Do Skinfold Measurements Provide Additional Information to Body Mass Index in the Assessment of Body Fatness Among Children and Adolescents?
OBJECTIVES. The purpose of this work was to validate the performance of age- and gender-specific BMI, triceps, and subscapular skinfold for the classification of excess of body fat in children and adolescents and to examine how much additional information these 2 skinfold measurements provide to BMI-for-age.
METHODS. The receiver operating characteristic curve was used to characterize the sensitivity and specificity of these 3 indices in classifying excess body fat. Percentage of body fat was determined by dual-energy radiograph absorptiometry. Both ≥85th and ≥95th percentile of percentage of body fat were used to define excess body fat. Data from the New York Pediatric Rosetta Body Composition Project were examined (n = 1196; aged 5–18 years).
RESULTS. For children aged 5 to 18 years, BMI-for-age, triceps skinfold-for-age, and subscapular skinfold-for-age each performed equally well alone in the receiver operating characteristic curves in the identification of excess body fat defined by either the 85th or 95th percentile of percentage of body fat by dual-energy radiograph absorptiometry. However, if BMI-for-age was already known and was >95th percentile, the additional measurement of skinfolds did not significantly increase the sensitivity or specificity in the identification of excess body fat.
CONCLUSIONS. In contrast to the recommendations of expert panels, skinfold measurements do not seem to provide additional information about excess body fat beyond BMI-for-age alone if the BMI-for-age is >95th percentile.
- dual-energy radiograph absorptiometry
- receiver operating characteristic curve
Total body fatness can be measured by a variety of methods,1,2 including laboratory techniques, such as underwater weighing, dual-energy radiograph absorptiometry (DXA), total-body water, total-body electrical conductivity, total-body potassium, and air displacement plethysmography. However, most of these methods are limited to research because of their complexity and cost.3–6 Studies of DXA measurements in animals and humans have demonstrated that DXA scans accurately capture regional and total body composition and may constitute a new reference method.7–12 However, the most frequently used tools in public health evaluations and clinical screening are anthropometric-based measurements, such as skinfold thickness or circumference, or various height- and weight-based indices, such as weight-for-height and BMI.1,2
BMI has been recommended and used worldwide to screen for overweight and obesity among both adults2,13,14 and adolescents,15–18 but as a measure of weight relative to height, it is only a proxy used to estimate body fatness. Body fatness also has been estimated from measurements of skinfold thicknesses, which correlate reasonably well with various laboratory estimates of body fatness.1,2,5 However, concerns have been expressed about the accuracy of this approach, because skinfold measurements are poorly reproducible, and, typically, only a few regional body sites are measured.19–21
Expert panels have recommended the measurement of triceps and subscapular skinfold thickness as a component of the in-depth medical assessment for children and adolescents with age- and gender-specific BMI ≥95th percentile or BMI ≥30 (whichever was smaller) or BMI ≥85th percentile but <95th percentile with complications of obesity.15,16 However, this recommendation was based on clinical judgment rather than on the ability of skinfolds to predict body fat. No pediatric studies have systematically examined the improvement in the prediction of body fat provided by skinfold measures when BMI-for-age is known. In addition, this recommendation preceded the publication of the Centers for Disease Control and Prevention (CDC) 2000 gender-specific BMI-for-age reference values.
The purpose of this analysis is to validate the performance of age- and gender-specific BMI, triceps, and subscapular skinfold percentiles in identifying children and adolescents with excess body fat and to determine how much information these skinfold measurements can contribute to that provided by BMI-for-age alone.
The potential study population consisted of 1208 healthy children and adolescents who were participants in the Pediatric Rosetta Study at St Luke's-Roosevelt Hospital Center in New York (1995–2000), a cross-sectional study of pediatric body composition. Healthy volunteers aged 5 to 18 years were recruited in the New York City area through local newspaper notices, announcements at schools and activity centers, and word of mouth. The study protocol was approved by the institutional review board of St Luke's-Roosevelt Hospital Center, and consent was obtained from each volunteer's parent or guardian. When appropriate, assent was also obtained from each volunteer. Participants' normal health status was confirmed by a medical history from a parent or guardian and a physical examination, both taken at the time of the body composition evaluation. The pubertal stage of subjects (as defined by the criteria of Tanner22) was determined by physical examination by the pediatric endocrinologist or nurse in children <11 or 12 years and by self-assessment in older subjects.23 Details of the study design and methods have been published elsewhere.11,24,25 After excluding 12 children with missing height, weight, or skinfold measurements, we were left with 1196 subjects for this analysis.
All of the anthropometric measurements were made by specially trained laboratory technicians who measured subjects' body height to the nearest 0.1 cm with a wall-mounted stadiometer (Holtain, Crosswell, Wales) and their weight to the nearest 0.1 kg using a balance-beam scale (Weight Tronix, New York, NY) with subjects wearing a hospital gown and foam slippers. They also measured subjects' triceps and subscapular skinfold thicknesses to the nearest 1 mm with a Lange caliper (Beta Technology, Inc, Cambridge, MD); both measurements were taken on subjects' right side in accordance with standard procedures.26 All of the anthropometric measurements were taken by 2 technicians with identical training and similar measurement precision. All of the intraclass (between technicians) correlation coefficients were ≥0.89.27
Whole body DXA scans were performed using Lunar models DPX with pediatric software 3.8G and DPX-L with pediatric software 1.5G (GE Lunar Corporation, General Electric, Madison, WI) in accordance with the manufacturer's instructions.28 Each scan provided estimates of subjects' fat mass and fat-free mass in kilograms and percentage of body fat (%BF). The coefficient of variation (CV) for repeated measures of percentage of fat by total body DXA scan in pediatric subjects (aged 5–17 years) in the St Luke's-Roosevelt Body Composition Laboratory is 2%.12
Each morning, before subjects were evaluated, technicians scanned an anthropomorphic spine phantom made up of calcium hydroxyapatite embedded in a 17.5 × 15 × 17.5 cm Lucite block with both DXA instruments for quality control. The phantom was also scanned immediately before and after all of the maintenance visits by the DXA system manufacturer. The measured phantom bone mineral density was stable throughout the study period at 1.166 to 1.196 g/cm2. Methanol and water bottles (8-L volume), simulating fat and fat-free soft tissues, respectively, were scanned as soft-tissue quality control markers monthly; the range in measured R values over the study period was 1.255 to 1.258 (CV = 0.127%) for ethanol and 1.367 to 1.371 (CV = 0.103%) for water.
We used the age- and gender-specific BMI reference values developed by CDC29 to assign BMI percentiles and z scores (BMI-for-age) to study participants. To determine skinfold percentiles and z scores for study subjects, we developed triceps and subscapular skinfold references for children aged 2 to 18 years (triceps and subscapular skinfold-for-age) using the same CDC growth reference data set and smoothing techniques that were used to develop the BMI percentiles and z scores. To create an internal DXA %BF reference for excess body fat, we first divided the study participants by gender into 6 age groups based on the sample size (5–7, 8–9, 10–11, 12–13, 14–15, and 16–18 years), then calculated the 85th and 95th percentiles for %BF from DXA by gender and age group. The 85th and 95th DXA cutoff points are presented in Table 1.
We used receiver operating characteristic (ROC) curves30 to assess the performance of the indices in identifying subjects with excess body fat as defined by the 85th and 95th percentile for %BF by DXA. We obtained the ROC curves by dichotomizing BMI-for-age and the skinfolds- for-age at all of the possible z score values and then plotting the resulting true-positive fractions (sensitivity) on the y-axis versus the corresponding false-positive fractions (1 − specificity) on the x-axis. An area under the curve of 0.5 indicates that the test is no better than chance in identifying cases, whereas an area of 1.0 indicates perfect prediction. We used MedCalc software31 to test the significance of the differences for the areas under the ROC curves (AUC). First, using the 85th or 95th percentile for %BF from DXA as a standard, we examined the ROC performance of BMI-for-age, triceps skinfold-for-age, and subscapular skinfold-for-age measurements independently to evaluate which was the best indicator of excess body fat with the population stratified by age group (5–11 and 12–18 years). We then repeated this analysis with subjects stratified by gender, race, and pubertal group. Second, to examine the additional information that the 2 skinfold measurements contribute to the identification of subjects with excess body fat if their BMI-for-age is known, we evaluated the ROC performance of the BMI-for-age measurements alone, of the triceps skinfold-for-age measurements among subjects with a BMI-for-age above the 85th or 95th percentile, and of the subscapular skinfold-for-age measurements among subjects with a BMI-for-age above the 85th or 95th percentile; for all of these evaluations, we used SAS 9.1.32
Subject characteristics are presented in Table 2. The study population was 52% boys, and the racial/ethnic distribution was 25% white, 22% black, 14% Hispanic, 30% Asian, and 8% other. The percentages of subjects in pubertal stages 1, 2, 3, 4, and 5 were 36%, 19%, 16%, 16%, and 13%, respectively, for boys and 38%, 11%, 17%, 19%, and 15%, respectively, for girls. Boys were taller and heavier (P < .01) and had higher z scores for height than girls (P < .05), but the mean BMI values of boys and girls were similar. Girls had greater triceps and subscapular skinfold thicknesses and higher percentages of body fat and total fat mass, whereas boys had more fat-free mass (P < .01). However, the z scores for skinfold thicknesses did not differ significantly between boys and girls.
Table 3 shows the ROC performance of BMI-for-age, triceps skinfold-for-age, and subscapular skinfold-for-age measurements among subjects stratified by age group, pubertal stage, gender, and race. There were no significant differences in the ROC performance of the 3 indicators in any of the stratified groups (P > .05). The areas under the curves were all >0.91.
Figure 1A compares the results obtained using ROC curves for BMI-for-age alone, triceps skinfold-for-age conditional on BMI ≥85th percentile, and triceps skinfold-for-age conditional on BMI ≥95th percentile to classify excess body fat in children and adolescents ages 5 to 18 years, using 95% for %BF by DXA as the standard. The key portion of the ROC curves from sensitivities ≥70% and specificities ≥70% (1 − specificities ≤30%) is amplified in Fig 1B. Selected sensitivities and specificities of BMI-for-age alone from Fig 1 are listed in Table 4, and selected sensitivities and specificities for triceps skinfold-for-age conditional on BMI ≥85th and 95th percentiles are listed in Table 5.
For the BMI ≥95th percentile group, the triceps skinfold-for-age (solid line with star markers in Fig 1B) does not provide useful information, because, whereas it improves the specificity for excess body fat compared with BMI-for-age alone at the 95%, the increase is actually less than would be found by simply increasing the BMI percentile cutoff. However, for the BMI ≥85th percentile group, triceps skinfold-for-age ≥95th percentile adds information to that from BMI-for-age alone by increasing the specificity for excess body fat (solid line with diamond markers in Fig 1B and Table 5). Results from the same analysis repeated using subscapular skinfold-for-age (shown in Fig 2 and Table 5) were quite similar, as were results from the same analyses in which the 85th percentile for %BF by DXA was used to define excess body fat (results not shown).
Because of the limitations of BMI as a measure of actual body fatness for overweight or obesity, expert panels have recommended measuring triceps and subscapular skinfold thicknesses as part of the in-depth medical assessment of children and adolescents with age- and gender-specific BMI ≥95th percentile or ≥30 (which ever was smaller) or age- and gender-specific BMI ≥85th percentile but <95th percentile or equal to 30 (which ever was smaller).15,16 However, the additional information provided by these skinfolds has not been rigorously examined. Our results indicate that for children aged 5 to 18 years, BMI-for-age, triceps skinfold-for-age, and subscapular skinfold-for-age individually performed equally well in the classification of excess body fat defined by total body DXA. However, if BMI-for-age is already known and is >95th percentile, the specificity in the identification of excess fat is similar for BMI-for-age alone and in combination with skinfold-for-age. On the other hand, for subjects with BMI-for-age of 85th to 95th percentile and skinfold-for-age ≥95th percentile, the specificity for identification of excess body fat is slightly improved.
Our results are based on standardized measurements of height, weight, and triceps and subscapular skinfold thickness on >1000 children and adolescents aged 5 to 18 years, as well as on DXA measurements for these children. Although the children and adolescents recruited in this study were from a convenience sample and, thus, do not represent the general US population, their mean height, weight, and BMI were only slightly different from children and adolescents examined in the National Health and Nutrition Examination Surveys at a fairly comparable time period.30 In our analysis, we assigned age- and gender-specific percentiles/z scores for BMI and for triceps and subscapular skinfold thicknesses and then examined the sensitivity and specificity of their ROC performance in the classification of excess body fat using %BF by DXA as the reference.
Skinfold measurements have been widely used to assess body composition in the past. They are simpler and less expensive than hydrostatic weighing or other laboratory-based techniques for body composition analysis. After the outlay for purchase of calipers, the costs are minimal. However, measurement can vary from tester to tester depending on skill and experience, unless cross-validation between testers and test-retest reliability evaluation are performed and monitored. On the other hand, BMI is calculated from weight and height measurements, both of which are routinely performed in pediatric clinical settings and are more reliable than skinfold thickness measurements. The procedure of measuring height and weight is simpler, the reliability is higher compared with that of measuring skinfolds, and both are routinely performed in pediatric clinical settings. Because BMI is calculated from weight and height, and reference curves for BMI-for-age are readily available,29 it is an appropriate screening test for excess adiposity. Furthermore, more recent studies concluded that BMI is an excellent proxy measure of adiposity in children and adolescents.33–35 Our results show that >90% of children were correctly classified as having high or low body fat using BMI-for-age ≥95th percentile cutoff (Table 4).
This study provides evidence that the current expert panel recommendation to measure the triceps and subscapular skinfolds as part of the in-depth assessment of pediatric patients with BMI-for-age ≥95th percentile to confirm excess body fat may need to be reevaluated.16 This recommendation preceded the publication of the CDC 2000 gender-specific BMI-for-age reference. Using these reference data and triceps and subscapular skinfold references based on the same data set, our evaluation suggests that skinfold measurements provide additional information on excess body fat for pediatric subjects ages 5 to 18 years with BMI-for-age of the 85th to 95th percentiles but not for those with BMI-for-age >95th percentile.
This work was supported by National Institutes of Health grant DK37352.
- Accepted December 12, 2006.
- Address correspondence to Zuguo Mei, MD, Centers for Disease Control and Prevention, Mailstop K-25, 4770 Buford Hwy, Atlanta, GA 30341-3724. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Himes JH. Anthropometric Assessment of Nutritional Status. New York, NY: Wiley-Liss, Inc; 1991
- ↵World Health Organization. Physical Status: The Use and Interpretation of Anthropometry. Geneva, Switzerland: World Health Organization; 1995
- ↵Bray GA, DeLany JP, Volaufova J, Harsha DW, Champagne C. Prediction of body fat in 12-y-old African American and white children: evaluation of methods. Am J Clin Nutr.2002;76 :980– 990
- ↵Pintauro S, Nagy TR, Duthie C, Goran MI. Cross-calibration of fat and lean measurements by dual energy X-ray absorptiometry to pig carcass analysis in the pediatric body weight range. Am J Clin Nutr.1996;63 :293– 299
- Ellis KJ, Shypailo RJ, Pratt JA, Pond WG. Accuracy of dual-energy x-ray absorptiometry for body-composition measurements in children. Am J Clin Nutr.1994;60 :660– 665
- Brunton JA, Bayley HS, Atkinson SA. Validation and application of dual-energy x-ray absorptiometry to measure bone mass and body composition in small infants. Am J Clin Nutr.1993;58 :839– 845
- Goran MI, Driscoll P, Johnson R, Nagy TR, Hunter G. Cross-calibration of body- composition against dual energy X-ray absorptiometry in young children. Am J Clin Nutr.1996;63 :299– 305
- ↵Sopher AS, Thornton JC, Wang J, Pierson RN Jr, Heymsfield SB, Horlick M. Measurement of percent body fat in 411 children and adolescents: a comparison of dual-energy x-ray absorptiometry to a four-compartment model. Pediatrics.2004;113 :1285– 1290
- ↵Garrow JS, Webster JD. Quetelet's index (W/H2) as a measure of fatness. Int J Obes (Lond).1985;9 :147– 153
- ↵Himes JH, Dietz WH. Guidelines for overweight in adolescent preventive services: recommendations from an expert committee. Am J Clin Nutr.1994;59 :307– 316
- ↵Barlow SE, Dietz WH. Obesity evaluation and treatment: expert committee recommendations. Pediatrics.1998;102(3) . Available at: www.pediatrics.org/cgi/content/full/102/3/e29
- Brook CGD. Determination of body composition of children from skinfold measurements. Arch Dis Child.1971;46 :182– 184
- ↵Roche AF, Siervogel RM, Chumlea WC, Webb P. Grading body fatness from limited anthropometric data. Am J Clin Nutr.1981;34 :2831– 2838
- ↵Tanner JM. Growth and Adolescence. 2nd ed. Oxford, United Kingdom: Blackwell; 1962
- ↵Duke PM, Litt IF, Gross RT. Adolescents' self-assessment of sexual maturation. Pediatrics.1980;66 :918– 920
- ↵Horlick M, Wang J, Pierson RN, Thornton JC. Prediction models for evaluation of total-body bone mass with dual-energy X-ray absorptiometry among children and adolescents. Pediatrics.2004;114(3) . Available at: www.pediatrics.org/cgi/content/full/114/3/e337
- ↵Lohman TG, Roche AF, Martorell R. Anthropometric Standardization Reference Manual. Champaign, IL: Human Kinetics; 1988
- ↵He Q, Horlick M, Wang J, et al. Trunk fat and blood pressure in children through puberty. Circulation.2002;105 :1093– 1098
- ↵Mazess RB, Barden HS, Bisek JP, Hanson J. Dual-energy x-ray absorptiometry for total-body and regional bone-mineral and soft-tissue composition. Am J Clin Nutr.1990;51 :1106– 1212
- ↵Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. Vital Health Stat. 2002;May:1–190
- ↵MedCalc for Windows: Statistics for Biomedical Research Software Manual [computer program]. Version 6. Mariakerke, Belgium: MedCalc Software; 2000
- ↵Base SAS 9.1 Procedures Guide [computer software]. Volumes 1–4. Cary, NC: SAS Institute Inc; 2004
- ↵Zimmermann MB, Gübeli C, Püntener C, Molinari L. Detection of overweight and obesity in a national sample of 6–12-y-old Swiss children: accuracy and validity of reference values for body mass index from the US Centers for Disease Control and Prevention and the International Obesity Task Force. Am J Clin Nutr.2004;79 :838– 843
- Mei Z, Grummer-Strawn LM, Pietrobelli A, Goulding A, Goran MI, Dietz WH. Validity of the body mass index compared to other body composition screening indices in the assessment of body fatness for children and adolescents. Am J Clin Nutr.2002;75 :978– 985
- ↵Taylor RW, Jones IE, Williams SM, Goulding A. Body fat percentages measured by dual-energy X-ray absorptiometry corresponding to recently recommended body mass index cutoffs for overweight and obesity in children and adolescents aged 3–18 y. Am J Clin Nutr.2002;76 :1416– 1421
- Copyright © 2007 by the American Academy of Pediatrics