Corticosteroid-Induced Diabetes in Palliative Care

Abstract

Background:

Corticosteroids are one of the most commonly used medications in palliative care. Although the benefit of corticosteroids generally outweighs the risk in the palliative population, side effects are common and necessitate careful consideration prior to prescribing. In March of 2010, a guideline for monitoring blood glucose values was implemented as part of our standard care within our two inpatient tertiary palliative care units.

Method:

A retrospective study was conducted, the aim of which was twofold. First, we hoped to determine a prevalence rate for steroid-induced diabetes mellitus (SDM) in palliative care and whether or not screening glucose levels twice weekly was appropriate or required. Second, we wanted to determine if possible predictors existed for the development of SDM in a palliative population, thereby identifying the patients most at risk who would benefit from ongoing glucose monitoring.

Results and discussion:

We found that SDM is more common in palliative care patients than previously thought. Our study showed a higher likelihood of developing hyperglycaemia with higher doses of dexamethasone. But although dose is correlated with hyperglycemia, patients without high doses were also at risk. Further study is currently underway with slight modifications to the guideline to more accurately assess the physical burden, as well as the emotional and financial cost of a hyperglycemia screening protocol.

Introduction

Corticosteroids are one of the most commonly used medications in palliative care.¹ Some of the indications for use include decreasing peritumoral edema associated with brain tumors,² bowel obstructions,^3,4 spinal cord compressions,⁵ superior vena cava syndrome,⁶ and ureteric obstruction. They may also be used in symptomatic treatment of nausea and dyspnea, anorexia, weight loss, fatigue,⁷ and improving a patient's general feeling of well-being.

Although the benefit of corticosteroids generally outweighs the risk in the palliative population, side effects are common^8,9 and necessitate careful consideration prior to prescribing. Metabolic-related adverse effects of steroids, such as impaired glucose metabolism, can have a negative impact on mortality and morbidity for patients near the end of their lives.¹⁰ These adverse effects may be particularly problematic in palliative patients, who may be already significantly compromised in their ability to maintain metabolic and nutritional homeostasis.¹¹

Review of the literature suggests that 30% to 60% of patients in palliative care settings internationally receive corticosteroids as part of their treatment.^8,9,12,13 Given the widespread use of corticosteroids in this population, the potential for adverse effects in palliative care patients secondary to corticosteroid use is considerable.

To date, there is very little literature on the prevalence of steroid-induced diabetes mellitus (SDM) in the palliative care population. Neurologic, renal, respiratory, and rheumatology patient populations have had more robust monitoring for SDM and prevalence rates in these populations have been found to be 52%, 40%, 14%, and 8%, respectively.^14,15,16,17 Mercadante and colleagues in their study of corticosteroid use in palliative care patients receiving home care estimated the prevalence to be around 3% based on patient reporting of symptoms.⁸ However, to our knowledge, there has been no literature on the prevalence of SDM in the palliative care population using routine screening of capillary blood glucose values.

Although there is a paucity of research in this area, a study by Quinn and coworkers conducted by end-of-life health care providers and endocrinologists recommended regular monitoring of blood glucose if corticosteroids were used; however, there was a lack of consensus on the frequency or most effective monitoring system.¹⁰ For those diagnosed with hyperglycemia as a result of treatment, the literature recommended that blood glucose levels be maintained in the range between 10 and 20 mmol/L to allow most patients to experience only mild symptoms of hyperglycemia, while avoiding the dangers of hypoglycemia.^10,11

Methods

In March 2010, a guideline for monitoring blood glucose values was implemented as part of our standard care within our two inpatient tertiary palliative care units. The guideline established a minimum of twice weekly screening of blood glucose levels for patients receiving corticosteroid treatment, with the target blood glucose range of between 10 and 20 mmol/L, as suggested by Quinn et al.¹⁰ Glucose monitoring and the treatment of hyperglycemia was only instituted if it was felt to be clinically appropriate and consistent with the patient/family's goals of care. The guideline did not recommend specific types of hypoglycemic agents to be used to treat hyperglycemia but rather left treatment decisions to the prescribing health care team (see Appendix 1).

Two months after the implementation of the guideline, a retrospective study was conducted to evaluate the change in practice. The aim of the study was twofold. First, we hoped to determine a prevalence rate for SDM in palliative care and whether or not screening glucose levels twice weekly was appropriate or required. Second, we wanted to determine if possible predictors existed for the development of SDM in a palliative population, thereby identifying the patients most at risk who would benefit from ongoing glucose monitoring. Previous literature, in other patient populations, has suggested the odds ratio for developing SDM is between 1.5 and 2.5, with dose, duration, and body mass index (BMI) being strong predictors of diabetes induction.^18,19,20,21 We were interested to know if risk factors could be determined for palliative care patients, including dose, duration, primary prescribing indication, and underlying diagnoses. Data were not collected to determine if BMI was a predictor, as this concept seemed less useful in palliative patients, where metabolic and nutritional homeostasis is difficult to maintain, cachexia often plays a role, and measurements of height and weight are not routinely performed. It was thought that if predictors exist for the development of SDM in palliative patients, our care could be modified to include glucose monitoring for only patients known to be at the highest risk of developing hyperglycemia, thereby reducing the monitoring burden for patients at low risk.

Possible participants for the study were identified by the pharmacy records or the electronic patient record. Patients were included in the study if they were receiving treatment with systemic corticosteroids, over the age of 18, and inpatients on the palliative care units. Patients were excluded if they were known to have preexisting diabetes. Ethical approval was obtained prior to implementing the study. Participant consent was not required as the study was primarily retrospective in design and involved analyzing data that were already obtained as part of our standard of care. Data collection included: glucose values, admission blood work values, diagnoses, the reason corticosteroids were started, evidence of organ failure, infection, medications, signs or symptoms of hyper- or hypoglycemia, treatments of hyper- or hypoglycemia, the Palliative Performance Scale Version 2 (PPSv2).²² See Appendix 2.

Diabetes was defined by a fasting morning glucose of >7.0 mmol/L or a random glucose >11.1 mmol/L with symptoms of hyperglycemia, according to 2008 criteria from the Canadian Journal of Diabetes.²³ Many patients in the study did not go on to have fasting morning glucose levels checked, and it was often difficult to determine if symptoms were attributable to hyperglycemia. Therefore, patients with a screening glucose of >11.1 mmol/L without documented symptoms were not diagnosed with diabetes but rather were labelled as having hyperglycemia.

Statistical analyses were performed using SPSS version 14.0 (SPSS Inc., Chicago, IL). Descriptive analyses of the differences were tested with the χ² distribution for categorical variables and with analysis of variance (ANOVA; t test) for comparison of means for continuous variables.

Results

Pharmacy and electronic patient records identified 149 patients prescribed corticosteroids over a 4-month period. These patient charts were screened by the research team, and 15 were excluded because the patients were known to have preexisting diabetes. The remaining 134 participants were included in the study, and data were collected for the duration of their hospitalizations. See Figure 1.

FIG. 1.

Enrollment of patients and their glucose values.

Patient demographics

There were 57 (42%) males and 77 (58%) females. Age ranged from 31 to 101 years (median 71), PPSv2 scores ranged from 10 to 80 (median 40), and length of stay ranged from 2-199 days (median 22). Of the 134 patients included in the study, 91 (68%) died during their hospitalization. The remainder were discharged home or to alternate care facilities, including hospice and long-term care. The most common cancer diagnoses among patients included lung cancer in 45 (34%), colon cancer in 16 (12 %), prostate cancer in 9 (7%), brain cancer and gynecologic cancers at 7 (5%) each, kidney cancer in 6 (4%), and breast cancer in 5 (4%). See Table 1.

Table 1.

Patient Characteristics

Demographic data
Total patients screened	149
Total analyzed (excluding preexisting DM2)	134
St. Boniface Hospital (%)	66 (49)
Riverview Health Centre (%)	68 (51)
Males (%)	57 (42)
Females (%)	77 (58)
Age range	31–101 (median 71)
PPS range	10–80 (median 40)
Length of stay (days)	2–199 (median 22)
Patients who died during admission (%)	91 (68)
Steroid use prior to admission (%)	83 (62)

Primary cancer diagnoses (%)
Lung	45 (34)	Gynecologic	7 (5)	Unknown primary	5 (4)
Colon	16 (12)	Kidney	6 (4)	Pancreatic	4 (3)
Prostate	9 (7)	Breast	5 (4)	Other	25 (18)
Brain	7 (5)	Hematologic	5 (4)

Prescribing indication (%)
Pain	33 (25)	Hematologic malignancy	4 (3)	Other	8 (6)
Brain tumor	25 (19)	Spinal cord compression	3 (2)	Unknown	32 (24)
Dyspnea	19 (14)	Nausea/vomiting	2 (1)
Small bowel obstruction	6 (5)	Superior vena cava obstruction	2 (1)

DM2, diabetes mellitus type 2; PPS, Palliative Performance Scale.

Five patients treated with corticosteroids were prescribed prednisone, and the rest were prescribed dexamethasone. Because the vast majority of the patients were prescribed dexamethasone, all calculations to determine correlations with dose were done solely on the patients with dexamethasone.

Findings

Forty-two of 134 participants (31%) were hyperglycemic (random glucose >11.1 mmol/L) but lacked documented symptoms or a fasting blood glucose, and 15 (11%) had a diagnosis of SDM, confirmed with a fasting blood glucose >7.0 mmol/L. Of the 15 with the diagnosis, 5 had glucose values >20.0 mmol/L and had documented symptoms clearly related to hyperglycemia. These patients were treated with hypoglycemic agents and had improvement in their symptoms. See Table 4.

All patients prescribed steroids undergo screening on our units unless they are believed to be close to death, as assessed by the health care team Screening was stopped by the health care team on all patient study participants as they approached death (n=24). Only two patients refused to have capillary blood glucose monitoring, and it was not initiated in these patients. No patients or families asked to have the monitoring stopped once it was initiated.

Many patients on corticosteroids had fluctuations in their prescribing doses. Sometimes this represented weaning and other times it represented dose escalation. For the hyperglycemic patients who underwent only weaning without any dose escalations (n=15), 11 (69%) did not have normalization of their blood sugars, although lower glucose values were observed in 3 of these 11.

The likelihood of developing hyperglycemia increased the higher the prescribed dose of dexamethasone. (See Table 2.) There was no obvious correlation for length of time on dexamethasone. (See Figure 2.)

FIG. 2.

Change in glucose values according to dose and duration.

Table 2.

Correlations with Dose of Dexamethasone

Average afternoon blood glucose values versus doses of dexamethasone
Dexamethasone dose (mg/day)	N	Mean afternoon glucose
4 mg	81	8.6
6 mg	18	8.8
8 mg	72	9.0
12 mg	15	9.5
16 mg	37	10.6

Primary cancer diagnosis and a secondary diagnosis of infection did not correlate with the likelihood of developing hyperglycemia when corrected for dose. Patients who developed hyperglycemia, and/or diabetes were no more likely to have a lower PPSv2 score or die during the admission, than their euglycemic counterparts. (See Table 3.) Steroids were prescribed for pain in 40 patients (30%), for primary and secondary brain tumor in 27(20%), and for dyspnea in 21 (17%). Significantly more patients prescribed steroids for brain tumor developed hyperglycemia (p=0.038). However, when corrected for dose, the results at 4- and 8-mg/day doses were no longer significant. At the 16-mg/day dose, the results fell just short of significance (p=0.051).

Table 3.

Patient Variables as Risk Factors for the Development of Hyperglycemia

	ANOVA correlations with mean daily doses of dexamethasone (p values)
Patient variables	4 mg	8 mg	16 mg
Primary cancer diagnosis	1.874	0.704	0.572
	(0.099)	(0.701)	(0.750)
Infection	0.609	0.012	0.562
	(0.440)	(0.913)	(0.458)
Indication for prescribing	1.138	0.226	1.090
	(0.364)	(0.976)	(0.400)
Brain Tumor versus other Indication	0.243	0.017	4.149
	(0.626)	(0.898)	(0.051)
Use prior to admission	0.000	0.201	2.948
	(0.984)	(0.656)	(0.094)
Age	0.455	0.822	0.375
	(0.960)	(0.687)	(0.982)
Length of stay	0.239	0.958	1.073
	(0.999)	(0.587)	(0.543)
Death during admission	2.154	0.561	0.642
	(0.150)	(0.458)	(0.428)
PPSv2	1.227	0.865	1.938
	(0.318)	(0.514)	(0.115)

	Pearson correlations with mean daily doses of dexamethasone (p values)
Patient variables	4 mg	8 mg	16 mg
Primary cancer diagnosis	0.065	0.077	0.001
	(0.684)	(0.609)	(0.995)
Infection	0.126	−0.017	0.311
	(0.440)	(0.913)	(0.054)
Indication for prescribing	0.188	−0.011	0.031
	(0.348)	(0.950)	(0.868)
Use prior to admission	−0.003	−0.670	0.182
	(0.984)	(0.656)	(0.266)
Age	0.126	0.196	0.196
	(0.434)	(0.193)	(0.231)
Sex	−0.142	0.195	0.087
	(0.377)	(0.193)	(0.597)
Length of stay	−0.020	0.134	0.129
	(0.903)	(0.380)	(0.434)
Death during admission	−0.229	−0.112	0.135
	(0.150)	(0.458)	(0.413)
PPSv2	−0.142	0.104	0.043
	(0.396)	(0.505)	(0.796)

ANOVA, analysis of variance; PPSv2, Palliative Performance Scale Version 2.

Table 4.

Characteristics of the Patients with Diabetes

Demographic data
Number of patients with diabetes	15
Males (%)	5 (33)
Females (%)	10 (67)
Age range	51–89 (median 78)
PPS range	10–50 (median 30)
Length of stay (days)	11–129 (median 33)
Patients who died during admission (%)	12 (80)
Steroid use prior to admission (%)	13 (87)
Patients treated with hypoglycaemic medications	5

Primary cancer diagnoses (%)
Lung	6 (40)	Brain	1 (7)
Colon	1 (7)	Pancreatic	1 (7)
Prostate	2 (13)	Other	4 (26)

Secondary diagnoses (%)
Infection	5 (33)

Prescribing indication (%)
Pain	6 (40)	Other	1 (7)
Brain Tumor(primary or secondary)	5 (33)	Unknown	2 (13)
Dyspnea	1 (7)

Hypoglycaemic treatments
Patient 1	Patient 2	Patient 3	Patient 4	Patient 5
Novolin™ 30/70	Metformin	Humulin R	NovoRapid	Humulin R
Humulin™ R	Novolin Toronto	Novolin NPH		Novolin NPH
NovoRapid™

PPS, Palliative Performance Scale.

Discussion

The prevalence of SDM and hyperglycemia in our palliative patients is higher than the previous estimate,⁸ but within the range of other patient populations prescribed corticosteroids.^14–17 It is difficult to ascertain whether our patients had SDM or previously undiagnosed type 2 diabetes. Although both conditions lie on the spectrum of impaired glucose metabolism, previously the distinction has been made on whether the diabetes resolves when the steroids are withdrawn.¹⁸ However, almost all of our patients were still on steroids at the time of their deaths, and so, in most, the distinction could not be made. For this reason, making a distinction between these two closely related diagnoses in palliative care may be a less relevant concept. Palliative patients who exhibit hyperglycemia, for whatever cause, should be more closely monitored and treated when symptomatic, if this approach is consistent with the goals of care.

Our higher prevalence rates are likely due to a difference in the screening protocol when compared with the previous study.⁸ Our protocol did not rely on patient reporting of symptoms but rather on a screening protocol that occurred regardless of the presence or absence of symptoms. This meant hyperglycemia was often picked up earlier.

Monitoring was not undertaken in our patients close to death, so as not to introduce additional burden to their care when they were less likely to benefit from treatment. By including these unmonitored patients in our analysis we may have underestimated the prevalence rates.

The monitoring did not seem to be overly burdensome to most patients. Only two patients refused to have monitoring initiated, and no patients or families asked for the monitoring to be discontinued once it was started. Patients close to death had their monitoring discontinued by the health care team.

When diagnosed, hyperglycemia was often treated without medications and was treated instead by attempting to reduce the steroid dose. Unfortunately, this strategy was not effective in many patients. Once hyperglycemia developed, it did not seem to abate in most patients with a weaning protocol. Steroids induce diabetes by decreasing insulin secretion from the pancreas, redistributing adipose tissue, decreasing skeletal muscle uptake, and promoting gluconeogenesis within the liver.^10,24 These mechanisms, once engaged, may take days to weeks to reset. This may explain why, in patients who may not have a long life expectancy, they remain hyperglycemic despite weaning steroid doses.

We found a correlation with dose of steroid; specifically, patients on high-dose dexamethasone were more likely to develop hyperglycemia. This has been shown previously in the literature and is likely related to a reduced sensitivity to insulin in the liver and skeletal muscle at all dosing levels, as well as a decreased secretory capacity for insulin at high corticosteroid doses.¹⁸

The diagnoses most likely to develop SDM were primary brain tumors and breast and lung cancers that had metastasized to brain. This is most likely due to dose effect, as patients with primary and secondary brain tumors were prescribed higher doses than patients with many other indications. When corrected for dose, there was no significant correlation with primary cancer diagnosis.

There was no obvious correlation, in our data, with duration of corticosteroid therapy. This is in contrast to a recent meta-analysis showing an increased risk with the length of treatment.¹⁸ Our data may not have shown this previously described correlation due to two possibilities. Most of our patients had frequent changes in their doses, reflecting worsening or improving control of symptoms at end of life. If symptoms improved, weaning doses may have confounded the data, negating an otherwise subtle increase in glucose values with length of treatment. This situation may not have been present in other nonpalliative populations that may have been on a more stable dose of corticosteroid over time. In may also have to due with the relatively short time palliative care patients are on corticosteroids (often only for days), compared with other patient populations that may be prescribed corticosteroids for years at a time. It is possible that our duration of treatments, even for patients initiated prior to admission, were too short to notice a correlation with length of treatment.

Although there is a correlation with dose, and a trend for prescribing indication, many other patients also developed hyperglycemia. It was therefore decided that the guideline will continue to be implemented on all patients receiving steroids who are admitted to palliative care regardless of dose and diagnosis.

Conclusion

We have found that SDM is more common in palliative care patients than previously thought. Our study, initially designed after the implementation of a hyperglycemia screening guideline was implemented on our ward, showed a higher likelihood of developing hyperglycemia with higher doses of dexamethasone. We had initially thought that the determination of risk factors for the development of the hyperglycemia may be used to modify our guideline. However, it became clear that, although dose is correlated with hyperglycemia, patients without high doses were also at risk. We therefore decided to continue our screening on all patients prescribed corticosteroids, regardless of dose. The capillary blood glucose screening did not seem to be overly burdensome to our patients, when used appropriately, and this method of screening for steroid-induced hyperglycemia may serve as a prototype for other palliative care programs that wish to implement a screening protocol.

Further study is currently underway with our palliative care patients in the community setting with slight modifications to the guideline. We hope this further study will more accurately assess the physical burden, as well as the emotional and financial cost, of a hyperglycemia screening protocol. Our hope is that by implementing screening for our palliative care patients in the community, we can improve quality of life by decreasing symptoms attributable to hyperglycemia and, ultimately, decrease the need for hospitalization. As the study progresses, it will be interesting to determine if iatrogenic deaths, which had otherwise been attributed to underlying disease progression, may be prevented and what, if any, the impact of this is for patients and families.

Footnotes

Acknowledgments

We would like to thank the editor and reviewers of the Journal of Palliative Medicine. Their thoughtful consideration and comments helped to improve the quality of this work. Thank you to the staff at St. Boniface General Hospital and Riverview Health Centre Palliative Care Units for allowing us time and access to the wards. Thank you to Ms. Rehana Durocher, pharmacist at the Riverview Health Centre, and Ms. Lindsay Filz, pharmacist at St. Boniface Hospital, who assisted with patient identification. Also thanks to Ms. Lori Embelton, Ms. Christine Newell, and Dr. Mike Harlos with the Winnipeg Regional Health Authority Palliative Care Program for advice and support in generating the guideline and the research protocol.

Author Disclaimer Statement

No competing financial interests exist.

Appendix 1.

Glucose Monitoring Guideline

Appendix 2.

Palliative Performance Scale (PPSv2)

References

Nauck

, Ostgathe

, Klaschik

, Bausewein

, Fuchs

, Lindena

, Neuwohner

, Schulenber

, Radbruch

. the Working Group on the Core Documentation for Palliative Care Units in Germany. Drugs in palliative care: Results from a representative survey in Germany. Palliat Med, 2004; 18:100–107.

Kaal

, Vecht

. The management of brain edema in brain tumors. Curr Opinion Oncol, 2004; 16:593–600.

Laval

, Girardier

, Lassauniere

, Leduc

, Haond

, Schaerer

. The use of steroids in the management of inoperable intestinal obstruction in terminal cancer patients: Do they remove the obstruction? Palliat Med, 2000; 14:3–10.

Feuer

, Broadly

. Corticosteroids for the resolution of malignant bowel obstruction in advanced gynaecological, gastrointestinal cancer (Review) The Cochrane Library. John Wiley & Sons, 2009.

Vecht

, Haaxma-Reiche

, van Putten

, de Visser

, Vries

, Twijnstra

. Initial bolus of conventional versus high-dose dexamethasone in metastatic spinal cord compression. Neurology, 1989; 39,9:1255–1257.

Rowell

, Gleeson

. Steroids, radiotherapy, chemotherapy, stents for superior vena caval obstruction in carcinoma of the bronchus (Review) The Cochrane Library 3 John Wiley & Sons, 2005.

Bruera

, Neumann

. Management of specific symptom complexes in patients receiving palliative care. Can Med Assoc J, 1998; 158:1717–1726.

Mercadante

, Fabio

, Casuccio

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

Pilkey

, Daeninck

. A retrospective analysis of dexamethasone use on a Canadian palliative care unit. Prog Palliat Care, 2008; 16:63–68.

10.

Quinn

, Hudson

, Dunning

. Diabetes management in patients receiving palliative care. J Pain Symptom Manage, 2006; 32:275–286.

11.

McCoubrie

, Jeffrey

, Paton

, Dawes

. Managing diabetes mellitus in patients with advanced cancer: A case note audit and guidelines. Eur J Cancer Care (Engl), 2005; 14:244–248.

12.

Hardy

, Ress

, Ling

, Burman

, Feuer

, Broadley

, Stone

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

13.

Lundström

, Fürst

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

14.

Kim

, Yoo

, Lee

, Chung

, Kim

, Han

, Shim

, Yim

. Incidence and risk factors of steroid-induced diabetes in patients with respiratory disease. J Korean Med Sci, 2011; 26:264–267.

15.

Iwamaoto

, Kagawa

, Naito

, Kuzuhara

, Kojima

. Steroid-induced diabetes mellitus and related risk factors in patients with neurologic diseases. Pharmacotherapy, 2004; 24:508–514.

16.

Uzu

, Harada

, Sakaguchi

, Kanasaki

, Isshiki

, Araki

, Sugiomoto

, Koya

, Haneda

, Kashiwagi

, Yamauchi

. Glucocorticoid-induced diabetes mellitus: Prevalence and risk factor in primary renal diseases. Nephron Clin Pract., 2007; 105:c54–c57.

17.

Panthakalam

, Bhatnagar

, Klimiuk

, Scott . The prevalence and management of hyperglycaemia in patients with rheumatoid arthritis on corticosteroid therapy. Med J, 2004; 49:139–141.

18.

Clore

, Thurby-Hay

. Glucocorticoid-induced hyperglycemia. Endocr Pract., 2009; 15:469–474.

19.

Gulliford

, Charlton

, Latinovic

. Risk of diabetes associated with prescribed glucocorticoids in a large population. Diabetes Care, 2006; 29:2728–2729.

20.

Gurwitz

, Bohn

, Glynn

, Monane

, Mogun

, Avorn

. Glucocorticoids and the risk for initiation of hypoglycemic therapy. Arch Intern Med, 1994; 154:97–101.

21.

Blackburn

, Hux

, Mamdani

. Quantification of the risk of corticosteroid-induced diabetes mellitus among the elderly. J Gen Intern Med, 2002; 17:717–720.

22.

Lau

, Maida

, Downing

, Lesperance

, Karlson

, Kuziemsky

. Use of the Palliative Performance Scale (PPS) for end-of-life prognostication in a palliative medicine consultation service. J Pain Symptom Manage, 2009; 37:965–972.

23.

Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada. Can J Diabetes, 2008; 32,Suppl 1:S1–S201.

24.

Vegiopoulos

, Herzig

. Glucocorticoids, metabolism and metabolic diseases. Mol Cell Endocrinol, 2007; 275:43–61.