2. Standards of Medical Care
Table 1—ADA evidence grading system for clinical practice recommendations 2-h value in the 75-g oral glucose toler-
ance test (OGTT) (4).
Level of In 2009, an International Expert
evidence Description Committee that included representatives
of the ADA, the International Diabetes
A Clear evidence from well-conducted, generalizable, randomized controlled Federation (IDF), and the European As-
trials that are adequately powered, including: sociation for the Study of Diabetes
• Evidence from a well-conducted multicenter trial (EASD) recommended the use of the A1C
• Evidence from a meta-analysis that incorporated quality ratings in the test to diagnose diabetes, with a threshold
analysis of 6.5% (5), and ADA adopted this cri-
Compelling nonexperimental evidence, i.e., “all or none” rule developed terion in 2010 (4). The diagnostic test
by Center for Evidence Based Medicine at Oxford should be performed using a method that
Supportive evidence from well-conducted randomized controlled trials is certified by the National Glycohemo-
that are adequately powered, including: globin Standardization Program (NGSP)
• Evidence from a well-conducted trial at one or more institutions and standardized or traceable to the Dia-
• Evidence from a meta-analysis that incorporated quality ratings in the betes Control and Complications Trial
analysis (DCCT) reference assay. Point-of-care
B Supportive evidence from well-conducted cohort studies A1C assays are not sufficiently accurate at
• Evidence from a well-conducted prospective cohort study or registry this time to use for diagnostic purposes.
• Evidence from a well-conducted meta-analysis of cohort studies Epidemiologic datasets show a simi-
Supportive evidence from a well-conducted case-control study lar relationship between A1C and risk of
C Supportive evidence from poorly controlled or uncontrolled studies retinopathy as has been shown for the
• Evidence from randomized clinical trials with one or more major or corresponding FPG and 2-h plasma glu-
three or more minor methodological flaws that could invalidate the cose thresholds. The A1C has several ad-
results vantages to the FPG and OGTT, including
• Evidence from observational studies with high potential for bias (such greater convenience, since fasting is not
as case series with comparison to historical controls) required; evidence to suggest greater pre-
• Evidence from case series or case reports analytical stability; and less day-to-day
Conflicting evidence with the weight of evidence supporting the perturbations during periods of stress and
recommendation illness. These advantages must be bal-
E Expert consensus or clinical experience anced by greater cost, the limited avail-
ability of A1C testing in certain regions of
the developing world, and the incomplete
tion is listed after each recommendation lin action, diseases of the exocrine pan- correlation between A1C and average glu-
using the letters A, B, C, or E. creas (such as cystic fibrosis), and drug- cose in certain individuals. In addition,
These standards of care are revised or chemical-induced (such as in the A1C levels can vary with patients’ ethnic-
annually by the ADA’s multidisciplinary treatment of HIV/AIDS or after organ ity (6) as well as with certain anemias and
Professional Practice Committee, incor- transplantation) hemoglobinopathies. For patients with an
porating new evidence. Members of the ● Gestational diabetes mellitus (GDM) abnormal hemoglobin but normal red cell
Professional Practice Committee and their (diabetes diagnosed during pregnancy turnover, such as sickle cell trait, an A1C
disclosed conflicts of interest are listed on that is not clearly overt diabetes) assay without interference from abnormal
page S97. Subsequently, as with all Posi- hemoglobins should be used (an updated
tion Statements, the standards of care are Some patients cannot be clearly classified list is available at www.ngsp.org/interf.
reviewed and approved by the Executive as having type 1 or type 2 diabetes. Clin- asp). For conditions with abnormal red
Committee of ADA’s Board of Directors. ical presentation and disease progression cell turnover, such as pregnancy, recent
vary considerably in both types of diabe- blood loss or transfusion, or some ane-
I. CLASSIFICATION AND tes. Occasionally, patients who otherwise mias, the diagnosis of diabetes must em-
DIAGNOSIS OF DIABETES have type 2 diabetes may present with ke- ploy glucose criteria exclusively.
toacidosis. Similarly, patients with type 1 The established glucose criteria for
A. Classification of diabetes diabetes may have a late onset and slow the diagnosis of diabetes (FPG and 2-h
The classification of diabetes includes (but relentless) progression of disease de- PG) remain valid as well (Table 2). Just as
four clinical classes: spite having features of autoimmune dis- there is less than 100% concordance be-
ease. Such difficulties in diagnosis may tween the FPG and 2-h PG tests, there is
● Type 1 diabetes (results from -cell de- not perfect concordance between A1C
occur in children, adolescents, and
struction, usually leading to absolute adults. The true diagnosis may become and either glucose-based test. Analyses of
insulin deficiency) more obvious over time. National Health and Nutrition Examina-
● Type 2 diabetes (results from a progres- tion Survey (NHANES) data indicate that,
sive insulin secretory defect on the assuming universal screening of the undi-
background of insulin resistance) B. Diagnosis of diabetes agnosed, the A1C cut point of 6.5%
● Other specific types of diabetes due to For decades, the diagnosis of diabetes was identifies one-third fewer cases of undiag-
other causes, e.g., genetic defects in based on plasma glucose criteria, either nosed diabetes than a fasting glucose cut
-cell function, genetic defects in insu- the fasting plasma glucose (FPG) or the point of 126 mg/dl (7.0 mmol/l) (7).
S12 DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 care.diabetesjournals.org
3. Position Statement
Table 2—Criteria for the diagnosis of variability of all the tests, it is also possible Table 3—Categories of increased risk for di-
diabetes that when a test whose result was above abetes (prediabetes)*
A1C 6.5%. The test should be performed
the diagnostic threshold is repeated, the FPG 100–125 mg/dl (5.6–6.9 mmol/l): IFG
in a laboratory using a method that is second value will be below the diagnostic or
NGSP certified and standardized to the cut point. This is least likely for A1C, 2-h plasma glucose in the 75-g OGTT 140–
DCCT assay.* somewhat more likely for FPG, and most 199 mg/dl (7.8–11.0 mmol/l): IGT
or likely for the 2-h PG. Barring a laboratory or
FPG 126 mg/dl (7.0 mmol/l). Fasting is error, such patients are likely to have test A1C 5.7–6.4%
defined as no caloric intake for at least results near the margins of the threshold
*For all three tests, risk is continuous, extending
8 h.* for a diagnosis. The healthcare profes- below the lower limit of the range and becoming
or sional might opt to follow the patient disproportionately greater at higher ends of the
2-h plasma glucose 200 mg/dl (11.1 closely and repeat the testing in 3– 6 range.
mmol/l) during an OGTT. The test should months.
be performed as described by the World The current diagnostic criteria for di- compared with an A1C of 5.0% (10). In a
Health Organization, using a glucose load abetes are summarized in Table 2. community-based study of black and
containing the equivalent of 75 g white adults without diabetes, baseline
anhydrous glucose dissolved in water.* C. Categories of increased risk for A1C was a stronger predictor of subse-
or diabetes (prediabetes) quent diabetes and cardiovascular events
In a patient with classic symptoms of In 1997 and 2003, The Expert Committee than fasting glucose (11). Other analyses
hyperglycemia or hyperglycemic crisis, a on Diagnosis and Classification of Diabe- suggest that an A1C of 5.7% is associated
random plasma glucose 200 mg/dl (11.1 tes Mellitus (8,9) recognized an interme- with diabetes risk similar to that of the
mmol/l) diate group of individuals whose glucose high-risk participants in the Diabetes Pre-
levels, although not meeting criteria for vention Program (DPP).
*In the absence of unequivocal hyperglycemia, re-
sult should be confirmed by repeat testing.
diabetes, are nevertheless too high to be Hence, it is reasonable to consider an
considered normal. These persons were A1C range of 5.7– 6.4% as identifying in-
defined as having impaired fasting glu- dividuals with high risk for future diabe-
However, in practice, a large portion of cose (IFG) (FPG levels 100 –125 mg/dl tes, a state that may be referred to as
the diabetic population remains unaware [5.6 – 6.9 mmol/l]) or impaired glucose prediabetes (4). As is the case for individ-
of their condition. Thus, the lower sensi- tolerance (IGT) (2-h PG values in the uals found to have IFG and IGT, individ-
tivity of A1C at the designated cut point OGTT of 140 –199 mg/dl [7.8 –11.0 uals with an A1C of 5.7– 6.4% should be
may well be offset by the test’s greater mmol/l]). It should be noted that the informed of their increased risk for diabe-
practicality, and wider application of a World Health Organization (WHO) and a tes as well as CVD and counseled about
more convenient test (A1C) may actually number of other diabetes organizations effective strategies to lower their risks (see
increase the number of diagnoses made. define the cutoff for IFG at 110 mg/dl (6.1 IV. PREVENTION/DELAY OF TYPE 2 DIABETES). As
As with most diagnostic tests, a test mmol/l). with glucose measurements, the contin-
result diagnostic of diabetes should be re- Individuals with IFG and/or IGT have uum of risk is curvilinear—as A1C rises,
peated to rule out laboratory error, unless been referred to as having prediabetes, in- the risk of diabetes rises disproportion-
the diagnosis is clear on clinical grounds, dicating the relatively high risk for the fu- ately (10). Accordingly, interventions
such as a patient with a hyperglycemic ture development of diabetes. IFG and should be most intensive and follow-up
crisis or classic symptoms of hyperglyce- IGT should not be viewed as clinical en- particularly vigilant for those with A1Cs
mia and a random plasma glucose 200 tities in their own right but rather risk above 6.0%, who should be considered to
mg/dl. It is preferable that the same test be factors for diabetes as well as cardiovas- be at very high risk.
repeated for confirmation, since there will cular disease (CVD). IFG and IGT are as- Table 3 summarizes the categories of
be a greater likelihood of concurrence in sociated with obesity (especially increased risk for diabetes.
this case. For example, if the A1C is 7.0% abdominal or visceral obesity), dyslipide-
and a repeat result is 6.8%, the diagnosis mia with high triglycerides and/or low II. TESTING FOR DIABETES
of diabetes is confirmed. However, if two HDL cholesterol, and hypertension. IN ASYMPTOMATIC
different tests (such as A1C and FPG) are As is the case with the glucose mea- PATIENTS
both above the diagnostic thresholds, the sures, several prospective studies that
diagnosis of diabetes is also confirmed. used A1C to predict the progression to Recommendations
On the other hand, if two different diabetes demonstrated a strong, continu- ● Testing to detect type 2 diabetes and
tests are available in an individual and the ous association between A1C and subse- assess risk for future diabetes in asymp-
results are discordant, the test whose re- quent diabetes. In a systematic review of tomatic people should be considered in
sult is above the diagnostic cut point 44,203 individuals from 16 cohort stud- adults of any age who are overweight or
should be repeated, and the diagnosis is ies with a follow-up interval averaging 5.6 obese (BMI 25 kg/m2) and who have
made on the basis of the confirmed test. years (range 2.8 –12 years), those with an one or more additional risk factors for
That is, if a patient meets the diabetes cri- A1C between 5.5 and 6.0% had a sub- diabetes (Table 4). In those without
terion of the A1C (two results 6.5%) but stantially increased risk of diabetes with these risk factors, testing should begin
not the FPG ( 126 mg/dl or 7.0 mmol/l), 5-year incidences ranging from 9 –25%. at age 45 years. (B)
or vice versa, that person should be con- An A1C range of 6.0 – 6.5% had a 5-year ● If tests are normal, repeat testing car-
sidered to have diabetes. risk of developing diabetes between 25– ried out at least at 3-year intervals is
Since there is preanalytic and analytic 50% and relative risk 20 times higher reasonable. (E)
care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 S13
4. Standards of Medical Care
Table 4—Criteria for testing for diabetes in A. Testing for type 2 diabetes and complications of diabetes within 3 years
asymptomatic adult individuals risk of future diabetes in adults of a negative test result. In the modeling
1. Testing should be considered in all adults Type 2 diabetes is frequently not diag- study, repeat screening every 3 or 5 years
who are overweight (BMI 25 kg/m2*) nosed until complications appear, and was cost-effective (12).
and have additional risk factors: approximately one-fourth of all people Because of the need for follow-up and
• physical inactivity with diabetes in the U.S. may be undiag- discussion of abnormal results, testing
• first-degree relative with diabetes nosed. The effectiveness of early identifi- should be carried out within the health
• high-risk race/ethnicity (e.g., African cation of prediabetes and diabetes care setting. Community screening out-
American, Latino, Native American, through mass testing of asymptomatic in- side a health care setting is not recom-
Asian American, Pacific Islander) dividuals has not been proven defini- mended because people with positive
• women who delivered a baby weighing tively, and rigorous trials to provide such tests may not seek, or have access to, ap-
9 lb or were diagnosed with GDM proof are unlikely to occur. However, propriate follow-up testing and care.
• hypertension ( 140/90 mmHg or on mathematical modeling studies suggest Conversely, there may be failure to ensure
therapy for hypertension) that screening independent of risk factors appropriate repeat testing for individuals
• HDL cholesterol level 35 mg/dl (0.90 beginning at age 30 or 45 years is highly who test negative. Community screening
mmol/l) and/or a triglyceride level 250 cost-effective ( $11,000 per quality- may also be poorly targeted, i.e., it may
mg/dl (2.82 mmol/l) adjusted life-year gained) (12). Prediabe- fail to reach the groups most at risk and
• women with polycystic ovarian tes and diabetes meet established criteria inappropriately test those at low risk (the
syndrome (PCOS) for conditions in which early detection is worried well) or even those already diag-
• A1C 5.7%, IGT, or IFG on previous appropriate. Both conditions are com- nosed.
testing mon and increasing in prevalence and im-
• other clinical conditions associated with pose significant public health burdens. B. Testing for type 2 diabetes in
insulin resistance (e.g., severe obesity, There is a long presymptomatic phase be- children
acanthosis nigricans) fore the diagnosis of type 2 diabetes is The incidence of type 2 diabetes in ado-
• history of CVD usually made. Relatively simple tests are lescents has increased dramatically in the
2. In the absence of the above criteria, testing available to detect preclinical disease. Ad- last decade, especially in minority popu-
for diabetes should begin at age 45 ditionally, the duration of glycemic bur- lations (21), although the disease remains
years. den is a strong predictor of adverse rare in the general pediatric population
3. If results are normal, testing should be outcomes, and effective interventions ex- (22). Consistent with recommendations
repeated at least at 3-year intervals, with ist to prevent progression of prediabetes for adults, children and youth at in-
consideration of more frequent testing to diabetes (see IV. PREVENTION/DELAY OF TYPE creased risk for the presence or the devel-
depending on initial results and risk 2 DIABETES) and to reduce risk of compli- opment of type 2 diabetes should be
status. cations of diabetes (see VI. PREVENTION AND tested within the health care setting. The
MANAGEMENT OF DIABETES COMPLICATIONS). recommendations of the ADA Consensus
*At-risk BMI may be lower in some ethnic groups.
Recommendations for testing for dia- Statement on Type 2 Diabetes in Children
betes in asymptomatic, undiagnosed and Youth (23), with some modifications,
adults are listed in Table 4. Testing should are summarized in Table 5.
● To test for diabetes or to assess risk of be considered in adults of any age with
future diabetes, A1C, FPG, or 2-h 75-g BMI 25 kg/m2 and one or more of the C. Screening for type 1 diabetes
OGTT is appropriate. (B) known risk factors for diabetes. Because Generally, people with type 1 diabetes
● In those identified with increased risk age is a major risk factor for diabetes, test- present with acute symptoms of diabetes
for future diabetes, identify and, if ap- ing of those without other risk factors and markedly elevated blood glucose lev-
propriate, treat other CVD risk factors. should begin no later than age 45 years. els, and most cases are diagnosed soon
(B) Either A1C, FPG, or the 2-h OGTT is after the onset of hyperglycemia. How-
appropriate for testing. The 2-h OGTT ever, evidence from type 1 prevention stud-
For many illnesses, there is a major dis- identifies people with either IFG or IGT ies suggests that measurement of islet
tinction between screening and diagnos- and thus more people at increased risk for autoantibodies identifies individuals who
tic testing. However, for diabetes, the the development of diabetes and CVD. It are at risk for developing type 1 diabetes.
same tests would be used for “screening” should be noted that the two tests do not Such testing may be appropriate in high-
as for diagnosis. Diabetes may be identi- necessarily detect the same individuals. risk individuals, such as those with prior
fied anywhere along a spectrum of clinical The efficacy of interventions for primary transient hyperglycemia or those who have
scenarios ranging from a seemingly low- prevention of type 2 diabetes (13–19) relatives with type 1 diabetes, in the context
risk individual who happens to have glu- have primarily been demonstrated among of clinical research studies (see, for ex-
cose testing, to a higher-risk individual individuals with IGT, not for individuals ample, http://www2.diabetestrialnet.org).
whom the provider tests because of high with IFG (who do not also have IGT) or Widespread clinical testing of asymptom-
suspicion of diabetes, to the symptomatic for individuals with specific A1C levels. atic low-risk individuals cannot currently
patient. The discussion herein is primar- The appropriate interval between be recommended, as it would identify
ily framed as testing for diabetes in those tests is not known (20). The rationale for very few individuals in the general popu-
without symptoms. Testing for diabetes the 3-year interval is that false negatives lation who are at risk. Individuals who
will also detect individuals at increased will be repeated before substantial time screen positive should be counseled
future risk for diabetes, herein referred to elapses, and there is little likelihood that about their risk of developing diabetes.
as having prediabetes. an individual will develop significant Clinical studies are being conducted to
S14 DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 care.diabetesjournals.org
5. Position Statement
Table 5—Testing for type 2 diabetes in for detection and classification of GDM, Table 6—Screening for and diagnosis of
asymptomatic children but its limitations were recognized for GDM
Criteria many years. As the ongoing epidemic of Perform a 75-g OGTT, with plasma glucose
• Overweight (BMI 85th percentile for obesity and diabetes has led to more type measurement fasting and at 1 and 2 h,
age and sex, weight for height 85th 2 diabetes in women of childbearing age, at 24–28 weeks of gestation in women
percentile, or weight 120% of ideal for the number of pregnant women with un- not previously diagnosed with overt
height) diagnosed type 2 diabetes has increased diabetes.
Plus any two of the following risk factors: (24). Because of this, it is reasonable to The OGTT should be performed in the
• Family history of type 2 diabetes in first- screen women with risk factors for type 2 morning after an overnight fast of at
or second-degree relative diabetes (Table 4) for diabetes at their ini- least 8 h.
• Race/ethnicity (Native American, African tial prenatal visit, using standard diagnos- The diagnosis of GDM is made when any of
American, Latino, Asian American, tic criteria (Table 2). Women with the following plasma glucose values are
Pacific Islander) diabetes found at this visit should receive exceeded:
• Signs of insulin resistance or conditions a diagnosis of overt, not gestational, dia- • Fasting 92 mg/dl (5.1 mmol/l)
associated with insulin resistance betes. • 1 h 180 mg/dl (10.0 mmol/l)
(acanthosis nigricans, hypertension, GDM carries risks for the mother and • 2 h 153 mg/dl (8.5 mmol/l)
dyslipidemia, PCOS, or small-for- neonate. The Hyperglycemia and Adverse
gestational-age birth weight) Pregnancy Outcomes (HAPO) study (25),
• Maternal history of diabetes or GDM a large-scale ( 25,000 pregnant women)
during the child’s gestation multinational epidemiologic study, dem- optimizing gestational outcomes for
Age of initiation: age 10 years or at onset of onstrated that risk of adverse maternal, women and their babies.
puberty, if puberty occurs at a younger fetal, and neonatal outcomes continu- Admittedly, there are few data from
age ously increased as a function of maternal randomized clinical trials regarding ther-
Frequency: every 3 years glycemia at 24 –28 weeks, even within apeutic interventions in women who will
ranges previously considered normal for now be diagnosed with GDM based on
pregnancy. For most complications, there only one blood glucose value above the
test various methods of preventing type 1 was no threshold for risk. These results specified cut points (in contrast to the
diabetes, or reversing early type 1 diabe- older criteria that stipulated at least two
have led to careful reconsideration of the
tes, in those with evidence of autoimmu- abnormal values.) Expected benefits to
diagnostic criteria for GDM. After delib-
nity. their pregnancies and offspring is inferred
erations in 2008 –2009, the International
Association of Diabetes and Pregnancy from intervention trials that focused on
III. DETECTION AND women with more mild hyperglycemia
Study Groups (IADPSG), an international
DIAGNOSIS OF than identified using older GDM diagnos-
consensus group with representatives
GESTATIONAL DIABETES tic criteria and that found modest benefits
MELLITUS from multiple obstetrical and diabetes or-
ganizations, including ADA, developed (27,28). The frequency of their follow-up
revised recommendations for diagnosing and blood glucose monitoring is not yet
Recommendations clear, but likely to be less intensive than
● Screen for undiagnosed type 2 diabetes GDM. The group recommended that all
women not known to have diabetes un- women diagnosed by the older criteria.
at the first prenatal visit in those with
dergo a 75-g OGTT at 24 –28 weeks of Additional well-designed clinical studies
risk factors, using standard diagnostic
gestation. Additionally, the group devel- are needed to determine the optimal in-
criteria. (B)
● In pregnant women not known to have oped diagnostic cut points for the fasting, tensity of monitoring and treatment of
diabetes, screen for GDM at 24 –28 1-h, and 2-h plasma glucose measure- women with GDM diagnosed by the new
weeks of gestation, using a 75-g 2-h ments that conveyed an odds ratio for ad- criteria (that would not have met the prior
OGTT and the diagnostic cut points in verse outcomes of at least 1.75 compared definition of GDM). It is important to note
Table 6. (B) with the mean glucose levels in the HAPO that 80 –90% of women in both of the
● Screen women with GDM for persistent study. Current screening and diagnostic mild GDM studies (whose glucose values
diabetes 6 –12 weeks postpartum. (E) strategies, based on the IADPSG state- overlapped with the thresholds recom-
● Women with a history of GDM should ment (26), are outlined in Table 6. mended herein) could be managed with
have lifelong screening for the develop- These new criteria will significantly lifestyle therapy alone.
ment of diabetes or prediabetes at least increase the prevalence of GDM, primar- Because some cases of GDM may rep-
every 3 years. (E) ily because only one abnormal value, not resent preexisting undiagnosed type 2 di-
two, is sufficient to make the diagnosis. abetes, women with a history of GDM
For many years, GDM was defined as any The ADA recognizes the anticipated sig- should be screened for diabetes 6 –12
degree of glucose intolerance with onset nificant increase in the incidence of GDM weeks postpartum, using nonpregnant
or first recognition during pregnancy (8), to be diagnosed by these criteria and is OGTT criteria. Women with a history of
whether or not the condition persisted af- sensitive to concerns about the “medical- GDM have a greatly increased subsequent
ter pregnancy, and not excluding the pos- ization” of pregnancies previously catego- risk for diabetes (29) and should be fol-
sibility that unrecognized glucose rized as normal. These diagnostic criteria lowed up with subsequent screening for
intolerance may have antedated or begun changes are being made in the context of the development of diabetes or prediabe-
concomitantly with the pregnancy. This worrisome worldwide increases in obe- tes, as outlined in II. TESTING FOR DIABETES IN
definition facilitated a uniform strategy sity and diabetes rates, with the intent of ASYMPTOMATIC PATIENTS.
care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 S15
6. Standards of Medical Care
Table 7—Therapies proven effective in diabetes prevention trials
Incidence in 3-Year
Mean control Relative risk number
age Duration Intervention subjects reduction (%) needed to
Study (ref.) n Population (years) (years) (daily dose) (%/year) (95% CI) treat
Lifestyle
Finnish DPS (14) 522 IGT, BMI 25 kg/m2 55 3.2 I-D&E 6 58 (30–70) 8.5
DPP (13) 2,161* IGT, BMI 24 kg/m2, 51 3 I-D&E 10.4 58 (48–66) 6.9
FPG 5.3 mmol/l
Da Qing (15) 259* IGT (randomized groups) 45 6 G-D&E 14.5 38 (14–56) 7.9
Toranomon Study 458 IGT (men), BMI 24 55 4 I-D&E 2.4 67 (P 0.043)† 20.6
(35) kg/m2
Indian DPP (19) 269* IGT 46 2.5 I-D&E 23 29 (21–37) 6.4
Medications
DPP (13) 2,155* IGT, BMI 24 kg/m2, 51 2.8 Metformin (1,700 10.4 31 (17–43) 13.9
FPG 5.3 mmol/l mg)
Indian DPP (19) 269* IGT 46 2.5 Metformin (500 mg) 23 26 (19–35) 6.9
STOP-NIDDM (17) 1,419 IGT, FPG 5.6 mmol/l 54 3.2 Acarbose (300 mg) 12.4 25 (10–37) 9.6
XENDOS (36) 3,277 BMI 30 kg/m2 43 4 Orlistat (360 mg) 2.4 37 (14–54) 45.5
DREAM (18) 5,269 IGT or IFG 55 3.0 Rosiglitazone (8 mg) 9.1 60 (54–65) 6.9
Voglibose Ph-3 1,780 IGT 56 3.0 (1-year Vogliobose (0.2 mg) 12.0 40 (18–57) 21 (1-year
(37) Rx) Rx)
Modified and reprinted with permission (38). Percentage points: Number needed to treat to prevent 1 case of diabetes, standardized for a 3-year period to improve
comparisons across studies. *Number of participants in the indicated comparisons, not necessarily in entire study. †Calculated from information in the article. DPP, Diabetes
Prevention Program; DPS, Diabetes Prevention Study; DREAM, Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication; STOP-NIDDM, Study to
Prevent Non-Insulin Dependent Diabetes; XENDOS, Xenical in the prevention of Diabetes in Obese Subjects. I, individual; G, group; D&E, diet and exercise.
IV. PREVENTION/DELAY abetes (13–19). These interventions loss and moderate physical activity of at
OF TYPE 2 DIABETES include intensive lifestyle modification least 150 min/week). Regarding the more
programs that have been shown to be very difficult issue of drug therapy for diabetes
Recommendations effective (58% reduction after 3 years) prevention, a consensus panel felt that
● Patients with IGT (A), IFG (E), or an and use of the pharmacologic agents met- metformin should be the only drug con-
A1C of 5.7– 6.4% (E) should be re- formin, -glucosidase inhibitors, orlistat, sidered (39). For other drugs, the issues of
ferred to an effective ongoing support and thiazolidinediones (TZDs), each of cost, side effects, and lack of persistence
program targeting weight loss of 7% of which has been shown to decrease inci- of effect in some studies led the panel to
body weight and increasing physical dent diabetes to various degrees. A sum- not recommend their use for diabetes pre-
activity to at least 150 min/week of mary of major diabetes prevention trials is vention. Metformin, which was signifi-
moderate activity such as walking. shown in Table 7. cantly less effective than lifestyle in the
● Follow-up counseling appears to be im- Follow-up of all three large studies of DPP and DPPOS, reasonably may be rec-
portant for success. (B) lifestyle intervention has shown sustained ommended for very-high-risk individuals
● Based on potential cost savings of diabe- reduction in the rate of conversion to type (those with risk factors for diabetes and/or
tes prevention, such programs should be 2 diabetes, with 43% reduction at 20 those with more severe or progressive hy-
covered by third-party payors. (E) years in the Da Qing study (30), 43% re- perglycemia). Of note, in the DPP met-
● Metformin therapy for prevention of duction at 7 years in the Finnish Diabetes formin was most effective compared to
type 2 diabetes may be considered in Prevention Study (DPS) (31) and 34% re- lifestyle in those with BMI of at least 35
those at the highest risk for developing duction at 10 years in the U.S. Diabetes kg/m2 and was not significantly better
diabetes, such as those with multiple Prevention Program Outcomes Study than placebo in those over age 60 years.
risk factors, especially if they demon- (DPPOS) (32). A cost-effectiveness analy-
strate progression of hyperglycemia sis suggested that lifestyle interventions as V. DIABETES CARE
(e.g., A1C 6%) despite lifestyle inter- delivered in the DPP are cost-effective
ventions. (B) (33). Group delivery of the DPP interven- A. Initial evaluation
● Monitoring for the development of di- tion in community settings has the poten- A complete medical evaluation should be
abetes in those with prediabetes should tial to be significantly less expensive while performed to classify the diabetes, detect
be performed every year. (E) still achieving similar weight loss (34). the presence of diabetes complications,
Based on the results of clinical trials review previous treatment and glycemic
Randomized controlled trials have shown and the known risks of progression of control in patients with established diabe-
that individuals at high risk for develop- prediabetes to diabetes, persons with an tes, assist in formulating a management
ing diabetes (those with IFG, IGT, or A1C of 5.7– 6.4%, IGT, or IFG should be plan, and provide a basis for continuing
both) can be given interventions that sig- counseled on lifestyle changes with goals care. Laboratory tests appropriate to the
nificantly decrease the rate of onset of di- similar to those of the DPP (7% weight evaluation of each patient’s medical con-
S16 DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 care.diabetesjournals.org
7. Position Statement
dition should be performed. A focus on Table 8—Components of the comprehensive diabetes evaluation
the components of comprehensive care Medical history
(Table 8) will assist the health care team to • Age and characteristics of onset of diabetes (e.g., DKA, asymptomatic laboratory finding)
ensure optimal management of the pa- • Eating patterns, physical activity habits, nutritional status, and weight history; growth
tient with diabetes. and development in children and adolescents
• Diabetes education history
B. Management • Review of previous treatment regimens and response to therapy (A1C records)
People with diabetes should receive med- • Current treatment of diabetes, including medications, meal plan, physical activity
ical care from a physician-coordinated patterns, and results of glucose monitoring and patient’s use of data
team. Such teams may include, but are • DKA frequency, severity, and cause
not limited to, physicians, nurse practitio- • Hypoglycemic episodes
ners, physician’s assistants, nurses, dieti- • Hypoglycemia awareness
tians, pharmacists, and mental health • Any severe hypoglycemia: frequency and cause
professionals with expertise and a special • History of diabetes-related complications
interest in diabetes. It is essential in this • Microvascular: retinopathy, nephropathy, neuropathy (sensory, including history of
collaborative and integrated team ap- foot lesions; autonomic, including sexual dysfunction and gastroparesis)
proach that individuals with diabetes as- • Macrovascular: CHD, cerebrovascular disease, PAD
sume an active role in their care. • Other: psychosocial problems*, dental disease*
The management plan should be Physical examination
formulated as a collaborative therapeu- • Height, weight, BMI
tic alliance among the patient and fam- • Blood pressure determination, including orthostatic measurements when indicated
ily, the physician, and other members of • Fundoscopic examination*
the health care team. A variety of strat- • Thyroid palpation
egies and techniques should be used to • Skin examination (for acanthosis nigricans and insulin injection sites)
provide adequate education and devel- • Comprehensive foot examination:
opment of problem-solving skills in the • Inspection
various aspects of diabetes manage- • Palpation of dorsalis pedis and posterior tibial pulses
ment. Implementation of the manage- • Presence/absence of patellar and Achilles reflexes
ment plan requires that each aspect is • Determination of proprioception, vibration, and monofilament sensation
understood and agreed to by the patient Laboratory evaluation
and the care providers and that the goals • A1C, if results not available within past 2–3 months
and treatment plan are reasonable. Any • If not performed/available within past year:
plan should recognize diabetes self- • Fasting lipid profile, including total, LDL and HDL cholesterol and triglycerides
management education (DSME) and • Liver function tests
ongoing diabetes support as an integral • Test for urine albumin excretion with spot urine albumin-to-creatinine ratio
component of care. In developing the • Serum creatinine and calculated GFR
plan, consideration should be given to • Thyroid-stimulating hormone in type 1 diabetes, dyslipidemia, or women over age 50
the patient’s age, school or work sched- years
ule and conditions, physical activity, Referrals
eating patterns, social situation and • Annual dilated eye exam
cultural factors, and presence of com- • Family planning for women of reproductive age
plications of diabetes or other medical • Registered dietitian for MNT
conditions. • DSME
• Dental examination
• Mental health professional, if needed
C. Glycemic control
*See appropriate referrals for these categories.
1. Assessment of glycemic control
Two primary techniques are available for ● For patients using less-frequent insulin in conjunction with intensive insulin
health providers and patients to assess the injections, noninsulin therapies, or regimens can be a useful tool to lower
effectiveness of the management plan on medical nutrition therapy (MNT) A1C in selected adults (age 25 years)
glycemic control: patient self-monitoring alone, SMBG may be useful as a guide to with type 1 diabetes. (A)
of blood glucose (SMBG) or interstitial the success of therapy. (E) ● Although the evidence for A1C-
glucose, and A1C. ● To achieve postprandial glucose tar- lowering is less strong in children,
gets, postprandial SMBG may be appro- teens, and younger adults, CGM may
a. Glucose monitoring priate. (E) be helpful in these groups. Success cor-
● When prescribing SMBG, ensure that relates with adherence to ongoing use
Recommendations patients receive initial instruction in, of the device. (C)
● SMBG should be carried out three or and routine follow-up evaluation of, ● CGM may be a supplemental tool to
more times daily for patients using mul- SMBG technique and their ability to use SMBG in those with hypoglycemia un-
tiple insulin injections or insulin pump data to adjust therapy. (E) awareness and/or frequent hypoglyce-
therapy. (A) ● Continuous glucose monitoring (CGM) mic episodes. (E)
care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 S17
8. Standards of Medical Care
Major clinical trials of insulin-treated pa- hyperglycemic excursions. Small studies Table 9—Correlation of A1C with average
tients that demonstrated the benefits of in selected patients with type 1 diabetes glucose
intensive glycemic control on diabetes have suggested that CGM use reduces the
complications have included SMBG as time spent in hypo- and hyperglycemic Mean plasma glucose
part of multifactorial interventions, sug- ranges and may modestly improve glyce-
gesting that SMBG is a component of ef- mic control. A larger 26-week random- A1C (%) mg/dl mmol/l
fective therapy. SMBG allows patients to ized trial of 322 type 1 patients showed 6 126 7.0
evaluate their individual response to ther- that adults age 25 years and older using 7 154 8.6
apy and assess whether glycemic targets intensive insulin therapy and CGM expe- 8 183 10.2
are being achieved. Results of SMBG can rienced a 0.5% reduction in A1C (from 9 212 11.8
be useful in preventing hypoglycemia and 7.6% to 7.1%) compared to usual in- 10 240 13.4
adjusting medications (particularly pran- tensive insulin therapy with SMBG (45). 11 269 14.9
dial insulin doses), MNT, and physical ac- Sensor use in children, teens, and adults 12 298 16.5
tivity. up to age 24 years did not result in signif- These estimates are based on ADAG data of 2,700
The frequency and timing of SMBG icant A1C lowering, and there was no sig- glucose measurements over 3 months per A1C mea-
should be dictated by the particular needs nificant difference in hypoglycemia in any surement in 507 adults with type 1, type 2, and no
and goals of the patient. SMBG is espe- group. Importantly, the greatest predictor diabetes. The correlation between A1C and average
cially important for patients treated with of A1C-lowering in this study for all age- glucose was 0.92 (51). A calculator for converting
A1C results into estimated average glucose (eAG), in
insulin to monitor for and prevent asymp- groups was frequency of sensor use, either mg/dl or mmol/l, is available at http://
tomatic hypoglycemia and hyperglyce- which was lower in younger age-groups. professional.diabetes.org/eAG.
mia. For most patients with type 1 In a smaller randomized controlled trial of
diabetes and pregnant women taking in- 129 adults and children with baseline
sulin, SMBG is recommended three or A1C 7.0%, outcomes combining A1C quency of A1C testing should be
more times daily. For these populations, and hypoglycemia favored the group uti- dependent on the clinical situation, the
significantly more frequent testing may be lizing CGM, suggesting that CGM is also treatment regimen used, and the judg-
required to reach A1C targets safely with- beneficial for individuals with type 1 dia- ment of the clinician. Some patients with
out hypoglycemia. The optimal frequency betes who have already achieved excellent stable glycemia well within target may do
and timing of SMBG for patients with type control with A1C 7.0 (46). Although well with testing only twice per year,
2 diabetes on noninsulin therapy is un- CGM is an evolving technology, emerging while unstable or highly intensively man-
clear. A meta-analysis of SMBG in non– data suggest that, in appropriately se- aged patients (e.g., pregnant type 1
insulin-treated patients with type 2 lected patients who are motivated to wear women) may be tested more frequently
diabetes concluded that some regimen of it most of the time, it may offer benefit. than every 3 months. The availability of
SMBG was associated with a reduction in CGM may be particularly useful in those the A1C result at the time that the patient
A1C of 0.4%. However, many of the stud- with hypoglycemia unawareness and/or is seen (point-of-care testing) has been re-
ies in this analysis also included patient frequent episodes of hypoglycemia, and ported to result in increased intensifica-
education with diet and exercise counsel- studies in this area are ongoing. tion of therapy and improvement in
ing and, in some cases, pharmacologic in- glycemic control (49,50).
tervention, making it difficult to assess the b. A1C The A1C test is subject to certain lim-
contribution of SMBG alone to improved itations. Conditions that affect erythro-
control (40). Several recent trials have Recommendations cyte turnover (hemolysis, blood loss) and
called into question the clinical utility and ● Perform the A1C test at least two times hemoglobin variants must be considered,
cost-effectiveness of routine SMBG in a year in patients who are meeting treat- particularly when the A1C result does not
non–insulin-treated patients (41– 43). ment goals (and who have stable glyce- correlate with the patient’s clinical situa-
Because the accuracy of SMBG is in- mic control). (E) tion (44). In addition, A1C does not pro-
strument and user dependent (44), it is ● Perform the A1C test quarterly in pa- vide a measure of glycemic variability or
important to evaluate each patient’s mon- tients whose therapy has changed or hypoglycemia. For patients prone to gly-
itoring technique, both initially and at who are not meeting glycemic goals. (E) cemic variability (especially type 1 pa-
regular intervals thereafter. In addition, ● Use of point-of-care testing for A1C al- tients, or type 2 patients with severe
optimal use of SMBG requires proper in- lows for timely decisions on therapy insulin deficiency), glycemic control is
terpretation of the data. Patients should changes, when needed. (E) best judged by the combination of results
be taught how to use the data to adjust of SMBG testing and the A1C. The A1C
food intake, exercise, or pharmacological Because A1C is thought to reflect average may also serve as a check on the accuracy
therapy to achieve specific glycemic goals, glycemia over several months (44), and of the patient’s meter (or the patient’s re-
and these skills should be reevaluated pe- has strong predictive value for diabetes ported SMBG results) and the adequacy of
riodically. complications (47,48), A1C testing the SMBG testing schedule.
CGM through the measurement of in- should be performed routinely in all pa- Table 9 contains the correlation be-
terstitial glucose (which correlates well tients with diabetes, at initial assessment tween A1C levels and mean plasma glu-
with plasma glucose) is available. These and then as part of continuing care. Mea- cose levels based on data from the
sensors require calibration with SMBG, surement approximately every 3 months international A1C-Derived Average Glu-
and the latter are still recommended for determines whether a patient’s glycemic cose (ADAG) trial utilizing frequent
making acute treatment decisions. CGM targets have been reached and main- SMBG and CGM in 507 adults (83% Cau-
devices also have alarms for hypo- and tained. For any individual patient, the fre- casian) with type 1, type 2, and no diabe-
S18 DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 care.diabetesjournals.org
9. Position Statement
tes (51). The American Diabetes the diagnosis of diabetes, is associated The Veterans Affairs Diabetes Trial
Association and American Association of with long-term reduction in macrovas- (VADT) showed significant reductions in
Clinical Chemists have determined that cular disease. Therefore, a reasonable albuminuria with intensive (achieved me-
the correlation (r 0.92) is strong A1C goal for many nonpregnant adults dian A1C 6.9%) compared to standard
enough to justify reporting both an A1C is 7%. (B) glycemic control, but no difference in ret-
result and an estimated average glucose ● Because additional analyses from sev- inopathy and neuropathy (60,61). The
(eAG) result when a clinician orders the eral randomized trials suggest a small Action in Diabetes and Vascular Disease:
A1C test. The table in previous versions of but incremental benefit in microvascu- Preterax and Diamicron Modified Release
the Standards of Medical Care in Diabetes lar outcomes with A1C values closer to Controlled Evaluation (ADVANCE) study
describing the correlation between A1C normal, providers might reasonably of intensive versus standard glycemic
and mean glucose was derived from rela- suggest more stringent A1C goals for control in type 2 diabetes found a statis-
tively sparse data (one 7-point profile selected individual patients, if this can tically significant reduction in albumin-
over 1 day per A1C reading) in the pri- be achieved without significant hypo- uria with an A1C target of 6.5%
marily Caucasian type 1 diabetic partici- glycemia or other adverse effects of (achieved median A1C 6.3%) compared
pants in the DCCT (52). Clinicians treatment. Such patients might include to standard therapy achieving a median
should note that the numbers in the table those with short duration of diabetes, A1C of 7.0% (62). Recent analyses from
are now different, as they are based on long life expectancy, and no significant the Action to Control Cardiovascular Risk
2,800 readings per A1C in the ADAG CVD. (B) in Diabetes (ACCORD) trial have shown
trial. ● Conversely, less stringent A1C goals lower rates of measures of microvascular
In the ADAG trial, there were no sig- may be appropriate for patients with a complications in the intensive glycemic
nificant differences among racial and eth- history of severe hypoglycemia, limited control arm compared with the standard
nic groups in the regression lines between life expectancy, advanced microvascu- arm (63,64).
A1C and mean glucose, although there lar or macrovascular complications, ex- Epidemiological analyses of the
was a trend toward a difference between tensive comorbid conditions, and those DCCT and UKPDS (47,48) demonstrate a
African/African American participants with longstanding diabetes in whom curvilinear relationship between A1C and
and Caucasian ones that might have been the general goal is difficult to attain de- microvascular complications. Such anal-
significant had more African/African spite DSME, appropriate glucose mon- yses suggest that, on a population level,
American participants been studied. A re- itoring, and effective doses of multiple the greatest number of complications will
cent study comparing A1C with CGM glucose-lowering agents including in- be averted by taking patients from very
data in 48 type 1 diabetic children found sulin. (C) poor control to fair or good control. These
a highly statistically significant correla- analyses also suggest that further lowering
tion between A1C and mean blood glu- Glycemic control is fundamental to the of A1C from 7 to 6% is associated with
cose, although the correlation (r 0.7) management of diabetes. The DCCT (47) further reduction in the risk of microvas-
was significantly lower than in the ADAG (in patients with type 1 diabetes), the Ku- cular complications, albeit the absolute
trial (53). Whether there are significant mamoto study (54), and the UK Prospec- risk reductions become much smaller.
differences in how A1C relates to average tive Diabetes Study (UKPDS) (55,56) Given the substantially increased risk of
glucose in children or in African Ameri- (both in patients with type 2 diabetes) hypoglycemia (particularly in those with
can patients is an area for further study. were prospective, randomized, controlled type 1 diabetes, but also in the recent type
For the time being, the question has not trials of intensive versus standard glyce- 2 trials), the concerning mortality find-
led to different recommendations about mic control in patients with relatively re- ings in the ACCORD trial (65), and the
testing A1C or to different interpretations cently diagnosed diabetes. These trials relatively much greater effort required to
of the clinical meaning of given levels of showed definitively that improved glyce- achieve near-normoglycemia, the risks of
A1C in those populations. mic control is associated with signifi- lower targets may outweigh the potential
For patients in whom A1C/eAG and cantly decreased rates of microvascular benefits on microvascular complications
measured blood glucose appear discrep- (retinopathy and nephropathy) and neu- on a population level. However, selected
ant, clinicians should consider the possi- ropathic complications. Follow up of the individual patients, especially those with
bilities of hemoglobinopathy or altered DCCT cohorts in the Epidemiology of Di- little comorbidity and long life expect-
red cell turnover, and the options of more abetes Interventions and Complications ancy (who may reap the benefits of fur-
frequent and/or different timing of SMBG (EDIC) study (57,58) and of the UKPDS ther lowering of glycemia below 7%) may,
or use of CGM. Other measures of chronic cohort (59) has shown persistence of at patient and provider judgment, adopt
glycemia such as fructosamine are avail- these microvascular benefits in previously glycemic targets as close to normal as pos-
able, but their linkage to average glucose intensively treated subjects, even though sible as long as significant hypoglycemia
and their prognostic significance are not their glycemic control has been equiva- does not become a barrier.
as clear as is the case for A1C. lent to that of previous standard arm sub- Whereas many epidemiologic studies
jects during follow-up. and meta-analyses (66,67) have clearly
2. Glycemic goals in adults Subsequent trials in patients with shown a direct relationship between A1C
more long-standing type 2 diabetes, de- and CVD, the potential of intensive glyce-
Recommendations signed primarily to look at the role of mic control to reduce CVD has been less
● Lowering A1C to below or around 7% intensive glycemic control on cardiovas- clearly defined. In the DCCT, there was a
has been shown to reduce microvascu- cular outcomes also confirmed a benefit, trend toward lower risk of CVD events
lar and neuropathic complications of although more modest, on onset or pro- with intensive control. However, 9-year
diabetes and, if implemented soon after gression of microvascular complications. post-DCCT follow-up of the cohort has
care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 S19
10. Standards of Medical Care
shown that participants previously ran- explanation for the excess mortality in the concept, data from an ancillary study of
domized to the intensive arm had a 42% intensive arm. The ACCORD investiga- the VADT demonstrated that intensive
reduction (P 0.02) in CVD outcomes tors subsequently published additional glycemic control was quite effective in re-
and a 57% reduction (P 0.02) in the analyses showing no increase in mortality ducing CVD events in individuals with
risk of nonfatal myocardial infarction in the intensive arm participants who less atherosclerosis at baseline (assessed
(MI), stroke, or CVD death compared achieved A1C levels 7% or in those who by coronary calcium) but not in persons
with those previously in the standard arm lowered their A1C quickly after trial en- with more extensive baseline atheroscle-
(68). The benefit of intensive glycemic rollment. In fact, the converse was ob- rosis (72).
control in this type 1 cohort has recently served—those at highest risk for mortality The evidence for a cardiovascular
been shown to persist for several decades were participants in the intensive arm benefit of intensive glycemic control pri-
(69). with the highest A1C levels (71). marily rests on long-term follow-up of
The UKPDS trial of type 2 diabetes The primary outcome of ADVANCE study cohorts treated early in the course
observed a 16% reduction in cardiovascu- was a combination of microvascular of type 1 and type 2 diabetes and subset
lar complications (combined fatal or non- events (nephropathy and retinopathy) analyses of ACCORD, ADVANCE, and
fatal MI and sudden death) in the and major adverse cardiovascular events VADT. A recent group-level meta-
intensive glycemic control arm, although (MI, stroke, and cardiovascular death). analysis of the latter three trials suggests
this difference was not statistically signif- Intensive glycemic control significantly that glucose lowering has a modest (9%)
icant (P 0.052), and there was no sug- reduced the primary end point, although but statistically significant reduction in
gestion of benefit on other CVD outcomes this was due to a significant reduction in major CVD outcomes, primarily nonfatal
such as stroke. However, 10 years of fol- the microvascular outcome, primarily de- MI, with no significant effect on mortality.
low-up of the UKPDS cohort demon- velopment of macroalbuminuria, with no A prespecified subgroup analysis sug-
strated, for participants originally significant reduction in the macrovascu- gested that major CVD outcome reduc-
randomized to intensive glycemic control lar outcome. There was no difference in tion occurred in patients without known
compared with those randomized to con- overall or cardiovascular mortality be- CVD at baseline (HR 0.84 [95% CI 0.74 –
ventional glycemic control, long-term re- tween the intensive compared with the 0.94]) (73). Conversely, the mortality
ductions in MI (15% with sulfonylurea or standard glycemic control arms (62). findings in ACCORD and subgroup anal-
insulin as initial pharmacotherapy, 33% The VADT randomized participants yses of VADT suggest that the potential
with metformin as initial pharmacother- with type 2 diabetes uncontrolled on in- risks of very intensive glycemic control
apy, both statistically significant) and in sulin or maximal dose oral agents (me- may outweigh its benefits in some pa-
all-cause mortality (13 and 27%, respec- dian entry A1C 9.4%) to a strategy of tients, such as those with very long dura-
tively, both statistically significant) (59). intensive glycemic control (goal A1C tion of diabetes, known history of severe
Results of three large trials 6.0%) or standard glycemic control, hypoglycemia, advanced atherosclerosis,
(ACCORD, ADVANCE, and VADT) sug- with a planned A1C separation of at least and advanced age/frailty. Certainly, pro-
gested no significant reduction in CVD 1.5%. The primary outcome of the VADT viders should be vigilant in preventing se-
outcomes with intensive glycemic control was a composite of CVD events. The cu- vere hypoglycemia in patients with
in these populations, who had more ad- mulative primary outcome was nonsig- advanced disease and should not aggres-
vanced diabetes than UKPDS partici- nificantly lower in the intensive arm (60). sively attempt to achieve near-normal
pants. Details of these three studies are Unlike the UKPDS, which was carried A1C levels in patients in whom such a
reviewed extensively in a recent ADA po- out in patients with newly diagnosed di- target cannot be reasonably easily and
sition statement (70). abetes, all three of the recent type 2 car- safely achieved.
The glycemic control arm of diovascular trials were conducted in Recommended glycemic goals for
ACCORD was halted early due to the participants with established diabetes many nonpregnant adults are shown in
finding of an increased rate of mortality in (mean duration 8 –11 years) and either Table 10. The recommendations are
the intensive arm compared with the stan- known CVD or multiple risk factors, sug- based on those for A1C values, with listed
dard arm (1.41% vs. 1.14% per year; HR gesting the presence of established ath- blood glucose levels that appear to corre-
1.22 [95% CI 1.01 to 1.46]); with a sim- erosclerosis. Subset analyses of the three late with achievement of an A1C of 7%.
ilar increase in cardiovascular deaths. The trials suggested a significant benefit of in- Less-stringent treatment goals may be ap-
primary outcome of ACCORD (MI, tensive glycemic control on CVD in par- propriate for adults with limited life ex-
stroke, or cardiovascular death) was ticipants with shorter duration of pectancies or advanced vascular disease.
lower in the intensive glycemic control diabetes, lower A1C at entry, and/or or Glycemic goals for children are provided
group, due to a reduction in nonfatal MI, absence of known CVD. The DCCT-EDIC in VII.A.1.a. Glycemic control. Severe or
but this reduction was not statistically sig- study and the long-term follow-up of the frequent hypoglycemia is an absolute in-
nificant when the study was terminated UKPDS cohort both suggest that intensive dication for the modification of treatment
(65). glycemic control initiated soon after diag- regimens, including setting higher glyce-
The potential cause of excess deaths nosis of diabetes in patients with a lower mic goals.
in the intensive group of the ACCORD level of CVD risk may impart long-term The issue of pre- versus postprandial
has been difficult to pinpoint. Explor- protection from CVD events. As is the SMBG targets is complex (74). Elevated
atory analyses of the mortality findings of case with microvascular complications, it postchallenge (2-h OGTT) glucose values
ACCORD (evaluating variables including may be that glycemic control plays a have been associated with increased car-
weight gain, use of any specific drug or greater role before macrovascular disease diovascular risk independent of FPG in
drug combination, and hypoglycemia) is well developed and minimal or no role some epidemiological studies. In diabetic
were reportedly unable to identify a clear when it is advanced. Consistent with this subjects, some surrogate measures of vas-
S20 DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 care.diabetesjournals.org
11. Position Statement
Table 10—Summary of glycemic recommendations for many nonpregnant adults with type 1 diabetes consists of the following
diabetes components: 1) use of multiple dose in-
A1C 7.0%* sulin injections (three to four injections
Preprandial capillary plasma glucose 70–130 mg/dl* (3.9–7.2 mmol/l) per day of basal and prandial insulin) or
Peak postprandial capillary plasma glucose† 180 mg/dl* ( 10.0 mmol/l) CSII therapy; 2) matching of prandial in-
• Goals should be individualized based on*: sulin to carbohydrate intake, premeal
• duration of diabetes blood glucose, and anticipated activity;
• age/life expectancy and 3) for many patients (especially if hy-
• comorbid conditions poglycemia is a problem), use of insulin
• known CVD or advanced microvascular analogs. There are excellent reviews avail-
complications able that guide the initiation and manage-
• hypoglycemia unawareness ment of insulin therapy to achieve desired
• individual patient considerations glycemic goals (3,79,81).
• More or less stringent glycemic goals may Because of the increased frequency of
be appropriate for individual patients. other autoimmune diseases in type 1 dia-
• Postprandial glucose may be targeted if betes, screening for thyroid dysfunction,
A1C goals are not met despite reaching vitamin B12 deficiency, or celiac disease
preprandial glucose goals. should be considered based on signs and
Postprandial glucose measurements should be made 1–2 h after the beginning of the meal, generally peak symptoms. Periodic screening in absence
levels in patients with diabetes. of symptoms has been recommended, but
the effectiveness and optimal frequency
are unclear.
cular pathology, such as endothelial dys- • 2-h postmeal 120 mg/dl (6.7
function, are negatively affected by mmol/l)
postprandial hyperglycemia (75). It is For women with preexisting type 1 or 2. Therapy for type 2 diabetes
clear that postprandial hyperglycemia, type 2 diabetes who become pregnant, a The ADA and the EASD published an ex-
like preprandial hyperglycemia, contrib- recent consensus statement (78) recom- pert consensus statement on the approach
utes to elevated A1C levels, with its rela- mended the following as optimal glyce- to management of hyperglycemia in indi-
tive contribution being higher at A1C mic goals, if they can be achieved without viduals with type 2 diabetes (82). High-
levels that are closer to 7%. However, out- excessive hypoglycemia: lights of this approach are: intervention at
come studies have clearly shown A1C to the time of diagnosis with metformin in
be the primary predictor of complica- ● premeal, bedtime, and overnight glu- combination with lifestyle changes (MNT
tions, and landmark glycemic control tri- cose 60 –99 mg/dl (3.3–5.4 mmol/l) and exercise) and continuing timely aug-
als such as the DCCT and UKPDS relied ● peak postprandial glucose 100 –129 mentation of therapy with additional
overwhelmingly on preprandial SMBG. mg/dl (5.4 –7.1mmol/l) agents (including early initiation of insu-
Additionally, a randomized controlled ● A1C 6.0% lin therapy) as a means of achieving and
trial in patients with known CVD found maintaining recommended levels of gly-
no CVD benefit of insulin regimens tar- D. Pharmacologic and overall cemic control (i.e., A1C 7% for most
geting postprandial glucose compared approaches to treatment patients). As A1C targets are not achieved,
with targeting preprandial glucose (76). A treatment intensification is based on the
reasonable recommendation for post- 1. Therapy for type 1 diabetes addition of another agent from a different
prandial testing and targets is that for in- The DCCT clearly showed that intensive class. The overall objective is to achieve
dividuals who have premeal glucose insulin therapy (three or more injections and maintain glycemic control and to
values within target but have A1C values per day of insulin, or continuous subcu- change interventions when therapeutic
above target, monitoring postprandial taneous insulin infusion (CSII) (insulin goals are not being met.
plasma glucose (PPG) 1–2 h after the start pump therapy) was a key part of im- The algorithm took into account the
of the meal and treatment aimed at reduc- proved glycemia and better outcomes evidence for A1C-lowering of the individ-
ing PPG values to 180 mg/dl may help (47,68). At the time of the study, therapy ual interventions, their additive effects,
lower A1C. was carried out with short- and interme- and their expense. The precise drugs used
As regards goals for glycemic control diate-acting human insulins. Despite bet- and their exact sequence may not be as
for women with GDM, recommendations ter microvascular outcomes, intensive important as achieving and maintaining
from the Fifth International Workshop- insulin therapy was associated with a high glycemic targets safely. Medications not
Conference on Gestational Diabetes (77) rate in severe hypoglycemia (62 episodes included in the consensus algorithm, ow-
were to target maternal capillary glucose per 100 patient-years of therapy). Since ing to less glucose-lowering effectiveness,
concentrations of: the time of the DCCT, a number of rapid- limited clinical data, and/or relative ex-
acting and long-acting insulin analogs pense, still may be appropriate choices in
• Preprandial 95 mg/dl (5.3 mmol/l) have been developed. These analogs are individual patients to achieve glycemic
and either associated with less hypoglycemia with goals. Initiation of insulin at time of diagno-
• 1-h postmeal 140 mg/dl (7.8 equal A1C-lowering in type 1 diabetes sis is recommended for individuals present-
mmol/l) (79,80). ing with weight loss or other severe
or Therefore, recommended therapy for hyperglycemic symptoms or signs.
care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 1, JANUARY 2011 S21