The current CDC guidelines for healthy weight in children and youth are in the range of the 5th to 85th percentiles of the age- and gender-specific BMI charts. Therefore, a child whose weight tracks in that range—that is, he or she does not cross to lower than the 5th or higher than the 85th percentiles—would be considered to be in the healthy weight range according to these definitions.

The CDC BMI charts are mathematically smoothed curves of the pooled growth parameters of children and adolescents sampled in cross-sectional national health surveys conducted from 1963 to 1994. An analogy would be to consider the curves as compiled from a series of “snapshots” of large national samples made at different times over three decades. But because the sample sizes at each age level get much smaller at the extremes of the distributions, the growth curves may be more prone to errors at the upper and lower ends.

Because of the increases in body weight that occurred in the 1980s and 1990s—after the second National Health and Nutrition Examination Survey (NHANES II) conducted in 1976-1980—a decision was made not to include the NHANES III (1988-1994) body-weight data in the revised 2000 BMI charts for children aged 6 years or older. The NHANES III data would have shifted the affected curves (weight-for-age and BMI-for-age) upward, which was considered to be biologically and medically undesirable. However, the fact that the CDC BMI charts were developed from data for a prior time period in which children were leaner, on average, leads to an occasionally confusing situation—for example, where more than 5 percent of the population is above the 95th percentile—but this is readily clarified in the context of the charts’ historical source.

The CDC BMI charts are derived from cross-sectional samples of children (data for different age groups are based on different children). That is, they do not directly represent the longitudinal growth trajectory for the same set of children who have been measured as they age.4 Therefore, it is not known whether an individual child’s height, weight, or BMI should be expected to follow along the same percentile curve over time in order to maintain health or whether there are health implications of variations throughout childhood (e.g., crossing percentiles by going from the 20th percentile at age 1 to the 60th percentile at age 5 to the 40th percentile at age 12). Mei and colleagues (2004) found that shifts in growth rates were


The latter approach has been used to develop longitudinal growth charts that are used in several other countries (Tanner and Davies, 1985; Cameron, 2002). These types of charts are generally developed from smaller, and potentially less representative, samples.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement