Hemoglobin A1c (HbA1c) is used as the cornerstone indicator of blood sugar monitoring to prevent diabetic complications and hypoglycemia. However, HbA1c has limitations including unexplained racial differences, and clinically significant mismatches between HbA1c and average glucose (AG). In the context of its limitations, improvements in convenience, cost and accuracy of continuous glucose monitoring (CGM) devices have caused a reexamination of the future role for HbA1c. However, CGM glucose data is 1) a challenge to interpret, 2) more invasive than a spot blood sample, and 3) of uncertain complications risk. Bergenstal, Beck and colleagues recently proposed a Glucose Management Indicator (GMI) that represents AG as a familiar ?calculated? HbA1c. However, the prevalence of discrepancies between measured HbA1c vs. GMI >0.5 percentage points is 28% of the GMI study reference population. As measured, HbA1c is the only prospectively evaluated measure for assessing diabetes complications risk. The challenge is to determine the mechanism of HbA1c-GMI mismatch in order to clarify whether GMI-calculated A1c is a valid AG surrogate linked faithfully to complication risk. We hypothesize that the predominant cause for mismatch between AG and HbA1c is normal variation in RBC lifespan. Using a modernized classic ?gold standard? stable isotope (SI) method, we directly measured mean RBC age MRBC (si-MRBC) in a small number of subjects and demonstrated heterogeneity in MRBC sufficient to cause 1-2% point differences (e.g. 7 vs 6 or 8%) in HbA1c. However, as the SI method is a research technique, Higgins developed semi-mechanistic models to derive MRBC from either CGM and HbA1c (cgm-MRBC) or RBC/ reticulocyte cell volume and cell Hb distributions obtained from widely available complete blood count (CBC) analyzers (cbc-MRBC). Using cgm-MRBC in 4 independent sets of subjects (age range 8-90), cgm-MRBC fully accounted for mismatches. For this R01 application, in Aim 1, we propose to perform the SI method on 70 subjects distributed equally between African-American and Caucasian subjects. si-MRBC will be repeated (n=10) to confirm reproducibility of the SI method. CGM will be performed throughout. CGM and HbA1c will be repeated in all subjects nine to twelve months after the initial evaluation to confirm temporally stable normalization. In collaboration with Dr. Bergenstal, we will determine if variations in MRBC fully account for differences between HbA1c and GMI within individuals.
In Aim 2, we will concurrently assess the validity of the Higgins semi-mechanistic models to allow more widespread application. We expect to demonstrate definitively that MRBC accounts for >80% of HbA1c/AG and HbA1c/GMI mismatches and readily available measures can routinely correct for MRBC variation. Mechanistic results from this study will put translation of risk prediction from decades of work with HbA1c on a more solid footing to this new era with rapidly evolving clinical information provided by CGM.
People with or being evaluated for the diagnosis of diabetes and their providers usually take for granted that their hemoglobin A1c (HbA1c) result faithfully reflects their blood sugar control because the amount of sugar on hemoglobin is determined by how much sugar it has been exposed to over its life time. However, new devices that can routinely measure blood sugar continuously for weeks with low cost and convenience confirm that in as great as 1 out of 3 people there is a significant mismatch between HbA1c and the blood sugar directly measured by the devices. Therefore, the purpose of this proposal is to demonstrate that these mismatches are primarily due to variation in the lifespan of the red blood cell, the cell in the blood in which Hemoglobin is located; this will personalize HbA1c allowing it to complement the new devices and be a better measure of blood sugar control improving the care of millions of people with diabetes.
