[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Research Letter
February 21, 2019

Association of Race/Ethnicity With Specific Glycemic Thresholds for Predicting Diabetic Retinopathy

Author Affiliations
  • 1Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, Florida
  • 2Department of Pediatrics, Floating Hospital for Children at Tufts Medical Center, Boston, Massachusetts
JAMA Ophthalmol. 2019;137(4):463-464. doi:10.1001/jamaophthalmol.2019.0024

Diabetic retinopathy (DR) is the leading cause of acquired blindness in the US adult population and one of the major microvascular complications of type 1/2 diabetes. Epidemiologic studies have used glycemic thresholds to suggest thresholds for predicting DR status.1 While there is extensive research on racial/ethnic differences in diabetes and on a glycemic threshold for DR in general, to our knowledge, there is limited literature on race/ethnicity–specific glycemic thresholds to predict risk for DR and DR status. Using a nationally representative sample, this study proposes race/ethnicity–specific glycemic thresholds to predict DR status among US adults.

Data (2005–2008) from the National Health and Nutrition Examination Survey were analyzed for US adults (age ≥40 years). The National Health and Nutrition Examination Survey is a nationally representative sample of the noninstitutionalized US adult population; the data are deidentified and publicly available on the Centers for Disease Control and Prevention website,2 making this study was exempt from institutional review board approval. To determine DR status, retinal images were graded by ophthalmologists using 45° nonmydriatic digital images of the retina. Race/ethnicity was self-reported and categorized as Hispanic (comprising Mexican-American and other Hispanic), white, or black. Logistic regression models were fitted to predict DR status using hemoglobin A1c (HbA1c). Receiver operator curves and the Youden index were used to determine the optimal threshold for HbA1c in predicting DR status by optimizing sensitivity, specificity, and positive and negative predictive values.