Genetic correlations between intraocular pressure, blood pressure and primary open-angle glaucoma: a multi-cohort analysis

Abstract

Primary open-angle glaucoma (POAG) is the most common chronic optic neuropathy worldwide. Epidemiological studies show a robust positive relation between intraocular pressure (IOP) and POAG and modest positive association between IOP and blood pressure (BP), while the relation between BP and POAG is controversial. The International Glaucoma Genetics Consortium (n=27 558), the International Consortium on Blood Pressure (n=69 395), and the National Eye Institute Glaucoma Human Genetics Collaboration Heritable Overall Operational Database (n=37 333), represent genome-wide data sets for IOP, BP traits and POAG, respectively. We formed genome-wide significant variant panels for IOP and diastolic BP and found a strong relation with POAG (odds ratio and 95% confidence interval: 1.18 (1.14-1.21), P=1.8 × 10) for the former trait but no association for the latter (P=0.93). Next, we used linkage disequilibrium (LD) score regression, to provide genome-wide estimates of correlation between traits without the need for additional phenotyping. We also compared our genome-wide estimate of heritability between IOP and BP to an estimate based solely on direct measures of these traits in the Erasmus Rucphen Family (ERF; n=2519) study using Sequential Oligogenic Linkage Analysis Routines (SOLAR). LD score regression revealed high genetic correlation between IOP and POAG (48.5%, P=2.1 × 10); however, genetic correlation between IOP and diastolic BP (P=0.86) and between diastolic BP and POAG (P=0.42) were negligible. Using SOLAR in the ERF study, we confirmed the minimal heritability between IOP and diastolic BP (P=0.63). Overall, IOP shares genetic basis with POAG, whereas BP has limited shared genetic correlation with IOP or POAG.

Calcium Channel Blockers and Risk of Primary Open-angle Glaucoma

“…3 Aschard and colleagues examined the relationship between BP, IOP and POAG using genotypic data; they found no evident genetic correlation between BP and POAG which strongly contrasted with the high genetic correlation they observed, as expected, between IOP and POAG. 4 This suggests that the observed phenotypic relationship between BP and POAG is not causal, and supports the hypothesis that the observed association is due to confounding by antihypertensive treatment.…”

Genetic Correlations Between Diabetes and Glaucoma: An Analysis of Continuous and Dichotomous Phenotypes

“…For example, we estimated a strong inverse relation between HDL and T2D (r g = -0.40; SE = 0.06; p = 4.2E-11; Supplemental Table 4), and a strong positive genetic correlation between IOP measured in various ways and POAG, as previously reported. 24 Weaknesses include the fact that the study was limited to European populations although some, but not all, data from Japan are consistent with our findings. 26 Second, the absence of a statistically significant genetic correlation does not rule out that a minority of genes are truly shared between diabetes-and glaucoma-related traits.…”

“…ACCEPTED MANUSCRIPT p = 3.0E-04). 24 Similarly there were strong positive genetic correlations between IOPcc and POAG (r g = 0.50; SE = 0.09; p = 5.5E-08) and between IOPg and POAG (r g = 0.60; SE = 0.15; p = 4.3E-05). None of the corneal features (CCT, CH or CRF) showed significant genetic correlation with CDR (p≥0.13) or POAG (p≥0.07).…”

“…66 This study has strengths and weaknesses. Strengths include the use of LD score regression, a novel unbiased approach, to assess correlations between many traits where strong positive associations are suspected such as IOP and POAG 24 and others where there is controversy such as T2D and POAG. 25,26 Furthermore, our genetic correlation analysis between diabetes and glaucoma was extensive as we considered nine diabetes-and eight glaucoma-related traits.…”

“…23 For glaucoma-related traits, a strong genetic correlation between IOP measured with Goldmann applanation tonometry and POAG was reported using GWAS summary data from two large European-derived consortia. 24 Using LD score regression in a Japanese population, Shiga et al 25 found a positive genetic correlation between T2D and POAG (r g =0.27; p=2.00E-04) but Kinai et al found no significant correlations between various quantitative diabetes traits and POAG in the same population. 26 Another approach is to form panels of genome-wide significant markers for a trait and test them in relation to another trait of interest.…”