Response to Comment on: Boronat et al. Differences in Cardiovascular Risk Profile of Diabetic Subjects Discordantly Classified by Diagnostic Criteria Based on Glycated Hemoglobin and Oral Glucose Tolerance Test. Diabetes Care 2010;33:2671–2673 Open Access

We appreciate the efforts of Kramer and Araneta (1) to investigate whether our findings are reproducible in the Rancho Bernardo cohort. In apparent contrast to our results (2), they observed that subjects with diabetic oral glucose tolerance test (OGTT) but A1C <6.5% displayed a more unfavorable cardiovascular risk profile than subjects with A1C ≥6.5% (1). However, as noted by the authors, groups used for comparison in both studies were not homogeneous. All analyzed subjects with A1C ≥6.5% in the Rancho Bernardo study had nondiabetic OGTT, whereas in our study 86% of subjects with A1C ≥6.5% also fulfilled criteria of diabetes based on plasma glucose. Probably, the cardiovascular risk profile is worse among subjects meeting both criteria of diabetes than among those that only fulfill the A1C-based criterion. However, according to the greatest studies performed in Western populations (3), the 6.5% cut-off point for A1C-based diagnosis of diabetes is highly specific, i.e., most individuals with A1C ≥6.5% also are diagnosed with diabetes following criteria based on plasma glucose. In this regard, the Rancho Bernardo study, in which 85% of participants with A1C ≥6.5% were classified as nondiabetic by OGTT (4), seems to be an exception. The authors attribute these findings to the advanced age of the studied population. Methodological issues related to the laboratory assay of A1C might also have contributed to overdiagnosis of diabetes in the Rancho Bernardo study. In this survey, A1C was assayed between 1984 and 1987, before the standardization of A1C measurement. As far as we know, the authors have not communicated alignment of their A1C values to the Diabetes Control and Complications Trial (DCCT) assay. The diagnostic cut-off point of 6.5% for the A1C test only should be applied using a method that is standardized or traceable to the DCCT reference. Therefore, if this was not the case, interpretation of the characteristics of subjects diagnosed of diabetes by A1C test in the Rancho Bernardo study should be taken with caution.