Predictors of Delirium in Corticosteroid-Treated Patients with Advanced Cancer: An Exploratory,Multicenter,Prospective,Observational Study

Abstract

Background:

Corticosteroids are often used to treat fatigue and anorexia, but occasionally produce delirium. Information on the predictors of delirium in corticosteroid-treated cancer patients remains limited.

Objective:

To identify potential factors predicting the development of delirium in corticosteroid-treated cancer patients.

Design:

An exploratory, multicenter, prospective, observational study.

Setting/Subjects:

Inclusion criteria for this study were patients who had metastatic or locally advanced cancer and a fatigue or anorexia intensity score of 4 or more on a 0–10 Numerical Rating Scale.

Measurement:

Univariate and multivariable analyses were performed to identify the predictors of delirium diagnosed by the Confusion Assessment Method (CAM) within three days of initiation of corticosteroids.

Results:

Among 207 patients administered corticosteroids, 35 (17%; 95% confidence interval [CI] 12%–23%) developed at least one episode of delirium diagnosed by the CAM. Factors predictive of the development of delirium were as follows: Palliative Performance Scale ≤20, Eastern Cooperative Oncology Group Performance Status (ECOG PS) = 4, the Support Team Assessment Schedule (STAS) score of drowsiness >1, concurrent opioid use, parenteral hydration volume ≤500 mL, and the absence of lung metastasis. A multivariable analysis identified the independent factors predicting responses as ECOG PS = 4 (odds ratio [OR] 4.0; 95% CI 1.7–9.3), STAS score of drowsiness >1 (OR 3.4; 95% CI 1.4–8.2), and concurrent opioid use (OR 3.7; 95% CI 1.0–13).

Conclusion:

Delirium in corticosteroid-treated advanced cancer patients may be predicted by PS, drowsiness, and concurrent opioid use. Larger prospective studies are needed to confirm these results.

Introduction

Delirium is a very distressing symptom among patients receiving palliative care. In oncology and palliative care settings, the incidence of delirium ranges between 18%¹ and 88%,² and increased in weeks to days before death.^2–4 The etiologies of delirium are multifactorial, including metabolic dysfunction, preexisting cognitive dysfunction (i.e., dementia),⁵ and medications.^1,2 Drugs have been reported to cause or at least contribute to the development of delirium in 11%–30% of cases.⁶ Exposure to corticosteroids and opioids is associated with an increased risk of delirium in hospitalized cancer patients.⁷ Corticosteroid-induced psychosis may result from corticosteroids increasing the release of dopamine in the nucleus accumbens through the overstimulation of dopamine neurons in the ventral tegmental area (VTA).⁸ On the contrary, opioids may be psychotogenic by enhancing the activity of VTA dopamine neurons through μ-opioid receptors located on gamma-aminobutyric acid (GABA) neurons within the VTA, thereby increasing dopamine release in the nucleus accumbens.⁹

Corticosteroids are often used in the treatment of fatigue and anorexia^10–14; but occasionally induce delirium.¹⁵ In a review of corticosteroid-induced psychiatric complications in cancer patients, the incidence of steroid psychiatric complications was estimated with a possible range of 3%–48%¹⁶ and was not specifically in palliative settings. Corticosteroid-induced delirium has not yet been studied in detail in advanced cancer patients.

An understanding of how often adverse effects occurs in real-world patients is important for assisting clinicians to make a medical decision whether an intervention is appropriate for a terminally ill patient.^17,18 Identifying the predictors of delirium in cancer patients who receive corticosteroid therapy will also be of value for clinicians to provide information regarding clinical decisions on whether the use of corticosteroids is appropriate. However, to the best of our knowledge, the potential predictors of delirium in corticosteroid-treated patients have not yet been systemically explored in palliative care settings.

Therefore, the aim of this prospective observational study was to identify the potential predictors of delirium with corticosteroid therapy in terminally ill cancer patients receiving specialized palliative care services. We also explored the prevalence of delirium with corticosteroids in this population.

Subjects and Methods

This was a part of another multicenter cohort study involving advanced cancer patients in Japan.¹⁹ All consecutive patients who received corticosteroids to treat fatigue or anorexia were enrolled and monitored. Attending physicians, all of whom were palliative care specialists, recorded all data. This study was conducted in accordance with the ethical standards of the Declaration of Helsinki and the ethical guidelines for epidemiological research of the Ministry of Health, Labour and Welfare in Japan, and was approved by the local institutional review boards of all participating institutions.

Subjects

Study subjects were advanced cancer patients receiving specialized palliative care services at 22 sites in Japan. Patients were enrolled consecutively from August 2012 to November 2014.

Patients were eligible if they were 20 years of age or older, had a histologically, cytologically, or clinically confirmed malignancy, had metastatic or locally advanced cancer, were receiving specialized palliative care services (i.e., admitted to inpatient palliative care units or referred to a palliative care consultation team), had a fatigue or anorexia intensity score of 4 or more on a 0–10 Numerical Rating Scale (NRS) (worst during the last 24 hours, 0 = no fatigue or anorexia to 10 = worst possible fatigue or anorexia), and initially receiving corticosteroid therapy. Before the administration of corticosteroids, patients with delirium, dementia, or organic brain disorders were excluded.

Intervention

Before starting this cohort study, the research group shared clinical practice, and the recommended doses of corticosteroids were determined based on the consensus among study sites with reference to international expert consensus^20–23: betamethasone 1–4 mg per day; dexamethasone 1–4 mg per day; prednisolone 10–30 mg per day, orally, intravenously, or subcutaneously. We calculated the dosage as betamethasone 1 mg = dexamethasone 1 mg = prednisolone 6.25 mg. Due to the observational nature of this study, each physician was allowed to adjust doses where clinically appropriate, and patients were permitted to concurrently receive palliative treatments for fatigue or anorexia throughout the study period.

Measurements

Endpoints

The primary endpoint was the development of delirium diagnosed using the Confusion Assessment Method (CAM) short-version. During the three days of administration of corticosteroids, patients were assessed daily for the presence of delirium using the CAM by palliative care specialists as a part of routine practice. Specialists directly examined patients at least once a day. The CAM is a valid and reliable tool for the identification of delirium.²⁴ In the CAM short-version, a diagnosis of delirium is suspected if patients have both (1) an acute onset and fluctuating course of mental status changes and (2) inattention, and had either (3) disorganized thinking or (4) an altered level of consciousness. It has a sensitivity of 94%–100% and specificity of 90%–95%.²⁴ Patients were divided into delirium and nondelirium groups based on the CAM. Patients who developed at least one episode of delirium were defined as the delirium group.

As a secondary endpoint, we used item 9 of the Memorial Delirium Assessment Scale (MDAS) to evaluate the frequency and proportion of the psychomotor activity of delirium.^25,26 MDAS is a valid and reliable instrument to assess the severity of delirium, and the psychometric properties of the Japanese version have also been established.²⁷ The sensitivity and specificity of the MDAS ≥10 were 82.4% and 75.0%, respectively.²⁵ Item 9 quantified increased psychomotor activity on a 4-point scale (0 = none, normal psychomotor activity; 1 = mild, hypoactivity is barely noticeable, expressed as slight slowing of movement. Hyperactivity is barely noticeable or appears as simple restlessness; 2 = moderate, hypoactivity is undeniable, with marked reduction in the number of movements or marked slowness of movement; subject rarely spontaneously moves or speaks. Hyperactivity is undeniable, subject moves almost constantly; in both cases, examination is prolonged as a consequence; and 3 = severe, hypoactivity is severe; patient does not move or speak without prodding or is catatonic. Hyperactivity is severe; patient is constantly moving, overreacts to stimuli, requires surveillance and/or restraint; getting through the examination is difficult or impossible).

Potential predictors

Potential predictive factors associated with delirium were recorded before the administration of corticosteroids by the treating palliative care physicians. The potential predictors investigated included (Table 3) age, gender, primary tumor sites, metastasis sites, comorbidities, indicators of general conditions (Eastern Cooperative Oncology Group Performance Status [ECOG PS],²⁸ Palliative Performance Scale [PPS]²⁹), baseline NRS of fatigue and anorexia, other symptoms (the Support Team Assessment Schedule, Japanese version: STAS-J of pain, dyspnea, insomnia, drowsiness, nausea, and constipation),³⁰ medical conditions (presence or absence of fever, infection, hypoxia, and fluid retention symptoms), palliative care setting (palliative care unit/inpatients hospice or palliative care team), concurrent medical interventions (hypnotics, opioids, anticholinergics, anxiolytics, parenteral hydration volume, and indwelling urinary catheter), and laboratory findings if patients received laboratory examinations during the last two weeks for clinical purposes.

The ECOG PS score runs from 0 to 5 (0 = Asymptomatic; 1 = Symptomatic but completely ambulatory; 2 = Symptomatic, <50% in bed during the day; 3 = Symptomatic, >50% in bed, but not bedbound; 4 = Bedbound; 5 = Death).²⁸

STAS-J is a valid and reliable proxy symptom assessment scale that rates the intensity of symptoms on a 5-point scale (0 = None; 1 = Occasional or grumbling single or few symptom(s). Patient has usual activity and is not bothered to be rid of symptom; 2 = Moderate distress, occasional bad days, symptoms limit some activity possible within extent of disease; 3 = Severe symptom(s) present often. Activities and concentration markedly affected by symptom(s); and 4 = Severe and continuous overwhelming symptom(s). Unable to think of other matters).³⁰ STAS-J was assessed by the treating palliative care physicians. Pleural effusion and ascites were defined as being present if they were physically detectable. Comorbidity and infection were diagnosed on the basis of the clinical judgment of the treating palliative care specialists. All concurrent treatments with the potential to cause delirium (i.e., opioids, anticholinergics, anxiolytics, and hypnotics) were also recorded. Due to the exploratory nature of this study, we selected these potential predictive factors associated with delirium from a literature review^31,32 and discussions among the authors.

Statistical analysis

We performed descriptive statistics for patient demographics, comorbidities, and treatment variables. We did not perform formal sample size calculation for this study aim, because this was a secondary analysis of another multicenter cohort study.¹⁹ We calculated the frequency and 95% confidence interval (CI) of the proportion of patients with a positive CAM test (delirium).

We compared the proportion of patients with a positive CAM test between patients with and without each potential predictor using the Fisher's exact test. Furthermore, we performed a logistic multivariable regression analysis to explore independent factors predicting the development of delirium. We entered all factors with p < 0.2 identified in a univariate analysis. We did not enter PPS, which potentially correlated with PS. We additionally calculated the frequency and proportion of patients with MDAS item 9 scores ≥1.

All results reported in this study were based on two-sided tests. p Values of 0.05 or lower were considered significant. Because of the exploratory nature of this study, we had decided not to use the Bonferroni correction for multiple comparisons. Analyses were performed using the Statistical Package for the Social Sciences (version 22.0, SPSS, Inc., Chicago, IL).

Results

A total of 207 cancer patients were enrolled in this study. Of these, 7 patients died due to progression of the underlying disease before day 3. Table 1 summarizes the baseline characteristics of the 207 patients analyzed. Median survival was 24 days, and the survival time after the administration of corticosteroids was <4 weeks in 104 patients (50%). A total of 165 out of 207 patients (80%) received betamethasone, followed by prednisolone (11%) and dexamethasone (9.7%). Median daily doses (betamethasone equivalent dosages; 1 mg betamethasone = 1 mg dexamethasone = 6.25 mg prednisolone) on days 1 and 3 were 2.4 mg (range 1–12 mg) and 3.0 mg (range 1–8 mg), respectively. Administration routes were oral 36%, subcutaneous 15%, and intravenous 49%.

Table 1.

Baseline Patient Demographic and Clinical Characteristics (N = 207)

Characteristics	N (%)
Age, years (median, range)	73 (38–97)
Sex
Male	98 (47)
Female	109 (53)
Primary tumor sites
Esophagus, stomach, colon, and rectum	68 (33)
Liver, pancreas, and bile duct	44 (21)
Lung	43 (21)
Kidney, urinary tract, and prostate	17 (8.2)
Ovary and uterus	12 (5.8)
Breast	7 (3.4)
Others	16 (7.7)
Metastasis sites
Lung	82 (40)
Liver	95 (46)
Bone	56 (27)
Brain	18 (8.7)
Comorbidity
Diabetes mellitus	30 (14)
Cerebral vascular disease	6 (2.9)
General condition
ECOG PS
0–3	158 (76)
4	49 (24)
Median PPS (range) (N = 206)	40 (10–70)
>20	174 (84)
≤20	32 (16)
Baseline intensity of physical symptoms (median, range)
NRS for fatigue (N = 200)	7 (0–10)
NRS for appetite loss (N = 200)	8 (0–10)
STAS score for pain	1 (0–4)
STAS score for dyspnea	0 (0–4)
STAS score for insomnia	1 (0–4)
STAS score for drowsiness	1 (0–4)
STAS score for nausea	0 (0–4)
STAS score for constipation	1 (0–4)
Medical conditions
Fever	7 (3.4)
Infection	9 (4.3)
Hypoxia	27 (13)
Ascites	78 (38)
Pleural effusion	62 (30)
Peripheral edema	109 (53)
Palliative care setting
Palliative care unit/inpatient hospice	165 (80)
Palliative care team	42 (20)
Concurrent medical interventions
Hypnotics	94 (45)
Opioids	148 (71)
Anticholinergics	8 (3.9)
Anxiolytics	22 (11)
Parenteral hydration volume, mL (median, range)	0 (0–1750)
>500 mL	29 (14)
Indwelling urinary catheter	36 (17)
Laboratory (median, range)
White blood cell count (/mm³, N = 192)	7560 (1400–49,880)
Lymphocytes (%, N = 184)	11 (0.9–41.7)
Hemoglobin (g/dL, N = 182)	10 (4.7–14.6)
Serum albumin (g/dL, N = 183)	2.8 (1.0–5.2)
Total bilirubin (mg/dL, N = 191)	0.6 (0.1–26.2)
Urea nitrogen (mg/dL, N = 191)	17.1 (2.8–66.5)
Creatinine (mg/dL, N = 192)	0.68 (0.21–3.97)
Serum sodium (mEq/L, N = 192)	135 (125–148)
Corrected calcium (mg/dL, N = 180)	9.8 (7.0–13.7)
C-reactive protein (mg/dL, N = 186)	5.3 (0–28)
Survival, days (median, 95% CIs)	24 (20–28)

CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group Performance Status; PPS, Palliative Performance Scale; NRS, Numerical Rating Scale; STAS, Support Team Assessment Schedule.

Prevalence of delirium

During the study period, 35 patients (17%; 95% CI 12%–23%) developed delirium as measured by the CAM, while 12 (5.8%) developed mild and moderate hyperactivity as measured with MDAS item 9 (Table 2). The agreement between a positive CAM test and MDAS item 9 ≥ 1 was good (Cohen's kappa coefficient = 0.98; 95% CI 0.95–1.02).

Table 2.

Prevalence of Delirium

	N (%)	95% CI
Delirium diagnosed using the CAM
Negative	172 (83)	77–88
Positive	35 (17)	12–23
MDAS item 9
Activity of delirium
Hypoactive	12 (5.8)	3.0–9.9
Mild	7 (3.4)	1.4–6.8
Moderate	4 (1.9)	0.5–4.9
Severe	1 (0.5)	0–2.7
Hyperactive	12 (5.8)	3.0–9.9
Mild	7 (3.4)	1.4–6.8
Moderate	5 (2.4)	0.8–5.5
Severe	0 (0)	0–1.4
Mixed	11 (5.3)	2.7–9.3
Mild	4 (1.9)	0.5–4.9
Moderate	4 (1.9)	0.5–4.9
Severe	3 (1.4)	0.3–4.2

CAM, Confusion Assessment Method; MDAS, Memorial Delirium Assessment Scale.

Twelve out of 35 patients stopped the administration of corticosteroids before day 3: the reasons were delirium caused by corticosteroids in five patients, death in four patients, and the ineffectiveness of corticosteroids in three patients (fatigue or anorexia not improved).

Predictors of delirium during corticosteroid therapy

In a univariate analysis (Table 3), the factors that positively correlated in the CAM test were as follows: PPS ≤20 (delirium rates 47% vs. 11%, p < 0.00001), ECOG PS = 4 (37% vs. 11%, p < 0.001), STAS score of drowsiness >1 (41% vs. 12%, p < 0.001), concurrent opioid use (22% vs. 5%, p < 0.01), parenteral hydration volume ≤500 mL (19% vs. 3.7%, p = 0.034), and absence of lung metastasis (22% vs. 9.8%, p = 0.036). No significant difference was observed in the initial dose of corticosteroids between patients who did and did not develop delirium (3.0 vs. 2.7 mg/day, p = 0.11). Furthermore, no significant difference was noted in the types of corticosteroids between the two groups (p = 0.14).

Table 3.

Prevalence of Delirium According to Patient Characteristics (Univariate Analysis, N = 207)

Item	Categories	Delirium rate N (%)	p	Missing N (%)
Age	>70 years (N = 120)	19 (16)	0.71
	≤70 years (N = 87)	16 (18)
Sex	Male (N = 98)	18 (18)	0.71
	Female (N = 109)	17 (16)
Primary tumor sites
Digestive cancer^a	Yes (N = 112)	23 (21)	0.14
	No (N = 95)	12 (13)
Lung cancer	Yes (N = 43)	6 (14)	0.65
	No (N = 164)	29 (18)
Metastasis sites
Lung	Present (N = 82)	8 (9.8)	0.036
	Absent (N = 125)	27 (22)
Liver	Present (N = 95)	15 (16)	0.71
	Absent (N = 112)	20 (18)
Bone	Present (N = 56)	11 (20)	0.53
	Absent (N = 151)	24 (16)
Brain	Present (N = 18)	4 (22)	0.52
	Absent (N = 189)	31 (16)
Comorbidity
Diabetes mellitus	Present (N = 30)	7 (23)	0.30
	Absent (N = 177)	28 (16)
Cerebral vascular disease	Present (N = 6)	2 (33)	0.27
	Absent (N = 201)	33 (16)
General condition
ECOG PS	0–3 (N = 158)	17 (11)	0.0001
	4 (N = 49)	18 (37)
PPS	>20 (N = 174)	20 (11)	0.000001	1 (0.5)
	≤20 (N = 32)	15 (47)
Baseline intensity of physical symptoms
NRS for fatigue	>7 (N = 80)	18 (23)	0.12	7 (3.4)
	≤7 (N = 120)	16 (13)
NRS for anorexia	>6 (N = 131)	25 (19)	0.56	7 (3.4)
	≤6 (N = 69)	10 (14)
STAS score for pain	>1 (N = 99)	18 (18)	0.71
	≤1 (N = 108)	17 (16)
STAS score for dyspnea	>1 (N = 45)	6 (13)	0.65
	≤1 (N = 162)	29 (18)
STAS score for insomnia	>1 (N = 62)	13 (21)	0.32
	≤1 (N = 145)	22 (15)
STAS score for drowsiness	>1 (N = 37)	15 (41)	0.0001
	≤1 (N = 170)	20 (12)
STAS score for nausea	>1 (N = 57)	8 (14)	0.54
	≤1 (N = 150)	27 (18)
STAS score for constipation	>1 (N = 66)	10 (15)	0.70
	≤1 (N = 141)	25 (18)
Medical condition
Fever	Present (N = 7)	3 (43)	0.096
	Absent (N = 200)	32 (16)
Infection	Present (N = 9)	2 (22)	0.65
	Absent (N = 198)	33 (17)
Hypoxia	Present (N = 27)	6 (22)	0.42
	Absent (N = 180)	29 (16)
Ascites	Present (N = 78)	16 (21)	0.34
	Absent (N = 129)	19 (15)
Pleural effusion	Present (N = 62)	10 (16)	1.0
	Absent (N = 145)	25 (17)
Peripheral edema	Present (N = 109)	22 (20)	0.20
	Absent (N = 98)	13 (13)
Palliative care setting	Palliative care unit/inpatient hospice (N = 165)	30 (18)	0.49
	Palliative care team (N = 42)	5 (12)
Initial daily doses as a betamethasone equivalent^b	>3 mg (N = 93)	15 (16)	0.85
	≤3 mg (N = 114)	20 (18)
Concurrent medical interventions
Hypnotics	Yes (N = 94)	20 (21)	0.19
	No (N = 111)	15 (14)
Opioids	Yes (N = 148)	32 (22)	0.0036
	No (N = 58)	3 (5.2)
Anticholinergics	Yes (N = 8)	1 (13)	1.0
	No (N = 199)	34 (17)
Anxiolytics	Yes (N = 22)	3 (14)	1.0
	No (N = 185)	32 (17)
Parenteral hydration volume	>500 mL (N = 29)	1 (3.4)	0.034
	≤500 mL (N = 178)	34 (19)
Indwelling urinary catheter	Yes (N = 36)	10 (28)	0.083
	No (N = 171)	25 (15)
Laboratory findings
White blood cell count (/mm³)	>8500 (N = 84)	16 (19)	0.71	15 (7.2)
	≤8500 (N = 108)	18 (17)
Lymphocytes (%)	>12 (N = 80)	17 (21)	0.24	23 (11)
	≤12 (N = 104)	15 (14)
Hemoglobin (g/dL)	>9 (N = 117)	18 (15)	0.68	25 (12)
	≤9 (N = 65)	12 (18)
Serum albumin (g/dL)	>3.0 (N = 72)	11 (15)	0.56	24 (12)
	≤3.0 (N = 111)	21 (19)
Total bilirubin (mg/dL)	>1.2 (N = 41)	8 (20)	0.82	16 (7.7)
	≤1.2 (N = 150)	26 (17)
Urea nitrogen (mg/dL)	>22 (N = 70)	10 (14)	0.44	16 (7.7)
	≤22 (N = 121)	24 (20)
Creatinine (mg/dL)	>0.8 (N = 66)	10 (15)	0.56	15 (7.2)
	≤0.8 (N = 126)	24 (19)
Serum sodium (mEq/L)	>135 (N = 95)	17 (18)	1.00	15 (7.2)
	≤135 (N = 97)	17 (18)
Corrected calcium (mg/dL)	>11 (N = 19)	2 (11)	0.53	27 (13)
	≤11 (N = 161)	30 (19)
C-reactive protein (mg/dL)	>1.0 (N = 158)	29 (18)	0.45	21 (10)
	≤1.0 (N = 28)	5 (18)

Digestive cancer includes cancers of the esophagus, stomach, colon, rectum, liver, pancreas, and bile duct.

Betamethasone equivalent (1 mg betamethasone = 1 mg dexamethasone = 6.25 mg prednisolone).

Multivariable analysis

A multivariable analysis identified the independent factors predicting delirium in corticosteroid-treated cancer patients as ECOG PS = 4, STAS score of drowsiness >1, and concurrent opioid use (Table 4). The initial doses and types of corticosteroids were not independent factors.

Table 4.

Predictors of Delirium in Corticosteroid-Treated Cancer Patients (Multivariate Analysis, N = 207)

Factor	OR	95% CI	p
ECOG PS 4	4.0	1.7–9.3	0.001
STAS score of drowsiness >1	3.4	1.4–8.2	0.008
Concurrent opioid use	3.7	1.0–13	0.043

R-square value = 0.28.

OR, odds ratio.

Discussion

To the best of our knowledge, this is the first multicenter prospective observational study to systemically explore the factors potentially predicting delirium in corticosteroid-treated advanced cancer patients.

The most important result of the present study was the identification of factors potentially predicting delirium in corticosteroid-treated patients. Significant predictors were classified into three categories: (1) general condition (ECOG PS = 4), (2) observer-rated severity of drowsiness, and (3) opioids as concurrent medication.

This study identified general conditions, such as the performance status, as one of the most important predictors of the development of delirium after the administration of corticosteroids. Since the incidence of delirium in terminally ill cancer patients in the palliative care unit increases before death,^2,3 we cannot assume that corticosteroids are the single cause of delirium; however, information that patients with PS of four have a fourfold higher risk of developing delirium will be useful for clinicians to determine whether the estimated benefits outweigh the risks for each patient.

The predictive value of drowsiness warrants further investigation. The etiologies of drowsiness are multifactorial, including electrolyte abnormalities, anemia, organ failure, and medications. In this study population, no significant differences were observed in the prevalence of electrolyte abnormalities, prevalence of organ failure, or level of hemoglobin between the delirium group and nondelirium group. Furthermore, medications inducing somnolence such as hypnotics, anxiolytics, and anticholinergics did not correlate with the development of delirium in this study. Only the concurrent administration of opioids was significantly higher in the delirium group. Drowsiness, as the predictive value for delirium in corticosteroid-treated patients, may be a simple marker of general cognitive dysfunction from multiple etiologies. This information stresses that careful monitoring is vital when corticosteroid therapy is initiated in patients with somnolence of STAS >1.

Furthermore, the concurrent administration of opioids was identified as an independent predictive factor for delirium in corticosteroid-treated cancer patients. Opioids themselves are a well-recognized risk factor for cognitive dysfunction.⁷ The metabolic drug interactions that occur between opioids and corticosteroids currently remain unknown. Previous studies showed that glucocorticoids facilitated the behavioral and dopaminergic effects of morphine in animal models.³³ The influence of the concurrent administration of corticosteroids and opioids on the development of delirium needs to be examined in future studies.

The strength of this study was its consecutive sampling, and enrollment from multiple institutions ensured generalizability. Nevertheless, our study had several limitations. Since this study did not have a control group, the development of delirium may be attributed to factors other than corticosteroids.³⁴ Furthermore, the prevalence of delirium appeared to be lower than that in previous studies, in which it increased from 28% to 88% of terminally ill cancer patients in weeks to days before death in the palliative care unit.^2–4 In our cohort, 16.9% of 207 patients developed delirium, as measured by the CAM, within three days of the administration of corticosteroids. A simple comparison is difficult because of differences in the measurement time frame; the study procedure in which patients with dementia, a well-known risk factor for delirium,⁵ were excluded may have resulted in the underestimation of delirium. Delirium following the administration of corticosteroids was only measured for three days. Corticosteroid-induced delirium may occur anywhere from one day to three months after starting corticosteroids.¹⁶ We consider this an acceptable limitation due to the short survival expected in this study population, and, thus, had planned to minimize patient withdrawal due to death and progression of the underlying disease. In addition, physicians treated patients within usual clinical practices and provided various cotreatments such as nonpharmacological treatments to avoid delirium. However, our primary aim was to identify predictors in usual clinical practices, and, thus, consider this limitation acceptable. In this study, physicians used lower doses of corticosteroids, a median of 2.4 mg per day as a betamethasone equivalent. Furthermore, we could not control the dose adjustment due to the large variance in clinical practice. A clinical trial comparing different dose protocols is needed to clarify the best dose of corticosteroids to minimize the risk of developing delirium associated with corticosteroid therapy. Finally, 95% CIs of predictors were relatively wide decreasing certainty in the results even though the p values are quite small.

In conclusion, delirium may occur in terminally ill cancer patients after the administration of corticosteroids if they have a poor general condition, drowsiness, and are being treated with opioids. These predictive factors may be useful for clinical decision making by physicians to determine whether corticosteroids are administered to a patient with specific symptoms. A larger study will confirm the results of the present study.

Footnotes

Acknowledgment

This research is (partially) supported by the Practical Research for Innovative Cancer Control from Japan Agency for Medical Research and Development, AMED.

Author Disclosure Statement

No competing financial interests exist.

References

Ljubisavljevic

, Kelly

: Risk factors for development of delirium among oncology patients. Gen Hosp Psychiatry, 2003; 25:345–352.

Lawlor

, Gagnon

, Mancini

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

Bruera

, Miller

, McCallion

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

Caraceni

, Nanni

, Maltoni

, et al.: Impact of delirium on the short term prognosis of advanced cancer patients. Italian Multicenter Study Group on Palliative Care. Cancer, 2000; 89:1145–1149.

Ahmed

, Leurent

, Sampson

: Risk factors for incident delirium among older people in acute hospital medical units: A systematic review and metaanalysis. Age Ageing, 2014; 43:326–333.

Moore

, O'Keeffe

: Drug-induced cognitive impairment in the elderly. Drugs Aging, 1999; 15:15–28.

Gaudreau

, Gagnon

, Harel

, et al.: Psychoactive medications and risk of delirium in hospitalized cancer patients. J Clin Oncol, 2005; 23:6712–6718.

Piazza

, Le Moal

: Glucocorticoids as a biological substrate of reward: Physiological and pathophysiological implications. Brain Res Brain Res Rev, 1997; 25:359–372.

Cowen

, Lawrence

: The role of opioid-dopamine interactions in the induction and maintenance of ethanol consumption. Prog Neuropsychopharmacol Biol Psychiatry, 1999; 23:1171–1212.

10.

Moertel

, Schutt

, Reitemeier

, et al.: Corticosteroid therapy of preterminal gastrointestinal cancer. Cancer, 1974; 33:1607–1609.

11.

Bruera

, Roca

, Cedaro

, et al.: Action of oral methylprednisolone in terminal cancer patients: A prospective randomized double-blind study. Cancer Treat Rep, 1985; 69:751–754.

12.

Hanks

, Trueman

, Twycross

: Corticosteroids in terminal cancer: A prospective analysis of current practice. Postgrad Med J, 1983; 59:702–706.

13.

Matsuo

, Morita

, Iwase

: Physician-reported corticosteroid therapy practices in certified palliative care units in Japan: A nationwide survey. J Palliat Med, 2012; 15:1011–1016.

14.

Denton

, Shaw

: Corticosteroid prescribing in palliative care settings: A retrospective analysis in New Zealand. BMC Palliat Care, 2014; 13:7.

15.

Warrington

, Bostwick

: Psychiatric adverse effects of corticosteroids. Mayo Clin Proc, 2006; 81:1361–1367.

16.

Stiefel

, Breitbart

, Holland

: Corticosteroids in cancer: Neuropsychiatric complications. Cancer Investiga, 1989; 7:479–491.

17.

Currow

, Plummer

, Frith

, et al.: Can we predict which patients with refractory dyspnea will respond to opioids?. J Palliat Med, 2007; 10:1031–1036.

18.

Currow

, Rowett

, Doogue

, et al.: An international initiative to create a collaborative for pharmacovigilance in hospice and palliative care clinical practice. J Palliat Med, 2012; 15:282–286.

19.

Matsuo

, Morita

, Matsuda

, et al.: Predictors of responses to corticosteroids for cancer-related fatigue in advanced cancer patients: A multicenter prospective observational study. J Pain Symptom Manage, 2016; 52:64–72.

20.

Lundström

, Fürst

: The use of corticosteroids in Swedish palliative care. Acta Oncol, 2006; 45:430–437.

21.

Hardy

, Rees

, Ling

, et al.: A prospective survey of the use of dexamethasone on a palliative care unit. Palliat Med, 2001; 15:3–8.

22.

Mercadante

, Fulfaro

, Casuccio

: The use of corticosteroids in home palliative care. Support Care Cancer, 2001; 9:386–389.

23.

Radbruch

, Strasser

, Elsner

, et al.: Fatigue in palliative care patients—an EAPC approach. Palliat Med, 2008; 22:13–32.

24.

Inouye

, van Dyck

, Alessi

, et al.: Clarifying confusion: The confusion assessment method. A new method for detection of delirium. Ann Intern Med, 1990; 113:941–948.

25.

Breitbart

, Rosenfeld

, Roth

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

26.

Lawlor

, Nekolaichuk

, Gagnon

, et al.: 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.

27.

Matsuoka

, Miyake

, Arakaki

, et al.: Clinical utility and validation of the Japanese version of Memorial Delirium Assessment Scale in a psychogeriatric inpatient setting. Gen Hosp Psychiatry, 2001; 23:36–40.

28.

Oken

, Creech

, Tormey

, et al.: Toxicity and response criteria of the eastern cooperative oncology group. Am J Clin Oncol, 1982; 5:649–655.

29.

Morita

, Tsunoda

, Inoue

, et al.: Validity of the Palliative Performance Scale from a survival perspective. J Pain Symptom Manage, 1999; 18:2–3.

30.

Miyashita

, Matoba

, Sasahara

, et al.: Reliability and validity of the Japanese version of the Support Team Assessment Schedule (STAS-J). Palliat Support Care, 2004; 2:379–385.

31.

Inouye

, Viscoli

, Horwitz

, et al.: A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Ann Intern Med, 1993; 119:474–481.

32.

Inouye

, Charpentier

: Precipitating factors for delirium in hospitalized elderly persons. Predictive model and interrelationship with baseline vulnerability. J Am Med Assoc, 1996; 275:852–857.

33.

Marinelli

, Aouizerate

, Barrot

, et al.: Dopamine-dependent responses to morphine depend on glucocorticoid receptors. Proc Natl Acad Sci U S A. 1998; 95:7742–7747.

34.

Jenkins

, Bruera

: Difficulties in diagnosing neuropsychiatric complications of corticosteroids in advanced cancer patients: Two case reports. J Pain Symptom Manage, 2000; 19:309–317.