Physician-Reported Corticosteroid Therapy Practices in Certified Palliative Care Units in Japan: A Nationwide Survey

Introduction

Corticosteroids are commonly used for symptomatic relief in the treatment of patients with advanced cancer; the earliest use of corticosteroids is reported in the 1970s.^1–3 Only a few controlled trials^1,2 of corticosteroid therapy in a palliative care setting have been performed. Corticosteroid use is mostly based on the experience of practitioners rather than on evidence collected in such trials, and the optimal protocol for corticosteroid use remains uncertain.

Long-term corticosteroid use without close monitoring may be harmful because of the high incidence of adverse effects.^3,6 However, long-term corticosteroid use has been reported.^4,7

Corticosteroid withdrawal has different effects such as fever and anorexia, ⁸ and may exacerbate terminal restlessness. ⁹ However, a universally accepted protocol for corticosteroid administration, especially for patients in the terminal stage, remains to be established. Abrupt corticosteroid withdrawal because of loss of the oral route of administration has frequently been reported.^10,11

Thus the clinical practices and attitudes about corticosteroid use are likely to vary widely among physicians. This discrepancy in corticosteroid use may cause unnecessary distress in terminal patients. To our knowledge, very little information^4,7 is available about physician-reported practices and attitudes toward corticosteroid therapy in palliative care settings nationwide. Clarification of current corticosteroid therapy practices is the first step toward determining the important areas for discussion and further research. This nationwide survey aimed to clarify (1) the administration method, (2) withdrawal when death is imminent, and (3) management of corticosteroid side effects, by surveying Japanese palliative care physicians about their strategies for relieving anorexia, fatigue, and dyspnea in terminal cancer patients.

Materials and Methods

This was a cross-sectional anonymous mailed survey of representative managing physicians in 178 certified palliative care units in Japan. The survey was divided into two parts. The first part was designed to evaluate the physicians' experiences with treatment effects and side effects as described in a separate article, ¹² and the second part included physician-reported corticosteroid therapy in clinical practice. This report describes the second part of the survey. The questionnaires were sent in November 2008, along with a cover letter explaining the purpose of the study. The respondents were requested to report their clinical practices and perceptions of corticosteroid treatment in terminally ill patients. In this study, terminally ill cancer patients were defined as those who were not on any anticancer treatment, with an estimated survival of 6 months or less. We sent a reminder 1 month after the initial mailing. We chose these physicians as our study population, because in Japan most palliative care units have a single or only a few such physicians; thus the physicians surveyed were representative of physicians involved in direct patient care.

Results

In all, 124 physicians returned the questionnaires (response rate, 70%). The characteristics of the participating palliative care units have previously been described in detail. ¹²

Corticosteroids were used at 123 (99%) institutions. Corticosteroids were not used in the remaining 1% of the surveyed institutions because of insufficient efficacy. Among all terminal cancer patients, 2% to 100% (median, 80%) inpatients received corticosteroids.

Discussion

To our knowledge, this is the first nationwide representative survey for the systematic investigation of clinical practices involving corticosteroid therapy for terminal cancer patients. The important findings of this study are as follows: (1) corticosteroids were prescribed in almost all institutions, although marked differences were observed in physician-reported percentages of prescription among palliative care specialists; (2) major variations were observed in physician practices about corticosteroid administration methods, especially corticosteroid withdrawal when the death of the patient was imminent; and (3) approaches for the management of side effects varied among the responding physicians.

Corticosteroids were extensively prescribed in Japanese palliative care units. The median rate of corticosteroid prescription in a palliative care setting in Japan is 80%, which is higher than the rates in other countries, 33% to 70%.^3,4,7,10 We did not determine the reasons for the higher prescription rates of corticosteroids seen in Japan. A Swedish survey indicates that steroid therapy has become a routine practice, ⁴ and 95% of the institutions in Japan have no guidelines for corticosteroid therapy. ¹²

However, physician-reported percentages of corticosteroid prescriptions for anorexia, fatigue, and dyspnea varied among respondents. These major differences in the clinical use of corticosteroids may cause serious problems. Overdoses of corticosteroids can cause unnecessary adverse effects leading to poor quality of life, particularly in the case of terminal patients. We believe that the lack of research on the efficacy of corticosteroids for each of the aforementioned symptoms is one of the main causes of the discrepancy seen in the empirical use of corticosteroids. Furthermore, to reduce the inconsistencies seen in corticosteroid prescription practices, well-designed clinical studies are urgently needed.

Unlike Sweden, ⁴ where dexamethasone is the drug of choice,^9,20 betamethasone was the most commonly prescribed corticosteroid for all indications in Japan. However, the two drugs have many similarities. The preference for betamethasone in Japan is based on tradition and a doctor's palliative care manual that is popular among Japanese palliative care physicians. Because the percentages of patients receiving chemotherapy were high at the institutions where physicians prescribed prednisolone, the physicians' experiences and strategies may have had an influence on the choice of drugs used.

In addition, our study showed that physician decisions about dose adjustment protocols, the timing of administration, and dosing differ. Twycross ¹³ recommended starting with a relatively high dose to avoid missing a possible treatment effect, and then reducing the dose to a low maintenance level. However, dose adjustment procedures varied among physicians, although most respondents reported a gradual escalation in the dose.

The most common indication for initiating corticosteroid therapy for anorexia and fatigue was a predicted survival of less than 3 months. Although a simple comparison is difficult because of differences in study design, the administration period in our study appears to be longer than those reported in previous studies (26 days ¹⁴ and 4–8 weeks ³ ), and those recommended by experts ¹⁵ (1–2 weeks). Japanese clinical tradition and lack of availability of alternative drugs such as megestrol acetate for anorexia and methylphenidate and modafinil for fatigue may be responsible for the differences in administration periods. Marked variability exists in initiating corticosteroid administration for dyspnea; half of the respondents answered that they prescribed corticosteroids regardless of the prognosis.

Dose selection is an important unresolved issue. In controlled trials for anorexia treatment, the dose varied between 2.1 and 5.7 mg of betamethasone equivalent per day.^1,2,5 In our survey, the physician-reported ranges of minimum and maximum dosing is from <2 to 2–4 mg, and 2–4 mg to >4 mg, respectively. Although few detailed reports are available on the optimal doses of corticosteroids for each symptom, 4 mg of dexamethasone seems to be used most frequently for anorexia and weakness. The doses used in this study were consistent with those used in previous studies. A formal study comparing dose adjustment protocols, the timing of administration, and optimal doses, is needed to clarify the best clinical practice.

Corticosteroid withdrawal in terminal patients in palliative care poses a new set of problems. The effects of corticosteroids diminish with time, either because of loss of drug efficacy, disease progression, or both.^1,2 Long-term corticosteroid use suppresses the hypothalamic-pituitary-adrenal axis for prolonged periods; some experts advocate that doses be tapered over several days or weeks according to circumstances, ¹³ because abrupt corticosteroid reduction might exacerbate terminal restlessness,^9,13 and acute withdrawal psychosis. ¹⁶ When oral corticosteroid administration became difficult because of dysphagia due to progression of the underlying disease, the rates of corticosteroid cessation were relatively low; however, a variation in practices for managing anorexia was observed. This is inconsistent with the results obtained in a retrospective study conducted in a different country, in which only 2% of the patients were switched to parenteral corticosteroids when oral administration became difficult. ¹⁰ However, it is noteworthy that physicians reported a wide variation in practices concerning corticosteroid withdrawal if death was expected within 1 week. The strategies adopted by the respondents to avoid hyperactive delirium included abrupt cessation (46%), reducing but not discontinuing corticosteroids (33%), and neither (21%). Corticosteroids appear to be one of the factors aggravating hyperactive delirium, ¹⁷ and thus cessation or reduction of corticosteroid around the time of death may be a common trend among Japanese palliative care physicians. A well-designed retrospective cohort study is urgently needed to address this issue.

Several studies^18,19 have reported the side effects of corticosteroid therapy for terminal cancer patients, but little evidence is available about prevention and management of such effects. ¹⁰ According to our survey, clinical practices for the management of hyperactive delirium varied widely among physicians. This finding may reflect the aforementioned attitude favoring corticosteroid withdrawal in the final stages of life. Although randomized controlled trials designed to determine the efficacy of preventive measures are very difficult to complete in such a patient population, a large-scale retrospective cohort study of patients with and without severe complications might allow a conclusions to be drawn about the roles of preventive medications.^4,7,10

The major limitation of this study is that the responding physicians might not have reported all data, and there could be significant observer bias. In addition, the data may have been influenced by cultural differences. Thus, generalized recommendations cannot be drawn from these findings.

In conclusion, corticosteroids are frequently prescribed for terminally ill cancer patients in Japan. Physician-reported practices about corticosteroid therapy varied among palliative care specialists. Considerable variations were identified in areas such as prescription rate, dose adjustment protocols, the timing of administration, doses, prevention of adverse effects, and in particular, withdrawal of corticosteroid therapy when patient death is imminent. Future studies should focus on these diverse areas, and clear clinical guidelines are required to establish consistent corticosteroid therapy practices in the field of palliative care.