| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communication |
1 From the Departments of Human Nutrition (RWT), Medical and Surgical Sciences (IEJ and AG), and Preventive and Social Medicine (SMW), University of Otago, Dunedin, New Zealand.
2 Supported by The Health Research Council of New Zealand and the Otago Medical Research Foundation.
3 Address reprint requests to RW Taylor, Department of Human Nutrition, University of Otago, PO Box 56, Dunedin, New Zealand. E-mail: rachael.taylor{at}stonebow.otago.ac.nz.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
25 but < 30) and obesity ( 30) in adults were published recently. Objective: The objective was to estimate the percentage body fat (%BF) values typically associated with these BMI cutoffs in children and adolescents.
Design: The %BF was measured by dual-energy X-ray absorptiometry in 661 subjects (49% male) aged 3–18 y. Regression equations using BMI, age, and sex were developed to predict the %BF associated with BMI cutoffs for overweight (age-specific BMI equivalent to a BMI of 25 in an 18-y-old) and obesity (age-specific BMI equivalent to a BMI of 30 in an 18-y-old) over this age range.
Results: Measurements classified 17.1% of males and 19.8% of females as overweight and 5.5% of males and 7.5% of females as obese. The %BF associated with an obese BMI tended to be higher in peripubertal males (34–36%) than in younger (24–30%) or older (27–30%) males. Although the predicted %BF of young females was similar to that of young males, values rose steadily with age, such that an 18-y-old female with a BMI of 30 had an estimated %BF of 42%, whereas that in males of similar age was 27%.
Conclusion: The %BF values associated with BMI classifications of overweight and obesity vary considerably with age in growing children, particularly in girls.
Key Words: Percentage body fat • body mass index • obesity • dual-energy X-ray absorptiometry • overweight • children • adolescents
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
25 but < 30) and obesity ( 30) in adults (2). Because the BMI percentile allocation tends to track in childhood (3, 4), these percentiles could be used to define overweight and obesity in subjects of all ages, thus ensuring continuity in recommended cutoffs from youth to adulthood. Cole et al (5) recently published such an analysis of an international database of children aged 2–18 y from 6 countries; BMI cutoffs were provided for every half year of age in both sexes. However, although BMI is strongly correlated with adiposity in children (6–8), it may not be suitable for use in some subgroups of the pediatric population, because of its inherent status as a weight-based index (9, 10). The health complications associated with obesity are related to the elevated deposition of body fat rather than to body weight per se, and thus the ability to measure actual adiposity would eliminate any potential misclassification based on BMI. To date, few studies have assessed the magnitude of health risk associated with excess body fat in children (11–13). Further work is required to ascertain whether measurements of percentage body fat (%BF) are superior to those of BMI for determining health risks. Until recently, the available methods that were suitable for assessing %BF in large numbers of children have been limited. The increasing availability of bioelectrical impedance analysis of body composition, which derives %BF, necessitates information regarding what constitutes excess adiposity during growth. Therefore, in an effort to describe the typical levels of adiposity that might be expected at the cutoffs denoting overweight or obesity, we intended for our study to link dual-energy X-ray absorptiometry (DXA ) measurements of %BF with BMI guidelines denoting overweight and obesity (5) in children and adolescents aged 3–18 y.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Body-composition analysis
The %BF was estimated with the use of DXA (Lunar DPX-L scanner; Lunar Corporation, Madison, WI), and DPX-L software, version 1.3z (Lunar Corporation) was used to analyze the DXA scans as has been used previously (15, 16). The scanner determines total fat mass, bone-free lean tissue mass, bone mineral content (in g), and areal bone mineral density (in g/cm2). Percentage fat mass (% fat mass) determined by DXA is calculated as [fat mass/(fat mass + bone-free lean tissue mass + bone mineral content) x 100]. Total body mass by DXA was highly correlated with total body weight measured by electronic scales in the entire study population (r > 0.99). All DXA scans were completed on the same scanner with the same software by 2 operators who had been fully trained in the operation of the scanner, the positioning of subjects, and the analysis of results according to manufacturer’s guidelines. Ethical approval to conduct estimates of precision for DXA body-composition analysis by performing repeat scans in children was not available. In our laboratory, the CVs for scanning precision completed on 10 repeat scans of adult subjects were 2.6% for total fat mass (kg), 2.5% for % fat mass, and 1.1% for bone-free lean tissue mass (kg).
Statistical analyses
Statistical analyses were performed with SPSS 10 software for the Macintosh (Language Systems Corp, Chicago), and results are presented as 
± SD. Cole et al (5) listed the BMI values for each half-year of age from 2 to 18 y, which correspond to the adult BMI cutoffs of 25 and 30. These cutoffs were created with the use of an international database of > 190 000 subjects aged 0–25 y from 6 countries. Sex-specific BMI centile curves were constructed for each country by the use of the LMS method (5). The percentiles corresponding to BMI cutoffs of 25 and 30 at the age of 18 y were used to denote overweight and obesity at all ages (2–18 y). By averaging of the distribution curves from each country, international cutoffs were obtained that are representative of the reference countries but independent of the level of obesity in each country (5). Because these cutoffs of Cole et al (5) were provided for each half-year of age, and because our subjects did not necessarily undergo the scanning near their birthday, we used linear interpolation to calculate the appropriate BMI cutoff for each child according to their exact age at measurement (in decimal years). The numbers and distribution of children above the overweight and obese BMI cutoffs were then examined according to age and sex.
In each sex, the relation between %BF (log transformed) and BMI (log transformed) was assessed with the use of Pearson’s correlation coefficients, before and after adjustment for age. Subsequently, multiple regression analysis was used to determine the relations between %BF (log-transformed, dependent variable), BMI (log transformed), age (decimal years), and interaction between BMI and age (independent variables). Age and BMI were centered (age -11 and lnBMI -3) to decrease the multicolinearity between the independent variables. Separate analyses were conducted in each sex because the relation between BMI and body fat differs in males and females during growth (9). The regression equations created were then used to estimate the %BF corresponding to the BMI cutoffs equivalent to BMIs of 25 and 30 in an 18-y-old for every year of age in males and females. The 95% CIs for these estimates are presented.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
according to age and sex. Fifty-six males (17.1%) and 66 females (19.8%) met the BMI criteria for overweight: 18 males (5.5%) and 25 females (7.5%) were classified as obese according to their BMI (5) (P > 0.05 for sex comparisons). Thus, in total, 22.6% of the males and 27.2% of the females in this study were heavier than is desirable for their health. As is shown in Figure 1, these overweight and obese children ranged in age from 3 to 18 y. In addition, the mean age of the children with normal BMI values did not differ from that of the children classified as overweight or obese (± SD: 11.3 ± 4.3 compared with 11.5 ± 4.5 y in males and 10.4 ± 4.1 compared with 10.9 ± 4.4 y in females, P > 0.05 for both comparisons).
|
|
 | (1) |
 | (2) |
with R2 of 0.83 and SE of 0.16 in females. In both sexes, all independent variables contributed significantly to the prediction of %BF. In males, lnBMI -3 explained 47.5% of the variance in %BF; age -11 explained an additional 10.5%, and the interaction between age -11 and lnBMI -3 explained yet a further 3.6%. In females, lnBMI -3 also explained the majority of the variance in percentage fat (81.1%); age -11 and the interaction term explained much smaller but still significant amounts (0.5 and 1.5%, respectively). The level of agreement between predicted and measured %BF was high in both the total group [ difference (95% CI): in males, 0.6% (0.0, 1.1%); in females, 0.5% (0.1, 0.9%)] and when restricted to the overweight (n = 165) subjects [males, 1.8% (0.2, 3.4%); females, 0.6% (-0.3, 1.6%)]. The predicted %BF values corresponding to the BMI cutoffs (5) for each year of age are shown in Table 2, and it can be seen that values were consistently higher in females than in males at all ages.
|
. Although some overlap in %BF values was apparent between children of the 3 BMI groupings (normal-weight, overweight, and obese), most subjects with low BMI also had low %BF. Similarly, most of the children with high BMI also had high adiposity.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The predicted %BF corresponding to the BMI cutoff classifying children and adolescents as overweight ranged from 18% to 23% in males and from 20% to 34% in females. As expected, the predicted %BF values designating obesity were higher: 24–36% in males and 26–46% in females. The %BF values we estimated for obese 18-y-olds are similar to those reported in a similar analysis by Gallagher et al (24) for white adults aged 20–39 y. In their group, a BMI of 30 was linked to a %BF of 26% in males and of 39% in females, as compared with our estimates of 27% and 42%, respectively. The higher predicted percentage fat seen in the 18-y-old females in our study than was seen in the females in the study by Gallagher et al (24) may have resulted from the fact that we conducted separate analyses in each sex, because the relations between BMI and adiposity differ in males and females during growth (9, 15). In contrast, Gallagher et al (24) used a single equation that also included ethnicity as an independent variable. Development of a combined equation in our study population gave estimates similar to those observed with the sex-specific equations (28% in males and 43% in females).
As expected, the DXA cutoffs predicting %BF in relation to age showed a different pattern in males than in females. The apparent increase in predicted %BF for a given BMI cutoff observed in the young teenaged males in our study supports earlier longitudinal studies reporting that peripubertal males typically show transient increases in adiposity (25, 26). In contrast, the percentage fat cutoffs in females increased steadily with age before beginning to plateau in the older subjects. This age-related variation occurred to a greater extent in the cutoffs used to classify obesity than in those used to classify overweight in both sexes. Because cross-sectional data represent average growth patterns (27), we cannot use our predicted %BF values to estimate the velocity of fat growth in individual children. Rather, those values were designed simply to estimate expected levels of adiposity in children classified as overweight or obese according to their BMI.
Our results show that a single percentage fat value may not be suitable across a wide age spectrum for classifying children as obese, but the use of such a value may be appropriate in studies that restricted their analyses to prepubertal children (23) or to subjects within a narrow age range (13). However, studies (11, 12) that have used a single %BF value (> 30%) to denote fatness in females aged 5–18 y may have underestimated excess adiposity in younger girls and overestimated the proportion of older females who should be classified as overfat. Several lines of evidence support our belief that age-specific definitions of adiposity might be required, at least in females. Few prepubertal girls have %BF that exceeds 30%, and yet adverse risk factors are apparent in children as young as 5 y old when they are classified as overweight according to their BMI ( 95th percentile) (28). BMI and DXA-measured %BF show similar associations with cardiovascular disease risk factors during growth (15), and both BMI and %BF increase with advancing age in females (6, 7, 29). These observations suggest that the level of adiposity denoting increased metabolic risk might be influenced by age. Although Dwyer and Blizzard (12) chose to use a single percentage fat cutoff (30%), the most appropriate value denoting increased risk of elevated systolic blood pressure ranged from 29 to 35% in the 9- and 15-y-old females in their study. Williams et al (11) divided their 5–18-y-old subjects by ranges of %BF (eg, 20–25%) and adjusted for the effect of age within their regression analyses. It would have been interesting to know the effect of including more complex terms for age, because our data and those of Cole et al (5) suggest an increasing variability in both BMI and percentage fat in older children than in younger children.
The strengths of our study include the large sample size and the use of DXA, a validated measure of %BF in children (30). However, because our study included children enrolled in various body-composition studies, they constitute a convenience sample of children. A recent national survey of Australian children (31) showed that 15.0% of males and 15.8% of females aged 2–18 y were overweight according to the criteria of Cole et al (5), and an additional 4.5% of males and 5.3% of females were classified as obese. The proportions of the males in our study who were overweight (17.1%) or obese (5.5%) were similar to the proportions in the Australian study, which suggests that our sample of males is representative. In contrast, in comparison with the Australian data, greater proportions of the females in our study were classified as overweight (19.8%) or obese (7.5%). Chi-square tests showed that these differences were not significant (P > 0.05 for all categories). It is possible, however, that this slight overrepresentation of overweight females may have contributed to the higher predicted percentage fat estimated in the young adult females in our study than was seen by Gallagher et al (24).
Because insufficient numbers of children from other ethnic groups were available for analysis, only white children were included in our study. It is possible and indeed feasible that different body fat values may be expected in children from other ethnic groups. We have provided estimates of the %BF values associated with BMI indexes of overweight in growing white New Zealand children, but we have made no attempt to calculate the %BF corresponding to BMI measures of underweight, because BMI cutoffs comparable with those used in adults do not yet exist for children. In addition, it seems possible that BMI cutoffs for underweight are inappropriate in youngsters because BMI is a poor indicator of relative adiposity in thin children (29).
Our results are not designed to be decisive values denoting the %BF expected in every child who meets the BMI classifications for overweight and obesity. Rather, they represent the typical levels of adiposity that might be expected in a child at the cutoffs classifying the child as overweight or obese on the basis of BMI. Further work is required to compare the usefulness of %BF cutoffs with that of BMI classifications of overweight and obesity in predicting increased metabolic risk in children and adolescents of all ages and ethnicities.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. C. Nooyens, L. L. Koppes, T. L. Visscher, J. W. Twisk, H. C. Kemper, A J. Schuit, W. van Mechelen, and J. C Seidell Adolescent skinfold thickness is a better predictor of high body fatness in adults than is body mass index: the Amsterdam Growth and Health Longitudinal Study Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1533 - 1539. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Mei, L. M. Grummer-Strawn, J. Wang, J. C. Thornton, D. S. Freedman, R. N. Pierson Jr, W. H. Dietz, and M. Horlick Do Skinfold Measurements Provide Additional Information to Body Mass Index in the Assessment of Body Fatness Among Children and Adolescents? Pediatrics, June 1, 2007; 119(6): e1306 - e1313. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Moreno, M. G. Blay, G. Rodriguez, V. A. Blay, M. I. Mesana, J. L. Olivares, J. Fleta, A. Sarria, M. Bueno, and AVENA-Zaragoza Study Group Screening performances of the international obesity task force body mass index cut-off values in adolescents. J. Am. Coll. Nutr., October 1, 2006; 25(5): 403 - 408. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kuh, R. Hardy, S. Butterworth, L. Okell, M. Richards, M. Wadsworth, C. Cooper, and A. A. Sayer Developmental Origins of Midlife Physical Performance: Evidence from a British Birth Cohort Am. J. Epidemiol., July 15, 2006; 164(2): 110 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kuh, R. Hardy, S. Butterworth, L. Okell, M. Wadsworth, C. Cooper, and A. Aihie Sayer Developmental origins of midlife grip strength: findings from a birth cohort study. J. Gerontol. A Biol. Sci. Med. Sci., July 1, 2006; 61(7): 702 - 706. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Hoffman, A. L. Sawaya, P. A. Martins, M. A. McCrory, and S. B. Roberts Comparison of Techniques to Evaluate Adiposity in Stunted and Nonstunted Children Pediatrics, April 1, 2006; 117(4): e725 - e732. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. L Burdette, R. C Whitaker, W. C Hall, and S. R Daniels Breastfeeding, introduction of complementary foods, and adiposity at 5 y of age Am. J. Clinical Nutrition, March 1, 2006; 83(3): 550 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. W. Demerath, C. M. Schubert, L. M. Maynard, S. S. Sun, W. C. Chumlea, A. Pickoff, S. A. Czerwinski, B. Towne, and R. M. Siervogel Do Changes in Body Mass Index Percentile Reflect Changes in Body Composition in Children? Data From the Fels Longitudinal Study Pediatrics, March 1, 2006; 117(3): e487 - e495. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kalantar-Zadeh, N. Kuwae, D. Y Wu, R. S Shantouf, D. Fouque, S. D Anker, G. Block, and J. D Kopple Associations of body fat and its changes over time with quality of life and prospective mortality in hemodialysis patients Am. J. Clinical Nutrition, February 1, 2006; 83(2): 202 - 210. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Mok, L. Beghin, P. Gachon, C. Daubrosse, J.-E. Fontan, J.-M. Cuisset, F. Gottrand, and R. Hankard Estimating body composition in children with Duchenne muscular dystrophy: comparison of bioelectrical impedance analysis and skinfold-thickness measurement Am. J. Clinical Nutrition, January 1, 2006; 83(1): 65 - 69. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. S. Cooke and A. Naaz Role of Estrogens in Adipocyte Development and Function Experimental Biology and Medicine, December 1, 2004; 229(11): 1127 - 1135. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G Neovius, Y. M Linne, B. S Barkeling, and S. O Rossner Sensitivity and specificity of classification systems for fatness in adolescents Am. J. Clinical Nutrition, September 1, 2004; 80(3): 597 - 603. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, BMI less than age-specific BMI equivalent to a BMI of 25 in an 18-y-old; 
, BMI greater than age-specific BMI equivalent to a BMI of > 25 but < 30 in an 18-y-old; and 
, BMI greater than age-specific BMI equivalent to a BMI of 