Mendelian randomization analysis of causal relationship between cheese intake and diabetic retinopathy.

Aim: To assess whether there is a possible causal link between the intake of cheese and the risk of diabetic retinopathy (DR) utilizing a two-sample Mendelian randomization (MR) analysis.

Methods: The research data were obtained from summary statistics of genome-wide association studies (GWAS). Genetic loci closely related to cheese intake were extracted as instrumental variables (IVs), and DR was the outcome variable. The data were extracted from individuals of European ethnicity. The data of cheese intake consisted of 451 486 samples with 9 851 867 single nucleotide polymorphisms (SNPs), while the DR data consisted of 206 234 samples with 16 380 446 SNPs. Sixty-one genetic loci closely related to cheese intake were selected as IVs. MR analysis was performed by inverse-variance weighted (IVW) method and MR-Egger regression respectively. The causal relationship between cheese intake and DR was evaluated using odds ratios (ORs) and 95% confidence intervals (CIs). Egger-intercept test was used to test horizontal pleiotropy and sensitivity analysis was performed by leave-one-out test.

Results: The P value of the IVW method was less than 0.05, indicating a significant negative correlation between cheese intake and DR. MR-Egger regression showed that the intercept was 0.01 with a standard error of 0.022, and a P-value of 0.634, indicating no evidence of horizontal pleiotropy affecting the IVs related to the exposure factors. Besides, heterogeneity tests confirmed the absence of heterogeneity, and the "leave-one-out" sensitivity analysis demonstrated that the results were stable.

Conclusion: Cheese intake is causally negatively correlated with the occurrence of DR, and cheese intake could reduce the risk of DR.