HbA1C Variability as an Independent Predictor of Diabetic Retinopathy: A 10-Year Retrospective Nested Case-Control...

Background Diabetic retinopathy (DR) is a leading cause of vision impairment worldwide and a major microvascular complication of type 2 diabetes mellitus. Although chronic hyperglycemia is a well-established risk factor, evidence is accumulating that long-term glycemic variability may independently contribute to DR development. Data from Middle Eastern and Gulf Cooperation Council populations remain scarce. Methods We conducted a retrospective case-control study nested within a 10-year diabetes cohort at the Dubai Diabetes Center, a specialized tertiary facility in the United Arab Emirates. Cases (n = 67) who developed DR during follow-up were matched 1:1 by age to DR-free controls (n = 67). Long-term glycemic variability was assessed using the standard deviation (SD) and coefficient of variation (CV) of serial HbA1c measurements collected twice yearly. Five multiple logistic regression models evaluated the independent association between HbA1c variability and DR, adjusting for diabetes duration, baseline HbA1c, neuropathy, smoking, and insulin use. Findings Patients who developed DR had significantly higher HbA1c variability than controls (median SD 1.00% vs 0.55%, p < 0.001; median CV 12.34% vs. 7.93%, p < 0.001). In multivariate analyses, HbA1c variability remained independently associated with DR across all five models, whether expressed as SD (adjusted OR up to 13.85 per 1% increase) or CV (adjusted OR 1.20 per 1% increase), after adjustment for follow-up mean HbA1c. Neuropathy, smoking, and insulin use were also independently associated with DR. ROC analysis showed acceptable discriminatory performance for both SD (AUC 0.746) and CV (AUC 0.716). Interpretation Long-term HbA1c variability was independently associated with DR in this Middle Eastern cohort, beyond mean glycemic control. These findings suggest that serial HbA1c measurements - already collected in routine diabetes care - may provide a scalable, low-cost tool for early DR risk stratification. Prospective validation in larger, diverse populations is needed before these findings can be incorporated into clinical algorithms.