report of the expert committee on the diagnosis and ... · dent diabetes mellitus (iddm, type 1...

R e v i e w s / C o m m e n t a r i e s / P o s i t i o n S t a t e m e n t s

C O M M I T T E E R E P O R T

Report of the Expert Committee onthe Diagnosis and Classification ofDiabetes MellitusTHE EXPERT COMMITTEE ON THE

DIAGNOSIS AND CLASSIFICATION OFDIABETES MELLITUS*

The current classification and diagnosisof diabetes used in the U.S. was devel-oped by the National Diabetes Data

Group (NDDG) and published in 1979(1). The impetus for the classification anddiagnosis scheme proposed then holds truetoday. That is

the growth of knowledge regarding the eti-ology and pathogenesis of diabetes has ledmany individuals and groups in the dia-betes community to express the need for arevision of the nomenclature, diagnosticcriteria, and classification of diabetes. As aconsequence, it was deemed essential todevelop an appropriate, uniform terminol-ogy and a functional, working classificationof diabetes that reflects the current knowl-edge about the disease. (1)

It is now considered to be particularlyimportant to move away from a systemthat appears to base the classification of thedisease, in large part, on the type of phar-macological treatment used in its manage-ment toward a system based on diseaseetiology where possible.

An international Expert Committee,working under the sponsorship of theAmerican Diabetes Association, was estab-lished in May 1995 to review the scientificliterature since 1979 and to decide ifchanges to the classification and diagnosisof diabetes were warranted. The Commit-tee met on multiple occasions and widelycirculated a draft report of their findings

and preliminary recommendations to theinternational diabetes community. Basedon the numerous comments and sugges-tions received, including the opportunity toreview unpublished data in detail, theCommittee discussed and revised numer-ous drafts of a manuscript that culminatedin this published document.

This report is divided into four sec-tions: definition and description of diabetes,classification of the disease, diagnostic cri-teria, and testing for diabetes. The aim ofthis document is to define and describe dia-betes as we know it today, present a classi-fication scheme that reflects its etiologyand/or pathogenesis, provide guidelines forthe diagnosis of the disease, develop rec-ommendations for testing that can helpreduce the morbidity and mortality associ-ated with diabetes, and review the diagno-sis of gestational diabetes.

DEFINITION AND DESCRIPTIONOF DIABETES MELLITUS— Dia-betes mellitus is a group of metabolic dis-eases characterized by hyperglycemiaresulting from defects in insulin secretion,insulin action, or both. The chronic hyper-glycemia of diabetes is associated withlong-term damage, dysfunction, and failureof various organs, especially the eyes, kid-neys, nerves, heart, and blood vessels.

Several pathogenic processes areinvolved in the development of diabetes.

From the American Diabetes Association, Alexandria, Virginia.Address correspondence and reprint requests to Richard Kahn, PhD, American Diabetes Association, 1660

Duke St., Alexandria, VA 22314.*For a complete list of committee members see p. 1194.ACOG, American College of Obstetricians and Gynecologists; FPG, fasting plasma glucose; GDM, gesta-

tional diabetes mellitus; HNF, hepatocyte nuclear factor; IFG, impaired fasting glucose; IGT, impaired glu-cose tolerance; MODY, maturity-onset diabetes of the young; NDDG, National Diabetes Data Group;NHANES III, Third National Health and Nutrition Examination Survey; OGTT, oral glucose tolerance test;PA1-1, plasminogen activator inhibitor-1; WHO, World Health Organization; 2hPG, 2-h postload glucose.

These range from autoimmune destructionof the (B-cells of the pancreas with conse-quent insulin deficiency to abnormalitiesthat result in resistance to insulin action. Thebasis of the abnormalities in carbohydrate,fat, and protein metabolism in diabetes isdeficient action of insulin on target tissues.Deficient insulin action results from inade-quate insulin secretion and/or diminishedtissue responses to insulin at one or morepoints in the complex pathways of hormoneaction. Impairment of insulin secretion anddefects in insulin action frequently coexist inthe same patient, and it is often unclearwhich abnormality, if either alone, is the pri-mary cause of the hyperglycemia.

Symptoms of marked hyperglycemiainclude polyuria, polydipsia, weight loss,sometimes with polyphagia, and blurredvision. Impairment of growth and suscep-tibility to certain infections may alsoaccompany chronic hyperglycemia. Acute,life-threatening consequences of diabetesare hyperglycemia with ketoacidosis or thenonketotic hyperosmolar syndrome.

Long-term complications of diabetesinclude: retinopathy with potential loss ofvision; nephropathy leading to renal failure;peripheral neuropathy with risk of footulcers, amputation, and Charcot joints; andautonomic neuropathy causing gastroin-testinal, genitourinary, and cardiovascularsymptoms and sexual dysfunction. Glyca-tion of tissue proteins and other macromol-ecules and excess production of polyolcompounds from glucose are among themechanisms thought to produce tissue dam-age from chronic hyperglycemia. Patientswith diabetes have an increased incidence ofatherosclerotic cardiovascular, peripheralvascular, and cerebrovascular disease.Hypertension, abnormalities of lipoproteinmetabolism, and periodontal disease areoften found in people with diabetes. Theemotional and social impact of diabetes andthe demands of therapy may cause signifi-cant psychosocial dysfunction in patientsand their families.

The vast majority of cases of diabetesfall into two broad etiopathogenetic cate-gories (discussed in greater detail below).In one category (type 1 diabetes), the cause

DIABETES CARE, VOLUME 20, NUMBER 7, JULY 1997 1183©

Committee Report

is an absolute deficiency of insulin secre-tion. Individuals at increased risk of devel-oping this type of diabetes can often beidentified by serological evidence of anautoimmune pathologic process occurringin the pancreatic islets and by geneticmarkers. In the other, much more preva-lent category (type 2 diabetes), the cause isa combination of resistance to insulinaction and an inadequate compensatoryinsulin secretory response. In the latter cat-egory, a degree of hyperglycemia sufficientto cause pathologic and functional changesin various target tissues, but without clini-cal symptoms, may be present for a longperiod of time before diabetes is detected.During this asymptomatic period, it is pos-sible to demonstrate an abnormality in car-bohydrate metabolism by measurement ofplasma glucose in the fasting state or aftera challenge with an oral glucose load.

CLASSIFICATION OF DIABETESMELLITUS AND OTHERCATEGORIES OF GLUCOSEREGULATION— A major require-ment for epidemiological and clinicalresearch and for the clinical management ofdiabetes is an appropriate system of classi-fication that provides a framework withinwhich to identify and differentiate its vari-ous forms and stages. While there havebeen a number of sets of nomenclature anddiagnostic criteria proposed for diabetes,no generally accepted systematic catego-rization existed until the NDDG classifica-tion system was published in 1979 (1).The World Health Organization (WHO)Expert Committee on Diabetes in 1980and later the WHO Study Group on Dia-betes Mellitus endorsed the substantive rec-ommendations of the NDDG (2). Thesegroups recognized two major forms of dia-betes, which they termed insulin-depen-dent diabetes mellitus (IDDM, type 1diabetes) and non-insulin-dependent dia-betes mellitus (NIDDM, type 2 diabetes),but their classification system went on toinclude evidence that diabetes mellitus wasan etiologically and clinically heteroge-neous group of disorders that share hyper-glycemia in common. The overwhelmingevidence in favor of this heterogeneityincluded the following:

1. There are several distinct disorders,most of them rare, in which glucoseintolerance is a feature.

2. There are large differences in the preva-lence of the major forms of diabetes

among various racial or ethnic groupsworldwide.

3. Patients with glucose intolerance pres-ent with great phenotypic variation;take for example the differencesbetween thin, ketosis-prone, insulin-dependent diabetes and obese, nonke-totic, insulin-resistant diabetes.

4. Evidence from genetic, immunologi-cal, and clinical studies shows that inwestern countries, the forms of dia-betes that have their onset primarily inyouth are distinct from those that havetheir onset mainly in adulthood.

5. A type of non-insulin-requiring dia-betes in young people, inherited in anautosomal dominant fashion, isclearly different from the classic acute-onset diabetes that typically occurs inchildren.

6. In tropical countries, several clinicalpresentations occur, including dia-betes associated with fibrocalcific pan-creatitis.These and other lines of evidence were

used to divide diabetes mellitus into fivedistinct types (IDDM, NIDDM, gestationaldiabetes mellitus [GDM], malnutrition-related diabetes, and other types). The dif-ferent clinical presentations and geneticand environmental etiologic factors of thefive types permitted discrimination amongthem. All five types were characterized byeither fasting hyperglycemia or elevatedlevels of plasma glucose during an oral glu-cose tolerance test (OGTT). In addition, the1979 classification included the category ofimpaired glucose tolerance (IGT), in whichplasma glucose levels during an OGTTwere above normal but below thosedefined as diabetes.

The NDDG/WHO classification high-lighted the heterogeneity of the diabeticsyndrome. Such heterogeneity has hadimportant implications not only for treat-ment of patients with diabetes but also forbiomedical research. This previous classifi-cation indicated that the disorders groupedtogether under the term diabetes differmarkedly in pathogenesis, natural history,response to therapy, and prevention. Inaddition, different genetic and environ-mental factors can result in forms of dia-betes that appear phenotypically similarbut may have different etiologies.

The classification published in 1979was based on knowledge of diabetes at thattime and represented some compromisesamong different points of view. It was basedon a combination of clinical manifestations

or treatment requirements (e.g., insulin-dependent, non-insulin-dependent) andpathogenesis (e.g., malnutrition-related,"other types," gestational). It was anticipated,however, that as knowledge of diabetes con-tinued to develop, the classification wouldneed revision. When the classification wasprepared, a definitive etiology had not beenestablished for any of the diabetes sub-classes, except for some of the "other types."Few susceptibility genes for diabetes hadbeen discovered, and an understanding ofthe immunological basis for most type 1diabetes was just beginning.

The current Expert Committee hascarefully considered the data and rationalefor what was accepted in 1979, along withresearch findings of the last 18 years, andwe are now proposing changes to theNDDG/WHO classification scheme (Table1). The main features of these changes areas follows:

1. The terms insulin-dependent diabetesmellitus and non-insulin-dependentdiabetes mellitus and their acronyms,IDDM and NIDDM, are eliminated.These terms have been confusing andhave frequently resulted in classifyingthe patient based on treatment ratherthan etiology.

2. The terms type 1 and type 2 diabetesare retained, with arabic numeralsbeing used rather than roman numer-als. We recommend adoption of arabicnumerals in part because the romannumeral II can easily be confused bythe public as the number 11. The class,or form, named type 1 diabetes encom-passes the vast majority of cases that areprimarily due to pancreatic islet (3-celldestruction and that are prone toketoacidosis. This form includes thosecases currently ascribable to an autoim-mune process and those for which anetiology is unknown. It does notinclude those forms of (3-cell destruc-tion or failure for which non-autoim-mune-specific causes can be assigned(e.g., cystic fibrosis). While most type 1diabetes is characterized by the pres-ence of islet cell, GAD, IA-2, IA-20, orinsulin autoantibodies that identify theautoimmune process that leads to IB-cell destruction, in some subjects, noevidence of autoimmunity is present;these cases are classified as type 1 idio-pathic.

3. The class, or form, named type 2 dia-betes includes the most prevalent formof diabetes, which results from insulin

1184 DIABETES CARE, VOLUME 20, NUMBER 7, JULY 1997©

Committee Report

resistance with an insulin secretorydefect.

4. A recent international meeting reviewedthe evidence for and characteristics ofmalnutrition-related diabetes (3). Whileit appears that malnutrition may influ-ence the expression of other types ofdiabetes, the evidence that diabetes canbe directly caused by protein deficiencyis not convincing. Therefore, the classmalnutrition-related diabetes mellitushas been eliminated. Fibrocalculouspancreatopathy (formerly a subtype ofmalnutrition-related diabetes) has beenreclassified as a disease of the exocrinepancreas.

5. The stage impaired glucose tolerance(IGT) has been retained. The analo-gous intermediate stage of fasting glu-cose is named impaired fasting glucose(IFG).

6. Gestational diabetes mellitus (GDM) isretained as defined by the WHO andNDDG, respectively. Selective ratherthan universal screening for glucoseintolerance in pregnancy is now rec-ommended.

7. The degree of hyperglycemia (if any)may change over time, depending onthe extent of the underlying diseaseprocess (Fig. 1). A disease process maybe present but may not have pro-gressed far enough to cause hyper-glycemia. The same disease process cancause IFG and/or IGT without fulfillingthe criteria for the diagnosis of dia-betes. In some individuals with dia-betes, adequate glycemic control canbe achieved with weight reduction,exercise, and/or oral agents. Theseindividuals therefore do not requireinsulin. Other individuals who havesome residual insulin secretion butrequire exogenous insulin for adequateglycemic control can survive withoutit. Individuals with extensive (3-celldestruction and therefore no residualinsulin secretion require insulin forsurvival. The severity of the metabolicabnormality can progress, regress, orstay the same. Thus, the degree ofhyperglycemia reflects the severity ofthe underlying metabolic process andits treatment more than the nature ofthe process itself.

8. Assigning a type of diabetes to an indi-vidual often depends on the circum-stances present at the time of diagnosis,and many diabetic individuals do noteasily fit into a single class. For exam-

Table 1—Etiologic classification of diabetes mellitus

I. Type 1 diabetes* ((3-cell destruction, usually leading to absolute insulin deficiency)A. Immune mediatedB. Idiopathic

II. Type 2 diabetes* (may range from predominantly insulin resistance with relative insulin deficiencyto a predominantly secretory defect with insulin resistance)

III. Other specific typesA. Genetic defects of (3-cell function

1. Chromosome 12, HNF-la (formerly M0DY3)2. Chromosome 7, glucokinase (formerly M0DY2)3. Chromosome 20, HNF-4a (formerly MODY1)4. Mitochondrial DNA5. Others

B. Genetic defects in insulin action1. Type A insulin resistance2. Leprechaunism3. Rabson-Mendenhall syndrome4. Lipoatrophic diabetes5. Others

C. Diseases of the exocrine pancreas1. Pancreatitis2. Trauma/pancreatectomy3. Neoplasia4. Cystic fibrosis5. Hemochromatosis6. Fibrocalculous pancreatopathy7. Others

D. Endocrinopathies1. Acromegaly2. Cushing's syndrome3. Glucagonoma4. Pheochromocytoma5. Hyperthyroidism6. Somatostatinoma7. Aldosteronoma8. Others

E. Drug- or chemical-inducedVacorPentamidineNicotinic acidGlucocorticoidsThyroid hormone

IV

6. Diazoxide7. (3-adrenergic agonists8. Thiazides9. Dilantin

10. a-Interferon11. Others

E Infections1. Congenital rubella2. Cytomegalovirus3. Others

G. Uncommon forms of immune-mediated diabetes1. "Stiff-man" syndrome2. Anti-insulin receptor antibodies3. Others

H. Other genetic syndromes sometimes associated with diabetes1. Down's syndrome2. Klinefelter's syndrome3. Turner's syndrome4. Wolfram's syndrome5. Friedreich's ataxia6. Huntington's chorea7. Lawrence Moon Beidel syndrome8. Myotonic dystrophy9. Porphyria

10. Prader Willi syndrome11. Others

Gestational diabetes mellitus (GDM)

*Patients with any form of diabetes may require insulin treatment at some stage of their disease. Such useof insulin does not, of itself, classify the patient.


Committee Report

Stages

Types

Normoglycemia

Normal glucose regulation

Hyperglycemia

Impaired Glucose Toleranceor

Impaired Fasting Glucose

Diabetes Mellitus

Not insulin Insulin requiring Insulin requiringrequiring for control for survival

Type 1*

Type 2

Other Specific Types'

Gestational Diabetes

Figure 1—Disorders ofglycemia: etiologic types and stages. *Even after presenting in ketoaddosis, these patients can briefly return to normoglycemia with-out requiring continuous therapy (i.e., "honeymoon" remission). **ln rare instances, patients in these categories (e.g., Vacor toxicity, type 1 diabetes pre-senting in pregnancy) may require insulin for survival.

pie, a person with GDM may continueto be hyperglycemic after delivery andmay be determined to have, in fact,type 1 diabetes. Alternatively, a personwho acquires diabetes because of largedoses of exogenous steroids maybecome normoglycemic once the glu-cocorticoids are discontinued, but thenmay develop diabetes many years laterafter recurrent episodes of pancreatitis.Another example would be a persontreated with thiazides who developsdiabetes years later. Because thiazidesin themselves seldom cause severehyperglycemia, such individuals prob-ably have type 2 diabetes that is exac-erbated by the drug. Thus, for theclinician and patient, it is less impor-tant to label the particular type of dia-betes than it is to understand thepathogenesis of the hyperglycemia andto treat it effectively.

Type 1 diabetes (pi-cell destruction,usually leading to absolute insulindeficiency)Immune-mediated diabetes. This form ofdiabetes, previously encompassed by theterms insulin-dependent diabetes, type 1diabetes, or juvenile-onset diabetes, resultsfrom a cellular-mediated autoimmunedestruction of the (B-cells of the pancreas

(4). Markers of the immune destruction ofthe P-cell include islet cell autoantibodies(ICAs), autoantibodies to insulin (IAAs),autoantibodies to glutamic acid decar-boxylase (GAD65), and autoantibodies tothe tyrosine phosphatases 1A-2 and lA-2(3(5-13). One and usually more of theseautoantibodies are present in 85-90% ofindividuals when fasting hyperglycemia isinitially detected. Also, the disease hasstrong HLA associations, with linkage tothe DQA and B genes, and it is influencedby the DRB genes (14,15). These HLA-DR/DQ alleles can be either predisposing orprotective.

In this form of diabetes, the rate of (3-cell destruction is quite variable, beingrapid in some individuals (mainly infantsand children) and slow in others (mainlyadults [16]). Some patients, particularlychildren and adolescents, may present withketoacidosis as the first manifestation of thedisease. Others have modest fasting hyper-glycemia that can rapidly change to severehyperglycemia and/or ketoacidosis in thepresence of infection or other stress. Stillothers, particularly adults, may retain resid-ual p-cell function sufficient to preventketoacidosis for many years. Many suchindividuals with this form of type 1 dia-betes eventually become dependent oninsulin for survival and are at risk for

ketoacidosis. At this latter stage of the dis-ease, there is little or no insulin secretion, asmanifested by low or undetectable levels ofplasma C-peptide. Immune-mediated dia-betes commonly occurs in childhood andadolescence, but it can occur at any age,even in the 8th and 9th decades of life.

Autoimmune destruction of 3-cells hasmultiple genetic predispositions and is alsorelated to environmental factors that arestill poorly defined. Although patients arerarely obese when they present with thistype of diabetes, the presence of obesity isnot incompatible with the diagnosis. Thesepatients are also prone to other autoim-mune disorders such as Graves' disease,Hashimotos thyroiditis, Addison's disease,vitiligo, and pernicious anemia.Idiopathic diabetes. Some forms of type 1diabetes have no known etiologies. Some ofthese patients have permanent insulinope-nia and are prone to ketoacidosis, but haveno evidence of autoimmunity Althoughonly a minority of patients with type 1 dia-betes fall into this category, of those whodo, most are of African or Asian origin.Individuals with this form of diabetes suf-fer from episodic ketoacidosis and exhibitvarying degrees of insulin deficiencybetween episodes. This form of diabetes isstrongly inherited, lacks immunologicalevidence for (3-cell autoimmunity, and is


Committee Report

not HLA associated. An absolute require-ment for insulin replacement therapy inaffected patients may come and go (17).

Type 2 diabetes (ranging frompredominantly insulin resistancewith relative insulin deficiency topredominantly an insulin secretorydefect with insulin resistance)This form of diabetes, previously referred toas non-insulin-dependent diabetes, type 2diabetes, or adult-onset diabetes, is a termused for individuals who have insulin resis-tance and usually have relative (rather thanabsolute) insulin deficiency (18-21). Atleast initially, and often throughout theirlifetime, these individuals do not needinsulin treatment to survive. There areprobably many different causes of this formof diabetes, and it is likely that the propor-tion of patients in this category willdecrease in the future as identification ofspecific pathogenic processes and geneticdefects permits better differentiation amongthem and a more definitive subclassifica-tion. Although the specific etiologies of thisform of diabetes are not known, autoim-mune destruction of (3-cells does not occurand patients do not have any of the othercauses of diabetes listed above or below.

Most patients with this form of dia-betes are obese, and obesity itself causessome degree of insulin resistance (22,23).Patients who are not obese by traditionalweight criteria may have an increased per-centage of body fat distributed predomi-nantly in the abdominal region (24).Ketoacidosis seldom occurs spontaneouslyin this type of diabetes; when seen, it usu-ally arises in association with the stress ofanother illness such as infection (25-27).This form of diabetes frequently goes undi-agnosed for many years because the hyper-glycemia develops gradually and at earlierstages is often not severe enough for thepatient to notice any of the classic symp-toms of diabetes (28-30). Nevertheless,such patients are at increased risk of devel-oping macrovascular and microvascularcomplications (30-34). Whereas patientswith this form of diabetes may have insulinlevels that appear normal or elevated, thehigher blood glucose levels in these diabeticpatients would be expected to result in evenhigher insulin values had their (3-cell func-tion been normal (35). Thus, insulin secre-tion is defective in these patients andinsufficient to compensate for the insulinresistance. Insulin resistance may improvewith weight reduction and/or pharmaco-

logical treatment of hyperglycemia but isseldom restored to normal (36-40). Therisk of developing this form of diabetesincreases with age, obesity, and lack of phys-ical activity (29,41). It occurs more fre-quently in women with prior GDM and inindividuals with hypertension or dyslipi-demia, and its frequency varies in differentracial/ethnic subgroups (29,30,41). It isoften associated with a strong genetic pre-disposition, more so than is the autoim-mune form of type 1 diabetes (42,43).However, the genetics of this form of dia-betes are complex and not clearly defined.

Other specific types of diabetesGenetic defects of the fJ-cell. Severalforms of diabetes are associated with mono-genetic defects in P-cell function. Theseforms of diabetes are frequently character-ized by onset of mild hyperglycemia at anearly age (generally before age 25 years).They were formerly referred to as maturity-onset diabetes of the young (MODY) andare characterized by impaired insulin secre-tion with minimal or no defects in insulinaction (44-46). They are inherited in anautosomal dominant pattern. Abnormalitiesat three genetic loci on different chromo-somes have been identified to date. Themost common form is associated withmutations on chromosome 12 in a hepatictranscription factor referred to as hepatocytenuclear factor (HNF)-la (47,48). A secondform is associated with mutations in theglucokinase gene on chromosome 7p andresults in a defective glucokinase molecule(49,50). Glucokinase converts glucose toglucose-6-phosphate, the metabolism ofwhich, in turn, stimulates insulin secretionby the P-cell. Thus, glucokinase serves asthe "glucose sensor" for the (3-cell. Becauseof defects in the glucokinase gene, increasedplasma levels of glucose are necessary toelicit normal levels of insulin secretion. Athird form is associated with a mutation inthe HNF-4a gene on chromosome 20q(51,52). HNF-4a is a transcription factorinvolved in the regulation of the expressionof HNF-la. The specific genetic defects ina substantial number of other individualswho have a similar clinical presentation arecurrently unknown.

Point mutations in mitochondrial DNAhave been found to be associated with dia-betes mellitus and deafness (53-55). Themost common mutation occurs at position3243 in the tRNA leucine gene, leading toan A-to-G transition. An identical lesionoccurs in the MELAS syndrome (mito-

chondrial myopathy encephalopathy, lacticacidosis, and stroke-like syndrome); how-ever, diabetes is not part of this syndrome,suggesting different phenotypic expressionsof this genetic lesion (56).

Genetic abnormalities that result in theinability to convert proinsulin to insulinhave been identified in a few families, andsuch traits are inherited in an autosomaldominant pattern (57,58). The resultantglucose intolerance is mild. Similarly, theproduction of mutant insulin moleculeswith resultant impaired receptor bindinghas also been identified in a few familiesand is associated with an autosomal inher-itance and only mildly impaired or evennormal glucose metabolism (59-61).Genetic defects in insulin action. Thereare unusual causes of diabetes that resultfrom genetically determined abnormalitiesof insulin action. The metabolic abnormal-ities associated with mutations of theinsulin receptor may range from hyperin-sulinemia and modest hyperglycemia tosevere diabetes (62,63). Some individualswith these mutations may have acanthosisnigricans. Women may be virilized andhave enlarged, cystic ovaries (64,65). Inthe past, this syndrome was termed type Ainsulin resistance (62). Leprechaunism andthe Rabson-Mendenhall syndrome are twopediatric syndromes that have mutations inthe insulin receptor gene with subsequentalterations in insulin receptor function andextreme insulin resistance (63). The formerhas characteristic facial features and is usu-ally fatal in infancy, while the latter is asso-ciated with abnormalities of teeth and nailsand pineal gland hyperplasia.

Alterations in the structure and func-tion of the insulin receptor cannot bedemonstrated in patients with insulin-resistant lipoatrophic diabetes. Therefore,it is assumed that the lesion(s) must residein the postreceptor signal transductionpathways.Diseases of the exocrine pancreas. Anyprocess that diffusely injures the pancreascan cause diabetes. Acquired processesinclude pancreatitis, trauma, infection, pan-createctomy, and pancreatic carcinoma(66-68). With the exception of cancer,damage to the pancreas must be extensivefor diabetes to occur. However, adenocarci-nomas that involve only a small portion ofthe pancreas have been associated with dia-betes. This implies a mechanism other thansimple reduction in (3-cell mass. If extensiveenough, cystic fibrosis and hemochromato-sis will also damage p-cells and impair


Committee Report

insulin secretion (69,70). Fibrocalculouspancreatopathy may be accompanied byabdominal pain radiating to the back andpancreatic calcifications on X ray (71). Pan-creatic fibrosis and calcium stones in theexocrine ducts have been found at autopsyEndocrinopathies. Several hormones (e.g.,growth hormone, cortisol, glucagon, epi-nephrine) antagonize insulin action. Excessamounts of these hormones (e.g.,acromegaly, Cushing's syndrome, glucagon-oma, pheochromocytoma) can cause dia-betes (72-75). This generally occurs inindividuals with preexisting defects ininsulin secretion, and hyperglycemia typi-cally resolves when the hormone excess isremoved.

Somatostatinoma- and aldosteronoma-induced hypokalemia can cause diabetes, atleast in part, by inhibiting insulin secretion(75,76). Hyperglycemia generally resolvesafter successful removal of the tumor.Drug- or chemical-induced diabetes.Many drugs can impair insulin secretion.These drugs may not cause diabetes bythemselves, but they may precipitate dia-betes in individuals with insulin resistance(77,78). In such cases, the classification isunclear because the sequence or relativeimportance of pi-cell dysfunction andinsulin resistance is unknown. Certain tox-ins such as Vacor (a rat poison) and intra-venous pentamidine can permanentlydestroy pancreatic 3-cells (79-82). Suchdrug reactions fortunately are rare. Thereare also many drugs and hormones that canimpair insulin action. Examples includenicotinic acid and glucocorticoids (77,78).Patients receiving a-interferon have beenreported to develop diabetes associatedwith islet cell antibodies and, in certaininstances, severe insulin deficiency (83,84).The list shown in Table 1 is not all-inclu-sive, but reflects the more commonly rec-ognized drug-, hormone-, or toxin-inducedforms of diabetes.Infections.Certain viruses have been asso-ciated with (3-cell destruction. Diabetesoccurs in patients with congenital rubella(85), although most of these patients haveHLA and immune markers characteristic oftype 1 diabetes. In addition, coxsackievirusB, cytomegalovirus, adeno virus, andmumps have been implicated in inducingcertain cases of the disease (86-88).Uncommon forms of immune-mediateddiabetes. In this category, there are twoknown conditions, and others are likely tooccur. The stiff-man syndrome is anautoimmune disorder of the central ner-

vous system characterized by stiffness ofthe axial muscles with painful spasms (89).Patients usually have high titers of the GADautoantibodies and approximately one-third will develop diabetes.

Anti-insulin receptor antibodies cancause diabetes by binding to the insulinreceptor, thereby blocking the binding ofinsulin to its receptor in target tissues (63).However, in some cases, these antibodiescan act as an insulin agonist after binding tothe receptor and can thereby cause hypo-glycemia. Anti-insulin receptor antibodiesare occasionally found in patients with sys-temic lupus erythematosus and otherautoimmune diseases (63). As in other statesof extreme insulin resistance, patients withanti-insulin receptor antibodies often haveacanthosis nigricans. In the past, this syn-drome was termed type B insulin resistance.Other genetic syndromes sometimesassociated with diabetes. Many geneticsyndromes are accompanied by anincreased incidence of diabetes mellitus(90). These include the chromosomalabnormalities of Down's syndrome, Kline-felter's syndrome, and Turners syndrome.Wolframs syndrome is an autosomal reces-sive disorder characterized by insulin-defi-cient diabetes and the absence of P-cells atautopsy (91). Additional manifestationsinclude diabetes insipidus, hypogonadism,optic atrophy, and neural deafness. Othersyndromes are listed in Table 1.

Gestational diabetes mellitus (GDM)GDM is defined as any degree of glucoseintolerance with onset or first recognitionduring pregnancy. The definition appliesregardless of whether insulin or only dietmodification is used for treatment or ofwhether the condition persists after preg-nancy. It does not exclude the possibilitythat unrecognized glucose intolerance mayhave antedated or begun concomitantlywith the pregnancy (92). Six weeks ormore after pregnancy ends, the womanshould be reclassified, as described below(see diagnostic criteria for diabetes melli-tus), into one of the following categories: 1)diabetes, 2) IFG, 3) IGT, or 4) normo-glycemia. In the majority of cases of GDM,glucose regulation will return to normalafter delivery.

GDM complicates ~4% of all pregnan-cies in the U.S., resulting in ~135,000 casesannually (93). The prevalence may rangefrom 1 to 14% of pregnancies, dependingon the population studied (93). GDM rep-resents nearly 90% of all pregnancies com-

plicated by diabetes (94). Clinical recogni-tion of GDM is important because therapy,including diet, insulin when necessary, andantepartum fetal surveillance, can reducethe well-described GDM-associated perina-tal morbidity and mortality (95). Maternalcomplications related to GDM also includean increased rate of cesarean delivery andchronic hypertension (95-97). Althoughmany patients diagnosed with GDM willnot develop diabetes later in life, others willbe diagnosed many years postpartum ashaving type 1 diabetes, type 2 diabetes,IFG, or IGT (98-103).

Deterioration of glucose toleranceoccurs normally during pregnancy, particu-larly in the 3rd trimester. The criteria forabnormal glucose tolerance in pregnancy,which are widely used in the U.S., wereproposed by O'Sullivan and Mahan (98) in1964 and were based on data obtained fromOGTTs performed on 752 pregnant women.Abnormal glucose tolerance was defined astwo or more blood glucose values out of fourthat were greater than or equal to two stan-dard deviations above the mean. These val-ues were set based on the prediction ofdiabetes developing later in life.

In 1979, the NDDG revised the O'Sul-livan and Mahan criteria, converting thewhole blood values to plasma values (Table2) (1). These criteria have been adopted bythe American Diabetes Association and theAmerican College of Obstetricians andGynecologists (ACOG) (104), but are atvariance with WHO criteria. The diagnosisof GDM is made if any two out of fourthreshold values are met or exceeded.Testing for gestational diabetes. Previ-ous recommendations have been thatscreening for GDM be performed in all preg-nancies. However, there are certain factorsthat place women at lower risk for the devel-opment of glucose intolerance during preg-nancy, and it is likely not cost-effective toscreen such patients. This low-risk groupcomprises women who are <25 years of ageand of normal body weight, have no familyhistory (i.e., first-degree relative) of diabetes,and are. not members of an ethnic/racialgroup with a high prevalence of diabetes(e.g., Hispanic, Native American, Asian,African-American) (105-107). Pregnantwomen who fulfill all of these criteria neednot be screened for GDM.

Most American obstetricians use ascreening test consisting of a 50-g oral glu-cose load followed by a plasma glucosedetermination 1 h later (92). Screeningshould be performed (unless otherwise


Committee Report

Table 2—Screening and diagnosis schemefor GDM

Plasma 50-g 100-gglucose screening test diagnostic test

Fasting1-h2-h3-h

140 mg/dl105 mg/dl190 mg/dl165 mg/dl145 mg/dl

Screening for GDM may not be necessary in preg-nant women who meet all of the following criteria:<25 years of age, normal body weight, no first-degree relative with diabetes, and not Hispanic,Native American, Asian-, or African-American. The100-g diagnostic test is performed on patients whohave a positive screening test. The diagnosis ofGDM requires any two of the four plasma glucosevalues obtained during the test to meet or exceedthe values shown above.

indicated) between 24 and 28 weeks of ges-tation, and the patient need not be fasting.A value of >140 mg/dl 1 h after the 50-gload indicates the need for a full diagnostic,100-g, 3-h OGTT performed in the fastingstate (Table 2). This two-step process—a50-g screening test and, if that is positive, a100-g diagnostic test—is the testing schemein current use in the U.S. The ACOG (104)recommends that in certain populations,the prevalence of GDM is so high that preg-nant women can be considered to have apositive screen and should proceed directlyto diagnostic testing.

There have been challenges to the diag-nostic scheme shown in Table 2. Carpenterand Coustan (108) have suggested that theNDDG conversion of the O'Sullivan andMahan values from the original Somogyi-Nelson determinations may have resultedin values that are too high. They proposedcutoff values for plasma glucose that appearto represent more accurately the originalO'Sullivan and Mahan determinations. Inthree studies, these criteria identified morepatients with GDM whose infants had peri-natal morbidity (109-111). In addition,efforts are being made to establish interna-tionally agreed-upon diagnostic testing pro-cedures; for example, using the 75-g oralglucose load, as recommended by theWHO, because that protocol identifies agreater number of pregnancies with mater-nal or perinatal complications associatedwith high plasma glucose (112-114).However, it is considered premature to rec-ommend any change from the current reg-imen because it is so widely accepted andpracticed in the U.S.

Impaired glucose tolerance (IGT)and impaired fasting glucose (IFG)The terms IGT and IFG refer to a metabolicstage intermediate between normal glucosehomeostasis and diabetes. This stageincludes individuals who have IGT andindividuals with fasting glucose levels>110 mg/dl but <126 mg/dl (IFG). Theterm IFG was coined by Charles et al. (115)to refer to a fasting plasma glucose (FPG)level >110 mg/dl (6.1 mmol/1) but <140mg/dl (7.8 mmol/1). We are using a similardefinition, but with the upper end loweredto correspond to the new diagnostic crite-ria for diabetes. A fasting glucose concen-tration of 110 mg/dl has been chosen as theupper limit of "normal." Although it is rec-ognized that this choice is somewhat arbi-trary, it is near the level above which acutephase insulin secretion is lost in response tointravenous administration of glucose(116) and is associated with a progressivelygreater risk of developing micro- andmacrovascular complications (117-121).

Note that many individuals with IGTare euglycemic in their daily lives (122) andmay have normal or near normal glycatedhemoglobin levels (123). Individuals withIGT often manifest hyperglycemia onlywhen challenged with the oral glucose loadused in the standardized OGTT.

In the absence of pregnancy, IFG andIGT are not clinical entities in their own rightbut rather risk factors for future diabetes andcardiovascular disease (117). They can beobserved as intermediate stages in any of thedisease processes listed in Table 1. IFG andIGT are associated with the insulin resistancesyndrome (also known as syndrome X or themetabolic syndrome), which consists ofinsulin resistance, compensatory hyperin-sulinemia to maintain glucose homeostasis,obesity (especially abdominal or visceralobesity), dyslipidemia of the high-triglyc-eride and/or low-HDL type, and hyperten-sion (124). Insulin resistance is directlyinvolved in the pathogenesis of type 2 dia-betes. IFG and IGT appear as risk factors forthis type of diabetes at least in part becauseof their correlation with insulin resistance. Incontrast, the explanation for why IFG andIGT are also risk factors for cardiovasculardisease is less clear. The insulin resistancesyndrome includes well-recognized cardio-vascular risk factors such as low HDL levelsand hypertension. In addition, it includeshypertriglyceridemia, which is highly cor-related with small dense LDL and increasedplasminogen activator inhibitor-1 (PAI-1)levels. The former is thought to have

enhanced atherogenicity, perhaps as a resultof its greater vulnerability to oxidation thannormal LDL. PAI-1 is a cardiovascular riskfactor probably because it inhibits fibri-noloysis. Thus, the insulin resistance syn-drome contains many features that increasecardiovascular risk. IFG and IGT may not inthemselves be directly involved in the patho-genesis of cardiovascular disease, but rathermay serve as statistical risk factors by asso-ciation because they correlate with thoseelements of the insulin resistance syndromethat are cardiovascular risk factors.

DIAGNOSTIC CRITERIA FORDIABETES MELLITUS

The new criteriaThe diagnostic criteria for diabetes mellitushave been modified from those previouslyrecommended by the NDDG (1) or WHO(2). The revised criteria for the diagnosis ofdiabetes are shown in Table 3. Three waysto diagnose diabetes are possible, and eachmust be confirmed, on a subsequent day, byany one of the three methods given in Table3. For example, one instance of symptomswith casual plasma glucose ^200 mg/dl,confirmed on a subsequent day by 1) FPG>126 mg/dl, 2) an OGTT with the 2-hpostload value ^200 mg/dl, or 3) symp-toms with a casual plasma glucose ^200mg/dl, warrants the diagnosis of diabetes.

For epidemiological studies, estimatesof diabetes prevalence and incidence shouldbe based on an FPG ^126 mg/dl. This rec-ommendation is made in the interest ofstandardization and also to facilitate fieldwork, particularly where the OGTT may bedifficult to perform and where the cost anddemands on participants' time may beexcessive. This approach will lead to slightlylower estimates of prevalence than wouldbe obtained from the combined use of theFPG and OGTT (Table 4).

The Expert Committee recognizes anintermediate group of subjects whose glu-cose levels, although not meeting criteriafor diabetes, are nevertheless too high to beconsidered altogether normal. This group,IFG, is defined as having FPG levels ^ 110mg/dl but < 126 mg/dl or 2-h values in theOGTT of >140 mg/dl but <200 mg/dl.Thus, the categories of FPG values are asfollows:

• FPG <110 mg/dl = normal fasting glu-cose;

• FPG >110 and <126 mg/dl = IFG;


Committee Report

Table 3—Criteria for the diagnosis of diabetes mellitus

Symptoms of diabetes plus casual plasma glucose concentration >200 mg/dl (11.1 mmol/1).Casual is defined as any time of day without regard to time since last meal. The classicsymptoms of diabetes include polyuria, polydipsia, and unexplained weight loss.

or2. FPG ^126 mg/dl (7.0 mmol/1). Fasting is defined as no caloric intake for at least 8 h.

or3. 2hPG ^200 mg/dl during an OGTT. The test should be performed as described by WHO

(2), using a glucose load containing the equivalent of 75-g anhydrous glucose dissolved inwater.

In the absence of unequivocal hyperglycemia with acute metabolic decompensation, these criteria shouldbe confirmed by repeat testing on a different day. The third measure (OGTT) is not recommended for rou-tine clinical use.

• FPG 2:126 mg/dl = provisional diagno-sis of diabetes (the diagnosis must beconfirmed, as described above).

The corresponding categories when theOGTT is used are the following:

• 2-h postload glucose (2hPG) <140mg/dl = normal glucose tolerance;

• 2hPG >140 and <200 mg/dl = IGT;• 2hPG >200 mg/dl = provisional diag-

nosis of diabetes (the diagnosis must beconfirmed, as described above).

Since the 2-h OGTT cutoff of 140mg/dl will identify more people as havingimpaired glucose homeostasis than will thefasting cutoff of 110 mg/dl, it is essentialthat investigators always report which testwas used.

Rationale for the revised criteria fordiagnosing diabetesThe revised criteria are still based on meas-ures of hyperglycemia. Whereas many dif-ferent diagnostic schemes have been used,all have been based on some measurementof blood or urine glucose, as reviewed byMcCance et al. (125). The metabolicdefects underlying hyperglycemia, such asislet cell autoimmunity or insulin resis-tance, should be referred to independentlyfrom the diagnosis of diabetes, i.e., in theclassification of the disease. Determiningthe optimal diagnostic level of hyper-glycemia depends on a balance betweenthe medical, social, and economic costs ofmaking a diagnosis in someone who is nottruly at substantial risk of the adverseeffects of diabetes and those of failing todiagnose someone who is (126). Unfortu-nately, not all these data are available, so werelied primarily on medical data.

Plasma glucose concentrations are dis-tributed over a continuum, but there is anapproximate threshold separating those sub-jects who are at substantially increased riskfor some adverse outcomes caused by dia-betes (e.g., microvascular complications)from those who are not. Based in part onestimates of the thresholds for microvasculardisease, the previous WHO criteria defineddiabetes by FPG >140 mg/dl, 2hPG >200mg/dl in the OGTT, or both. These criteriaeffectively defined diabetes by the 2hPGalone because the fasting and 2-h cutpointvalues are not equivalent. Almost all indi-viduals with FPG >140 mg/dl have 2hPG2:200 mg/dl if given an OGTT, whereasonly about one-fourth of those with 2hPG2:200 mg/dl and without previously knowndiabetes have FPG >140 mg/dl (127). Thus,the cutpoint of FPG 2:140 mg/dl defined agreater degree of hyperglycemia than did thecutpoint of 2hPG 2:200 mg/dl. It is the con-sensus of the Expert Committee that this dis-crepancy is unwarranted and that thecutpoint values for both tests should reflecta similar degree of hyperglycemia and risk ofadverse outcomes.

Under the previous WHO and theNDDG criteria, the diagnosis of diabetes islargely a function of which test is per-formed. Many individuals who would have2hPG >200 mg/dl in an OGTT are nottested with an OGTT because they lacksymptoms or because they have an FPG<140 mg/dl. Thus, if it is desired that allpeople with diabetes be diagnosed and theprevious criteria are followed, OGTTs mustbe performed periodically in everyone.However, in ordinary practice, not only isthe OGTT performed infrequently, but it isusually not used even to confirm suspectedcases (128). In summary, the diagnosticcriteria are now revised to 1) avoid the dis-crepancy between the FPG and 2hPG cut-point values and 2) facilitate and encouragethe use of a simpler and equally accuratetest—fasting plasma glucose—for diagnos-ing diabetes.

The cutpoint for the 2hPG has beenjustified largely because at approximatelythat point the prevalence of the microvas-cular complications considered specific fordiabetes (i.e., retinopathy and nephropa-thy) increases dramatically. This property ofthe 2hPG has been compared with the FPGin population studies of the Pima Indians inthe U.S., among Egyptians, and in theThird National Health and Nutrition Exam-ination Survey (NHANES III) in the U.S. Inother studies, the relationships betweenglycemia and macrovascular disease havealso been examined.

The relationships of FPG and 2hPG tothe development of retinopathy were eval-uated in Pima Indians over a wide range ofplasma glucose cutpoints (Fig. 2A) (129).Both variables were similarly associatedwith retinopathy, indicating that by thiscriterion, each could work equally well fordiagnosing diabetes. The authors con-cluded that both measures were equivalent

Table 4—Estimated prevalence of diabetes in the U.S. in individuals 40-74 years old usingdata from the NHANES III

Diabetes diagnostic criteria

Prevalence (%) of diabetes byglucose criteria without a

medical history of diabetes*Total diabetes

prevalence (%)t

Medical history of diabetesWHO (2) criteria for diabetes:

FPG >140 mg/dl (7.8 mmol/1) or2hPG >200 mg/dl (11.1 mmol/1)

FPG >126 mg/dl (7.0 mmol/1)

6.34

4.35

7.92

14.26

12.27Data are from K. Flegal, National Center for Health Statistics, personal communication. *Diabetes preva-lence (by glucose criteria) in those without a medical history of diabetes X (100% — prevalence of diabetesby medical history). TFirst column of data plus 7.92.


Committee Report

in terms of the properties previously usedto justify diagnostic criteria.

These findings were confirmed in a sim-ilar study in Egypt, in which the FPG and2hPG were each strongly and equally asso-ciated with retinopathy (Fig. IB) (130). Forboth the FPG and the 2hPG, the prevalenceof retinopathy was markedly higher abovethe point of intersection of the two compo-nents of the bimodal frequency distribution(FPG = 129 mg/dl and 2hPG = 207 mg/dl).

In the NHANES III, 2,821 individualsaged 40-74 years received an OGTT, ameasurement of HbAlo and an assessmentof retinopathy by fundus photography (K.Flegal, personal communication). Figure1C shows that all three measures ofglycemia (FPG, 2hPG, and HbAlc) arestrongly associated with retinopathy, whichis similar to the relationships found in thePima Indians (129) and Egyptians (130),although the relationship was strongest for2hPG. As in other studies, the prevalencerose dramatically in the highest decile ofeach variable, corresponding to FPG ^120mg/dl, 2hPG >195 mg/dl, and HbAlc

>6.2%. As in the Pima Indian (129) andEgyptian (130) studies, estimates of these"thresholds" for retinopathy are somewhatimprecise. More precision cannot easily beobtained by using narrower glycemic inter-vals (e.g., 20 instead of the 10 shown inFig. 2) because of the limited numbers ofcases of retinopathy in each sample (32cases in the Pima study, 146 in the Egypt-ian study, and 111 in NHANES III). Thereare no absolute thresholds because someretinopathy occurred at all glucose levels,presumably because of measurement ordisease variability and because of nondia-betic causes of retinopathy.

The associations between FPG and2hPG and macrovascular disease have beenexamined in adults without known dia-betes (131). The 2hPG was somewhat moreclosely associated with major coronary heartdisease, but there was no significant differ-ence in the association of the FPG or the2hPG with other indexes of macrovasculardisease. Similarly, the relationship betweenglycemia and peripheral arterial disease wasstudied in 50- to 74-year-old Caucasians(132). The prevalence of arterial diseasewas strongly related to the FPG and 2hPG.The associations appeared to be of the samestrength for both variables.

In a recent analysis of the Paris Prospec-tive Study, the incidence of fatal coronaryheart disease was related to both FPG and2hPG determined at a baseline examination

15-

%Q.O

5-

•• 2hPG

• HbAlc

FPG (mg/dl) 70- 89- 93- 97- 100- 105- 109- 116- 136- 226-

2hPG (mg/dl) 38- 94- 106- 116- 126- 138- 156- 185- 244- 364-

HbA1C (%) 3.4- 4.8- 5.0- 5.2- 5.3- 5.5- 5.7- 6.0- 6.7- 9.5-

FPG (mg/dl) 57- 79- 84- 89- 93- 99- 108-130-178-258-

2hPG (mg/dl) 39- 80- 90- 99- 110- 125- 155- 218- 304-386-

HbA1C (%) 2.2- 4.7- 4.9- 5.1- 5.4- 5.6- 6.0- 6.9- 8.5- 10.3-

15-

~ 10Q.Oc

5-

FPG (mg/dl) 42- 87- 90- 93- 96- 98- 101- 104- 109- 120-

2hPG (mg/dl) 34- 75- 86- 94- 102- 112- 120- 133- 154- 195-

HbA1C (%) 3.3- 4.9- 5.1- 5.2- 5.4- 5.5- 5.6- 5.7- 5.9- 6.2-

Figure 2—Prevalence of retinopathy by deciles of the distribution of FPG, 2hPG, and HbAlc in PimaIndians (A) described in McCance et al. (129), Egyptians (B) described in Engelgau et al. (130), andin 40- to 74-year-old participants in NHANES III (C) (K. Flegal, National Center for Health Statis-tics, personal communication). The x-axis labels indicate the lower limit of each decile group. Note thatthese deciles and the prevalence rates of retinopathy differ considerably among the studies, especiallythe Egyptian study, in which diabetic subjects were oversampled. Retinopathy was ascertained by dif-ferent methods in each study; therefore, the absolute prevalence rates are not comparable between stud-ies, but their relationships with FPG, 2hPG, and HbAlc are very similar within each population.


Committee Report

Table 5—FPG cutpoints equivalent to the WHO 2-h plasma glucose criterion of 200 mg/dl

Study and reference Method Fasting plasma glucose4

Pima Indians (129)Pima Indians (129)Several Pacific populations (134)NHANES III§

ROC curves!Equal prevalence!Equal prevalence!Equal prevalence!

123 mg/dl (6.8 mmol/1)120 mg/dl (6.7 mmol/1)126 mg/dl (7.0 mmol/1)121 mg/dl (6.7 mmol/1)

*The results for the receiver operating characteristics (ROC) curve analysis of the Pima Indian data and thosefrom the Pacific populations appear in the cited publications (in millimoles per liter). The other results havenot been published. ^Equivalent to the WHO criterion of 2hPG S200 mg/dl in sensitivity and specificityfor retinopathy from analysis of ROC curves. !The method is described by Finch et al. (134). §NHANESIII subjects ages 40-74 years, excluding users of insulin and oral hypoglycemic agents, weighted accord-ing to sampling plan (K. Flegal, National Center for Health Statistics, personal communication).

(118). Incidence rates were markedlyincreased at FPG ^6.9 mmol/1 (125 mg/dl)or 2hPG >7.8 mmol/1 (140 mg/dl). Simi-larly, the incidence of coronary artery dis-ease and the all-cause mortality rates werepredicted by the FPG in the Baltimore Lon-gitudinal Study of Aging (R. Andres, C.Coleman, D. Elahi, J. Fleg, D.C. Muller,J.D. Sorkin, J.D. Tobin, personal communi-cation). The incidence rates of both theseoutcomes increased markedly and almostlinearly above FPG levels in the range of110-120 mg/dl. In conclusion, both theFPG and 2hPG provide important informa-tion regarding risk of both micro- andmacro vascular disease, and the approximatethresholds for increased risk correspondwith those for retinopathy and with therevised diagnostic criteria.

Reproducibility is another importantproperty of a diagnostic test, a property forwhich the FPG appears to be preferable.When OGTTs were repeated in adults dur-ing a 2- to 6-week interval, the intra-indi-vidual coefficients of variation were 6.4%for the FPG and 16.7% for the 2hPG (133).

It is important to review the rationalefor retaining the diagnostic cutpoint of 200mg/dl for the 2hPG. This cutpoint wasoriginally adopted for three reasons (1,2).First, 200 mg/dl has been found to approx-imate the cutpoint separating the two com-ponents of the bimodal distribution of2hPG. Second, in several studies, theprevalence of microvascular disease sharplyincreased above 2hPG levels of ~200mg/dl. Third, an enormous body of clinicaland epidemiological data has been col-lected based on the 2hPG cutpoint of 200mg/dl. Thus, this value has been retainedfor the diagnosis of diabetes because itwould be very disruptive, and add littlebenefit, to alter the well-accepted 2hPGdiagnostic level of >200 mg/dl.

Changing the diagnostic cutpoint forthe FPG to 126 mg/dl is based on the beliefthat the cutpoints for the FPG and 2hPGshould diagnose similar conditions, giventhe equivalence of the FPG and the 2hPGin their associations with vascular compli-cations and their discrimination betweentwo components of a bimodal frequencydistribution (129,130). McCance et al.(129) computed the FPG level equivalent(in sensitivity and specificity for retinopa-thy) to the 1985 WHO criterion of the2hPG >200 mg/dl and found it to be anFPG of > 123 mg/dl (Table 5). Finch et al.(134) approached the problem in each of13 Pacific populations surveyed withOGTTs by determining the value in theFPG that, when used alone as a diagnosticcriterion, gave the same prevalence of dia-betes as did 2hPG >200 mg/dl. The sum-mary estimate from all these populationswas a cutpoint of 126 mg/dl. The samemethod was applied to data derived fromthe Pima Indians and resulted in an FPGcutpoint of 120 mg/dl. In NHANES III, thecorresponding cutpoint was 121 mg/dl(Table 5). These values and the 2hPG cut-point of 200 mg/dl are also quite similar tothe values of FPG 129 mg/dl and 2hPG 207mg/dl that separated the components of thebimodal frequency distributions and iden-tified individuals with a high prevalence ofretinopathy among Egyptians (130).Because the standard errors of these esti-mates are not known, the small differencesin the estimates shown in Table 5 may beconsistent with sampling variability.

We chose a cutpoint at the upper endof these estimates (FPG >126 mg/dl, 7.0mmol/1). This value is slightly higher thanmost of the estimated cutpoints that wouldgive the same prevalence of diabetes as thecriterion of 2hPG >200 mg/dl. That is,slightly fewer people will be diagnosed

with diabetes if the new FPG criterion isused alone than if either the FPG or theOGTT is used and interpreted by the pre-vious WHO and NDDG criteria (Table 4).

As noted above, although the OGTT isan acceptable diagnostic test and has beenan invaluable tool in research, it is not rec-ommended for routine use. Because of itsinconvenience to patients and the percep-tion by many physicians that it is unneces-sary, the OGTT is already not widely usedfor diagnosing diabetes. In addition, it ismore costly and time-consuming than theFPG, and the repeat test reproducibility ofthe 2hPG is worse than that of the FPG(133). If the OGTT is used, either for clin-ical or research purposes, the test proce-dure methods recommended by the WHO(2) and the diagnostic criterion in Table 3should be employed.

HbAlc measurement is not currentlyrecommended for diagnosis of diabetes,although some studies have shown thatthe frequency distributions for HbAlc havecharacteristics similar to those of the FPGand the 2hPG. Moreover, these studieshave defined an HbAlc level above whichthe likelihood of having or developingmacro- or microvascular disease risessharply (Fig. 2) (129-132). Furthermore,HbAlc and FPG (in type 2 diabetes) havebecome the measurements of choice inmonitoring the treatment of diabetes, anddecisions on when and how to implementtherapy are often made on the basis ofHbAlc. These observations have led someto recommend HbAlc measurement as adiagnostic test (126,135).

On the other hand, there are many dif-ferent methods for the measurement ofHbAlc and other glycosylated proteins, andnationwide standardization of the HbAlc

test has just begun (136). Studies of theutility of the test compared with the FPGand 2hPG have used different assays,thereby making it difficult to assign anappropriate cutpoint. Also, the FPG, 2hPG,and HbA]c tests are imperfectly correlated.In most clinical laboratories, a "normal"HbAlc is usually based on a statistical sam-pling of healthy, presumably nondiabeticindividuals. In conclusion, HbAlc remainsa valuable tool for monitoring glycemia,but it is not currently recommended for thediagnosis of diabetes.

The revised criteria are for diagnosisand are not treatment criteria or goals oftherapy. No change is made in the Ameri-can Diabetes Association's recommenda-tions of FPG <120 mg/dl and HbAlc <7%


Committee Report

as treatment goals (137). The new diag-nostic cutpoint (FPG ^126 mg/dl) is basedon the observation that this degree ofhyperglycemia usually reflects a seriousmetabolic abnormality that has been shownto be associated with serious complica-tions. The treatment of nonpregnantpatients with hyperglycemia near the cut-point should begin with an individualizedlifestyle-modification regimen (i.e., mealplanning and exercise). Initiation of phar-macological therapy in these patients hasnot yet been shown to improve prognosisand may lead to an unacceptably high inci-dence of hypoglycemic reactions with cer-tain drugs (e.g., sulfonylureas, insulin).

The new criteria have implications forestimates of the prevalence of diabetes.Although an FPG > 126 mg/dl and a 2hPG2:200 mg/dl have similar predictive valuefor adverse outcomes, the two tests are notperfectly correlated with each other. Agiven person may have one glucose valueabove one cutpoint and another valuebelow the other cutpoint. Thus, simultane-ous measurement of both FPG and 2hPGwill inevitably lead to some diagnostic dis-crepancies and dilemmas. Although diag-nosing diabetes by either test will result ina similar number of "cases," different indi-viduals in different hyperglycemic stagesmay be identified. (This situation would beeven more complicated if a third diagnos-tic test, such as HbAlc, were used.) How-ever, according to the data reviewed above,there is no basis for concluding that the2hPG is more reliable than the FPG. Thus,the FPG alone should be used for estimat-ing the comparative prevalence of diabetesin different populations.

Table 4 shows the effect of the newdiagnostic criteria on the estimated preva-lence of diabetes in the U.S. populationaged 40-74 years using data fromNFIANES III. Diagnosing diabetes in thosewithout a medical history of diabetes byusing only the FPG test would result in alower prevalence of diabetes than wouldusing WHO criteria (4.35 vs. 6.34%). Thetotal prevalence of diabetes (includingthose with a medical history) would be12.27%, or 14% lower than the prevalenceof 14.26% by the WHO criteria. Of note,these prevalence estimates refer to results oftesting on one occasion. The prevalence ofdiabetes confirmed by a second test will belower regardless of which criteria are used.

Widespread adoption of the new crite-ria may, however, have a large impact onthe number of people actually diagnosed

Table 6—Criteria for testing for diabetes in asymptomatic, undiagnosed individuals

1. Testing for diabetes should be considered in all individuals at age 45 years and above and,if normal, it should be repeated at 3-year intervals.

2. Testing should be considered at a younger age or be carried out more frequently in indi-viduals who:• are obese (>120% desirable body weight or a BMI ^27 kg/m2)• have a first-degree relative with diabetes• are members of a high-risk ethnic population (e.g., African-American, Hispanic, Native

American)• have delivered a baby weighing >9 lb or have been diagnosed with GDM• are hypertensive (^140/90)• have an HDL cholesterol level ^35 mg/dl and/or a triglyceride level ^250 mg/dl• on previous testing, had IGT or IFG

The OGTT or FPG test may be used to diagnose diabetes; however, in clinical settings the FPG test is greatlypreferred because of ease of administration, convenience, acceptability to patients, and lower cost.

with diabetes. Presently, about half theadults with diabetes in the U.S. are undi-agnosed (127), but many might now bediagnosed if the simpler FPG test werealways used.

TESTING FOR DIABETES INPRESUMABLY HEALTHYINDIVIDUALS — Type 1 diabetes isusually an autoimmune disease, character-ized by the presence of a variety of autoan-tibodies to protein epitopes on the surfaceof or within the (3-cells of the pancreas.The presence of such markers before thedevelopment of overt disease can identifypatients at risk (138). For example, thosewith more than one autoantibody (i.e., ICA,IAA, GAD, 1A-2) are at high risk (139-141).At this time, however, many reasons pre-clude the recommendation to test individu-als routinely for the presence of any of theimmune markers outside of a clinical trialssetting. First, cutoff values for some of theassays for immune markers have not beencompletely established for clinical settings.Second, there is no consensus yet as to whataction should be taken when a positiveautoantibody test is obtained. Thus, autoan-tibody testing may identify people at risk ofdeveloping type 1 diabetes without offeringany proven measures that might prevent ordelay the clinical onset of disease. Of note,however, is that there are a number of ongo-ing well-controlled clinical studies testingvarious means of preventing type 1 dia-betes. These studies conducted in high-risksubjects may one day offer an effectivemeans to prevent type 1 diabetes, in whichcase screening may become appropriate.Last, because the incidence of type 1 dia-

betes is low, routine testing of healthy chil-dren will identify only the small number(<0.5%) who at that moment may be "pre-diabetic." Thus, the cost-effectiveness ofsuch screening is questionable, at least untilan effective therapy is available. For theabove reasons, the clinical testing of indi-viduals for autoantibodies related to type 1diabetes, outside of research studies, cannotbe recommended at this time. Similarly,antibody testing of high-risk individuals(e.g., siblings of type 1 patients) is also notrecommended until the efficacy and safetyof therapies to prevent or delay type 1 dia-betes have been demonstrated. On theother hand, the autoantibody tests may beof value to identify which newly diagnosedpatients have immune-mediated type 1 dia-betes in circumstances where it is not obvi-ous, particularly when therapies becomeavailable to preserve (3-cell mass.

Undiagnosed type 2 diabetes is com-mon in the U.S. As many as 50% of the peo-ple with the disease, or about 8 millionindividuals, are undiagnosed (127). Of con-cern, there is epidemiological evidence thatretinopathy begins to develop at least 7 yearsbefore the clinical diagnosis of type 2 dia-betes is made (142). Because hyperglycemiain type 2 diabetes causes microvascular dis-ease and may cause or contribute tomacro vascular disease, undiagnosed dia-betes is a serious condition. Patients withundiagnosed type 2 diabetes are at signifi-cantly increased risk for coronary heart dis-ease, stroke, and peripheral vascular disease.In addition, they have a greater likelihood ofhaving dyslipidemia, hypertension, and obe-sity (143).

Thus, early detection, and consequentlyearly treatment, might well reduce the bur-


Committee Report

den of type 2 diabetes and its complications.However, to increase the cost-effectivenessof testing undiagnosed, otherwise healthyindividuals, testing should be considered inhigh-risk populations. Suggested criteria fortesting are given in Table 6. Factors leadingto these recommendations include: 1) thesteep rise in the incidence of the disease afterage 45 years, 2) the negligible likelihood ofdeveloping any of the complications of dia-betes within a 3-year interval of a negativescreening test, and 3) knowledge of the well-documented risk factors for the disease.Although the OGTT and FPG are both suit-able tests, in clinical settings, the FPG isstrongly recommended because it is easierand faster to perform, more convenient andacceptable to patients, more reproducible,and less expensive.

MEMBERS OF THE EXPERTCOMMITTEE ON THEDIAGNOSIS ANDCLASSIFICATION OF DIABETESMELLITUS— James R. Gavin III, MD,PhD (Chair), Howard Hughes MedicalCenter, Bethesda, MD; K.G.M.M. Alberti,MD, University of Newcastle, Newcastle,U.K.; Mayer B. Davidson, MD, Universityof California, Los Angeles, CA; Ralph A.DeFronzo, MD, University of Texas, SanAntonio, TX; Allan Drash, MD, Universityof Pittsburgh, Pittsburgh, PA; Steven G.Gabbe, MD, University of Washington,Seattle, WA; Saul Genuth, MD, Case West-ern Reserve University, Cleveland, OH;Maureen I. Harris, PhD, MPH, NationalInstitutes of Health, Bethesda, MD; RichardKahn, PhD, American Diabetes Associa-tion, Alexandria, VA; Harry Keen, MD,FRCP Guys Hospital, London, U.K.;William C. Knowler, MD, DrPH, NationalInstitutes of Health, Phoenix, AZ; HaroldLebovitz, MD, State University of NewYork, Brooklyn, NY; Noel K. Maclaren,MD, University of Florida, Gainesville, FL;Jerry R Palmer, MD, University of Wash-ington, Seattle, WA; Philip Raskin, MD,University of Texas, Dallas, TX; Robert A.Rizza, MD, Mayo Clinic, Rochester, MN;Michael R Stern, MD, University of Texas,San Antonio, TX.

Acknowledgments — We gratefully acknowl-edge the invaluable assistance of Robert Misbin,MD, in the development of the manuscript;Katherine Flegal, PhD, for her analysis of theNHANES III data set; Reubin Andres, MD, forsharing unpublished data from the Baltimore

Longitudinal Study of Aging; and MichaelEngelgau, MD, for providing the raw data fromthe Egyptian Study (130).

References

1. National Diabetes Data Group: Classifica-tion and diagnosis of diabetes mellitusand other categories of glucose intoler-ance. Diabetes 28:1039-1057, 1979

2. World Health Organization: Diabetes Mel-litus: Report of a WHO Study Group.Geneva, World Health Org., 1985 (Tech.Rep. Ser.,no. 727)

3. Hoet JJ, Tripathy BB, Rao RH, Yajnik CS:Malnutrition and diabetes in the tropics.Diabetes Care 19:1014-1017, 1996

4. Atkinson MA, Maclaren NK: The patho-genesis of insulin dependent diabetes. NEnglJ Med 331:1428-1436, 1994

5. Baekkeskov S, NeilsenJH, Mamer B, BildeT, Ludvigsson J, Lernmark A: Autoanti-bodies in newly diagnosed diabetic chil-dren with immunoprecipitate humanpancreatic islet cell proteins. Nature298:167-169, 1982

6. Atkinson MA, Maclaren NK, Riley WJ,Winter WE, Fisk DD, Spillar RP: Areinsulin autoantibodies markers forinsulin-dependent mellitus? Diabetes 35:894-898, 1986

7. Kaufman D, Erlander M, Clare-Salzler M,Atkinson M, Maclaren N, Tobin A:Autoimmunity to two forms of glutamatedecarboxylase in insulin-dependent mel-litus. J Clxn Invest 89:283-292, 1992

8. Christie MR, Tun RY, Lo SSS, Cassidy D,Brown TJ, Hollands J, Shattock M, Bot-tazzo GF, Leslie RDG: Antibodies to GADand tryptic fragments of islet 64K antigenas distinct markers for development ofIDDM: studies with identical twins. Dia-betes 41:782-787, 1992

9. Schott M, Schatz D, Atkinson M, KrischerJ, Mehta H, Void B, Maclaren N: GAD65

autoantibodies increase the predictabilitybut not the sensitivity of islet cell andinsulin autoantibodies for developinginsulin dependent diabetes mellitus. JAutoimmunity 7:865-872, 1994

10. Schmidli RS, Colman PG, Harrison LC:Do glutamic acid decarboxylase antibod-ies improve the prediction of IDDM infirst-degree relatives at risk for IDDM? JAutoimmunity 7:873-879, 1994

11. Myers MA, Rabin DU, Rowley MJ: Pancre-atic islet cell cytoplasmic antibody in dia-betes is represented by antibodies to isletcell antigen 512 and glutamic acid decar-boxylase. Diabetes 44:1290-1295, 1995

12. Lan MS, Wasserfall C, Maclaren NK,Notkins AL: IA-2, a transmembrane pro-tein of the protein tyrosine phosphatasefamily, is a major autoantigen in insulin-dependent diabetes mellitus. Proc Natl

Acad Sci USA 93:6367-6370, 199613. Lu J, Li Q, Xie H, Chen Z, Borovitskaya

AE, Maclaren NK, Notkins AL, Lan MS:Identification of a second transmembraneprotein tyrosine phosphatase, IA-2(3, as anautoantigen in insulin-dependent diabetesmellitus: precursor of the 37-kDa trypticfragment. Proc Natl Acad Sci USA 93:2307-2311, 1996

14. Cantor AB, Krischer JP, Cuthbertson DD,Schatz DA, Riley WJ, Malone J, SchwartzS, Quattrin T, Maclaren NK: Age and fam-ily relationship accentuate the risk ofIDDM in relatives of patients with insulindependent diabetes. J Clin EndocrinolMetab 80:3739-3743, 1995

15. Huang W, Connor E, DelaRosa T, Muir A,Schatz D, Silverstein J, Crockett S, She JX,Maclaren NK: Although DR3-DQB1* maybe associated with multiple componentdiseases of the autoimmune polyglandularsyndromes, the human leukocyte antigenDR4-DQB110302 haplotype is implicatedonly in beta cell autoimmunity J ClinEndocrinol Metab 81:1-5, 1996

16. Zimmet PZ, Tuomi T, Mackay R, RowleyMJ, Knowles W, Cohen M, Lang DA:Latent autoimmune diabetes mellitus inadults (LADA): the role of antibodies toglutamic acid decarboxylase in diagnosisand prediction of insulin dependency.DiabetMed 11:299-303, 1994

17. Banerji M, Lebovitz H: Insulin sensitiveand insulin resistant variants in IDDM.Diabetes 38:784-792,1989

18. Reaven GM, Bernstein R, Davis B, OlefskyJM: Nonketotic diabetes mellitus: insulindeficiency or insulin resistance? Am] Med60:80-88, 1976

19. Olefsky JM, Kolterman OG, Scarlett, JA:Insulin action and resistance in obesityand noninsulin-dependent type II diabetesmellitus. AmJPhysiol 243:E15-E30,1982

20. DeFronzo R, Deibert D, Hendler R, FeligP: Insulin sensitivity and insulin bindingto monocytes in maturity-onset diabetes. JClin Invest 63:939-946,1979

21. Turner RC, Holman, RR, Matthews D,Hockaday TDR, Peto J: Insulin deficiencyand insulin resistance interaction in dia-betes: estimation of their relative contri-bution by feedback analysis from basalplasma insulin and glucose concentra-tions. Metabolism 28:1086-1096,1979

22. Kolterman OG, Gray RS, Griffin J,Burstein P, Insel J, Scarlett JA, Olefsky JM:Receptor and postreceptor defects con-tribute to the insulin resistance in nonin-sulin-dependent diabetes mellitus. J ClinInvest 68:957-969, 1981

23. Bogardus C, Lillioja S, Mott DM, Hollen-beck C, Reaven G: Relationship betweendegree of obesity and in vivo insulinaction in man. Am J Physio! 248:E286-E291,1985

24. Kissebah AH, Vydelingum N, Murray R,


Committee Report

Evans DF, Hartz AJ, Kalkhoff RK, AdamsPW: Relationship of body fat distributionto metabolic complications of obesity JGin Endocrinol Metab 54:254-260, 1982

25. Butkiewicz EK, Leibson C, O'Brien PC,Palumbo PJ, Rizza RA: Insulin therapy fordiabetic ketoacidosis. Diabetes Care18:1187-1190, 1995

26. Banerji MA, Chaiken RL, Huey H, TuomiT, Norin AJ, Mackay 1R, Rowley MJ, Zim-met P, Lebovitz H: GAD antibody negativeNIDDM in adult black subjects with dia-betic ketoacidosis and increased frequencyof human leukocyte antigen DR3 andDR4. Diabetes 43:741-745, 1994

27. Umpierrez GE, Casals MMC, Gebhart SSP,Mizon PS, Clark WS, Phillips LS: Diabeticketoacidosis in obese African-Americans.Diabetes 44:79-85, 1995

28. Harris MI: Impaired glucose tolerance inthe U.S. population. Diabetes Care 12:464-474, 1989

29. Zimmet PZ: Kelly West Lecture 1991: chal-lenges in diabetes epidemiology: from westto the rest. Diabetes Care 15:232-252,1992

30. Fujimoto WY, Leonetti DL, Kinyoun JL,Shuman WP, Stolov WC, Wahl PW:Prevalence of complications among sec-ond-generation Japanese-American menwith diabetes, impaired glucose toleranceor normal glucose tolerance. Diabetes36:730-739,1987

31. Moss SE, Klein R, Klein BEK, Meuer MS:The association of glycemia and cause-specific mortality in a diabetic popula-tion. Arch Int Med 154:2473-2479, 1984

32. Kuusisto J, Mykknen L, Pyorala K, LaaksoM: NIDDM and its metabolic control pre-dict coronary heart disease in elderly sub-jects. Diabetes 43:960-967, 1994

33. Andersson DKG, Svaardsudd K: Long-term glycemic control relates to mortalityin type II diabetes. Diabetes Care 18:1534-1543, 1995

34. Uusitupaa MIJ, Niskanen LK, Siitonen O,Voutilainen E, Pyorala K: Ten year cardio-vascular mortality in relation to risk factorsand abnormalities in lipoprotein composi-tion in type 2 (non-insulin-dependent)diabetic and non-diabetic subjects. Dia-betologia 18:1534-1543, 1993

35. Polonsky KS, Sturis J, Bell GI: Non-insulin-dependent diabetes mellitus: agenetically programmed failure of the betacell to compensate for insulin resistance. NEngl] Med 334:777-784, 1996

36. Scarlett JA, Gray RS, Griffin J, OlefskyJM,Kolterman OG: Insulin treatment reversesthe insulin resistance of type II diabetesmellitus. Diabetes Care 5:353-363, 1982

37. Firth RG, Bell PM, Rizza RM: Effects oftolazamide and exogenous insulin oninsulin action in patients with non-insulin-dependent diabetes mellitus. NEnglJ Med 314:1280-1286, 1986

38. Simonson DC, Ferrannini E, Bevilacqua S,

Smith D, Barrett E, Carlson R, DeFronzoRA: Mechanism of improvement in glu-cose metabolism after chronic glyburidetherapy. Diabetes 33:838-845, 1984

39. Henry RR, Wallace P, OlefskyJM: Effectsof weight loss on mechanisms of hyper-glycemia in obese non-insulin-dependentdiabetes mellitus. Diabetes 35:990-998,1986

40. Wing RR, Blair EH, Bononi P, Marcus MD,Watanabe R, Bergman RN: Caloric restric-tion per se is a significant factor inimprovements in glycemic control andinsulin sensitivity during weight loss inobese NIDDM patients. Diabetes Care17:30-36, 1994

41. Harris MI, Couric CC, Reiber G, Boyko E,Stern M, Bennett P (Eds.): Diabetes inAmerica. 2nd ed. Washington, DC, U.S.Govt. Printing Office, 1995 (NIH publ.no. 95-1468)

42. Newman B, SelbyJV, Slemenda C, FabsitzR, Friedman GD: Concordance for type 2(non-insulin-dependent) diabetes melli-tus in male twins. Diabetologia 30:763-738,1987

43. Barnett AH, Eff C, Leslie RDG, Pyke DA:Diabetes in identical twins. Diabetologia20:87-93, 1981

44. Herman WH, Fajans SS, Oritz FJ, SmithMJ, Sturis J, Bell GI, Polonsky KS, HalterJB: Abnormal insulin secretion, not insulinresistance, is the genetic or primary defectof MODY in the RW pedigree. Diabetes43:40-46, 1994

45. Byrne MM, SturisJ, Menzel S, Yamagata K,Fajans SS, Dronsfield MJ, Bain SC, Hatter-sley AT, Velho G, Froguel P, Bell GI, Polon-sky KS: Altered insulin secretory responseto glucose in diabetic and nondiabetic sub-jects with mutations in the diabetes sus-ceptibility gene MODY3 on chromosome12. Diabetes 45:1503-1510, 1996

46. Clement K, Pueyo ME, Vaxillaire M, Rako-toambinina B, Thuillier F, Passa P, FroguelP, Roberts J, Velho G: Assessment ofinsulin sensitivity in glucokinase-deficientsubjects. Diabetologia 39:82-90, 1996

47. Vaxillaire M, Boccio V, Philippi A,Vigouroux C, Terwilliger J, Passa P, Beck-man JS, Velho G, Lathrop GM, Froguel P:A gene for maturity onset diabetes of theyoung (MODY) maps to chromosome12q. Nature Genet 9:418-423, 1995

48. Yamagata K, Oda N, Kaisaki PJ, Menzel S,Furuta H, Vaxillaire M, Southam L, CoxRD, Lathrop GM, Boriraj W, Chen X, CoxNJ, Oda Y, Yano H, Le Beau MM, YamadaS, Nishigori H, TakedaJ, Fajans SS, Hat-tersley AT, Iwasaki N, Hansen T, PedersenO, Polonsky KS, Bell GI: Mutations in thehepatocyte nuclear factor-la gene inmaturity-onset diabetes of the young(MODY 3). Nature 384:455-458, 1996

49. Froguel P, Vaxillaire M, Sun F, Velho G,Zouali H, Butel MO, Lesage S, Vionnet N,

Clement K, Fougerousse F, et al.: Closelinkage of glucokinase locus on chromo-some 7p to early-onset non-insulin-dependent diabetes mellitus. Nature356:162-164, 1992

50. Vionnet N, Stoffel M, TakedaJ, Yasuda K,Bell GI, Zouali H, Lesage S, Velho G, Iris F,Passa P, et al.: Nonsense mutation in theglucokinase gene causes early-onset non-insulin-dependent diabetes mellitus.Nature 356:721-1488, 1992

51. Bell GI, Xiang K, Newman MV, Wu S,Wright LG, Fajans SS, Spielman RS, CoxNJ: Gene for non-insulin-dependent dia-betes mellitus (maturity-onset diabetes ofthe young subtype) is linked to DNA poly-morphism on human chromosome 20q.Proc Natl Acad Sci 88:1484-1488, 1991

52. Yamagata K, Furuta H, Oda N, Kaisaki PJ,Menzel S, Cox NJ, Fajans SS, Signorini S,Stoffel M, Bell GI: Mutations in the hepa-tocyte factor-4a gene in maturity-onsetdiabetes of the young (MODY 1). Nature384:458-460, 1996

53. Reardon W, Ross RJM, Sweeney MG,Luxon LM, Pembrey ME, Harding AE,Trembath RC: Diabetes mellitus associ-ated with a pathogenic point mutation inmitochondrial DNA. Lancet 340:1376-1379,1992

54. van den Ouwenland JMW, Lemkes HHPJ,Ruitenbeek W, Sandkuijl LA, de VijlderMF, Struyvenberg PAA, van de Kamp,Maassen JA: Mutation in mitochondrialtRNA (Leu(URR) gene in a large pedigreewith maternally transmitted type II dia-betes mellitus and deafness. Nature Genet1:368-371, 1992

55. Kadowaki T, Kadowaki H, Mori Y, Tobe K,Sakuta R, Suzuki Y, Tanabe Y, Sakura H,Awata T, Goto Y, et al.: A subtype of dia-betes mellitus associated with a mutationof mitochondrial DNA. N Engl J Med330:962-968, 1994

56. Johns DR: Mitochondrial DNA and dis-ease. N EnglJ Med 333:638-644, 1995

57. Gruppuso PA, Gorden P, Kahn CR, Corn-blath M, Zeller WP, Schwartz R: Familialhyperproinsulinemia due to a proposeddefect in conversion of proinsulin toinsulin. N EnglJ Med 311:629-634,1984

58. Robbins DC, Shoelson SE, Rubenstein AH,Tager HS: Familial hyperproinsulinemia:two cohorts secreting indistinguishabletype II intermediates of proinsulin conver-sion.; Clin Invest 73:714-719, 1984

59. Tager H, Given B, Baldwin D, Mako M,Markese J, Rubenstein A, Olefsky J,Kobayashi M, Kolterman O, Poucher R: Astructurally abnormal insulin causinghuman diabetes. Nature 281:122-125,1979

60. Haneda M, Chan SJ, Kwok SCM, Ruben-stein AH, Steiner DF: Studies on mutanthuman insulin genes: identification andsequence analysis of a gene encoding


Committee Report

[Ser82-1] insulin. Proc Natl Acad Sri USA80:6366-6370,1983

61. Given BD, Mako ME, Tager HS, BaldwinD, MarkeseJ, Rubenstein AH, OlefskyJ,Kobayashi M, Kolterman O, Poucher R:Diabetes due to secretion of an abnormalinsulin. NEnglJMed 302:129-135,1980

62. Kahn CR, FlierJS, Bar RS, ArcherJA, Gor-den P, Martin MM, Roth J: The syndromesof insulin resistance and acanthosis nigri-cans. N Engl] Med 294:739-745, 1976

63. Taylor SI: Lilly Lecture: molecular mecha-• nisms of insulin resistance: lessons from

patients with mutations in the insulin-receptor gene. Diabetes 41:1473-1490,1992

64. Ciaraldi TP, El-Roeiy A, Madar Z, ReichartD, Olefsky JM, Yen SSC: Cellular mecha-nisms of insulin resistance in polycysticovarian syndrome. J Clin Endocrinol Metab75:577-583, 1992

65. Dunaif A, Segal KR, Shelley DR, Green G,Dobrjansky A, Licholai T: Evidence fordistinctive and intrinsic defects in insulinaction in polycystic ovaiy syndrome. Dia-betes 41:1257-1266, 1992

66. Schwartz SS, Zeidler A, Moossa AR, KukuSF, Rubenstein AH: A prospective study ofglucose tolerance, insulin, C-peptide, andglucagon responses in patients with pan-creatic carcinoma. Digestive Dis 23:1107-1114,1978

67. Cersosimo E, Pister PWT, Pesola G,McDermott K, Bajorunas D, Brennan MF:Insulin secretion and action in patientswith pancreatic cancer. Cancer 67:486-493,1991

68. Larsen S, HilstedJ, Tronier B, Worning H:Metabolic control and B cell function inpatients with insulin-dependent diabetesmellitus secondary to chronic pancreatitis.Metabolism 36:964-967, 1987

69. Phelps G, Chapman I, Hall P Braund W,Mackinnon M: Prevalence of genetichaemochromatosis among diabeticpatients. Lancet ii:233-234, 1989

70. Handwerger S, Roth J, Gorden P, Di Sant'Agnese P, Carpenter DF, Peter G: Glucoseintolerance in cystic fibrosis. N Engl] Med281:451-461,19'69

71. Yajnik CS, Shelgikar KM, Naik SS, Kan-itkar SY Orskov H, Alberti KGMM, Hock-aday TDR: The ketoacidosis-resistance infibro-calculous-pancreatic-diabetes. Dia-betes Res Clin Pract 15:149-156, 1992

72. Soffer LJ, Iannaccone A, Gabrilove JL:Cushing's syndrome. Am J Med 30:129-146, 1961

73. Jadresic A, Banks LM, Child DF, DiamantL, Doyle FH, Fraser TR, Joplin GF: Theacromegaly syndrome. QJ Med 202:189-204, 1982

74. Stenstrom G, Ernest I, Tisell L: Long-termresults in 64 patients operated upon forpheochromocytoma. Ada Med Scan223:345-352, 1988

75. Berelowitz M, Eugene HG: Non-insulindependent diabetes mellitus secondary toother endocrine disorders. In DiabetesMellitus. LeRoith D, Taylor SI, Olefsky JM,Eds. New York, Lippincott-Raven, 1996,p.496-502

76. Conn JW: Hypertension, the potassiumion and impaired carbohydrate tolerance.NEnglJMed 273:1135-1143, 1965

77. Pandit MK, Burke J, Gustafson AB,Minocha A, Peiris AN: Drug-induced dis-orders of glucose tolerance. Ann Int Med118:529-540, 1993

78. O'Byrne S, FeelyJ: Effects of drugs on glu-cose tolerance in non-insulin-dependentdiabetes (parts 1 and II). Drugs 40:203-219,1990

79. Bouchard P, Sai P, Reach G, Caubarrere I,Ganeval D, Assan R: Diabetes mellitus fol-lowing pentamidine-induced hypo-glycemia in humans. Diabetes 31:40-45,1982

80. Assan R, Perronne C, Assan D, Chotard L,Mayaud C, Matheron S, Zucman D: Pen-tamidine-induced derangements of glucosehomeostasis. Diabetes Care 18:47-55,1995

81. Gallanosa AG, Spyker DA, Curnow RT:Diabetes mellitus associated with auto-nomic and peripheral neuropathy afterVacor poisoning: a review. Clin Toxicol18:441-449, 1981

82. Esposti MD, Ngo A, Myers MA: Inhibitionof mitochondrial complex I may accountfor IDDM induced by intoxication withrodenticide Vacor. Diabetes 45:1531-1534,1996

83. Fabris P, Betterle C, Floreani A, GreggioNA, de Lazzari F, Naccarato R, Chiara-monte M: Development of type 1 diabetesmellitus during interferon alpha therapyfor chronic HCV hepatitis. Lancet340:548, 1992

84. Shiba T, Morino Y, Tagawa K, Fujino H,Unuma T: Onset of diabetes with hightiter anti-GAD antibody after IFN therapyfor chronic hepatitis. Diabetes Res ClinPract 30:237-241, 1996

85. Forrest, JA, Menser MA, Burgess JA: Highfrequency of diabetes mellitus in youngpatients with congenital rubella. Lancetii:332-334, 1971

86. King ML, Bidwell D, Shaikh A, Voller A,Banatvala JE: Coxsackie-B-virus-specificIgM responses in children with insulin-dependent (juvenile-onset; type I) dia-betes mellitus. Lancet i: 1397-1399, 1983

87. Karjalainen J, Knip M, Hyoty H, Linikki P,llonen J, Kaar M-L, Akerblom HK: Rela-tionship between serum insulin antibod-ies, islet cell antibodies and Coxsackie-B4and mumps virus-specific antibodies atthe clinical manifestation of type I(insulin-dependent) diabetes. Diabetologia31:146-152, 1988

88. Pak CY, Eun H, McArthur RG, Yoon J:Association of cytomegalovirus infection

with autoimmune type 1 diabetes. Lancetii:l-4,1988

89. Solimena M, Folli, Aparisi R, Pozza G, DeCamilli P: Autoantibodies to GABA-nergicneurons and pancreatic beta cells in stiff-man syndrome. N Engl] Med 41:347-353,1992

90. Rimoin DL: Genetic syndromes associatedwith glucose intolerance. In The Genetics ofDiabetes Mellitus. Berlin, Springer-Verlag,1976

91. Barrett TG, Bundey SE, Macleod AF: Neu-rodegeneration and diabetes: UK nation-wide study of Wolfram (D1DMOAD)syndrome. Lancet 346:1458-1463, 1995

92. Metzger BE, Organizing Committee: Sum-mary and recommendations of the ThirdInternational Workshop-Conference onGestational Diabetes Mellitus. Diabetes40:197-201, 1991

93. Engelgau MM, Herman WH, Smith PJ,German RR, Aubert RE: The epidemiol-ogy of diabetes and pregnancy in the U.S.,1988. Diabetes Care 18:1029-1033,1995

94. Coustan DR: Gestational diabetes. In Dia-betes in America. 2nd ed. Washington, DC,U.S. Govt. Printing Office, 1995 (NIHpubl. no. 95-1468), p. 703-717

95. Langer O, Rodriguez DA, Xenakis EMJ,McFarland MB, Berkus MD, ArrendondoF: Intensified versus conventional man-agement of gestational diabetes. Am ]ObstetGynecol 170:1036-1047, 1994

96. Magee MS, Walden CE, Benedetti TJ:Influence of diagnostic criteria on the inci-dence of gestational diabetes and perinatalmorbidity. JAMA 269:609-615, 1993

97. Cousins L: Obstetric complications. InDiabetes Mellitus and Pregnancy: Principlesand Practice. 2nd ed. New York, ChurchillLivingstone, 1995, p. 455-468

98. O'SullivanJB, Mahan CM: Criteria for theoral glucose tolerance test in pregnancy.Diabetes 13:278, 1964

99. O'SullivanJB: The Boston Gestational Dia-betes Studies: review and perspective. InCarbohydrate Metabolism in Pregnancy andthe Newborn. Sutherland HW, StowersJM,Pearson DWM, Eds. London, Springer-Verlag, 1989, p. 187-294

100. O'Sullivan JB: Diabetes after GDM. Dia-betes 40 (Suppl. 2): 131-135, 1991

101. Metzger BE, Cho NH, Roston SM, Rad-vany R: Prepregnancy weight and antepar-tum insulin secretion predict glucosetolerance five years after gestational dia-betes mellitus. Diabetes Care 16:1598-1605, 1993

102. Coustan DR, Carpenter MVJ, O'SullivanPS, Carr SR: Gestational diabetes: predic-tors of subsequent disordered glucosemetabolism. Am J Obstet Gynecol 168:1139-1145,1993

103. Kjos SL, Buchanan TA, Greenspoon JS,Montoro M, Bernstein GS, Mestman JH:Gestational diabetes: the prevalence of


Committee Report

glucose intolerance and diabetes mellitusin the first two months postpartum. Am]Obstet Gynecol 163:93-98, 1990

104. Diabetes and pregnancy. In ACOG Techni-cal Bulletin 200, 1994

105. Marquette GP, Klein VR, Niebyl JR: Effi-cacy of screening for gestational diabetes.AmJPerinatology 2:7-14, 1985

106. Dietrich ML, Dolnicek TF, Rayburn WR:Gestational diabetes screening in a private,midwestern American population. Am JObstet Gynecol 156:1403-1408,1987

107. Lucas MJ, Lowe TW, Bowe L, MclntireDD: Class A! gestational diabetes: a mean-ingful diagnosis? Obstet Gynecol 82:260-265,1993

108. Carpenter MW, Coustan DR: Criteria forscreening tests for gestational diabetes.Am] Obstet Gynecol 144:768-773, 1982

109. Sacks DA, Abu-Fadil S, Greenspoon JS:Do the current standards for glucose tol-erance testing in pregnancy represent avalid conversion of O'Sullivans originalcriteria? Am J Obstet Gynecol 161:638—641, 1989

110. Neiger R, Coustan DR: Are the currentACOG glucose tolerance test criteria sen-sitive enough? Obstet Gynecol 78:1117—1120,1991

111. Naylor CD, Sermer M, Chen E, Sykora K:Caesarean delivery in relation to birthweight and gestational glucose tolerance.JAMA 275:1265-1270, 1996

112. Pettitt DJ, Bennett PH, Hanson RL,Narayan KMV, Knowler WC: Compari-son of World Health Organization andNational Diabetes Data Group proceduresto detect abnormalities of glucose toler-ance during pregnancy. Diabetes Care17:1264-1268, 1994

113. Sacks DA, Greenspoon JS, Abu-Fadil S,Henry HM, Wolde-Tsadik G, Yao JF:Toward universal criteria for gestationaldiabetes: the 75-gram glucose tolerancetest in pregnancy. Am J Obstet Gynecol172:607-614, 1995

114. Deerochanawong C, Putiyanum C, Wong-suryrat M, Serirat S, Jinayon P: Compari-son of National Diabetes Data Group andWorld Health Organization criteria fordetecting gestational diabetes mellitus.Diabetologia 39:1070-1073, 1996

115. Charles MA, Fontboune A, Thibult N,Warnet JM, Rosselin GE, Eschwege E:Risk factors for N1DDM in white popula-tion: Paris Prospective Study. Diabetes40:796-799, 1991

116. BrunzellJD, Robertson RP, Lerner RL, Haz-zard WR, Ensink JW, Bierman EL, Porte DJr: Relationships between fasting plasmaglucose levels and insulin secretion duringintravenous glucose tolerance tests. J ClinEndocrinol Metab 42:222-229, 1976

117. Fuller JH, Shipley MJ, Rose G, Jarrett RJ,Keen H: Coronary-heart disease risk andimpaired glucose tolerance: the Whitehall

Study. Lancet i: 1373-1376, 1980118. Charles MA, Balkau B, Vauzelle-Kervoe-

den F, Thibult N, Eschwege E: Revision ofdiagnostic criteria for diabetes (Letter).Lancet 348:1657-1658, 1996

119. Jarrett RJ, Kahn H: Hyperglycemia and dia-betes mellitus. Lancet ii: 1009-1012,1976

120. Klein, R, Comor EB, Blount BA, WingardDL: Visual impairment and retinopathy inpeople with normal glucose tolerance,impaired glucose tolerance and newlydiagnosed NIDDM. Diabetes Care 14:914-918,1991

121. McCartney P, Keen H, Jarrett RJ: The Bed-ford Survey: observations on retina andlens of subjects with impaired glucose tol-erance and in controls with normal glucosetolerance. Diabete Metab 9:303-305,1983

122. Reaven GM, OlefskyJ, Farquhar JW: Doeshyperglycemia or hyperinsulinaemia char-acterize the patient with chemical dia-betes? Lancet i: 1247-1249, 1972

123. Little RR, England JD, Wiedmeyer H-M,McKenzie EM, Pettit DJ, Knowler WC,Goldstein DE: Relationship of glycosy-lated hemoglobin to oral glucose toler-ance: implications for diabetes screening.Diabetes 37:60-64, 1988

124. Reaven GM: Role of insulin resistance inhuman disease. Diabetes 37:1595-607,1988

125. McCance, DR, Hanson RL, Pettitt DJ, Ben-nett PH, Hadden DR, Knowler WC: Diag-nosing diabetes mellitus: do we need newcriteria? Diabetologia 40:247-255, 1997

126. Knowler, WC: Screening for NIDDM:opportunities for detection, treatment andprevention. Diabetes Care 17:445-450,1994

127. Harris MI, Hadden WC, Knowler WC,Bennett PH: Prevalence of diabetes andimpaired glucose tolerance and plasmaglucose levels in the U.S. population aged20-74 yr. Diabetes 36:523-534, 1987

128. Stolk RP, Orchard TJ, Grobbee DE: Whyuse the oral glucose tolerance test? Dia-betes Care 18:1045-1049; 1995

129. McCance DR, Hanson RL, Charles MA,Jacobsson LTH, Pettitt DJ, Bennett PH,Knowler WC: Comparison of tests for gly-cated haemoglobin and fasting and twohour plasma glucose concentrations asdiagnostic methods for diabetes. BMJ308:1323-1328, 1994

130. Engelgau MM, Thompson TJ, HermanWH, Boyle JP, Aubert RE, Kenny SJ, Bad-ran A, Sous ES, Ali MA: Comparison offasting and 2-hour glucose and HbAlc lev-els for diagnosing diabetes: diagnostic cri-teria and performance revisited. DiabetesCare 20:785-791, 1997

131. Jackson CA, Yudkin JS, Forrest RD: Acomparison of the relationships of the glu-cose tolerance test and the glycatedhaemoglobin assay with diabetic vasculardisease in the community: the Islington

Diabetes Survey. Diabetes Res Clin Pract17:111-123, 1992

132. Beks PJ, Mackay AJC, de NeelingJND, deVries H, Bouter LM, Heine RJ: Peripheralarterial disease in relation to glycaemiclevel in elderly Caucasian population: theHoorn Study. Diabetologia 38:86-96,1995

133. Mooy JM, Gootenhuis PA, de Vries H,Kostense PJ, Popp-Snijders C, Bouter LM,Heine RJ: Intra-individual variation of glu-cose, specific insulin and proinsulin con-centrations measured by two oral glucosetolerance tests in general Caucasian pop-ulation: the Hoorn Study. Diabetologia39:298-305, 1996

134. Finch CF, Zimmet PA, Alberti KGMM:Determining diabetes prevalence: a ratio-nal basis for the use of fasting plasma glu-cose concentrations? Diabet Mcd 7:603-610,1990

135. Peters AL, Davidson MB, Schriger DL,Hasselblad V, the Meta-analysis ResearchGroup on the Diagnosis of Diabetes UsingGlycated Hemoglobin Levels 1996: A clin-ical approach for the diagnosis of diabetesmellitus: an analysis using glycated hemo-globin levels. JAMA 276:1246-1252,1996

136. American Diabetes Association: Tests ofglycemia in diabetes (Position Statement).Diabetes Care 20 (Suppl. l):S18-S20,1997

137. American Diabetes Association: Standardsof medical care for patients with diabetesmellitus (Position Statement). DiabetesCare 20 (Suppl. 1):S5-S13, 1997

138. Mayrhofer M, Rabin DU, Messenger L,Standl E, Ziegler AG: Value of ICA512antibodies for prediction and diagnosis oftype 1 diabetes. Exp Clin Endocrinol Diab104:228-234, 1996

139. Bingley PJ, Christie MR, Bonifacio E, Bon-fanti R, Shattock M, Fonte M-T, BottazzoG-F, Gale EAM: Combined analyses ofautoantibodies improves prediction ofIDDM in islet cell antibody-positive rela-tives. Diabetes 43:1304-1310, 1994

140. Verge CF, Gianani R, Kawasaki E, ChaseH, Eisenbarth GS: Prediction of type I dia-betes in first degree relatives using a com-bination of insulin, GAD andICA512bc/IA-2 autoantibodies. Diabetes45:926-933, 1996

141. Hagopian WA, Sanjeevi CB, Kockum I,Landin-Olsson M, Karlsen AE, SundkvistG, Dahlquist G, Palmer J, Lernmark A:Glutamate decarboxylase-, insulin-, andislet cell-antibodies and HLA typing todetect diabetes in a general population-based study of Swedish children. J ClinInvest 95:1505-1511, 1995

142. Harris MI: Undiagnosed NIDDM: clinicaland public health issues. Diabetes Care16:642-652, 1993

143. Klein R: Hyperglycemia and microvascu-lar and macrovascular disease in diabetes.Diabetes Care 18:258-268, 1995


report of the expert committee on the diagnosis and ... · dent diabetes mellitus (iddm, type 1...

Documents