Abstract
Purpose :
Previous studies have reported an increased risk for open-angle glaucoma (OAG) in patients with a thinner central corneal thickness (CCT). However, due to suspected collider bias with intraocular pressure (IOP), this relation remains disputed. Similarly, the co-occurrence of Fuchs’ endothelial corneal dystrophy (FECD) and OAG has been reported previously, but it is unclear if there is any true causal effect. We combine classical association analyses with Mendelian Randomization experiments to clarify the relation between these traits.
Methods :
Utilizing data from the Rotterdam Study, a prospective population-based cohort study, we tested the effect of CCT, a CCT genetic risk score (GRS), and a FECD GRS on OAG risk and IOP. The GRS were calculated using estimates from the latest meta-GWAS for CCT and FECD, respectively. We used multivariable Cox proportional hazard and linear regression models for OAG and IOP, respectively, adjusted for age and sex. Finally, CCT and the GRS were stratified by quartile to evaluate non-linear effects, using the lowest quartile as the reference group.
Results :
A total of 303 participants with OAG and 10598 without OAG were included in our analyses.
Increased CCT was associated with lower risk of OAG, with a HR (95%) of 0.71 (0.57, 0.90) per standard deviation (SD) increase in CCT (p = 0.003), and highest vs. lowest quartile HR (95% CI) of 0.48 (0.24, 0.96). In contrast, we did not find a significant association between the CCT GRS and OAG, with a HR (95%) for one SD increase in the CCT GRS of 1.01 (0.87, 1.17; p = 0.944). Both measured CCT and the CCT GRS were associated with IOP, with a mean difference (95% CI) per SD of 0.53 (0.43, 0.64) and 0.22 (0.14, 0.29) mmHg, respectively. We found no significant association between the FECD GRS and OAG (p = 0.946). One SD increase in the FECD GRS was, however, associated with a reduction of -0.16 (-0.27, -0.06) mmHg in IOP (p < 0.001).
Conclusions :
Increased measured CCT is strongly associated with reduced OAG risk, but genetically predicted CCT is not. This suggests that the previously reported relation between CCT and OAG is most likely a spurious association resulting from collider bias and not a true causal effect. Genetically predicted FECD was not associated with OAG. Although CCT could still be a useful parameter for the purpose of prediction, neither CCT nor FECD should be considered a causal determinant for OAG in future research.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.