The causal effect of obesity on diabetic retinopathy: A two-sample Mendelian randomization study

Background: The causal effect of obesity on diabetic retinopathy (DR) remains controversial. The aim of this study was to assess the causal association of generalized obesity evaluated by body mass index (BMI) and abdominal obesity evaluated by waist or hip circumference with DR, background DR, and proliferative DR using a two-sample Mendelian randomization (MR) analysis.

Methods: Genetic variants associated with obesity at the genome-wide significance (P<5×10−8) level were derived using GWAS summary statistics from the UK Biobank (UKB) with a sample size of 461 460 individuals for BMI, 462 166 individuals for waist circumference, and 462 117 individuals for hip circumference. We obtained genetic predictors of DR (14 584 cases and 202 082 controls), background DR (2026 cases and 204 208 controls), and proliferative DR (8681 cases and 204 208 controls) from FinnGen. Univariable and multivariable Mendelian randomization analyses were conducted. Inverse variance weighted (IVW) was the main method used to analyze causality, accompanied by several sensitivity MR analyses.

Results: Genetically predicted increased BMI [OR=1.239; 95% CI=(1.134, 1.353);P=1.94×10-06], waist circumference [OR=1.402; 95% CI=(1.242, 1.584); P=5.12×10-08], and hip circumference [OR=1.107; 95% CI=(1.003, 1.221); P=0.042] were associated with increased risk of DR. BMI [OR=1.625; 95% CI=(1.285, 2.057); P=5.24×10-05], waist circumference [OR=2.085; 95% CI=(1.54, 2.823); P=2.01×10-06], and hip circumference [OR=1.394; 95% CI=(1.085, 1.791); P=0.009] were correlated with the risk of background DR. MR analysis also supported a causal association between BMI [OR=1.401; 95% CI=(1.247, 1.575); P=1.46×10-08], waist circumference [OR=1.696; 95% CI=(1.455, 1.977); P=1.47×10-11], and hip circumference [OR=1.221; 95% CI=(1.076, 1.385); P=0.002] and proliferative DR. The association of obesity with DR continued to be significant after adjustment for type 2 diabetes.

Conclusion: This study using two-sample MR analysis indicated that generalized obesity and abdominal obesity might increase the risk of any DR. These results suggested that controlling obesity may be effective in DR development.