Measuring kidney function without muscle mass finds loci eGFR had missed

Design

genome-wide association meta-analysis with prespecified sex-specific analyses

Population

58,976 individuals of European ancestry from the Lifelines Cohort Study

Primary outcome

genetic loci associated with creatinine clearance, a muscle-mass-independent measure of kidney function

Effect

16 independent loci and 21 lead SNPs; two new loci, rs146465192 near RP1-249F5.3 (P=3.38x10-9) and rs117014836 near AGPAT4 (P=5.42x10-9); two female-specific loci at GPC6 and IGF1R

Genome-wide association studies of kidney function have almost all used creatinine-based estimated GFR, which carries a known problem: serum creatinine reflects muscle mass as well as filtration, so any genetic variant that influences muscle influences the phenotype. This meta-analysis used creatinine clearance instead, a measure independent of muscle mass, in 58,976 people of European ancestry from the Lifelines cohort, with prespecified sex-specific analyses.

Sixteen independent loci reached genome-wide significance, carrying 21 lead variants. Two had not previously been reported in kidney function studies: rs146465192 near RP1-249F5.3, with an effect allele frequency of 0.01 and P of 3.38 x 10-9, and rs117014836 near AGPAT4, frequency 0.02 and P of 5.42 x 10-9. Both were also associated with eGFRcrea within Lifelines but had not appeared in previously published eGFR studies. In silico annotation linked the first to plasma IGF2R protein and the second to blood AGPAT4 expression. Two further loci were female-specific, at GPC6 and IGF1R.

The methodological point is more useful than any single locus. If a phenotype is contaminated by muscle mass, genetic signals that act on filtration can be diluted and signals that act on muscle can masquerade as kidney signals. Using a muscle-independent measure recovered variants that were present in the same cohort's eGFR data but had not surfaced in much larger eGFR studies.

Nothing here changes a clinic decision today. Both new variants are rare, with allele frequencies of 1% to 2%, and their effects on kidney function have not been quantified clinically. The female-specific loci are the part most worth following, since the sex difference in chronic kidney disease progression remains poorly explained and is usually attributed to hormones and lifestyle rather than to genetics.

• Remember that serum creatinine reflects muscle mass as well as filtration whenever an eGFR looks out of keeping with the patient
• Do not order genetic testing on the strength of this: the variants are rare and their clinical effects are unquantified
• Note that two loci were female-specific, which is a rare finding in kidney genetics and worth watching
• Read GWAS annotations, such as the IGF2R and AGPAT4 links here, as leads for mechanism rather than as established biology
• Recognise that this cohort is entirely of European ancestry, which limits transfer to Indian and other populations

The statistics, in plain English

P values of around 10-9 clear the conventional genome-wide threshold of 5 x 10-8, which is set that low precisely because millions of variants are tested and ordinary significance would produce thousands of false positives. Effect allele frequencies of 0.01 and 0.02 mean these are rare variants, so the associations rest on relatively few carriers and the effect estimates will be imprecise even when the P value is small. That both variants replicated within this cohort's own eGFR data but not in larger published eGFR studies is consistent with dilution by muscle mass rather than with a false positive, though independent replication in another cohort would settle it.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.