam is the drug most frequently reported as useful because of its rapid onset and ease of titration. Choice of medication is often determined by clinician preference and/or institutional policy.
Given the common association between delirium and sedation, it is important to have some understanding of the extensive literature on the ethical validity of using sedation management. 38 Numerous articles have addressed the importance of the double-effect argument as applied to the practice of sedation in palliative care. The concept of the double effect distinguishes between the compelling primary intent to relieve suffering and the unavoidable consequence of potentially accelerating death. 38 Legal opinions tend to support the doctrine of double effect as a major ethical foundation for the distinction between palliative care and euthanasia. 52 One study proposed physician intent, proportionality, and autonomy as the ethical principles relevant to palliative sedation therapy. A prospective study of 102 palliative care patients in 21 (out of 56) palliative care units in Japan found that these principle