Ada 2010 full


Published on

Published in: Education, Health & Medicine
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Ada 2010 full

  1. 1. P O S I T I O N S T A T E M E N TStandards of Medical Care in Diabetes—2010AMERICAN DIABETES ASSOCIATIOND iabetes is a chronic illness that re- more detailed information about manage- by the Executive Committee of ADA’s quires continuing medical care and ment of diabetes, refer to references 1–3. Board of Directors. ongoing patient self-management The recommendations included areeducation and support to prevent acute screening, diagnostic, and therapeutic ac- I. CLASSIFICATION ANDcomplications and to reduce the risk of tions that are known or believed to favor- DIAGNOSISlong-term complications. Diabetes care is ably affect health outcomes of patients A. Classificationcomplex and requires that many issues, with diabetes. A grading system (Table 1), The classification of diabetes includesbeyond glycemic control, be addressed. A developed by the American Diabetes As- four clinical classes:large body of evidence exists that sup- sociation (ADA) and modeled after exist-ports a range of interventions to improve ing methods, was used to clarify and ● type 1 diabetes (results from -cell de-diabetes outcomes. codify the evidence that forms the basis struction, usually leading to absolute These standards of care are intended for the recommendations. The level of ev- insulin deficiency)to provide clinicians, patients, research- idence that supports each recommenda- ● type 2 diabetes (results from a progres-ers, payors, and other interested individ- sive insulin secretory defect on the tion is listed after each recommendationuals with the components of diabetes background of insulin resistance) using the letters A, B, C, or, general treatment goals, and tools to ● other specific types of diabetes due toevaluate the quality of care. While indi- These standards of care are revised annually by the ADA multidisciplinary other causes, e.g., genetic defects invidual preferences, comorbidities, and -cell function, genetic defects in insu-other patient factors may require modifi- Professional Practice Committee, and new evidence is incorporated. Members lin action, diseases of the exocrine pan-cation of goals, targets that are desirable creas (such as cystic fibrosis), and drug-for most patients with diabetes are pro- of the Professional Practice Committee or chemical-induced diabetes (such asvided. These standards are not intended and their disclosed conflicts of interest are in the treatment of AIDS or after organto preclude clinical judgment or more ex- listed in the Introduction. Subsequently, transplantation)tensive evaluation and management of the as with all position statements, the stan- ● gestational diabetes mellitus (GDM)patient by other specialists as needed. For dards of care are reviewed and approved (diabetes diagnosed during pregnancy)● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●Originally approved 1988. Most recent review/revision October 2009. Some patients cannot be clearly classifiedDOI: 10.2337/dc10-S011 as having type 1 or type 2 diabetes. Clin-Abbreviations: ABI, ankle-brachial index; ACCORD, Action to Control Cardiovascular Risk in Diabetes; ADAG, A1C-Derived Average Glucose Trial; ADVANCE, Action in Diabetes and Vascular Disease: Pre- ical presentation and disease progression terax and Diamicron Modified Release Controlled Evaluation; ACE, angiotensin converting enzyme; ARB, vary considerably in both types of diabe- angiotensin receptor blocker; ACT-NOW, ACTos Now Study for the Prevention of Diabetes; BMI, body tes. Occasionally, patients who otherwise mass index; CBG, capillary blood glucose; CFRD, cystic fibrosis–related diabetes; CGM, continuous have type 2 diabetes may present with ke- glucose monitoring; CHD, coronary heart disease; CHF, congestive heart failure; CCM, chronic care toacidosis. Similarly, patients with type 1 model; CKD, chronic kidney disease; CMS, Centers for Medicare and Medicaid Services; CSII, continuous subcutaneous insulin infusion; CVD, cardiovascular disease; DASH, Dietary Approaches to Stop Hyper- diabetes may have a late onset and slow tension; DCCT, Diabetes Control and Complications Trial; DKA, diabetic ketoacidosis; DMMP, diabetes (but relentless) progression despite hav- medical management plan; DPN, distal symmetric polyneuropathy; DPP, Diabetes Prevention Program; ing features of autoimmune disease. Such DPS, Diabetes Prevention Study; DREAM, Diabetes Reduction Assessment with Ramipril and Rosiglita- difficulties in diagnosis may occur in chil- zone Medication; DRS, Diabetic Retinopathy Study; DSME, diabetes self-management education; DSMT, diabetes self-management training; eAG, estimated average glucose; eGFR, estimated glomerular filtration dren, adolescents, and adults. The true rate; ECG, electrocardiogram; EDIC, Epidemiology of Diabetes Interventions and Complications; ERP, diagnosis may become more obvious over education recognition program; ESRD, end-stage renal disease; ETDRS, Early Treatment Diabetic Reti- time. nopathy Study; FDA, Food and Drug Administration; FPG, fasting plasma glucose; GDM, gestational diabetes mellitus; GFR, glomerular filtration rate; HAPO, Hyperglycemia and Adverse Pregnancy Out- B. Diagnosis of diabetes comes; ICU, intensive care unit; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; Look AHEAD, Action for Health in Diabetes; MDRD, Modification of Diet in Renal Disease; MNT, medical Recommendations nutrition therapy; NDEP, National Diabetes Education Program; NGSP, National Glycohemoglobin Stan- For decades, the diagnosis of diabetes has dardization Program; NPDR, nonproliferative diabetic retinopathy; OGTT, oral glucose tolerance test; PAD, peripheral arterial disease; PCOS, polycystic ovarian syndrome; PDR, proliferative diabetic retinop- been based on plasma glucose (PG) crite- athy; PPG, postprandial plasma glucose; RAS, renin-angiotensin system; SMBG, self-monitoring of blood ria, either fasting PG (FPG) or 2-h 75-g glucose; STOP-NIDDM, Study to Prevent Non-Insulin Dependent Diabetes; SSI, sliding scale insulin; oral glucose tolerance test (OGTT) values. TZD, thiazolidinedione; UKPDS, U.K. Prospective Diabetes Study; VADT, Veterans Affairs Diabetes Trial; In 1997, the first Expert Committee on XENDOS, XENical in the prevention of Diabetes in Obese Subjects.© 2010 by the American Diabetes Association. Readers may use this article as long as the work is properly the Diagnosis and Classification of Diabe- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons. tes Mellitus revised the diagnostic criteria org/licenses/by-nc-nd/3.0/ for details. using the observed association DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S11
  2. 2. Standards of Medical CareTable 1 —ADA evidence grading system for clinical practice recommendations be used (an updated list of A1C assays and whether abnormal hemoglobins impactLevel of them is available at Description index3.html). For conditions with abnor- mal red cell turnover, such as pregnancy orA Clear evidence from well-conducted, generalizable, randomized controlled trials that anemias from hemolysis and iron defi- are adequately powered, including: ciency, the diagnosis of diabetes must use ● Evidence from a well-conducted multicenter trial glucose criteria exclusively. ● Evidence from a meta-analysis that incorporated quality ratings in the analysis The established glucose criteria for Compelling nonexperimental evidence, i.e., all or none rule developed by Center the diagnosis of diabetes (FPG and 2-h for Evidence Based Medicine at Oxford PG) remain valid. Patients with severe hy- Supportive evidence from well-conducted randomized controlled trials that are perglycemia such as those who present adequately powered, including: with severe classic hyperglycemic symp- ● Evidence from a well-conducted trial at one or more institutions toms or hyperglycemic crisis can continue ● Evidence from a meta-analysis that incorporated quality ratings in the analysis to be diagnosed when a random (or ca-B Supportive evidence from well-conducted cohort studies: sual) PG of 200 mg/dl (11.1 mmol/l) is ● Evidence from a well-conducted prospective cohort study or registry found. It is likely that in such cases the ● Evidence from a well-conducted meta-analysis of cohort studies health care professional would also con- Supportive evidence from a well-conducted case-control study duct an A1C test as part of the initial as-C Supportive evidence from poorly controlled or uncontrolled studies sessment of the severity of the diabetes ● Evidence from randomized clinical trials with one or more major or three or and that it would be above the diagnostic more minor methodological flaws that could invalidate the results cut point. However, in rapidly evolving ● Evidence from observational studies with high potential for bias (such as case diabetes such as the development of type series with comparison to historical controls) 1 in some children, the A1C may not be ● Evidence from case series or case reports significantly elevated despite frank Conflicting evidence with the weight of evidence supporting the recommendation diabetes.E Expert consensus or clinical experience Just as there is 100% concordance between the FPG and 2-h PG tests, there is not perfect concordance between A1Cglucose levels and presence of retinopa- the A1C test to diagnose diabetes with a and either glucose-based test. Analyses ofthy as the key factor with which to iden- threshold of 6.5%, and ADA affirms this National Health and Nutrition Examina-tify threshold FPG and 2-h PG levels. The decision (6). The diagnostic test should tion Survey (NHANES) data indicate that,committee examined data from three be performed using a method certified by assuming universal screening of the undi-cross-sectional epidemiologic studies that the National Glycohemoglobin Standard- agnosed, the A1C cut point of 6.5%assessed retinopathy with fundus photog- ization Program (NGSP) and standard- identifies one-third fewer cases of undiag-raphy or direct ophthalmoscopy and ized or traceable to the Diabetes Control nosed diabetes than a fasting glucose cutmeasured glycemia as FPG, 2-h PG, and and Complications Trial (DCCT) refer- point of 126 mg/dl (7.0 mmol/l) (E.HbA1c (A1C). The studies demonstrated ence assay. Point-of-care A1C assays are Gregg, personal communication). How-glycemic levels below which there was lit- not sufficiently accurate at this time to use ever, in practice, a large portion of thetle prevalent retinopathy and above for diagnostic purposes. diabetic population remains unaware ofwhich the prevalence of retinopathy in- Epidemiologic datasets show a rela- their condition. Thus, the lower sensitiv-creased in an apparently linear fashion. tionship between A1C and the risk of ret- ity of A1C at the designated cut point mayThe deciles of FPG, 2-h PG, and A1C at inopathy similar to that which has been well be offset by the test’s greater practi-which retinopathy began to increase were shown for corresponding FPG and 2-h PG cality, and wider application of a morethe same for each measure within each thresholds. The A1C has several advan- convenient test (A1C) may actually in-population. The analyses helped to in- tages to the FPG, including greater conve- crease the number of diagnoses made.form a then-new diagnostic cut point of nience, since fasting is not required; As with most diagnostic tests, a test 126 mg/dl (7.0 mmol/l) for FPG and evidence to suggest greater preanalytical result diagnostic of diabetes should be re-confirmed the long-standing diagnostic stability; and less day-to-day perturba- peated to rule out laboratory error, unless2-h PG value of 200 mg/dl (11.1 tions during periods of stress and illness. the diagnosis is clear on clinical grounds,mmol/l) (4). These advantages must be balanced by such as a patient with classic symptoms of ADA has not previously recom- greater cost, limited availability of A1C hyperglycemia or hyperglycemic crisis. Itmended the use of A1C for diagnosing testing in certain regions of the develop- is preferable that the same test be repeateddiabetes, in part due to lack of standard- ing world, and incomplete correlation be- for confirmation, since there will be aization of the assay. However, A1C assays tween A1C and average glucose in certain greater likelihood of concurrence in thisare now highly standardized, and their re- individuals. In addition, the A1C can be case. For example, if the A1C is 7.0% andsults can be uniformly applied both tem- misleading in patients with certain forms a repeat result is 6.8%, the diagnosis ofporally and across populations. In a of anemia and hemoglobinopathies. For diabetes is confirmed. However, there arerecent report (5), after an extensive review patients with a hemoglobinopathy but scenarios in which results of two differentof both established and emerging epide- normal red cell turnover, such as sickle tests (e.g., FPG and A1C) are available formiological evidence, an international ex- cell trait, an A1C assay without interfer- the same patient. In this situation, if thepert committee recommended the use of ence from abnormal hemoglobins should two different tests are both above the di-S12 DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010
  3. 3. Position StatementTable 2—Criteria for the diagnosis of diabetes costs of false positives (falsely identifying1. A1C 6.5%. The test should be performed in a laboratory using a method and then spending intervention resources that is NGSP certified and standardized to the DCCT assay.* on those who were not going to develop OR diabetes anyway).2. FPG 126 mg/dl (7.0 mmol/l). Fasting is defined as no caloric intake for at Linear regression analyses of nation- least 8 h.* ally representative U.S. data (NHANES OR 2005–2006) indicate that among the3. Two-hour plasma glucose 200 mg/dl (11.1 mmol/l) during an OGTT. nondiabetic adult population, an FPG of The test should be performed as described by the World Health 110 mg/dl corresponds to an A1C of Organization, using a glucose load containing the equivalent of 75 g 5.6%, while an FPG of 100 mg/dl corre- anhydrous glucose dissolved in water.* sponds to an A1C of 5.4%. Receiver op- OR erating curve analyses of these data4. In a patient with classic symptoms of hyperglycemia or hyperglycemic indicate that an A1C value of 5.7%, com- crisis, a random plasma glucose 200 mg/dl (11.1 mmol/l). pared with other cut points, has the best combination of sensitivity (39%) and*In the absence of unequivocal hyperglycemia, criteria 1–3 should be confirmed by repeat testing. specificity (91%) to identify cases of IFG (FPG 100 mg/dl [5.6 mmol/l]) (R.T.agnostic threshold, the diagnosis of dia- or impaired glucose tolerance (IGT) (2-h Ackerman, Personal Communication).betes is confirmed. OGTT values of 140 mg/dl [7.8 mmol/l] Other analyses suggest that an A1C of On the other hand, if two different to 199 mg/dl [11.0 mmol/l]). 5.7% is associated with diabetes risk sim-tests are available in an individual and the Individuals with IFG and/or IGT have ilar to that of the high-risk participants inresults are discordant, the test whose re- been referred to as having pre-diabetes, the Diabetes Prevention Program (DPP)sult is above the diagnostic cut point indicating the relatively high risk for the (R.T. Ackerman, personal communica-should be repeated, and the diagnosis is future development of diabetes. IFG and tion). Hence, it is reasonable to considermade on the basis of the confirmed test. IGT should not be viewed as clinical en- an A1C range of 5.7– 6.4% as identifyingThat is, if a patient meets the diabetes cri- tities in their own right but rather risk individuals with high risk for future dia-terion of the A1C (two results 6.5%) but factors for diabetes as well as cardiovas- betes and to whom the term pre-diabetesnot the FPG ( 126 mg/dl or 7.0 mmol/l), cular disease (CVD). IFG and IGT are may be applied (6).or vice versa, that person should be con- associated with obesity (especially As is the case for individuals found tosidered to have diabetes. Admittedly, in abdominal or visceral obesity), dyslipide- have IFG and IGT, individuals with anmost circumstance the “nondiabetic” test mia with high triglycerides and/or low A1C of 5.7– 6.4% should be informed ofis likely to be in a range very close to the HDL cholesterol, and hypertension. their increased risk for diabetes as wellthreshold that defines diabetes. Structured lifestyle intervention, aimed at as CVD and counseled about effective Since there is preanalytic and analytic increasing physical activity and produc- strategies to lower their risks (see IV. PRE-variability of all the tests, it is also possible ing 5–10% loss of body weight, and cer- VENTION/DELAY OF TYPE 2 DIABETES).that when a test whose result was above tain pharmacological agents have been As with glucose measurements, the contin-the diagnostic threshold is repeated, the demonstrated to prevent or delay the de- uum of risk is curvilinear, so that as A1Csecond value will be below the diagnostic velopment of diabetes in people with IGT rises, the risk of diabetes rises dispropor-cut point. This is least likely for A1C, (see Table 7). It should be noted that the tionately. Accordingly, interventionssomewhat more likely for FPG, and most 2003 ADA Expert Committee report re- should be most intensive and follow-uplikely for the 2-h PG. Barring a laboratory duced the lower FPG cut point to define should be particularly vigilant for thoseerror, such patients are likely to have test IFG from 110 mg/dl (6.1 mmol/l) to 100 with an A1C 6.0%, who should be con-results near the margins of the threshold mg/dl (5.6 mmol/l), in part to make the sidered to be at very high risk. However,for a diagnosis. The healthcare profes- prevalence of IFG more similar to that of just as an individual with a fasting glucose ofsional might opt to follow the patient IGT. However, the World Health Organi- 98 mg/dl (5.4 mmol/l) may not be at negli-closely and repeat the testing in 3– 6 zation (WHO) and many other diabetes gible risk for diabetes, individuals with anmonths. organizations did not adopt this change. A1C 5.7% may still be at risk, depending The current diagnostic criteria for di- As the A1C becomes increasingly on the level of A1C and presence of otherabetes are summarized in Table 2. used to diagnose diabetes in individuals risk factors, such as obesity and family with risk factors, it will also identify those history.C. Categories of increased risk for at high risk for developing diabetes in thediabetes future. As was the case with the glucose Table 3—Categories of increased risk forIn 1997 and 2003, The Expert Committee measures, defining a lower limit of an in- diabetes*on the Diagnosis and Classification of Di- termediate category of A1C is somewhat FPG 100–125 mg/dl (5.6–6.9 mmol/l)abetes Mellitus (4,7) recognized an inter- arbitrary, since risk of diabetes with any IFGmediate group of individuals whose measure or surrogate of glycemia is a con- 2-h PG on the 75-g OGTT 140–199 mg/dlglucose levels, although not meeting cri- tinuum extending well into the normal (7.8–11.0 mmol/l) IGTteria for diabetes, are nevertheless too ranges. To maximize equity and efficiency A1C 5.7–6.4%high to be considered normal. This group of preventive interventions, such an A1C *For all three tests, risk is continuous, extendingwas defined as having impaired fasting cut point, should balance the costs of false below the lower limit of the range and becomingglucose (IFG) (FPG levels of 100 mg/dl negatives (failing to identify those who are disproportionately greater at higher ends of the[5.6 mmol/l] to 125 mg/dl [6.9 mmol/l]) going to develop diabetes) against the DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S13
  4. 4. Standards of Medical CareTable 4—Criteria for testing for diabetes in asymptomatic adult individuals Recommendations for testing for dia-1. 2 Testing should be considered in all adults who are overweight (BMI 25 kg/m *) and betes in asymptomatic undiagnosed have additional risk factors: adults are listed in Table 4. Testing should ● physical inactivity be considered in adults of any age with ● first-degree relative with diabetes BMI 25 kg/m2 and one or more risk fac- ● members of a high-risk ethnic population (e.g., African American, Latino, Native tors for diabetes. Because age is a major American, Asian American, Pacific Islander) risk factor for diabetes, testing of those ● women who delivered a baby weighing 9 lb or were diagnosed with GDM without other risk factors should begin no ● hypertension ( 140/90 mmHg or on therapy for hypertension) later than at age 45 years. ● HDL cholesterol level 35 mg/dl (0.90 mmol/l) and/or a triglyceride level 250 Either A1C, FPG, or 2-h OGTT is ap- mg/dl (2.82 mmol/l) propriate for testing. The 2-h OGTT identi- ● women with polycystic ovary syndrome fies people with either IFG or IGT and thus ● A1C 5.7%, IGT, or IFG on previous testing more people at increased risk for the devel- ● other clinical conditions associated with insulin resistance (e.g., severe obesity, opment of diabetes and CVD. It should be acanthosis nigricans) noted that the two tests do not necessarily ● history of CVD detect the same individuals (10). The effi-2. In the absence of the above criteria, testing diabetes should begin at age 45 years cacy of interventions for primary preven-3. If results are normal, testing should be repeated at least at 3-year intervals, with tion of type 2 diabetes (11–17) has consideration of more frequent testing depending on initial results and risk primarily been demonstrated among indi- status. viduals with IGT, but not for individuals with IFG (who do not also have IGT) or*At-risk BMI may be lower in some ethnic groups. those with specific A1C levels. The appropriate interval between Table 3 summarizes the categories of who the provider tests because of high tests is not known (18). The rationale forincreased risk for diabetes. suspicion of diabetes, to the symptomatic the 3-year interval is that false negatives patient. The discussion herein is primar- will be repeated before substantial timeII. TESTING FOR DIABETES ily framed as testing for diabetes in indi- elapses, and there is little likelihood thatIN ASYMPTOMATIC viduals without symptoms. Testing for an individual will develop significantPATIENTS diabetes will also detect individuals at in- complications of diabetes within 3 years creased future risk for diabetes, herein re- of a negative test result.Recommendations ferred to as pre-diabetic. Because of the need for follow-up and● Testing to detect type 2 diabetes and discussion of abnormal results, testing assess risk for future diabetes in asymp- A. Testing for type 2 diabetes and should be carried out within the health tomatic people should be considered in risk of future diabetes in adults care setting. Community screening out- adults of any age who are overweight or Type 2 diabetes is frequently not diag- side a health care setting is not recom- obese (BMI 25 kg/m2) and who have nosed until complications appear, and mended because people with positive one or more additional risk factors for approximately one-fourth of all people tests may not seek, or have access to, ap- diabetes (Table 4). In those without with diabetes in the U.S. may be undiag- propriate follow-up testing and care. these risk factors, testing should begin nosed. Although the effectiveness of early Conversely, there may be failure to ensure at age 45 years. (B) identification of pre-diabetes and diabetes appropriate repeat testing for individuals● If tests are normal, repeat testing should through mass testing of asymptomatic in- who test negative. Community screening be carried out at least at 3-year inter- dividuals has not been proven definitively may also be poorly targeted, i.e., it may vals. (E) (and rigorous trials to provide such proof fail to reach the groups most at risk and● To test for diabetes or to assess risk of are unlikely to occur), pre-diabetes and inappropriately test those at low risk (the future diabetes, either A1C, FPG , or diabetes meet established criteria for con- worried well) or even those already diag- 2-h 75-g OGTT are appropriate. (B) ditions in which early detection is appro- nosed (19,20).● In those identified with increased risk priate. Both conditions are common, are for future diabetes, identify and, if ap- increasing in prevalence, and impose sig- propriate, treat other CVD risk factors. nificant public health burdens. There is a B. Testing for type 2 diabetes in (B) long presymptomatic phase before the di- children agnosis of type 2 diabetes is usually made. The incidence of type 2 diabetes in ado-For many illnesses there is a major dis- Relatively simple tests are available to de- lescents has increased dramatically in thetinction between screening and diagnos- tect preclinical disease (9). Additionally, last decade, especially in minority popu-tic testing. However, for diabetes the same the duration of glycemic burden is a lations (21), although the disease remainstests would be used for “screening” as for strong predictor of adverse outcomes, rare in the general pediatric populationdiagnosis. Type 2 diabetes has a long and effective interventions exist to pre- (22). Consistent with recommendationsasymptomatic phase and significant clin- vent progression of pre-diabetes to diabe- for adults, children and youth at in-ical risk markers. Diabetes may be identi- tes (see IV. PREVENTION/DELAY OF creased risk for the presence or the devel-fied anywhere along a spectrum of clinical TYPE 2 DIABETES) and to reduce risk of opment of type 2 diabetes should bescenarios ranging from a seemingly low- complications of diabetes (see VI. PRE- tested within the health care setting (23).risk individual who happens to have glu- VENTION AND MANAGEMENT OF DI- The recommendations of the ADA con-cose testing, to a higher-risk individual ABETES COMPLICATIONS). sensus statement on type 2 diabetes inS14 DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010
  5. 5. Position StatementTable 5—Testing for type 2 diabetes in asymptomatic children within ranges previously considered nor-Criteria: Overweight (BMI 85th percentile for age and sex, weight for height mal for pregnancy. For most complica- 85th percentile, or weight 120% of ideal for height) tions there was no threshold for risk.Plus any two of ● Family history of type 2 diabetes in first- or second-degree relative These results have led to careful reconsid- the following ● Race/ethnicity (Native American, African American, Latino, Asian eration of the diagnostic criteria for GDM. risk factors: American, Pacific Islander) The IADPSG recommended that all ● Signs of insulin resistance or conditions associated with insulin women not known to have prior diabetes resistance (acanthosis nigricans, hypertension, dyslipidemia, undergo a 75-g OGTT at 24 –28 weeks of polycystic ovary syndrome, or small for gestational age gestation. The group developed diagnos- birthweight) tic cut points for the fasting, 1-h, and 2-h ● Maternal history of diabetes or GDM during the child’s gestation PG measurements that conveyed an oddsAge of Age 10 years or at onset of puberty, if puberty occurs at a younger ratio for adverse outcomes of at least 1.75 initiation: age compared with women with the meanFrequency: Every 3 years glucose levels in the HAPO study. At the time of this update to the Stan- dards of Medical Care in Diabetes, ADA ischildren and youth, with some modifica- set or first recognition during pregnancy planning to work with U.S. obstetrical or-tions, are summarized in Table 5. (4). Although most cases resolve with de- ganizations to consider adoption of the livery, the definition applied whether the IADPSG diagnostic criteria and to discussC. Screening for type 1 diabetes condition persisted after pregnancy and the implications of this change. While thisGenerally, people with type 1 diabetes did not exclude the possibility that unrec- change will significantly increase thepresent with acute symptoms of diabetes ognized glucose intolerance may have an- prevalence of GDM, there is mounting ev-and markedly elevated blood glucose lev- tedated or begun concomitantly with the idence that treating even mild GDM re-els, and most cases are diagnosed soon pregnancy. This definition facilitated a duces morbidity for both mother andafter the onset of hyperglycemia. How- uniform strategy for detection and classi- baby (27).ever, evidence from type 1 diabetes pre- fication of GDM, but its limitations were Because women with a history ofvention studies suggests that measurement recognized for many years. As the ongo- GDM have a greatly increased subsequentof islet autoantibodies identifies individ- ing epidemic of obesity and diabetes has risk for diabetes (28), they should beuals who are at risk for developing type 1 led to more type 2 diabetes in women of screened for diabetes 6 –12 weeks post-diabetes. Such testing may be appropriate childbearing age, the number of pregnant partum, using nonpregnant OGTT crite-in high-risk individuals, such as those women with undiagnosed type 2 diabetes ria, and should be followed up withwith prior transient hyperglycemia or has increased (24). After deliberations in subsequent screening for the develop-those who have relatives with type 1 dia- 2008 –2009, the International Associa- ment of diabetes or pre-diabetes, as out-betes, in the context of clinical research tion of Diabetes and Pregnancy Study lined in II. TESTING FOR DIABETES INstudies (see, for example, http://www2. Groups (IADPSG), an international con- ASYMPTOMATIC PATIENTS. Widespread clini- sensus group with representatives from tion on the National Diabetes Educationcal testing of asymptomatic low-risk multiple obstetrical and diabetes organi- Program (NDEP) campaign to preventindividuals cannot currently be recom- zations, including ADA, recommended type 2 diabetes in women with GDM canmended, as it would identify very few in- that high-risk women found to have dia- be found at in the general population who betes at their initial prenatal visit using NeverTooEarly_Tipsheet.pdf.are at risk. Individuals who screen posi- standard criteria (Table 2) receive a diag-tive should be counseled about their risk nosis of overt, not gestational, diabetes. IV. PREVENTION/DELAYof developing diabetes. Clinical studies Approximately 7% of all pregnancies OF TYPE 2 DIABETESare being conducted to test various meth- (ranging from 1 to 14% depending on theods of preventing type 1 diabetes or re- population studied and the diagnostic Recommendationsversing early type 1 diabetes in those with tests used) are complicated by GDM, re- ● Patients with IGT (A), IFG (E), or anevidence of autoimmunity. sulting in more than 200,000 cases A1C of 5.7– 6.4% (E) should be re- annually. ferred to an effective ongoing supportIII. DETECTION AND Because of the risks of GDM to the program for weight loss of 5–10% ofDIAGNOSIS OF GDM mother and neonate, screening and diag- body weight and an increase in physical nosis are warranted. Current screening activity of at least 150 min/week ofRecommendations and diagnostic strategies, based on the moderate activity such as walking.● Screen for GDM using risk factor anal- 2004 ADA position statement on GDM ● Follow-up counseling appears to be im- ysis and, if appropriate, an OGTT. (C) (25), are outlined in Table 6. portant for success. (B)● Women with GDM should be screened Results of the Hyperglycemia and Ad- ● Based on potential cost savings of dia- for diabetes 6 –12 weeks postpartum verse Pregnancy Outcomes (HAPO) study betes prevention, such counseling and should be followed up with subse- (26), a large-scale ( 25,000 pregnant should be covered by third-party pay- quent screening for the development of women) multinational epidemiologic ors. (E) diabetes or pre-diabetes. (E) study, demonstrated that risk of adverse ● In addition to lifestyle counseling, met- maternal, fetal, and neonatal outcomes formin may be considered in those whoFor many years, GDM has been defined as continuously increased as a function of are at very high risk for developing di-any degree of glucose intolerance with on- maternal glycemia at 24 –28 weeks, even abetes (combined IFG and IGT DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S15
  6. 6. Standards of Medical CareTable 6—Screening for and diagnosis of GDM other drugs, the issues of cost, side effects,Carry out diabetes risk assessment at the first prenatal visit. and lack of persistence of effect in someWomen at very high risk should be screened for diabetes as soon as possible after the studies led the panel to not recommend confirmation of pregnancy. Criteria for very high risk are: use for diabetes prevention. Metformin ● Severe obesity use was recommended only for very- ● Prior history of GDM or delivery of large-for-gestational-age infant high-risk individuals (those with com- ● Presence of glycosuria bined IGT and IFG who are obese and ● Diagnosis of PCOS have at least one other risk factor for dia- ● Strong family history of type 2 diabetes betes) who are under 60 years of age. InScreening/diagnosis at this stage of pregnancy should use standard diagnostic testing (Table addition, the panel highlighted the evi- 2). dence that in the DPP, metformin wasAll women of greater than low risk of GDM, including those above not found to have diabetes most effective compared with lifestyle in early in pregnancy, should undergo GDM testing at 24–28 weeks of gestation. Low- individuals with BMI 35 kg/m2 and risk status, which does not require GDM screening, is defined as women with ALL of those under age 60 years. the following characteristics: ● Age 25 years V. DIABETES CARE ● Weight normal before pregnancy ● Member of an ethnic group with a low prevalence of diabetes A. Initial evaluation ● No known diabetes in first-degree relatives A complete medical evaluation should be ● No history of abnormal glucose tolerance performed to classify the diabetes, detect ● No history of poor obstetrical outcome the presence of diabetes complications,Two approaches may be followed for GDM screening at 24–28 weeks: review previous treatment and glycemic 1. Two-step approach: control in patients with established diabe- A. Perform initial screening by measuring plasma or serum glucose 1 h after a 50-g load tes, assist in formulating a management of 140 mg/dl identifies 80% of women with GDM, while the sensitivity is further plan, and provide a basis for continuing increased to 90% by a threshold of 130 mg/dl. care. Laboratory tests appropriate to the B. Perform a diagnostic 100-g OGTT on a separate day in women who exceed the chosen evaluation of each patient’s medical con- threshold on 50-g screening. dition should be performed. A focus on 2. One-step approach (may be preferred in clinics with high prevalence of GDM): Perform the components of comprehensive care a diagnostic 100-g OGTT in all women to be tested at 24–28 weeks. (Table 8) will assist the health care team to The 100-g OGTT should be performed in the morning after an overnight fast of at least 8 ensure optimal management of the pa- h. tient with diabetes.To make a diagnosis of GDM, at least two of the following plasma glucose values must be found: B. Management ● Fasting 95 mg/dl People with diabetes should receive med- ● 1-h 180 mg/dl ical care from a physician-coordinated ● 2-h 155 mg/dl team. Such teams may include, but are ● 3-h 140 mg/dl not limited to, physicians, nurse practitio- ners, physician’s assistants, nurses, dieti- tians, pharmacists, and mental health other risk factors such as A1C 6%, been shown to decrease incident diabetes professionals with expertise and a special hypertension, low HDL cholesterol, el- to various degrees. A summary of major interest in diabetes. It is essential in this evated triglycerides, or family history of diabetes prevention trials is shown in Ta- collaborative and integrated team ap- diabetes in a first-degree relative) and ble 7. proach that individuals with diabetes as- who are obese and under 60 years of Two studies of lifestyle intervention sume an active role in their care. age. (E) have shown persistent reduction in the The management plan should be for-● Monitoring for the development of di- role of conversion to type 2 diabetes with mulated as a collaborative therapeutic al- abetes in those with pre-diabetes 3 years (29) to 14 years (30) of postinter- liance among the patient and family, the should be performed every year. (E) vention follow-up. physician, and other members of the Based on the results of clinical trials health care team. A variety of strategiesRandomized controlled trials have shown and the known risks of progression of and techniques should be used to providethat individuals at high risk for develop- pre-diabetes to diabetes, an ADA Consen- adequate education and development ofing diabetes (those with IFG, IGT, or sus Development Panel (36) concluded problem-solving skills in the various as-both) can be given interventions that sig- that people with IGT and/or IFG should pects of diabetes management. Imple-nificantly decrease the rate of onset of di- be counseled on lifestyle changes with mentation of the management planabetes (11–17). These interventions goals similar to those of the DPP (5–10% requires that each aspect is understoodinclude intensive lifestyle modification weight loss and moderate physical activ- and agreed to by the patient and the careprograms that have been shown to be very ity of 30 min/day). Regarding the more providers and that the goals and treat-effective (58% reduction after 3 years) difficult issue of drug therapy for diabetes ment plan are reasonable. Any planand use of the pharmacologic agents met- prevention, the consensus panel felt that should recognize diabetes self-manage-formin, -glucosidase inhibitors, orlistat, metformin should be the only drug con- ment education (DSME) and on-going di-and thiazolidinediones, each of which has sidered for use in diabetes prevention. For abetes support as an integral componentS16 DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010
  7. 7. Position StatementTable 7—Therapies proven effective in diabetes prevention trials Incidence in 3-Year Mean control Relative risk number age Duration Intervention subjects reduction (%) needed toStudy (ref.) n Population (years) (years) (daily dose) (%/year) (95% CI) treat*Lifestyle Finnish DPS (12) 522 IGT, BMI 25 kg/m2 55 3.2 I-D&E 6 58 (30–70) 8.5 DPP (11) 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 (13) 259† IGT (randomized 45 6 G-D&E 14.5 38 (14–56) 7.9 groups) Toranomon Study (31) 458 IGT (men), BMI 24 55 4 I-D&E 2.4 67 (P 0.043)‡ 20.6 kg/m2 Indian DPP (17) 269† IGT 46 2.5 I-D&E 23 29 (21–37) 6.4Medications DPP (11) 2,155† IGT, BMI 24 kg/m2, 51 2.8 Metformin 10.4 31 (17–43) 13.9 FPG 5.3 mmol/l (1,700 mg) Indian DPP (17) 269† IGT 46 2.5 Metformin 23 26 (19–35) 6.9 (500 mg) STOP NIDDM (15) 1,419 IGT, FPG 5.6 54 3.2 Acarbose 12.4 25 (10–37) 9.6 mmol/l (300 mg) XENDOS (32) 3,277 BMI 30 kg/m2 43 4 Orlistat (360 2.4 37 (14–54) 45.5 mg) DREAM (16) 5,269 IGT or IFG 55 3.0 Rosiglitazone 9.1 60 (54–65) 6.9 (8 mg) Voglibose Ph-3 (33) 1,780 IGT 56 3.0 (1-year Rx) Vogliobose 12.0 40 (18–57) 21 (1-year (0.2 mg) Rx) ACT-NOW (34) 602 IGT or IFG 52 2.6 Pioglitizone 6.8 81 (61–91) 6.3 (45 mg)Modified and reprinted with permission (35). Percentage points: *Number needed to treat to prevent 1 case of diabetes, standardized for a 3-year period to improvecomparisons across studies. †Number of participants in the indicated comparisons, not necessarily in entire study. ‡Calculated from information in the article.ACT-NOW, ACTos Now Study for the Prevention of Diabetes; DPP, Diabetes Prevention Program; DPS, Diabetes Prevention Study; DREAM, Diabetes ReductionAssessment with Ramipril and Rosiglitazone Medication; STOP NIDDM, Study to Prevent Non-Insulin Dependent Diabetes; XENDOS, Xenical in the prevention ofDiabetes in Obese Subjects. I, individual; G, group; D&E, diet and exercise.of care. In developing the plan, consider- tiple insulin injections or insulin pump adherence to ongoing use of the device.ation should be given to the patient’s age, therapy. (A) (C)school or work schedule and conditions, ● For patients using less frequent insulin ● CGM may be a supplemental tool tophysical activity, eating patterns, social injections, noninsulin therapies, or SMBG in those with hypoglycemia un-situation and cultural factors, and pres- medical nutrition therapy (MNT) awareness and/or frequent hypoglyce-ence of complications of diabetes or other alone, SMBG may be useful as a guide to mic episodes. (E)medical conditions. the success of therapy. (E) ● To achieve postprandial glucose tar- The ADA consensus and position state-C. Glycemic control gets, postprandial SMBG may be appro- ments on SMBG provide a comprehensive priate. (E) review of the subject (37,38). Major clin- ● When prescribing SMBG, ensure that ical trials of insulin-treated patients that1. Assessment of glycemic control patients receive initial instruction in, demonstrated the benefits of intensiveTwo primary techniques are available for and routine follow-up evaluation of, glycemic control on diabetes complica-health providers and patients to assess the SMBG technique and using data to ad- tions have included SMBG as part ofeffectiveness of the management plan on just therapy. (E) multifactorial interventions, suggestingglycemic control: patient self-monitoring ● Continuous glucose monitoring (CGM) that SMBG is a component of effectiveof blood glucose (SMBG) or interstitial in conjunction with intensive insulin therapy. SMBG allows patients to eval-glucose and A1C. regimens can be a useful tool to lower uate their individual response to ther- A1C in selected adults (age 25 years) apy and assess whether glycemic targetsa. Glucose monitoring with type 1 diabetes (A). are being achieved. Results of SMBG can ● Although the evidence for A1C lower- be useful in preventing hypoglycemiaRecommendations ing is less strong in children, teens, and and adjusting medications (particularly● SMBG should be carried out three or younger adults, CGM may be helpful in prandial insulin doses), MNT, and more times daily for patients using mul- these groups. Success correlates with physical DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S17
  8. 8. Standards of Medical CareTable 8—Components of the comprehensive diabetes evaluation strument and user dependent (43), it isMedical history important to evaluate each patient’s mon- ● Age and characteristics of onset of diabetes (e.g., DKA, asymptomatic laboratory finding) itoring technique, both initially and at ● Eating patterns, physical activity habits, nutritional status, and weight history; growth regular intervals thereafter. In addition, and development in children and adolescents optimal use of SMBG requires proper in- ● Diabetes education history terpretation of the data. Patients should ● Review of previous treatment regimens and response to therapy (A1C records) be taught how to use the data to adjustCurrent treatment of diabetes, including medications, meal plan, physical activity patterns, food intake, exercise, or pharmacological and results of glucose monitoring and patient’s use of data therapy to achieve specific glycemic goals, ● DKA frequency, severity, and cause and these skills should be reevaluated ● Hypoglycemic episodes periodically. ● Hypoglycemia awareness CGM through the measurement of in- ● Any severe hypoglycemia: frequency and cause terstitial glucose (which correlates well ● History of diabetes-related complications with PG) is available. These sensors re- ● Microvascular: retinopathy, nephropathy, neuropathy (sensory, including history of quire calibration with SMBG, and the lat- foot lesions; autonomic, including sexual dysfunction and gastroparesis) ter are still recommended for making ● Macrovascular: CHD, cerebrovascular disease, PAD acute treatment decisions. CGM devices ● Other: psychosocial problems*, dental disease* also have alarms for hypo- and hypergly-Physical examination cemic excursions. Small studies in se- ● Height, weight, BMI lected patients with type 1 diabetes have ● Blood pressure determination, including orthostatic measurements when indicated suggested that CGM use reduces the time ● Fundoscopic examination* spent in hypo- and hyperglycemic ranges ● Thyroid palpation and may modestly improve glycemic con- ● Skin examination (for acanthosis nigricans and insulin injection sites) trol. A larger 26-week randomized trial of ● Comprehensive foot examination: 322 type 1 diabetic patients showed that ● Inspection adults age 25 years and older using inten- ● Palpation of dorsalis pedis and posterior tibial pulses sive insulin therapy and CGM experi- ● Presence/absence of patellar and Achilles reflexes enced a 0.5% reduction in A1C (from ● Determination of proprioception, vibration, and monofilament sensation 7.6 to 7.1%) compared with usual in-Laboratory evaluation tensive insulin therapy with SMBG (44). ● A1C, if results not available within past 2–3 months Sensor use in children, teens, and adults ● If not performed/available within past year: to age 24 years did not result in significant ● Fasting lipid profile, including total, LDL- and HDL cholesterol and triglycerides A1C lowering, and there was no signifi- ● Liver function tests cant difference in hypoglycemia in any ● Test for urine albumin excretion with spot urine albumin/creatinine ratio group. Importantly, the greatest predictor ● Serum creatinine and calculated GFR of A1C lowering in this study for all age- ● TSH in type 1 diabetes, dyslipidemia, or women over age 50 years groups was frequency of sensor use,Referrals which was lower in younger age-groups. ● Annual dilated eye exam In a smaller randomized controlled trial of ● Family planning for women of reproductive age 129 adults and children with baseline ● Registered dietitian for MNT A1C 7.0%, outcomes combining A1C ● DSME and hypoglycemia favored the group us- ● Dental examination ing CGM, suggesting that CGM is also ● Mental health professional, if needed beneficial for individuals with type 1 dia- betes who have already achieved excellent* See appropriate referrals for these categories. control with A1C 7.0% (45). Although CGM is an evolving technology, emerging The frequency and timing of SMBG clear. A meta-analysis of SMBG in non– data suggest that it may offer benefit inshould be dictated by the particular needs insulin-treated patients with type 2 appropriately selected patients who areand goals of the patient. SMBG is espe- diabetes concluded that some regimen of motivated to wear it most of the time.cially important for patients treated with SMBG was associated with a reduction in CGM may be particularly useful in thoseinsulin in order to monitor for and pre- A1C of 0.4%. However, many of the stud- with hypoglycemia unawareness and/orvent asymptomatic hypoglycemia and hy- ies in this analysis also included patient frequent episodes of hypoglycemia, andperglycemia. For most patients with type education with diet and exercise counsel- studies in this area are ongoing.1 diabetes and pregnant women taking ing and, in some cases, pharmacologic in-insulin, SMBG is recommended three or tervention, making it difficult to assess the b. A1Cmore times daily. For these populations, contribution of SMBG alone to improvedsignificantly more frequent testing may be control (39). Several recent trials have Recommendationsrequired to reach A1C targets safely with- called into question the clinical utility ● Perform the A1C test at least two timesout hypoglycemia. The optimal frequency and cost-effectiveness of routine SMBG in a year in patients who are meeting treat-and timing of SMBG for patients with type non–insulin-treated patients (40 – 42). ment goals (and who have stable glyce-2 diabetes on noninsulin therapy is un- Because the accuracy of SMBG is in- mic control). (E)S18 DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010
  9. 9. Position Statement● Perform the A1C test quarterly in pa- Table 9—Correlation of A1C with average ant, clinicians should consider the possi- tients whose therapy has changed or glucose bilities of hemoglobinopathy or altered who are not meeting glycemic goals. (E) red cell turnover and the options of more● Use of point-of-care testing for A1C al- frequent and/or different timing of SMBG Mean plasma glucose lows for timely decisions on therapy or use of CGM. Other measures of chronic changes, when needed. (E) A1C (%) mg/dl mmol/l glycemia such as fructosamine are avail- 6 126 7.0 able, but their linkage to average glucoseBecause A1C is thought to reflect average 7 154 8.6 and their prognostic significance are notglycemia over several months (43) and 8 183 10.2 as clear as is the case for A1C.has strong predictive value for diabetes 9 212 11.8complications (11,46), A1C testing 10 240 13.4 2. Glycemic goals in adultsshould be performed routinely in all pa- ● Lowering A1C to below or around 7% 11 269 14.9tients with diabetes, at initial assessment 12 298 16.5 has been shown to reduce microvascu-and then as part of continuing care. Mea- These estimates are based on ADAG data of 2,700 lar and neuropathic complications ofsurement approximately every 3 months glucose measurements over 3 months per A1C mea- type 1 and type 2 diabetes. Therefore,determines whether a patient’s glycemic surement in 507 adults with type 1, type 2, and no for microvascular disease prevention,targets have been reached and main- diabetes. The correlation between A1C and average the A1C goal for nonpregnant adults intained. For any individual patient, the fre- glucose was 0.92 (49). A calculator for converting general is 7%. (A) A1C results into estimated average glucose (eAG),quency of A1C testing should be ● In type 1 and type 2 diabetes, random- in either mg/dl or mmol/l, is available atdependent on the clinical situation, the ized controlled trials of intensive versustreatment regimen used, and the judg- standard glycemic control have notment of the clinician. Some patients with shown a significant reduction in CVDstable glycemia well within target may do porting both an A1C result and an esti- outcomes during the randomized por-well with testing only twice per year, mated average glucose (eAG) result when tion of the trials. Long-term follow-upwhile unstable or highly intensively man- a clinician orders the A1C test. In previ- of the DCCT and UK Prospective Dia-aged patients (e.g., pregnant type 1 dia- ous versions of the Standards of Medical betes Study (UKPDS) cohorts suggestsbetic women) may be tested more Care in Diabetes, the table describing the that treatment to A1C targets below orfrequently than every 3 months. The correlation between A1C and mean glu- around 7% in the years soon after theavailability of the A1C result at the time cose was derived from relatively sparse diagnosis of diabetes is associated withthat the patient is seen (point-of-care test- data (one seven-point profile over 1 day long-term reduction in risk of macro-ing) has been reported to result in in- per A1C reading) in the primarily Cauca- vascular disease. Until more evidencecreased intensification of therapy and sian type 1 participants in the DCCT (50). becomes available, the general goal ofimprovement in glycemic control Clinicians should note that the numbers 7% appears reasonable for many(47,48). in the table are now different, as they are adults for macrovascular risk reduc- The A1C test is subject to certain lim- based on 2,800 readings per A1C in the tion. (B)itations. Conditions that affect erythro- ADAG trial. ● Subgroup analyses of clinical trials suchcyte turnover (hemolysis, blood loss) and In the ADAG trial, there were no sig- as the DCCT and UKPDS, and evidencehemoglobin variants must be considered, nificant differences among racial and eth- for reduced proteinuria in the Action inparticularly when the A1C result does not nic groups in the regression lines between Diabetes and Vascular Disease: Preteraxcorrelate with the patient’s clinical situa- A1C and mean glucose, although there and Diamicron Modified Release Con-tion (43). In addition, A1C does not pro- was a trend toward a difference between trolled Evaluation (ADVANCE) trialvide a measure of glycemic variability or Africans/African Americans participants suggest a small but incremental benefithypoglycemia. For patients prone to gly- and Caucasians that might have been sig- in microvascular outcomes with A1Ccemic variability (especially type 1 dia- nificant had more Africans/African Amer- values closer to normal. Therefore, forbetic patients, or type 2 diabetic patients icans been studied. A recent study selected individual patients, providerswith severe insulin deficiency), glycemic comparing A1C to CGM data in 48 type 1 might reasonably suggest even lowercontrol is best judged by the combination diabetic children found a highly statisti- A1C goals than the general goal ofof results of SMBG testing and the A1C. cally significant correlation between A1C 7%, if this can be achieved withoutThe A1C may also serve as a check on the and mean blood glucose, although the significant hypoglycemia or other ad-accuracy of the patient’s meter (or the pa- correlation (r 0.7) was significantly verse effects of treatment. Such patientstient’s reported SMBG results) and the ad- lower than in the ADAG trial (51). might include those with short dura-equacy of the SMBG testing schedule. Whether there are significant differences tion of diabetes, long life expectancy, Table 9 contains the correlation be- in how A1C relates to average glucose in and no significant CVD. (B)tween A1C levels and mean PG levels children or in African American patients ● Conversely, less-stringent A1C goalsbased on data from the international A1C- is an area for further study. For the time than the general goal of 7% may beDerived Average Glucose (ADAG) trial being, the question has not led to different appropriate for patients with a historyusing frequent SMBG and CGM in 507 recommendations about testing A1C or of severe hypoglycemia, limited life ex-adults (83% Caucasian) with type 1, type different interpretations of the clinical pectancy, advanced microvascular or2, and no diabetes (49). ADA and the meaning of given levels of A1C in those macrovascular complications, and ex-American Association of Clinical Chem- populations. tensive comorbid conditions and thoseists have determined that the correlation For patients in whom A1C/eAG and with longstanding diabetes in whom(r 0.92) is strong enough to justify re- measured blood glucose appear discrep- the general goal is difficult to attain DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S19
  10. 10. Standards of Medical Care spite diabetes self-management educa- analyses also suggest that further lowering such as stroke. In an epidemiologic anal- tion, appropriate glucose monitoring, of A1C from 7 to 6% is associated with ysis of the study cohort, a continuous as- and effective doses of multiple glucose- further reduction in the risk of microvas- sociation was observed such that for every lowering agents including insulin. (C) cular complications, albeit the absolute percentage point lower median on-study risk reductions become much smaller. A1C (e.g., 8 –7%), there was a statisticallyGlycemic control is fundamental to the The ADVANCE study of intensive versus significant 18% reduction in CVD events,management of diabetes. The DCCT, a standard glycemic control in type 2 dia- again with no glycemic threshold. A re-prospective, randomized, controlled trial betes found a statistically significant re- cent report of 10 years of follow-up of theof intensive versus standard glycemic duction in albuminuria with an A1C UKPDS cohort described, for the partici-control in patients with relatively recently target of 6.5% (achieved median A1C pants originally randomized to intensivediagnosed type 1 diabetes, showed defin- 6.3%) compared with standard therapy glycemic control compared with thoseitively that improved glycemic control is achieving a median A1C of 7.0% (63). randomized to conventional glycemicassociated with significantly decreased Given the substantially increased risk of control, long-term reductions in MI (15%rates of microvascular (retinopathy and hypoglycemia (particularly in those with with sulfonylurea or insulin as initialnephropathy) as well as neuropathic type 1 diabetes, but also in the recent type pharmacotherapy, 33% with metformincomplications (53). Follow-up of the 2 diabetes trials described below), the as initial pharmacotherapy, both statisti-DCCT cohorts in the Epidemiology of Di- concerning mortality findings in the Ac- cally significant) and in all-cause mortal-abetes Interventions and Complications tion to Control Cardiovascular Risk in Di- ity (13 and 27%, respectively, both(EDIC) study has shown persistence of abetes (ACCORD) trial described below statistically significant) (59).this effect in previously intensively and the relatively much greater effort re- Because of ongoing uncertainty re-treated subjects, even though their glyce- quired to achieve near-normoglycemia, garding whether intensive glycemic con-mic control has been equivalent to that of the risks of lower targets may outweigh trol can reduce the increased risk of CVDprevious standard arm subjects during the potential benefits on microvascular events in people with type 2 diabetes, sev-follow-up (54,55). complications on a population level. eral large long-term trials were launched In type 2 diabetes, the Kumamoto However, selected individual patients, es- in the past decade to compare the effectsstudy (56) and the UKPDS (57,58) dem- pecially those with little comorbidity and of intensive versus standard glycemiconstrated significant reductions in micro- long life expectancy (who may reap the control on CVD outcomes in relativelyvascular and neuropathic complications benefits of further lowering glycemia be- high-risk participants with establishedwith intensive therapy. Similar to the low 7%) may, at patient and provider type 2 diabetes. In 2008, results of threeDCCT-EDIC findings, long-term fol- judgment, adopt glycemic targets as close large trials (ACCORD, ADVANCE, andlow-up of the UKPDS cohort has recently to normal as possible as long as significant VADT) suggested no significant reductiondemonstrated a “legacy effect” of early in- hypoglycemia does not become a barrier. in CVD outcomes with intensive glycemictensive glycemic control on long-term Whereas many epidemiologic studies control in these populations. Details ofrates of microvascular complications, and meta-analyses (64,65) have clearly these three studies are shown in Table 10,even with loss of glycemic separation be- shown a direct relationship between A1C and their results and implications are re-tween the intensive and standard cohorts and CVD, the potential of intensive glyce- viewed more extensively in a recent ADAafter the end of the randomized con- mic control to reduce CVD has been less position statement (52).trolled trial (59). The more recent Veter- clearly defined. In the DCCT, there was a The ACCORD study randomizedans Affairs Diabetes Trial (VADT) in type trend toward lower risk of CVD events 10,251 participants with either history of2 diabetes also showed significant reduc- with intensive control (risk reduction a CVD event or significant CVD risk to ations in albuminuria with intensive 41%, 95% CI 10 – 68%), but the number strategy of intensive glycemic control (tar-(achieved median A1C 6.9%) compared of events was small. However, 9-year get A1C 6.0%) or standard glycemicwith standard glycemic control but no post-DCCT follow-up of the cohort has control (A1C target 7.0 –7.9%). Investiga-difference in retinopathy and neuropathy shown that participants previously ran- tors used multiple glycemic medications(60,61). domized to the intensive arm had a 42% in both arms. From a baseline median In each of these large randomized reduction (P 0.02) in CVD outcomes A1C of 8.1%, the intensive arm reached aprospective clinical trials, treatment regi- and a 57% reduction (P 0.02) in the median A1C of 6.4% within 12 months ofmens that reduced average A1C to 7% risk of nonfatal myocardial infarction randomization, while the standard group(1% above the upper limits of normal) (MI), stroke, or CVD death compared reached a median A1C of 7.5%. Otherwere associated with fewer markers of with participants previously in the stan- risk factors were treated aggressively andlong-term microvascular complications; dard arm (66). The benefit of intensive equally in both groups. The intensive gly-however, intensive control was found to glycemic control in this type 1 diabetic cemic control group had more use of in-increase the risk of severe hypoglycemia cohort has recently been shown to persist sulin in combination with multiple oraland led to weight gain (46,60,62). for up to 30 years (67). agents, significantly more weight gain, Epidemiological analyses of the The UKPDS trial of type 2 diabetes and more episodes of severe hypoglyce-DCCT and UKPDS (46,53) demonstrate a observed a 16% reduction in cardiovascu- mia than the standard group.curvilinear relationship between A1C and lar complications (combined fatal or non- In early 2008, the glycemic controlmicrovascular complications. Such anal- fatal MI and sudden death) in the arm of ACCORD was halted on the rec-yses suggest that, on a population level, intensive glycemic control arm, although ommendation of the study’s data safetythe greatest number of complications will this difference was not statistically signif- monitoring board due to the finding of anbe averted by taking patients from very icant (P 0.052), and there was no sug- increased rate of mortality in the intensivepoor control to fair or good control. These gestion of benefit on other CVD outcomes arm compared with the standard armS20 DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010
  11. 11. Position StatementTable 10—Comparison of the three trials of intensive glycemic control and CVD outcomes ACCORD ADVANCE VADTParticipant characteristics n 10,251 11,140 1,791 Mean age (years) 62 66 60 Duration of diabetes (years) 10 8 11.5 History of CVD (%) 35 32 40 Median baseline A1C (%) 8.1 7.2 9.4 On insulin at baseline (%) 35 1.5 52Protocol characteristics A1C goals (%) (I vs. S)* 6.0 vs. 7.0–7.9 6.5 vs. “based on local guidelines” 6.0 (action if 6.5) vs. planned separation of 1.5 Protocol for glycemic control (I vs. S)* Multiple drugs in both Multiple drugs added to gliclizide vs. Multiple drugs in both arms arms multiple drugs with no gliclizide Management of other risk factors Embedded blood pressure Embedded blood pressure trial Protocol for intensive and lipid trials treatment in both armsOn-study characteristics Achieved median A1C (%) (I vs. S) 6.4 vs. 7.5 6.3 vs. 7.0 6.9 vs. 8.5 On insulin at study end (%) (I vs. S)* 77 vs. 55* 40 vs. 24 89 vs. 0.74Weight changes (kg) Intensive glycemic control arm 3.5 0.1 7.8 Standard glycemic control arm 0.4 1.0 3.4 Severe hypoglycemia (participants with one or more episodes during study) (%) Intensive glycemic control arm 16.2 2.7 21.2 Standard glycemic control arm 5.1 1.5 9.9Outcomes Definition of primary outcome Nonfatal MI, nonfatal Microvascular plus macrovascular Nonfatal MI, nonfatal stroke, stroke, CVD death (nonfatal MI, nonfatal stroke, CVD CVD death, death) outcomes hospitalization for heart failure, revascularization HR for primary outcome (95% CI) 0.90 (0.78–1.04) 0.9 (0.82–0.98); 0.88 (0.74–1.05) macrovascular 0.94 (0.84–1.06) HR for mortality findings (95% CI) 1.22 (1.01–1.46) 0.93 (0.83–1.06) 1.07 (0.81–1.42)*Insulin rates for ACCORD are for any use during the study. I, intensive glycemic control; S, standard glycemic control. Abridged from ref. 52.(1.41 vs. 1.14%/year, hazard ratio 1.22, The cause of excess deaths in the in- achieved A1C levels 7% or in those who95% CI 1.01–1.46), with a similar in- tensive group of the ACCORD has been lowered their A1C quickly after trial en-crease in cardiovascular deaths. The pri- difficult to pinpoint (and is discussed in rollment. In fact, the converse was ob-mary outcome of ACCORD (MI, stroke, some detail in a 2009 ADA position state- served: those at highest risk for mortalityor cardiovascular death) was lower in the ment [52]). However, exploratory analy- were participants in the intensive armintensive glycemic control group due to a ses of the mortality findings of ACCORD with the highest A1C levels.reduction in nonfatal MI, although this (evaluating variables including weight The ADVANCE study randomizedfinding was not statistically significant gain, use of any specific drug or drug participants to a strategy of intensive gly-when the study was terminated (68). Of combination, and hypoglycemia) were re- cemic control (with primary therapy be-note, prespecified subset analyses showed portedly unable to identify a clear expla- ing the sulfonylurea gliclizide andthat participants with no previous CVD nation for the excess mortality in the additional medications as needed toevent and those who had a baseline A1C intensive arm. At the 69th Scientific Ses- achieve a target A1C of 6.5%) or to stan- 8% had a statistically significant reduc- sions of the American Diabetes Associa- dard therapy (in which any medicationtion in the primary CVD outcome, al- tion, the ACCORD investigators but gliclizide could be used and the gly-though overall mortality was not reduced presented additional analyses showing no cemic target was according to “localin these groups. increase in mortality in participants who guidelines”). ADVANCE DIABETES CARE, VOLUME 33, SUPPLEMENT 1, JANUARY 2010 S21