Comment on Nathan et al. Relationship Between Average Glucose Levels and HbA_1c Differs Across Racial Groups: A Substudy of the GRADE Randomized Trial. Diabetes Care 2024;47:2155–2163

I read the article by Nathan et al. (1) with great interest. This is an important topic, and the study confirms the long-standing finding that individuals who self-identify as Black or African American have higher HbA_1c, on average, than individuals who self-identify as White. The study also replicates previous work in type 1 diabetes using up to 12 weeks of continuous glucose monitoring, which found that self-identified Black individuals had higher average HbA_1c levels at any given average glucose level than those who self-identified as White (2). Nathan et al. suggest that racial differences in HbA_1c should be considered in making treatment decisions and setting glycemic goals in diabetes. This conclusion is troubling.

Race is a social construct used historically to marginalize certain groups of people. Racial disparities in the burden of diabetes and its complications in the U.S. are primarily driven by historical factors, including slavery, segregation, Jim Crow laws, redlining and other racist policies, environmental exposures such as differences in the built environment and food availability, and other social determinants of health and health care.

Nathan et al. (1) state that the fact that the relationship between average glucose levels and HbA_1c differs across racial groups “. . . should be considered in setting treatment goals and diagnostic levels.” The small racial differences we see in HbA_1c independent of average glucose are likely driven wholly or in part by genetic variants that also happen to differ by race (3,4). Using race to adjust HbA_1c could help line up HbA_1c with mean glucose, on average, in the population. However, because race does not reliably reflect individual genetic or other biological information, using race or a race-based adjustment to guide treatment can result in substantial misclassification of individual patients. If the concern is that genetic variants may alter HbA_1c, we need to quantify those genetic determinants. We should not rely on race to address genetic differences.

If we treat people based on their skin color and not direct determinants of their health, some patients could be denied appropriate treatment, and some could be managed too aggressively. The recommendations by Nathan et al. (1) have the potential to worsen, not ameliorate, long-standing racial and ethnic disparities in diabetes.

In the future, we may be able to use genetic information to refine and improve the accuracy ofHbA_1c as a measure of hyperglycemia (5). In the meantime, if there is a concern about the accuracy of an HbA_1c test result, we should use it in combination with other clinical information, especially information concerning glucose, to make the best decisions for individual patients.