Adiposity, or the amount of body fat measured directly or indirectly, is used to determine overweight status. There are three key issues regarding the identification of overweight children: (1) whether the same measures of adiposity should be used for clinical and research purposes; (2) whether indirect measures of adiposity are adequate for screening purposes; and (3) the recommended cutoff points for defining overweight. From the point of view of those in epidemiology and public health, measures of overweight should be internationally standardized and accepted to ensure comparability of prevalence estimates within and between studies. The measures should have reasonable predictive value for morbidity and mortality, and should be easily and inexpensively obtained, especially in the clinical practice setting. In addition, if the assessment tool is not a direct measure of body fat, it should correlate well with a child's total fat.
Clinicians would no doubt share some of these preferences and would likely add that the measure should be highly specific. Having a specific measure is probably more important than having one that is highly sensitive (24) because of the great concern of psychosocial dysfunction associated with the diagnosis of overweight and the high prevalence of eating disorders among certain subsets of the pediatric population, such as adolescent girls. Thus, clinicians are likely to favor a measure that minimizes false positives, even if it does not detect all children who are overweight.
Pediatricians commonly used an indirect measure of adiposity, a weight-for-height index, to screen for overweight in children. Previously, weight-for-height was used. However, the new Centers for Disease Control and Prevention (CDC) growth charts provide weight-for-height2, or BMI, percentiles for children ages 2-20 years, allowing a single screening tool to be used (and compared) throughout the life span after 2 years of age (25). For children < 2 years of age, weight-for-height is still the screening tool of choice to assess weight status (25). With the exception of 3- to 5-year-olds, BMI is more correlated with body fat than is weight-for-height. Among children and adolescents ages 3-19 years, the correlation coefficients between BMI and percentage of body fatness defined by pooled dual-energy x-ray absorptiometry (DXA), a direct measure of adipostiy, range from 0.78 to 0.88, which are higher than those found between DXA and weight-for-height (26). This indicates that BMI-for-age is better than weight-for-height for predicting both underweight and overweight, except among 2- to 5-year-olds, for whom the two approaches are equivalent (26). Thus, for indirectly assessing overweight, BMI can be used for all children ages 2 years and above, although weight-for-height might be substituted for those aged 2-5 years.
More direct assessment of adiposity is possible by several methods including DXA (27,28), underwater weighing, bioelectrical impedance analyses, total body water, and measurement of triceps skinfold thickness. With the exception of triceps skinfold thickness, these methods are still primarily used for research, because they are more complex to obtain and require expensive equipment. Each of these measurement techniques has different assumptions and limitations, yet the percent fat values derived from the methods are highly correlated (28-30). Unfortunately, measurement of triceps skinfold thickness requires calipers, which are not commonly available in pediatrician's offices. More importantly, reproducible skinfold measurements may be a problem. Obtaining reliable triceps skinfold measurements requires practice, and between-observer measurements are not as reproducible as measurements of height and weight. Furthermore, skinfold measurement may become less reliable as body fatness increases. Because skinfold measurements are currently the only practical direct measure of adiposity available to most clinical practice settings, we recommend using triceps skinfold measurements as the primary diagnostic method for confirming that those overweight are overfat. However, adequate staff training and monitoring of the accuracy of their measurements is crucial to obtain accurate results.
Although both triceps skinfold measures and BMI are correlated with morbidity (3,23), BMI is easier to use as an initial screening tool. BMI and weight-for-height also predict the persistence of obesity into adulthood (10,11). Thus, given all their attributes, the use of BMI (10,11,25,31) for children ages 2-20 years or weight-for-height (25,26,32) for those ages from birth to 5 years appears well suited to estimating the risks for chronic diseases, persistent obesity, or other long-term sequelae of childhood overweight.
These observations led the International Obesity Task Force (33-35) to suggest that the internationally accepted scheme of BMI cutpoints for adult morbidity be used for children and adolescents. Cole et a1. (36) used these criteria to create centile curves based on the centile values identified by the adult cutpoints of 25 and 30 kg/m2 at age 18 years. These standards have been further refined and clarified by an Expert Committee on Pediatric Obesity established by the Maternal and Child Health Bureau (37), which recommended that children and adolescents with a BMI >85th centile be screened for complications such as the presence of additional risk factors, including family history, blood pressure, cholesterol, and be evaluated and possibly treated (Fig. 1).
This committee recommended the following defining criteria: children with a BMI >85th centile and <95th centile for the same age and gender be categorized as at risk for overweight and those with a BMI
>95th centile for the same age and gender be defined as overweight. Among children ages 2-20 years, if BMI is at or exceeds the 95th centile for the same age and gender, the diagnosis of excess adipose tissue should be confirmed by triceps skinfold thickness. In addition, it is also useful for selecting intervention strategies to know if the child or adolescent is concerned about his or her weight. Although these cutpoints were originally based on BMI cutpoints for adult morbidity, Freedman et al. (3) found these cutpoints were able to identify adverse lipid, insulin, and blood pressure in childhood as well.
Was this article helpful?