Background: The fat mass and obesity-associated (FTO) single nucleotide polymorphisms (SNPs; rs1421085, rs17817449, rs9939609, rs8050136) and macronutrient intake (carbohydrate, protein, fat, total calories) are associated with body mass index (BMI). However, the mechanism for this relationship has not been fully elucidated.
Objective: This study examined whether macronutrient intake mediates the association between FTO SNPs and BMI.
Design: Baseline cross-sectional data from the Atherosclerosis Risk in Communities (ARIC) study of whites (n = 10,176) and African Americans (n = 3641) aged 45 to 64 years were analyzed.
Results: In linear regression models with BMI as the dependent variable, FTO SNPs were significantly associated with higher BMI after adjusting for covariates. The addition of energy-adjusted macronutrients attenuated the FTO effect estimates, indicating partial mediation. In whites, β ranged from 0.40 (95% confidence interval [CI], 0.20, 0.60) for rs17817449 heterozygous carriers to 0.93 (95% CI, 0.64, 122) for rs8050136 homozygous carriers; for African Americans rs17817449 homozygous carriers β was 0.65 (95% CI, 0.03, 1.27). In models with macronutrient intake as the dependent variable, all FTO SNPs were associated with higher protein intake for homozygous carriers after adjusting for BMI and other covariates. Among whites, β ranged from 1.44 (95% CI, 0.51, 2.37) for rs8050136 to 1.73 (95% CI, 0.85, 2.61) for rs17817449; among African American rs8050136 homozygous carriers β was 2.46 (95% CI, 0.77, 4.14). In mediation analysis, in whites only, FTO high-risk alleles were associated with higher BMI partly through their small effects on carbohydrate and protein intake.
Conclusions: These findings suggest that in adults, the relationship between FTO variants and BMI is not primarily through mediation of food intake.
Keywords: FTO; body mass index; macronutrient; mediation; race.