A recent article in Diabetes Care (1) found that hemoglobin A1c did not perform as well for the detection of risk for diabetes as fasting plasma glucose or the 2-h glucose on an OGTT. Do you agree or disagree? Is this the only interpretation of the data? What else might be studied?
My comments: The study group was composed of 1625 subjects from the Insulin Resistance Atherosclerosis Study (IRAS). At follow-up [after an average of 5.2 years (range 4.5– 6.6 years)], fasting plasma glucose (FPG), 2-h glucose following a 75 g glucose challenge (2-h Glu) and hemoglobin A1c were measured. In the absence of diabetes, increased risk for diabetes was defined as a FPG of 100-125 mg/dL (i.e., impaired fasting glucose; IFG), a 2-h Glu of 140-199 mg/dL (impaired glucose tolerance; IGT) or an A1c of 5.7-6.4%. Based upon this definition, IFG was present in 69% of the "at risk subjects", IGT was present in 60% of "at risk" individuals and A1c between 5.7 and 6.4% was present in only 31% of the "at risk" individuals.
A concern is that the rates of eventual conversion to full-fledged diabetes were not reported. It is possible that a borderline elevated A1c is more predictive of diabetes that either IFG or IGT. In the natural history of type 2 diabetes, it is known that IGT precedes IFG and a borderline elevation in A1c may be a later event. To this end, the authors reference the literature that an A1c of =>6.5% only recognized one third of people with diabetes. IFG, IGT and a borderline A1c do not describe equivalent stages in the transition from non-diabetic to diabetic. Is this article prematurely critical of A1c for diabetes risk detection? While A1c is more expensive than a plasma glucose measurement, it is more reproducible and does not require a fasting specimen.