THE HEMOGLOBIN A1C (HBA1C) ASSAY WAS FIRST USED more than 30 years ago as a marker of diabetes control. Despite its availability for 2 decades, HbA1c testing did not assume a central role in diabetes care until the publication of 2 major clinical trials. The Diabetes Control and Complications Trial (DCCT) and the UK Prospective Diabetes Study (UKPDS) demonstrated that HbA1c level strongly predicts risk of microvascular complications associated with type 1 and type 2 diabetes, respectively. Largely consequent to these findings, HbA1c level has become the primary basis for diabetes diagnosis, treatment decisions, and assessment of quality health care. As clinical application of HbA1c measurement expands, so does the need for caution in its interpretation and appreciation of its substantial limitations. The unreliability of HbA1c findings has been recognized in a number of clinical settings. Among the most frequently cited are hematologic conditions, including anemia, hemolysis, variable red blood cell life span, reticulocytosis, and hemoglobinopathies. A variety of less appreciated clinical factors can affect the concordance of HbA1c level with mean blood glucose (MBG) level. These include the use of certain medications (dapsone, erythropoietin), mechanical heart valves, and hypothyroidism. Differential glycation rates also can produce variability in HbA1c values, and rates of protein glycation appear to vary to a much greater extent across individuals than historically has been appreciated. When HbA1c values are compared systematically with MBG values generated by continued glucose monitoring, the range of corresponding MBG for an HbA1c of 8% spans 70 mg/dL; this variability increases even further at higher HbA1c values and may derive substantially from heterogeneity in glycation rates. In a pediatric population of patients with type 1 diabetes, roughly equivalent MBG values corresponded to an HbA1c level of 7.6% among “low glycators” and 9.6% among “high glycators.” These data derive from a study population that simply was divided into tertiles on the basis of differential glycation. Further, differential glycation appears not only a source of HbA1c measurement variability but possibly an independent predictor of diabetes-related microvascular complications. Cohen et al reported that the glycation gap—the difference between glycated fructosamine and glycated hemoglobin—is associated with risk of incident retinopathy in a diabetic population. Differential glycation may have a substantial hereditary basis. The hemoglobin glycation index also has been proposed as an independent predictor of microvascular complications, although this has been refuted on the basis of the strong correlation between HbA1c values and the hemoglobin glycation index. More recently, greater recognition has evolved of the effects of race on HbA1c level, with particular focus on the implications for new diagnosis of diabetes. For comparable MBG values, HbA1c values vary by 0.2% to 0.65% across racial backgrounds. These findings suggest that an HbA1c level of 6.5% may be too sensitive a threshold for diagnosis in some populations and an inadequate threshold in others. The DCCT did not address the possibility that risk of microvascular complications may be conferred at variant HbA1c values across race; it included only patients with type 1 diabetes, a racially homogeneous population. This racial bias was skewed further by the exclusion of study participants with an abnormal hemoglobin electrophoresis and calls into question the fundamental applicability of the DCCT findings to nonwhite patients. The UKPDS similarly lacked racial heterogeneity, as 81% of study participants were white, whereas only 10% were Asian and 8% African Caribbean. The continual identification of sources of variability in HbA1c level has generated interest in the use of alternative, integrative measures of MBG. Expanded use of either fructosamine or glycated albumin, for example, has been advocated for the diagnosis of diabetes and treatment monitoring. Fructosamine offers the advantage of gauging shorterterm glycemic control; however, fructosamine does not correlate strongly with fasting plasma glucose and is a rela-
Search from 250 M+ research papersSearch from 250 M+ research papers
Open Access
Nov 11, 2014
Jesús López-Torres + 5
