Objective: Although the body mass index (BMI, mass index, kg/m2) is widely used as a surrogate measure of adiposity, it is moderately associated (r approximately 0.3) with height among children. We examined whether the resulting preferential classification of taller children as overweight is appropriate.
Design: Cross-sectional analyses of children (ages, 3-17 y) examined the relation of height to adiposity (as assessed by BMI and skinfold thicknesses) and fasting levels of insulin. Longitudinal analyses examined the relation of childhood height and weight-height indices to adult (mean age, 25 y) levels of adiposity and fasting insulin.
Subjects: Children (n=11,406) and adults (n=2911) who had participated in the Bogalusa Heart Study.
Measurements: We constructed three weight-height indices: BMI, W/H3, and W/Hp. The triceps and subscapular skinfolds, as well as fasting levels of insulin, were also measured.
Results: The classification of children as overweight (BMI-for-age > or =95th percentile) varied markedly by height, with a 10-fold difference in the prevalence of overweight across quintiles of height between the ages of 3 and 10 y. Childhood height, however, was also related to skinfold thicknesses and insulin levels, and all associations were modified in a similar manner by age. Furthermore, childhood height was related to adult adiposity, and of the three childhood weight-height indices, BMI showed the strongest associations with adult adiposity.
Conclusions: Because BMI reflects the positive association between height and adiposity among children, it is a better weight-height index than is either W/H3 or W/Hp.