See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/15182578

Asians have lower body mass index (BMI) but higher percent body fat than do
whites: Comparisons of anthropometric measurements

Article  in  American Journal of Clinical Nutrition · August 1994

DOI: 10.1093/ajcn/60.1.23 · Source: PubMed

CITATIONS READS

761 3,933

6 authors, including:

Jiantian Wang John C Thornton

Kean University AOL
165 PUBLICATIONS   12,921 CITATIONS    235 PUBLICATIONS   11,602 CITATIONS   

SEE PROFILE SEE PROFILE

Steven B Heymsfield Richard N Pierson

Pennington Biomedical Research Center Columbia University
1,043 PUBLICATIONS   86,831 CITATIONS    279 PUBLICATIONS   21,352 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Metabolic Syndrome View project

Quality control in medical testing View project

All content following this page was uploaded by Steven B Heymsfield on 18 August 2014.

The user has requested enhancement of the downloaded file.

Asians have lower body mass index (BMI) but higher
percent body fat than do whites: comparisons of
anthropometric measurements3
Jack Wang, John C Thornton, Mary Russell, Santiago Burastero, Steven Heymsfield,
and Richard N Pierson Jr

ABSTRACT We studied the correlations between body comparisons between whites and Chinese, but the data for the
mass index (BMI) and percent body fat (fat%) measured by dual- Chinese group were collected almost three decades ago (13).
photon absorptiometry (DPA) in 445 white and 242 Asian adults For this report we studied the correlations of BMI and percent
aged 1 8-94 y. In addition, comparisons in six circumferences body fat (%fat) measured by dual-photon absorptiometry
and eight skinfold-thickness measurements between whites and (DPA) in whites and Asians with wide ranges of age and body
Asians were made to explain the discrepancies. Although Asians fatness. The DPA method bypasses the traditional assumptions

Downloaded from www.ajcn.org by guest on July 14, 2011

had lower BMI, they were fatter than whites of both sexes. The of consistency of water, potassium, or density of fat-free mass
correlations between fat% and BMI varied by BMI and sex and (3-5), and therefore is independent of age, sex, and race (14).
race. Comparisons in anthropometry show that Asians had more In addition, comparisons between whites and Asians in anthro-
subcutaneous fat than did whites and had different fat distribu- pometric measurements were made in an effort to explain the
tions from whites. Asians had more upper-body subcutaneous fat discrepancies.
than did whites. The magnitude of differences between the two
races was greater in females than in males. Prediction equations
developed for each sex and race, based on BMI alone, gave SEEs
ranging from 4.4% to 5.7%. All were significantly improved to
Subjects and methods
the range of 3.5-4.4% when age and several skinfold-thickness
measurements were added. Am J Clin Nutr l994;60:23-8. Six hundred eighty-seven healthy volunteers (Table 1) were
recruited in the New York City area; 445 whites and 242 Asians
KEY WORDS Body mass index, body fat, anthropometry, aged 18-94 y with BMIs of 15-38 [wt (in kg)/ht’ (in m)] were
ethnicity studied. Ethnicity was defined by the background of both parents
and all grandparents. Asians included 225 Chinese, 9 Japanese,
6 Koreans, and 2 Filipinos, 97% of whom were born in Asia.
Introduction Whites were of European origin.
Posters, local newspapers, radio stations, local community
Weight-height indexes are well accepted as indexes for body-
centers, and a census-tract-based mailing list provided by an
fat content. Therefore, most researchers in nutrition use body
election computer service were used for recruitment. All subjects
mass index (BMI) for assessment of nutritional status. But recent
were ambulatory. The study was approved by the St Luke’s-
studies by Lohman (1) and Smalley et al (2) demonstrate the
Roosevelt Hospital Institutional Review Board. Each subject
unreliability of BMI for predicting body fat.
gave informed consent.
Two observations motivated this study. First, in earlier refer-
After an overnight fast, body weight was measured to ±0.2 kg
ence data, body fat was measured by traditional methods: under-
with a standard physician’s office scale, with subjects in a hos-
water weighing (UWW), total body water, or total body potas-
pital gown and no shoes, and height to ±0.5 cm with a height-
sium (3-5). However, the accuracy of these methods is affected
bar fixed on the wall, with subjects standing with back, buttocks,
not only by sex and age but also by race (6). Therefore, evalua-
and heels together pressed to the wall. Other anthropometric mea-
tion of BMI as an index for body fat requires an update using
modern methods that include adjustment for race. Second, the
I From the Body Composition Unit, Center for Research in Clinical
Asian population is the largest population in the world and the
second fastest growing minority group in the United States, next Nutrition, St Luke’s-Roosevelt Hospital Center, Columbia University,
New York.
only to Hispanics (7). Significant differences in anthropometrics
Supported
2 by NIH grant DK-37352.
between races other than Asians have been well recognized, but
3 Address reprint requests to J Wang, Body Composition Unit, St
reference data on anthropometry in Asians in the literature is
Luke’s-Roosevelt Hospital Center, Columbia University, Amsterdam
limited to young subjects, and most of the data were obtained Avenue at 1 14th Street, New York, NY, 10025.
more than two decades ago (8- 12). Comparisons between whites Received June 4, 1993.
and Asians are few. A recent report by Jiang et al (13) made Accepted for publication January 4, 1994.

Am J Clin Nutr l994;60:23-8. Printed in USA. © 1994 American Society for Clinical Nutrition 23
24 WANG ET AL

TABLE I
Physical characeristics of studied subjects’

Males Females

Whites Asians Whites Asian

Variable (n = 187) (n = 1 10) (,‘ = 258) (P1 = 132)

Age(y) 51±19 52±18 51±19 51±17

Height (cm) 175 ± 7 170 ± 62 163 ± 7 157 ± 72

Weight (kg) 77 ± I I 68 ± 10’ 61 ± 9 54 ± 82

BMI3 25.1 ± 3.0 23.4 ± 3.0 23.9 ± 3.4 22.5 ± 334
Percent body fatS (%) 19.3 ± 6.4 21.4 ± 6.36 30.1 ± 8.7 31.6 ± 6.5

‘1± SD.
3 In kg/m’.
5 Measured by dual-photon absorptiometry.
2.4.6.7 Significantly different from whites of same sex: 2 p 0.0001, p = 0.002, p 0.006, p = 0.05.

surements were measured as described by Steinkamp et al (15): Skinfold-fat free area = pi X skinfold fat-free diameter’/4 (4)
eight skinfold-thickness measurements to ±2 mm with a Lange
caliper, and six circumferences to ±5 mm with a flex-resisting Skinfold-fat area = total area - skinfold fat-free area (5)
heavy-duty sewing tape. The average of two readings was re-

Downloaded from www.ajcn.org by guest on July 14, 2011

corded. More than 90% of the subjects were measured by one where the skinfold-thickness measurement for midarm was the
investigator, and the rest by another investigator after 1 -mo train- average of triceps and biceps; for trunk was the average of chest,
ing with validation. subscapular, umbilicus, abdomen, and suprailiac; and for thigh
Two DPA systems (Lunar DP4 and Lunar DPX) were used for was for the midthigh.
measuring fat%. Both systems were calibrated by ground-beef The arm circumference was the measured midarm circumfer-
standards as previously described, and the measurement differ- ence; the trunk circumference was the average circumference of
ences between these two systems were rationalized by a highly upper chest, chest, waist, and iliac crest; and the midthigh cir-
satisfactory linear-regression model (16): cumference was the thigh circumference. Equations I -4 were
generated based on the following three basic equations: area
Fat%DPA 0.95 x fat%DPX 0.8802. = (circumference)2/4 pi, or area = pi X (diameter/2)’, and di-
ameter = circumference/pi.
This model has an r of 0.94 and an SEE of 2.6%, and the
slope is not different from 1 (P = 0.06).
Statistical analysis
Skinfold-fat-area calculation
Regression analyses were used to investigate the relationships
The total area, the skinfold-fat area, and the skinfold fat-free of BMI to fat%, and to investigate the improvement of prediction
area at the midarm, the trunk, and the midthigh were calculated from adding measured skinfolds. The level of significance for all
by the following equations: statistical tests was set at 0.05. Statistical calculations were per-
formed by using STATA (Computing Resource Center, Santa
Total diameter = circumference/pi (1) Monica, CA) and SAS (SAS Institute Inc., Cary, NC) software
packages.
Total area = circumference2/4 pi (2)

Skinfold-fat free diameter

TABLE 3
Estimated percent body fat (fat %) at three levels of BMI by using
= total diameter - skinfold thickness (3)
linear equations relating BMI to fat % in whites (W) and Asians (A)

Fat %
TABLE 2
Linear equations relating percent body fat to body mass index in
BMI = 15 BMI = 25 BMI = 35
whites and Asians
Sex and race (Lean) (Normal) (Obese)
Coefficients
,.22 SEE %
Sex Race BMI’ Constant r’

Males
M White 1.270 -12.58 0.35 0.35 5.2
Whites 6.5 19.2 31.9
M Asian 1.356 -10.31 0.40 0.39 4.9
Asians 10.0 23.6 37.2
F White I .923 - I 3.85 0.56 0.56 5.7
Differences (A-W) 3.5 4.4 5.3
F Asian 1.638 -4.13 0.55 0.54 4.4
Females
Whites 15.0 34.2 53.5
I In kg/m’. Asians 20.4 36.8 53.0
2 Adjusted value was adjusted for number of variables used in the
Differences(A-W) 4.6 2.6 -0.3
model.
 BODY MASS INDEX IN WHITES AND ASIANS 25

TABLE 4 TABLES
Comparisons of circumferences at six sites between whites and Asians’ Comparisons for skinfold-thickness measurements at eight sites
between whites and Asians’
Males Females
Males Females
Whites Asians White Asian
Upper chest (n = 187) (n = 1 10) (n = 258) (n = 132) Whites Asians White Asian
Variables (n = 187) (n = 1 10) (n = 258) (n = 132)
mm
mm
Arm 310 ± 34 297 ± 332 273 ± 32 269 ± 29
Upperchest 1007 ± 67 946 ± 60’ 861 ± 55 838 ± 553 Triceps 12±5 13±5 21±7 22±7
Chest 993 ± 73 920 ± 66 890 ± 80 854 ± 70 Biceps 4±3 5#{247}32 8±5 953
Waist 878 ± 89 831 ± 84 734 ± 85 731 ± 84 Chest 11±6 11±6 12±8 13±7
Iliac crest 937 ± 83 897 ± 72 919 ± 85 887 ± 733 Umbilicus 25 ± 12 24 ± 11 25 ± 13 29 ± ll
Thigh 515 ± 41 492 ± 4l 494 ± 44 467 ± 4&’ Abdomen 18 ± 10 20 ± l0 25 ± 12 29 ± 11’
Suprailiac 11 ± 8 13 ± 8 13 ± 8 18 ± 9
i± SD. Subscapular 15 ± 7 17 ± 7 14 ± 8 19 ± 88
2.3 Significantly different from whites of same sex: 2 p 0.0014, -‘ p Thigh 14±6 13±5 26±10 22±7k
= 0.0001.
‘1± SD.
2-9 Significantly different from whites of same sex: 2 p 0.0194, p
Results = 0.0148, P = 0.0008, P = 0.0427, p 0.0015, P = 0.0197, 8 p
= 0.0001, P = 0.0034.

Downloaded from www.ajcn.org by guest on July 14, 2011

Table 1 shows that there were no significant differences in age
between study cohorts by either sex or race. Whites were signif-
icantly taller, heavier, and had higher BMIs than Asians, but Comparisons for skinfold-fat area at midarm, midthigh, and
Asians were significantly fatter than whites in both sexes. trunk are summarized in Table 7. In males, the only differ-
The linear relationships between BMI and fat% were sig- ences between whites and Asians were in the thigh, which
nificant for all studied groups, by sex and race (Table 2). To tended to be larger in whites than in Asians (P = 0.06). In
demonstrate variation in the relationship between fat% and females, Asians had larger trunk skinfold-fat area but whites
BMI at different BMIs, Table 3 shows the estimated fat% us- had larger thigh skinfold-fat area. Arm fatness did not differ
ing equations in Table 2 for simulated subjects in each sex of between the two races. Females had larger skinfold-fat areas
the two studied races at three different BMIs (BMI = 15 for than did males in both races, except for the trunk, where in
lean, BMI = 25 for normal, and BMI = 35 for obese). For whites there were no significant differences between males
lean and normal BMIs, Asians were fatter than whites in both and females.
sexes, but the differences in estimated fat% between whites When subcutaneous-fat area was expressed as a proportion
and Asians varied by BMI in different directions for males of the total cross-sectional area (in midarm, trunk, and mid-
and females: fat% increased with BMI for males but decreased thigh), Asians had significantly higher values than whites in
with BMI for females. arm and trunk in both sexes (Table 8). But in the thigh, white
Comparisons of anthropometric measurements were made to females had significantly higher values than Asian females.
explore the reasons for the differences between the two races. There were no significant differences between races in
Table 4 shows the comparisons for circumference measurements. males.
All six measured circumferences in white males were signifi- Comparisons for the ratios of fat-free mass (FFM) to height
cantly larger than in Asian males. In females, arm and waist between study groups are also shown in Table 8, where FFM was
circumferences were not different between whites and Asians, calculated from the standard equation
but other circumferences were larger in whites.
Comparisons in subcutaneous-fat thickness (skinfold thick-
ness) showed that subcutaneous-fat distributions varied by sex TABLE 6
and race: Asians had more upper-body fat (trunk) and white fe- Significant differences (P < 0.05) in skinfold-thickness measurements
males had more lower-body fat (thigh) (Table 5). Asians had at eight sites between whites (W) and Asians (A) when adjusted for
body fat at 20% in males and 30% in females
significantly thicker biceps, abdomen, subscapular and suprailiac
skinfolds than did whites in both sexes. In the thigh, subcutane- Variable Males’ Females’
ous fat was significantly higher in white females. There were no
significant differences in triceps and chest skinfold-thicknesses Triceps W=A W=A
in males. Biceps W<A W<A
Subcutaneous-fat thicknesses were also compared when Chest W=A W=A
Umbilicus W = A W < A
skinfold thicknesses were adjusted for the same level of fat%
Abdomen W<A W<A
in each sex between the two races. In this instance skinfold
Suprailiac W < A W < A
thicknesses were adjusted for fat at 20% and 30% in males
Subscapular W < A W < A
and females, respectively (Table 6). Differences in all skinfold Thigh W=A W>.A
thicknesses for females and in most skinfold thicknesses for
males have similar patterns as for the entire study group as ‘n = 187W, 110 A.
listed in Table 5. 2, 258W, 132 A.
26 WANG ET AL

TABLE 7
Comparisons of skinfold-fat areas between Asians and whites’

Males Females

Whites Asians Whites Asian

Variables (n = 187) (n = 1 10) (n = 258) (n = 132)

Midarm 1238 ± 613 1275 ± 512 1816 ± 847 1913 ± 808

Trunk 7530 ± 3969 7641 ± 3479 7601 ± 4217 8807 ± 3799
Midthigh 3421 ± 1584 3079 ± 1461 6042 ± 2546 4858 ± l807
‘ .F ± SD. Arm, crosssection at midarm; trunk, average crosssection at trunk; thigh, crosssection of midthigh.
2 p 0.002.

3P = 0.0001.

FFMkg body weight,g X (1 fat%DPA/lOO) dexes by Smalley et al (2). The comparisons are summarized in
Table 10.
Whites had higher ratios of FFM to height in both sexes, and
males’ ratios were higher than those of females in both races.
In model development, age and skinfold-thickness measure- Discussion
ments significantly improved the BMI-based fat% prediction
equations for each of the studied cohorts from SEEs ranging from Most reported equations to predict body fat from weight and

Downloaded from www.ajcn.org by guest on July 14, 2011

4.4-5.7% to 3.5-4.4% (Table 9), but the degree of improvement stature have been developed in white populations. Our data show
varied widely. BMI had higher coefficients for females than for that substantial errors occur when these equations are applied to
males in both races. Age coefficients were the lowest in the four Asians, because many factors affect body-fat content besides
groups. Thigh skinfold-thickness measurements improved all weight-to-height relationships. Hampton et al (19) realized that
equations, whereas triceps improved all but that for white males. equations developed in adults are inappropriate for children in
Others varied by sex and race. All equations have comparable r screening for obesity. Lohman (1, 20), Roche (21), Garrow (22),
and SEE values. Addition of (BMI)2 did not improve any of the Smalley et al (2), and many others called for special caution in
BMI-based fat% prediction equations, which suggested that BMI using general prediction equations for any clinical application.
had simple linear relationships with fat% in the current study. The validity of prediction equations must be tested by using cor-
We have also tested the hypothesis that there is a general pre- relations with fat% as the ruler, and the method for measuring
diction equation for each sex in these two races by assuming in body fat should be independent of sex, age, and race.
Table 9 the following: equation for white males = equation for A recent report by Barlett et al (23) studied the fat-free mass
Asian males and equation for white females = equation for Asian in relation to stature in 1 103 subjects from childhood to old age
females. The test gave P = 0.2 for hypothesis 1 and P = 0.0001 by pooling many studies over the past 20 y, most of them in
for hypothesis 2. Therefore, there was no general equation for whites (23). They used the traditional UWW method for esti-
each sex in the two races studied. mating fat%, and then calculated the FFM whereas we used DPA.
The equations listed in Table 9 were compared with the equa- As the authors pointed out, errors may occur in the estimation of
tions in the literature for predicting fat% by UWW by using the fat% with UWW, to the extent that the density of FFM varies.
sum of chest, triceps, and subscapular skinfold thicknesses and DPA does not rely on that traditional assumption of chemical
age by Jackson and Pollock (17); the sum ofbiceps, triceps, sub- consistency (14). Comparisons between DPA and UWW in our
scapular, and suprailiac skinfold thicknesses for subjects 17-76 recent study documented that UWW underestimates fat% in
y old by Womersley and Durnin (18); using weight-height in- young subjects and overestimates fat% in elderly people (6).

TABLE 8
Comparisons of ratios of skinfold-fat area with total cross-sectional area in arm, trunk, and thigh, and ratios of fat-free mass (FFM) to height
between Asians and whites’

Males Females

Whites Asians Whites Asian

Variable (n = 187) (n = 1 10) (n = 258) (n = 132)

Arm 0.158 ± .062 0.179 ± .0582 0.291 ± .08 0.318 ± .076

Trunk 0. 100 ± .041 0. 1 15 ± 0.04l 0. 123 ± .05 1 0. 156 ± .047
Thigh 0. 161 ± .066 0. 155 ± .054 0.301 ± .085 0.274 ± 0743
FFMJht 30.5 ± 3.9 26.2 ± 3,46 20.9 ± 3.6 19.9 ± 4,97

‘ I ± SD. Arm, crosssection at midarm; trunk, average crosssection at trunk; thigh, crosssection at midthigh; FFM/ht, ratio of fat-free mass to height
(kg/cm).
2-7 Significantly different from whites of same sex: P = 0.004, P = 0.001, P = 0.002, P = 0.0001, 6 p 0.000, ‘ p = 0.023.
 BODY MASS INDEX IN WHITES AND ASIANS 27

TABLE 9
Equations for predicting percent body fat’

Coefficients

Sex Race BMI Age Tri Scp Umb Sic Abd Thi Constant r r22 SEE

M White 0.399 0.067 - 0.128 0.124 - 0.133 0.257 -4.98 0.71 0.70 3.5
M Asian 0.471 0.082 0.327 - - - 0.132 0.289 -4.40 0.67 0.66 3.7
F White 0.682 0.086 0.247 - - - 0.197 0.170 -4.40 0.75 0.75 4.4
F Asian 0.899 0.029 0.279 -0.1 17 - 0.172 - 0.188 -0.57 0.73 0.72 3.5

‘ Tri, tricep; Scp, subscapular; Umb, umbilicus; Sic, suprailiac; Abd, abdomen; Thi, thigh.
2 Adjusted ,2 value was adjusted for number of variables used in the model.

Thus, the relationships between BMI and either fat% or FFM The current study is limited to anthropometnic measurements
estimated by UWW will be in error because of the methodolog- for investigating fat distribution. The anthropometric methods,
ical limitations of UWW. being entirely regional, are less accurate than, for example, mul-
The current study documents differences in fat distribution tislice CT. The fat areas calculated from anthropometry in this
between whites and Asians and among the four study groups, report may not be accurate but the technique is appropriate for
by sex and race. Shorter stature and smaller body circumfer- comparisons for population studies. It would be ideal to include
ences (except in the waist) in Asians were expected. But the CT in future studies to investigate the fat distribution between

Downloaded from www.ajcn.org by guest on July 14, 2011

higher fat% and the thicker subcutaneous fat in Asians for a subcutaneous and mesenteric for the two races.
given BMI at most of the measured sites (excepting the thigh) There are many prediction equations in the literature but most
are original observations. In this respect, Asians are similar to of these equations are for predicting UWW-measured fat% in
Mexican-Americans, who have more trunk subcutaneous fat whites. We have documented that UWW-measured fat tended to
than do whites, but the other subcutaneous-fat distributions in be lower than the DPA-measured fat, and the differences are sex-,
Asians are different from Mexican-Americans aged 2- 17 y age-, and race-dependent (6). Comparisons between prediction
old (24). equations for DPA-measured fat developed from this study (Ta-
Anthropometric methods for estimating fat and muscle areas ble 9) and equations for UWW-measured fat developed by
have been widely used for nutritional assessments since Jackson and Pollock (17) and Womersley and Durnin (18) show
Frisancho’s report (25). However, computed tomography (CT) indeed that their predicted fat% values are significantly lower
measurements indicate that anthropometry overestimates muscle than ours (Table 10). These results agree with our previous find-
and bone area in the arms, and therefore might underestimate the ings (6). The prediction equations based on weight-height in-
fat (26, 27). dexes developed by Smalley et al (2) give much larger fat values
than do our equations. The differences are larger in females than
in males, and the largest differences are in Asians.
TABLE 10
Acceptance of an anthropometric method for nutritional-status
Comparison of percent body fat with three prediction equations, by
assessment in field studies will be determined by the simplicity
race and sex
of the method as well as by its accuracy. However, simplicity
Method, sex, and race Mean SD P that provides invalid results provides no benefit to health care
services. Indeed, a technique that was consistently inaccurate in
Womersley and Durnin (18)
predicting absolute fat content, but which was precise (repeata-
Female
ble) and correlated well with changes in fat, would be acceptable
Asian -0.93 4.01 0.009
White -2.90 5.05 0.0001 and have high utility.
Male BMI has been used by many investigators for estimating body
Asian -0.76 4.27 0.06 fatness in the past. However, many recent publications have dem-
White -1.02 4.19 0.001 onstrated the unreliability of BMI for estimating body fat. Be-
Smalley et al (2) cause body weight is the sum of fat, muscle, visceral organs, and
Female bone, subjects with long trunks and short legs for height have
Asian -9.69 4.52 0.0001 higher BMIs regardless of their fat content (1). It is well known
White -7.40 6.08 0.0001
that the length of trunk and legs relative to height varies by race.
Male
Asian This study highlights the necessity for race specificity in meth-
-4.43 4.95 0.0001
White -1.17 5.27 0.003 ods for estimating body composition by anthropometric methods.
Jackson and Pollock (17) We present two levels of BMI-based anthropometric methodol-
Female ogy for estimating fat% in whites and Asians: equations based
Asian -2.93 4.29 0.0001 on BMI alone require less technical expertise but give relatively
White -4.18 4.85 0.0001 large SEEs (>4.4%), which may be sufficient for field studies
Male
(Table 2), and equations requiring additional selected skinfold-
Asian -2.72 4.14 0.0001
thickness measurements, which provide more accurate results
White -2.14 4.11 0.0001
suitable for clinical studies (Table 9). U
 28 WANG ET AL

References 15. Steinkamp RC, Cohen NL, Siri WB, Sargent W, Walsh HE. Mea-
surement of body fat and related factors in normals-Il. J Chron Dis
1965;l8:1292-307.
1 . Lohman TG. Advances in body composition assessment, current is-
16. Russell-Aulet M, Wang J, Thornton JC, RN Pierson Jr. Comparison
sues in exercise science series. Champaign, IL: Human Kinetics Pub-
of dual-photon absorptiornetry systems for total body bone and soft
lishers, 1992.
tissue measurements: dual-energy x-rays versus gadolinium 153. J
2. Smalley IU, Knerr AN, Kendrick ZV, Colliver JA, Owen OE. Reas-
Bone Miner Res 199l;6:4l 1-S.
sessment of body mass indices. Am J Clin Nutr 1990;52:405-8.
17. Jackson AS, Pollock ML. Practical assessment of body composition.
3. Pace N, Rathbun EN. Studies on body composition; body water and
Physician Sport Med 1985;l3:76-90.
chemically combined nitrogen content in relation to fat content. J
18. Womersley J, Durnin JVGA. A comparison of the skinfold
Biol Chem l945;l58:685-9l.
method with extent of ‘overweight’ and various weight-height
4. Forbes GB, Lewis AM. Total sodium, potassium and chloride in
adult man. J Clin Invest 1956;6:596-600. relationships in the assessment of obesity. Br J Nutr
1977;38:27l -84.
5. Behnke AR, Feen BG, Welham WC. Specific gravity of healthy
man. J Am Med Assoc 1942;l 18:495-8. 19. Hampton MC, Huenemann RL, Shapiro LR, Mitcheal BW, Behnke
6. Wang J, Russell-Aulet M, Mazariegos M, et al. Body fat by dual AR. A longitudinal study of grossbody composition and body con-
photon absorptiometry (DPA): comparisons with traditional meth- formation and their association with food and activity in a teen-age
ods in Asians, Blacks and Caucasians. Am J Human Biol population. Am J Clin Nutr l966;19:422-35.
l992;4:501 -10. 20. Lohman TC. Preface to body composition assessment: a re-evalua-
7. US Department of Commerce. Census Bureau Releases 1990 census tion of our past and a look toward the future. Med Sci Sports Exerc
counts on specific racial groups. United States Department of Corn- 1984;16:578-53.
rnerce News June 12, 1991. (CB91-215.) 21. Roche AF. Anthropometric methods: new and old, what they tell us.
8. Chen KP, Damon A, Elliot 0. Body form, composition and some Int J Obes 1984:8:09-11.
physiological functions of Chinese on Taiwan. l963;ll0:760-77. 22. Garrow JS. Treatment of obesity. Lancet 1992;340:409- 13.

Downloaded from www.ajcn.org by guest on July 14, 2011

9. Chien 5, Peng MT. Chen KP, Huang TF, Chang C, Fang HS. Lon- 23. Barlett H, Puhl SM, Hodgson JL, Buskirk ER. Fat-free mass in re-
gitudinal studies on adipose tissue and its distribution in human sub- lation to stature: ratios of fat-free mass to height in children, adults,
jects. J Appl Physiol l975;39:825-30. and elderly subjects. Am J Clin Nutr 1991;53:l I 12-6.
10. Fry PC, Fox HM, Fry El. Effect of diet on skinfold measurements 24. Martorell R, Mama RM, Castillo RO, Mendoza FS, Pawson 1G.
and creatinine and urea excretion of Hong Kong Chinese children. Body proportions in three ethnic groups: children and youths
Am J Clin Nutr l968;2l:l 197-201. 2- 17 years in NHANES II and HHANES. Hum Biol l988;60:205-
1 1. Yoshinori S. Studies on the obesity incidence and the skinfold thick- 22.
ness of pupils in northern Taiwan and in northern Kyushu (Japan). 25. Frisancho AR. New norms of upper limb fat and muscle areas for
J Formos Med Assoc 1982;81 :997-1011. assessment of nutritional status. Am J Clin Nutr 1981 ;34:2540-S.
12. Wang WC, Kou DH. Estimation of body fat in Chinese young men. 26. Forbes GB, Brown MR. Griffiths HJL. Arm muscle plus bone area:
Aerospace Med l974;45:307-9. anthropometry and CAT scan compared. Am J Clin Nutr
13. Jiang ZM, Yang NF, Chou C, et al. Body composition in Chinese l988;47:929-3 1.
subjects: comparison with data from North America. World J Surg 27. Heymsfield SB, McManus C, Smith J, Stevens V, Nixon DW. An-
l99l;lS:95- 102. thropometric measurement of muscle mass: revised equations for
14. Mazess RB, Peppler WW, Gibbons H. Total body composition by dual- calculating bone-free arm muscle area. Am J Clin Nutr
photon (#{176}3Gd)
absorptiometry. Am J Clin Nutr 1984;40:834-9. 1982;36:680-90.

View publication stats