Neobook — clinical reference

Introduction

Interpretation of diagnostic tests for diabetes and other categories of glycemia

Diagnosis of diabetes

HbA1c vs. plasma glucose for diagnosis

Diagnosis in symptomatic individuals and unequivocal hyperglycaemia

Diagnosis in asymptomatic individuals or doubtful hyperglycaemia

Diagnosis of intermediate hyperglycaemia/categories of increased risk for diabetes (IFG and IGT)

IFG, IGT and prediabetes – the controversies

Screening for type 2 diabetes in adults

Screening for diabetes vs. screening for diabetes and prediabetes

SEMDSA 2017 Recommendations

The diagnosis of diabetes is confirmed:
a. In patients with symptoms of hyperglycaemia (polyuria, polydipsia, blurred vision, weight loss) or metabolic decompensation (diabetic ketoacidosis or hyperosmolar non-ketotic state), when any one single test confirms that the:

Random plasma glucose is ≥ 11.1 mmol/L
Fasting plasma glucose is ≥ 7.0 mmol/L
HbA_1c is ≥ 6.5%
2-hour post-load glucose is ≥ 11.1 mmol/L. However, a GTT is rarely needed in this category of patient

b. In an asymptomatic individual, when any one of the following tests, repeated on separate days within a 2 week period confirms that the:

Fasting plasma glucose is ≥ 7.0 mmol/L
2 hr-post load glucose (OGTT) is ≥ 11.1 mmol/L
HbA_1c is >6.5%

If the diagnosis of diabetes is not confirmed with the repeated test, institute lifestyle modification and retest in 3 to 6 months

HbA_1c can be used as a diagnostic test for diabetes providing that stringent quality assurance tests are in place and assays are standardised to criteria aligned to the international reference values, and there are no conditions present which preclude its accurate measurement

Bedside or point-of-care devices (for glucose or HbA_1c) must not be used to diagnose diabetes

HbA_1c of 6.5% is recommended as the cut-point for diagnosing diabetes. A value of less than 6.5% does not exclude diabetes diagnosed using glucose tests. A glucose based measurement is desirable in individuals with HbA_1cvalues close to the diagnostic cut-point (e.g. 6.0 to 6.4%)

The diagnosis of type 2 diabetes is confirmed when all other causes of diabetes are reasonably excluded (Refer to: Definition and Classification of Diabetes Mellitus)

Impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) are categories of intermediate hyperglycaemia that identify individuals at risk for future diabetes and cardiovascular disease. IFG and IGT are modifiable risk factors. Refer to: HIV and Diabetes for management of these risk factors

Impaired fasting glucose is present when 2 consecutive tests performed on different days confirm that the fasting plasma glucose is 6.1 to 6.9 mmol/L, in the absence of diabetes and impaired glucose tolerance by other tests

Impaired glucose tolerance is present when 2 consecutive tests performed on different days confirm that the 2-hour post-load plasma glucose is 7.8 to 11.0 mmol/L, in the absence of diabetes by any other test

Screening for type 2 diabetes: Screen all overweight adults at any age if they have at least one other risk factor for diabetes. For all other adults, start screening for diabetes at age 45. The frequency of re-screening for diabetes depends on individual risk and can range from 3 months (e.g. the obese individual with IGT and multiple other risk factors for diabetes), to 3 years (e.g. the normal weight individual with no risk factors for diabetes). The preferred screening test for high-risk individuals is the OGTT as it is more sensitive and is the only method for detecting impaired glucose tolerance

INTRODUCTION

There have been no changes to the diagnostic criteria for the diagnosis of diabetes, impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) since the SEMDSA 2012 Guideline. The different recommended diagnostic cut-points for IFG from the American Diabetes Association (ADA) [5.6 to 6.9 mmol/L] and the World Health Organisation (WHO) [6.1 to 6.9 mmol/L] was reconsidered.

The WHO position is retained and endorsed. The categorisation of individuals with non-diabetic but abnormal HbA_1c (5.7 to 6.4%), was also reconsidered. The WHO position, that there is insufficient evidence at this time, make any firm recommendations for the classification of this category of abnormal HbA_1c, is endorsed.

A glucose-based measurement is recommended for risk stratification in these individuals. The use of the term prediabetes to describe the high-risk categories of intermediate hyperglycaemia (IFG and IGT), was debated; although inaccurate and not ideal, the term prediabetes has become entrenched in medical literature, and it is thought to improve communication with professionals and patients regarding diabetes risk. The term prediabetes is therefore an acceptable alternative for IFG and/or IGT.

Table I: interpretation of tests used for screening and diagnosis of diabetes

Fasting plasma glucose^a (FPG)^a (mmol/L)	<5.6 Diabetes excluded	6.0 - 6.9 Impaired fasting glucose^e	≥ 7.0 Diabetes^e
2hr-plasma glucose (2-hr PG)^b(mmol/L)	<7.8 Normal glucose tolerance	7.8 - 11.0 Impaired glucose tolerance^e	≥ 11.1 Diabetes^e
Glycated haemoglobin A_1c (HbA1c)^c(%)	<6.5 Inconclusive^f		≥ 6.5 Diabetes^e
Random plasma glucose (RPG)^d(mmol/L)	<5.6 Diabetes excluded	5.6 - 11.0 Inconclusive^f	≥ 11.1 Diabetes^e

^a Fasting is defined as no caloric intake for at least 8 hours

^b2-h PG is measured during an oral glucose tolerance test (OGTT). The test should be performed as described by the WHO, using a glucose load containing the equivalent of 75 g anhydrous glucose dissolved in 250 ml water and ingested over five minutes

^a,b Glucose should be measured immediately after collection by near-patient testing, or if a blood sample is collected, plasma should be immediately separated, or the sample should be collected into a container with glycolytic inhibitors and placed in ice-water until separated prior to analysis

^cHbA1_c Refers to glycated haemoglobin measurement performed using a DCCT standardised assay and an NGSP certified laboratory, in the absence of conditions that preclude its accurate measurement

^dRandom (casual) is defined as any time of day, without regard to time of last meal. RPG should only be performed in patients with classic symptoms (polyuria, polydipsia, weight loss) or those with hyperglycaemic crises (diabetic ketoacidosis or hyperosmolar non-ketotic state). It is not a recommended screening test for
asymptomatic individuals. A RPG < 11.1 mmol/L does not exclude the diagnosis of diabetes

^eIn non-pregnant individuals with classic symptoms of diabetes or hyperglycaemic crises, a single abnormal test result confirms the diagnosis of diabetes. The diagnosis of impaired fasting glucose (IFG), impaired glucose tolerance (IGT) or diabetes in asymptomatic individuals must be confirmed with a second measurement using the same test method as soon as is convenient

^fFPG or 2-h PG should be measured in those individuals with inconclusive results from RPG or HbA_1c

Note: if two or more test methods yield discordant results, the individual is classified with the more abnormal disorder of glucose regulation. The test with the more abnormal result will need to be repeated on a different day in asymptomatic individuals

INTERPRETATION OF DIAGNOSTIC TEST FOR DIABETES AND OTHER CATEGORIES OF GLYCAEMIA

Plasma glucose, (both fasting and 2-h PG), is a continuous variable that predicts premature mortality and increased risk of microvascular and cardiovascular complications. The current diagnostic criteria for diabetes distinguishes a group of individuals with significantly increased risk.

Normoglycaemia is defined arbitrarily by glucose values that carry a low risk (not zero risk) of cardiovascular complications and future progression to diabetes. IFG and IGT define intermediate hyperglycaemia and risk, i.e. glucose values that are associated with a higher risk of cardiovascular complications and progression to diabetes than normoglycaemia, but a lower risk of microvascular and cardiovascular complications than diabetes.⁵

Unlike diabetes, IFG and IGT are not clinical entities but rather risk factors for future diabetes and adverse outcomes. They are by no means the only risk factors for future diabetes and cardiovascular complications, and ideally these categories of intermediate hyperglycaemia will be superceded by an overall risk assessment (which will include glucose as a continuous variable) for diabetes and cardiovascular disease.⁵

The diagnostic tests, criteria and cut-points for the diagnosis of diabetes mellitus (diabetes) and categories of intermediate hyperglycaemia (categories of increased risk for diabetes) are outlined in Table I.^1,2,3,4,5,6 The footnotes in Table I are extremely important. The same tests are used to diagnose and screen for diabetes and to detect individuals with intermediate hyperglycaemia (IFG and IGT).

DIAGNOSIS OF DIABETES

Diabetes may be identified anywhere along the clinical spectrum: in apparently low-risk asymptomatic individuals who happen to have glucose testing; in those tested based on diabetes risk assessment; and in symptomatic patients.

Diabetes can be diagnosed based on the plasma glucose criteria, either the fasting plasma glucose (FPG), the 2-h plasma glucose (2-h PG) value after a 75 g oral glucose tolerance test (OGTT), a random plasma glucose in symptomatic individuals, or the HbA_1c criteria.^1,2,3,4,5,6No one test is preferred over another for diagnosis.²

For clinical purposes, the diagnosis of diabetes should always be confirmed by repeating the test on another day, preferably, the same test, unless there is unequivocal hyperglycaemia with acute metabolic decompensation or classic (obvious) symptoms (i.e. polyuria, polydipsia and weight loss). The diagnosis of diabetes must be based on formal laboratory testing and not point-of-care or bedside instruments, (e.g. glucose reflectance meters or single-use HbA_1c kits). For glucose-based diagnosis, laboratory venous plasma glucose is preferred.

Capillary blood glucose measurements should only be used for diagnosis in the rare event that laboratory measurements are unavailable. In this event, if the reflectance meter is not pre-calibrated to convert capillary glucose to plasma glucose, the plasma glucose value will need to be derived using the following conversion factor: Plasma glucose (mmol/l) = 0.102 + 1.066 x capillary blood glucose.

HbA_1c vs. plasma glucose for diagnosis

(Refer to: The use of HbA1C for the Diagnosis of type 2 Diabetes)
The HbA_1c test used for diagnosis of diabetes must meet the following conditions: ^2,3,4,6

The test method must meet stringent quality-assurance criteria
The assay must be standardised to criteria aligned to international reference values (i.e. NGSP certified)
The assay must be standardised or traceable to the DCCT reference assay
There must be no conditions present which preclude the accurate measure of HbA_1c.

Point of care HbA_1c assays for diagnostic purposes is not recommended, because although some such assays maybe NGSP certified, proficiency testing is not mandated for performing the test. For a complete list of laboratories that are NGSP certified, the reader is referred to the NGSP website.⁷

HbA_1c has some advantages compared with plasma glucose-based criteria (FPG and OGTT), including greater convenience (fasting not required), greater pre-analytic stability, and less day-to-day perturbations during stress and illness. However, these advantages may be off set by the following: lower sensitivity (of HbA_1c) at the designated cut-point, greater cost, limited availability in certain developing countries and imperfect correlation between HbA_1c and average glucose in certain individuals.

National Health and Nutrition Examination Survey (NHANES) data from the USA indicate that an HbA_1c cut-point of > 6.5% (48 mmol/mol) identifies one third fewer cases of undiagnosed diabetes than a FPG cut-point of > 7.0 mmol/l.⁸

It is important to take age, race/ethnicity, anaemia, haemoglobinopathy, and other states of increased red cell turnover into consideration when using HbA_1c to diagnose diabetes. In conditions associated with increased red cell turnover, such as pregnancy (2nd and 3rd trimester), recent blood loss or transfusion, erythropoeitin therapy, or haemolysis, only blood glucose criteria should be used to diagnose diabetes.^2,3,4

Regarding the plasma glucose/glucose-based criteria, both FPG and 2h-PG may be used to diagnose diabetes. The concordance between FPG and 2h-PG tests is imperfect, as is the concordance between HbA_1c and either glucose-based test. Several studies have confirmed that, compared with FPG and HbA_1ccut-points, the 2-h PG value diagnoses more people with diabetes.^2,3

DIAGNOSIS IN SYMPTOMATIC INDIVIDUALS AND UNEQUIVOCAL HYPERGLYCAEMIA

In patients who have the classic symptoms of hyperglycaemia (polyuria, polydipsia and weight loss), or unequivocal hyperglycaemia i.e. hyperglycaemic crisis (diabetic ketoacidosis or hyperosmolar non-ketotic hyperglycaemia) a single abnormal test is sufficient to confirm the diagnosis of diabetes.

The OGTT can result in severe hyperglycaemia and a random plasma glucose will usually be adequate in this clinical situation. Note that even severe hyperglycaemia detected under conditions of acute infective, traumatic, cardiovascular or other stress including corticosteroid therapy, may be transitory and should not be regarded as diagnostic of diabetes until confirmed subsequently.

DIAGNOSIS IN ASYMPTOMATIC INDIVIDUALS OR DOUBTFUL HYPERGLYCAEMIA

In asymptomatic individuals and in those where there is doubt about the presence of persistent hyperglycaemia, the diagnosis of diabetes (or other categories of intermediate hyperglycaemia) should not be based on a single abnormal test result. It is advisable to perform either a glucose-based test (fasting or 2-hPG), or the HbA_1c test. If the test is abnormal, then the same test must be repeated on another day (preferably within 2 weeks) to confirm the diagnosis.

In the event that both a glucose-based test and the HbA_1ctest are measured, if both are “diagnostic” for diabetes, then the diagnosis of diabetes is confirmed. If only one of these tests is abnormal, a second abnormal result of the same testing method is required to confirm the diagnosis of diabetes on a different day, preferably within two weeks. Should the repeat test not confirm diabetes, the test should then be repeated after 3-6 months.^2,3,4

A test result below the diagnostic threshold for diabetes does not exclude the diagnosis of diabetes. So although the diagnostic cut-point for fasting plasma glucose is 7.0 mmol/Land 6.1 mmol/l for diabetes and IFG respectively, one can only exclude the diagnosis of diabetes if the fasting plasma glucose is ≤ 5.6 mmol/l.

The level of HbA_1c below which the diagnosis of diabetes is excluded is not known. The American Diabetes Association (ADA) uses a cut-point of < 5.7%, but the World Health Organisation (WHO) report does not endorse this. This raises a dilemma in individuals in whom the HbA_1c test result is close to, but does not exceed, the diagnostic cut-point of 6.5%. SEMDSA recommends a glucose-based measurement (FPG or 2-h PG) for high-risk individuals whose HbA_1c is abnormal but not diagnostic of diabetes (Refer to: Diabetic Eye Disease for risk factors).

DIAGNOSIS OF INTERMEDIATE HYPERGLYCAEMIA/CATEGORIES OF INCREASED RISK FOR DIABETES (IFG & IGT)

Table I above outlines the diagnostic criteria for IFG and IGT. By definition, these categories of glycaemia mandate a FPG or OGTT for diagnosis. As is the case with diabetes, there is also considerable discordance between the FPG and 2-h PG test in identifying intermediate hyperglycaemia. Nevertheless, a fair proportion of individuals will have both IFG and IGT.⁵

Glucose-based measurements have variable reproducibabilty which emphasises the need for repeat testing to confirm the diagnosis. Approximately 40-60% of individuals with IFG and or IGT will have confirmatory tests on retesting, about 10% will meet the diagnostic criteria for diabetes and the rest will revert to normal.⁵ The reproducibility of fasting glucose is somewhat better than the 2-h PG.

The risk of progression to diabetes varies among ethnic groups. A meta-analysis of prospective studies showed that the annual incidence of diabetes in individuals with IFG and IGT was 6-9% and 4-6% respectively. The incidence in persons with both IFG and IGT was significantly higher at 15-19%/yr.¹³ In South African Indians with IGT, the annual incidence of diabetes was found to be 12.6% per year over four years.¹⁴ Refer to: Prevention or Delay of type 2 Diabetes Mellitus for a discussion on the management of IFG and IGT.

IFG, IGT AND PREDIABETES - THE CONTROVERSIES
(Optional reading)

The 2011 WHO report only recommended an HbA_1c cut-point (> 6.5%/ 48mmol/mol) for the diagnosis of diabetes and emphasised that a value < 6.5% does not exclude diabetes diagnosed using glucose tests. The report concluded that there was insufficient evidence at the time to make any formal recommendation on the interpretation of HbA_1c level < 6.5%.⁴

ThE report also endorsed the previous WHO recommendations^1,5 of retaining the cut-point plasma glucose for IFG at 6.1 mmol/L vs. the 2003 ADA recommended cut-point of 5.6 mmol/L^2,3; the reasons included lack of evidence of any benefit in reducing adverse outcomes or progression to diabetes, and the possible impact on individuals and health systems with the significant increase in IFG prevalence.⁴

The ADA has since 2010, recommended that an HbA_1c 5.7-6.4% (39-46 mmol/mol) can be used to identify categories at increased risk for diabetes in addition to IFG and IGT, and uses the term prediabetes to describe these three categories; the argument being that for all these tests, the risk continues, extending below the lower limit of the range and becoming disproportionately greater at higher ends of the range.^2,3,9

In addition, the ADA has recommended screening individuals with risk factors for the purpose of detecting prediabetes (as opposed to only screening to detect diabetes). The counter argument is that there is no universal lower level of either glucose or HbA_1c that eliminates risk.

The question needs to evolve from “is a certain level of glucose associated with risk” to ” is intervention at that level likely to be beneficial, cost-effective and affordable”. Regarding IFG, lowering the FPG threshold to 5.6 mmol/l means that ~30% of people would be classified as having prediabetes. There is little evidence that this is very useful.

The same is true for lowering the HbA_1c threshold because as the threshold goes lower than 6.0 %, one is on the steep section of the normal distribution and many people get labelled. In addition, HbA_1c in the range 5.7 to 6.4% appears to identify a much smaller population of prediabetes than does IFG or IGT, it includes people who do not IFG or IGT, IFG and IGT, are stronger predictors of CVD than HbA_1c ,and intervention studies in prediabetes have included patients based on IGT (predominantly) and IFG. There have been no prospective intervention studies based on HbA_1c.^5,10,11

The question to ask is what is the purpose of labelling someone with prediabetes? What are you going to do on the basis of that knowledge? How many people can you afford to offer individual prevention to? Using this logic, one would probably keep the thresholds high. In the case of FPG there is no evidence that people in the range 5.5-6.1 mmol/L are at substantial cardiovascular risk. There is also no trial evidence of benefit of lifestyle intervention; and the population is large with an unknown cost-benefit ratio for any intervention.

SEMDSA’s position is that the cut point of FPG and IFG should remain at 6.1 mmol/l, and that the previous position of not recommending HbA_1c for intermediate hyperglycaemia should be retained until there is substantial evidence from prospective studies that there is a defined threshold of HbA_1c, above which there is increased/absolute risk of progression to diabetes or development of its complications, and until it can be shown that no individual with HbA_1c < 6.5% has diabetes.

The same is true for screening for IFG and IGT, i.e. until there is further evidence, screening for intermediate hyperglycaemia should not be recommended. It is also important that serum insulin should not be used for diagnosis of intermediate hyperglycaemia.

The use of the term prediabetes for individuals with IFG/IGT is problematic because it implies that all of these individuals will develop diabetes, and that those who do not meet the criteria for prediabetes will not, neither of which is true. For example, it ignores people with other non-glucose based risk factors such as the woman with a history of gestational diabetes, whose risk of diabetes is quite high (20-60%) over the five to ten years after pregnancy.¹²

Nevertheless the term prediabetes has become sufficiently entrenched in medical and lay literature so as not to be ignored or discarded; a PubMed search yields more literature for the search term “prediabetes” than it does for “intermediate hyperglycaemia”.

The term prediabetes also facilitates an easier conversation with professionals and patients to help understand the high risk status. If the term prediabetes is used, it should be confined to IFG and IGT, must not include HbA_1c and those using the term must not ignore the risk of people who do not have intermediate hyperglycaemia.

SCREENING FOR TYPE 2 DIABETES IN ADULTS

As described above, the distinction between diagnostic testing and screening for diabetes is somewhat blurred. The same tests are used for “screening” and for diagnosis.

Diabetes may be identified anywhere along a spectrum of clinical presentations, ranging from low-risk individuals who happen to have glucose testing incidentally (random screening), to individuals identified as having a high risk for diabetes during routine consultations for unrelated health matters (opportunistic screening), to those who are deliberately identified and tested because of their high risk status (targeted screening).

The spectrum then extends to the higher-risk individual with clinical features suggestive of diabetes (e.g. obese adult with recurrent urinary infections and nocturia) whom the provider tests because of a high suspicion of diabetes, and finally to the patient with classic symptoms or metabolic decompensation. The latter two scenarios would be considered diagnostic testing.

Because of the need for follow-up, screening should only be carried out within the health care setting. Community screening outside a health care setting (e.g. fun days, shopping centres) is not recommended, because individuals with abnormal (positive) tests may not seek or have access to appropriate follow-up testing and care. Or, for those who test negative, there may be failure to ensure appropriate repeat testing. Such screening may also be poorly targeted, i.e. may fail to reach groups most at risk, and inappropriately test those at low risk (the “worried well”) or those already diagnosed.

Similarly, random screening for all adults is not recommended until after the age of 45 years. The indications for targeted and opportunistic screening are described in Table below.

Table II: Criteria for screening for type 2 diabetes in asymptomatic adults ^1,2,3,⁵

Indications	1. High risk individuals: All adults (any age) who are overweight (BMI> 25 kg/m² or > 23 kg/m² in Asians), plus one or more additional risk factors^b: Physical inactivity Hypertension [Blood pressure (BP) ≥ 140/90 mmHg] or treatment for hypertension First degree relative with diabetes Dyslipidaemia^c Polycystic ovarian syndrome High-risk race/ethnicity (Asian Indian, Coloured) Cardiovascular disease history Gestational diabetes or baby > 4 kg Previous IFG or IGT Other conditions associated with insulin resistance (severe obesity, acanthosis nigricans) 2. If no risk factors: Age ≥ 45 years
Frequency	At three-year intervals, if normal More frequently, based on initial result and risk status
Test Method	FPG, 2-h PG (OGTT) or HBA_1c The OGTT is the preferred test in high risk individuals

a Only to be done within the healthcare setting
b Risk factors for future diabetes
c Serum high-density lipoprotein (HDL) cholesterol < 0.90 mmol/l, or triglycerides > 2.82 mmol/l

SCREENING FOR DIABETES VS. SCREENING FOR DIABETES AND PREDIABETES

In 2016, the ADA² recommends guidelines for screening tests for both type 2 diabetes and prediabetes in asymptomatic adults (vs. screening tests only for diabetes). This has been discussed in HbA_1c vs. plasma glucose for diagnosis.

SEMDSA advocates screening for the detection of diabetes; individuals identified with IFG/IGT during this process can be managed appropriately. In any case the risk factors for prediabetes are the same as the risk factors for diabetes.

Author: Ayesha A. Motala
Editor: Aslam Amod

References

Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med. 1998;15(7):539-553.
American Diabetes Association. Classification and diagnosis of diabetes. Sec. 2. In Standards of Medical Care in Diabetesd2016. Diabetes Care 2016; 39(Suppl. 1):S13–S22.
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2014;37(Suppl. 1):S81–S90.
World Health Organization. Use of glycated haemoglobin (HbA1c) inthe diagnosis of diabetes mellitus: http://www.who.int/diabetes/publications/reporhba1c_2011.pdf.(accessed 25 January 2016).
World Health Organization. Definition and diagnosis of diabetesmellitus and intermediate hyperglycaemia: http://whqlibdoc.who.int/publications/2006/9241594934_eng.pdf(accessed 21 January 2016).
Canadian Diabetes Association Clinical Practice Guidelines Expert Committee.Canadian Diabetes Association 2013 Clinical Practice Guidelines for the
Prevention and Management of Diabetes in Canada. Can J Diabetes 2013;37(suppl 1):S1-S212.
http://www.ngsp.org/certified.asp (accessed 13 February 2012).
Cowie CC, Rust KF, Byrd-Holt DD, et al. Prevalence of diabetes and high risk for diabetes using A1C criteria in the U.S. population in 1988-2006. Diabetes Care. 2009;32:1327-1334.
American Diabetes Association: Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 2010; 33(Suppl.1):S62–S69.
Buysschaert M, Medina J-L, Buysschaert B, Bergman M. Definitions (and Current Controversies) of Diabetes and Prediabetes. Curr Diabetes Rev. 2015;12(1):8-13. doi:10.2174/1573399811666150122150233.
Bonora E, Tuomilehto J. The Pros and Cons of Diagnosing Diabetes With A1C. Diabetes Care. 2011;34(Supplement_2):S184-S190. doi:10.2337/dc11-s216.12.
Tabák AG, Herder C, Rathmann W, Brunner EJ, Kivimäki M. Prediabetes: a high-risk state for diabetes development. Lancet. 2012;379(9833):2279-2290.
doi:10.1016/S0140-6736(12)60283-9.
Gerstein HC, Santaguida P, Raina P, et al. Annual incidence and relative risk of diabetes in people with various categories of dysglycemia: A systematic overview and meta-analysis of prospective studies. Diabetes Res Clin Pract. 2007;78(3):305-312. doi:10.1016/j.diabres.2007.05.004.
Motala AA, Omar MAK, Gouws E. High Risk of Progression to NIDDM in South-African Indians with Impaired Glucose Tolerance. Diabetes. 1993;42(4):556-563. doi:10.2337/diab.42.4.556.