Association of symptoms of attention-deficit/hyperactivity disorder with childhood overweight adjusted for confounding parental variables.

ABSTRACT

OBJECTIVE: There have been numerous reports on association between attention-deficit/hyperactivity disorder (ADHD) and overweight/obesity in children and adolescents; however, most studies adjusted only for a limited number of possible confounders. METHODS: We analyzed the data of 11,159 six through seventeen-year-old participants in the German Health Interview and Examination Survey for Children and Adolescents. We determined weight status based on measured anthropometry and national reference data by International Obesity Task Force criteria. The parent-rated hyperactivity/inattention subscale of the Strengths and Difficulties Questionnaire (SDQ-HI) was used as a continuous measure of ADHD symptoms. We examined whether the putative confounders socioeconomic status, migrant status, parental body mass index (BMI) and parental smoking were associated with both SDQ-HI and overweight/obesity. Associations between SDQ-HI and overweight/obesity vs normal weight were analyzed by binary logistic regression analyses. In the first model, we adjusted for age and sex only and in the second model also for the parental confounders. RESULTS: SDQ-HI was associated with an increased risk for overweight/obesity in both sexes adjusting for age and sex. However, after adjusting for all confounders SDQ-HI was associated with an increased risk for overweight/obesity only in adolescent females. Socioeconomic status, parental BMI and parental smoking each were relevant confounders. Migrant status was also significantly associated with both SDQ-HI and overweight/obesity, thus qualifying as a confounder but contributed only weakly to the association. CONCLUSIONS: The association between ADHD symptoms and overweight/obesity is due to confounding by family background variables in all but adolescent girls. Possible reasons for the increased risk for overweight/obesity in this subgroup are discussed. We also propose possible mechanisms for confounding by parental socioeconomic status, BMI and smoking.