Abstract: FR-OR036

Effect Modification Detected in a Genome-Wide Comparison of Kidney Function Variants between Over 270,000 Diabetic and Non-Diabetic Participants: The Million Veterans Program

Session Information

Category: Genetic Diseases of the Kidney

  • 803 Genetic Epidemiology and Other Genetic Studies of Common Kidney Diseases

Authors

  • Hung, Adriana, VA TVHS Nashville, Nashville, Tennessee, United States
  • Roumie, Christianne, VA TVHS Nashville, Nashville, Tennessee, United States
  • O'Donnell, Christopher Joseph, Boston Veterans Administration, Boston, Massachusetts, United States
  • Edwards, Todd L., Vanderbilt University Medical Center, Nashville, Tennessee, United States
  • Giri, Ayush, Vanderbilt University, Nashville, Tennessee, United States
  • Velez edwards, Digna R, Vanderbilt University, Nashville, Tennessee, United States
  • Hellwege, Jacklyn N, Vanderbilt University, Nashville, Tennessee, United States
  • Wilson, Otis D, Vanderbilt University, Nashville, Tennessee, United States
  • Torstenson, Eric S, Vanderbilt University, Nashville, Tennessee, United States
  • Kovesdy, Csaba P., University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Birdwell, Kelly A., Vanderbilt University, Nashville, Tennessee, United States
  • Siew, Edward D., Vanderbilt University, Nashville, Tennessee, United States

Group or Team Name

  • On behalf of VA Million Veteran Program
Background

Diabetes is the number one cause of end stage kidney disease worldwide and diabetic kidney disease is often thought to be distinct from non-diabetic kidney disease. To investigate the extent of shared genetic etiology between these two subgroups in relation to kidney function

Methods

We conducted two genome-wide association studies (GWAS) in the Million Veteran Program (MVP) cohort to detect associations with eGFR: in self-reported black and white, non-diabetic veterans (N = 181,315) and diabetic veterans (N = 91,523).

Results

In total, 1,991 SNPs at 36 loci reached genome-wide significance (p-value <5x10-8) in the diabetic subgroup, of which 15 were novel loci. UMOD represented the most striking signal in diabetics (p-value = 2.43x10-82). In the non-diabetic group, 7,357 SNPs at 101 loci, of which 52 were novel loci, reached genome-wide significance, with SPATA5L1 having the strongest association (p-value = 6.39x10-106). Notably, 24 genome-wide significant loci were shared between diabetics and non-diabetics, including 5 novel loci: TPPP, NRIP1, MAS1L, HFE, and C15orf54. Comparison of effect estimates for significant loci showed a positive correlation between estimates across diabetics and nondiabetics (R2 = 0.69). Correlations were strongest for known loci SPATA5L1 and UMOD (r2=0.93 and 0.98, respectively). Differences in effect sizes between diabetics and non-diabetics were also observed at two distinct loci: MAF and C2. Index SNP rs11644758 near MAF was associated with eGFR in non-diabetics (beta: 0.018; p-value = 1.1x10-9), but not in diabetics (beta = -0.003; p-value = 0.54), while index SNP rs644045 in C2 was only associated in diabetics

Conclusion


In the largest comparative investigation of GWAS loci for eGFR between diabetics and non-diabetics known to date, this study reports numerous known and novel loci for eGFR, with mostly shared, and a few distinct loci across the two groups.

Funding

  • Veterans Affairs Support