Primary Open-Angle Glaucoma Polygenic Risk Score and Risk of Disease Onset: A Post Hoc Analysis

Academic Background

Primary open-angle glaucoma (POAG) is the most common form of glaucoma, often associated with elevated intraocular pressure (IOP). It is an irreversible optic neuropathy that, if untreated, can lead to vision loss. While IOP-lowering medications can delay or prevent the onset of POAG, not all individuals with elevated IOP will necessarily develop glaucoma. Thus, accurately identifying high-risk patients while avoiding overtreatment of low-risk individuals represents a critical challenge in clinical practice.

Polygenic risk scores (PRS) aggregate the cumulative effects of multiple genetic variants to evaluate an individual’s disease risk. PRS has demonstrated its potential for risk stratification across multiple diseases, particularly in identifying low-risk individuals. However, the application of PRS in POAG has not been fully explored. This study aims to evaluate the utility of PRS in identifying low-risk POAG individuals and to assess the efficacy of early intervention in high-risk individuals.

Paper Origin

This study was conducted by Sayuri Sekimitsu, Nabil Ghazal, Kanza Aziz, and others from various institutions, including the Tufts University School of Medicine, Massachusetts Eye and Ear, and Harvard Medical School. The paper was published on November 7, 2024, in JAMA Ophthalmology under the title “Primary Open-Angle Glaucoma Polygenic Risk Score and Risk of Disease Onset: A Post Hoc Analysis of a Randomized Clinical Trial.”

Study Design and Methods

Study Dataset

This study is based on data from the Ocular Hypertension Treatment Study (OHTS), a multicenter randomized clinical trial conducted at 22 centers in the United States, involving 1,636 participants with ocular hypertension. The study period ranged from February 1994 to April 2019. Of the 1,636 original participants, 1,077 had available genetic data. After excluding 67 individuals with missing data or non-European or non-African ancestry, 1,010 participants were included in the final analysis.

POAG PRS Calculation

The PRS was calculated from summary statistics derived from a cross-ancestry genome-wide association study (GWAS) for POAG. The PRS was computed using the lassosum penalized regression method and normalized within each genetically inferred ancestral cohort. The PRS was dichotomized using a threshold determined by the Youden index to maximize predictive power for POAG onset during the 20-year follow-up.

Statistical Analysis

Statistical methods included mixed-effects Cox proportional hazards models and stratified Kaplan-Meier curves to predict POAG onset and evaluate disease-free survival probabilities. Statistical analyses were performed using R software, adhering to the CONSORT reporting guidelines.

Key Results

PRS and POAG Risk

Among the 1,010 participants, 563 (65.8%) were female, with a mean age of 55.9 years. It was found that individuals with a PRS below the 48th percentile had a 1.49-times higher likelihood of remaining disease-free over 20 years of follow-up compared to the high genetic risk group (95% CI, 1.04–2.15; p = 0.03). In the highest baseline clinical risk tertile of OHTS, those with low PRS had a 20-year POAG incidence rate of 23.8% compared to 61.1% in the high-PRS group (p = 0.01).

Effect of Early Intervention

In individuals randomized to early treatment, the 10-year, 15-year, and 20-year conversion rates to POAG in the low-PRS group were 3.2%, 6.2%, and 7.1%, respectively, compared to 6.8%, 13.7%, and 15.2% in the high-PRS group. In the observation group, the conversion rates were also significantly lower in the low-PRS group than in the high-PRS group. This suggests that early treatment can partially mitigate high genetic risk, but its benefit is limited for individuals with low genetic risk.

Integration of PRS and Clinical Risk Models

The study found that combining PRS with the OHTS clinical risk model improved risk prediction. Adding PRS to the model increased its predictive ability (AUC improved from 0.67 to 0.72; p < 0.001).

Conclusions and Implications

The study supports the use of a PRS threshold for POAG to identify low-risk individuals, who exhibited lower 20-year disease conversion rates. Early treatment can partially offset high genetic risk in the highest clinical risk tier but provides limited benefits for individuals with low genetic risk. The application of PRS in clinical practice could optimize resource allocation and reduce unnecessary screening and treatment.

Study Highlights

1. Application of PRS in POAG Risk Stratification: This study is the first to demonstrate the potential of PRS in identifying low-risk POAG individuals. Combining PRS with clinical risk models significantly improves the accuracy of risk stratification.
2. Efficacy of Early Intervention: Early treatment benefits high-genetic-risk individuals significantly but has limited effects on low-genetic-risk individuals, providing a basis for personalized treatment strategies.
3. Reduction in Overdiagnosis and Overtreatment: PRS can help identify individuals who are unlikely to develop POAG, reducing unnecessary medical interventions and conserving healthcare resources.

Additional Insights and Limitations

While the study is notable for its use of a prospective dataset with a 20-year follow-up period and the inclusion of both structural and functional POAG endpoints, there are limitations. The OHTS cohort with genetic data was relatively small, and the study included only participants of European and African ancestry, limiting its generalizability to other populations. Furthermore, the diagnosis of POAG in OHTS predated the widespread use of optical coherence tomography (OCT), which is now a standard clinical tool.

Summary

This study underscores the utility of PRS in improving POAG risk prediction, particularly when combined with clinical risk models. As genetic testing becomes more accessible, PRS promises to be a valuable tool for personalized medicine, enhancing outcomes for patients while reducing the burden on healthcare systems.