abstract
- BACKGROUND: Large population-based studies suggest that systematic measures of maternal sensitivity predict later risk for overweight and obesity. More work is needed to establish the developmental timing and potential moderators of this association. The current study examined the association between maternal sensitivity at 6 months of age and BMI z score measures at 48 months of age, and whether sex moderated this association. DESIGN: Longitudinal Canadian cohort of children from birth (the MAVAN project). METHODS: This analysis was based on a dataset of 223 children (115 boys, 108 girls) who had structured assessments of maternal sensitivity at 6 months of age and 48-month BMI data available. Mother-child interactions were videotaped and systematically scored using the Maternal Behaviour Q-Sort (MBQS)-25 items, a standardized measure of maternal sensitivity. Linear mixed-effects models and logistic regression examined whether MBQS scores at 6 months predicted BMI at 48 months, controlling for other covariates. RESULTS: After controlling for weight-relevant covariates, there was a significant sex by MBQS interaction (P=0.015) in predicting 48 month BMI z. Further analysis revealed a strong negative association between MBQS scores and BMI in girls (P=0.01) but not boys (P=0.72). Logistic regression confirmed that in girls only, low maternal sensitivity was associated with the higher BMI categories as defined by the WHO (i.e. "at risk for overweight" or above). CONCLUSIONS: A significant association between low maternal sensitivity at 6 months of age and high body mass indices was found in girls but not boys at 48 months of age. These data suggest for the first time that the link between low maternal sensitivity and early BMI z may differ between boys and girls.