Skip to main content

Diagnosing diabetes—time for a change?

Recently an International Expert Committee including the authors of this paper proposed a revision of the diagnostic criteria for diabetes, recommending that HbA1c may be a better means of diagnosing diabetes than measures of glucose (fasting and/or post-challenge) and that it be adopted as a diagnostic criterion for diabetes [1]. The international expert committee acknowledges that practical, medical, methodological and financial factors may prevent implementation of the recommendation, and in these cases recommends that current standard diagnostic methods are retained.

The recommendation is based on the association of microvascular complications with HbA1c being at least as strong as those with fasting or post-challenge glucose, that HbA1c is subject to less day-to-day variability than fasting or post-challenge glucose, and can be measured at any time of the day without preparations such as fasting or a glucose challenge. So far, this is a proposal to be considered by organisations such as the EASD, American Diabetes Association (ADA), International Diabetes Federation (IDF) and the WHO, as well as by authorities in individual countries worldwide. This decision is not trivial. Re-defining diagnostic criteria may re-define the severity of the disease [2], and if the same criteria are not used in all countries (as has been the case since 1997), a transatlantic trip may cure or cause diabetes simply as a result of small but important differences in diagnostic criteria [3, 4]. Re-defining diagnostic criteria usually means that one set of criteria is replaced by another. Having different diagnostic procedures in place at the same time potentially creates confusion, not clarity. Consequently, now is the time for a careful evaluation of the existing evidence and the public health, economic and practical implications of redefining the diagnostic criteria for diabetes.

Identifying a diagnostic cut-off point

Although the clinical manifestations of diabetes were first described more than 2,000 years ago, the first common and internationally accepted criteria defining diabetes were agreed by a WHO expert committee less than 30 years ago [5]. Before that, different diagnostic thresholds, different glucose loads and time points during the OGTT had been used to classify an individual as having diabetes. This first set of diagnostic criteria was revised by new WHO expert committees in 1985, 1997, 2006 [3, 6, 7], and again this year a WHO expert committee has convened to review the diagnostic criteria for, and classification of, diabetes. These frequent revisions can only be partly explained by progression in our understanding of the underlying mechanisms leading to diabetes and significant new data related to the association between glucose and micro- or macrovascular complications.

Why is defining a diagnostic cut-off point for diabetes so difficult? The simplest answer (and perhaps also the correct answer) is that a cut-off point below which the risk of diabetes-related complications is zero, and beyond which it progressively increases may not exist. The risk of an event like a myocardial infarction increases progressively with increasing levels of the specific associated risk factor (e.g. blood pressure, cholesterol). Why should it be different for glucose? For macrovascular disease, glucose seems to be the same as other risk factors, with risk increasing continuously with increasing levels of 2 h plasma glucose after an OGTT [8] as well as with HbA1c [9]. For stroke (fatal and non-fatal), a non-linear relationship with HbA1c has been suggested, with those with an HbA1c level of >7.0% showing an increased risk [10], but the number of events in each HbA1c stratum is low, and each stratum is very wide. The two factors in combination may well produce artificial cut-off points, and, consequently, further studies would help identify whether thresholds really exist. The fact that macrovascular disease is not specific to diabetes may limit its usefulness in defining a diagnostic cut-off point, despite the fact that it represents by far the most frequent cause of death in patients with type 2 diabetes. The presence of microvascular complications seems to be more specifically related to the presence of diabetes according to the current diagnostic criteria [3, 11]. In a recent analysis of 48,418 people based on the database of the DETECT-2 collaboration [12], there was a glucose and HbA1c range below which diabetes-specific moderate or advanced retinopathy was rare and above which it clearly increased with increasing levels of glycaemia. However, currently available statistical and mathematical models were not able to identify a clear cut-off point (K. Borch-Johnsen, S. Colagiuri, unpublished results). In the absence of a clear, objective cut-off point, all past and present cut-off points for diabetes represent a combination of the best available evidence and a round-table consensus among experts.

Do we need to define diabetes?

Do we need the condition ‘diabetes’, or is glucose just a continuous risk factor like all others? Clearly, glucose is a continuous risk factor, but it is also apparent that if an individual happens to fall in the diabetes category then the impact of other risk factors is greater, as demonstrated in the Multiple Risk Factor Intervention Trial (MRFIT) [13]. This has therapeutic consequences. The absolute risk reduction associated with antihypertensive treatment is higher in people with diabetes than in non-diabetic individuals with the same blood pressure level [14], and lipid-lowering produces a greater reduction in the risk of macrovascular events in individuals with diabetes than in non-diabetic individuals [15]. Furthermore, the classical microvascular complications are specific consequences of prolonged hyperglycaemia, and therefore prevention of and screening for microvascular complications is relevant only in individuals with high glucose levels. Consequently, diabetes is a disease in its own right, not just a risk factor.

Diagnosing diabetes by HbA1c—a possibility?

The idea of using HbA1c to diagnose diabetes is not new. It has been discussed by previous WHO expert committees [7], but only now is it close to being adopted.

The first WHO expert committee report was published in 1980 [5]. At that time our understanding of the association between mean glucose levels and HbA1c was still limited, and again in 1985 no data were available linking HbA1c to the risk of complications in large cohorts including diabetic and non-diabetic individuals. The 1997 ADA expert committee [4] was the first to highlight that the association between an increase in HbA1c from the normal to the diabetic range with retinopathy was similar to the relationships of fasting and 2 h plasma glucose with this diabetic complication. According to their data, the three different measures of glycaemia showed similar decile distribution relationships with retinopathy. In 1997, standardisation of the method used to determine HbA1c was a major concern. This issue remained until the International Federation of Clinical Chemistry (IFCC) together with the EASD and ADA developed an anchor method and thereby provided the possibility for international standardisation of the HbA1c measurement [16].

Diagnosing diabetes by HbA1c—what are the pros?

Undiagnosed diabetes is a major problem in most countries. Epidemiological surveys have consistently demonstrated that 30–50% of all cases of diabetes are undiagnosed [17, 18], and this is a major reason why some countries and organisations recommend screening for diabetes. Diagnosing diabetes is, however, not simple at present. Fasting blood sampling interferes with daily activities such as work, increasing the likelihood that a person with few or no symptoms will remain undiagnosed. Furthermore, one in three cases of undiagnosed diabetes in Europe will have non-diabetic fasting values [19], and therefore an OGTT will be required. This test is time-consuming, costly and inconvenient, and is therefore rarely used outside scientific studies and surveys. Measurement of HbA1c would offer a convenient alternative, as the test can be taken at any time of the day. Once the blood has been drawn, it can be measured locally or sent for analysis further afield without altering the performance of the test. Therefore, shifting to HbA1c could allow the earlier detection of diabetes, which could have a major public health impact.

Accuracy is another concern in relation to diagnosing diabetes based on plasma glucose. To provide a correct measure of plasma glucose, the sample should be placed on ice-water immediately and spun to isolate the plasma within 30–60 min [7]. This is not the standard procedure followed in routine clinical practice. If the sample is left unprocessed, the cells in the blood will continue to live on glucose, and consequently the glucose concentration falls by a median of at least 0.5 mmol/l. Furthermore the intra-individual biological variability in fasting plasma glucose is substantial at 12–15% [20, 21], and is even greater for 2 h plasma glucose, while it is only 2% for HbA1c [22]. This is one reason why a diagnosis of diabetes requires two positive glucose-based tests.

Diagnosing diabetes by HbA1c—what are the cons?

Testing for diabetes by measuring the HbA1c level rather than by glucose testing is more expensive and therefore unaffordable in many, if not most, parts of the world. However, this consideration may not be as significant when viewed from a societal perspective. At present, the asymptomatic person undergoing diagnostic testing requires two morning fasting tests to confirm a diagnosis of diabetes. A substantial proportion will require an OGTT. For many, this results in losing one to three working hours twice before a diagnosis is made. For HbA1c, the test can be performed outside normal working hours, and the increased costs of HbA1c may be more than balanced by the decreased costs carried by the patient. Furthermore, if HbA1c is introduced as a diagnostic test, market forces and competition should lead to a reduction in the costs of an HbA1c test. Projected costs and effectiveness could easily be modelled to examine the economic impact under differing scenarios.

Accuracy in measurement and standardisation of assays remains a concern with regard to HbA1c measurement. Although the IFCC together with EASD and ADA have developed a ‘reference anchor’ and thereby provided the possibility for international standardisation of the HbA1c measurement [16], this has not been fully implemented at present, but an updated examination of the laboratory measurements of glucose and HbA1c by the current International Expert Committee indicates that with advances in instrumentation and standardisation, the accuracy and precision of HbA1c assays at least match those of glucose assays.

The HbA1c test is not available in many parts of the world. This is primarily a problem in low- and middle-income countries, and is definitely a major concern. Introduction of HbA1c as the diagnostic test in this situation is not possible. These countries must have the option (as recommended in the report of the International Expert Committee [1]) of continuing with the current diagnostic criteria until the alternative becomes accessible and affordable.

Testing for diabetes by measuring HbA1c is not possible in all patients with abnormal haemoglobin traits, such as HbS, HbC, HbF and HbE, as they interfere with some HbA1c assay methods [24]. Many assay methods can correct for the presence of the most common haemoglobin traits (see, accessed 8 August 2009). Alternatively, affinity assays that are unaffected by haemoglobin traits may be used [23]. However, where HbA1c testing is limited or not currently available, the use of different routine methodologies to deal with these situations will not be practical in the short term.

Conditions that change erythrocyte turnover, such as haemolytic anaemia, chronic malaria, major blood loss or blood transfusions will lead to spurious HbA1c results. As for settings where HbA1c assays are unavailable, the traditional glucose-based diagnostic tests must be used in individuals for whom interpreting the HbA1c level is problematic.

How to reach a conclusion

The basis for the current diagnostic criteria for diabetes, namely the relationship between glucose and retinopathy, is as strong for HbA1c as other measures of glycaemia. Therefore, the adoption of HbA1c as a method for diagnosing diabetes becomes a discussion not about science but about practical considerations. In this regard, HbA1c has some clear advantages but also a number of significant disadvantages, especially in less developed countries.

The ultimate goal would be to have a single universal diagnostic test for diabetes. Unfortunately, given the clinical situations that interfere with its accurate measurement, HbA1c is unlikely to be that test in the near future without significant technological advances. Consequently, a recommendation to use HbA1c as the only diagnostic criterion for diabetes on a global scale is not possible at present. For the foreseeable future, glucose measurement will continue to have an important role in the diagnosis of diabetes. This means that on the global scale we will see different diagnostic tests being used in different countries. However, within an individual country there should be a nationally endorsed protocol for diagnosing diabetes to avoid confusion among patients and health professionals. When making the decision of whether to adopt HbA1c as a diagnostic test, the country will need to consider the availability, cost and performance of local HbA1c assays and the prevalence of clinical situations that interfere with its measurement.

While the world continues to struggle with the dilemma of how best to diagnose diabetes, the introduction of an additional diagnostic criterion will add to the current problem: that whether or not an individual has diabetes will vary by country and the test used to diagnose it. Nevertheless, a move to include HbA1c as a diagnostic criterion is likely to be a positive step in the journey towards the ultimate goal of having a single universal diagnostic test for diabetes, and to have a test that can reduce the number of people with undiagnosed diabetes worldwide. Making progress will be challenging, but persevering with old habits will not take us forward. As Nirmal in The Hungry Tide finds himself torn ‘between the quiet persistence of everyday change and the heady excitement of revolution—between prose and poetry’ [24] we are also torn between established criteria and convenience against potential—but somewhat uncertain—advantages.



American Diabetes Association


International Federation of Clinical Chemistry


  1. The International Expert Committee (2009) International Expert Committee report on the role of the HbA1c assay in the diagnosis of diabetes. Diabetes Care 32:1327–1334

    Article  Google Scholar 

  2. DECODE-Study (1999) Glucose tolerance and mortality. Comparison of the WHO and ADA diagnostic criteria. Lancet 354:617–621

    Article  Google Scholar 

  3. WHO/IDF (2006) Definition and diagnosis of diabetes mellitus and intermediate hyperglycemia. Report of a WHO/IDF consultation. WHO, Geneva

    Google Scholar 

  4. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (1997) Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 20:1183–1197

    Google Scholar 

  5. WHO (1980) World Health Organization Expert Committee on Diabetes Mellitus: second report. WHO Technical Report 646. WHO, Geneva

    Google Scholar 

  6. WHO Study Group (1985) Diabetes mellitus. WHO Technical Report Series 727. WHO, Geneva

    Google Scholar 

  7. WHO (1999) Definition, diagnosis and classification of diabetes mellitus and its complications: report of a WHO Consultation. Part 1. Diagnosis and classification of diabetes mellitus. WHO, Geneva

    Google Scholar 

  8. The Decode Study Group on behalf of the European Diabetes Epidemiology Group (2001) Glucose tolerance and cardiovascular mortality. comparison of fasting and 2-hour diagnostic criteria. Arch Intern Med 161:397–404

    Article  Google Scholar 

  9. Khaw KT, Wareham N, Luben R et al (2001) Glycated haemoglobin, diabetes and mortality in men in Norfolk cohort of European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk). BMJ 322:15–18

    Article  CAS  PubMed  Google Scholar 

  10. Myinth PK, Sinha S, Wareham NJ et al (2007) Glycated hemoglobin and risk of stroke in people without known diabetes in the European Prospective Investigation into Cancer (EPIC)-Norfolk prospective population study. A threshold relationship. Stroke 38:271–275

    Article  Google Scholar 

  11. Tapp RJ, Tikellis G, Wong TY, Harper C, Zimmet PZ, Shaw JE (2008) Longitudinal association of glucose metabolism with retinopathy. Diabetes Care 31:1349–1354

    Article  CAS  PubMed  Google Scholar 

  12. Colagiuri S, Borch-Johnsen K (2003) DETECT-2: early detection of type 2 diabetes and IGT. Diabetes Voice 48:11–13

    Google Scholar 

  13. Stamler J, Vaccaro O, Neaton JD, Wentworth D (1993) Diabetes, other risk factors and 12-yr cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial. Diabetes Care 16:434–444

    Article  CAS  PubMed  Google Scholar 

  14. Tuomilehto J, Rastenyte D, Birkhager WH et al (1999) Effects of calcium-channel blocker in older patients with diabetes and systolic hupertension. Systolic hypertension in Europe Trial Investigators. N Engl J Med 340:677–684

    Article  CAS  PubMed  Google Scholar 

  15. Pyörälä K, Pedersen TR, Kjekshusn J, Færgeman O, Olsson AG, Thorgeirsson G (1997) Cholesterol lowering with simvastatin improves prognosis of diabetic patients with coronary heart disease. A subgroup analysis of the Scandinavian Simvastatin Survival Study (4S). Diabetes Care 20:614–620

    Article  PubMed  Google Scholar 

  16. American Diabetes Association, European Association for the Study of Diabetes, International Federation of Clinical Chemistry and Laboratory Medicine, International Diabetes Federation (2007) Consensus statement on the worldwide standardization of the hemoglobin A1C measurement. Diabetes Care 30:2399–2400

    Article  Google Scholar 

  17. Mooy JM, Grootenhuis PA, de Vries H et al (1995) Prevalence and determinants of glucose intolerance in a Dutch Caucasian population. The Hoorn study. Diabetes Care 18:1270–1273

    Article  CAS  PubMed  Google Scholar 

  18. Glümer C, Jørgensen T, Borch-Johnsen K (2003) Prevalence of diabetes and impaired glucose regulation in a Danish population - The Inter99 study. Diabetes Care 26:2335–2340

    Article  PubMed  Google Scholar 

  19. The DECODE-study group on behalf of the EDESG (1998) Will new diagnostic criteria for diabetes mellitus change phenotype of patients with diabetes? Reanalysis of European epidemiological data. BMJ 317:371–375

    Google Scholar 

  20. Gambino R (2007) Glucose: a simple molecule that is not simple to quantify. Clin Chem 53:2040–2041

    Article  CAS  PubMed  Google Scholar 

  21. Ollerton RL, Playle R, Ahmed K, Dunstan FD, Luzio SD, Owens DR (1999) Day-to-day variability of fasting plasma glucose in newly diagnosed type 2 diabetic subjects. Diabetes Care 22:394–398

    Article  CAS  PubMed  Google Scholar 

  22. Little RR, Rohlfing CL, Tennill AL, Connolly S, Hanson S (2007) Effects of sample storage conditions on glycated hemoglobin measurement: evaluation of five different high performance liquid chromatography methods. Diabetes Technol Ther 9:36–42

    Article  CAS  PubMed  Google Scholar 

  23. Roberts WL, Safa-Pour S, De BK, Rohlfing CL, Weykamp CW, Little RR (2005) Effects of hemoglobin C and S traits on glycohemoglobin measurements by eleven methods. Clin Chem 51:776–778

    Article  CAS  PubMed  Google Scholar 

  24. Ghosh A (2005) The hungry tide. Harper Collins, New York

    Google Scholar 

Download references

Duality of interest

K. Borch-Johnsen is head of the Steno Diabetes Center, a hospital integrated in the Danish National Health Care Service, but owned by Novo Nordisk. K. Borch-Johnsen holds shares in Novo Nordisk. S. Colagiuri has no duality of interest relevant to this paper.

Author information

Authors and Affiliations


Corresponding author

Correspondence to K. Borch-Johnsen.

Additional information

K. Borch-Johnsen (assigned by the EASD) and S. Colagiuri (assigned by the IDF) were both members of the International Expert Committee on HbA1c and Diagnosis of Diabetes.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Borch-Johnsen, K., Colagiuri, S. Diagnosing diabetes—time for a change?. Diabetologia 52, 2247–2250 (2009).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Diagnostic test
  • Diabetes
  • Diagnosis
  • Diagnostic test
  • Glucose
  • HbA1c
  • OGTT