Evidence of the Month

Commentaries on both new and classic studies of importance for the treatment of diabetes are posted here monthly.

Petra-Maria Schumm-Draeger Klinik für Endokrinologie, Diabetologie und Angiologie Klinikum Bogenhausen Munich, Germany	Click here to download PowerPoint slides of this commentary.
	Click here for the programme evaluation and post-test.

Background
Direct measurement of glucose has been the basis for diabetes diagnosis for more than 30 years. In the absence of extreme symptoms, typically those seen with the clinical onset of Type 1 diabetes, diagnosis has relied mainly on the fasting plasma glucose (FPG) assay or, less frequently, the oral glucose tolerance test (OGTT). Diagnostic cut points with these tests were based initially on the association of glucose levels with the risk of symptomatic diabetes, rather than the risk of long-term vascular complications. That changed in 1997 when the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus scrutinized epidaemiological data on diabetic retinopathy in 3 populations: Pima Indians of the southwestern United States, an Egyptian cohort, and participants in the US National Health and Nutrition Examination Survey. The committee found that the deciles of glucose at which retinopathy risk began to increase were the same within each population and the same for each test used—FPG, OGTT, or glycated haemoglobin (HbA_1c) (Expert Committee, 1997).

The HbA_1c is a measure of the percentage of haemoglobin molecules that are glycated. It is a reflection of blood glucose exposure in the preceding 2 to 3 months and a well-established index of long-term microvascular risks. The use of HbA_1c for diabetes diagnosis has advantages on both practical and clinical levels, so much so that expert committees considered recommending it for diagnosis in 1997 and 2003 (Expert Committee, 1997, 2003). In each instance, the committees chose not to recommend the HbA_1c, in large part because of concerns about assay standardization. Now, an International Expert Committee appointed by 3 leading diabetes associations has endorsed diagnosis by HbA_1c. In presenting its case, this committee pointed out that its views on HbA_1c are not necessarily those of the appointing organizations. Moreover, the committee declined to label HbA_1c or any other single test as the ‘gold standard’ for distinguishing pathological from benign glycaemic states.

Methods and Key Results
The International Expert Committee was appointed by the American Diabetes Association (ADA), the European Association for the Study of Diabetes, and the International Diabetes Federation. The committee convened in 2008 to revisit evidence on diagnostic assays. Its 2009 report recommending HbA_1c use in diabetes diagnosis covered the following key points:

Rationale for HbA_1c—The HbA_1c assay captures long-term glycaemic exposure; therefore, it stands to be a better predictor of diabetic complications than single or episodic measures of glucose, as obtained by FPG and OGTT. In observational studies, HbA_1c results have shown stronger and more consistent relationships with retinopathy than fasting glucose levels. These relationships have been seen in both Type 1 (Diabetes Control and Complications Trial [DCCT]) and Type 2 (United Kingdom Prospective Diabetes Study [UKPDS]) diabetic populations.

Standardization and accuracy—The International Expert Committee concluded that ‘the accuracy and precision of A1C assays at least match those of glucose assays’ due to advances in instrumentation and standardization. (Specifically, the HbA_1c assay is now standardized and aligned to the DCCT and UKPDS.) In fact, glucose assays are subject to more inaccuracies than the HbA_1c. Day-to-day glucose values vary to a greater extent than HbA_1c values; glucose samples are relatively unstable even before processing; and many laboratory instruments that measure glucose have significant bias from the reference method—enough for potential misclassification of 12% of patients tested.

Diagnostic cut point—An HbA_1c of 6.5 % is associated with a substantial increase in the prevalence of diabetic retinopathy. Therefore, the International Expert Committee deemed an HbA_1c of ≥6.5 % to be ‘sufficiently sensitive and specific’ to identify individuals with diabetes. In the absence of clinical symptoms of diabetes or a blood glucose level >200 mg/dl (11.1 mmol/l), a diagnostic HbA_1c assay should be repeated for confirmation.

Subdiabetic states—Individuals with an HbA_1c between 6 % and 6.5 % are likely to be at very high risk for retinopathy. However, the specific point at which the risk for diabetes becomes clinically important is unclear. The International Expert Committee advised against use of the terms prediabetes, impaired fasting glucose, and impaired glucose tolerance, saying only that the risk for diabetes becomes greater as HbA_1c approaches 6.5 % and that preventive interventions should begin for individuals with HbA_1c ≥6 %.

Clinical Implications
When the International Expert Committee published its recommendations, the ADA released a statement ‘endorsing in principle’ the use of the HbA_1c test for diabetes diagnosis. Rather than issuing a position statement at the time, the organization said it would refer the committee findings to practice groups to review the implications. There are reasons to proceed thoughtfully. The HbA_1c assay is a relatively expensive test, and in many parts of the world it might not be routinely available. (Current in-office, point-of-care HbA_1c tests are not considered reliable for making a diagnosis.) Further, some countries have high rates of severe chronic anaemia, haemolytic anaemia, chronic malaria, and other conditions that affect red cell turnover, which may render the HbA_1c inaccurate. The committee acknowledged that traditional diagnostic tests should be used in these circumstances but emphasized that glucose levels and HbA_1c reflect different aspects of glucose metabolism and should not be mixed in the diagnostic process.

The International Expert Committee recommended that individuals with an HbA_1c≥6.0 % and <6.5 % receive proven interventions. Furthermore, the committee stated that if the HbA_1c is <6.0 %, other risk factors should be considered in determining when to initiate interventions. So in essence, for both situations the measurement of glucose (or even an OGTT) is indicated to determine the kind of necessary intervention.

Also it has to be considered that for practical therapy, distinct glucose determinations and targets are essential. It will be difficult to communicate to a patient that although he or she is not considered diabetic (HbA_1c <6.5 %), he or she has to achieve stringent glucose targets. Better to tell the patient that he or she is in a prediabetic state, which will become overt diabetes if not treated.

Although I acknowledge that HbA_1c could be a convenient but rough first-line diagnostic tool to identify a diabetic state (if done with a point-of-care meter), in my clinical practice, and probably this applies to all diabetologists in Germany, glucose determinations cannot and should not be omitted. If clinically observed risk factors (obesity, hypertension, dyslipidaemia, etc.) are present at HbA_1c levels between 5.5 % and 6.0 %, only the determination of blood glucose levels (OGTT) can discriminate between normal and impaired glucose tolerance. In addition, glucose levels have to be known to the clinician in order to initiate appropriate therapeutic intervention(s).

The International Expert Committee based the diagnostic decision for diabetes solely on the occurrence of microangiopathic changes (retinopathy). In reality, the majority of diabetic patients will die of macroangiopathy or cardiovascular disease. Clinical experience clearly shows that changes in the macrovascular system precede the development of microangiopathy. Therefore, it does not make sense to take microvascular changes as a cut-off point for the diagnosis of diabetes.

Perhaps the data demonstrating the contribution of hyperglycaemia (and especially temporary glucose peaks) to the development of macrovascular diseases is not totally clear or not really conclusive, but there is abundant evidence that hyperglycemia and prediabetes or diabetes itself are linked in a causal manner to the development of cardiovascular disease. As an expert panel convened by the ADA in 2006 to focus on whether prediabetes interventions should be recommended and implemented stated, ‘Even in the absence of direct data regarding the benefits of diabetes prevention on long-term complications, the Panel believes in principle that early intervention is justified based on the following: the goal of delaying the onset of diabetes and postponing its requirements for treatment, which is often complex; the prospect of preserving β-cell function; and the likelihood that microvascular, and perhaps cardiovascular, complications will be delayed or prevented’ (Nathan et al, 2007).