Review Endocrinol Metab 2017;32:180-189

https://doi.org/10.3803/EnM.2017.32.2.180
Article pISSN 2093-596X · eISSN 2093-5978

Glucocorticoid-Induced Diabetes Mellitus: An Important

but Overlooked Problem
Sunghwan Suh, Mi Kyoung Park

Division of Endocrinology and Metabolism, Department of Internal Medicine, Dong-A University Hospital, Dong-A University
College of Medicine, Busan, Korea

Glucocorticoids are widely used as potent anti-inflammatory and immunosuppressive drugs to treat a wide range of diseases. How-
ever, they are also associated with a number of side effects, including new-onset hyperglycemia in patients without a history of dia-
betes mellitus (DM) or severely uncontrolled hyperglycemia in patients with known DM. Glucocorticoid-induced diabetes mellitus
(GIDM) is a common and potentially harmful problem in clinical practice, affecting almost all medical specialties, but is often diffi-
cult to detect in clinical settings. However, scientific evidence is lacking regarding the effects of GIDM, as well as strategies for pre-
vention and treatment. Similarly to nonsteroid-related DM, the principles of early detection and risk factor modification apply.
Screening for GIDM should be considered in all patients treated with medium to high doses of glucocorticoids. Challenges in the
management of GIDM stem from wide fluctuations in postprandial hyperglycemia and the lack of clearly defined treatment proto-
cols. Together with lifestyle measures, hypoglycemic drugs with insulin-sensitizing effects are indicated. However, insulin therapy is
often unavoidable, to the point that insulin can be considered the drug of choice. The treatment of GIDM should take into account
the degree and pattern of hyperglycemia, as well as the type, dose, and schedule of glucocorticoid used. Moreover, it is essential to
instruct the patient and/or the patient’s family about how to perform the necessary adjustments. Prospective studies are needed to an-
swer the remaining questions regarding GIDM.

Keywords: Glucocorticoids; Hyperglycemia; Diabetes mellitus

INTRODUCTION tients may need to use corticosteroids for more than 6 months
[2]. Glucocorticoids are extensively used in almost every sub-
Glucocorticoids have been shown to be potent anti-inflammato- specialty of medicine [3]. Although they are widely prescribed
ry and immunosuppressive drugs; they started to be used for for their anti-inflammatory and immunosuppressive properties,
therapeutic purposes in the mid-20th century, and are currently glucocorticoids have a range of common metabolic side effects
widely used in the treatment of many diseases [1]. At any point including hypertension, osteoporosis, and diabetes [2]. Gluco-
in time, up to 0.9% of the general population may be using cor- corticoid-induced diabetes mellitus (GIDM) has been recog-
ticosteroids, with the highest rate of use (2.5%) seen in individ- nized as a complication of glucocorticoid use for over 60 years
uals 70 to 79 years of age. Moreover, nearly one-quarter of pa- [4]. Glucocorticoids are the drug group most often associated

Received: 12 March 2017, Revised: 20 March 2017, Accepted: 27 March 2017 Copyright © 2017 Korean Endocrine Society
Corresponding author: Mi Kyoung Park This is an Open Access article distributed under the terms of the Creative Com­
Division of Endocrinology and Metabolism, Department of Internal Medicine, mons Attribution Non-Commercial License (http://creativecommons.org/
Dong-A University Hospital, Dong-A University College of Medicine, 26 licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribu­
Daesingongwon-ro, Seo-gu, Busan 49201, Korea tion, and reproduction in any medium, provided the original work is properly
Tel: +82-51-240-2815, Fax: +82-51-242-5852, E-mail: pmk02@dau.ac.kr cited.

180 www.e-enm.org
 Glucocorticoid-Induced DM

with the onset of hyperglycemia or diabetes mellitus [1]. Gluco-

Table 1. Risk Factors for Glucocorticoid-Induced Diabetes Mel-
corticoids exacerbate hyperglycemia in patients with diabetes litus
mellitus, unmask undiagnosed diabetes mellitus or may precipi-
Higher dose of glucocorticoid treatment (prednisolone >20 mg, hydro-
tate the appearance of GIDM, which is an independent risk fac- cortisone >50 mg, dexamethasone >4 mg)
tor for other complications associated with the use of these Longer duration of glucocorticoid treatment
drugs [2]. Furthermore, although it is expected that the blood Advanced age
glucose levels of non-diabetic patients should normalize after High body mass index
discontinuing glucocorticoid use, this does not always happen, Previous glucose intolerance or impaired glucose tolerance
and such patients require close monitoring due to the risk of de- Personal history of gestational diabetes or previous glucocorticoid-in-
veloping diabetes mellitus in the future [5]. duced hyperglycemia
The diagnosis and treatment of GIDM are surprisingly under- Family history of diabetes mellitus
valued by many health-care professionals, although GIDM re- Hemoglobin A1c ≥6%
sults in an increased use of medical care and the development of
associated morbidities that affect medical costs and patients’
quality of life [2]. Despite the high prevalence of corticosteroid ing glucocorticoids in various studies has been reported to range
use in clinical practice and the strong association of glucocorti- from 1.36 to 2.31 [10].
coids with new-onset hyperglycemia, limited information is It is unlikely that the development of hyperglycemia is related
available regarding the pathophysiology, frequency, and diagno- only to the existence of an underlying disease, because abnor-
sis of this condition. Similarly, prospective studies assessing the malities in glucose metabolism have also been observed in ex-
effectiveness of preventive measures taken against this condi- perimental studies with healthy subjects. The incidence of hy-
tion and comparing the effectiveness of different treatment perglycemia (defined as blood glucose >200 mg/dL) in hospi-
schedules are lacking. Therefore, the primary goal of this review talized patients without a known history of diabetes who are
article was to highlight the most recent evidence regarding the treated with corticosteroids is over 50% [7]. Not all patients
clinical aspects of GIDM. who use glucocorticoids develop GIDM, suggesting that GIDM
occurs only in vulnerable populations. A number of predictors
PREVALENCE of the onset of GIDM have been identified [1,2,9,15]. These
predictors include the dose and duration of glucocorticoid treat-
The incidence of glucocorticoid-induced hyperglycemia has ment, age, weight, previous glucose intolerance, reduced sensi-
been reported to be 12% [6]. Other studies of the prevalence of tivity to insulin or impaired insulin secretion stimulated by glu-
GIDM associated with different diseases have reported varying cose, a family history of diabetes, or race, as shown in Table 1.
results [7-9]. However, the exact prevalence of hyperglycemia Gender does not seem to be a predictor and, conversely, early
secondary to glucocorticoid therapy is not known, making this withdrawal from corticosteroids is a protective factor [1,2,9,15].
an unpredictable challenge for general practitioners and endo- Various glucocorticoids differ in their pharmacological proper-
crinologists alike [5]. Gulliford et al. [10] reported that 2% of ties and duration of action, and can be administered in different
incident cases of diabetes were associated with oral glucocorti- schedules depending on the therapeutic indication. These drugs
coids in a primary care population. The prevalence of abnormal may cause hyperglycemia when administered at supraphysio-
glucose metabolism in patients with an organ transplant who logical doses by any route (topical, oral, inhaled, intramuscular,
undergo glucocorticoid therapy has been reported to be 17% to intravenous, or intra-articular) [10].
32% [11]. GIDM has been reported to have an even greater
prevalence among patients treated with glucocorticoids for PATHOPHYSIOLOGY
rheumatoid arthritis [12]. A recent meta-analysis indicated that
the rates of glucocorticoid-induced hyperglycemia and diabetes The effects of glucocorticoids on glucose homeostasis are com-
were 32.3% and 18.6%, respectively [13]. The length of time on plex and not completely understood. Their negative effects are
glucocorticoids, the relative potency of the glucocorticoid and believed to be caused by a variety of factors, including increased
the absolute dose all play a role in the occurrence of GIDM [14]. insulin resistance, increased glucose intolerance, reduced β-cell
The odds ratio for presenting with new-onset diabetes after tak- mass from β-cell dysfunction, and increased hepatic insulin re-

Copyright © 2017 Korean Endocrine Society www.e-enm.org  181

 Suh S, et al.

sistance leading to impaired suppression of hepatic glucose pro- CLINICAL COURSE

duction [16,17]. The metabolic effects of glucocorticoids on
glucose metabolism are seen at numerous stages in the insulin- There are several reasons why glucocorticoid-induced hyper-
signaling cascade [5]. Prednisolone administration for 7 days in glycemia is important in clinical practice. First, it is quite com-
healthy volunteers led to a 50% reduction in insulin sensitivity, mon, and the problems of hyperglycemia itself, which are often
as assessed using the insulin clamp methodology [18]. The ef- severe, may cause hospitalization, prolonged hospital stays, or
fect of glucocorticoids on glucose metabolism likely results repeated emergency room visits. Second, glucocorticoids in-
from the impairment of multiple pathways, including β-cell dys- crease the risk of infection due to persistent hyperglycemia,
function (sensitivity to glucose and ability to release insulin) and which has a deleterious effect on patients’ prognoses. Moreover,
insulin resistance in other tissues [3]. Glucocorticoids reduce the control of transient hyperglycemia in similar situations dur-
peripheral glucose uptake at the level of the muscle and adipose ing hospitalization has been associated with decreased mortality
tissue [19,20]. Skeletal muscle is primarily responsible for the and complication rates [26-28]. Most inpatients given glucocor-
insulin-mediated capture of postprandial glucose and corticoste- ticoids at a dose at least equivalent to 40 mg/day for more than
roids can induce insulin resistance by interfering directly with 2 days develop hyperglycemia [11]. It is well known that gluco-
various components of the insulin signaling cascade [19,21]. corticoid therapy may provoke new-onset type 2 diabetes melli-
Chronic glucocorticoid overexposure alters body composition, tus (T2DM) and invariably worsens hyperglycemia in patients
including the expansion of adipose tissue depots in the trunk, with preexisting diabetes mellitus [10]. Glucocorticoids are a
and impairs metabolism and insulin action, resulting in hyper- well-known cause of hyperosmolar hyperglycemic nonketotic
glycemia and dyslipidemia [16]. The ability of glucocorticoids syndrome [28], which requires admission to the hospital for ag-
to induce adipose tissue lipolysis depends on their concentra- gressive hydration and insulin therapy. The tendency for high-
tion, duration of exposure, and the specific adipose tissue depot risk patients to develop new-onset hyperglycemia is often not
[16]. The liver plays a major role in the control of glucose me- anticipated, and outpatients with new-onset GIDM frequently
tabolism, maintaining fasting euglycemia. Corticosteroids in- require hospitalization for intravenous hydration and initiation
crease endogenous glucose production directly by activating of insulin therapy. In addition, hyperglycemia in hospitalized
numerous genes involved in the hepatic metabolism of carbohy- patients (with or without diabetes) has been associated with
drates, leading to increased gluconeogenesis [19,22]. Glucocor- poor outcomes regardless of the underlying cause [17]. Inpatient
ticoids also inhibit the production and secretion of insulin from hyperglycemia is known to be associated an increased length of
pancreatic β-cells [23,24]. The abilities of glucocorticoids to in- stay and a greater incidence of infections [27]. In addition, sus-
duce hyperglycemia depend on their dose and the duration of tained glucocorticoid treatment increases the potential for future
exposure [16,25]. In addition to insulin resistance, inflammatory cardiovascular disease through multiple pathways [11]. Gluco-
disease states themselves may induce β-cell dysfunction through corticoids are administered using different schedules, doses, and
indirect mechanisms [11]. The mechanisms of glucocorticoid- methods, and it is possible that the mode of administration (e.g.,
induced hyperglycemia are summarized in Table 2. inhaled, injected, or topical) and dosing regimen may affect the
development of hyperglycemia [2]. Knowledge of these hyper-
glycemia patterns is essential to establish the most appropriate
strategy for diagnosis and therapy.
Table 2. Mechanisms of Glucocorticoid-Induced Diabetes Mel-
litus
Reduced peripheral insulin sensitivity and/or promotion of weight gain
DEFINITION AND DIAGNOSIS
Increase in glucose production through promotion of hepatic gluconeo-
genesis
GIDM is defined as an abnormal increase in blood glucose as-
Destruction of pancreatic cells, leading to β-cell injury (inflammation)
sociated with the use of glucocorticoids in a patient with or
β-Cell dysfunction
without a prior history of diabetes mellitus. The diagnosis of
Impaired insulin release
GIDM, as well as other types of diabetes, is set according to the
Inhibited glyceroneogenesis
criteria established by expert committees [29]. The current cri-
teria for diagnosing diabetes are an 8-hour fasting blood glucose
Increase in fatty acids
level ≥7.0 mmol/L (126 mg/dL), a 2-hour post-75 g oral glu-

182 www.e-enm.org Copyright © 2017 Korean Endocrine Society

 Glucocorticoid-Induced DM

cose tolerance test (OGTT) ≥11.1 mmol/L (200 mg/dL), a he- may underestimate the true incidence of GIDM. Therefore, the
moglobin A1c (HbA1c) percentage ≥6.5%, or in patients with most useful criterion for the diagnosis of GIDM in most patients
symptoms of hyperglycemia, a random plasma glucose reading is a glucose level >200 mg/dL at any time of the day [2].
≥11.1 mmol/L (200 mg/dL) [29]. In patients with pre-existing diabetes or risk factors for
It is generally thought that glucocorticoids result mainly in an GIDM, screening should be performed even when low doses of
increase in postprandial blood glucose levels [30]. The use of a corticosteroids are indicated [2]. However, no glucose monitor-
continuous blood glucose monitor in chronic obstructive pul- ing was performed in 24% of patients receiving high-dose glu-
monary disease patients treated with prednisolone demonstrated cocorticoid therapy [7]. Failure to appreciate the frequency of
that hyperglycemia predominantly occurred in the afternoon hyperglycemia in patients without diabetes is associated with
and evening, indicating that this would be the most appropriate the time course of corticosteroid action with prednisone or pred-
time to screen for GIDM as well as the preferable time to imple- nisolone given once daily [7]. No consensus exists for the opti-
ment specific treatments [31]. In patients treated with most glu- mal screening frequency. Some authors recommend screening
cocorticoids, measuring fasting blood glucose can underesti- with a once- to twice-weekly 2-hour post-lunch capillary blood
mate glucocorticoid-induced hyperglycemia and diabetes, par- glucose test in patients who are not known to be diabetic, but
ticularly in intermediate-acting treatments that are administered who are at a high risk of GIDM [5]. The American Diabetes As-
in single morning doses. However, given the pathophysiology sociation further recommended that glucose monitoring be con-
and pattern of glucocorticoid-induced hyperglycemia, it can be ducted in patients without known diabetes who receive thera-
inferred that the established criteria [29] provide a low diagnos- pies with a high risk of hyperglycemia, such as parenteral or en-
tic sensitivity in most patients with GIDM. For most glucocorti- teral nutrition, glucocorticoids, immunosuppressive therapy, or
coids, measuring fasting blood glucose can underestimate glu- octreotide [26]. The Endocrine Society recommended that all
cocorticoid-induced hyperglycemia and diabetes, especially in patients have blood glucose levels tested on admission, regard-
cases of intermediate-acting glucocorticoid treatment with sin- less of whether they have a pre-existing diagnosis of diabetes
gle morning doses [11]. On the contrary, postprandial glycemia [27]. This approach is believed to be warranted by the opportu-
after lunch offers the greatest diagnostic sensitivity. Preprandial nity to diagnose new diabetes and to assess glycemic control
glycemia at dinner offers less sensitivity, but is easier to stan- early course of hospitalization. Gonzalez-Gonzalez et al. [34]
dardize [11]. Blood glucose monitoring should be considered reported that, 50% of the time, GIDM occurred between the 2nd
for non-diabetic patients who are at high risk of developing and 4th week. Therefore, continued blood glucose monitoring is
GIDM as shown in Table 1 [5]. In addition, an OGTT should advisable before meals and at bedtime in patients who exhibit
also be performed as early as to detect diabetes in those deemed hyperglycemia, and is definitely recommended for those who
to be at risk [32]. However, the OGTT does not seem suitable require anti diabetic medications.
for the diagnosis of GIDM because, apart from the difficulties
associated with its implementation, it is performed in fasting TREATMENT
patients and may underestimate the increase in glucose levels
that occurs predominantly in the evening [2]. HbA1c may be a Efforts to minimize hyperglycemic episodes are beneficial for
suitable method for diagnosis in patients treated with corticoste- patients exposed to glucocorticoids. Fluctuations in plasma glu-
roids for >2 months, but it is not useful for patients whose treat- cose concentrations have shown an increased risk of cardiovas-
ment has been initiated more recently. Recently, Japanese scien- cular mortality [35]. As is the case for diagnosis, no clear evi-
tists developed a detection algorithm for GIDM and suggested dence exists for the establishment of therapeutic goals in pa-
that monitoring changes in HbA1c levels is important for de- tients with GIDM. From a practical standpoint, we recommend
tecting GIDM [33]. In patients where HbA1c levels could be that the treatment of corticosteroid-induced hyperglycemia be
unreliable, such as those with hemoglobinopathies, renal failure, considered when the preprandial and postprandial capillary glu-
or anemia/recent blood transfusion, fructosamine measurements cose levels are ≥140 and ≥200 mg/dL, respectively. In the case
would be a better alternative [5]. In order to appreciate the mag- of chronic treatment with glucocorticoids at more or less stable
nitude of GIDM, one needs to consider that glucocorticoids doses, the control goals and the need for drug treatment can be
cause predominantly postprandial hyperglycemia and therefore, based on the recommended control aims for most patients with
using impaired fasting glucose as the sole diagnostic criterion, diabetes mellitus: preprandial glycemia <130 mg/dL, postpran-

Copyright © 2017 Korean Endocrine Society www.e-enm.org  183

 Suh S, et al.

dial glycemia <180 mg/dL, and HbA1c <7% [26,27]. Similarly ORAL HYPOGLYCEMIC AGENTS
to the recommendations for T2DM, the introduction of thera-
peutic measures should be progressive and additive. The selec- Theoretically, of the entire range of oral agents that can be used
tion strategy for hypoglycemic drugs should prioritize those to treat glucocorticoid-induced or glucocorticoid-exacerbated
with a mechanism of action that fits with the pathophysiology diabetes, preference should be given to agents that target post-
of the process and the patient’s hyperglycemic profile. In the prandial hyperglycemia and have a rapid onset of action. The
posttransplant setting, as more studies will be conducted with information currently available in the literature suggests that
these and other agents, it is essential to focus attention to on oral agents do not adequately control hyperglycemia in hospital-
drug—drug interactions is essential. ized patients, and that sliding-scale regimens have resulted in
At present, no consensus guidelines exist for the optimal man- hypoglycemia [17]. For these reasons, the use of oral hypogly-
agement of hyperglycemia secondary to glucocorticoids, al- cemic drugs is reserved for the treatment of mild glucocorti-
though various opinions have been published by international coid-induced hyperglycemia (glycemia <200 mg/dL) in pa-
organizations. There is no evidence to confirm which hypogly- tients without known diabetes or with diabetes adequately con-
cemic drugs and treatment regimens are more effective in trolled by lifestyle measures or oral hypoglycemic drugs [2]. In
achieving adequate glycemic control and lowering complication the absence of studies comparing different strategies, the choice
rates in patients with glucocorticoid-induced hyperglycemia [2]. of the oral hypoglycemic drug should depend on the type and
Therefore, the recommendations proposed in the present review schedule of the corticosteroid and on the potential advantages
are based on the pathophysiology of the process, the mechanism and disadvantages of oral hypoglycemic drugs.
of action of various hypoglycemic drugs, and few clinical stud-
ies. First, it is important to evaluate the degree of pre-existing INSULIN SENSITIZERS
glucose intolerance, the patient’s clinical condition, and the de-
gree of hyperglycemia. Second, it is essential to determine the Together or after lifestyle measures, metformin and, in cases of
type, dose, and frequency of administration of the corticosteroid intolerance or contraindication to metformin, pioglitazone are
compound. Third, it is necessary to recognize the mechanism of useful and can be considered the first choice in patients who re-
action, pharmacokinetics, and pharmacodynamics of various ceive chronic treatment with low doses of intermediate-acting
hypoglycemic drugs. The final relevant aspect of addressing the glucocorticoids administered in the morning [1]. The character-
treatment of GIDM is to differentiate between temporary and in- istics of metformin and pioglitazone that contribute to their se-
definite treatment with glucocorticoids. All these aspects condi- lection include their insulin-sensitizing effect, which avoids the
tion the selection and schedule of hypoglycemic measures, as risk of hypoglycemia, and the low cost of metformin. The dose
well as the goals set in terms of glycemic control. Together with, and use of these drugs in GIDM does not differ from other clini-
or after, lifestyle measures, hypoglycemic drugs with insulin- cal settings. Metformin is an attractive option because it en-
sensitizing effects are indicated. Other oral hypoglycemic drugs hances insulin sensitivity; thus, preventing metabolic side ef-
or insulin therapy can be considered as the second drug of choice fects during systemic glucocorticoid therapy [36]. However,
[2]. As second drugs of choice, pioglitazone, sulfonylureas, many patients treated with glucocorticoids receive treatment for
glinides, dipeptidylpeptidase 4 (DPP-4) inhibitors, glucagon-like conditions associated with hypoxia and renal insufficiency,
peptide 1 (GLP-1) receptor agonists or insulin therapy can be which are both at least relative contraindications to the use of
considered, taking into account their profile of action, tolerance, metformin because of the increased risk of lactic acidosis. In
cost, and risk for hypoglycemia, especially at night [1]. patients requiring long-term glucocorticoid use, metformin
could be a reasonable choice given acceptable renal and liver
NON-PHARMACOLOGIC INTERVENTION function. Thiazolidinediones (TZDs) may improve diabetes-re-
lated parameters by antagonizing the pathways of glucocorti-
As with all types of diabetes, the initial steps to improve glyce- coid-induced insulin resistance and by reversing the adverse ef-
mic control incorporate lifestyle modifications, which include fects of glucocorticoids on β-cell function [37]. Therefore, the
exercise and dietary counseling to provide options that may re- use of TZDs, such as pioglitazone in GIDM has also been sug-
duce postprandial hyperglycemia [3]. gested [38]. These agents have been used for the long-term
treatment of transplant-induced diabetes mellitus with some

184 www.e-enm.org Copyright © 2017 Korean Endocrine Society

 Glucocorticoid-Induced DM

success in combination with other agents [38]. Vu et al. [39] and inhibiting glucagon secretion in the fasting and postprandial
also reported the added benefit of increased progression-free setting. Drugs with incretin effects should probably be the drug
survival with the use of metformin and TZDs in acute lympho- of choice because of their immediate onset of action, their pre-
blastic leukemia. However, the usefulness of metformin and pi- dominant effect on postprandial glycemia, and their low risk of
oglitazone in the treatment of transient corticosteroid-induced hypoglycemia related to glucose-dependent effects [40]. A recent
hyperglycemia is limited due to their slow onset of action [2]. study found that concomitant treatment with a DPP-4 inhibitor
improved various aspects of pancreatic islet-cell function in pa-
INSULIN SECRETAGOGUES tients receiving high-dose prednisolone [41]. Furthermore, an in-
travenous infusion of exenatide prevented glucose intolerance
Oral secretagogues such as sulfonylurea therapy do not specifi- induced by high doses of prednisolone in healthy patients and
cally target postprandial hyperglycemia and thus long-acting may be explored as a potential strategy to prevent GIDM [42].
agents may be associated with hypoglycemia if the patient does
not eat meals regularly. In cases where hyperglycemia is in- INSULIN
duced by intermediate-acting glucocorticoids in 2 or more daily
doses, by long-term preparations such as dexamethasone, or by Insulin therapy is often unavoidable when hyperglycemia is the
intra-articular glucocorticoids, sulfonylureas may also be a use- result of the combined effects of insulin resistance induced by
ful therapeutic option because their hyperglycemic effect lasts glucocorticoids and inhibition of the biosynthesis and release of
for 24 hours and the risk of nocturnal hypoglycemia is relatively insulin [16]. Transient corticosteroid use is one of the most
low [2]. Second-generation sulfonylureas, such as gliclazide, common schedules of treatment, and is characterized by high
are mainstays of oral treatment due to their rapid onset of action initial doses and a gradual reduction as the underlying disease
[5]. Doses are most appropriately given at lunch time to target ameliorates. These types of glucocorticoid schedules can lead to
postprandial hyperglycemia. If hyperglycemia remains an issue initial moderate to severe hyperglycemia with rapid changes in
at a near-maximum dose, initiation of insulin should be contem- glycemia in response to changes in the glucocorticoid dose.
plated rather than the addition of an alternative oral hypoglyce- This hyperglycemia is temporary according to the duration of
mic agent as they act too slowly to be beneficial in this circum- corticosteroid treatment. Therefore, the ideal hypoglycemic
stance. At this point, it would be appropriate to discuss treat- drug to address these situations should be potent, immediate-
ment options with a team of diabetes specialists [5]. However, acting, and with an unlimited hypoglycemic action in order to
the prolonged duration of action of these agents may increase control this rapid-onset hyperglycemia. In general, the effective-
the risk of hypoglycemia when short-term and tapering doses of ness of the various oral hypoglycemic drugs in the treatment of
glucocorticoids are administered. Shorter-acting agents, such as new-onset glucocorticoid-induced hyperglycemia is limited [2].
meglinitide, nateglinide, or repaglinide, might be suitable in this This is due to their limited hypoglycemic power, which does not
regard [30]. Glinides allow minimal dose titration and have an allow for the correction of hyperglycemia in many circumstanc-
immediate onset of action and a short effect duration, which al- es. Most available oral hypoglycemic drugs have a slow onset
lows adaptation to the hyperglycemic profile of the corticoste- of action and/or a very limited or nonexistent ability to be titrat-
roids and reduces the risk of hypoglycemia in the morning, co- ed, leaving little capacity to adapt to major changes in the re-
inciding with the disappearance of the hyperglycemic action of quirements of hypoglycemic action. Furthermore, the action
corticosteroids [2]. For patients with mild hyperglycemia who profile of oral hypoglycemic drugs throughout the day does not
are unable or unwilling to perform injections of insulin, a trial usually coincide with the pattern of glucocorticoid-induced hy-
of short-acting secretagogues such as nateglinide or repaglinide perglycemia. We should also note that many situations that re-
taken before meals could be considered [5]. Their major disad- quire corticosteroids represent a formal contraindication to the
vantage is the requirement for multiple daily doses. use of oral hypoglycemic drugs. Patients with glycemia >200
mg/dL are frequently encountered and, in these cases, insulin is
INCRETIN MIMETICS generally the drug of choice. In all other situations, insulin is
usually the treatment of choice because of its efficacy and safe-
Incretin-based therapies, such as GLP-1 receptor agonists and ty. It provides an immediate onset of action, an unlimited hypo-
DPP-4 inhibitors control glucose levels by stimulating insulin glycemic power, and can be easily titrated. The different types

Copyright © 2017 Korean Endocrine Society www.e-enm.org  185

 Suh S, et al.

Table 3. Estimation of the Initial Dose of Insulin in Glucocorti- MANAGEMENT OF HYPERGLYCEMIA IN

coid-Induced Hyperglycemia, According to the Type and Dose PRE-EXISTING DIABETES
of Glucocorticoids
Prednisone dose, Dexamethasone dose, Insulin NPH, glargine/
Glucocorticoids cause hyperglycemia in patients with or with-
mg/day mg/day detemir dose, IU/kg/day out pre-existing diabetes, but it remains to be determined how
≥40 ≥8 0.4 glucocorticoid-induced hyperglycemia in such patients should
30 6 0.3 be managed and what the effects of management are. The con-
20 4 0.2 trol goals are those recommended for most patients with diabe-
10 2 0.1 tes mellitus: preprandial glycemia <130 mg/dL, postprandial
glycemia <180 mg/dL, and HbA1c <7% without hypoglyce-
NPH, neutral protamine Hagedorn.
mia [26,27]. Patients who are diet-controlled may require an
oral hypoglycemic agent, similarly to patients at a high risk of
of insulin available allow clinicians to prepare administration developing GIDM. Patients who are already on an oral hypogly-
schedules with hypoglycemic effects appropriate for different cemic regimen may require dose intensification, addition of an
corticosteroid schedules. The choice of the type and schedule of alternative agent (with preference given to sulfonylureas and
insulin should be determined by the previous hypoglycemic up-titrating the dose to the maximum), or initiation of insulin if
schedules of the patient and the type and schedule of corticoste- despite these measures, blood glucose levels persistently remain
roid established. The initial insulin dose and increments in insu- >180 mg/dL. However, there is a risk of hypoglycemia when
lin dosing should take into account the main variables that in- corticosteroid doses are reduced, and this risk is very high when
crease insulin resistance, which are the patient’s weight and the a nocturnal dose of corticosteroids is removed. This therapeutic
dose of corticosteroid administered (Table 3) [30,43]. The algo- strategy can also be applied in T2DM patients who are well con-
rithm for the management of corticosteroid-induced hypergly- trolled with lifestyle measures or with oral hypoglycemic drugs
cemia that we propose takes into account all these aspects. Basal that exacerbate their hyperglycemia during corticosteroid treat-
bolus insulin therapy remains the most flexible option for pa- ment. In these cases, the hypoglycemic drug selected would be
tients, and includes three components: basal insulin, prandial in- added to the standard therapy, always considering drugs with
sulin, and a supplemental correction-factor insulin [44]. Both synergistic effects with previous treatments and accounting for
prandial and basal insulin regimens and premixed insulin are contraindications [2]. On starting insulin, the dose of sulfonyl-
equally effective [11]. The conventional use of long-acting basal ureas should be reduced to prevent hypoglycemia, while other
insulin with traditional weight-based dosing may cause noctur- oral hypoglycemic agents can be continued [5]. In patients pre-
nal hypoglycemia [31]. More studies exploring the dose titra- viously treated with insulin, the usual schedule should be modi-
tion of insulin in patients on glucocorticoids possibly utilizing fied. The increment in the usual daily insulin dose will be esti-
technology such as continuous glucose monitoring systems are mated considering the patient’s body weight, the preparation,
needed. In general, however, a judicious step would be to time and the glucocorticoid [30,43]. The calculated dose increment is
glucocorticoid administration to a midday or an evening meal added to the patient’s usual total dose, distributing it in different
with concomitant administration of intermediate-acting insulin. administrations of insulin according to the usual insulin sched-
Subsequently, the insulin dose can be adjusted according to cap- ule and the glucocorticoid schedule [45]. Re-evaluation will
illary glycemia with increases and/or decreases of around 20%. need to take place daily as the effect of glucocorticoids on gly-
Changes in the dosage of glucocorticoids require parallel and cemic control is cumulative. In patients on a basal-bolus regi-
proportional adjustments of the insulin dose [5,30]. In the out- men (4 to 5 doses per day), it may be necessary to change the
patient setting, it is essential to instruct the patient and/or the pa- ratio from 50% basal and 50% bolus to 30% basal and 70% bo-
tient’s family regarding how to adjust the dose of insulin ac- lus, or to keep the same basal dose and increase the meal insulin
cording to glycemia and changes in the dose of glucocorticoids. [5]. In patients only on basal insulin with suboptimal glycemic
Once insulin therapy is initiated, it is recommended to adjust the control, the addition of rapid-acting insulin analogs with lunch
basal insulin dose daily based on the glucose pattern and total and evening meals may be required [2]. For patients with persis-
amount of correction insulin in the past 24 hours [11]. tent significant hyperglycemia with glucose levels >300 mg/dL,
more aggressive insulin therapy is indicated, such as intrave-

186 www.e-enm.org Copyright © 2017 Korean Endocrine Society

 Glucocorticoid-Induced DM

nous insulin infusion, a higher dose of insulin, or more frequent dose, and schedule of glucocorticoid used. Treatment should
correction with rapid-acting insulin [46]. then be adjusted based on capillary glycemia and on changes in
the dose of glucocorticoids. Moreover, it is essential to instruct
PATIENT EDUCATION the patient and/or the patient’s family about how to perform
these adjustments. Further investigation into the precise mecha-
It should be ensured that all patients with diabetes and high-risk nism of glucocorticoid-induced insulin resistance will provide
patients for GIDM have access to blood glucose monitoring, in insights for future diabetes prevention efforts and targeted thera-
order to prevent the development of hyperglycemic emergen- pies.
cies [5]. All patients should be educated on typical hyperglyce-
mic symptoms that should prompt them to check their capillary CONFLICTS OF INTEREST
blood glucose. Once the doses of glucocorticoids are being ta-
pered down, the dosage of oral hypoglycemic agents or insulin No potential conflict of interest relevant to this article was re-
should be appropriately reduced to prevent hypoglycemia. The ported.
appropriate follow-up of patients with previously unknown dia-
betes includes an HbA1c test 12 weeks following completion of ACKNOWLEDGMENT
glucocorticoid therapy to re-assess their diabetes status [17]. Pa-
tients on agents that can cause hypoglycemia need to check This work was supported by the Dong-A University research
their blood glucose levels more frequently than usual 1 to 3 fund.
days after a glucocorticoid dose reduction, because it may take
this amount of time for the glycemic effect of the glucocorticoid ORCID
to diminish and for them to adjust their diabetes medication to
the appropriate dose [47]. Sunghwan Suh https://orcid.org/0000-0001-6865-966X

CONCLUSIONS REFERENCES

GIDM is a common and potentially harmful problem in clinical 1. Fathallah N, Slim R, Larif S, Hmouda H, Ben Salem C. Drug-
practice, affecting almost all medical specialties. However, its induced hyperglycaemia and diabetes. Drug Saf 2015;
diagnosis and treatment are surprisingly undervalued by most 38:1153-68.
healthcare professionals. The knowledge that hyperglycemia 2. Perez A, Jansen-Chaparro S, Saigi I, Bernal-Lopez MR,
leads to longer hospital stays, delayed wound healing, increased Minambres I, Gomez-Huelgas R. Glucocorticoid-induced
infections, and higher mortality rates should encourage health- hyperglycemia. J Diabetes 2014;6:9-20.
care providers to monitor its occurrence and to manage it appro- 3. Hwang JL, Weiss RE. Steroid-induced diabetes: a clinical
priately. Similarly to T2DM in general, the principles of early and molecular approach to understanding and treatment. Di-
detection and risk factor modification apply. Screening for abetes Metab Res Rev 2014;30:96-102.
GIDM should be considered in all patients (especially in those 4. Conn JW, Fajans SS. Influence of adrenal cortical steroids
with risk factors) treated with medium to high doses of gluco- on carbohydrate metabolism in man. Metabolism 1956;5:
corticoids. Diagnosing impaired fasting glucose or impaired 114-27.
glucose tolerance prior to the initiation of chronic glucocorti- 5. Mills E, Devendra S. Steroid-induced hyperglycaemia in
coids will better identify those who would benefit from gluco- primary care. London J Prim Care (Abingdon) 2015;7:103-
corticoid-sparing treatment, or if this is not an option, blood 6.
glucose monitoring while starting therapy. Together with, or af- 6. Blackburn D, Hux J, Mamdani M. Quantification of the risk
ter, lifestyle measures, hypoglycemic drugs with important in- of corticosteroid-induced diabetes mellitus among the elder-
sulin-sensitizing effects are indicated. Other oral hypoglycemic ly. J Gen Intern Med 2002;17:717-20.
drugs or insulin therapy can be considered to be the drug of 7. Donihi AC, Raval D, Saul M, Korytkowski MT, DeVita
choice. Treatment, in most cases with insulin, should take into MA. Prevalence and predictors of corticosteroid-related hy-
account the degree and pattern of hyperglycemia and the type, perglycemia in hospitalized patients. Endocr Pract 2006;12:

Copyright © 2017 Korean Endocrine Society www.e-enm.org  187

 Suh S, et al.

358-62. than insulin signal transduction. Diabetes 2000;49:1700-8.

8. Kasiske BL, Snyder JJ, Gilbertson D, Matas AJ. Diabetes 21. Weinstein SP, Wilson CM, Pritsker A, Cushman SW. Dexa-
mellitus after kidney transplantation in the United States. methasone inhibits insulin-stimulated recruitment of GLUT4
Am J Transplant 2003;3:178-85. to the cell surface in rat skeletal muscle. Metabolism 1998;
9. Kim SY, Yoo CG, Lee CT, Chung HS, Kim YW, Han SK, et 47:3-6.
al. Incidence and risk factors of steroid-induced diabetes in 22. Vegiopoulos A, Herzig S. Glucocorticoids, metabolism and
patients with respiratory disease. J Korean Med Sci 2011;26: metabolic diseases. Mol Cell Endocrinol 2007;275:43-61.
264-7. 23. Boden G, Shulman GI. Free fatty acids in obesity and type 2
10. Gulliford MC, Charlton J, Latinovic R. Risk of diabetes as- diabetes: defining their role in the development of insulin
sociated with prescribed glucocorticoids in a large popula- resistance and beta-cell dysfunction. Eur J Clin Invest 2002;
tion. Diabetes Care 2006;29:2728-9. 32 Suppl 3:14-23.
11. Kwon S, Hermayer KL. Glucocorticoid-induced hypergly- 24. Delaunay F, Khan A, Cintra A, Davani B, Ling ZC, Anders-
cemia. Am J Med Sci 2013;345:274-7. son A, et al. Pancreatic beta cells are important targets for the
12. Panthakalam S, Bhatnagar D, Klimiuk P. The prevalence diabetogenic effects of glucocorticoids. J Clin Invest 1997;
and management of hyperglycaemia in patients with rheu- 100:2094-8.
matoid arthritis on corticosteroid therapy. Scott Med J 2004; 25. van Raalte DH, Nofrate V, Bunck MC, van Iersel T, Elas-
49:139-41. saiss Schaap J, Nassander UK, et al. Acute and 2-week ex-
13. Liu XX, Zhu XM, Miao Q, Ye HY, Zhang ZY, Li YM. Hy- posure to prednisolone impair different aspects of beta-cell
perglycemia induced by glucocorticoids in nondiabetic pa- function in healthy men. Eur J Endocrinol 2010;162:729-35.
tients: a meta-analysis. Ann Nutr Metab 2014;65:324-32. 26. American Diabetes Association. 14. Diabetes care in the
14. Gurwitz JH, Bohn RL, Glynn RJ, Monane M, Mogun H, hospital. Diabetes Care 2017;40(Suppl 1):S120-7.
Avorn J. Glucocorticoids and the risk for initiation of hypo- 27. Umpierrez GE, Hellman R, Korytkowski MT, Kosiborod M,
glycemic therapy. Arch Intern Med 1994;154:97-101. Maynard GA, Montori VM, et al. Management of hypergly-
15. Katsuyama T, Sada KE, Namba S, Watanabe H, Katsuyama cemia in hospitalized patients in non-critical care setting: an
E, Yamanari T, et al. Risk factors for the development of endocrine society clinical practice guideline. J Clin Endocri-
glucocorticoid-induced diabetes mellitus. Diabetes Res Clin nol Metab 2012;97:16-38.
Pract 2015;108:273-9. 28. Baldwin D, Apel J. Management of hyperglycemia in hospi-
16. Geer EB, Islam J, Buettner C. Mechanisms of glucocorti- talized patients with renal insufficiency or steroid-induced
coid-induced insulin resistance: focus on adipose tissue func- diabetes. Curr Diab Rep 2013;13:114-20.
tion and lipid metabolism. Endocrinol Metab Clin North Am 29. American Diabetes Association. 2. Classification and diag-
2014;43:75-102. nosis of diabetes. Diabetes Care 2017;40(Suppl 1):S11-24.
17. Brady VJ, Grimes D, Armstrong T, LoBiondo-Wood G. 30. Clore JN, Thurby-Hay L. Glucocorticoid-induced hypergly-
Management of steroid-induced hyperglycemia in hospital- cemia. Endocr Pract 2009;15:469-74.
ized patients with cancer: a review. Oncol Nurs Forum 2014; 31. Burt MG, Roberts GW, Aguilar-Loza NR, Frith P, Stranks
41:E355-65. SN. Continuous monitoring of circadian glycemic patterns
18. Pagano G, Cavallo-Perin P, Cassader M, Bruno A, Ozzello A, in patients receiving prednisolone for COPD. J Clin Endo-
Masciola P, et al. An in vivo and in vitro study of the mecha- crinol Metab 2011;96:1789-96.
nism of prednisone-induced insulin resistance in healthy sub- 32. Valderhaug TG, Jenssen T, Hartmann A, Midtvedt K, Hold-
jects. J Clin Invest 1983;72:1814-20. aas H, Reisaeter AV, et al. Fasting plasma glucose and gly-
19. van Raalte DH, Ouwens DM, Diamant M. Novel insights cosylated hemoglobin in the screening for diabetes mellitus
into glucocorticoid-mediated diabetogenic effects: towards after renal transplantation. Transplantation 2009;88:429-34.
expansion of therapeutic options? Eur J Clin Invest 2009;39: 33. Imatoh T, Sai K, Hori K, Segawa K, Kawakami J, Kimura M,
81-93. et al. Development of a novel algorithm for detecting gluco-
20. Sakoda H, Ogihara T, Anai M, Funaki M, Inukai K, Katagiri corticoid-induced diabetes mellitus using a medical infor-
H, et al. Dexamethasone-induced insulin resistance in 3T3- mation database. J Clin Pharm Ther 2017;42:215-20.
L1 adipocytes is due to inhibition of glucose transport rather 34. Gonzalez-Gonzalez JG, Mireles-Zavala LG, Rodriguez-

188 www.e-enm.org Copyright © 2017 Korean Endocrine Society

 Glucocorticoid-Induced DM

Gutierrez R, Gomez-Almaguer D, Lavalle-Gonzalez FJ, Ta- MW, Pouwels PJ, Ouwens DM, et al. Does dipeptidyl pepti-
mez-Perez HE, et al. Hyperglycemia related to high-dose dase-4 inhibition prevent the diabetogenic effects of gluco-
glucocorticoid use in noncritically ill patients. Diabetol corticoids in men with the metabolic syndrome? A random-
Metab Syndr 2013;5:18. ized controlled trial. Eur J Endocrinol 2014;170:429-39.
35. Suh S, Kim JH. Glycemic variability: how do we measure it 42. van Raalte DH, van Genugten RE, Linssen MM, Ouwens
and why is it important? Diabetes Metab J 2015;39:273-82. DM, Diamant M. Glucagon-like peptide-1 receptor agonist
36. Seelig E, Meyer S, Timper K, Nigro N, Bally M, Pernicova I, treatment prevents glucocorticoid-induced glucose intoler-
et al. Metformin prevents metabolic side effects during sys- ance and islet-cell dysfunction in humans. Diabetes Care
temic glucocorticoid treatment. Eur J Endocrinol 2017;176: 2011;34:412-7.
349-58. 43. Saigi I, Perez A. Management of glucocorticoid induced hy-
37. Willi SM, Kennedy A, Brant BP, Wallace P, Rogers NL, perglycemia. Rev Clin Esp 2010;210:397-403.
Garvey WT. Effective use of thiazolidinediones for the treat- 44. Burt MG, Drake SM, Aguilar-Loza NR, Esterman A,
ment of glucocorticoid-induced diabetes. Diabetes Res Clin Stranks SN, Roberts GW. Efficacy of a basal bolus insulin
Pract 2002;58:87-96. protocol to treat prednisolone-induced hyperglycaemia in
38. Luther P, Baldwin D Jr. Pioglitazone in the management of hospitalised patients. Intern Med J 2015;45:261-6.
diabetes mellitus after transplantation. Am J Transplant 45. Grommesh B, Lausch MJ, Vannelli AJ, Mullen DM, Ber-
2004;4:2135-8. genstal RM, Richter SA, et al. Hospital insulin protocol
39. Vu K, Busaidy N, Cabanillas ME, Konopleva M, Faderl S, aims for glucose control in glucocorticoid-induced hyper-
Thomas DA, et al. A randomized controlled trial of an inten- glycemia. Endocr Pract 2016;22:180-9.
sive insulin regimen in patients with hyperglycemic acute 46. Hawkins K, Donihi AC, Korytkowski MT. Glycemic man-
lymphoblastic leukemia. Clin Lymphoma Myeloma Leuk agement in medical and surgical patients in the non-ICU
2012;12:355-62. setting. Curr Diab Rep 2013;13:96-106.
40. Yanai H, Masui Y, Yoshikawa R, Kunimatsu J, Kaneko H. 47. Lansang MC, Hustak LK. Glucocorticoid-induced diabetes
Dipeptidyl peptidase-4 inhibitor for steroid-induced diabe- and adrenal suppression: how to detect and manage them.
tes. World J Diabetes 2010;1:99-100. Cleve Clin J Med 2011;78:748-56.
41. van Genugten RE, van Raalte DH, Muskiet MH, Heymans

Copyright © 2017 Korean Endocrine Society www.e-enm.org  189