Background: Parental feeding behaviors (restriction/pressure) are widely assumed to have a causal influence on a child’s weight; however, interest has grown in the idea that the child’s weight may also ‘elicit’ feeding behaviors in the parents. Longitudinal data support this interpretation but cannot definitely identify causation. A novel approach is to examine associations between the child’s genetic risk for obesity and parental feeding behaviors (gene-environment correlation).
Methods: BMI genome-wide polygenic scores (GPS) were created for 6710 unrelated children from the Twins Early Development Study using genome-wide association study summary statistics from an independent meta-analysis. Parental ‘pressure’ and ‘restriction’ were assessed using the Child Feeding Questionnaire when children were 10 years old. Child BMI standard deviation scores (BMI SDS) were calculated from parent-reported height and weight at 10 years. Correlations and polynomial contrasts tested associations between the GPS and parental feeding behaviors. In addition, we performed Genome-wide Complex Trait Analysis (GCTA) and twin analyses to calculate both narrow and broad-sense heritability estimates, as well as genetic correlations between BMI SDS and parental feeding styles.
Results: As expected, the GPS was correlated with BMI SDS (r = 0.192, p = 1.96 x 10-34). Consistent with the gene-environment correlation hypothesis, child BMI GPS was significantly positively associated with parental ‘restriction’ (r = 0.076,p = 6.52 x 10-07) and negatively associated with ‘pressure’ (r = -0.079,p = 2.80 x 10-07). Effects were linear across the GPS distribution. Our results remained unchanged after controlling for parental SES and parental weight. SNP-heritability estimates for parental feeding behaviors were negligible and non-significant, whereas twin-heritability estimates were moderate for parental ‘pressure’ and ‘restriction’. Genetic correlations based on the twin model were moderate and negative between BMI SDS and ‘pressure’ and moderate and positive between BMI SDS and parental ‘restriction’.
Conclusions: Results suggest an evocative gene-environment correlation, whereby heritable characteristics in the child ‘elicit’ parental behavior. These findings suggest that part of the observed association between children’s BMI and parental feeding style is due to parents responding to genetically determined characteristics of their child.
