Abstract MP19: An eGFR Polygenic Score Predicts Chronic Kidney Disease in African Americans

Abstract

Chronic kidney disease (CKD) is a risk factor for cardiovascular disease and early death. Genetic factors contribute to CKD, and recently, polygenic scores (PGS) have been developed to quantify risk for complex diseases, such as CKD. However, African ancestry populations are underrepresented in both CKD genetic studies and PGS development overall; moreover, European-ancestry derived PGSs demonstrate diminished predictive performance in African ancestry populations. This study aimed to develop a PGS for CKD using genotype and phenotype data from African American (AA) participants of observational cohort studies. We obtained score weights from a meta-analysis of genome-wide association studies (GWAS) for estimated glomerular filtration rate (eGFR) in the Million Veteran Program (MVP) and Reasons for Geographical and Racial Differences in Stroke (REGARDS) Study (total n~66,000). We then optimized the PGS in a cohort of AAs from the Hypertension Genetic Epidemiology Network (HyperGEN) Study (n~1,900) using the PRS-CS software and evaluated the predictive performance of the PGS at multiple global shrinkage parameters. We further adjusted the PGS for APOL1 risk status. In HyperGEN, the eGFR-based PGS was significantly associated with the odds of prevalent CKD—defined as baseline eGFR <60 mL/min/1.73m 2 — in logistic regression models adjusted for age, sex, and population structure. Further, accounting for APOL1 risk status—a putative variant for CKD unique to those of sub-Saharan African descent—improved the score’s accuracy, with the APOL1 -adjusted PGS explaining 1.9% (1.1% without APOL1 ) of the variance in CKD and an area under the curve (AUC) of 58.9% [95% CI: 53.0%-64.9%] (without APOL1 , 58.2% [95% CI: 52.3%-64.1%]). Sensitivity analyses and validation in external cohorts, as well as comparisons to previously published PGS, are ongoing. In this study, we developed a PGS that was significantly associated with CKD with improved predictive accuracy in AAs over previously published PGS.