Kidney Week

Abstract: SA-PO792

Association of Polygenic Risk Scores and CKD in the UK Biobank

Session Information

• CKD: Epidemiology, Risk Factors, Prevention - III
  October 27, 2018 | Location: Exhibit Hall, San Diego Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: CKD (Non-Dialysis)

• 1901 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention

Authors

• Vy, Thi ha my, Icahn School of Medicine at Mount Sinai, New York, New York, United States

Thi ha my Vy,

• Chan, Lili, Icahn School of Medicine at Mount Sinai, New York, New York, United States

Lili Chan,

• Chaudhary, Kumardeep, Icahn School of Medicine at Mount Sinai, New York, New York, United States

Kumardeep Chaudhary,

• Coca, Steven G., Icahn School of Medicine at Mount Sinai, New York, New York, United States

Steven G. Coca,

• Do, Ron, Icahn School of Medicine at Mount Sinai, New York, New York, United States

Ron Do,

• Nadkarni, Girish N., Icahn School of Medicine at Mount Sinai, New York, New York, United States

Girish N. Nadkarni,

Background

Genome wide association studies (GWAS) have tested for associations between single nucleotide polymorphisms (SNPs) and chronic kidney disease (CKD). Polygenic risk scores (PRS) have shown significant promise for risk discrimination in complex disease but there are no studies on PRS in CKD.

Methods

We generated per-individual genome-wide additive polygenic risk scores (PRS) using published GWAS statistics and tested them for association with CKD in 8,432 cases and 400,529 controls of self-identified white British ancestry in the UK Biobank. We derived PRS using GWAS summary statistics from four different sets of SNP’s; (1) 33 loci significantly associated with eGFR_creatinine; (2) 53 loci significantly associated with CKD; (3) >38,000 independent SNVs using a linkage disequilibrium (LD) pruning approach; and (4)~1,800,000 SNVs with re-assigned coefficients adjusted for linkage disequilibrium (LD) by LDpred.

Results

Among the four PRS, the area under the curve (AUC) was slightly higher for PRS score 4 using ~1,800,000 SNPs (AUC=0.703) compared to PRS score 1 (0.7); 2 (0.700) or 3 (0.701). The odds ratios (OR) for CKD prevalence increased significantly at higher quintiles of genome wide PRS score 4 compared to bottom quintile (Table 1). Individuals with highest decile of PRS score had two-fold odds of CKD vs. lowest decile (OR=2.0; 97.5% Confidence Interval 1.7-2.3; p=2.28x10^-24).

Conclusion

PRS incorporating higher number of SNPs had slightly higher discrimination compared to those incorporating only GWAS significant SNPs, indicating polygenic architecture of CKD. PRS may provide modest improvements in risk prediction, especially in the extremes of genetic risk distribution.

Prevalence of CKD in higher polygenic score bins in comparison with the bottom 20% of polygenic score distribution