Efficacy and Undesirable Effects of Corticosteroid Therapy Experienced by Palliative Care Specialists in Japan: A Nationwide Survey

Abstract

Background and Methods:

Corticosteroids are commonly used for symptom relief in the treatment of patients with advanced cancer. Consistent efficacy of corticosteroid treatment in palliative care remains controversial. A cross-sectional anonymous survey was mailed to representative managing physicians in certified palliative care units in Japan to clarify the physician-perceived efficacy of steroid treatment on anorexia, fatigue, and dyspnea in terminal cancer patients, to clarify physicians' experience of side effects of corticosteroid use, and to determine the Japanese palliative care physician-reported predictive factors for efficacy and lack of efficacy.

Results and Conclusions:

Many Japanese palliative care specialists perceived that corticosteroids are effective for each of the symptoms, are aware of the prevalence and importance of serious adverse effects, and predict the effectiveness of steroid therapy by etiological factors.

Introduction

Corticosteroids are commonly used for symptom relief in the treatment of patients with advanced cancer, with their initial use reported in the 1970s.^1–4 Indications for corticosteroid use in cancer patients differ according to the broad range of effects of this class of drugs. Specific indications include spinal cord compression, lymphatic carcinomatosis, and raised intracranial pressure, and nonspecific indications include pain, nausea, anorexia, fatigue, a general feeling of being unwell, and malaise. Treatment with corticosteroids is also sometimes associated with potentially serious side effects. Therefore, close monitoring of the patient using the minimum effective dose is recommended, with cessation if no benefit is obtained.

Consistent efficacy of corticosteroid treatment in palliative care remains controversial, with only limited evidence of efficacy from uncontrolled studies.^5,6 There are also only scant data^5,7 on physician-perceived effectiveness as well as side effects of corticosteroid therapy in palliative care settings nationwide. Assessment of treatment effectiveness in terminally ill cancer patients with severe symptoms and a short life expectancy is indeed difficult; thus, analysis of the clinical practice experience and perceptions of palliative care experts may be useful. The aims of this nationwide survey were to (1) clarify the physician-perceived efficacy of steroid treatment on anorexia, fatigue, and dyspnea in terminal cancer patients, (2) clarify physicians' experience of side effects of corticosteroid use, and (3) determine the Japanese palliative care physician-reported predictive factors for efficacy and lack of efficacy.

Materials and Methods

This was a cross-sectional, anonymous mailed survey of representative managing physicians in all 178 certified palliative care units in Japan. The questionnaires were sent out in November 2008, along with a cover letter explaining the purpose of the study. The respondents were requested to report their clinical practice experience and perceptions regarding corticosteroid treatment in terminally ill patients. We sent a reminder 1 month after the initial mailing. We chose this population as our study population because most palliative care units in Japan have a single or only a few physicians and representative physicians are usually involved in direct patient care.

Questionnaire

The authors developed a 15-item questionnaire (see Appendix) on the basis of a literature review and discussions with 3 palliative care physicians.^5,7,8

The physicians were first asked to report their rate of agreement regarding corticosteroid therapy using a non-Likert-type scale and to report on their actual clinical experiences regarding corticosteroid treatment of adult cancer patients with an estimated survival of 6 months or less who were not on anticancer treatment. The questions focused on corticosteroid therapy for the specific symptoms of anorexia, fatigue, and dyspnea, which were the most common indications for steroid therapy reported in the literature, because we speculated that the physicians' attitudes might differ for each of these symptoms.^8,9 The questionnaire was divided into four sections.

The first section related to the physician-perceived efficacy. The respondents were asked to report the physician-perceived effective rate and timing and to evaluate the treatment effect. The physician-perceived effective rate was defined as the proportion of patients showing a positive effect that lasted at least a week among all patients treated for the particular indication. The evaluation of the treatment effect was not based on the patient's records, but on the clinical impression of the respondents.

The second section related to the side effects of corticosteroid therapy. The physicians were asked to report the physician-perceived rate of patients who experienced 10 major side effects within 1 week and over 1 month, and was defined as the percentage of patients who experienced each side effect among all the patients who had received corticosteroids. The evaluation of the side effects was based on the clinical impression of the respondents. Respondents were then requested to report their clinical experience of serious side effects. The definition of each side effect was based on the clinical diagnosis, and we had not provided any operational definitions, which are not clearly established.

In the third section, the participants reported the predictive factor(s) for efficacy or lack of efficacy, on an empirical basis, in relation to each symptom (e.g., the impression that corticosteroid therapy is effective for fatigue associated with hepatic failure). These questions were in a multiple-choice format, with additional open-answer questions.

Next, the respondents were asked to report the frequency of corticosteroid use. The frequency of corticosteroid use was defined as the percentage of patients receiving corticosteroid therapy among the inpatients.

Background data were obtained, including the characteristics of the participating institutions (average patient age, median death rate, median percentage of patients receiving chemotherapy, median length of hospitalization, availability of palliative care outpatient services, and availability of local guidelines).

Statistical analyses

All analyses were performed using the Statistical Package for the Social Sciences (version 12.0; SPSS Japan Inc., Tokyo, Japan).

Results

A total of 124 physicians returned the questionnaires (response rate 70%) by January 2009. The characteristics of the participating palliative care units are detailed in Table 1.

Table 1.

Characteristics of the Participating Institutions

Variable
Average age of inpatients (years) (n = 116)	70 [4.3]
Median death rate (n = 120)^*	90% (4.2–100)
Median percentage of patients receiving chemotherapy (n = 115)^*	0% (0–90)
Median duration from admission to death (days) (n = 119)^*	35 (2–87)
Availability of palliative care outpatient services	90% (111/124)
Availability of local guidelines on treatment with corticosteroids in terminal cancer	5.0% (6/121)

Range in brackets.

Prevalence of corticosteroid prescription

Corticosteroids were used at 123 institutions (99%). The reported reason for the lack of use of corticosteroids at 1% of the surveyed institutions was insufficient efficacy. The percentage of patients who received corticosteroids among all terminally ill cancer inpatients ranged from 2% to 100% (average 71%).

Physician-reported efficacy

The highest physician-perceived effective rate as perceived by the responding physicians was for anorexia (57%), whereas more than 97% reported a positive effect within 7 days for each of the three symptoms (Table 2).

Table 2.

Responses Regarding Efficacy of Corticosteroid Use (n = 120)

	Anorexia	Fatigue	Dyspnea
Average physician-perceived effective rate^*	57 [21]%	50 [21]%	52 [22]%
Positive effect came
Within 24 hours	7.5% (n = 9)	10% (n = 12)	37% (n = 44)
1–2 days	44% (n = 53)	34% (n = 41)	38% (n = 45)
3–7 days	48% (n = 58)	53% (n = 63)	24% (n = 29)
8–14 days	0%	2.5% (n = 3)	0.8% (n = 1)

Some percentages do not add up to 100% due to missing values.

The empirical effective rate was defined as the proportion of patients showing a positive effect that lasted for at least 1 week among all patients treated for each indication.

Side effects

The physician-perceived rates (%) of 10 common side effects as reported by the physicians are indicated in Table 3. The side effects seen in more than 10% of patients within 1 week of the start of treatment were insomnia and hyperglycemia, whereas those seen in more than 20% of patients after over 1 month of the start of treatment were oral candidiasis, moon face, and hyperglycemia.

Table 3.

Physician-perceived Rates of 10 Common Side Effects Related to Corticosteroid Therapy^* ( n = 105)

Side effect	Median rate within 1 week after initial administration (%) ^†	Median rate over 1 month after initial administration (%) ^†
Aggravated/triggered diabetes mellitus	10 (0–100)	20 (0–100)
Insomnia	10 (0–80)	10 (0–80)
Delirium	5 (0–50)	10 (0–80)
Oral candidiasis	1 (0–60)	30 (0–100)
Moon face	0 (0–20)	30 (0–100)
Depression	0 (0–30)	10 (0–50)
Electrolyte abnormalities	0 (0–20)	10 (0–40)
Peptic ulcer	0 (0–20)	7.5 (0–30)
Osteoporosis	0 (0–10)	5 (0–90)
Myopathy	0 (0–14)	5 (0–70)

The physician-perceived side-effect rates were defined as the percentage of patients with the side effect among patients who received corticosteroids, based on the respondents' clinical impressions.

†

Range in brackets.

Experiences with serious side effects were reported by 23% (n = 28) of the respondents, as follows: gastrointestinal bleeding (n = 7, 5.8%), gastrointestinal perforation (n = 6, 5.0%), potentially fatal hyperglycemia (n = 10, 8.3%), Pneumocystis jiroveci pneumonia (n = 3, 2.5%), potentially fatal bacterial pneumonia (n = 2, 1.7%), pulmonary tuberculosis (n = 2, 1.7%), sepsis (n = 1, 0.83%), serious delirium (n = 3, 2.5%), suicidal tendency (n = 1, 0.83%), severe myopathy (n = 2, 1.7%), and compression fracture (n = 1, 0.83%).

Physician-reported predictive factors of effectiveness

The physician-reported predictive factors, on an empirical basis, for efficacy and lack of efficacy for each symptom are listed in Table 4. The predictive factors for efficacy against anorexia identified by more than 50% of the respondents were lung cancer and digestive cancer. For fatigue, tumor fever, digestive cancer, and lung cancer were identified as the predictive factors of efficacy, whereas liver failure, renal failure, and cachexia were identified as predictive factors for lack of efficacy. In relation to dyspnea, lymphantic carcinomatosis, airway obstruction, and multiple lung metastases were identified as predictive factors for steroid efficacy, whereas the predictive factor identified for lack of efficacy by 53% of the respondents was pleural effusion.

Table 4.

Physician-reported Predictive Factors for the Empirical Effectiveness of Corticosteroid Treatment of the Three Symptoms

	Effective predictive factor (n = 111)		Ineffective predictive factor (n = 92)
Anorexia
	Lung cancer	64% (n = 71)	Cachexia	26% (n = 24)
	Digestive cancer	58% (n = 64)
	Cachexia	32% (n = 35)
	<70 years old	28% (n = 31)
Fatigue
	Tumor fever related	69% (n = 75)	Liver failure	37% (n = 34)
	Digestive cancer	53% (n = 58)	Renal failure	36% (n = 33)
	Lung cancer	52% (n = 57)	Cachexia	36% (n = 33)
	Chemotherapy induced	30% (n = 33)	Anemia	30% (n = 28)
	Liver failure	29% (n = 30)	Depression	27% (n = 25)
Dyspnea
	Lymphangitic carcinomatosis	83% (n = 91)	Pleural effusion	53% (n = 47)
	Airway obstruction	67% (n = 74)	Bronchial secretion	47% (n = 42)
	Multiple lung metastases	66% (n = 73)	Anemia	42% (n = 37)
	Weakness	32% (n = 35)	Pneumonia	40% (n = 36)
	Bronchial secretion	27% (n = 30)	Weakness	39% (n = 35)

The questions were multiple-choice models and open answers.

This table represents the opinions reported by more than 25% of respondents.

Discussion

To the best of our knowledge, this is the first representative nationwide survey to systematically investigate physicians' experience of corticosteroid therapy for terminally ill cancer patients. The important findings of this study were that many Japanese palliative care specialists (1) perceived that corticosteroids are effective for each of the symptoms, (2) are aware of the prevalence and importance of serious adverse effects, and (3) predict the effectiveness of steroid therapy by etiological factors.

In our study, more than 50% of the patients were assessed as showing good response to steroid treatment administered for different indications. Although a simple comparison is difficult because of differences in the study design, the results of our study are consistent with earlier findings.^1,4,10,11 Thus, corticosteroid therapy may be found to be generally effective for these indications in many terminally ill cancer patients.

Many respondents stated that a positive response was noted within 1 week in cases in which the treatment was effective, consistent with previous studies.^3,10,12 This finding may indicate that the efficacy of corticosteroid therapy should be evaluated within 1 week after it is initiated. This information is very useful in terms of the cost-effectiveness, economic resources, and prevention of unnecessary adverse effects.¹²

Large variations exist in physicians' estimation of side effects related to treatment with corticosteroids. As compared with previous reports,^7,13 the actual percentages of the side effects in our study were relatively low. In contrast, this study revealed that 23% of the respondents encountered serious side effects, and the side effects experienced by palliative care specialists were similar to those reported previously: gastrointestinal bleeding,¹² severe myopathy,^5,14 severe osteoporosis,⁵ severe infection,⁵ and severe neuropsychiatric complications.^15,16 To obtain the greatest benefit with the fewest adverse effects, a large prospective cohort study to clarify the frequency of side effects is strongly needed. However, it cannot be overemphasized that clinicians should carefully monitor and manage treatment-related complications on an individual basis to avoid side effects.

Although the present study identified some predictive factors that many respondents regarded as useful, opinions varied among physicians regarding other predictive factors. Predicting the effectiveness of steroid therapy would allow avoidance of unnecessary use in patients. A future prospective study is needed for investigation of each predictive factor; this would contribute to using steroid therapy to obtain maximum benefits with minimum adverse effects.

One of the limitations of this study is that it was based on reports of physicians rather than on assessments by patients. Because of this limitation, we need to be cautious in making conclusions on the basis of the findings of this study.

In conclusion, corticosteroids are frequently administered to terminally ill cancer patients in Japan. Of special interest among the findings of this study is that physicians often predict whether corticosteroids would be effective based on etiological factors. A future study to validate the identified predictive factors might allow avoidance of potential over-prescription of steroid therapy. Because palliative care specialists encounter serious side effects in clinical settings, a future controlled trial to determine the efficacy of steroid therapy for each symptom of homogeneous etiology, as well as a large cohort study to clarify the prevalence of serious side effects is strongly needed to obtain optimal effect of corticosteroid therapy and minimize the risk of development of side effects in terminal cancer patients.

Footnotes

Acknowledgments

This study was supported by the Japan Palliative Care Audit Network (JPCAN). We would like to thank Yujiro Kuroda, Clinical Psychologist, Department of Palliative Medicine, The University of Tokyo Hospital, for the data collection.

Author Disclosure Statement

No competing financial interests exist.

A ppendix . Q uestionnaire Sheet

	Effective predictive factors	Ineffective predictive factors
Anorexia	Digestive cancer, Lung cancer, Chemotherapy induced, Hepatomegaly, Hypercalcemia, Depression, Cachexia, Over 70 years old, Less than 70 years old, Male, Female, Predicted survival, Others (please state)	Digestive cancer, Lung cancer, Chemotherapy induced, Hepatomegaly, Hypercalcemia, Depression, Cachexia, Over 70 years old, Less than 70 years old, Male, Female, Predicted survival, Others (please state)
Fatigue	Digestive cancer, Lung cancer, Chemotherapy induced, Liver failure, Renal failure, Depression, Tumor fever, Elevated C-reactive protein, Anemia, Cachexia, Over 70 years old, Less than 70 years old, Male, Female, Predicted survival, Others (please state)	Digestive cancer, Lung cancer, Chemotherapy induced, Liver failure, Renal failure, Depression, Tumor fever, Elevated C-reactive protein, Anemia, Cachexia, Over 70 years old, Less than 70 years old, Male, Female, Predicted survival, Others (please state)
Dyspnea	Multiple lung metastasis, Air way obstruction, Pleural effusion, Lymphangitic carcinomatosis, General weakness, Bronchial secretion, Pneumonia, Anemia, Others (please state)	Multiple lung metastasis, Air way obstruction, Pleural effusion, Lymphangitic carcinomatosis, General weakness, Bronchial secretion, Pneumonia, Anemia, Others (please state)

References

Moertel

, Schutt

, Reitemeier

, Hahn

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

Della Cuna

, Pellegrini

, Piazzi

. Effect of methylprednisolone sodium succinate on quality of life in preterminal cancer patients: A placebo-controlled, multicenter study. Eur J Cancer Clin Oncol, 1989; 25:1817–1821.

Bruera

, Roca

, Cedaro

, Carraro

, Chacon

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

Hanks

, Trueman

, Twycross

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

Needham

, Daley

, Lennard

. Steroids in advanced cancer: Survey of current practice. BMJ, 1992; 305:999.

Twycross

. The risks and benefits of corticosteroids in advanced cancer. Drug Safety, 1994; 11:163–178.

Lundström

, Fürst

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

Gannon

, McNamara

. A retrospective observation of corticosteroid use at the end of life in a hospice. J Pain Symptom Manage, 2002; 24:328–334.

Twycross

. Corticosteroids in advanced cancer. BMJ, 1992; 305:969–970.

10.

Radbruch

, Strasser

, Elsner

, Gonçalves

, Løge

, Kaasa

, Nauck

, Stone

. Research Steering Committee of the European Association for Palliative Care (EAPC): Fatigue in palliative care patients—an EAPC approach. Palliat Med, 2008; 22:13–32.

11.

Wooldridge

, Anderson

, Perry

. Corticosteroids in advanced cancer. Oncology (Williston Park), 2001; 15:225–234.

12.

Mercadante

, Fulfaro

, Casuccio

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

13.

Hardy

, Rees

, Ling

, Burman

, Feuer

, Broadley

, Stone

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

14.

Batchelor

, Taylor

, Thaler

, Posner

, DeAngelis

. Steroid myopathy in cancer patients. Neurology, 1997; 48:1234–1238.

15.

Stiefel

, Breitbart

, Holland

. Corticosteroids in cancer: neuropsychiatric complications. Cancer Invest, 1989; 7:479–491.

16.

Jenkins

, Bruera

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