Influence of diet, physical activity and parents' obesity on children's adiposity:: a four-year longitudinal study

OBJECTIVE: To assess the relationships between diet, body composition, physical activity, parents' obesity and adiposity in children at the age of 8 y and four years later. STUDY DESIGN: Prospective observational study of anthropometric measures initiated in 1992, follow-up examination in 1996. METHODS: 112 prepubertal (age: 8.6 +/- 1.0 y) children were studied. Energy and nutrient intakes were assessed by diet history, body composition by anthropometry and physical activity, by a questionnaire. Obesity was defined as relative body mass index (BMI) (rel BMI) > 120%, where rel BMI = (BMI/BMI at 50th centile for age and gender) x 100. RESULTS: Prevalence of obesity was not statistically different at baseline (22.3%) than four years later (19.8%): rel BMI at the age of 8 y was positively self-related with rel BMI at the age of 12 y (r = 0.73, P < 0.001). After four years, eight (32%) obese children became non obese and five (6%) non obese children became obese. Multiple regression analysis (stepwise procedure) revealed that, in the final equation, the mother's BMI and TV viewing (independent variables) accounted for 17% of the children's rel BMI variance at the age of 8 y (R = 0.42, P < 0.001) while the parents' BMIs accounted for 13.5% of the children's rel BMI variance at the age of 12 y (R = 0.37, P < 0.001). Other variables such as total energy intake, nutrient intake percentage and amount of physical activity, were all rejected. An autoregressive unbalanced measures model regression analysis recognised the mother's and father's BMIs as the only variables able to predict rel BMI in the children (mother's BMI coeff. 2.53 (s.e.m. 0.26), P < 0.0001; father's BMI coeff. 2.07 (s.e.m. 0.23), P < 0.0001). A multivariate logistic regression analysis was also performed. The children who participated in the follow-up, were divided into two groups based on the positive or negative change in the rel BMI between final and baseline measurements. Of all the variables considered, only rel BMI at baseline was selected in the final equation. Other variables such as age, gender, energy and nutrient intake, TV viewing and amount of physical activity, as well as the parents' BMI, were all removed. CONCLUSIONS: The parents' obesity was the main risk factor for obesity in this group of children. Sedentary behaviour (TV viewing) was independently associated with overweight at the age of 8 y. Physical activity and energy and nutrient intakes did not significantly affect the change in rel BMI over the four-year period when the parents' obesity was taken into account.