Relationships among physical activity, diet, and obesity measures during adolescence

Abstract

Today's high prevalence of obesity is a concern especially in youth. Physical activity and diet are both important factors associated with weight management, and current recommendations are to consume a diet low in saturated fat and high in fiber, fruit and vegetables and to participate in frequent and regular physical activity. Adherence to recommendations is low, a factor that is strongly correlated with development of obesity and associated chronic diseases such as hypertension and cardiovascular disease. While associations between diet and physical activity are well established, investigation of changes in their association during growth is lacking. This thesis uses five years of diet, physical activity, and anthropometric data from 2379 adolescent girls in the National Heart Lung and Blood Institutes, Growth and Health Study to explore associations among diet, physical activity, and obesity cross-sectionally and with age. Variables representing physical activity, diet quality, and obesity, as well as income, maturation stage, and other potential confounders, were evaluated pair-wise for correlation, and bivariate statistics were examined for longitudinal trends. For further evaluation of relationships between groups of variables we used a canonical correlation analysis. First, physical activity variables were grouped with confounders and examined in relationship to diet quality variables; next, we grouped physical activity, diet quality, and confounders and examined the relationship to obesity measures. We found a moderately increasing correlation between physical activity and diet with age and an age-related decrease in correlation of all health behaviors and confounding variables with obesity measures, indicating that obesity measures become less sensitive to behaviors and socioeconomic factors with age at the same time as health behaviors become more tightly linked. These results suggest that while health behaviors continue to interact during growth, and in fact become more intertwined, measures of obesity become more static and may be less responsive to potential interventions with increasing age. These findings should motivate intervention work to aim for youth as potential impact would be greater before health behaviors and obesity measures become "locked in" to the more static frame of adulthood.