test and health do not exist in youth or adults or when a cut-point exists in adults but extrapolation to youth is not appropriate (i.e., waist circumference, sum of skinfolds), the percentiles from another fitness measure (such as the 85th and 95th percentiles used for BMI) can be used temporarily to derive interim cut-points. As research progresses, cut-points based on the measure’s relationship to health in youth should be developed.

The committee concludes that these two approaches are appropriate for interpretation of body composition measurements administered in the context of a national youth fitness survey. For these approaches, a cut-point is determined specifically for each body composition test recommended. Obtaining information for the different indicators of body composition in this manner allows for a more complete and accurate description of an individual’s body composition. As a result, the interpretation of the tests in terms of health risks is expanded and possibly more accurate than if only one test is administered.

A third approach, which involves transforming the raw data to a measure of percent body fat by using prediction equations, has been used for interpreting body composition test items. This approach makes it possible to compare results from more than one measurement with a selected standard for percent body fat. This approach may be appropriate when test administrators must select one measure of body composition from multiple alternatives, such as when a battery of tests is applied in schools and other educational settings.

Body Mass Index

Cut-points for BMI have been calculated by age and gender from percentiles developed using the CDC growth charts based on data from large national surveys. The CDC growth charts are based primarily on data from the National Health Examination Survey (NHES) and NHANES from 1963 to 1994 (NHES II and III and NHANES I, II, and III). Data on body weight from NHANES III for subjects ≥6 years of age were not used so as to avoid the influence of an increase in body weight from previous years.1 Details are described in the CDC growth charts (Kuczmarski et al., 2000, 2002). Sample sizes were sufficiently large in the national surveys, which


1Questions arose over which population was appropriate for establishing such percentiles related to health given concern for the increasing prevalence of obesity between NHANES II (1976-1980) and NHANES III (1988-1994) (Kuczmarski et al., 2002; Ryan et al., 1999). Developing percentiles for weight using elevated values from NHANES III would have raised the percentiles and thus resulted in a false sense of having a satisfactory weight, specifically relative to stature.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement