Intraocular pressure, primary open-angle glaucoma and the risk of retinal vein occlusion: A Mendelian randomization mediation analysis

Abstract

Background: The etiological connection between intraocular pressure (IOP) and the risk of retinal vein occlusion (RVO) remains elusive, particularly regarding whether this risk emanates from the direct influence of elevated intraocular pressure (IOP), irrespective of the presence of primary open-angle glaucoma (POAG), or if it arises as a consequence of the sequelae of POAG. Therefore, we conducted a Mendelian Randomization (MR) mediation analysis to elucidate the mediating role of POAG in the association between IOP and RVO.

Methods: We identified 47 single-nucleotide polymorphisms (SNPs) associated with IOP (P-value < 5 × 10^-8) leveraging data from a genome-wide association study (GWAS) (N = 97,653) obtained from the UK Biobank and 50 SNPs associated with POAG (P-value < 5 × 10^-8) from a GWAS meta-analysis (16,677 cases and 199,580 controls). We related these SNPs with RVO using a GWAS of 775 RVO cases and 376,502 controls from FinnGen. By utilizing univariable and multivariable MR analyses we calculated the total effect of IOP on RVO and estimated the degree to which POAG mediates this association.

Results: MR analyses showed that higher IOP is associated with higher RVO risk (odds ratio of RVO per 1 mmHg increase in IOP: 1.53; 95% confidence interval: 1.04 to 2.26; p-value = 0.03). Moreover, our MR mediation analysis suggested that 91.6% of the total effect of IOP on RVO risk was mediated through POAG. The primary results were consistent with estimates of pleiotropy-robust MR methods.

Conclusion: Our findings suggest that higher IOP increases the risk of RVO and that the majority of this effect is mediated through POAG.

Fig. 1. Directed acyclic graphs of the mediation analysis with Mendelian randomization *.

* The indirect effect of intraocular pressure (IOP) on retinal vein occlusion (RVO) can be calculated by multiplying α times β, where α is the effect of IOP on primary open-angle glaucoma (POAG), and β the effect of POAG on RVO. The proportion mediated can be estimated by dividing the indirect effect by the total effect of IOP on RVO. All estimates are shown as the difference in the logarithm of odds of the outcome, per 1 unit increase of the exposure (continuous variables: IOP) or between the two exposure groups (binary variables: POAG).

Fig. 2. Mendelian randomization estimates for the effect of intraocular pressure on retinal vein occlusion.

Estimates are reported as changes in odds of retinal vein occlusion per 1 mmHg increase in intraocular pressure *. * SNP single nucleotide polymorphism, CI confidence interval, MR-PRESSO Mendelian randomization pleiotropy residual sum, and outlier.