Genome-wide association study of serum creatinine as a surrogate of muscle size

Abstract

The present study aimed to find out whether age and sex affect serum creatinine (sCr) levels as well as to discover the relationship between sCr and muscle size, and finally to search for genetic variants associating with serum sCr. The present study is involved in a larger study setting, in which biomarkers and genetic data were collected by The Finnish Institute for Health and Welfare via Finrisk cohort studies of the years 1997, 2002, 2007 and 2012. The omics data used consisted of sCr, age and sex variables and other additive information. The most important variables in the present study were sCr concentration, age, sex, fat-free mass (FFM) and/or lean mass (LM) by bioimpedance analysis (BIA), and beta values determined for the four genome-widely associative single-nucleotide polymorphisms (SNPs). Preparation and analyses of the data were mostly performed by using R (Studio version for Windows) and MS Excel. Glomerular filtration rates (eGFR) were estimated for the whole study population in the raw data to exclude individuals with decreased renal function based on minimum value of 60 ml/min/1.73 m2 (eGFR60). In addition to eGFR60 group (n = 21 067), cut-off groups eGFR70, eGFR80 and eGFR90 were formed. To determine the effects of age, sex and FFM on sCr, several statistical comparisons were performed with different settings and finally, genome-widely associative SNPs were tested via beta values that represent the relative effects of the variants and genetic risk scores (GRSs) calculated based on the beta values. In the present study, FFM correlated significantly with sCr (r = 0.38, p < 0.001). Statistically significant increase was observed in sCr levels during aging (p < 0.001) in the cut-off eGFR60 (mean 74.44 ± 12.75 μmol/l) which can be explained by age-dependent decline in renal function. Instead, the cut-offs eGFR80 and eGFR90 showed statistically significant decrease in sCr levels related to aging (p < 0.001) and this may represent muscle loss. The cut-off eGFR70 can be considered as a plateau group of age comparison regarding sCr since no significant difference was observed while comparing youth and older participants. In the GWAS analysis, four suggestively or strongly genome-widely associative loci were discovered of which PRKAG2/RHEB locus showed the strongest associations in all (p = 1.36 x 10-11) as well as in under 50 (p = 1.10 x 10-7) and over 50 years old (p = 7.28 x 10-5) participants. However, weaker associations were shown among older compared to younger participants. Estimated GRSs in comparison with absolute sCr data were mostly supportive to the relative allelic effects of the four SNPs. The present study is supportive to earlier findings regarding influences of age and sex on sCr and it showed the importance of considering renal efficiency via cut-offs. GWAS results strengthened the idea of younger compared to older individuals as more favourable participants for studying genetic background of muscle phenotype.