Abstract
Keratoconus (KC) is an intricate disease involving multiple factors. It is widely accepted that the pathogenesis of KC is linked to oxidative stress (OS). Nevertheless, the precise causal relationship between OS and KC remains uncertain due to the presence of confounding factors. This study aims to ascertain whether there was a potential causal effect between OS and KC. A bidirectional 2-sample Mendelian randomization (MR) analysis was conducted, utilizing genetic instrumental variables as substitutes for 12 oxidative stress injury biomarkers (OSIBs). The available summarized data for OSIBs were obtained through the published genome-wide association study. Data for KC was collected from the FinnGen cohort, comprising 311 cases and 209,287 controls of European population. The primary MR analysis employed the inverse variance weighted (IVW) method. To evaluate the reliability of the observed associations, sensitivity analysis and reverse MR analysis were performed. The MR analysis revealed significant associations. The IVW method suggested that elevated genetically predicted glutathione peroxidase levels were correlated with the diminished KC risk (odds ratio [OR]: 0.660, 95% confidence interval [CI]: 0.500-0.873, P = .004). Total bilirubin was also found to be associated with a decreased risk of KC through the IVW method (OR: 0.912, 95% CI: 0.845-0.984, P = .017), which was consistent with the result from the MR-Egger method (OR: 0.910, 95% CI: 0.835-0.992, P = .034). Reverse MR analysis did not suggest causal relationship of KC on OSIBs. These findings provide a robust support for the causal link between OSIBs and the development of KC, indicating that targeting OS pathways may become a potential therapeutic approach for KC.