Glaucoma polygenic risk score predicts treatment intensity and RNFL loss in glaucoma suspects

Abstract

Purpose : Assess the clinical utility of a glaucoma polygenic risk score (PRS) to predict disease progression in glaucoma suspects

Methods : We developed a glaucoma polygenic risk score (PRS) by characterising glaucoma endophenotypes on 67,040 UK Biobank (UKBB) participants and performing a genome-wide association study (GWAS). The vertical cup to disc ratio (VCDR) and vertical disc diameter (VDD) were measured from digital retinal photographs. These endophenotypes were then combined with GWAS data on intraocular pressure (IOP) and glaucoma, using multi-trait analysis of GWAS (MTAG). We used this PRS for prediction in a prospectively monitored cohort of glaucoma suspects examined with 6 monthly visits and OCT obtained using Cirrus HD-OCT from the PROGRESSA study. We used the PRS to predict the structural thinning of peripapillary retinal nerve fibre layer (RNFL) and intensity of medical therapy — as measured by the number of topical glaucoma drops. A linear mixed effects model was used to analyse OCT data accounting for correlated data from both eyes, and the ordinary least squared linear regression was used for treatment intensity (maximal treatment per patient as the outcome variable).

Results : In total, 467 glaucoma suspects had a mean duration of 3.6 ± 1.7 years of clinical follow-up after the baseline visit, with 5.01 ± 2.04 years of OCT data. The glaucoma PRS was strongly associated with the intensity of treatment in glaucoma suspects measured as the number of topical glaucoma medications or selective laser trabeculoplasty (SLT) procedures, even after adjustment for the maximum recorded IOP and age of the patient (P = 0.0036). There was genetic dose-response relationship where the first decile of PRS (lowest risk) were uncommonly treated (mean number of medications = 0.2; 95% CI 0.02 – 0.38), compared to the tenth decile (highest risk; mean number of medications = 0.94; 95% CI 0.69 – 1.19). Furthermore, the PRS was associated with progressive peripapillary RNFL thinning (estimated effect per decile of PRS -0.29; P = 0.033). This association remains significant after adjustment for IOP and the age at baseline (P = 0.026).

Conclusions : This comprehensive glaucoma PRS is associated with structural progression in glaucoma suspects, and is predictive of the intensity of medical therapy in a prospectively monitored patient cohort even after correction for known risk factors age and IOP.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.