2. Standards of Medical Care
Table 1—ADA evidence grading system for clinical practice recommendations same tests would be used for “screening”
as for diagnosis. Type 2 diabetes has a
Level of long asymptomatic phase and significant
evidence Description clinical risk markers. Diabetes may be
identified anywhere along a spectrum of
A Clear evidence from well-conducted, generalizable, randomized controlled clinical scenarios ranging from a seem-
trials that are adequately powered, including: ingly low-risk individual who happens to
● Evidence from a well-conducted multicenter trial have glucose testing, to a higher-risk in-
● Evidence from a meta-analysis that incorporated quality ratings in the dividual whom the provider tests because
analysis
Compelling nonexperimental evidence, i.e., “all or none” rule developed of high suspicion of diabetes, to the symp-
by the Centre for Evidence-Based Medicine at Oxford tomatic patient. The discussion herein is
Supportive evidence from well-conducted randomized controlled trials primarily framed as testing for diabetes in
that are adequately powered, including: those without symptoms. Testing for dia-
● Evidence from a well-conducted trial at one or more institutions
● Evidence from a meta-analysis that incorporated quality ratings in the
betes will also detect individuals with pre-
analysis diabetes.
B Supportive evidence from well-conducted cohort studies, including:
● Evidence from a well-conducted prospective cohort study or registry A. Testing for pre-diabetes and type
● Evidence from a well-conducted meta-analysis of cohort studies 2 diabetes in adults
Supportive evidence from a well-conducted case-control study Type 2 diabetes is frequently not diag-
C Supportive evidence from poorly controlled or uncontrolled studies nosed until complications appear, and
● Evidence from randomized clinical trials with one or more major or approximately one-third of all people
three or more minor methodological flaws that could invalidate the with diabetes may be undiagnosed. Al-
results though the effectiveness of early identifi-
● Evidence from observational studies with high potential for bias (such
as case series with comparison to historical controls) cation of pre-diabetes and diabetes
● Evidence from case series or case reports through mass testing of asymptomatic in-
Conflicting evidence with the weight of evidence supporting the dividuals has not been definitively proven
recommendation (and rigorous trials to provide such proof
E Expert consensus or clinical experience are unlikely to occur), pre-diabetes and
diabetes meet established criteria for con-
ditions in which early detection is appro-
rized as either impaired fasting glucose be carried out at least at 3-year inter- priate. Both conditions are common,
(IFG) or impaired glucose tolerance vals. (E) increasing in prevalence, and impose sig-
(IGT), depending on whether it is identi- ● To test for pre-diabetes or diabetes, an nificant public health burdens. There is a
fied through the FPG or the OGTT: FPG test or 2-h OGTT (75-g glucose long presymptomatic phase before the di-
load) or both are appropriate. (B) agnosis of type 2 diabetes is usually made.
● IFG FPG 100 mg/dl (5.6 mmol/l) to ● An OGTT may be considered in pa- Relatively simple tests are available to de-
125 mg/dl (6.9 mmol/l) tients with IFG to better define the risk tect preclinical disease (8). Additionally,
● IGT 2-h plasma glucose 140 mg/dl of diabetes. (E) the duration of glycemic burden is a
(7.8 mmol/l) to 199 mg/dl (11.0 ● In those identified with pre-diabetes, strong predictor of adverse outcomes,
mmol/l) identify and, if appropriate, treat other and effective interventions exist to pre-
CVD risk factors. (B) vent progression of pre-diabetes to diabe-
IFG and IGT have been officially termed tes (see Section IV) and to reduce risk of
“pre-diabetes.” Both categories of pre- For many illnesses, there is a major dis- complications of diabetes (see Section
diabetes are risk factors for future diabetes tinction between screening and diagnos- VI).
and for cardiovascular disease (CVD) (7). tic testing. However, for diabetes, the Recommendations for testing for pre-
II. TESTING FOR PRE-
Table 2—Criteria for the diagnosis of diabetes
DIABETES AND DIABETES
IN ASYMPTOMATIC 1. FPG 126 mg/dl (7.0 mmol/l). Fasting is defined as no caloric intake for at
PATIENTS least 8 h.*
OR
Recommendations 2. Symptoms of hyperglycemia and a casual (random) plasma glucose 200
● Testing to detect pre-diabetes and type mg/dl (11.1 mmol/l). Casual (random) is defined as any time of day without
2 diabetes in asymptomatic people regard to time since last meal. The classic symptoms of hyperglycemia
should be considered in adults of any include polyuria, polydipsia, and unexplained weight loss.
age who are overweight or obese (BMI OR
25 kg/m2) and who have one or more 3. 2-h plasma glucose 200 mg/dl (11.1 mmol/l) during an OGTT. The test
additional risk factors for diabetes (Ta- should be performed as described by the World Health Organization using
ble 3). In those without these risk fac- a glucose load containing the equivalent of 75-g anhydrous glucose
tors, testing should begin at age 45 dissolved in water.*
years. (B) *In the absence of unequivocal hyperglycemia, these criteria should be confirmed by repeat testing on a
● If tests are normal, repeat testing should different day (5).
S14 DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009
3. Position Statement
Table 3—Criteria for testing for pre-diabetes and diabetes in asymptomatic adult individuals relatives with type 1 diabetes, in the context
1. 2
Testing should be considered in all adults who are overweight (BMI 25 kg/m *) and
of clinical research studies (see, for exam-
have additional risk factors:
ple, http://www2.diabetestrialnet.org).
● physical inactivity
Widespread clinical testing of asymptom-
● first-degree relative with diabetes atic low-risk individuals cannot currently
● members of a high-risk ethnic population (e.g., African American, Latino, Native be recommended, as it would identify very
American, Asian American, Pacific Islander) few individuals in the general population
● women who delivered a baby weighing 9 lb or were diagnosed with GDM who are at risk. Individuals who screen pos-
● hypertension ( 140/90 mmHg or on therapy for hypertension)
● HDL cholesterol level 35 mg/dl (0.90 mmol/l) and/or a triglyceride level 250
itive should be counseled about their risk of
mg/dl (2.82 mmol/l) developing diabetes. Clinical studies are be-
● women with polycystic ovarian syndrome (PCOS) ing conducted to test various methods of
● IGT or IFG on previous testing preventing type 1 diabetes, or reversing
● other clinical conditions associated with insulin resistance (e.g., severe obesity,
acanthosis nigricans)
early type 1 diabetes, in those with evidence
● history of CVD of autoimmunity.
2. In the absence of the above criteria, testing for pre-diabetes and diabetes should begin
at age 45 years III. DETECTION AND
3. If results are normal, testing should be repeated at least at 3-year intervals, with DIAGNOSIS OF GDM
consideration of more frequent testing depending on initial results and risk status.
*At-risk BMI may be lower in some ethnic groups.
Recommendations
● Screen for GDM using risk factor anal-
ysis and, if appropriate, use of an
diabetes and diabetes in asymptomatic, tests may not seek, or have access to, ap- OGTT. (C)
undiagnosed adults are listed in Table 3. propriate follow-up testing and care. ● Women with GDM should be screened
Testing should be considered in adults of Conversely, there may be failure to ensure for diabetes 6 –12 weeks postpartum
any age with BMI 25 kg/m2 and one or appropriate repeat testing for individuals and should be followed up with subse-
more risk factors for diabetes. Because age who test negative. Community screening quent screening for the development of
is a major risk factor for diabetes, testing may also be poorly targeted, i.e., it may diabetes or pre-diabetes. (E)
of those without other risk factors should fail to reach the groups most at risk and
begin no later than age 45 years. inappropriately test those at low risk (the GDM is defined as any degree of glucose
Either FPG testing or the 2-h OGTT is worried well) or even those already diag- intolerance with onset or first recognition
appropriate for testing. The 2-h OGTT nosed (18,19). during pregnancy (4). Although most
identifies people with either IFG or IGT, cases resolve with delivery, the definition
and thus, more pre-diabetic people at in- B. Testing for type 2 diabetes in applies whether or not the condition per-
creased risk for the development of dia- children sists after pregnancy and does not exclude
betes and CVD. It should be noted that The incidence of type 2 diabetes in ado- the possibility that unrecognized glucose
the two tests do not necessarily detect the lescents has increased dramatically in the intolerance may have antedated or begun
same pre-diabetic individuals (9). The ef- last decade, especially in minority popu- concomitantly with the pregnancy. Ap-
ficacy of interventions for primary pre- lations (20), although the disease remains
vention of type 2 diabetes (10 –16) has rare in the general adolescent population
primarily been demonstrated among in- (21). Consistent with recommendations Table 4—Testing for type 2 diabetes in
dividuals with IGT, not individuals with for adults, children and youth at in- asymptomatic children
IFG (who do not also have IGT). As noted creased risk for the presence or the devel- Criteria:
in the diagnosis section (Section I.B), the opment of type 2 diabetes should be ● Overweight (BMI 85th percentile for
FPG test is more convenient, more repro- tested within the health care setting (22). age and sex, weight for height 85th
ducible, less costly, and easier to admin- The recommendations of the ADA con- percentile, or weight 120% of ideal for
ister than the 2-h OGTT (4,5). An OGTT sensus statement on type 2 diabetes in height)
may be useful in patients with IFG to bet- children and youth, with some modifica- Plus any two of the following risk factors:
ter define the risk of diabetes. tions, are summarized in Table 4. ● Family history of type 2 diabetes in first-
or second-degree relative
The appropriate interval between ● Race/ethnicity (Native American, African
tests is not known (17). The rationale for C. Screening for type 1 diabetes American, Latino, Asian American, Pacific
the 3-year interval is that false-negatives Generally, people with type 1 diabetes Islander)
will be repeated before substantial time present with acute symptoms of diabetes ● Signs of insulin resistance or conditions
elapses, and there is little likelihood that and markedly elevated blood glucose lev- associated with insulin resistance
(acanthosis nigricans, hypertension,
an individual will develop significant els, and most cases are diagnosed soon dyslipidemia, PCOS, or small-for-
complications of diabetes within 3 years after the onset of hyperglycemia. How- gestational-age birthweight)
of a negative test result. ever, evidence from type 1 prevention ● Maternal history of diabetes or GDM
Because of the need for follow-up and studies suggests that measurement of islet during the child’s gestation
discussion of abnormal results, testing autoantibodies identifies individuals who Age of initiation: age 10 years or at onset of
should be carried out within the health are at risk for developing type 1 diabetes. puberty, if puberty occurs at a younger
care setting. Community screening out- Such testing may be appropriate in high- age
side a health care setting is not recom- risk individuals, such as those with prior Frequency: every 3 years
mended because people with positive transient hyperglycemia or those who have Test: FPG preferred
DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009 S15
4. Standards of Medical Care
Table 5—Screening for and diagnosis of GDM IV. PREVENTION/DELAY
Carry out GDM risk assessment at the first prenatal visit.
OF TYPE 2 DIABETES
Women at very high risk for GDM should be screened for diabetes as soon as possible after
the confirmation of pregnancy. Criteria for very high risk are: Recommendations
● Patients with IGT (A) or IFG (E) should
● severe obesity
● prior history of GDM or delivery of large-for-gestational-age infant
be referred to an effective ongoing sup-
● presence of glycosuria
port program for weight loss of 5–10%
● diagnosis of PCOS
of body weight and for increasing phys-
● strong family history of type 2 diabetes
ical activity to at least 150 min per week
Screening/diagnosis at this stage of pregnancy should use standard diagnostic testing (Table 2). of moderate activity such as walking.
● Follow-up counseling appears to be im-
All women of greater than low risk of GDM, including those above not found to have
diabetes early in pregnancy, should undergo GDM testing at 24–28 weeks of gestation. portant for success. (B)
● Based on potential cost savings of dia-
Low risk status, which does not require GDM screening, is defined as women with ALL of
the following characteristics: betes prevention, such counseling
● age 25 years
should be covered by third-party pay-
● weight normal before pregnancy
ors. (E)
● In addition to lifestyle counseling, met-
● member of an ethnic group with a low prevalence of diabetes
● no known diabetes in first-degree relatives
formin may be considered in those who
● no history of abnormal glucose tolerance
are at very high risk for developing di-
● no history of poor obstetrical outcome
abetes (combined IFG and IGT plus
Two approaches may be followed for GDM screening at 24–28 weeks: other risk factors such as A1C 6%,
1. Two-step approach: hypertension, low HDL cholesterol, el-
A. Perform initial screening by measuring plasma or serum glucose 1 h after a 50-g oral evated triglycerides, or family history of
glucose load. A glucose threshold after 50-g load of 140 mg/dl identifies 80% of diabetes in a first-degree relative) and
women with GDM, while the sensitivity is further increased to 90% by a threshold of who are obese and under 60 years of
130 mg/dl. age. (E)
● Monitoring for the development of di-
B. Perform a diagnostic 100-g OGTT on a separate day in women who exceed the
chosen threshold on 50-g screening. abetes in those with pre-diabetes
2. One-step approach (may be preferred in clinics with high prevalence of GDM): Perform should be performed every year. (E)
a diagnostic 100-g OGTT in all women to be tested at 24–28 weeks.
The 100-g OGTT should be performed in the morning after an overnight fast of at least 8 h. Randomized controlled trials have shown
To make a diagnosis of GDM, at least two of the following plasma glucose values must be found: that individuals at high risk for develop-
Fasting: 95 mg/dl ing diabetes (those with IFG, IGT, or
1 h: 180 mg/dl both) can be given interventions that sig-
2 h: 155 mg/dl nificantly decrease the rate of onset of di-
3 h: 140 mg/dl abetes (10 –16). These interventions
include intensive lifestyle modification
programs that have been shown to be very
proximately 7% of all pregnancies (rang- These results have led to careful reconsid- effective ( 58% reduction after 3 years)
ing from 1 to 14% depending on the eration of the diagnostic criteria for GDM. and use of the pharmacologic agents met-
population studied and the diagnostic An international group representing mul- formin, acarbose, orlistat, and thiazo-
tests employed) are complicated by GDM, tiple obstetrical and diabetes organiza- lidinediones (TZDs), each of which has
resulting in more than 200,000 cases tions, including ADA, is currently been shown to decrease incident diabetes
annually. working on consensus toward 1) a world- to various degrees. A summary of major
Because of the risks of GDM to the wide standard for which diagnostic test to diabetes prevention trials is shown in Ta-
mother and neonate, screening and di- use for GDM and 2) rational diagnostic ble 6.
agnosis are warranted. The screening cut points. Two studies of lifestyle intervention
and diagnostic strategies, based on the Because women with a history of have shown persistent reduction in the
2004 ADA position statement on gesta- rate of conversion to type 2 diabetes with
GDM have a greatly increased subse-
tional diabetes mellitus (23), are outlined in 3 (26) to 14 years (27) of postintervention
quent risk for diabetes (25), they should
Table 5. follow-up.
Results of the Hyperglycemia and Ad- be screened for diabetes 6 –12 weeks Based on the results of clinical trials
verse Pregnancy Outcomes study (24), a postpartum, using nonpregnant OGTT and the known risks of progression of
large-scale (including 25,000 pregnant criteria, and should be followed up with pre-diabetes to diabetes, an ADA Consen-
women) multinational epidemiologic subsequent screening for the develop- sus Development Panel (7) concluded
study, demonstrated that risk of adverse ment of diabetes or pre-diabetes, as out- that persons with pre-diabetes (IGT
maternal, fetal, and neonatal outcomes lined in Section II. For information on and/or IFG) should be counseled on life-
continuously increased as a function of the National Diabetes Education Pro- style changes with goals similar to those of
maternal glycemia at 24 –28 weeks, even gram (NDEP) campaign to prevent type the Diabetes Prevention Program (DPP)
within ranges previously considered nor- 2 diabetes in women with GDM, go to (5–10% weight loss and moderate physi-
mal for pregnancy. For most complica- www.ndep.nih.gov/diabetes/pubs/ cal activity of 30 min per day). Regard-
tions, there was no threshold for risk. NeverTooEarly_Tipsheet.pdf. ing the more difficult issue of drug
S16 DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009
5. Position Statement
Table 6—Therapies proven effective in diabetes prevention trials
Mean Conversion in
age Duration Intervention control subjects
Study (ref.) n Population (years) (years) (daily dose) (%/year) Relative risk
Lifestyle
Finnish DPS (11) 522 IGT, BMI 25 kg/m2 55 3.2 Individual 6 0.42 (0.30–0.70)
diet/exercise
DPP (10) 2,161* IGT, BMI 24 kg/m2, 51 3 Individual 10 0.42 (0.34–0.52)
FPG 5.3 mmol/l diet/exercise
Da Qing (12) 259* IGT (randomized 45 6 Group diet/exercise 16 0.62 (0.44–0.86)
groups)
Toranomon study (28) 458 IGT (men), BMI 24 55 4 Individual 2 0.33 (0.10–1.0)†
kg/m2 diet/exercise
Indian DPP (16) 269* IGT 46 2.5 Individual 22 0.71 (0.63–0.79)
diet/exercise
Medications
DPP (10) 2,155* IGT, BMI 24 kg/m2, 51 2.8 Metformin (1,700 mg) 10 0.69 (0.57–0.83)
FPG 5.3 mmol/l
Indian DPP (16) 269* IGT 46 2.5 Metformin (500 mg) 22 0.74 (0.65–0.81)
STOP NIDDM (14) 1,419 IGT, FPG 5.6 mmol/l 54 3.2 Acarbose (300 mg) 13 0.75 (0.63–0.90)
XENDOS (29) 3,277 BMI 30 kg/m2 43 4 Orlistat (360 mg) 2 0.63 (0.46–0.86)
DREAM (15) 5,269 IGT or IFG 55 3.0 Rosiglitazone (8 mg) 9 0.40 (0.35–0.46)
*Number of participants in the indicated comparisons, not necessarily in the entire study. †Calculated from information in the article. DPP, Diabetes Prevention
Program; DREAM, Diabetes REduction Assessment with ramipril and rosiglitazone Medication; DPS, Diabetes Prevention Study; STOP NIDDM, Study to Prevent
Non-Insulin Dependent Diabetes; XENDOS, Xenical in the prevention of Diabetes in Obese Subjects. This table has been reprinted with permission (30) with some
modification.
therapy for diabetes prevention, the con- ensure optimal management of the pa- patient’s age, school or work schedule
sensus panel felt that metformin should tient with diabetes. and conditions, physical activity, eating
be the only drug considered for use in patterns, social situation and personality,
diabetes prevention. For other drugs, the B. Management cultural factors, and presence of compli-
issues of cost, side effects, and lack of per- People with diabetes should receive med- cations of diabetes or other medical con-
sistence of effect in some studies led the ical care from a physician-coordinated ditions.
panel to not recommend their use for di- team. Such teams may include, but are
abetes prevention. Metformin use was not limited to, physicians, nurse practitio- C. Glycemic control
recommended only for very-high-risk in- ners, physician’s assistants, nurses, dieti-
dividuals (those with combined IGT and tians, pharmacists, and mental health 1. Assessment of glycemic control
IFG who are obese and under 60 years of professionals with expertise and a special Two primary techniques are available for
age with at least one other risk factor for interest in diabetes. It is essential in this health providers and patients to assess the
diabetes). In addition, the panel high- collaborative and integrated team ap- effectiveness of the management plan on
lighted the evidence that in the DPP, met- proach that individuals with diabetes as- glycemic control: patient self-monitoring
formin was most effective compared to sume an active role in their care. of blood glucose (SMBG) or of interstitial
lifestyle in those with BMI of at least 35 The management plan should be for- glucose and measurement of A1C.
kg/m2 and those under age 60 years. mulated as an individualized therapeutic
alliance among the patient and family, the a. Glucose monitoring
V. DIABETES CARE physician, and other members of the
health care team. A variety of strategies Recommendations
A. Initial evaluation and techniques should be used to provide ● SMBG should be carried out three or
A complete medical evaluation should be adequate education and development of more times daily for patients using mul-
performed to classify the diabetes, detect problem-solving skills in the various as- tiple insulin injections or insulin pump
the presence of diabetes complications, pects of diabetes management. Imple- therapy. (A)
review previous treatment and glycemic mentation of the management plan ● For patients using less frequent insulin
control in patients with established diabe- requires that each aspect is understood injections, noninsulin therapies, or
tes, assist in formulating a management and agreed on by the patient and the care medical nutrition therapy (MNT) and
plan, and provide a basis for continuing providers and that the goals and treat- physical activity alone, SMBG may be
care. Laboratory tests appropriate to the ment plan are reasonable. Any plan useful as a guide to the success of ther-
evaluation of each patient’s medical con- should recognize diabetes self-manage- apy. (E)
dition should be performed. A focus on ment education (DSME) as an integral ● To achieve postprandial glucose tar-
the components of comprehensive care component of care. In developing the gets, postprandial SMBG may be appro-
(Table 7) will assist the health care team to plan, consideration should be given to the priate. (E)
DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009 S17
6. Standards of Medical Care
Table 7—Components of the comprehensive diabetes evaluation their individual response to therapy and
Medical history assess whether glycemic targets are being
● age and characteristics of onset of diabetes (e.g., DKA, asymptomatic laboratory finding)
achieved. Results of SMBG can be useful
● eating patterns, physical activity habits, nutritional status, and weight history; growth
in preventing hypoglycemia and adjust-
and development in children and adolescents ing medications (particularly prandial in-
● diabetes education history sulin doses), MNT, and physical activity.
● review of previous treatment regimens and response to therapy (A1C records) The frequency and timing of SMBG
● current treatment of diabetes, including medications, meal plan, physical activity should be dictated by the particular needs
patterns, and results of glucose monitoring and patient’s use of data and goals of the patients. SMBG is espe-
● DKA frequency, severity, and cause cially important for patients treated with
● hypoglycemic episodes insulin to monitor for and prevent asymp-
● hypoglycemia awareness tomatic hypoglycemia and hyperglyce-
● any severe hypoglycemia: frequency and cause mia. For most patients with type 1
● history of diabetes-related complications diabetes and pregnant women taking in-
● microvascular: retinopathy, nephropathy, neuropathy (sensory, including history of sulin, SMBG is recommended three or
foot lesions; autonomic, including sexual dysfunction and gastroparesis) more times daily. For this population, sig-
● macrovascular: CHD, cerebrovascular disease, PAD nificantly more frequent testing may be
● other: psychosocial problems,* dental disease* required to reach A1C targets safely with-
Physical examination out hypoglycemia. The optimal frequency
● height, weight, BMI and timing of SMBG for patients with type
● blood pressure determination, including orthostatic measurements when indicated 2 diabetes on noninsulin therapy is un-
● fundoscopic examination* clear. A meta-analysis of SMBG in non–
● thyroid palpation insulin-treated patients with type 2
● skin examination (for acanthosis nigricans and insulin injection sites) diabetes concluded that some regimen of
● comprehensive foot examination: SMBG was associated with a reduction in
● inspection A1C of 0.4%. However, many of the
● palpation of dorsalis pedis and posterior tibial pulses studies in this analysis also included pa-
● presence/absence of patellar and Achilles reflexes tient education with diet and exercise
● determination of proprioception, vibration, and monofilament sensation counseling and, in some cases, pharma-
Laboratory evaluation cologic intervention, making it difficult to
● A1C, if results not available within past 2–3 months assess the contribution of SMBG alone to
If not performed/available within past year: improved control (33). Several recent tri-
● fasting lipid profile, including total, LDL- and HDL-cholesterol and triglycerides als have called into question the clinical
● liver function tests utility and cost-effectiveness of routine
● test for urine albumin excretion with spot urine albumin/creatinine ratio SMBG in non–insulin-treated patients
● serum creatinine and calculated GFR (34 –36).
● thyroid-stimulating hormone in type 1 diabetes, dyslipidemia or women over age 50 Because the accuracy of SMBG is instru-
Referrals ment and user dependent (37), it is impor-
● annual dilated eye exam tant to evaluate each patient’s monitoring
● family planning for women of reproductive age technique, both initially and at regular in-
● registered dietitian for MNT tervals thereafter. In addition, optimal use
● diabetes self-management education of SMBG requires proper interpretation of
● dental examination the data. Patients should be taught how to
● mental Health professional, if needed use the data to adjust food intake, exercise,
*See appropriate referrals for these categories. or pharmacological therapy to achieve spe-
cific glycemic goals, and these skills should
be reevaluated periodically.
● When prescribing SMBG, ensure that ● CGM may be a supplemental tool to CGM through the measurement of in-
patients receive initial instruction in, SMBG in those with hypoglycemia un- terstitial glucose (which correlates well
and routine follow-up evaluation of, awareness and/or frequent hypoglyce- with plasma glucose) is available. These
SMBG technique and their ability to use mic episodes. (E) sensors require calibration with SMBG,
data to adjust therapy. (E) and the latter are still recommended for
● Continuous glucose monitoring (CGM) The ADA’s consensus and position state- making acute treatment decisions. CGM
in conjunction with intensive insulin ments on SMBG provide a comprehensive devices also have alarms for hypo- and
regimens can be a useful tool to lower review of the subject (31,32). Major clin- hyperglycemic excursions. Small studies
A1C in selected adults (age 25 years) ical trials of insulin-treated patients that in selected patients with type 1 diabetes
with type 1 diabetes (A). demonstrated the benefits of intensive have suggested that CGM use reduces the
● Although the evidence for A1C lowering glycemic control on diabetes complica- time spent in hypo- and hyperglycemic
is less strong in children, teens, and tions have included SMBG as part of mul- ranges and may modestly improve glyce-
younger adults, CGM may be helpful in tifactorial interventions, suggesting that mic control. A larger 26-week random-
these groups. Success correlates with ad- SMBG is a component of effective ther- ized trial of 322 type 1 patients showed
herence to ongoing use of the device. (C) apy. SMBG allows patients to evaluate that adults age 25 years and older using
S18 DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009
7. Position Statement
intensive insulin therapy and CGM expe- Table 8—Correlation of A1C with average ferent, as they are based on 2,800 read-
rienced a 0.5% reduction in A1C (from glucose ings per A1C in the ADAG trial.
7.6 to 7.1%) compared with usual in- In the ADAG study, there were no sig-
tensive insulin therapy with SMBG (38). Mean plasma glucose nificant differences among racial and eth-
Sensor use in children, teens, and adults nic groups in the regression lines between
to age 24 years did not result in significant A1C (%) mg/dl mmol/l A1C and mean glucose, although there
A1C lowering, and there was no signifi- 6 126 7.0 was a trend toward a difference between
cant difference in hypoglycemia in any 7 154 8.6 African/African-American and Caucasian
group. Importantly, the greatest predictor 8 183 10.2 participants’ regression lines that might
of A1C lowering in this study for all age- 9 212 11.8 have been significant had more African/
groups was frequency of sensor use, 10 240 13.4 African-American participants been stud-
which was lower in younger age-groups. 11 269 14.9 ied. A recent study comparing A1C to
Although CGM is an evolving technology, 12 298 16.5 CGM data in 48 type 1 children found a
emerging data suggest that, in appropri- Estimates based on ADAG data of 2,700 glucose
highly statistically significant correlation
ately selected patients who are motivated measurements over 3 months per A1C measure- between A1C and mean blood glucose,
to wear it most of the time, it may offer ment in 507 adults with type 1, type 2, and no dia- although the correlation (r 0.7) was sig-
benefit. CGM may be particularly useful betes. Correlation between A1C and average nificantly lower than in the ADAG trial
in those with hypoglycemia unawareness glucose: 0.92 (42). A calculator for converting A1C (44). Whether there are significant differ-
results into eAG, in either mg/dl or mmol/l, is avail-
and/or frequent episodes of hypoglyce- able at http://professional.diabetes.org/eAG. ences in how A1C relates to average glu-
mia, and studies in this area are ongoing. cose in children or in African-American
patients is an area for further study. For
b. A1C itations. Conditions that affect erythro- the time being, the question has not led to
cyte turnover (hemolysis, blood loss) and different recommendations about testing
Recommendations hemoglobin variants must be considered, A1C or to different interpretations of the
● Perform the A1C test at least two times particularly when the A1C result does not clinical meaning of given levels of A1C in
a year in patients who are meeting treat- correlate with the patient’s clinical situa- those populations.
ment goals (and who have stable glyce- tion (37). In addition, A1C does not pro- For patients in whom A1C/eAG and
mic control). (E) vide a measure of glycemic variability or measured blood glucose appear discrep-
● Perform the A1C test quarterly in pa- hypoglycemia. For patients prone to gly- ant, clinicians should consider the possi-
tients whose therapy has changed or cemic variability (especially type 1 pa- bilities of hemoglobinopathy or altered
who are not meeting glycemic goals. (E) tients, or type 2 patients with severe red cell turnover and the options of more
● Use of point-of-care testing for A1C al- insulin deficiency), glycemic control is frequent and/or different timing of SMBG
lows for timely decisions on therapy best judged by the combination of results or use of CGM. Other measures of chronic
changes, when needed. (E) of SMBG testing and the A1C. The A1C glycemia such as fructosamine are avail-
may also serve as a check on the accuracy able, but their linkage to average glucose
Because A1C is thought to reflect average of the patient’s meter (or the patient’s re- and their prognostic significance are not
glycemia over several months (37), and ported SMBG results) and the adequacy of as clear as is the case for A1C.
has strong predictive value for diabetes the SMBG testing schedule.
complications (10,39), A1C testing Table 8 contains the correlation be-
should be performed routinely in all pa- tween A1C levels and mean plasma glu- 2. Glycemic goals in adults
tients with diabetes at initial assessment cose levels based on data from the ● Lowering A1C to below or around 7%
and then as part of continuing care. Mea- international A1C-Derived Average Glu- has been shown to reduce microvascu-
surement approximately every 3 months cose (ADAG) trial utilizing frequent lar and neuropathic complications of
determines whether a patient’s glycemic SMBG and continuous glucose monitor- type 1 and type 2 diabetes. Therefore,
targets have been reached and main- ing in 507 adults (83% Caucasian) with for microvascular disease prevention,
tained. For any individual patient, the fre- type 1, type 2, and no diabetes (49) The the A1C goal for nonpregnant adults in
quency of A1C testing should be ADA and American Association of Clini- general is 7%. (A)
dependent on the clinical situation, the cal Chemists have determined that the ● In type 1 and type 2 diabetes, random-
treatment regimen used, and the judg- correlation (r 0.92) is strong enough to ized controlled trials of intensive versus
ment of the clinician. Some patients with justify reporting both an A1C result and standard glycemic control have not
stable glycemia well within target may do an estimated average glucose (eAG) result shown a significant reduction in CVD
well with testing only twice per year, when a clinician orders the A1C test. The outcomes during the randomized por-
while unstable or highly intensively man- table in previous versions of the Stan- tion of the trials. Long-term follow-up
aged patients (e.g., pregnant type 1 dards of Medical Care in Diabetes describ- of the DCCT and UK Prospective Dia-
women) may be tested more frequently ing the correlation between A1C and betes Study (UKPDS) cohorts suggests
than every 3 months. The availability of mean glucose was derived from relatively that treatment to A1C targets below or
the A1C result at the time that the patient sparse data (one seven-point profile over around 7% in the years soon after the
is seen (point-of-care testing) has been re- 1 day per A1C reading) in the primarily diagnosis of diabetes is associated with
ported to result in increased intensifica- Caucasian type 1 participants in the Dia- long-term reduction in risk of macro-
tion of therapy and improvement in betes Control and Complications Trial vascular disease. Until more evidence
glycemic control (40,41). (DCCT) trial (43). Clinicians should note becomes available, the general goal of
The A1C test is subject to certain lim- that the numbers in the table are now dif- 7% appears reasonable for many
DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009 S19
8. Standards of Medical Care
adults for macrovascular risk reduc- rates of microvascular complications, with those previously in the standard arm
tion. (B) even with loss of glycemic separation be- (55).
● Subgroup analyses of clinical trials such tween the intensive and standard cohorts The UKPDS trial of type 2 diabetes
as the DCCT and UKPDS and the mi- after the end of the randomized con- observed a 16% reduction in cardiovascu-
crovascular evidence from the Action in trolled (51). lar complications (combined fatal or non-
Diabetes and Vascular Disease: Preterax In each of these large randomized fatal MI and sudden death) in the
and Diamicron MR Controlled Evalua- prospective clinical trials, treatment regi- intensive glycemic control arm, although
tion (ADVANCE) trial suggest a small mens that reduced average A1C to 7% this difference was not statistically signif-
but incremental benefit in microvascu- ( 1% above the upper limits of normal) icant (P 0.052), and there was no sug-
lar outcomes with A1C values closer to were associated with fewer long-term mi- gestion of benefit on other CVD outcomes
normal. Therefore, for selected individ- crovascular complications; however, in- such as stroke. In an epidemiologic anal-
ual patients, providers might reason- tensive control was found to increase the ysis of the study cohort, a continuous as-
ably suggest even lower A1C goals than risk of severe hypoglycemia, most notably sociation was observed, such that for
the general goal of 7%, if this can be in the DCCT, and led to weight gain every percentage point lower median on-
achieved without significant hypogly- (39,52). study A1C (e.g., 8 to 7%) there was a sta-
cemia or other adverse effects of treat- Epidemiological analyses of the tistically significant 18% reduction in
ment. Such patients might include DCCT and UKPDS (39,45) demonstrate a CVD events, again with no glycemic
those with short duration of diabetes, curvilinear relationship between A1C and threshold. A recent report of 10 years of
long life expectancy, and no significant microvascular complications. Such anal- follow-up of the UKPDS cohort describes,
CVD. (B) yses suggest that, on a population level, for the participants originally randomized
● Conversely, less stringent A1C goals to intensive glycemic control compared
the greatest number of complications will
than the general goal of 7% may be be averted by taking patients from very with those randomized to conventional
appropriate for patients with a history poor control to fair or good control. These glycemic control, long-term reductions in
of severe hypoglycemia, limited life ex- analyses also suggest that further lowering MI (15% with sulfonylurea or insulin as
pectancy, advanced microvascular or of A1C from 7 to 6% is associated with initial pharmacotherapy, 33% with met-
macrovascular complications, exten- further reduction in the risk of microvas- formin as initial pharmacotherapy, both
sive comorbid conditions, and those cular complications, albeit the absolute statistically significant) and in all-cause
with longstanding diabetes in whom mortality (13 and 27%, respectively, both
risk reductions become much smaller.
the general goal is difficult to attain de- statistically significant) (51).
Given the substantially increased risk of
spite DSME, appropriate glucose mon- Because of ongoing uncertainty re-
hypoglycemia (particularly in those with
itoring, and effective doses of multiple garding whether intensive glycemic con-
type 1 diabetes) and the relatively much
glucose-lowering agents including in- trol can reduce the increased risk of CVD
greater effort required to achieve near-
sulin. (C) events in people with type 2 diabetes, sev-
normoglycemia, the risks of lower targets
eral large long-term trials were launched
Glycemic control is fundamental to the may outweigh the potential benefits on in the past decade to compare the effects
management of diabetes. The DCCT, a microvascular complications on a popu- of intensive versus standard glycemic
prospective, randomized, controlled trial lation level. However, selected individual control on CVD outcomes in relatively
of intensive versus standard glycemic patients, especially those with little co- high-risk participants with established
control in patients with relatively recently morbidity and long life expectancy (who type 2 diabetes.
diagnosed type 1 diabetes, showed defin- may reap the benefits of further lowering The Action to Control Cardiovascular
itively that improved glycemic control is of glycemia below 7%) may, at patient Risk in Diabetes (ACCORD) study ran-
associated with significantly decreased and provider judgment, adopt glycemic domized 10,251 participants with either
rates of microvascular (retinopathy and targets as close to normal as possible as history of a CVD event (ages 40 –79 years)
nephropathy) as well as neuropathic long as significant hypoglycemia does not or significant CVD risk (ages 55–79) to a
complications (45). Follow-up of the become a barrier. strategy of intensive glycemic control (tar-
DCCT cohorts in the Epidemiology of Di- Whereas many epidemiologic studies get A1C 6.0%) or standard glycemic
abetes Interventions and Complications and meta-analyses (53,54) have clearly control (A1C target 7.0 –7.9%). Investiga-
(EDIC) study has shown persistence of shown a direct relationship between A1C tors used multiple glycemic medications
this effect in previously intensively and CVD, the potential of intensive glyce- in both arms. ACCORD participants were
treated subjects, even though their glyce- mic control to reduce CVD has been less on average 62 years old and had a mean
mic control has been equivalent to that of clearly defined. In the DCCT, there was a duration of diabetes of 10 years, with 35%
previous standard arm subjects during trend toward lower risk of CVD events already treated with insulin at baseline.
follow-up (46,47). with intensive control (risk reduction From a baseline median A1C of 8.1%, the
In type 2 diabetes, the Kumamoto 41%, 95% CI 10 – 68%), but the number intensive arm reached a median A1C of
study (48) and the UKPDS (49,50) dem- of events was small. However, 9-year 6.4% within 12 months of randomiza-
onstrated significant reductions in micro- post-DCCT follow-up of the cohort has tion, while the standard group reached a
vascular and neuropathic complications shown that participants previously ran- median A1C of 7.5%. Other risk factors
with intensive therapy. Similar to the domized to the intensive arm had a 42% were treated aggressively and equally in
DCCT-EDIC findings, long-term fol- reduction (P 0.02) in CVD outcomes both groups. The intensive glycemic con-
low-up of the UKPDS cohort has recently and a 57% reduction (P 0.02) in the trol group had more use of insulin in com-
demonstrated a “legacy effect” of early in- risk of nonfatal myocardial infarction bination with multiple oral agents,
tensive glycemic control on long-term (MI), stroke, or CVD death compared significantly more weight gain, and more
S20 DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009
9. Position Statement
episodes of severe hypoglycemia than the Intensive glycemic control significantly use of insulin, TZDs, other drugs, and
standard group. reduced the primary endpoint (HR 0.90 drug combinations; and greater weight
In February 2008, the glycemic con- [95% CI 0.82– 0.98]; P 0.01), although gain. Such factors may be associated sta-
trol study of ACCORD was halted on the this was due to a significant reduction in tistically with the higher mortality rate in
recommendation of the study’s data safety the microvascular outcome (0.86 [0.77– the intensive arm but may not be caus-
monitoring board due to the finding of an 0.97], P 0.01), primarily development ative. It is biologically plausible that se-
increased rate of mortality in the intensive of macroalbuminuria, with no significant vere hypoglycemia could increase the risk
arm compared with the standard arm reduction in the macrovascular outcome of cardiovascular death in participants
(1.41%/year vs. 1.14%/year; HR 1.22 (0.94 [0.84 –1.06]; P 0.32). There was with high underlying CVD risk. Other
[95% CI 1.01–1.46]), with a similar in- no difference in overall or cardiovascular plausible mechanisms for the increase in
crease in cardiovascular deaths. The pri- mortality between the intensive and the mortality in ACCORD include weight
mary outcome of ACCORD (MI, stroke, standard glycemic control arms (57). gain, unmeasured drug effects or interac-
or cardiovascular death) was lower in the The VADT randomized 1,791 partic- tions, or the overall “intensity” of the AC-
intensive glycemic control group due to a ipants with type 2 diabetes uncontrolled CORD intervention (use of multiple oral
reduction in nonfatal MI, although this on insulin or maximal dose oral agents glucose-lowering drugs along with multi-
finding was not statistically significant (median entry A1C 9.4%) to a strategy of ple doses of insulin, frequent therapy ad-
when the study was terminated (HR 0.90 intensive glycemic control (goal A1C justments to push A1C and self-
[95% CI 0.78 –1.04]; P 0.16) (56). 6.0%) or standard glycemic control, monitored blood glucose to very low
Exploratory analyses of the mortality with a planned A1C separation of at least targets, and an intense effort to aggres-
findings of ACCORD (evaluating vari- 1.5%. Medication treatment algorithms sively reduce A1C by 2% in participants
ables including weight gain, use of any were used to achieve the specified glyce- entering the trial with advanced diabetes
specific drug or drug combination, and mic goals, with a goal of using similar and multiple comorbidities).
hypoglycemia) were reportedly unable to medications in both groups. Median A1C Since the ADVANCE trial did not
identify an explanation for the excess levels of 6.9 and 8.4% were achieved in show any increase in mortality in the in-
mortality in the intensive arm. Prespeci- the intensive and standard arms, respec- tensive glycemic control arm, examining
fied subset analyses showed that partici- tively, within the first year of the study. the differences between ADVANCE and
pants with no previous CVD event and Other CVD risk factors were treated ag- ACCORD supports additional hypothe-
those who had a baseline A1C 8% had a gressively and equally in both groups. ses. ADVANCE participants on average
statistically significant reduction in the The primary outcome of the VADT appeared to have earlier or less advanced
primary CVD outcome. was a composite of CVD events (MI, diabetes, with shorter duration by 2–3
The ADVANCE study randomized stroke, cardiovascular death, revascular- years and lower A1C at entry despite very
11,140 participants to a strategy of inten- ization, hospitalization for heart failure, little use of insulin at baseline. A1C was
sive glycemic control (with primary ther- and amputation for ischemia). During a also lowered less and more gradually in
apy being the sulfonylurea gliclizide and mean 6-year follow-up period, the cumu- the ADVANCE trial, and there was no sig-
additional medications as needed to lative primary outcome was nonsignifi- nificant weight gain with intensive
achieve a target A1C of 6.5%) or to stan- cantly lower in the intensive arm (HR glycemic therapy. Although severe hypo-
dard therapy (in which any medication 0.87 [95% CI 0.73–1.04]; P 0.12). glycemia was defined somewhat differ-
but gliclizide could be used and the gly- There were more CVD deaths in the in- ently in the three trials, it appears that this
cemic target was according to “local tensive arm than in the standard arm (40 occurred in fewer than 3% of intensively
guidelines”). ADVANCE participants vs. 33; sudden deaths 11 vs. 4), but the treated ADVANCE participants for the
(who had to be at least 55 years of age difference was not statistically significant. entire study duration (median 5 years)
with either known vascular disease or at Post hoc subgroup analyses suggested compared with 16% of intensively
least one other vascular risk factor) were that duration of diabetes interacted with treated subjects in ACCORD and 21% in
slightly older and of similar high CVD risk randomization such that participants VADT.
as those in ACCORD. However, they had with duration of diabetes less than about It is likely that the increase in mortal-
an average duration of diabetes 2 years 12 years appeared to have a CVD benefit ity in ACCORD was related to the overall
shorter, lower baseline A1C (median of intensive glycemic control while those treatment strategies for intensifying glyce-
7.2%), and almost no use of insulin at with longer duration of disease before mic control in the study population, not
enrollment. The median A1C levels study entry had a neutral or even adverse the achieved A1C per se. The ADVANCE
achieved in the intensive and standard effect of intensive glycemic control. Other study achieved a median A1C in its inten-
arms were 6.3 and 7.0%, respectively, exploratory analyses suggested that se- sive arm similar to that in the ACCORD
and maximal separation between the vere hypoglycemia within the past 90 study, with no increased mortality haz-
arms took several years to achieve. Use of days was a strong predictor of the primary ard. Thus, the ACCORD mortality find-
other drugs that favorably impact CVD outcome and of CVD mortality (58). ings do not imply that patients with type 2
risk (aspirin, statins, ACE inhibitors) was The cause of the excess deaths in the diabetes who can easily achieve or main-
lower in ADVANCE than in the ACCORD intensive glycemic control arm of AC- tain low A1C levels with lifestyle modifi-
or Veterans Affairs Diabetes Trial (VADT). CORD compared with the standard arm cations with or without pharmacotherapy
The primary outcome of ADVANCE has been difficult to pinpoint. By design of are at risk and need to “raise” their A1C.
was a combination of microvascular the trial, randomization to the intensive The three trials compared treatments
events (nephropathy and retinopathy) arm was associated with or led to many to A1C levels in the “flatter” part of the
and major adverse cardiovascular events downstream effects, such as higher rates observational glycemia-CVD risk curves
(MI, stroke, and cardiovascular death). of severe hypoglycemia; more frequent (median A1C of 6.4 – 6.9% in the inten-
DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009 S21
10. Standards of Medical Care
Table 9—Summary of glycemic recommendations for non-pregnant adults with diabetes gate measures of vascular pathology, such
A1C 7.0%* as endothelial dysfunction, are negatively
Preprandial capillary plasma glucose 70–130 mg/dl (3.9–7.2 mmol/l) affected by postprandial hyperglycemia
Peak postprandial capillary plasma glucose 180 mg/dl ( 10.0 mmol/l) (60). It is clear that postprandial hyper-
Key concepts in setting glycemic goals: glycemia, like preprandial hyperglyce-
● A1C is the primary target for glycemic control.
mia, contributes to elevated A1C levels,
● Goals should be individualized based on:
with its relative contribution being higher
● duration of diabetes
at A1C levels that are closer to 7%. How-
● age/life expectancy
ever, outcome studies have clearly shown
● comorbid conditions
A1C to be the primary predictor of com-
● known CVD or advanced microvascular plications, and landmark glycemic con-
complications trol trials such as the DCCT and UKPDS
● hypoglycemia unawareness relied overwhelmingly on preprandial
● individual patient considerations SMBG. Additionally, a randomized con-
● More or less stringent glycemic goals may be trolled trial presented at the 68th Scien-
appropriate for individual patients. tific Sessions of the American Diabetes
● Postprandial glucose may be targeted if A1C goals are Association in June 2008 found no CVD
not met despite reaching preprandial glucose goals. benefit of insulin regimens targeting post-
*Referenced to a nondiabetic range of 4.0 – 6.0% using a DCCT-based assay. Postprandial glucose measure- prandial glucose compared with those
ments should be made 1–2 h after the beginning of the meal, generally peak levels in patients with diabetes. targeting preprandial glucose. A reason-
able recommendation for postprandial
testing and targets is that for individuals
sive arms compared with 7.0 – 8.4% in events with intensive glycemic control in who have premeal glucose values within
the standard arms). Importantly, their re- ACCORD, ADVANCE, and VADT should target but have A1C values above target,
sults should not be extrapolated to imply not lead clinicians to abandon the general monitoring postprandial plasma glucose
that there would be no cardiovascular target of an A1C 7.0% and thereby dis- (PPG) 1–2 h after the start of the meal and
benefit of glucose lowering from very count the benefit of good control on what treatment aimed at reducing PPG values
poor control (e.g., A1C 9%) to good are serious and debilitating microvascular to 180 mg/dl may help lower A1C.
control (e.g., A1C 7%). complications. As noted above, less stringent treat-
All three trials were carried out in par- The evidence for a cardiovascular ment goals may be appropriate for adults
ticipants with established diabetes (mean benefit of intensive glycemic control pri- with limited life expectancies or advanced
duration 8 –11 years) and either known marily rests on long-term follow-up of vascular disease. Glycemic goals for chil-
CVD or multiple risk factors suggesting study cohorts treated early in the course dren are provided in Section VII.A.1.a.
the presence of established atherosclero- of type 1 and type 2 diabetes and subset Severe or frequent hypoglycemia is an ab-
sis. Subset analyses of the three trials sug- analyses of ACCORD, ADVANCE, and solute indication for the modification of
gested a significant benefit of intensive VADT. Conversely, the mortality findings treatment regimens, including setting
glycemic control on CVD in participants in ACCORD suggest that the potential higher glycemic goals.
with shorter duration of diabetes, lower risks of very intensive glycemic control Regarding goals for glycemic control
A1C at entry, and/or or absence of known may outweigh its benefits in some pa- for women with GDM, recommendations
CVD. The DCCT-EDIC study and the tients, such as those with very long dura- from the Fifth International Workshop-
long-term follow-up of the UKPDS cohort tion of diabetes, known history of severe Conference on Gestational Diabetes Mel-
both suggest that intensive glycemic con- hypoglycemia, advanced atherosclerosis, litus (61) were to target the following
trol initiated soon after diagnosis of dia- and advanced age/frailty. Certainly, pro- maternal capillary glucose concentra-
betes in patients with a lower level of CVD viders should be vigilant in preventing se- tions:
risk may impart long-term protection vere hypoglycemia in patients with
from CVD events. As is the case with mi- advanced disease and should not aggres-
crovascular complications, it may be that sively attempt to achieve near-normal ● preprandial: 95 mg/dl (5.3 mmol/l)
glycemic control plays a greater role be- A1C levels in patients in whom such a and either
fore macrovascular disease is well devel- target cannot be reasonably easily and ● 1-h postmeal: 140 mg/dl (7.8 mmol/l)
oped and minimal or no role when it is safely achieved. or
advanced. Recommended glycemic goals for ● 2-h postmeal: 120 mg/dl (6.7 mmol/l)
The benefits of intensive glycemic nonpregnant adults are shown in Table 9.
control on microvascular and neuro- The recommendations are based on those For women with preexisting type 1 or
pathic complications are well established for A1C, with listed blood glucose levels type 2 diabetes who become pregnant, a
for both type 1 and type 2 diabetes. The that appear to correlate with achievement
recent consensus statement (62) recom-
ADVANCE trial has added to that evi- of an A1C of 7%. The issue of pre- ver-
mended the following as optimal glyce-
dence base by demonstrating a significant sus postprandial SMBG targets is complex
mic goals, if they can be achieved without
reduction in the risk of new or worsening (59). Elevated postchallenge (2-h OGTT)
excessive hypoglycemia:
albuminuria when A1C was lowered to glucose values have been associated with
6.3% compared with standard glycemic increased cardiovascular risk indepen-
control achieving an A1C of 7.0%. The dent of FPG in some epidemiological ● premeal, bedtime, and overnight glu-
lack of significant reduction in CVD studies. In diabetic subjects, some surro- cose 60 –99 mg/dl
S22 DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009
11. Position Statement
● peak postprandial glucose 100 –129 The algorithm took into account the dietary strategies including reduced
mg/dl evidence for A1C-lowering of the individ- calories and reduced intake of dietary
● A1C 6.0% ual interventions, their additive effects, fat, can reduce the risk for developing
and their expense. The precise drugs used diabetes and are therefore recom-
3. Approach to treatment and their exact sequence may not be as mended. (A)
important as achieving and maintaining ● Individuals at high risk for type 2 diabetes
a. Therapy for type 1 diabetes. The glycemic targets safely. Medications not should be encouraged to achieve the U.S.
DCCT clearly showed that intensive insu- included in the consensus algorithm, ow- Department of Agriculture recommenda-
lin therapy (three or more injections per ing to less glucose-lowering effectiveness, tion for dietary fiber (14 g fiber/1,000
day of insulin or continuous subcutane- limited clinical data, and/or relative ex- kcal) and foods containing whole grains
ous insulin infusion (CSII, or insulin pense, still may be appropriate choices in (one-half of grain intake). (B)
pump therapy) was a key part of im- individual patients to achieve glycemic
proved glycemia and better outcomes goals. Initiation of insulin at time of diagno- Dietary fat intake in diabetes
(45). At the time of the study, therapy was sis is recommended for individuals present- management
carried out with short- and intermediate- ing with weight loss or other severe ● Saturated fat intake should be 7% of
acting human insulins. Despite better mi- hyperglycemic symptoms or signs. For a list total calories. (A)
crovascular outcomes, intensive insulin of currently approved diabetes medica- ● Intake of trans fat should be minimized. (B)
therapy was associated with a marked in- tions, see http://ndep.nih.gov/diabetes/
crease in severe hypoglycemia (62 epi- pubs/Drug_tables_supplement.pdf. Carbohydrate intake in diabetes
sodes per 100 patient-years of therapy). management
Since the time of the DCCT, a number of D. MNT ● Monitoring carbohydrate, whether by
rapid-acting and long-acting insulin ana- carbohydrate counting, exchanges, or
logs have been developed. These analogs General recommendations experience-based estimation, remains a
were designed to be more “physiological” ● Individuals who have pre-diabetes or key strategy in achieving glycemic con-
in their pharmacokinetics and pharmaco- diabetes should receive individualized trol. (A)
dynamics and are associated with less hy- MNT as needed to achieve treatment ● For individuals with diabetes, the use of
poglycemia with equal A1C lowering in goals, preferably provided by a regis- the glycemic index and glycemic load
type 1 diabetes (63,64). tered dietitian familiar with the compo- may provide a modest additional bene-
Therefore, recommended therapy for nents of diabetes MNT. (B) fit for glycemic control over that ob-
type 1 diabetes consists of the following ● MNT should be covered by insurance served when total carbohydrate is
components: 1) use of multiple dose in- and other payors. (E) considered alone. (B)
sulin injections (3– 4 injections per day of
basal and prandial insulin) or CSII ther- Energy balance, overweight, and Other nutrition recommendations
apy; 2) matching of prandial insulin to obesity ● Sugar alcohols and nonnutritive sweet-
carbohydrate intake, premeal blood glu- ● In overweight and obese insulin- eners are safe when consumed within
cose, and anticipated activity; and 3) for resistant individuals, modest weight the acceptable daily intake levels estab-
many patients (especially if hypoglycemia loss has been shown to reduce insulin lished by the Food and Drug Adminis-
is a problem), use of insulin analogs. resistance. Thus, weight loss is recom- tration (FDA). (A)
There are excellent reviews available that mended for all overweight or obese in- ● If adults with diabetes choose to use alco-
guide the initiation and management of dividuals who have or are at risk for hol, daily intake should be limited to a
insulin therapy to achieve desired glyce- diabetes. (A) moderate amount (one drink per day or
mic goals (3,63,65). ● For weight loss, either low-carbohy- less for adult women and two drinks per
b. Therapy for type 2 diabetes. The ADA drate or low-fat calorie restricted diets day or less for adult men). (E)
and the European Association for the may be effective in the short-term (up ● Routine supplementation with antioxi-
Study of Diabetes published a consensus to 1 year). (A) dants, such as vitamins E and C and
statement on the approach to manage- ● For patients on low-carbohydrate diets, carotene, is not advised because of lack
ment of hyperglycemia in individuals monitor lipid profiles, renal function, of evidence of efficacy and concern re-
with type 2 diabetes (66) and recently and protein intake (in those with ne- lated to long-term safety. (A)
published an update (67). Highlights of phropathy) and adjust hypoglycemic ● Benefit from chromium supplementa-
this approach are: intervention at the time therapy as needed. (E) tion in people with diabetes or obesity
of diagnosis with metformin in combina- ● Physical activity and behavior modifica- has not been conclusively demon-
tion with lifestyle changes (MNT and tion are important components of weight strated and, therefore, cannot be rec-
exercise) and continuing timely augmen- loss programs and are most helpful in ommended. (E)
tation of therapy with additional agents maintenance of weight loss. (B)
(including early initiation of insulin ther- MNT is an integral component of diabetes
apy) as a means of achieving and main- Primary prevention of diabetes prevention, management, and self-
taining recommended levels of glycemic ● Among individuals at high risk for de- management education. In addition to its
control (i.e., A1C 7% for most patients). veloping type 2 diabetes, structured role in preventing and controlling diabe-
The overall objective is to achieve and programs that emphasize lifestyle tes, ADA recognizes the importance of
maintain glycemic control and to change changes that include moderate weight nutrition as an essential component of an
interventions when therapeutic goals are loss (7% body weight) and regular overall healthy lifestyle. A full review of
not being met. physical activity (150 min/week), with the evidence regarding nutrition in pre-
DIABETES CARE, VOLUME 32, SUPPLEMENT 1, JANUARY 2009 S23