Glaucoma, Primary Open Angle

The correspondence of the determined allele and genotype frequencies to Hardy–Weinberg equilibrium (HWE) was estimated with the chi-square test. The logistic regression approach with adjustment for covariates was applied to analyze associations of the genetic variants with XFG and POAG. The three standard genetic models were assumed: additive, recessive, and dominant. The following covariates were applied: age, body mass index (BMI), systolic and diastolic blood pressures as quantitative variables and a family history of glaucoma, the presence of essential hypertension, heart atherosclerosis, heart ischemia, and diabetes mellitus (either type I or type II) as qualitative variables (Table 1). Adjustment for multiple comparisons for several SNPs was performed using the adaptive permutation test [19 (link)], and that for the number of genetic models (three) and groups compared (three), using the Bonferroni correction. Thus, the aggregated level of significance was set at p_perm<0.006 (0.05/9). Quanto 1.2.4 (Hydra 2009) was used to compute statistical power for each SNP.
The confidence intervals algorithm at D’>0.8 as implemented in HaploView v.4.2 was applied to identify haplotype blocks. Adjustment for multiple comparisons was performed using the permutation test (1,000 permutations). Taking account of the number of haplotypes studied (seven) and the number of cohorts (three), the aggregate significance level after the Bonferroni correction was set at p_perm<0.0024 (0.05/21). The computations above were performed using the respective algorithms implemented in the gPLINK v. 2.050 software.

Patients were enrolled and tested as part of the baseline assessment from the Falls in Glaucoma Study (FIGS), an observational prospective cohort of patients with glaucoma or suspect glaucoma conducted at the Wilmer Eye Institute at Johns Hopkins. Inclusion criteria for participants were: (1) age 60 and older (or turning age 60 over the course of the study), (2) glaucoma suspect or diagnosis of primary open angle glaucoma, primary angle closure glaucoma, pseudoexfoliation glaucoma, or pigmentary glaucoma, (3) residence within a 60-mile radius from the Wilmer Eye Institute, and (4) ability to perform VF testing. Exclusion criteria for participants were: (1) presence of visually significant concurrent eye disease, (2) any ocular surgery in the past 2 months, (3) any hospitalization in the past month, and (4) confinement to a bed or wheelchair, (5) history of stroke or other neurological disorders causing VF damage.^{23 (link)}

We classified POAG cases and controls with a previously published algorithm developed in the VA (17 ) and applied to the MVP as previously described (13 ). Ancestry groups were defined using the Harmonized Ancestry and Race/Ethnicity (HARE) algorithm (18 (link)), which classifies an individual’s HARE group based on the correspondence of their self-identified race/ethnicity and genetically inferred ancestry.

HIS Veterans are more genetically admixed than EUR Veterans (18 (link)); thus, we evaluated GRS performance in a subset of HIS Veterans with less than 50% EUR global genetic ancestry (GGA) as determined via the ADMIXTURE software program (20 (link)). We compared these subset results to the full MVP HIS POAG case-control dataset.