1. Reviews/Consensus Reports/ADA Statements
P O S I T I O N S T A T E M E N T
Management of Hyperglycemia in Type 2
Diabetes: A Patient-Centered Approach
Position Statement of the American Diabetes Association (ADA) and
the European Association for the Study of Diabetes (EASD)
SILVIO E. INZUCCHI, MD1 MICHAEL NAUCK, MD6 These recommendations should be
RICHARD M. BERGENSTAL, MD2 ANNE L. PETERS, MD7 considered within the context of the needs,
JOHN B. BUSE, MD, PHD3 APOSTOLOS TSAPAS, MD, PHD8 preferences, and tolerances of each patient;
MICHAELA DIAMANT, MD, PHD4 RICHARD WENDER, MD9 individualization of treatment is the cor-
ELE FERRANNINI, MD5 DAVID R. MATTHEWS, MD, DPHIL10,11,12
nerstone of success. Our recommenda-
tions are less prescriptive than and not as
algorithmic as prior guidelines. This fol-
G
lycemic management in type 2 di- information on the benefits/risks of glycemic lows from the general lack of comparative-
abetes mellitus has become increas- control, recent evidence concerning efficacy effectiveness research in this area. Our
ingly complex and, to some extent, and safety of several new drug classes intent is therefore to encourage an appre-
controversial, with a widening array of (16,17), the withdrawal/restriction of others, ciation of the variable and progressive
pharmacological agents now available (1–5), and increasing calls for a move toward more nature of type 2 diabetes, the specific role
mounting concerns about their potential patient-centered care (18,19). of each drug, the patient and disease
adverse effects and new uncertainties re- This statement has been written in- factors that drive clinical decision making
garding the benefits of intensive glycemic corporating the best available evidence (20–23), and the constraints imposed by
control on macrovascular complications and, where solid support does not exist, age and comorbidity (4,6). The implemen-
(6–9). Many clinicians are therefore per- using the experience and insight of the tation of these guidelines will require
plexed as to the optimal strategies for their writing group, incorporating an extensive thoughtful clinicians to integrate current
patients. As a consequence, the American review by additional experts (acknowl- evidence with other constraints and im-
Diabetes Association (ADA) and the European edged below). The document refers to peratives in the context of patient-specific
Association for the Study of Diabetes (EASD) glycemic control; yet this clearly needs to factors.
convened a joint task force to examine the be pursued within a multifactorial risk
evidence and develop recommendations for reduction framework. This stems from the PATIENT-CENTERED
antihyperglycemic therapy in nonpregnant fact that patients with type 2 diabetes are at APPROACHdEvidence-based advice
adults with type 2 diabetes. Several guide- increased risk of cardiovascular morbidity depends on the existence of primary
line documents have been developed by and mortality; the aggressive management source evidence. This emerges only
members of these two organizations (10) of cardiovascular risk factors (blood pres- from clinical trial results in highly selected
and by other societies and federations sure and lipid therapy, antiplatelet treat- patients, using limited strategies. It does
(2,11–15). However, an update was ment, and smoking cessation) is likely to not address the range of choices available,
deemed necessary because of contemporary have even greater benefits. or the order of use of additional therapies.
Even if such evidence were available, the
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c data would show median responses and
From the 1Section of Endocrinology, Yale University School of Medicine and Yale-New Haven Hospital, not address the vital question of who
New Haven, Connecticut; the 2International Diabetes Center at Park Nicollet, Minneapolis, Minnesota; the responded to which therapy and why (24).
3
Division of Endocrinology, University of North Carolina School of Medicine, Chapel Hill, North Carolina; Patient-centered care is defined as an ap-
the 4Diabetes Center/Department of Internal Medicine, VU University Medical Center, Amsterdam, the
Netherlands; the 5Department of Medicine, University of Pisa School of Medicine, Pisa, Italy; 6Diabeteszentrum
proach to “providing care that is respectful
Bad Lauterberg, Bad Lauterberg im Harz, Germany; the 7Division of Endocrinology, Keck School of Medicine, of and responsive to individual patient
University of Southern California, Los Angeles, California; the 8Second Medical Department, Aristotle Uni- preferences, needs, and values and ensur-
versity Thessaloniki, Thessaloniki, Greece; the 9Department of Family and Community Medicine, Jefferson ing that patient values guide all clinical de-
Medical College, Thomas Jefferson University, Philadelphia, Pennsylvania; the 10Oxford Centre for Diabetes, cisions” (25). This should be the organizing
Endocrinology and Metabolism, Churchill Hospital, Headington, Oxford, U.K.; the 11National Institute for
Health Research (NIHR), Oxford Biomedical Research Centre, Oxford, U.K.; and the 12Harris Manchester
principle underlying health care for indi-
College, University of Oxford, Oxford, U.K. viduals with any chronic disease, but given
Corresponding author: Silvio E. Inzucchi, silvio.inzucchi@yale.edu. our uncertainties in terms of choice or se-
DOI: 10.2337/dc12-0413 quence of therapy, it is particularly appro-
This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10 priate in type 2 diabetes. Ultimately, it is
.2337/dc12-0413/-/DC1.
S.E. Inzucchi and D.R. Matthews were co-chairs for the Position Statement Writing Group. R.M. Bergenstal, patients who make the final decisions re-
J.B. Buse, A.L. Peters, and R. Wender were the Writing Group for the ADA. M. Diamant, E. Ferrannini, garding their lifestyle choices and, to some
M. Nauck, and A. Tsapas were the Writing Group for the EASD. degree, the pharmaceutical interventions
This article is being simultaneously published in 2012 in Diabetes Care and Diabetologia by the American they use; their implementation occurs in
Diabetes Association and the European Association for the Study of Diabetes.
© 2012 by the American Diabetes Association and Springer-Verlag. Readers may use this article as long as the the context of the patients’ real lives and
work is properly cited, the use is educational and not for profit, and the work is not altered. See http:// relies on the consumption of resources
creativecommons.org/licenses/by-nc-nd/3.0/ for details. (both public and private).
care.diabetesjournals.org DIABETES CARE 1
Diabetes Care Publish Ahead of Print, published online April 19, 2012

2. Position Statement
Patient involvement in the medical or insulin, with a subset of overweight were suggestions in these trials that patients
decision making constitutes one of the patients randomized to metformin. The without overt CVD, with shorter duration
core principles of evidence-based medi- overall HbA1c achieved was 0.9% lower of disease, and lower baseline HbA 1c,
cine, which mandates the synthesis of best in the intensive policy group compared benefited from the more intensive strat-
available evidence from the literature with with the conventional policy arm (7.0% egies. Modest improvements in some
the clinician’s expertise and patient’s own vs. 7.9%). Associated with this difference in microvascular end points in the studies
inclinations (26). During the clinical encoun- glycemic control was a reduction in the risk were likewise demonstrated. Finally, a
ter, the patient’s preferred level of involve- of microvascular complications (retinopa- meta-analysis of cardiovascular out-
ment should be gauged and therapeutic thy, nephropathy, neuropathy) with inten- comes in these trials suggested that every
choices explored, potentially with the uti- sive therapy. A trend toward reduced rates HbA1c reduction of ;1% may be associ-
lization of decision aids (21). In a shared of myocardial infarction in this group did ated with a 15% relative risk reduction in
decision-making approach, clinician and not reach statistical significance (30). By nonfatal myocardial infarction, but
patient act as partners, mutually exchanging contrast, substantially fewer metformin- without benefits on stroke or all-cause
information and deliberating on options, in treated patients experienced myocardial mortality (36).
order to reach a consensus on the therapeu- infarction, diabetes-related and all-cause
tic course of action (27). There is good ev- mortality (32), despite a mean HbA1c only Overview of the pathogenesis of
idence supporting the effectiveness of this 0.6% lower than the conventional policy type 2 diabetes
approach (28). Importantly, engaging pa- group. The UKPDS 10-year follow-up Any rise in glycemia is the net result of
tients in health care decisions may enhance demonstrated that the relative benefit of glucose influx exceeding glucose outflow
adherence to therapy. having been in the intensive management from the plasma compartment. In the fast-
policy group was maintained over a de- ing state, hyperglycemia is directly related
BACKGROUND cade, resulting in the emergence of statisti- to increased hepatic glucose production.
cally significant benefits on cardiovascular In the postprandial state, further glucose
Epidemiology and health care disease (CVD) end points and total mortality excursions result from the combination
impact in those initially assigned to sulfonylurea/ of insufficient suppression of this glucose
Both the prevalence and incidence of type 2 insulin, and persistence of CVD benefits output and defective insulin stimulation
diabetes are increasing worldwide, particu- with metformin (33), in spite of the fact of glucose disposal in target tissues, mainly
larly in developing countries, in conjunction that the mean HbA1c levels between the skeletal muscle. Once the renal tubular
with increased obesity rates and westerni- groups converged soon after the ran- transport maximum for glucose is excee-
zation of lifestyle. The attendant economic domized component of the trial had ded, glycosuria curbs, though does not
burden for health care systems is skyrocket- concluded. prevent, further hyperglycemia.
ing, owing to the costs associated with treat- In 2008, three shorter-term studies Abnormal islet cell function is a key
ment and diabetes complications. Type 2 [Action to Control Cardiovascular Risk in and requisite feature of type 2 diabetes. In
diabetes remains a leading cause of car- Diabetes (ACCORD) (34), Action in Dia- early disease stages, insulin production is
diovascular disorders, blindness, end-stage betes and Vascular Disease: Preterax and normal or increased in absolute terms,
renal failure, amputations, and hospitaliza- Diamicron Modified-Release Controlled but disproportionately low for the degree
tions. It is also associated with increased risk Evaluation (ADVANCE) (35), Veterans Af- of insulin sensitivity, which is typically
of cancer, serious psychiatric illness, cogni- fairs Diabetes Trial (VADT) (36)] reported reduced. However, insulin kinetics, such
tive decline, chronic liver disease, acceler- the effects of two levels of glycemic control as the ability of the pancreatic b-cell to
ated arthritis, and other disabling or deadly on cardiovascular end points in middle- release adequate hormone in phase with
conditions. Effective management strategies aged and older individuals with well- rising glycemia, are profoundly compro-
are of obvious importance. established type 2 diabetes at high risk for mised. This functional islet incompetence
cardiovascular events. ACCORD and VADT is the main quantitative determinant of
Relationship of glycemic control aimed for an HbA1c ,6.0% using complex hyperglycemia (37) and progresses over
to outcomes combinations of oral agents and insulin. time. In addition, in type 2 diabetes, pan-
It is well established that the risk of micro- ADVANCE aimed for an HbA1c #6.5% creatic a-cells hypersecrete glucagon, fur-
vascular and macrovascular complications using a less intensive approach based on ther promoting hepatic glucose production
is related to glycemia, as measured by the sulfonylurea gliclazide. None of the (38). Importantly, islet dysfunction is not
HbA1c; this remains a major focus of ther- trials demonstrated a statistically signif- necessarily irreversible. Enhancing insulin
apy (29). Prospective randomized trials icant reduction in the primary combined action relieves b-cell secretory burden, and
have documented reduced rates of micro- cardiovascular end points. Indeed, in any intervention that improves glycemiad
vascular complications in type 2 diabetic ACCORD, a 22% increase in total mortality from energy restriction to, most strikingly,
patients treated to lower glycemic targets. with intensive therapy was observed, bariatric surgerydcan ameliorate b-cell
In the UK Prospective Diabetes Study mainly driven by cardiovascular mortality. dysfunction to an extent (39). More re-
(UKPDS) (30,31), patients with newly di- An explanation for this finding has re- cently recognized abnormalities in the in-
agnosed type 2 diabetes were randomized mained elusive, although rates of hypogly- cretin system (represented by the gut
to two treatment policies. In the standard cemia were threefold higher with intensive hormones, glucagon-like peptide 1 [GLP-1],
group, lifestyle intervention was the main- treatment. It remains unclear, however, if and glucose-dependent insulinotropic
stay with pharmacological therapy used hypoglycemia was responsible for the ad- peptide [GIP]) are also found in type 2
only if hyperglycemia became severe. In the verse outcomes, or if other factors, such as diabetes, but it remains unclear whether
more intensive treatment arm, patients were more weight gain, or simply the greater these constitute primary or secondary de-
randomly assigned to either a sulfonylurea complexity of therapy, contributed. There fects (40). In most patients with type 2
2 DIABETES CARE care.diabetesjournals.org

3. Inzucchi and Associates
diabetes, especially the obese, insulin re-
sistance in target tissues (liver, muscle,
adipose tissue, myocardium) is a promi-
nent feature. This results in both glucose
overproduction and underutilization.
Moreover, an increased delivery of fatty acids
to the liver favors their oxidation, which
contributes to increased gluconeogenesis,
whereas the absolute overabundance of lip-
ids promotes hepatosteatosis (41).
Antihyperglycemic agents are directed
at one or more of the pathophysiological
defects of type 2 diabetes, or modify
physiological processes relating to appetite
or to nutrient absorption or excretion.
Ultimately, type 2 diabetes is a disease
that is heterogeneous in both pathogenesis
and in clinical manifestationda point to be
considered when determining the optimal
therapeutic strategy for individual patients.
ANTIHYPERGLYCEMIC
THERAPY
Glycemic targets
The ADA’s “Standards of Medical Care in
Diabetes” recommends lowering HbA1c
to ,7.0% in most patients to reduce the Figure 1dDepiction of the elements of decision making used to determine appropriate efforts to
incidence of microvascular disease (42). achieve glycemic targets. Greater concerns about a particular domain are represented by in-
This can be achieved with a mean plasma creasing height of the ramp. Thus, characteristics/predicaments toward the left justify more
glucose of ;8.3–8.9 mmol/L (;150–160 stringent efforts to lower HbA1c, whereas those toward the right are compatible with less
mg/dL); ideally, fasting and premeal glu- stringent efforts. Where possible, such decisions should be made in conjunction with the patient,
cose should be maintained at ,7.2 mmol/L reflecting his or her preferences, needs, and values. This “scale” is not designed to be applied
(,130 mg/dL) and the postprandial glu- rigidly but to be used as a broad construct to help guide clinical decisions. Adapted with per-
cose at ,10 mmol/L (,180 mg/dL). mission from Ismail-Beigi et al. (20).
More stringent HbA1c targets (e.g., 6.0–
6.5%) might be considered in selected
patients (with short disease duration, long patient should also be considered, since general diabetes education (individual or
life expectancy, no significant CVD) if this the achievement of any degree of glucose group, preferably using an approved cur-
can be achieved without significant hypo- control requires active participation and riculum), with a specific focus on dietary
glycemia or other adverse effects of treat- commitment (19,23,45,46). Indeed, any interventions and the importance of in-
ment (20,43). Conversely, less stringent target could reflect an agreement between creasing physical activity. While encourag-
HbA 1c goalsde.g., 7.5–8.0% or even patient and clinician. An important related ing therapeutic lifestyle change is important
slightly higherdare appropriate for pa- concept is that the ease with which more at diagnosis, periodic counseling should
tients with a history of severe hypoglycemia, intensive targets are reached influences also be integrated into the treatment
limited life expectancy, advanced complica- treatment decisions; logically, lower tar- program.
tions, extensive comorbid conditions and gets are attractive if they can be achieved Weight reduction, achieved through
those in whom the target is difficult to attain with less complex regimens and no or dietary means alone or with adjunctive
despite intensive self-management educa- minimal adverse effects. Importantly, uti- medical or surgical intervention, improves
tion, repeated counseling, and effective lizing the percentage of diabetic patients glycemic control and other cardiovascular
doses of multiple glucose-lowering agents, who are achieving an HbA1c ,7.0% as a risk factors. Modest weight loss (5–10%) con-
including insulin (20,44). quality indicator, as promulgated by vari- tributes meaningfully to achieving improved
The accumulated results from the ous health care organizations, is inconsis- glucose control. Accordingly, establishing a
aforementioned type 2 diabetes cardio- tent with the emphasis on individualization goal of weight reduction, or at least weight
vascular trials suggest that not everyone of treatment goals. maintenance, is recommended.
benefits from aggressive glucose man- Dietary advice must be personalized
agement. It follows that it is important to Therapeutic options (49). Patients should be encouraged to eat
individualize treatment targets (5,34–36). Lifestyle. Interventions designed to im- healthy foods that are consistent with the
The elements that may guide the clinician pact an individual’s physical activity lev- prevailing population-wide dietary rec-
in choosing an HbA1c target for a specific els and food intake are critical parts of ommendations and with an individual’s
patient are shown in Fig. 1. As mentioned type 2 diabetes management (47,48). All preferences and culture. Foods high in fiber
earlier, the desires and values of the patients should receive standardized (such as vegetables, fruits, whole grains, and
care.diabetesjournals.org DIABETES CARE 3

4. 4
Table 1dProperties of currently available glucose-lowering agents that may guide treatment choice in individual patients with type 2 diabetes mellitus
Primary physiological
Class Compound(s) Cellular mechanism action(s) Advantages Disadvantages Cost
Biguanides c Metformin Activates AMP-kinase c ↓ Hepatic glucose c Extensive experience c Gastrointestinal side effects Low
DIABETES CARE
production c No weight gain (diarrhea, abdominal
Position Statement
cramping)
c No hypoglycemia
c Lactic acidosis risk (rare)
c Likely ↓ CVD events
c Vitamin B12 deficiency
(UKPDS)
c Multiple contraindications:
CKD, acidosis, hypoxia,
dehydration, etc.
Sulfonylureas 2nd generation Closes KATP channels c ↑ Insulin secretion c Extensive experience c Hypoglycemia Low
c Glyburide/ on b-cell plasma c ↓ Microvascular risk c Weight gain
glibenclamide membranes (UKPDS) c ? Blunts myocardial ischemic
c Glipizide preconditioning
c Gliclazideb c Low durability
c Glimepiride
Meglitinides c Repaglinide Closes KATP channels c ↑ Insulin secretion c ↓ Postprandial glucose c Hypoglycemia High
(glinides) c Nateglinide on b-cell plasma excursions c Weight gain
membranes c Dosing flexibility c ? Blunts myocardial ischemic
preconditioning
c Frequent dosing schedule
Thiazolidinediones c Pioglitazone Activates the nuclear c ↑ Insulin sensitivity c No hypoglycemia c Weight gain Highe
c Rosiglitazonec transcription factor c Durability c Edema/heart failure
PPAR-g c Bone fractures
c ↑ HDL-C
c ↓ Triglycerides c ↑ LDL-C (rosiglitazone)
(pioglitazone) c ? ↑ MI (meta-analyses,
c ? ↓ CVD events rosiglitazone)
(ProACTIVE, c ? ↑ Bladder cancer
pioglitazone) (pioglitazone)
a-Glucosidase c Acarbose Inhibits intestinal c Slows intestinal c No hypoglycemia c Generally modest HbA1c Moderate
inhibitorsa c Miglitol a-glucosidase carbohydrate c ↓ Postprandial glucose efficacy
c Vogliboseb,d digestion/absorption excursions c Gastrointestinal side effects
c ? ↓ CVD events (flatulence, diarrhea)
(STOP-NIDDM) c Frequent dosing schedule
c Nonsystemic
DPP-4 inhibitors c Sitagliptin Inhibits DPP-4 activity, c ↑ Insulin secretion c No hypoglycemia c Generally modest HbA1c High
c Vildagliptina increasing postprandial (glucose-dependent) c Well tolerated efficacy
c Saxagliptin active incretin (GLP-1, c ↓ Glucagon secretion c Urticaria/angioedema
c Linagliptin GIP) concentrations (glucose-dependent) c ? Pancreatitis
c Alogliptinb,d
Continued on p. 5
care.diabetesjournals.org

5. Table 1dContinued
Primary physiological
Class Compound(s) Cellular mechanism action(s) Advantages Disadvantages Cost
Bile acid c Colesevelam Binds bile acids in c Unknown c No hypoglycemia c Generally modest HbA1c High
sequestrantsa intestinal tract, c ? ↓ Hepatic glucose c ↓ LDL-C efficacy
increasing hepatic production c Constipation
care.diabetesjournals.org
bile acid production; c ? ↑ Incretin levels c ↑ Triglycerides
? activation of farnesoid c May ↓ absorption of other
X receptor (FXR) in liver medications
Dopamine-2 c Bromocriptine Activates dopaminergic c Modulates hypothalamic c No hypoglycemia c Generally modest HbA1c High
agonistsa (quick-release)d receptors regulation of metabolism c ? ↓ CVD events efficacy
c ↑ Insulin sensitivity (Cycloset Safety c Dizziness/syncope
Trial) c Nausea
c Fatigue
c Rhinitis
GLP-1 receptor c Exenatide Activates GLP-1 c ↑ Insulin secretion c No hypoglycemia c Gastrointestinal side effects High
agonists c Exenatide receptors (glucose-dependent) c Weight reduction (nausea/vomiting)
extended c ↓ Glucagon secretion c ? Potential for c ? Acute pancreatitis
release (glucose-dependent) improved b-cell c C-cell hyperplasia/medullary
c Liraglutide c Slows gastric emptying mass/function thyroid tumors in animals
c ↑ Satiety c ? Cardiovascular c Injectable
protective actions c Training requirements
Amylin mimeticsa c Pramlintided Activates amylin c ↓ Glucagon secretion c ↓ Postprandial glucose c Generally modest HbA1c High
receptors c Slows gastric emptying excursions efficacy
c ↑ Satiety c Weight reduction c Gastrointestinal side effects
(nausea/vomiting)
c Hypoglycemia unless
insulin dose is
simultaneously reduced
c Injectable
c Frequent dosing schedule
Insulins c Human NPH Activates insulin c ↑ Glucose disposal c Universally effective c Hypoglycemia Variablef
c Human Regular receptors c ↓ Hepatic glucose c Theoretically unlimited c Weight gain
c Lispro production efficacy c ? Mitogenic effects
c Aspart c ↓ Microvascular risk c Injectable
c Glulisine (UKPDS) c Training requirements
c Glargine c “Stigma” (for patients)
c Detemir
c Premixed
(several types)
a
Limited use in the U.S./Europe. bNot licensed in the U.S. cPrescribing highly restricted in the U.S.; withdrawn in Europe. dNot licensed in Europe. eTo be available as a generic product in 2012, with expected significant
reductions in cost. fDepends on type (analogs . human insulins) and dosage. CKD, chronic kidney disease; CVD, cardiovascular disease; DPP-4, dipeptidyl peptidase 4; GIP, glucose-dependent insulinotropic peptide;
DIABETES CARE
GLP-1, glucagon-like peptide 1; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol; PPAR, peroxisome proliferator–activated receptor; ProACTIVE, Prospective Pioglitazone Clinical Trial in Macrovascular Events (60);
STOP-NIDDM, Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (134); UKPDS, UK Prospective Diabetes Study (29–33).
5
Inzucchi and Associates

6. Position Statement
legumes), low-fat dairy products, and fresh production (54,55). It is generally consid- can be significant in some. A limiting side
fish should be emphasized. High-energy ered weight-neutral with chronic use and effect is nausea and vomiting, particularly
foods, including those rich in saturated does not increase the risk of hypoglycemia. early in the course of treatment. Concerns
fats, and sweet desserts and snacks should Metformin is associated with initial gastro- regarding an increased risk of pancreatitis
be eaten less frequently and in lower intestinal side effects, and caution is ad- remain unresolved. The oral dipeptidyl
amounts (50–52). Patients who eventually vised to avoid its use in patients at risk for peptidase 4 (DPP-4) inhibitors enhance
lose and keep weight off usually do so after lactic acidosis (e.g., in advanced renal in- circulating concentrations of active
numerous cycles of weight loss and relapse. sufficiency, alcoholism), a rare complica- GLP-1 and GIP (64). Their major effect
The health care team should remain non- tion of therapy. As noted earlier, there appears to be in the regulation of insulin
judgmental but persistent, revisiting and may be some cardiovascular benefits and glucagon secretion; they are weight
encouraging therapeutic lifestyle changes from this drug, but the clinical trial data neutral. Typically, neither of the incretin-
frequently, if needed. are not robust. based classes cause hypoglycemia by
As much physical activity as possible The oldest oral agent class is the sulfo- themselves.
should be promoted, ideally aiming for at nylurea insulin secretagogues. Through Two agents that are used infrequently
least 150 min/week of moderate activity the closure of ATP-sensitive potassium in the U.S. and Europe are the a-glucosidase
including aerobic, resistance, and flexi- channels on b-cells, these drugs stimulate inhibitors (AGIs), which retard gut carbohy-
bility training (53). In older individuals, insulin release (56). While effective in con- drate absorption (65), and colesevelam, a
or those with mobility challenges, so trolling glucose levels, their use is associ- bile acid sequestrant whose mechanism of
long as tolerated from a cardiovascular ated with modest weight gain and risk of glucose-lowering action remains poorly
standpoint, any increase in activity level hypoglycemia. In addition, studies have understood and whose major additional
is advantageous. demonstrated a secondary failure rate benefit is LDL-cholesterol reduction (66).
At diagnosis, highly motivated pa- that may exceed other drugs, ascribed to Both have gastrointestinal effects, mainly
tients with HbA1c already near target (e.g., an exacerbation of islet dysfunction (57). flatulence with AGIs and constipation
,7.5%) could be given the opportunity Shorter-acting secretagogues, the megliti- with colesevelam. The dopamine agonist
to engage in lifestyle change for a period nides (or glinides), stimulate insulin re- bromocriptine is only available in the U.S.
of 3–6 months before embarking on lease through similar mechanisms but as an antihyperglycemic agent (67). Its
pharmacotherapy (usually metformin). may be associated with less hypoglycemia mechanism of action and precise role
Those with moderate hyperglycemia or (58). They require more frequent dosing, are unclear. The amylin agonist, pramlintide,
in whom lifestyle changes are anticipated however. is typically reserved for patients treated
to be unsuccessful should be promptly Thiazolidinediones (TZDs) are per- with intensive insulin therapy, usually in
started on an antihyperglycemic agent oxisome proliferator–activated receptor g type 1 diabetes mellitus; it decreases post-
(also usually metformin) at diagnosis, activators (59) that improve insulin sen- prandial glucose excursions by inhibiting
which can later be modified or possibly sitivity in skeletal muscle and reduce he- glucagon secretion and slowing gastric
discontinued if lifestyle changes are suc- patic glucose production (54,55). They emptying (68).
cessful. do not increase the risk of hypoglycemia The glucose-lowering effectiveness of
Oral agents and noninsulin injectables. and may be more durable in their effective- noninsulin pharmacological agents is said
Important properties of antihyperglyce- ness than sulfonylureas and metformin to be high for metformin, sulfonylureas,
mic agents that play a role in the choice of (57). Pioglitazone appeared to have a mod- TZDs, and GLP-1 agonists (expected
drug(s) in individual patients are summa- est benefit on cardiovascular events as a HbA1c reduction ;1.0–1.5%) (1,69,70),
rized in Table 1. Ultimately, the aims of secondary outcome in one large trial in- and generally lower for meglitinides,
controlling glycemia are to avoid acute volving patients with overt macrovascular DPP-4 inhibitors, AGIs, colesevelam,
osmotic symptoms of hyperglycemia, to disease (60). Another agent of this class, and bromocriptine (;0.5–1.0%). How-
avoid instability in blood glucose over rosiglitazone, is no longer widely available ever, older drugs have typically been
time, and to prevent/delay the develop- owing to concerns of increased myocardial tested in clinical trial participants with
ment of diabetes complications without infarction risk (61). Pioglitazone has re- higher baseline HbA1c, which is itself as-
adversely affecting quality of life. Infor- cently been associated with a possible in- sociated with greater treatment emergent
mation on whether specific agents have creased risk of bladder cancer (62). glycemic reductions, irrespective of ther-
this ability is incomplete; an answer to Recognized side effects of TZDs include apy type. In head-to-head studies, any
these questions requires long-term, large- weight gain, fluid retention leading to differential effects on glucose control are
scale clinical trialsdnot available for most edema and/or heart failure in predisposed small. So agent- and patient-specific prop-
drugs. Effects on surrogate measures for individuals, and increased risk of bone erties, such as dosing frequency, side-effect
glycemic control (e.g., HbA1c) generally fractures (57,60). profiles, cost, and other benefits often
reflect changes in the probability of de- Drugs focused on the incretin system guide their selection.
veloping microvascular disease but not have been introduced more recently (63). Insulin. Due to the progressive b-cell
necessarily macrovascular complications. The injectable GLP-1 receptor agonists dysfunction that characterizes type 2 di-
Particularly from a patient standpoint, mimic the effects of endogenous GLP-1, abetes, insulin replacement therapy is fre-
stability of metabolic control over time thereby stimulating pancreatic insulin se- quently required (71). Importantly, most
may be another specific goal. cretion in a glucose-dependent fashion, patients maintain some endogenous insu-
Metformin, a biguanide, remains the suppressing pancreatic glucagon output, lin secretion even in late stages of disease.
most widely used first-line type 2 diabetes slowing gastric emptying, and decreasing Accordingly, the more complex and in-
drug; its mechanism of action predomi- appetite. Their main advantage is weight tensive strategies of type 1 diabetes are
nately involves reducing hepatic glucose loss, which is modest in most patients but not typically necessary (72).
6 DIABETES CARE care.diabetesjournals.org

7. Inzucchi and Associates
Ideally, the principle of insulin use is glucotoxicity resolved, and the metabolic
the creation of as normal a glycemic profile
KEY POINTS state stabilized, it may be possible to taper
as possible without unacceptable weight insulin partially or entirely, transferring to
c Glycemic targets and glucose-lowering
gain or hypoglycemia (73). As initial ther- noninsulin antihyperglycemic agents, per-
therapies must be individualized.
apy, unless the patient is markedly hyper- haps in combination.
c Diet, exercise, and education remain
glycemic and/or symptomatic, a “basal” If metformin cannot be used, another
the foundation of any type 2 diabetes
insulin alone is typically added (74). Basal oral agent could be chosen, such as a
treatment program.
insulin provides relatively uniform insulin sulfonylurea/glinide, pioglitazone, or a
c Unless there are prevalent contra-
coverage throughout the day and night, DPP-4 inhibitor; in occasional cases where
indications, metformin is the op-
mainly to control blood glucose by sup- weight loss is seen as an essential aspect of
timal first-line drug.
pressing hepatic glucose production in therapy, initial treatment with a GLP-1
c After metformin, there are limited
between meals and during sleep. Either receptor agonist might be useful. Where
data to guide us. Combination
intermediate-acting (neutral protamine available, less commonly used drugs (AGIs,
therapy with an additional 1–2 oral
Hagedorn [NPH]) or long-acting (insulin colesevelam, bromocriptine) might also be
or injectable agents is reasonable,
glargine [A21Gly,B31Arg,B32Arg hu- considered in selected patients, but their
aiming to minimize side effects
man insulin] or insulin detemir [B29Lys modest glycemic effects and side-effect
where possible.
(´-tetradecanoyl),desB30 human insulin]) profiles make them less attractive candi-
c Ultimately, many patients will require
formulations may be used. The latter two dates. Specific patient preferences, char-
insulin therapy alone or in com-
are associated with modestly less overnight acteristics, susceptibilities to side effects,
bination with other agents to
hypoglycemia (insulin glargine, insulin de- potential for weight gain and hypoglycemia
maintain glucose control.
temir) than NPH and possibly slightly less should play a major role in drug selection
c All treatment decisions, where possi-
weight gain (insulin detemir), but are (20,21). (See Supplementary Figs. for ad-
ble, should be made in conjunction
more expensive (75,76). Of note, the dos- aptations of Fig. 2 that address specific
with the patient, focusing on his/her
ing of these basal insulin analogs may differ, patient scenarios.)
preferences, needs, and values.
with most comparative trials showing a Advancing to dual combination therapy.
c Comprehensive cardiovascular risk
higher average unit requirement with insu- Figure 2 (and Supplementary Figs.) also
reduction must be a major focus of
lin detemir (77). depicts potential sequences of escalating
therapy.
Although the majority of patients glucose-lowering therapy beyond met-
with type 2 diabetes requiring insulin formin. If monotherapy alone does not
therapy can be successfully treated with achieve/maintain an HbA1c target over
basal insulin alone, some, because of pro- ;3 months, the next step would be to
gressive diminution in their insulin secre- Implementation strategies add a second oral agent, a GLP-1 recep-
tory capacity, will require prandial insulin Initial drug therapy. It is generally tor agonist, or basal insulin (5,10). No-
therapy with shorter-acting insulins. This agreed that metformin, if not contraindi- tably, the higher the HbA1c, the more
is typically provided in the form of the cated and if tolerated, is the preferred and likely insulin will be required. On average,
rapid insulin analogs, insulin lispro most cost-effective first agent (42) (Fig. 2 any second agent is typically associated
(B28Lys,B29Pro human insulin), insulin and Supplementary Figs.). It is initiated at, with an approximate further reduction in
aspart (B28Asp human insulin), or insulin or soon after, diagnosis, especially in pa- HbA1c of ;1% (70,79). If no clinically
glulisine (B3Lys,B29Glu human insulin), tients in whom lifestyle intervention alone meaningful glycemic reduction (i.e., “non-
which may be dosed just before the meal. has not achieved, or is unlikely to achieve, responder”) is demonstrated, then, adher-
They result in better postprandial glucose HbA1c goals. Because of frequent gastroin- ence having been investigated, that agent
control than the less costly human regular testinal side effects, it should be started at a should be discontinued, and another
insulin, whose pharmacokinetic profile low dose with gradual titration. Patients with a different mechanism of action
makes it less attractive in this setting. with a high baseline HbA1c (e.g., $9.0%) substituted. With a distinct paucity of
Ideally, an insulin treatment program have a low probability of achieving a near- long-term comparative-effectiveness trials
should be designed specifically for an in- normal target with monotherapy. It may available, uniform recommendations on
dividual patient, to match the supply of therefore be justified to start directly the best agent to be combined with metfor-
insulin to his or her dietary/exercise hab- with a combination of two noninsulin min cannot be made (80). Thus, advantages
its and prevailing glucose trends, as revealed agents or with insulin itself in this circum- and disadvantages of specific drugs for each
through self-monitoring. Anticipated glucose- stance (78). If a patient presents with sig- patient should be considered (Table 1).
lowering effects should be balanced with nificant hyperglycemic symptoms and/or Some antihyperglycemic medications
the convenience of the regimen, in the has dramatically elevated plasma glucose lead to weight gain. This may be associ-
context of an individual’s specific therapy concentrations (e.g., .16.7–19.4 mmol/L ated with worsening markers of insulin
goals (Fig. 1). [.300–350 mg/dL]) or HbA 1c (e.g., resistance and cardiovascular risk. One
Proper patient education regarding $10.0–12.0%), insulin therapy should be exception may be TZDs (57); weight gain
glucose monitoring, insulin injection strongly considered from the outset. Such associated with this class occurs in asso-
technique, insulin storage, recognition/ treatment is mandatory when catabolic ciation with decreased insulin resistance.
treatment of hypoglycemia, and “sick features are exhibited or, of course, if Although there is no uniform evidence that
day” rules is imperative. Where available, ketonuria is demonstrated, the latter re- increases in weight in the range observed
certified diabetes educators can be in- flecting profound insulin deficiency. Im- with certain therapies translate into a sub-
valuable in guiding the patient through portantly, unless there is evidence of type 1 stantially increased cardiovascular risk, it
this process. diabetes, once symptoms are relieved, remains important to avoid unnecessary
care.diabetesjournals.org DIABETES CARE 7

8. Position Statement
Figure 2dAntihyperglycemic therapy in type 2 diabetes: general recommendations. Moving from the top to the bottom of the figure, potential
sequences of antihyperglycemic therapy. In most patients, begin with lifestyle changes; metformin monotherapy is added at, or soon after, diagnosis
(unless there are explicit contraindications). If the HbA1c target is not achieved after ;3 months, consider one of the five treatment options combined
with metformin: a sulfonylurea, TZD, DPP-4 inhibitor, GLP-1 receptor agonist, or basal insulin. (The order in the chart is determined by historical
introduction and route of administration and is not meant to denote any specific preference.) Choice is based on patient and drug characteristics, with
the over-riding goal of improving glycemic control while minimizing side effects. Shared decision making with the patient may help in the selection of
therapeutic options. The figure displays drugs commonly used both in the U.S. and/or Europe. Rapid-acting secretagogues (meglitinides) may be
used in place of sulfonylureas. Other drugs not shown (a-glucosidase inhibitors, colesevelam, dopamine agonists, pramlintide) may be used where
available in selected patients but have modest efficacy and/or limiting side effects. In patients intolerant of, or with contraindications for, metformin,
select initial drug from other classes depicted and proceed accordingly. In this circumstance, while published trials are generally lacking, it is
reasonable to consider three-drug combinations other than metformin. Insulin is likely to be more effective than most other agents as a third-line
therapy, especially when HbA1c is very high (e.g., $9.0%). The therapeutic regimen should include some basal insulin before moving to more
complex insulin strategies (Fig. 3). Dashed arrow line on the left-hand side of the figure denotes the option of a more rapid progression from a two-
drug combination directly to multiple daily insulin doses, in those patients with severe hyperglycemia (e.g., HbA1c $10.0–12.0%). DPP-4-i, DPP-4
inhibitor; Fx’s, bone fractures; GI, gastrointestinal; GLP-1-RA, GLP-1 receptor agonist; HF, heart failure; SU, sulfonylurea. aConsider beginning at this
stage in patients with very high HbA1c (e.g., $9%). bConsider rapid-acting, nonsulfonylurea secretagogues (meglitinides) in patients with irregular
meal schedules or who develop late postprandial hypoglycemia on sulfonylureas. cSee Table 1 for additional potential adverse effects and risks, under
“Disadvantages.” dUsually a basal insulin (NPH, glargine, detemir) in combination with noninsulin agents. eCertain noninsulin agents may be
continued with insulin (see text). Refer to Fig. 3 for details on regimens. Consider beginning at this stage if patient presents with severe hyperglycemia
($16.7–19.4 mmol/L [$300–350 mg/dL]; HbA1c $10.0–12.0%) with or without catabolic features (weight loss, ketosis, etc.).
weight gain by optimal medication selec- may be tolerated by some, but not others. glucose-lowering agents in many environ-
tion and dose titration. Fluid retention may pose a clinical or ments. For resource-limited settings, less
For all medications, consideration merely an aesthetic problem (82). The expensive agents should be chosen. How-
should also be given to overall tolerability. risk of bone fractures may be a specific con- ever, due consideration should be also
Even occasional hypoglycemia may be cern in postmenopausal women (57). given to side effects and any necessary
devastating, if severe, or merely irritating, It must be acknowledged that costs monitoring, with their own cost impli-
if mild (81). Gastrointestinal side effects are a critical issue driving the selection of cations. Moreover, prevention of morbid
8 DIABETES CARE care.diabetesjournals.org

9. Inzucchi and Associates
long-term complications will likely reduce
long-term expenses attributed to the disease.
Advancing to triple combination ther-
apy. Some studies have shown advantages
of adding a third noninsulin agent to a
two-drug combination that is not yet or
no longer achieving the glycemic target
(83–86). Not surprisingly, however, at
this juncture, the most robust response
will usually be with insulin. Indeed, since
diabetes is associated with progressive
b-cell loss, many patients, especially those
with long-standing disease, will eventually
need to be transitioned to insulin, which
should be favored in circumstances where
the degree of hyperglycemia (e.g., $8.5%)
makes it unlikely that another drug will be
of sufficient benefit (87). If triple combina-
tion therapy exclusive of insulin is tried, the
patient should be monitored closely, with
the approach promptly reconsidered if it
proves to be unsuccessful. Many months
of uncontrolled hyperglycemia should
specifically be avoided.
In using triple combinations the es-
Figure 3dSequential insulin strategies in type 2 diabetes. Basal insulin alone is usually the
sential consideration is obviously to use optimal initial regimen, beginning at 0.1–0.2 units/kg body weight, depending on the degree of
agents with complementary mechanisms hyperglycemia. It is usually prescribed in conjunction with one to two noninsulin agents. In
of action (Fig. 2 and Supplementary Figs.). patients willing to take more than one injection and who have higher HbA1c levels ($9.0%), twice-
Increasing the number of drugs heightens daily premixed insulin or a more advanced basal plus mealtime insulin regimen could also be
the potential for side effects and drug–drug considered (curved dashed arrow lines). When basal insulin has been titrated to an acceptable
interactions, raises costs, and negatively im- fasting glucose but HbA1c remains above target, consider proceeding to basal plus mealtime in-
pacts patient adherence. The rationale, ben- sulin, consisting of one to three injections of rapid-acting analogs (see text for details). A less
efits, and side effects of each new medication studied alternativedprogression from basal insulin to a twice-daily premixed insulindcould be
should be discussed with the patient. The also considered (straight dashed arrow line); if this is unsuccessful, move to basal plus mealtime
insulin. The figure describes the number of injections required at each stage, together with the relative
clinical characteristics of patients more or
complexity and flexibility. Once a strategy is initiated, titration of the insulin dose is important, with
less likely to respond to specific combina- dose adjustments made based on the prevailing glucose levels as reported by the patient. Noninsulin
tions are, unfortunately, not well defined. agents may be continued, although insulin secretagogues (sulfonylureas, meglitinides) are typically
Transitions to and titrations of insulin. stopped once more complex regimens beyond basal insulin are utilized. Comprehensive education
Most patients express reluctance to be- regarding self-monitoring of blood glucose, diet, exercise, and the avoidance of, and response to,
ginning injectable therapy, but, if the hypoglycemia are critical in any patient on insulin therapy. Mod., moderate.
practitioner feels that such a transition is
important, encouragement and education
can usually overcome such reticence. In- weekly if the fasting glucose levels are above glucose excursions (e.g., to .10.0 mmol/L
sulin is typically begun at a low dose (e.g., the preagreed target is a reasonable ap- [.180 mg/dL]) occur. This is suggested
0.1–0.2 U kg21 day21), although larger proach (89). As the target is neared, dosage when the fasting glucose is at target but
amounts (0.3–0.4 U kg21 day21) are rea- adjustments should be more modest and the HbA1c remains above goal after 3–6
sonable in the more severely hyperglyce- occur less frequently. Downward adjust- months of basal insulin titration (91). The
mic. The most convenient strategy is ment is advisable if any hypoglycemia oc- same would apply if large drops in glucose
with a single injection of a basal insulin, curs. During self-titration, frequent contact occur during overnight hours or in be-
with the timing of administration depen- (telephone, e-mail) with the clinician may tween meals, as the basal insulin dose is
dent on the patient’s schedule and overall be necessary. Practitioners themselves can, increased. In this scenario, the basal insulin
glucose profile (Fig. 3). of course, also titrate basal insulin, but this dose would obviously need to be simulta-
Although extensive dosing instruc- would involve more intensive contact with neously decreased as prandial insulin is ini-
tions for insulin are beyond the scope of the patient than typically available in rou- tiated. Although basal insulin is titrated
this statement, most patients can be taught tine clinical practice. Daily self-monitoring primarily against the fasting glucose, gen-
to uptitrate their own insulin dose based of blood glucose is of obvious importance erally irrespective of the total dose, practi-
on several algorithms, each essentially in- during this phase. After the insulin dose is tioners should be aware that the need for
volving the addition of a small dose increase stabilized, the frequency of monitoring prandial insulin therapy will become likely the
if hyperglycemia persists (74,76,88). For should be reviewed (90). more the daily dose exceeds 0.5 U kg21
example, the addition of 1–2 units (or, in Consideration should be given to the day21, especially as it approaches 1 U kg21
those already on higher doses, increments addition of prandial or mealtime insulin day21. The aim with mealtime insulin is to
of 5–10%) to the daily dose once or twice coverage when significant postprandial blunt postprandial glycemic excursions,
care.diabetesjournals.org DIABETES CARE 9

10. Position Statement
which can be extreme in some individuals, 2. All insulins are associated with some events. It follows that glycemic targets for
resulting in poor control during the day. weight gain and some risk of hypo- elderly with long-standing or more com-
Such coverage may be provided by one of glycemia. plicated disease should be less ambitious
two methods. 3. The larger the doses and the more than for the younger, healthier individuals
The most precise and flexible prandial aggressive the titration, the lower the (20). If lower targets cannot be achieved
coverage is possible with “basal-bolus” HbA1c, but often with a greater likeli- with simple interventions, an HbA1c of
therapy, involving the addition of premeal hood of adverse effects. ,7.5–8.0% may be acceptable, transition-
rapid-acting insulin analog to ongoing 4. Generally, long-acting insulin analogs ing upward as age increases and capacity
basal insulin. One graduated approach reduce the incidence of overnight hy- for self-care, cognitive, psychological and
is to add prandial insulin before the poglycemia, and rapid-acting insulin economic status, and support systems
meal responsible for the largest glucose analogs reduce postprandial glucose decline.
excursiondtypically that with the greatest excursions as compared with corre- While lifestyle modification can be
carbohydrate content, often, but not always, sponding human insulins (NPH, Reg- successfully implemented across all age-
the evening meal (92). Subsequently, a sec- ular), but they generally do not result in groups, in the aged, the choice of anti-
ond injection can be administered before the clinically significantly lower HbA1c. hyperglycemic agent should focus on
meal with the next largest excursion (often drug safety, especially protecting against
breakfast). Ultimately, a third injection may Metformin is often continued when hypoglycemia, heart failure, renal dys-
be added before the smallest meal (often basal insulin is added, with studies dem- function, bone fractures, and drug–drug
lunch) (93). The actual glycemic benefits onstrating less weight gain when the two interactions. Strategies specifically mini-
of these more advanced regimens after are used together (95). Insulin secretago- mizing the risk of low blood glucose may
basal insulin are generally modest in typical gues do not seem to provide for additional be preferred.
patients (92). So, again, individualization HbA1c reduction or prevention of hypo- In contrast, healthier patients with
of therapy is key, incorporating the degree glycemia or weight gain after insulin is long life expectancy accrue risk for vas-
of hyperglycemia needing to be ad- started, especially after the dose is titrated cular complications over time. Therefore,
dressed and the overall capacities of the and stabilized. When basal insulin is lower glycemic targets (e.g., an HbA1c
patient. Importantly, data trends from used, continuing the secretagogue may ,6.5–7.0%) and tighter control of body
self-monitoring may be particularly help- minimize initial deterioration of glycemic weight, blood pressure, and circulating
ful in titrating insulins and their doses control. However, secretagogues should lipids should be achieved to prevent or
within these more advanced regimens to be avoided once prandial insulin regi- delay such complications. This usually re-
optimize control. mens are employed. TZDs should be re- quires combination therapy, the early in-
A second, perhaps more convenient duced in dose (or stopped) to avoid stitution of which may have the best
but less adaptable method involves “pre- edema and excessive weight gain, al- chance of modifying the disease process
mixed” insulin, consisting of a fixed com- though in certain individuals with large and preserving quality of life.
bination of an intermediate insulin with insulin requirements from severe insulin
regular insulin or a rapid analog. Tradi- resistance, these insulin sensitizers may be Weight
tionally, this is administered twice daily, very helpful in lowering HbA1c and mini- The majority of individuals with type 2
before morning and evening meals. In mizing the required insulin dose (96). diabetes are overweight or obese (;80%)
general, when compared with basal insu- Data concerning the glycemic benefits of (102). In these, intensive lifestyle inter-
lin alone, premixed regimens tend to incretin-based therapy combined with vention can improve fitness, glycemic
lower HbA1c to a larger degree, but often basal insulin are accumulating; combina- control, and cardiovascular risk factors
at the expense of slightly more hypogly- tion with GLP-1 receptor agonists may be for relatively small changes in body
cemia and weight gain (94). Disadvan- helpful in some patients (97,98). Once weight (103). Although insulin resistance
tages include the inability to titrate the again, the costs of these more elaborate is thought of as the predominate driver of
shorter- from the longer-acting compo- combined regimens must be carefully diabetes in obese patients, they actually
nent of these formulations. Therefore, considered. have a similar degree of islet dysfunction
this strategy is somewhat inflexible but to leaner patients (37). Perhaps as a result,
may be appropriate for certain patients OTHER CONSIDERATIONS the obese may be more likely to require
who eat regularly and may be in need combination drug therapy (20,104).
of a simplified approach beyond basal in- Age While common practice has favored met-
sulin (92,93). (An older and less commonly Older adults (.65–70 years) often have a formin in heavier patients, because of
used variation of this two-injection strategy higher atherosclerotic disease burden, re- weight loss/weight neutrality, this drug
is known as “split-mixed,” involving a fixed duced renal function, and more comor- is as efficacious in lean individuals (75).
amount of intermediate insulin mixed by bidities (99,100). Many are at risk for TZDs, on the other hand, appear to be
the patient with a variable amount of regu- adverse events from polypharmacy and more effective in those with higher BMIs,
lar insulin or a rapid analog. This allows for may be both socially and economically although their associated weight gain
greater flexibility in dosing.) disadvantaged. Life expectancy is reduced, makes them, paradoxically, a less attractive
The key messages from dozens of especially in the presence of long-term option here. GLP-1 receptor agonists are
comparative insulin trials in type 2 diabetes complications. They are also more likely associated with weight reduction (38),
include the following: to be compromised by hypoglycemia; for which in some patients may be substantial.
example, unsteadiness may result in falls Bariatric surgery is an increasingly
1. Any insulin will lower glucose and and fractures (101), and a tenuous cardiac popular option in severe obesity. Type 2
HbA1c. status may deteriorate into catastrophic diabetes frequently resolves rapidly after
10 DIABETES CARE care.diabetesjournals.org

11. Inzucchi and Associates
these procedures. The majority of patients (112), but the actual clinical relevance of this ongoing debate, however, as to whether
are able to stop some, or even all, of their remains unproven. Metformin may have these thresholds are too restrictive and
antihyperglycemic medications, although some cardiovascular benefits and would that those with mild–moderate renal im-
the durability of this effect is not known appear to be a useful drug in the setting pairment would gain more benefit than
(105). of CAD, barring prevalent contraindica- harm from using metformin (124,125). In
In lean patients, consideration should tions (32). In a single study, pioglitazone the U.K., the National Institute for Health
be given to the possibility of latent autoim- was shown to reduce modestly major ad- and Clinical Excellence (NICE) guidelines
mune diabetes in adults (LADA), a slowly verse cardiovascular events in patients are less proscriptive and more evidence-
progressive form of type 1 diabetes. These with established macrovascular disease. based than those in the U.S., generally al-
individuals, while presenting with mild It may therefore also be considered, unless lowing use down to a GFR of 30 mL/min,
hyperglycemia, often responsive to oral heart failure is present (60). In very pre- with dose reduction advised at 45 mL/min
agents, eventually develop more severe liminary reports, therapy with GLP-1 re- (14). Given the current widespread report-
hyperglycemia and require intensive insu- ceptor agonists and DPP-4 inhibitors has ing of estimated GFR, these guidelines
lin regimens (106). Measuring titres of islet- been associated with improvement in ei- appear very reasonable.
associated autoantibodies (e.g., anti-GAD) ther cardiovascular risk or risk factors, but Most insulin secretagogues undergo
may aid their identification, encouraging a there are no long-term data regarding clin- significant renal clearance (exceptions in-
more rapid transition to insulin therapy. ical outcomes (113). There are very limited clude repaglinide and nateglinide) and
data suggesting that AGIs (114) and bromo- the risk of hypoglycemia is therefore
Sex/racial/ethnic/genetic differences criptine (115) may reduce cardiovascular higher in patients with chronic kidney
While certain racial/ethnic features that events. disease (CKD). For most of these agents,
increase the risk of diabetes are well recog- Heart failure. With an aging population extreme caution is imperative at more severe
nized [greater insulin resistance in Latinos and recent decreases in mortality after degrees of renal dysfunction. Glyburide
(107), more b-cell dysfunction in East myocardial infarction, the diabetic patient (known as glibenclamide in Europe),
Asians (108)], using this information to with progressive heart failure is an in- which has a prolonged duration of
craft optimal therapeutic strategies is in its creasingly common scenario (116). This action and active metabolites, should
infancy. This is not surprising given the population presents unique challenges be specifically avoided in this group.
polygenic inheritance pattern of the dis- given their polypharmacy, frequent hos- Pioglitazone is not eliminated renally, and
ease. Indeed, while matching a drug’s pitalizations, and contraindications to therefore there are no restrictions for use
mechanism of action to the underlying various agents. TZDs should be avoided in CKD. Fluid retention may be a concern,
causes of hyperglycemia in a specific patient (117,118). Metformin, previously contra- however. Among the DPP-4 inhibitors,
seems logical, there are few data that com- indicated in heart failure, can now be used sitagliptin, vildagliptin, and saxagliptin
pare strategies based on this approach if the ventricular dysfunction is not se- share prominent renal elimination. In the
(109). There are few exceptions, mainly vere, if patient’s cardiovascular status is face of advanced CKD, dose reduction is
involving diabetes monogenic variants of- stable, and if renal function is normal necessary. One exception is linagliptin,
ten confused with type 2 diabetes, such as (119). As mentioned, cardiovascular ef- which is predominantly eliminated enter-
maturity-onset diabetes of the young fects of incretin-based therapies, includ- ohepatically. For the GLP-1 receptor ago-
(MODY), several forms of which respond ing those on ventricular function, are nists exenatide is contraindicated in stage
preferentially to sulfonylureas (110). currently under investigation (120). 4–5 CKD (GFR ,30 mL/min) as it is re-
While there are no prominent sex differ- Chronic kidney disease. Kidney disease nally eliminated; the safety of liraglutide is
ences in the response to various antihyper- is highly prevalent in type 2 diabetes, and not established in CKD though pharmaco-
glycemic drugs, certain side effects (e.g., moderate to severe renal functional im- kinetic studies suggest that drug levels are
bone loss with TZDs) may be of greater pairment (eGFR ,60 mL/min) occurs in unaffected as it does not require renal func-
concern in women. approximately 20–30% of patients tion for clearance.
(121,122). The individual with progres- More severe renal functional impair-
Comorbidities sive renal dysfunction is at increased risk ment is associated with slower elimina-
Coronary artery disease. Given the fre- for hypoglycemia, which is multifactorial. tion of all insulins. Thus doses need to be
quency with which type 2 diabetic patients Insulin and, to some degree, the incretin titrated carefully, with some awareness
develop atherosclerosis, optimal manage- hormones are eliminated more slowly, as for the potential for more prolonged
ment strategies for those with or at high risk are antihyperglycemic drugs with renal activity profiles.
for coronary artery disease (CAD) are excretion. Thus, dose reduction may be Liver dysfunction. Individuals with type
important. Since hypoglycemia may ex- necessary, contraindications need to be 2 diabetes frequently have hepatosteatosis
acerbate myocardial ischemia and may observed, and consequences (hypoglyce- as well as other types of liver disease
cause dysrhythmias (111), it follows that mia, fluid retention, etc.) require careful (126). There is preliminary evidence that
medications that predispose patients to evaluation. patients with fatty liver may benefit from
this adverse effect should be avoided, if Current U.S. prescribing guidelines treatment with pioglitazone (45,127,128).
possible. If they are required, however, to warn against the use of metformin in It should not be used in an individual with
achieve glycemic targets, patients should patients with a serum creatinine $133 active liver disease or an alanine transami-
be educated to minimize risk. Because of mmol/L ($1.5 mg/dL) in men or 124 nase level above 2.5 times the upper limit of
possible effects on potassium channels in mmol/L ($1.4 mg/dL) in women. Metfor- normal. In those with steatosis but milder
the heart, certain sulfonylureas have been min is eliminated renally, and cases of lactic liver test abnormalities, this insulin sensi-
proposed to aggravate myocardial ischemia acidosis have been described in patients tizer may be advantageous. Sulfonylureas
through effects on ischemic preconditioning with renal failure (123). There is an can rarely cause abnormalities in liver tests
care.diabetesjournals.org DIABETES CARE 11

12. Position Statement
but are not specifically contraindicated; and life-limiting complications, especially CA; Robert Ratner, MedStar Health Research
meglitinides can also be used. If hepatic CVD (19,23,70). Another issue about which Institute/Georgetown University School of
disease is severe, secretagogues should be more data are needed is the concept of du- Medicine, Washington, DC; Julio Rosenstock,
avoided because of the increased risk of rability of effectiveness (often ascribed to Dallas Diabetes and Endocrine Center at Medi-
hypoglycemia. In patients with mild he- b-cell preservation), which would serve cal City, Dallas, TX; Guntram Schernthaner,
Rudolfstiftung Hospital, Vienna, Austria; Robert
patic disease, incretin-based drugs can be to stabilize metabolic control and decrease Sherwin, Yale University School of Medicine,
prescribed, except if there is a coexisting the future treatment burden for patients. New Haven, CT; Jay Skyler, Miller School of
history of pancreatitis. Insulin has no re- Pharmacogenetics may very well inform Medicine, University of Miami, Miami, FL;
strictions for use in patients with liver im- treatment decisions in the future, guiding Geralyn Spollett, Yale University School of
pairment and is indeed the preferred choice the clinician to recommend a therapy for an Nursing, New Haven, CT; Ellie Strock, In-
in those with advanced disease. individual patient based on predictors of ternational Diabetes Center, Minneapolis, MN;
Hypoglycemia. Hypoglycemia in type 2 response and susceptibility to adverse ef- Agathocles Tsatsoulis, University of Ioannina,
diabetes was long thought to be a trivial fects. We need more clinical data on how Ioannina, Greece; Andrew Wolf, University of
issue, as it occurs less commonly than in phenotype and other patient/disease char- Virginia School of Medicine, Charlottesville,
type 1 diabetes. However, there is emerg- acteristics should drive drug choices. As VA; Bernard Zinman, Mount Sinai Hospital/
University of Toronto, Toronto, ON, Canada.
ing concern based mainly on the results new medications are introduced to the
The American Association of Diabetes Edu-
of recent clinical trials and some cross- type 2 diabetes pharmacopeia, their benefit cators, American College of Physicians, and
sectional evidence of increased risk of and safety should be demonstrated in stud- The Endocrine Society, and several other or-
brain dysfunction in those with repeated ies versus best current treatment, substan- ganizations who wished to remain anony-
episodes. In the ACCORD trial, the fre- tial enough both in size and duration to mous nominated reviewers who provided
quency of both minor and major hypo- provide meaningful data on meaningful input on the final draft. Such feedback does
glycemia was high in intensively managed outcomes. It is appreciated, however, that not constitute endorsement by these groups
patientsdthreefold that associated with head-to-head comparisons of all combina- or these individuals. The final draft was also
conventional therapy (129). It remains tions and permutations would be impossi- peer reviewed and approved by the Professional
unknown whether hypoglycemia was bly large (133). Informed judgment and the Practice Committee of the ADA and the Panel for
expertise of experienced clinicians will Overseeing Guidelines and Statements of the
the cause of the increased mortality in
EASD. We are indebted to Dr. Sue Kirkman of
the intensive group (130,131). Clearly, therefore always be necessary. the ADA for her guidance and support during
however, hypoglycemia is more danger- this process. We also thank Carol Hill and Mary
ous in the elderly and occurs consistently Merkin for providing administrative assistance.
more often as glycemic targets are low- AcknowledgmentsdThis position statement Funding
ered. Hypoglycemia may lead to dys- was written by joint request of the ADA and The three face-to-face meetings and the
rhythmias, but can also lead to accidents the EASD Executive Committees, which have travel of some of the writing group were sup-
and falls (which are more likely to be dan- approved the final document. The process in- ported by the EASD and ADA. D.R. Matthews
gerous in the elderly) (132), dizziness volved wide literature review, three face-to-face acknowledges support from the National In-
(leading to falls), confusion (so other ther- meetings of the Writing Group, several tele- stitute for Health Research.
apies may not be taken or taken incor- conferences, and multiple revisions via e-mail Duality of interest
communications. During the past 12 months, the following
rectly), or infection (such as aspiration We gratefully acknowledge the following relationships with companies whose products
during sleep, leading to pneumonia). Hy- experts who provided critical review of a or services directly relate to the subject matter
poglycemia may be systematically under- draft of this statement: James Best, Melbourne in this document are declared:
reported as a cause of death, so the true Medical School, The University of Melbourne, R.M. Bergenstal: membership of scientific
incidence may not be fully appreciated. Melbourne, Australia; Henk Bilo, Isala Clinics, advisory boards and consultation for or clinical
Perhaps just as importantly, additional con- Zwolle, the Netherlands; John Boltri, Wayne research support with Abbott Diabetes Care,
sequences of frequent hypoglycemia in- State University School of Medicine, Detroit, Amylin, Bayer, Becton Dickinson, Boehringer
clude work disability and erosion of the MI; Thomas Buchanan, Keck School of Ingelheim, Calibra, DexCom, Eli Lilly, Halozyme,
confidence of the patient (and that of family Medicine, University of Southern California, Helmsley Trust, Hygieia, Johnson & Johnson,
or caregivers) to live independently. Accord- Los Angeles, CA; Paul Callaway, University of Medtronic, NIH, Novo Nordisk, Roche, Sanofi,
Kansas School of Medicine-Wichita, Wichita, and Takeda (all under contracts with his
ingly, in at-risk individuals, drug selection employer). Inherited stock in Merck (held
KS; Bernard Charbonnel, University of Nantes,
should favor agents that do not precipitate Nantes, France; Stephen Colagiuri, The Uni- by family)
such events and, in general, blood glucose versity of Sydney, Sydney, Australia; Samuel J.B. Buse: research and consulting with
targets may need to be moderated. Dagogo-Jack, The University of Tennessee Health Amylin Pharmaceuticals, Inc.; AstraZeneca;
Science Center, Memphis, TN; Margo Farber, Biodel Inc.; Boehringer Ingelheim; Bristol-
FUTURE DIRECTIONS/ Detroit Medical Center, Detroit, MI; Cynthia Myers Squibb Company; Diartis Pharmaceuticals,
RESEARCH NEEDSdFor antihyper- Fritschi, College of Nursing, University of Illi- Inc.; Eli Lilly and Company; F. Hoffmann-La
glycemic management of type 2 diabetes, nois at Chicago, Chicago, IL; Rowan Hillson, Roche Ltd; Halozyme Therapeutics; Johnson
the comparative evidence basis to date is The Hillingdon Hospital, Uxbridge, U.K.; & Johnson; Medtronic MiniMed; Merck &
relatively lean, especially beyond metfor- Faramarz Ismail-Beigi, Case Western Reserve Co., Inc.; Novo Nordisk; Pfizer Inc.; Sanofi;
University School of Medicine/Cleveland and TransPharma Medical Ltd (all under
min monotherapy (70). There is a significant
VA Medical Center, Cleveland, OH; Devan contracts with his employer)
need for high-quality comparative- Kansagara, Oregon Health & Science University/ M. Diamant: member of advisory boards of
effectiveness research, not only regarding Portland VA Medical Center, Portland, OR; Abbott Diabetes Care, Eli Lilly, Merck Sharp &
glycemic control, but also costs and those Ilias Migdalis, NIMTS Hospital, Athens, Dohme (MSD), Novo Nordisk, Poxel Pharma.
outcomes that matter most to patientsd Greece; Donna Miller, Keck School of Medicine, Consultancy for: Astra-BMS, Sanofi. Speaker
quality of life and the avoidance of morbid University of Southern California, Los Angeles, engagements: Eli Lilly, MSD, Novo Nordisk.
12 DIABETES CARE care.diabetesjournals.org