Risk Factors for Proliferative Diabetic Retinopathy and Diabetic Macular Edema in the South Indian Population

Abstract

Purpose : To identify non-genetic risk factors for diabetic retinopathy (DR) in a study sample that will be used for whole exome sequencing (WES), to detect genetic variants associated with DR and diabetic macular edema (DME) in south Indian population.

Methods : Participants were patients with type 2 diabetes and DR grading who provided consent and blood for genetic studies. There were two groups of patients, the first group was recruited prospectively from Sankara Nethralaya, Chennai, India and the second group was retrospectively enrolled from SN-DREAMS. Fundus photographs and/or clinical examination were used to grade for DR severity and DME according to the International DR Severity Scale. Other covariates collected were age, sex, duration of diabetes, hemoglobin A1c (HbA1c), total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, and microalbuminuria. Stepwise logistic regression was used to determine risk factors associated with proliferative DR (PDR) and DME using Stata (StataCorp, College Station, TX). For each patient, the eye with the worst DR and DME grade, respectively, were used for the analyses.

Results : 1731 participants were included. For the PDR analysis, 95 participants with PDR were compared to 1636 individuals with no DR or non-proliferative DR. For the DME analysis, 132 participants with DME were compared to 1599 individuals without DME. In the univariate PDR analyses, presence of PDR was associated with male gender (P=0.00009), longer duration of diabetes (p=1.42×10^-7), higher HbA1c (p=3.7×10^-6), and greater microalbuminuria (p= 6.06×10^-10). In the multivariate PDR analysis, only male gender [odds ratio (OR)=5.57, p=0.03] and greater microalbuminuria (OR=1.01, p=0.002) to a statistically significant degree. In the univariate DME analyses, presence of DME was associated with male gender (p=0.0005), longer duration of diabetes (p=8.72×10^-11), and greater microalbuminuria (p=6.62×10^-8). In the multivariate DME analysis, only male gender (OR=4.67, p=0.01) and HbA1c (OR=1.26, p=0.03).

Conclusions : In south Indian population, male gender and microalbuminuria were associated with increased risk of PDR while male gender and poorer glycemic control were associated with increased risk of DME. It will be important to account for these non-genetic risk factors in the WES analyses for variants implicated in DR.

This is a 2020 ARVO Annual Meeting abstract.