Purchase this article with an account.
Marianne T Kennedy Neary, Oliver J Kennedy, George Davey Smith, Pirro G Hysi, Louis R Pasquale, Janey Wiggs, Anthony P Khawaja; Evidence for a causal link between pseudoexfoliation syndrome and myocardial infarction using Mendelian randomization. Invest. Ophthalmol. Vis. Sci. 2020;61(7):2738.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Observational studies suggest that pseudoexfoliation syndrome (PEX) is associated with a 70% increased risk of cardiovascular disease (CVD), but it remains unclear whether this association is causal. The microscopic amyloid-like protein fibers observed in PEX are visible in the eye and histologically in other structures, including coronary arteries where they may cause endothelial dysfunction. If the association between PEX and CVD is causal, this may provide insight into CVD mechanisms and prompt improved risk prediction incorporating PEX status. We aimed to investigate the causality of the PEX-CVD association using mendelian randomization (MR).
Genetic variants associated with PEX (P<1x10-6) at five genes were identified from a genome wide association study (GWAS) in a European population: FLT1-POMP (rs7329408), TMEM136 (rs11827818), RBMS3 (rs12490863), SEMA6A (rs10072088), LOXL1 (rs16958477). Using two sample MR analyses, we investigated whether PEX was causally linked to the risk of self-reported myocardial infarction among 461,880 unrelated men and women of European ancestry from the UK Biobank cohort, aged between 40-69 at baseline. The primary analysis was performed using inverse variance weighted (IVW) MR. Sensitivity analyses included Egger, weighted median and weighted mode MR analyses, which are less susceptible to confounding but less precise. Heterogeneity and horizontal pleiotropy were assessed using Cochran’s Q and Egger regression intercept, respectively.
There was evidence of a causal association between PEX and myocardial infarction (IVW odds ratio = 1.07 [1.03-1.10] per unit increase in log odds ratio of PEX; P=0.0002). The major driver for the association was LOXL1 rs16958477 (OR = 1.07 [1.04-1.11] p = 0.0002). Results from MR Egger (P=0.04), weighted median (P=0.02) and weighted mode (P=0.04) sensitivity analyses were similar in magnitude and direction to the IVW results. There was no evidence of heterogeneity or directional pleiotropy.
Our results suggest a causal relationship between PEX and myocardial infarction in Europeans. This may provide insight into the mechanisms underlying CVD and suggests that further studies examining the predictive ability of PEX status for CVD are warranted.
This is a 2020 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only