Submitted to: International Society for Behavioral Nutrition and Physical Activity
Publication Type: Abstract only
Publication Acceptance Date: 4/20/2006
Publication Date: 7/13/2006
Citation: Simpson, P.M., Huang, B.E., Gossett, J.M., Jo, C.H., Bogle, M.L., Weber, J., Shaw, J.L., Thompson, J.W. 2006. Age and gender specfic BMI percentiles are limited for tracking the childhood obesity epidemic [abstract]. In: Proceedings of the International Society for Behavioral Nutrition and Physical Activity, July 13-16, 2006, Boston, Massachusetts. p. 261. Interpretive Summary:
Technical Abstract: Purpose: To evaluate pediatric nutrition and physical activity interventions a reliable and feasible way of tracking change in body status is needed. Historically, body mass index (BMI) has been used in adults. BMI percentiles or Z scores, which are theoretically age and gender adjusted, have been used in children because of growth. Using CDC recommendations, over 30% of children are in the top 15 percentiles. We hypothesized that because of the extreme shift in population distribution, children above the 85th percentile may substantially reduce their BMI but may still minimally change percentiles, making it difficult to track meaningful changes with percentiles. Methods: Using the CDC programs to calculate percentiles, we examined the NHANES 1998-2002, the Delta NIRI FOODS 2000, and the Arkansas school census data from 2004. We used regression analysis to investigate the relationship of age, gender, and BMI percentile. For all percentiles above the 85th, we calculated the maximum change in BMI that would still result in no change in percentile. Result: We showed that percentiles are not adjusted correctly for age and gender in the current population and an upper percentile can span a large range of BMIs. For example, an overweight child can reduce BMI by more than 30 units or 100 pounds in a year and still be classified as >95th percentile. Conclusion: Using current pediatric populations to develop new classification systems instead of BMI percentiles will avoid masking of clinically important changes in body status.