Introduction

A worldwide pandemic of obesity and diabetes is well advanced. In the United States alone, diabetes now affects an estimated 24.1 million people, an increase of more than 3 million in approximately 2 years. Twenty-five percent of persons with diabetes in the United States do not know they have diabetes. Another 57 million people in the United States have prediabetes,[1] defined as people with impaired fasting glucose (IFG) or impaired glucose tolerance (IGT), some of whom in fact already have the characteristic microvascular changes resulting from diabetes itself.[2,3] Worldwide, the number of people with prediabetes is estimated to be 314 million and is projected to be 418 million in 2025.[4] As the prevalence of and progression to diabetes continue to increase, diabetes-related morbidity and mortality have emerged as major public health care issues. Diabetes is expensive—the associated yearly cost of diabetes in the United States is $174 billion. Direct costs related to diabetes, diabetes complications, and general medical care are $116 billion, and indirect costs are $58 billion from illness, disability, and premature mortality.[5]
Prediabetes raises short-term absolute risk of type 2 diabetes by 3-to 10-fold, with some populations exhibiting greater risk than others.[6,7] People with diabetes are vulnerable to multiple and complex medical complications. These complications involve both cardiovascular disease (CVD) (heart disease, stroke, and peripheral vascular disease) and microvascular disease (ie, retinopathy, neuropathy, and microalbuminuria). Most patients with diabetes die of CVD.[8]
Epidemiologic evidence suggests that the complications of diabetes begin early in the progression from normal glucose tolerance to frank diabetes. Early identification and treatment of persons with prediabetes have the potential to reduce or delay the progression to diabetes[9-13] and related CVD[14,15] and microvascular disease.[16]

Despite the clear origins of diabetes-related complications early in the prediabetic state, few recommendations have been made for the diagnosis and management of patients with prediabetes. No medications are approved by the US Food and Drug Administration for addressing either IFG or IGT. Most insurance companies deny payment for lifestyle treatment to prevent diabetes. There are differences in opinion among health care professionals regarding the therapeutic approach to treating people with prediabetes. Many of these people already have diabetes related complications, yet there are no defined goals and targets of treatment in prediabetes for the many risk factors, which include glucose levels, weight, blood pressure, and lipid levels.

It is clear that the risks and adverse consequences of high blood glucose occur at much lower glucose levels than those at which we currently define as diabetes. Acknowledging these many challenges, there are major questions that health care professionals must address such as: "When do the risks of diabetes begin?"; "What can we do to prevent diabetes?"; "What strategies are necessary to reduce the vascular complications related to diabetes?"; and "How does society pay for the preventive costs of diabetes in the large number of patients at risk?"

Considerations
The American College of Endocrinology (ACE) Task Force on the Prevention of Diabetes was convened under the auspices of ACE. This group formulated 6 specific diagnostic and management questions. Over a 2-day period, 23 international experts reviewed all available scientific data to assist the committee in addressing these questions (Appendix 1).
The consensus conference examined the current status of prediabetes, the facts about related complications, what happens to people who progress to diabetes, available intervention trials, economic implications of early intervention, and what future studies are needed.
The consensus conference's recommendations are primarily based upon analysis of the available scientific evidence; expert opinion was used when necessary.
These recommendations are aimed at the general medical community and are especially directed at general primary care physicians, health care providers, and educators because they are at the forefront of treating this condition. The message also calls upon national and local community leaders and governments to increase efforts to curtail the obesity and diabetes epidemics and for further research in this high-risk population.
ACE and the American Association of Clinical Endocrinologists (AACE) are available to support medical societies and public efforts in implementing these recommendations.

Question 1
What is the Spectrum Between Normal Glucose Tolerance, Prediabetes, and Diabetes, and What Should be the Diagnostic Criteria for Each?

Prediabetes currently refers to people who have IFG (100-125 mg/dL [5.6-6.9 mmol/L]), IGT (2-hour postglucose load, 140-199 mg/dL [7.8-11 mmol/L]), or both.

There is a continuous spectrum of glucose levels between those considered normal (fasting < 100 mg/dL [< 5.6 mmol/L]; postchallenge < 140 mg/dL [< 7.8 mmol/L]) and those that are considered diagnostic for diabetes (fasting ≥ 126 mg/dL [≥ 7 mmol/L]; postchallenge ≥ 200 mg/dL [≥ 11.1 mmol/L]). IGT should be considered more important for risk than IFG.

Presently, diabetes is diagnosed somewhat arbitrarily on the basis of the glucose level associated with the eventual appearance of characteristic end-organ complications, specifically retinopathy. Currently, diabetes may be diagnosed at a fasting plasma glucose level of 126 mg/dL (7 mmol/L) or higher or a 2-hour postglucose challenge plasma glucose concentration of 200 mg/dL (11.1 mmol/L) or higher.[17,18] However, in large population studies, values for both normal fasting and 2-hour plasma glucose levels are considerably lower than these thresholds for diagnosing diabetes. The upper limit of normal fasting plasma glucose is widely believed to be 99 mg/dL (5.5 mmol/L) although metabolic and vascular abnormalities have been described recently at values less than that. Similarly, 2-hour postglucose levels less than 140 mg/dL (7.8 mmol/L) are believed to be within the reference range. Whatever label is given to the "gap" in glycemic status between normal and diabetes, the data indicate that, for many individuals, these glucose levels are not benign and may herald overt type 2 diabetes and CVD,[19,20] as well as microvascular complications.[2,3] Thus, the ill-defined area in fasting glucose of 100 to 125 mg/dL (5.6-6.9 mmol/L) and 2-hour levels of 140 to 199 mg/dL (7.8-11 mmol/L) is thought to describe a prediabetic range, where some degree of increased microvascular and macrovascular complications of diabetes has been described.[21,22]

This intermediate state of prediabetes constitutes inherent disease risk. The progression to diabetes for patients with IGT is 6% to 10% per year, and for persons with both IFG and IGT, the cumulative incidence of diabetes by 6 years may be as high as 65% (compared with levels on the order of 5% for those with normal glucose levels at baseline).[23] Approximately half of patients with IGT meet the National Cholesterol Education Program (NCEP) criteria for the diagnosis of metabolic syndrome.[24]

Numerous investigations indicate that the risk of CVD maintains a linear association with glycemia well below the present diagnostic threshold for type 2 diabetes and extends to lower glucose levels than those defined by the criteria for the diagnosis of IFG and IGT[25,26] into the range of glucose otherwise considered normal.[27] In addition, the CVD event rate in epidemiologic studies, such as AusDiab (Australian Diabetes, Obesity, and Lifestyle Study)[25] and Framingham[20] and intervention studies such as STOP-NIDDM (Study to Prevent Non–Insulin-Dependent Diabetes Mellitus)[14] and DREAM (Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication),[10] suggest nearly a doubling of cardiovascular risk in prediabetes compared with what would be expected for individuals without IFG or IGT. The Nurses' Health Study demonstrated that women destined to convert to type 2 diabetes (a "true" prediabetes population) have nearly 3 times the risk of a cardiovascular event compared with those who remained nondiabetic over an extended follow-up period.[28]

Conversion of IFG to diabetes further increases CVD mortality 2-fold.[29] Similarly, the DECODE study (Diabetes Epidemiology: Collaborative Analysis of Diagnostic Criteria in Europe) found a higher coronary heart disease risk with elevated 2-hour postglucose levels even in the presence of normal fasting glucose levels.[22] The syndrome of multiple cardiovascular risk factors or the metabolic syndrome described by NCEP characterizes a group of individuals at increased risk of diabetes, as well as CVD. The age-adjusted risk in the Framingham offspring study with metabolic syndrome was 2.54 for coronary heart disease and 6.92 for diabetes in men, whereas in women, coronary heart disease risk was lower at 1.5 while diabetes risk was similar at 6.5.[7] Thus, IFG, IGT, and metabolic syndrome may each describe a prediabetic state that appears to have coincident heightened coronary heart disease risk. For example, in the San Antonio Heart Study, IGT increased future diabetes risk by approximately 5-fold, as did a diagnosis of the metabolic syndrome.[30] Individuals with combinations of these high risk states have increased absolute risk for type 2 diabetes compared with individuals who meet criteria for only a single risk category. In the Insulin Resistance Atherosclerosis Study, Haffner et al (unpublished data, 2008) found that patients with a diagnosis of IGT, IFG, or metabolic syndrome converted to diabetes at a rate of 8% to 10% per year, and if all 3 diagnoses were present, conversion rates far exceeded 10% per year.

More traditional diagnoses of prediabetes are future based risk predictions and include women with a history of poly cystic ovary syndrome or gestational diabetes, offspring of parents with type 2 diabetes, and individuals with abdominal adiposity. Patients with CVD also have an increased prevalence of prediabetes. Type 2 diabetes is also being observed with increased frequency in adolescents, but is uncommon in children younger than 10 years.[31]
The panel recommends targeted screening for populations at high risk for the development of diabetes. Risk factors include the following:

• Family history of diabetes[18]
• Cardiovascular Disease[18]
• Being overweight or obese[18]
• Sedentary lifestyle[18]
• Non-white ancestry[18]
• Previously identified IGT, IFG, and/or metabolic syndrome[18]
• Hypertension[18]
• Increased levels of triglycerides, low concentrations of high-density lipoprotein cholesterol, or both[18]
• History of gestational diabetes[18]
• Delivery of a baby weighing more than 9 lb (4 kg)[18]
• Polycystic ovary syndrome[18]
• Receiving antipsychotic therapy for schizophrenia and severe bipolar disease[32]

A diagnosis of prediabetes can be made by any of 3 criteria:[3] (a) IFG with glucose levels of 100 to 125 mg/dL (5.6-6.9 mmol/L). IFG should be determined after an overnight fast (8 hours minimum). Patients should not be active or have had caffeine or any other factor known to affect carbohydrate metabolism. (b) IGT with glucose levels of 140 to 199 mg/dL (7.8-11 mmol/L) after a 75-g oral glucose load given in the morning (after an appropriate overnight fast).[33] Patients should be on an adequate carbohydrate intake before the test, should not be physically active during the test, and must not smoke. For purposes of diagnosing IGT, a single sample drawn after a 2-hour glucose load is sufficient. The benefit to be gained by a 2-hour glucose tolerance test was considerable in the EUROHEART survey. Patients with impaired glucose metabolism identified by 2-hour oral glucose tolerance testing were greater in number than patients identified by routine determination of fasting glucose alone.[34] In patients with IFG, a 2-hour glucose tolerance test may further clarify the level of risk while also detecting undiagnosed diabetes. (c) Metabolic syndrome diagnosed by the NCEP criteria[24] should be considered a prediabetes equivalent. It predicts future diabetes better than IFG. Three of 5 metabolic syndrome criteria are sufficient; recent evidence suggests even 2 of 5 metabolic syndrome criteria may be adequate as well. Thus, it seems clear that prediabetic states may represent heterogeneous etiologies. These states not only entail increased risk of diabetes, but also increased risk of CVD. Progression rates of metabolic syndrome, IFG, or IGT to diabetes vary according to degrees of initial hyperglycemia, racial and ethnic backgrounds, and environmental influences. The higher the glucose values, the greater the risk of progression to diabetes and diabetic complications.[35]

Question 2 What are the Clinical Risks of Not Treating Prediabetes? In order to assess the clinical risk of not treating prediabetes, there are 2 obvious sources of data: (a) observational data from populations of patients with prediabetes and (b) data from interventional studies comparing placebo with active treatment. Most, if not all, diabetic complications progressively worsen as glycemia worsens. In the DECODE study of more than 22 000 patients, 2-hour postload glucose levels were associated with a linear increase in hazard ratio for all-cause mortality as the 2-hour blood glucose concentration increased from 95 to 200 mg/dL (5.3-11.1 mmol/L).[36] Over this range of 2-hour glucose levels, the risk doubled and approached that of patients treated for diabetes. A 10-year follow-up of a population-based cohort of Finnish subjects[37] comparing participants with normal glucose tolerance with participants who had IGT at baseline, participants with nonprogressive IGT, and participants with IGT that progressed to diabetes, showed a 130% increase in cardiovascular mortality in those who did not progress to diabetes compared with only a 70% increase in individuals who developed diabetes during follow-up. In a 23-year follow-up of the Honolulu Heart Study, an increase in sudden death was associated with postchallenge hyperglycemia.[38] With regard to interventional trials, in the Diabetes Prevention Program, diabetic retinopathy was observed in 7.9% of patients with IGT compared with 12.6% in patients with IGT that later progressed to diabetes. Furthermore, in the placebo IGT group, there was a progressive increase in the prevalence of hypertension from 29% to 38%, an increase in the prevalence of dyslipidemia from 6% to 16%,[15] and an increase in the prevalence of clinical CVD events by approximately 50% (relative risk 0.47 over 4 years).[14] IGT was also associated with impaired indices of autonomic function.[39] In other recent studies, the incidence of retinopathy in IFG has been higher (9% to 16%) than that described for the Diabetes Prevention Program (7.9% to 12.6%).[3] The prevalence of retinopathy has been observed to increase dramatically in the highest deciles of each glycemic measure. In the STOP-NIDDM trial, there was a 16% cumulative increase in hypertension (> 140/90 mm Hg) in the placebo- treated IGT participants over a 3-year period. More gradual increases in microalbuminuria prevalence for patients with IGT have also been observed.[2] In patients presenting with idiopathic peripheral neuropathy, approximately 40% have IGT.[40] Findings from these studies suggest that patients with IGT are at risk when IGT is identified, and when untreated, these patients experience progression in their incidence of diabetes, as well as in microvascular and macrovascular risk.