Meta-analysis of genome-wide association studies for body fat distribution in 694,649 individuals of European ancestry

Obesity has reached pandemic proportions: 25% of adults worldwide are either overweight (body mass index, BMI≥25 kg/m²) or obese (BMI≥30) and >10% of children meet obesity criteria. While obese individuals are at higher risk of some cancers, infertility, and cardiometabolic disease, epidemiological studies show that the location of fat storage, rather than general adiposity, is more informative for predicting disease risk. Independent of BMI, individuals with higher central adiposity have higher cardiometabolic disease risk and Mendelian randomisation studies indicate a putative causal effect of higher central adiposity on cardiometabolic disease.

To pinpoint genetic variants associated to fat distribution, we performed a meta-analysis of waist-hip ratio adjusted for BMI (WHRadjBMI). WHR is an easily-measured fat distribution phenotype well-correlated with imaging-based fat distribution measures. Implementing a linear mixed model to account for relatedness and ancestral heterogeneity, we performed genome-wide association studies (GWAS) of WHRadjBMI in UK Biobank, followed by meta-analysis with publicly-available data. The meta-analysis comprised 694,649 samples and ~27.4M SNPs and identified 346 associated loci (p<5x10^-9) containing 463 independent signals.

Given the known sex-dimorphism of fat distribution in humans, we also performed sex-specific meta-analyses. We identified ~3x more loci associated in women (266 loci) than men (91) and found SNP-based heritability to be strikingly different between the sexes (h_women²=25.6%, h_men²=16.7%, p=9x10^-85). Of all index SNPs (p<5x10^-9 in the combined or sex-specific analyses), 105 were sex-dimorphic (Bonferroni p<3.3 x 10^-5) and 97 of these 105 showed stronger effects in women than men.

In the combined analysis, index SNPs in 12 loci correlated with nearby loss-of-function variants, including stop gains in an insulin like growth factor binding protein (IGFBP3) and a gene expressed in testis (AKR1E2). Pathway analyses mapping associated SNPs from the combined analysis to eQTLs in 53 tissues revealed enrichment of association signal in adipose tissue, while analysis of associated SNPs in women revealed an enrichment not only for adipose but also in female reproductive tissues including uterus, cervix and vagina.

Additional work, including fine-mapping and further tissue and cell-enrichment analyses, will help elucidate the causal mechanisms and biological interplay underlying fat distribution and its comorbidities.