Impact of African-enriched GATM regulatory variants on creatinine-derived estimated glomerular filtration rate

Serum creatinine is a commonly used biomarker for estimating glomerular filtration rate (eGFR), but is influenced by determinants unrelated to kidney function, including muscle mass, age, sex, diet, and genetic variation. Biological drivers of persistent eGFR bias in African ancestry populations remain understudied. Here, we investigated whether African-enriched genetic variation at the glycine amidinotransferase (GATM) locus, encoding the rate limiting enzyme in creatine biosynthesis, contributes to systematic bias in creatinine-based estimates of GFR.

We analyzed whole-genome sequencing and phenotypic data from population-based cohorts in Uganda (5,760 individuals) and South Africa (827 individuals). Genome-wide association studies were integrated with expression quantitative trait locus and metabolomic analyses to characterize biological mechanisms. The iohexol measured GFR was used to quantify systematic bias for creatinine-based estimates (eGFRCr). Diagnostic implications were evaluated in African American participants of the Million Veteran Program.

We described variants associated with eGFRCr at the GATM locus that were common in African populations (allele frequency about 50%) but rare or absent in Europeans. These variants explained 1.0%-1.7% of the variance in eGFRCr and were associated with increased GATM expression and enhanced creatine synthesis, resulting in higher serum creatinine and consequently lower eGFRCr. Each allele was associated with an approximately 2 ml/min per 1.73 m² lower eGFRCr, and the association was independent of measured GFR, age, sex, and body composition. In contrast, these variants showed no association with measured GFR or cystatin C-based eGFR, supporting a biomarker-specific effect. In African American participants, GATM variants were associated with increased odds of a CKD diagnosis using eGFRCr thresholds but not with end-stage kidney disease.

African-enriched GATM variants are associated with measurable bias in creatinine-based GFR estimation without evidence of altered kidney function. These findings improve understanding of differences in kidney function biomarkers and may be relevant when interpreting creatinine-based estimates, particularly in populations where the variant is common.