Haemoglobin A1c: not yet ready for diabetes diagnosis

R Recommendations to use haemoglobin A1c as a sole laboratory test to diagnose type 2 diabetes are premature, according to international experts including Australia’s Professor Paul Zimmet. Writing in the BMJ, they noted that a committee representing the American Diabetes Association, the European Association for the Study of Diabetes and the International Diabetes Federation had suggested an HbA1c of 6.5% or more should be sufficient to define the presence of diabetes without the need for a fasting blood glucose level or oral glucose tolerance test. Superficially, the plan was appealing, the doctors said. There were undoubted benefits in using a single test that did not require fasting, and HbA1c was less variable than blood glucose levels. It also provided a more complete picture of glycaemia, which its proponents claimed was a “common sense” justification for its use. However, there were numerous difficulties with HbA1c. Results were influenced by the presence of haemoglobin disorders, which occurred in at least 10% of African Americans and had a high prevalence in people of Mediterranean and South East Asian descent. The level could also be affected by iron deficiency anaemia, renal failure, medications including antiretrovirals, and advancing age. HbA1c and conventional blood glucose levels could potentially identify different groups of patients as having diabetes, and there was no strong evidence on the value of glycated haemoglobin levels in predicting microvascular complications, Zimmet and colleagues said. Like a fasting blood glucose level of 7 mmol/l, the proposed HbA1c threshold of 6.5% for a diagnosis of diabetes was arbitrary and needed further discussion. BMJ 2009, published online....