Predictive Factors for Agitation Severity of Hyperactive Delirium in Terminally Ill Cancer Patients in a General Hospital Using Ordered Logistic Regression Analysis

Abstract

Background:

Despite the fact that many cancer patients worldwide die in general hospitals, there are few reports of the analysis of delirium in terminally ill cancer patients in this setting.

Purpose:

This study aimed to identify predictive factors for agitation severity of hyperactive delirium in terminally ill cancer patients in a general hospital.

Methods:

Participants were 182 consecutively admitted terminally ill cancer patients who died in a Japanese general hospital between April 2009 and March 2011. Variables present one week before death were extracted from the clinical records for regression analysis of factors potentially related to agitation severity of delirium. The prevalence and agitation severity of delirium were evaluated retrospectively. Multivariate ordered logistic regression analysis was performed to identify predictive factors.

Results:

Male sex [odds ratio (OR)=2.125, 95% confidence interval (CI)=1.111–4.067; P=0.0227]; total bilirubin (T-bil) [OR=1.557, CI=1.082–2.239; P=0.017]; antibiotics [OR=0.450, CI=0.219–0.925; P=0.0298]; nonsteroidal antiinflammatory drugs (NSAIDs) [OR=2.608, CI=1.374–4.950; P=0.0034]; and hematological malignancy [OR=3.903, CI=1.363–11.179; P=0.0112] were found to be statistically significant predictors for agitation severity of hyperactive delirium.

Conclusions:

Our study indicates that male sex, T-bil, antibiotic therapy, NSAID therapy, and hematological malignancy are significant predictors for agitation severity of hyperactive delirium in terminally ill cancer patients in a general hospital setting.

Introduction

Delirium is one of the psychiatric symptoms that occurs frequently (25% – 85%) in terminally ill cancer patients just before death.^1–5 Delirium is an acute brain syndrome with consciousness disturbance, psychomotor excitement, cognitive deficits, and psychomotor retardation, as opposed to dementia, in which a chronic organic cause affecting the brain is usually identified or likely.⁶ From a behavioral point of view, delirium can be classified as hypoactive; hyperactive (i.e., associated with the hypovigilant or hypervigilant level of consciousness); or mixed.⁶ In terminally ill cancer patients, delirium can be induced by several factors, such as metabolic disturbance, organ failure, and drugs;^6–8 and the same factors disturb recovery. Even though it is difficult to treat delirium in this population, pharmacological treatments using antipsychotics and/or sedative drugs may be appropriate for agitated delirium, because the condition may cause severe distress for both patients and family members.^9–11 Therefore, although delirium might be considered by some to be a natural part of the dying process,^9,12 the ability to predict the agitation severity of delirium and prevention of this aspect would represent an advance in clinical management of these patients.

Multiple studies have statistically identified predictive factors for delirium in these patients.^1–5,13 However, despite the fact that many cancer patients worldwide die in general hospitals, to the best of our knowledge there are few reports of the analysis of agitation severity of delirium in terminally ill cancer patients in this setting.^14,15 Therefore, a retrospective study was carried out with the primary aim of identifying predictive factors for agitation severity of hyperactive delirium in terminally ill cancer patients in a general hospital.

Methods

Study term and participants

Consecutively admitted adult cancer patients who died at the University Hospital of Kyoto Prefectural University of Medicine between April 2009 and March 2011 were enrolled in this study. This is a 1065-bed core hospital in the Kyoto prefecture, an acute care hospital with no palliative care unit. The inclusion criterion for this study was death due to cancer at the hospital after a stay of seven days or more, because the aim was to identify predictive factors during the week before death. Exclusion criteria were as follows: (1) patients maintained under medical continuous sedation during the final week of life; (2) patients who died suddenly due to unexpected causes (e.g., fatal arrhythmia, pulmonary embolism); (3) patients with dementia; and (4) patients who experienced delirium for more than one week before death.

Previous reports have indicated that in patients with cancer, delirium occurs with greater frequency during the few days before death;^1–5 therefore, we focused on agitation severity during the final week of life. This study was performed with the approval of the Ethics Review Board of Kyoto Prefectural University of Medicine.

Delirium assessment

The hyperactive delirium diagnoses were made by a psychiatrist with 10 years clinical experience using clinical records from the patients' last week. Mental status descriptions in the records were most often written by the treating physician and/or primary nurse. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) diagnostic criteria were used to define delirium.¹⁶ DSM-IV criteria include acute onset of consciousness disturbance, acute onset of cognition disturbance, and fluctuating symptoms. Referring to a previous study,¹⁷ delirium was coded as “present” if any key terms such as inattention, disorientation, hallucinations, agitation, and inadequate behavior were present and acute onset or acute change of symptoms was present. Because of the retrospective nature of the investigation and the great possibility of underestimation of hypoactive delirium, we defined only hyperactive delirium. The agitation severity of hyperactive delirium was also assessed from the records of the last week by the psychiatrist using the “psychomotor activity” item (Item 9) of the Memorial Delirium Assessment Scale (MDAS) following the previous studies.^13,18,19 The rationale for adopting this end-point was that (1) there is no validated method for assessing severity of agitation retrospectively, (2) several studies showed that MDAS Item 9 was associated with neurobehavioral dimension and severity of agitation,^20–22 and (3) it was assumed that remarkable events related to the patients with hyperactive delirium were usually well described in the medical records as a part of routine practice. Then the agitation severity (response variable) was categorized according to an ad hoc scale referring MDAS as follows: 0, no agitation (no episode of delirium or hypoactive delirium); 1, mild; 2, moderate; and 3, severe. The most severe symptom during the final week was ascertained. To calculate interrater reliability of the assessment of both diagnosis of delirium and severity of agitated delirium, 40 patients' records were randomized and assessed under blinded conditions by another psychiatrist with seven years of clinical experience. A kappa coefficient was calculated.

Extraction of variables

Variables possibly related to agitation severity of delirium were extracted from patients' clinical records for regression analysis. According to previous studies,^1–8 categories were chosen. Palliative care team intervention was defined as that occurring before the final week. Laboratory profiles were obtained from blood tests taken one week before death. Medication use was ascertained from prescriptions written one week before death, and anticancer drugs during the final three weeks. Morphine or fentanyl doses were converted to oral or transdermal daily doses, using the standard conversion ratios.²³

Statistical analyses

Multivariate ordered logistic regression analysis was used, since scoring for agitation severity of delirium was evaluated by a graded scale, and multiple factors involved in scoring were evaluated simultaneously. Variables were screened by examining for multicollinearity (correlation coefficient |r|>0.7), which occurs when correlations existing among variables results in use of an inappropriate model. A multivariate logistic regression model was constructed using forward stepwise selection among several candidate variables with a variable entry criterion of 0.25 and a variable retention criterion of 0.1 (JMP^® version 10; SAS Institute, Cary, NC). All statistical analyses were performed at a two-sided significance level of 0.05. Serum creatinine was categorized as either normal (<1.0 mg/dL) or abnormal (≥1.0 mg/dL), and T-bil was categorized as either normal (<1.0 mg/dL), a little high (≥1.0 mg/dL, <3.0 mg/dL), or high (≥3.0 mg/dL). These criteria were based on a previous study.²⁴ Hydration volume was categorized as none, peripheral hydration, or total parenteral nutrition. Statistical data were analyzed with JMP^® version 10 (SAS Institute, Cary, NC). P value<0.05 was considered to be statistically significant.

Results

Of 317 adult cancer patients who were consecutively admitted to, and subsequently died at, the University Hospital of Kyoto Prefectural University of Medicine during the study period, 135 patients (42.6 %) were excluded based on the defined exclusion criteria, as follows: 4 died of causes not related to cancer (2 heart failures, 1 cerebral infarction, and 1 postsurgical complication); 43 died less than one week after admission; 45 required medical continuous sedation during the final week; 5 died due to unexpected sudden change (2 brain hemorrhages and 2 unknown causes of death); 3 had dementia; and 35 developed prolonged delirium more than one week before death.

Table 1 shows the clinical characteristics of the 182 patients who were ultimately enrolled in this study, as well as various candidate factors possibly related to agitation severity of delirium in terminally ill cancer patients.

Table 1.

Patient Characteristics and All Extracted Variables and the Results of Univariate Analyses (N=182)

	Hyperactive delirium (n=80)	No hyperactive delirium (n=102)	P	OR (95% CI)	Entered into stepwise model
Demographic factors
Sex (male), n (%)	55 (69)	53 (52)	0.03	2.0 (1.1–3.5)	Yes
Age	63.7 (12.4)	63.5 (14.3)	0.56	1.1 (0.9–1.1)
PCT intervention final week, n (%)	22 (28)	28 (27)	0.66	0.9 (0.5–1.6)
Family support, n (%)	72 (90)	88 (86)	0.42	1.4 (0.6–3.5)
Marriage status, n (%)	66 (83)	75 (74)	0.16	1.7 (0.8–3.4)	Yes
Medical condition
Bone metastasis, n (%)	24 (30)	29 (28)	0.84	1.1 (0.6–2.0)
Liver metastasis, n (%)	33 (41)	37 (36)	0.66	1.1 (0.6–2.0)
Meningeal infiltration or brain metastasis, n (%)	4 (5)	11 (11)	0.11	0.4 (0.1–1.2)	Yes
Hepatic encephalopathy, n (%)	10 (13)	10 (10)	0.50	1.3 (0.6–3.2)
Laboratory test and physical measurement
BMI	20.4 (3.7)	20.5 (3.6)	0.62	0.9 (0.9–1.1)
CRP, mg/dL	7.6 (0.2–30.2)	5.7 (0.1–27.3)	0.03	1.1 (1.1–1.2)	Yes
AST, U/L	72 (8–477)	44 (9–671)	0.55	1.1 (0.9–1.1)
ALT, U/L	43 (6–382)	31 (8–524)	0.58	1.1 (0.9–1.1)
Albumin, g/dL	2.6 (0.5)	2.7 (0.6)	0.18	0.7 (0.4–1.2)	Yes
Total protein, mg/dL	5.7 (1.1)	5.6 (3.4)	0.23	1.1 (0.9–1.1)	Yes
T-bil (continuous), mg/dL	1.4 (0.2–32.1)	0.8 (0.3–24.7)	0.16	1.1 (0.9–1.1)
T-bil (category)	34/20/26	61/19/22	0.03	1.4 (1.1–2.0)	Yes
BUN, mg/dL	24.7 (6.3–105.8)	24.7 (6.2–151.1)	0.65	0.9 (0.9–1.1)
SCr (continuous), mg/dL	0.71 (0.2–4.7)	0.63 (0.2–7.2)	0.31	0.9 (0.7–1.2)
SCr (category), n (%)	26 (33)	29 (28)	0.89	1.1 (0.6–1.9)
Serum sodium, mEq/L	134 (5.5)	135 (6.3)	0.87	0.9 (0.9–1.1)
Serum potassium, mEq/L	4.5 (0.91)	4.4 (0.76)	0.24	1.2 (0.9–1.7)	Yes
Serum calcium, mg/dL	8.8 (2.0)	8.6 (1.0)	0.79	1.1 (0.8–1.4)
WBC,×10³ /μL	9.7 (0.1–62.9)	9.9 (1.2–68.2)	0.87	0.9 (0.9–1.1)
Lymphocyte,×10³ /μL	0.76 (0.09–7.8)	0.88 (0.1–3.2)	0.84	1.1 (0.7–1.5)
PLT,×10³ /μL	134 (7–774)	189 (7–601)	0.15	0.9 (0.9–1.1)	Yes
Hb, g/dL	9.0 (2.2)	9.5 (2.4)	0.07	0.9 (0.8–1.1)	Yes
Concomitant medications
Anticancer drugs (within two weeks), n (%)	15 (19)	14 (14)	0.31	1.5 (0.7–3.1)
Metoclopramide or domperidone, n (%)	6 (8)	10 (10)	0.84	0.9 (0.3–2.5)
Histamine 1-antihistamines, n (%)	4 (5)	8 (8)	0.44	0.6 (0.2–2.1)
Histamin 2-antihistamines, n (%)	30 (38)	38 (37)	0.82	1.1 (0.6–1.9)
Antibiotics, n (%)	25 (31)	44 (43)	0.21	0.7 (0.4–1.2)	Yes
Antiviral, n (%)	4 (5)	5 (5)	0.98	0.9 (0.3–3.6)
Hydration volume (category)	9/53/18	26/56/20	0.07	1.5 (1.0–2.4)	Yes
Antidepressants exclude TCA, n (%)	4 (5)	1 (1)	0.25	2.6 (0.5–13.2)	Yes
Antipsychotics (excluding prochlorperazine), n (%)	10 (13)	11 (11)	0.87	1.1 (0.5–2.6)
Benzodiazepines in last 24 hours, n (%)	32 (40)	42 (41)	0.91	0.9 (0.6–1.7)
Anticonvulsants, n (%)	4 (5)	3 (3)	0.79	1.2 (0.3–5.1)
Steroids, n (%)	40 (50)	48 (47)	0.89	1.1 (0.6–1.8)
NSAIDs, n (%)	39 (49)	34 (33)	0.02	1.9 (1.1–3.4)	Yes
Opioid, n (%)	42 (53)	54 (53)	0.75	0.9 (0.5–1.6)
Daily dosage of opioid
morphine, mg (oral morphine equivalents)	120 (20–600)	60 (5–360)	0.37	1.1 (0.9–1.1)
oral oxycodone, mg	12.5 (2.5–60)	20 (2.5–40)	0.43	1.1 (0.9–1.1)
fentanyl, μg/h (transdermal fentanyl equivalents)	4.2 (1.05–50.4)	4.2 (2.1–8.4)	0.32	1.1 (0.9–1.1)
Primary sites of malignancy
Lung, n (%)	6 (8)	15 (15)	0.19	0.5 (0.2–1.4)	Yes
Gastric, n (%)	8 (10)	11 (11)	0.68	0.8 (0.3–2.1)
Hematological malignancies, n (%)	12 (15)	11 (11)	0.19	1.7 (0.8–3.9)	Yes
Breast, n (%)	3 (4)	5 (5)	0.78	0.8 (0.2–3.3)
Colon, n (%)	6 (8)	8 (8)	0.73	0.8 (0.3–2.4)
Pancreas, n (%)	4 (5)	7 (7)	0.58	0.7 (0.2–2.4)
Esophageal, n (%)	5 (6)	7 (7)	0.92	0.9 (0.3–2.9)
Liver, n (%)	11 (14)	11 (11)	0.64	1.2 (0.5–2.9)
Cholangiocarcinoma, n (%)	5 (6)	5 (5)	0.95	0.9 (0.3–3.3)
Gynecologic, n (%)	5 (6)	4 (4)	0.66	1.3 (0.4–4.7)
Ontological, n (%)	5 (6)	8 (8)	0.58	0.7 (0.2–2.3)
Urological, n (%)	7 (9)	4 (4)	0.21	2.0 (0.7–6.3)	Yes

Values are median (range) or mean (SD) when appropriate.

Binary scales were female=0 and male=1 for sex;<1.0 mg/dL=0 and ≥1.0 mg/dL=1 for serum creatinine (2); and absent=0 and present=1 for others.

Ordinal scales were <1.0 mg/dL=0, >1.0 mg/dL but ≤3.0 mg/dL=1, and >3.0 mg/dL=2 for T-bil (2); none=0, peripheral parenteral nutrition=1, and total parenteral nutrition=2 for hydration volume.

ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body-mass index; BUN, blood urea nitrogen; CRP, C-reactive protein; Hb, hemoglobin; NSAIDs, nonsteroidal antiinflammatory drugs; PCT, palliative care team; PLT, platelet; SCr, serum creatinine; T-bil, total bilirubin; TCA, tricyclic antidepressant; WBC, white blood cell.

Table 2 presents agitation severity in all patients analyzed. The total prevalence in the study population of hyperactive delirium during the final week of life was 44.0% (n=80). There was good reliability between raters for the assessment of hyperactive delirium (A kappa coefficient=0.832; 95%CI=0.565–0.922) and ratings agitation severity of MDAS Item 9 (A kappa coefficient=0.605; 95% CI=0.361–0.85).

Table 2.

Categorization of the Agitation Severity in Terminally Ill Cancer Patients

Response	n=182
0	102
1	39
2	35
3	6

The response was categorized according to an ad hoc scale referring MDAS as follows: 0, no agitation (no episode of delirium or hypoactive delirium); 1, mild; 2, moderate; and 3, severe.

This analysis identified five independent predictors: male sex, T-bil, antibiotics, NSAIDs, and hematological malignancy (see Table 3).

Table 3.

Results of Ordered Logistic Regression Analysis for Variables Extracted by Forward Selection

			CI of OR
Variable	P	OR	Lower 95%	Upper 95%
Sex (male)	0.0227	2.125	1.111	4.067
Meningeal infiltration or brain metastasis	0.1907	0.421	0.115	1.539
T-Bil (category)	0.017	1.557	1.082	2.239
Hb	0.0622	0.873	0.757	1.007
Antibiotics	0.0298	0.450	0.219	0.925
Hydration volume	0.1095	1.507	0.912	2.490
NSAIDs	0.0034	2.608	1.374	4.950
Hematological malignancies	0.0112	3.903	1.363	11.179

Data p<0.05 indicated in bold and italic.

CI, confidence interval; Hb, hemoglobin; NSAIDs, nonsteroidal antiinflammatory drugs; OR, odds ratio; T-bil, total bilirubin.

Discussion

The multivariate logistic regression analysis used in this study demonstrated that male sex, T-bil, antibiotics, NSAIDs, and hematological malignancy were significant predictors for agitation severity of hyperactive delirium in terminally ill cancer patients.

Consistent with previous reports, the current analysis showed that the agitation severity of delirium in terminally ill cancer patients tended to increase with elevated T-bil level^4,13 and in males.¹³ Previous studies also clarified that T-bil or male sex were risk factors for delirium.^25,26 Clinicians need to be alert to the greater risk of agitation of delirium in terminally ill cancer patients having these characteristics.

Regarding a correlation with antibiotics use, a previous study found that infection is a risk factor for delirium.²⁶ This result might suggest that fever caused by infection may be an actual predictive factor for agitation severity of delirium, and that minimizing its potential occurrence may be an additional reason to use antipyretics for patients with infection.

Previous studies demonstrated that NSAIDs are a risk factor for delirium in terminally ill cancer patients or those with a progressive deterioration of cognitive function.^7,8,27,28 On the other hand, the antiinflammatory effects were effective for the prevention of cognitive impairment even in chronic situations.²⁹ In the current study, NSAIDs were identified as a predictive factor for agitation severity of delirium in terminally ill cancer patients. The reason might be because NSAIDs are usually used in patients with uncontrollable pain³⁰ and fever. NSAIDs also might cause a potential accumulation of toxic metabolites due to decreased renal function, or anemia from gastrointestinal tract disturbances. To the best of our knowledge, there are few previous reports identifying hematological malignancy as a predictor of delirium in terminally ill cancer patients.³¹ Patients with hematological malignancy are sometimes in an isolated environment in their terminal stage, with the medical intention of avoiding the risk of infection; the environment factors could contribute to severity of agitation. Also, our patients with hematological malignancy were isolated in a private room or observation room (data not shown). Clinicians need to be alert to the greater risk of agitated delirium in patients with hematological malignancy. Caraceni and colleagues reviewed all drugs or toxic effects that affect central nervous system (CNS) cholinergic neurons are candidates for causing delirium.⁶ The blood-brain barrier of patients with hematological malignancy might be broken down due to polypharmacy or inflammation with high-dose chemotherapy or radiation, and so on. Thus, medicinal products and endogenous substances such as bilirubin may gain access to the CNS with resultant toxicity.³²

Other studies have demonstrated that factors such as hypoalbuminemia, hydration status, and medications are commonly associated with delirium in this patient population. However, laboratory profiles were obtained from blood tests taken one week before death in our study, showing that most patients suffered from malnutrition. Therefore, there might be no significant difference between the delirium and the no-delirium group in mean albumin levels one week before death. Although hydration status was not associated with severity of delirium significantly, it showed a high odds ratio. Medication use was ascertained from prescriptions written one week before death; opioids or steroids were already prescribed to more than half the patients (opioids, 52%; steroids, 53%). Thus delirium might not have occurred due to new prescription of these drugs, or the dose of opioids (see Table 1) and steroids (data not shown) were not so high, thus delirium might not have occurred due to these drugs.

Limitations

This study has several limitations. First, the retrospective nature of the investigation may have decreased the reliability of the data collected. The assessment of delirium depends on the descriptions or terms in the chart. The assessment of agitation severity also depends on the chart descriptions. Therefore, there is a possibility of misclassification (especially about assessment of hypoactive delirium) due to no documentation about delirium or agitation. Second, this study was performed at a single institute and involved a relatively small number of patients, so the results should be confirmed in a further multicenter study.

Conclusion

Male sex, hematological malignancy, T-bil, antibiotics, and NSAIDs were shown to be predictors for agitation severity of hyperactive delirium in terminally ill cancer patients in a general hospital setting. These findings should be considered preliminary and in need of further refinement and study.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Bruera

, Miller

, McCallion

, Macmillan

, Krefting

, Hanson

. Cognitive failure in patients with terminal cancer: A prospective study. J Pain Symptom Manage, 1992; 7:192–195.

Pereira

, Hanson

, Bruera

. The frequency and clinical course of cognitive impairment in patients with terminal cancer. Cancer, 1997; 79:835–842.

Lawlor

, Gagnon

, Mancini

, Pereira

, Hanson

, Suarez-Almazor

, Bruera

. Occurrence, causes, and outcome of delirium in patients with advanced cancer: A prospective study. Arch Intern Med, 2000; 160:786–794.

Morita

, Tei

, Tsunoda

, Inoue

, Chihara

. Underlying pathologies and their associations with clinical features in terminal delirium of cancer patients. J Pain Symptom Manage, 2001; 22:997–1006.

Gagnon

, Allard

, Gagnon

, Mérette

, Tardif

. Delirium prevention in terminal cancer: Assessment of a multicomponent intervention. Psychooncology, 2012; 21:187–194.

Caraceni

, Simonetti

. Palliating delirium in patients with cancer. Lancet Oncol, 2009; 10:164–172.

White

, McCann

, Jackson

. First do no harm.. Terminal restlessness or drug-induced delirium. J Palliat Med, 2007; 10:345–351.

Alagiakrishnan

, Wiens

. An approach to drug induced delirium in the elderly. Postgrad Med J, 2004; 80:388–393.

Breitbart

, Gibson

, Tremblay

. The delirium experience: Delirium recall and delirium-related distress in hospitalized patients with cancer, their spouses/caregivers, and their nurses. Psychosomatics, 2002; 43:183–194.

10.

Morita

, Hirai

, Sakaguchi

, Tsuneto

, Shima

. Family-perceived distress from delirium-related symptoms of terminally ill cancer patients. Psychosomatics, 2004; 45:107–113.

11.

Morita

, Akechi

, Ikenaga

, Inoue

, Kohara

, Matsubara

, Matsuo

, Namba

, Shinjo

, Tani

, Uchitomi

. Terminal delirium: Recommendations from bereaved families' experiences. J Pain Symptom Manage, 2007; 34:579–589.

12.

Breitbart

, Chochinov

, Passik

. Psychiatric aspects of palliative care. Doyle

, Hanks

, MacDonald

. Oxford Textbook of Palliative Medicine, 2nd. New York: Oxford University Press, 1998; 216–247.

13.

Morita

, Tei

, Inoue

. Impaired communication capacity and agitated delirium in the final week of terminally ill cancer patients: Prevalence and identification of research focus. J Pain Symptom Manage, 2003; 26:827–834.

14.

Fried

, van Doorn

, O'Leary

, Tinetti

, Drickamer

. Older persons' preferences for site of terminal care. Ann Intern Med, 1999; 131:109–112.

15.

Cohen

, Houttekier

, Onwuteaka-Philipsen

, Miccinesi

, Addington-Hall

, Kaasa

, Bilsen

, Deliens

. Which patients with cancer die at home? A study of six European countries using death certificate data. J Clin Oncol, 2010; 28:2267–2273.

16.

American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Text revision. DSM IV–TR. Washington, DC: American Psychiatric Press, 2000.

17.

Inouye

, Leo-Summers

, Zhang

, Bogardus

Jr , Leslie

, Agostini

. A chart-based method for identification of delirium: Validation compared with interviewer ratings using the confusion assessment method. J Am Geriatr Soc, 2005; 53:312–318.

18.

Breitbart

, Rosenfeld

, Roth

, Smith

, Cohen

, Passik

. The Memorial Delirium Assessment Scale. J Pain Symptom Manage, 1997; 13:128–137.

19.

Nagase

, Okamoto

, Tsuneto

, Tanimukai

, Matsuda

, Okishiro

, Oono

, Tsugane

, Takagi

, Uejima

. A retrospective chart review of terminal patients with cancer with agitation and their risk factors. J Palliat Med, 2012; 15:1185–1190.

20.

Morita

, Tsunoda

, Inoue

, Chihara

, Oka

. Communication Capacity Scale and Agitation Distress Scale to measure the severity of delirium in terminally ill cancer patients: A validation study. Palliat Med, 2001; 15:197–206.

21.

Morita

, Chinone

, Ikenaga

, Miyoshi

, Nakaho

, Nishitateno

, Sakonji

, Shima

, Suenaga

, Takigawa

, Kohara

, Tani

, Kawamura

, Matsubara

, Watanabe

, Yagi

, Sasaki

, Higuchi

, Kimura

, Abo

, Ozawa

, Kizawa

, Uchitomi

. Japan Pain, Palliative Medicine, Rehabilitation, Psycho-Oncology Study Group: Efficacy and safety of palliative sedation therapy: A multicenter, prospective, observational study conducted on specialized palliative care units in Japan. J Pain Symptom Manage, 2005; 30:320–328.

22.

Lawlor

, Nekolaichuk

, Gagnon

, Mancini

, Pereira

, Bruera

. Clinical utility, factor analysis, and further validation of the memorial delirium assessment scale in patients with advanced cancer: Assessing delirium in advanced cancer. Cancer, 2000; 88:2859–2867.

23.

Hanks

GCW

, Cherny

. Opioid analgesic therapy. Doyle

, Hanks

GCW

, MacDonald

. Oxford Textbook of Palliative Medicine, 2nd. New York: Oxford Medical Publications, 1998; 331–355.

24.

Durand

, Valla

. Assessment of the prognosis of cirrhosis: Child-Pugh versus MELD. J Hepatol, 2005; 42:S100–S107.

25.

Dubois

, Bergeron

, Dumont

, Dial

, Skrobik

. Delirium in an intensive care unit: A study of risk factors. Intensive Care Med, 2001; 27:1297–1304.

26.

Schor

, Levkoff

, Lipsitz

, Reilly

, Cleary

, Rowe

, Evans

. Risk factors for delirium in hospitalized elderly. JAMA, 1992; 267:827–831.

27.

Macknight

, Rojas-Fernandez

. Celecoxib- and rofecoxib-induced delirium. J Neuropsychiatry Clin Neurosci, 2001; 13:305–306.

28.

Hoppmann

, Peden

, Ober

. Central nervous system side effects of nonsteroidal anti-inflammatory drugs: Aseptic meningitis, psychosis, and cognitive dysfunction. Arch Intern Med, 1991; 151:1309–1313.

29.

Cunningham

, Skelly

. Non-steroidal anti-inflammatory drugs and cognitive function: Are prostaglandins at the heart of cognitive impairment in dementia and delirium? J Neuroimmune Pharmacol, 2012; 7:60–73.

30.

Davis

, Walsh

, Lagman

, LeGrand

. Controversies in pharmacotherapy of pain management. Lancet Oncol, 2005; 6:696–704.

31.

Fadul

, El Osta

, Dalal

, Poulter

, Bruera

. Comparison of symptom burden among patients referred to palliative care with hematologic malignancies versus those with solid tumors. J Palliat Med, 2008; 11:422–427.

32.

Zeevi

, Pachter

, McCullough

, Wolfson

, Kuchel

. The blood-brain barrier: Geriatric relevance of a critical brain-body interface. Am Geriatr Soc, 2010; 58:1749–1757.