ORIGINAL ARTICLE

Assessment of the Prevalence and Risk Factors Associated With Glucocorticoid-

Induced Diabetes Mellitus in Pemphigus Vulgaris Patients
Abbas Darjani1, Nahid Nickhah1, Mohammad Hassan Hedayati Emami2, Narges Alizadeh1, Rana Rafiei1, Hojat Eftekhari1,
and Kaveh Gharaei Nejad1
1
Department of Dermatology, Skin Research Center, Guilan University of Medical Sciences, Rasht, Iran
2
Department of Internal Medicine, Endocrinology and Metabolism Research Center, Guilan University of Medical Sciences, Rasht, Iran

Received: 07 Apr. 2016; Revised: 07 Nov. 2016; Accepted: 01 Feb. 2017

Abstract- Pemphigus vulgaris is a chronic autoimmune disease and glucocorticoids are one of the main
treatments. Our study investigates the prevalence and associated factors of glucocorticoid-induced diabetes
mellitus in these patients under different glucocorticoid regimens. 36 patients with first diagnosed Pemphigus
vulgaris based on pathological and direct immunofluorescence findings who had received different
glucocorticoid regimens (1-2 mg/kg oral or 1-2 mg/kg oral with 1g methylprednisolone pulse daily for 3
consecutive days with or without azathioprine) were evaluated during 2014-2016. Our study found that
22.2% of patients had impaired fasting glucose and incidence of corticosteroid-induced diabetes mellitus was
22.2% with no difference between oral and pulse therapy of corticosteroid. The first day after pulse therapy
19 patients of 21 had post bolus hyperglycemia that 36% of them became diabetic after 8 weeks. None of the
variables, including age, BMI, HbA1c, LDL, HDL, TG, cholesterol, family history and blood pressure were
associated with diabetes. Pretreatment FBS was the factor that would increase the likelihood of
glucocorticoid-induced diabetes mellitus, 42.2% of patients with pretreatment FBS 100-126 developed
diabetes in comparison with 17.2% in normal pretreatment FBS. Although the group who received
azathioprine was associated with increased incidence of diabetes, the overall corticosteroid dose in this group
was significantly higher than the other group (P=0.012), and controversy with other studies could be because
of difference in corticosteroid dosage and small number of patients. The incidence of diabetes was not
different between the group with glucocorticoid pulses and oral prednisolone without pulse therapy. Higher
pretreatment FBS can be related to increased incidence of diabetes, but results from this study due to small
number of patients are preliminary and multicenter studies are needed.
© 2017 Tehran University of Medical Sciences. All rights reserved.
Acta Med Iran 2017;55(6):375-380.

Keywords: Glucocorticoid; Diabetes mellitus; Pemphigus vulgaris

Introduction Men and women both get pemphigus at the same rate. It
is most common in middle-aged and older adults, but it
Pemphigus vulgaris is a rare chronic autoimmune can occur in young adults and children (1). Prevalence
blistering disease affecting the skin and mucosal of pemphigus is 0.076 to 1.6 in 100,000. Before the
surfaces which are mediated by circulating antibodies advent of systemic corticosteroids, Pemphigus vulgaris
against a protein called Desmoglein 3 on the cell surface was usually a fatal disease with 50% mortality in 2 years
of keratinocytes. Clinically all patients with Pemphigus because large areas of the skin lost their epidermal
vulgaris have painful mucosal erosions, cutaneous barrier function, leading to the loss of fluids or to
erosions and blisters are also seen in more than 50% of secondary infections (2,3). The introduction of systemic
patients. The histologic changes include acantholysis corticosteroids and immunosuppressive agents has
with intraepidermal blisters due to loss of cell-cell greatly improved the prognosis of pemphigus; however,
adhesion of keratinocytes and deposits of IgG antibodies the morbidity, and occasional mortality are still
indirect immunofluorescence. significant due to complications of therapy. Diabetes
Pemphigus affects people of all races and cultures. mellitus is one of the major complications, which occur

Corresponding Author: A. Darjani

Department of Dermatology, Skin Research Center, Guilan University of Medical Sciences, Rasht, Iran
Tel: +98 911 1361570, Fax: +98 21 55419113, E-mail address: [email protected]
Glucocorticoid induced diabetes mellitus

because of impaired glucose metabolism by pathological and immunological tests (direct

corticosteroids due to increased insulin resistance in immunofluorescence) referred to the Departments of
tissues, increased glucose production in the liver and Dermatology of Razi Hospital at Guilan province, north
impaired glucose consumption in muscles and adipose of Iran, from March 2014 through March 2016. At first,
cells. Diabetes is one of the predisposing factors of all patients after admission (first hospitalization for
mortality in these patients (4,5). Up to 50% of patients Pemphigus vulgaris) underwent a precise assessment
have increased insulin resistance in various studies and including basic tests (FBS, HbA1C, Cholesterol, TG,
prevalence of corticosteroid-induced hyperglycemia is HDL, and LDL), Blood pressure measurement
12.6% to 40% depending on different factors such as (measured in the supine position and from left hand after
dose, duration of therapy and route of administration 30 minutes rest), and height and weight measurement.
(6,7,8,9, and 10). Immunosuppressive agents such as Demographic data such as age, sex, past medical history
azathioprine, mycophenolate mofetil, and cyclosporine and diabetic risk factors were extracted from the
are often used. Azathioprine is used at a dose of 2-4 questionnaire. All participants signed an informed
mg/kg (usually 100-300 mg/day), and its common side consent. Patients were categorized into four groups
effects are nausea and dose-dependent myelosuppression according to the type of the therapy. Group A and B
(11). received oral prednisolone 1-2 mg/kg daily without and
The aim of our study is to determine the incidence of with 100 mg azathioprine respectively. Group C and D
glucocorticoid-induced hyperglycemia and associated underwent methylprednisolone pulse therapy 1000 mg
possible risk factors (age, sex, HDL, LDL, TG, daily for 3 consecutive days plus 60mg oral
cholesterol, HbA1c, blood pressure, family history and prednisolone without and with 100 mg azathioprine
type of adjuvant) in different corticosteroid regimens in respectively (Table 1). Patients with known diabetes
order to prevent from incidence or complications in mellitus or FBS more than 126 mg/dl or HbA1c more
these patients. In this study we compare the incidence of than 6.5% before pemphigus treatment, previous
diabetes between oral and pulse prednisolone, and also prednisolone therapy for more than 7 days in the last
we investigate post-bolus hyperglycemia by measuring two years and Patients without pathological
FBS levels after every pulse therapy. confirmation of Pemphigus vulgaris were excluded from
the study. FBS was checked after every pulse therapy
Materials and Methods and then 2 times a week and patients were evaluated for
8 weeks and patients were divided into diabetic and non-
In this cross-sectional study, all patients with first diabetic groups.
diagnosed Pemphigus vulgaris according to the

Table 1. Therapeutic regimes in treatment groups

A B C D
Oral prednisolone without pulse
therapy + + -- --
(1-2 mg/kg/day)
Azathioprine (100 mg/d) - + - +
Oral prednisolone with
methylprednisolone, Pulse Therapy. - - + +
(1000 mg/day * 3day)
Patients number 5 4 9 18
Follow up duration (week) 8 8 8 8
Patients number with the second
-- -- 3 8
pulse
Total numbers of pulses -- -- 12 26

The statistical analysis of all data was performed X2 test depending on the variable. Correlations between
using SPSS version 19. The prevalence of diabetes quantitative variables were made with Pearson’s
mellitus was calculated. All quantitative data were correlation coefficient, and for comparing the groups,
expressed as mean and standard deviation. For ANOVA was used. P less than 0.05 were considered
comparative analysis, we used Student’s t-test and the significant.

376 Acta Medica Iranica, Vol. 55, No. 6 (2017)

 A. Darjani, et al.

Results The baseline LDL and HDL levels were 118±44

mg/dl and 41±8 mg/dl respectively. Increased level of
Forty-two Pemphigus vulgaris patients were LDL was found in 61.1% of patients. The baseline TG
admitted at the hospital during study time. 4 of them and cholesterol levels were 119±47mg/dl and 181±35
excluded from the study because of previous
respectively. Increased level of TG and cholesterol were
corticosteroid usage and 2 patients left the study. The
remaining 36 patients were divided into 4 groups. There seen in 36% and 25% of patients. 58.3% of patients had
were 5 patients in group A, 4 patients in group B, 9 normal blood pressure.
patients in group C and 18 patients in group D. 3 A day after each pulse of methylprednisolone,
patients in group C and 8 patients in group D underwent glucose levels increased to 155±28 mg/dl, 152±24
second pulse therapy after 21-30 days from the first one mg/dl, 148±30 mg/dl respectively (increased
(See Table 1). significantly compared to the FBS level before
Gender distribution was 17 women (47.2%) and 19 treatment) but none of them were significantly different
male (52.8%). The mean time of admission was 23 days. between group C and D (Table 2). 19 patients of 21
Because the period of study was eight weeks, the patients underwent pulse therapy had hyperglycemia
follow-up was done on an outpatient basis on Patients after the first pulse that 7 of them (36%) became
who were discharged earlier. The average age was diabetic after eight weeks. In our study, the patients who
48±15 years. BMI in 58.3% of patients was in normal had higher FBS level at first and third day; their FBS
range, and overweight was found in 30.6% and obesity level was also high after 8 weeks, but it was not true
in 8.3% and morbid obesity in 2.8% of patients. A about the second day FBS level (Graph1).
family history of diabetes mellitus was found in 30.6%.

Table 2. Mean FBS levels a day after each pulse therapy in C and D groups
Lower confidence Upper confidence
* Mean Standard deviation P
interval interval
C:152.90 C:23.269
FBS 1st day(mg/dl) 0.701 -23.679 24.714
D:157.24 D:30.363

C:157.80 C:22.695
FBS 2nd day(mg/dl) 0.402 -17.686 19.651
D:149.41 D:25.742

C:150.70 C:32.469
FBS 3rd day(mg/dl) 0.784 -27.968 29.086
D:147.24 D:30.667
* FBS mean difference after the First, second and third days of pulse between the two groups, C (treated with methylprednisolone pulse then oral
prednisolone) and D (treated with methylprednisolone pulse then oral prednisolone and azathioprine), is investigated

Graph 1. Relationship of FBS levels after every pulse with FBS level after 8 weeks

8 weeks after treatment; 55.5%of patients were non- patient in group B and 7 patients in group D). Our study
diabetic, 8 patients (22.2%) were pre-diabetic (2 patients showed that administration of corticosteroids
in group A and 1 patient in group C and 5 patients in significantly increases incidence if diabetes (P=0.034).
group D), and 8 patients (22.2%) were diabetic (1 Pretreatment FBS was a strong predictor of

Acta Medica Iranica, Vol. 55, No. 6 (2017) 377

Glucocorticoid induced diabetes mellitus

corticosteroid-induced diabetes. 17.2% of patients with (P=0.866).

pretreatment FBS fewer than 100 became diabetic The mean HbA1c level in the non-diabetic group
(42.2% in patients with pretreatment FBS 100-126) and was 4.76 and in pre-diabetic and diabetic groups was
17.2% of them developed impaired glucose intolerance 5.04 and 5.16 respectively, and there was no difference
(42.2% in patients with pretreatment FBS 100-126) in HbA1c level between these groups (P=0.219). Mean
(P=0.05). levels of LDL (116.85 in non-diabetic, 130 in pre-
After 8 weeks the mean FBS level was only diabetic and 109.88 in diabetic), HDL (42.65 in non-
significantly different between group C and D (P=0.012) diabetic, 37.38 in pre-diabetic and 40.63 in diabetic), TG
and the relationship between high FBS level and (115.05 in non-diabetic, 119.63 in pre-diabetic and
adjunction of azathioprine to pulse therapy was seen 128.38 in diabetic), and cholesterol (184.90 in non-
(P=0.009). As there was no difference in FBS levels diabetic, 172.38 in pre-diabetic and 182.38 in diabetic),
after every pulse therapy between group C and D, we were not significantly different between non-diabetic
evaluated maintenance glucocorticoid dosage during 8 and diabetic patients (Table 3). Family history
weeks, and we found that group D had taken (P=0.225), BMI (P=0.403) and blood pressure
significantly higher (up to 20 mg in some patients) (P=0.072) showed no association with incidence of
glucocorticoid dose than group C (0.012). There was no diabetes.
difference in FBS level according to number of pulses

Table 3. Association between variables and incidence of diabetes

Dependent 95% Confidence Interval
(I) FBS (J) FBS P
Variable Lower Bound Upper Bound
100-125 <99 0.225 -13.07 2.52
HD >126 <99 0.787 -9.82 5.77
100-125 <99 0.732 -31.03 57.33
LDL >126 <99 0.915 -51.15 37.2
100-125 <99 0.637 -47.35 22.3
CHOLE >126 <99 0.981 -37.35 32.3
100-125 <99 0.967 -42.97 52.12
TG >126 <99 0.758 -34.22 60.87
100-125 <99 0.620 -44.24 23.43
Age >126 <99 0.840 -33.2 29.98
100-125 <99 0.224 -13.12 10.45
Sex >126 <99 0.437 -18.35 14.77
100-125 <99 0.403 -35.55 16.67
BMI >126 <99 0.688 -46.11 23.11
100-125 <99 0.0702 -1.233 2.33
BP
>126 <99 0.0801 -1.33 2.16
100-125 <99 0.225 -13.06 2.54
FH >126 <99 0.488 -14.01 11.02
100-125 <99 0.219 -12.89 3.34
HbA1c >126 <99 0.413 -14.33 12.22
100-125 <99 0.879 -45.32 54.89
Pulse or oral >126 <99 0.923 -48.99 48.23
100-125 <99 0.009 -1.21 2.12
Azathioprine >126 <99 0.009 -1.23 2.13
Pulse 100-125 <99 0.866 -56.35 34.78
number >126 <99 0.899 -59.33 31.98
100-125 <99 0.050 -3.42 2.33
Pretreatment
FBS >126 <99 0.050 -3.43 2.11

By performing Pearson’s correlation between the incidence of diabetes mellitus. But the positive
type of therapy and other risk factors, we found that correlation between azathioprine and corticosteroid
none of them was statistically predictive factor in the dosage was seen (Table 4).

378 Acta Medica Iranica, Vol. 55, No. 6 (2017)

 A. Darjani, et al.

Table 4. Correlation between risk factors and type of therapy (with or without Azathioprine)
Therapy HDL TG CHOLE HbA1c BP BMI FH Corticoid dose
Pearson -0.331 0.061 -0.166 0.301 0.195 0.286 0.145 0.595
correlation
P 0.069 0.722 0.334 0.075 0.255 0.091 0.277 0.050

Discussion preventive effect on the onset of diabetes mellitus and

no study was found in the relationship of azathioprine
In spite of effectiveness of corticosteroids in the and increased risk of diabetes mellitus (16,17). In
treatment of patients with Pemphigus vulgaris, contrast to azathioprine, a study found relationship
development of severe side effects like diabetes has between administration of mycophenolate mofetil and
limited their use. Our study on 36 patients with increased incidence of diabetes mellitus (18). In this
Pemphigus vulgaris, who underwent different regimens study pretreatment FBS was the factor that would
of corticosteroid therapy (oral or pulse therapy) with or increase the likelihood of glucocorticoid-induced
without azathioprine showed that 22.2% of patients had diabetes mellitus, 42.2% of patients with pretreatment
impaired fasting glucose and incidence of corticosteroid- FBS 100-126 developed diabetes in comparison with
induced diabetes mellitus were 22.2% with no difference 17.2% in normal pretreatment FBS.
between oral and pulse therapy of corticosteroid. Other Type of therapy, age, BMI, family history, HbA1C,
studies had similar results, Valikhani et al., reported that LDL, HDL, TG, and cholesterol were not associated
the incidence of corticosteroid-induced diabetes mellitus with diabetes. Valikhani et al. found that HbA1c (mean
was 27.9% among Pemphigus vulgaris patients (12). level of 5.4 in diabetic patients compared to 4.98 in non-
Generally, the rate of corticosteroid-induced diabetic) and triglyceride (159mg/dl in diabetic patients
hyperglycemia was 30-40% in the majority of studies compared to 120 in non-diabetic) were risk factors for
(13,14). developing diabetes (12). But Alavi et al. reported
In this study, we found a significant increase in post similar findings in our study and in their study age and
bolus glucose level, more evident after the first pulse BMI and family history were not associated with
with the elevation of about 50 mg/dl. 81% of patients hyperglycemia, and glucocorticoid dose and duration of
after the first pulse, 96% after the second pulse and 96% treatment were the only effective factors for developing
after the third pulse had hyperglycemia. Perez et al. diabetes (14). There was also another study by Esmaeili
reported similar elevation of glucose level after the first et al., in which age, BMI, HbA1c, and HDL were strong
pulse (40 mg/dl) and some different from our study, risk factors for developing diabetes (19). Although we
68% of patients after the first pulse, 94% after the found no association between age and diabetes, we
second pulse and 98% after the third pulse had should consider that average age was under 50 years in
hyperglycemia (15). our study and it may influence the association between
In our study mean level of blood glucose after pulse age and diabetes. About the lack of association between
therapy was not different according to adjunction of family history and diabetes, it is possible that screening
azathioprine but at the end of treatment a significant for diabetes is not complete in Iranian families. Results
difference was seen at the rate of developing diabetes from this study due to small number of patients are
mellitus between groups with and without azathioprine preliminary, Due to the rarity of pemphigus in the world
(0% in pulse therapy without azathioprine compared to in order to investigate incidence of diabetes according to
25% in the opposite group). As patients underwent different corticosteroid regimens performing multicenter
azathioprine therapy had taken significantly higher dose studies is recommended.
of maintenance glucocorticoid (average level of
corticosteroid: 63.8 mg/dl in group D and 59.98 mg/dl in References
group C), we can assume that the difference in the rate
of diabetes between group C and D may be related to 1. Wojnarowska F, Venning VA, Burge SM,
small number of patients in our study or the different Immunobullous disease. In: Burns T, Breathnach S, Cox
maintenance doses of glucocorticoid between dose N, Griffiths C, eds. Rook’s Textbook of dermatology. 7th
groups not to the adjunction of azathioprine (Table 4). ed. Oxford: Blackwell Science 2004:41.3-41,12.
Almost other studies had shown that azathioprine had a

Acta Medica Iranica, Vol. 55, No. 6 (2017) 379

Glucocorticoid induced diabetes mellitus

2. Irwin MF, Arthur ZE, Kalus W, Frank A, Lowll AG, Textbook of Dermatology. 3rd Edition, Eds (2012)
Stephen LK, eds. Fitz patrick’s dermatology in general Elsevier Publishing, Chapters 29, Page 472.
medicine. 6th ed. New York: MC Graw-holl, 2004:558- 12. Valikhani M, Khoshniat Niko M, Tork AN. Risk factors
67. and frequency of steroid - induced diabetes in pemphigus
3. Rucocco E, Baroni A, Wolf R, Ruocco V,Life-threatening vulgaris patients during 1 year study, IJDD 2007;6:301-7.
bullous dermatoses: Pemphigus vulgaris. Clin Dermatol 13. Gonzalez-Gonzalez JG, Mireles-Zavala LG, Rodriguez-
2005;23:223-6. Gutierrez R, Gomez-Almaguer D, Lavalle-Gonzalez F J,
4. Zeina B, MD, PhD Consulting Staff, Department of Tamez-Perez HE, et al. Hyperglycemia related to high-
Dermatology, Milton Keynes Hospital, UK. (Accessed dose glucocorticoid use in noncritically ill patients.
March 2017, 12, at Diabetol Metab Syndr 2013;5:18.
http://emedicine.medscape.com/article/1064187-overview 14. Alavi A, Lowe J, Walsh S, Juurlink D, Mortaz-Hedjri S
#aw2aab6b2b4aa). and Neil H, Corticosteroid-induced hyperglycemia is
5. CeCilia lansang M, Hustak LK, Glucocorticoid-induced increased 10-fold in patients with pemphigus. Int J
diabetes and adrenal suppression: how to detect and Dermatol 2012;51:1248-52.
manage them. Cleve Clin J Med 2011;78:748-56. 15. Perez HET, De-Ossio MDG, Flores DLQ, Coria MIH,
6. Ha YJ, Lee KH, Jung S, Lee SW, Lee SK, Park YB. Peña ALT, Pérez GJC, et al. Glucose disturbances in non-
Glucocorticoid-induced diabetes mellitus in patients with diabetic patients receiving acute treatment with
systemic lupus erythematous treated with high-dose methylprednisolone pulses. Rev Assoc Med Bras
glucocorticoid therapy. Lupus 2011;20:1027-34. 2012;58:125-8.
7. Kim SY, Yoo CG, Lee CT, Chung HS, Kim YW, Han 16. Calafiiore R, Basta G, Falorni A, Pietropaolo M, Picchio
SK, et al. Incidence and risk factors of steroid-induced ML, Calcinaro F, et al. Preventive effects of azathioprine
diabetes in patients with respiratory disease. J Korean on the onset of diabetes mellitus in NOD mice. J
Med Sci 2011;26:264-7. Endocrinol Invest 1993;16:869-73.
8. Pagano G, Cavallo-Perin P, Cassader M, Bruno A, 17. Silverstein J, Maclaren N, Riley W, Spillar R.
Ozzello A, Masciola P. An in vivo and in vitro study of Immunosuppression with Azathioprine and Prednisolone
the mechanism of prednisone-induced insulin resistance in Recent-Onset Insulin Dependent Diabetes Mellitus. N
in healthy subjects. J Clin Invest 1983;72;1814-20. Engl J Med 1988;319:599-604.
9. Abdelmannan D, Tahboub R, Genuth S, Ismail-Beigi F. 18. Zabihi Yeganeh M, Sadeghi S. Risk Factors of
Effect of dexamethasone on oral glucose tolerance in Glucocorticoid-Induced Diabetes Mellitus in Systemic
normal individuals, Endocr Pract 2010;16:770-7. Lupus Erythematosus. GMJ 2013;2:39-43.
10. Nicod N, Giusti V, Besse C, Tappy L. Metabolic 19. Esmaeili N, Soori T, Shirzad N, Vahid-Moghadam M,
adaptations to dexamethasone-induced insulin resistance Karimi A. Frequency and risk factors for steroid-induced
in healthy volunteers. Obes Res 2003;11:625-31. diabetes in pemphigus vulgaris patients in Razi Hospital,
11. Bolognia, JL., Jorizzo, JL. and Schaffer, JV., Bolognia Tehran. Dermatol Cosmet 2015;6:140-6.

380 Acta Medica Iranica, Vol. 55, No. 6 (2017)