To the Editor Analyzing data from the 2005-2008 National Health and Nutrition Examination Survey, Moore et al1 reported that the optimal predictive hemoglobin A1c threshold for diabetic retinopathy was higher in black and Hispanic persons (6.5% and 6.4%, respectively) than in non-Hispanic white persons (6.0%). This finding is important, but we believe it is premature to conclude that there are race/ethnicity–specific differences in the association of glycemic levels with development of retinopathy. Rather than being attributable to differences in glycemic levels, the observed differences in hemoglobin A1c threshold are likely owing to biologic differences among racial/ethnic groups in red blood cell lifespan or other factors influencing hemoglobin glycation.2 Using continuous glucose monitoring, we have shown that, for the same mean glucose concentration, mean hemoglobin A1c will be 0.4% to 0.6% higher in black individuals compared with white individuals. Thus, although the hemoglobin A1c threshold for retinopathy may differ among races/ethnicities, it is likely that the glycemic control thresholds would be similar. The study also showed that, within a racially/ethnically defined group, there is a substantial degree of variability among persons with diabetes in the association of mean glucose level and hemoglobin A1c level, which affects the ability to establish relevant thresholds for an individual based on hemoglobin A1c alone.