Effect of choice of outcome measure on studies of the etiology of obesity in children

Epidemiologic studies of the etiology of childhood obesity often use proxies for adiposity as outcomes. This study aimed to compare the ability of a range of proxy measures to detect associations with sedentary behavior. Longitudinal study of children in the Gateshead Millennium Study who were 7 years at baseline and 9 years at follow-up. Associations between 2-year changes in objectively measured sedentary behavior and changes in proxies for adiposity (waist circumference, waist Z score, body mass index, body mass index Z score) and measurement of body composition (fat mass index from bioelectric impedance) were examined. Associations were tested with linear regression. Associations between 2-year increases in sedentary behavior and increases in adiposity were detectable by using the fat mass index as the outcome, but not the simple proxy measures of adiposity, as outcomes. Proxy measures are inferior to measures of body composition as outcomes in epidemiologic studies.