Identification of Independent Predictors for Underweight and Overweight in Adolescents with Overweight Mothers

Background: The nutrition and epidemiologic transition is characterised by changes in diet and activity patterns, leading to the development of a double burden of malnutrition. At the household level this often manifests as an overweight mother and underweight child. However, child and adolescent overweight and obesity also appear to be on the rise in transitional countries, although it is not well studied. Parent and child overweight are associated, and overweight children have a greater risk of becoming overweight adults. Understanding why these two extremes of child malnutrition occur in households with an overweight mother is necessary for policy and programmes to target both effectively.

Methods: This study examined Young Lives data on 256 adolescents with overweight (BMI>23kg/m2) or obese mothers (BMI>27kg/m2) from India. Child weight status was determined using BMI-for-age z-scores based on the 2007 WHO reference median. The aim was to identify independent predictors of child over- and underweight in two categories of dual burden households: 61(23.8%) underweight child/overweight mother (U/O) households, 21(8.2%) overweight child/overweight mother (O/O) households. The reference category was 174(68.0%) normal weight child/overweight mother (N/O) households. Bivariate analysis was used to examine associations between child, parent and household-level indicators and household nutrition status. Polychotomous logistic regression techniques were employed to identify independent predictors of U/O and O/O households compared to the reference N/O households.

Results: After controlling for nutritional and sociodemographic variables, several indicators emerged as being significant. Households were less likely to be U/O compared to N/O if: the child was female (OR=0.35, p<0.01) or post-pubertal (OR=0.26, p<0.01) or BCG immunised (a proxy for care practice) (OR=0.44, p<0.05), or the maternal support network was larger (OR=0.20, p<0.05). Households were also less likely to be O/O compared to N/O if the child was female (OR=0.23, p<0.05). Relative risk of O/O was increased if the child was post-pubertal (OR=3.47, p<0.05) or if the child had a moderate diet diversity (OR=4.97, p<0.1). Conclusion: Both U/O and O/O double burden households are present in this sample. Longitudinal data suggests under- and overweight among children in this cohort are both increasing. Indicators for either type of household imply that social and physiological factors may be involved, but more research is required to better define these pathways. In the meantime, nutrition intervention programmes must recognise the existence of both extremes of malnutrition at population and household level, and incorporate this into their targeting strategies in order to manage the dual burden of malnutrition effectively.