Comparison of predicted body fat percentage from anthropometric methods and from impedance in university students.

Abstract

The objective of the present study was to compare different methods for evaluating body fat percentage (BF%) (anthropometric methods and bioelectrical impedance analysis) in university students. Subjects were 653 healthy students whose mean age, body height, body weight and BMI were 21.1 (SD 2.5) years, 166.0 (SD 8.4) cm, 62.8 (SD 11.0) kg and 22.7 (SD 3.1) kg/m(2), respectively. Results showed that BMI is a poor predictor of body fatness since the sensitivity was low in comparison with the reference method (Siri equation). The lowest values of BF% were obtained using the reference method (Siri equation) (21.8 (SD 6.8)%). The two methods with the highest agreement were Siri and Lean (mean difference, -0.5), followed by Brozek (mean difference, -1.4) and Deurenberg (mean difference, -1.5). The largest mean difference for BF% was between Siri and impedance (-4.5). Although the methods and/or equations used in the present study have been commonly utilised to estimate BF% in young adults, the results must be interpreted with caution in the diagnosis and monitoring of overweight and obesity.

Full text

Page 1

Comparison of predicted body fat percentage from anthropometric
methods and from impedance in university students
Marta Arroyo1*, Ana M. Rocandio1, Laura Ansotegui1, Hector Herrera2, Itziar Salces2and
Esther Rebato2
1Department of Nutrition and Food Science, Faculty of Pharmacy, University of the Basque Country, Vitoria-Gasteiz, Spain
2Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Sciences,
University of the Basque Country, Bilbao, Spain
(Received 12 February 2004 – Revised 29 July 2004 – Accepted 3 August 2004)
The objective of the present study was to compare different methods for evaluating body fat percentage (BF%) (anthropometric methods
and bioelectrical impedance analysis) in university students. Subjects were 653 healthy students whose mean age, body height, body
weight and BMI were 21·1 (SD 2·5) years, 166·0 (SD 8·4) cm, 62·8 (SD 11·0) kg and 22·7 (SD 3·1) kg/m2, respectively. Results showed
that BMI is a poor predictor of body fatness since the sensitivity was low in comparison with the reference method (Siri equation).
The lowest values of BF% were obtained using the reference method (Siri equation) (21·8 (SD 6·8) %). The two methods with the highest
agreement were Siri and Lean (mean difference, 20·5), followed by Brozek (mean difference, 21·4) and Deurenberg (mean difference,
21·5). The largest mean difference for BF% was between Siri and impedance (24·5). Although the methods and/or equations used in the
present study have been commonly utilised to estimate BF% in young adults, the results must be interpreted with caution in the diagnosis
and monitoring of overweight and obesity.
Body fat: Bioelectrical impedance analysis: Anthropometry: University students
Given the rising incidence of obesity in the young
population of Western countries and, therefore, the import-
ance of measuring body fat, there has been a resurgence of
interest in the evaluation of different body composition
methods (Gruber et al. 2001; Kitano et al. 2001).
The World Health Organization (1995, 1998) defines
overweight and obesity at BMI cut-off points of 25 and
30kg/m2, respectively, in adult populations. However,
there is increasing evidence that these cut-off values are
not valid for all populations (Luke et al. 1997; Deurenberg
et al. 1998; Deurenberg-Yap et al. 2000) as the relationship
between BMI and body fat percentage (BF%) differs
between population groups. Furthermore, it is the amount of
body fat, rather than the amount of excess weight, that
determines the health risks of obesity (World Health
Organization,1998).This
interest of scientists and the general public in body fat
measurements.
AlthoughthereareseveralmethodstoestimateBF%,there
is no ‘gold standard’ for both epidemiological studies and
personal use. However, some scientific societies such as
the Spanish Society for Obesity Research recommend the
explainstheincreasing
Siri equation, based on anthropometric measures, to deter-
mine BF% (SEEDO, 1996). There are other more accurate
methods, for example, the underwater weighing method or
dual-energy X-ray absorptiometry, but their cost and com-
plexity limit widespread use (Bray et al. 1998).
Comparisons between body composition methods have
been made in healthy populations of both children (Ellis,
1996; Gutin et al. 1996; Treuth et al. 2001; Fors et al.
2002) and adults (Heymsfield et al. 1990; Wellens et al.
1994). Although there is literature on the differences in
body composition between university athletes and non-
athletic subjects (Emslander et al. 1998; Mitsuzono &
Komiya, 1991), as far as we know, no study has compared
the estimates of BF% by using different methods on
university students.
Therefore, the purpose of the present study was to com-
pare estimates of BF% by different methods (anthropo-
metry and bioelectrical impedance analysis) in university
students. Additionally, we investigated the association
between BF% and BMI to evaluate the screening
performance of the BMI focused on individual preventive
medicine.
*Corresponding author: Dr Marta Arroyo, fax þ34 945 013014, email knparizm@vc.ehu.es
Abbreviations: BF%, body fat percentage; BF%BROZEK, body fat percentage predicted by Brozek equation; BF%DEURENBERG, body fat percentage
predicted by Deurenberg equation; BF%IMP, body fat percentage evaluated by impedance; BF%LEAN, body fat percentage predicted by Lean
equation; BF%SIRI, body fat percentage predicted by Siri equation; WHR, waist:hip ratio.
British Journal of Nutrition (2004), 92, 827–832
q The Authors 2004
DOI: 10.1079/BJN20041273

Page 2

Methods
Subjects
A cross-sectional study of students from the University of
the Basque Country (Spain) was carried out. The sample
was composed of 653 individuals (190 males and 463
females) aged 18–30 years and recruited from different
degrees. The sample size was considered as representative
according to the estimations performed by means of the
formula by Martin & De Dios (1993) and according to
the deviation of BF% from a previous study in populations
of the Basque Community (Servicio Central de Publica-
ciones del Gobierno Vasco, 1994). A stratified recruitment
design was used according to demographic data (age, sex
and number of students per campus).
Theaimofthepresentstudyandthekindofmeasurements
were explained to the participants, who gave their informed
written consent. The experimental protocol was approved
by the University Ethical Committee on Human Research.
All measurements were done on the same visit at the
nutrition and physical anthropology laboratories (Univer-
sity of the Basque Country) and at least 3h after a meal
(including drink). Apart from these measurements, the sub-
jects were requested to refrain from strenuous exercise 12h
before the measurements and they were asked to empty
their bladders before the evaluation. Females were not
measured during their menstrual period.
Anthropometric measurements
A well-trained anthropometrist performed all the measure-
ments. Body weight was measured to 0·1kg usinga standard
beam balance (An ˜o ´-Sayolw; Atla ´ntida, An ˜o ´ Sayol, Barce-
lona, Spain). Height was measured to the nearest 1mm
using a Harpenden stadiometer (Holtain Ltd, Crymych,
Wales, UK). With these data we calculated the BMI
(weight(kg)/height(m)2).
(biceps, triceps, subscapular and supra-iliac) were measured
in duplicate on the left side of the body to the nearest 0·1mm
with a Holtain skinfold caliper (Holtain Ltd).
Circumferences of the waist and hip were also taken in
duplicate to the nearest 1mm with a tape measure. BF%
was calculated using some prediction equations from the lit-
erature:Siri(1961),Brozek(Brozeketal.1963),Deurenberg
(Deurenberg et al. 1991) and Lean (Lean et al. 1996)
equations (BF%SIRI, BF%BROZEK, BF%DEURENBERG and
BF%LEANrespectively). In the Siri and Brozek equations,
density was predicted using Durnin & Womersley’s formula
(Durnin & Womersley, 1974).
The subjects were classified according to BF% using the
criteria of Bray et al. (1998) for the classification of obes-
ity. BMI was classified according to the categories of obes-
ity and overweight of the World Health Organization
(1998). Additionally, we evaluated the regional adiposity
using the waist:hip ratio (WHR) and the waist circumfer-
ence. The volunteers were classified at risk according to
Heymsfield et al. (1998) and National Institutes of Health
criteria (National Institutes of Health, National Heart,
Lung and Blood Institute, 1999), respectively. The descrip-
tive statistics of the anthropometric traits and the age of the
subjects are displayed in Table 1.
The skinfoldthicknesses
Bioelectrical impedance analysis
Bioelectrical impedance analysis measurements were per-
formed using a tetrapolar multi-frequency impedanci-
ometer (MediSystem-SanoCare Human Systems S.L.,
Madrid, Spain). All measurements were performed in
accordance with the manufacturer’s instruction manual.
BF% evaluated by impedance (BF%IMP) was estimated
using Lohman’s formula (Lohman, 1992) based on pre-
entered personal particulars (weight, height, age and sex)
and impedance value.
Statistical analysis
Data were gathered using SPSS 10.0 (SPSS Inc., Chicago,
IL, USA) with significance set at P,0·05 and presented as
mean values and standard deviations. Bland–Altman ana-
lysis (Bland & Altman, 1986) was used to test for bias
(mean difference) and limits of agreement among all the
methods. Measures of BF% from the Siri equation were
used as the reference method, according to the recommen-
dation of the Spanish Society for Obesity Research
(SEEDO, 1996). Sensitivity, specificity and predictive
values were calculated to evaluate the classification of
obesity using the BMI as compared with the reference
method. In the present study, test sensitivity was the pro-
portion of obesity cases, as diagnosed by the reference
method, found by BMI. Specificity refers to the proportion
of subjects identified by the reference method as non-obese
and that BMI classified correctly. The positive predictive
value is the probability that a student classified as obese
by BMI actually is found to be so by the reference
method. The negative predictive value gives us the prob-
ability that a subject classified as non-obese by BMI is
also defined as non-obese by the reference method.
Results
Subjects’ characteristics
In total, 653 subjects participated in the present study, the
463 females ranging in age from 18 to 30 years, in BMI
from 16·7 to 34·2kg/m2and in BF%SIRIfrom 12 to 41%.
The 190 males ranged in age from 18 to 29 years, in
BMI from 17·5 to 38·6kg/m2and in BF%SIRIfrom 7 to
31%. The characteristics of the subjects are given in
Table 1. An acceptable BMI existed for 75·7% of the sub-
jects and 3·8% of the total sample were classified as low
weight (one male and twenty-four females). Significant
differences between the sexes were found for all par-
ameters (P,0·01), except for subscapular skinfold.
According to BF%SIRI, 6·1% of the subjects were classi-
fied as obese (seven males and thirty-three females) and
9·7% as overweight (twenty-four males and thirty-nine
females). However, according to the BMI classification,
2·5% were obese (eight males and eight females) and
17·2% of the total sample were overweight (forty-five
males and sixty-seven females).
The results showed the strongest specificity (1) and posi-
tive prediction (1) of BMI in the identification of obese
subjects (BMI $30·0kg/m2). However, the sensitivity
was low (0·4) and negative predictive value was 0·96.
M. Arroyo et al. 828

Page 3

In the total sample, 3·7% of the subjects would be falsely
classified as non-obese with BMI.
According to the waist circumference, 9·8% of the total
were classified as at risk (3·7% of the males and 12·3% of
the females). Additionally, according to the WHR, 31·2%
of the total sample were classified as at risk (2·1% of the
males and 43·2% of the females).
Comparisons of the body fat assessment methods
For the overall female population BF%BROZEK(26·2 (SD
4·2)), BF%DEURENBERG (26·0 (SD 3·3)), BF%LEAN (25·2
(SD 4·0)) and BF%IMP(29·0 (SD 4·1)) were significantly
different from BF%SIRI (24·3 (SD 5·8)) (P,0·001). For
the overall male population BF%BROZEK(16·1 (SD 4·4)),
BF%DEURENBERG(16·7 (SD 3·6)) and BF%IMP(20·0 (SD
5·3)) were significantly different from BF%SIRI(15·9 (SD
5·0)) (P,0·01).
The lowest values of BF% were obtained with Siri (21·8
(SD 6·8) %), followed by Lean (22·3 (SD 6·2) %), Brozek
(23·3 (SD 6·3) %) and Deurenberg (23·3 (SD 5·4) %),
with impedance giving the highest value of percentage
fat (26·4 (SD 6·1) %). The percentage fat mass from the
Lean and Deurenberg equations suggested that 6·9 and
10·0% were overweight or obese, respectively. BF%
from the Brozek equation and from impedance indicated
a high tendency towards overweight or obesity (15·6 and
36·5%, respectively).
When obesity was defined using the BF%SIRIin males,
Deurenberg and Brozek equations underestimated BF% at
higher values of body fat (BF% $25) (BF%SIRI, 28·1 (SD
2·6); BF%DEURENBERG, 25·6 (SD 4·3); BF%BROZEK, 26·6
(SD 2·2)) (in all cases, P,0·05). In males, the Lean, Deur-
enberg and Brozek formulas underestimated BF% at higher
values of body fat (BF% $33) (BF%SIRI, 35·4 (SD 2·1);
BF%LEAN, 30·3 (SD 3·9); BF%DEURENBERG, 31·7 (SD 3·2);
BF%BROZEK, 34·0 (SD 1·7)) (in all cases, P,0·001).
Limits of agreement
The biases (mean differences) for BF% and the limits of
agreement between the five methods are shown in
Table 2. For BF%, the two methods with the highest
agreement were Siri and Lean, the mean difference being
Table 1. Anthropometric data
(Mean values and standard deviations)
Total (n 653)Males (n 190)
Females
(n 463)
Mean
SD
Mean
SD
Mean
SD
Age (years)
Weight (kg)
Height (cm)
BMI (kg/m2)
Waist circumference (cm)
Hip circumference (cm)
WHR
Skinfolds (mm)
Triceps
Biceps
Subscapular
Supra-iliac
Total of four skinfolds (mm)
21·1
62·8
166·0
22·7
80·5
94·5
0·9
2·5
11·0
8·4
3·1
7·7
6·9
0·1
21·2
73·1
175·2
23·9
83·4
95·7
0·9
2·3
9·9
6·8
3·0
8·0
6·8
0·1
21·0
58·6
162·3
22·2
79·4
94·1
0·8
2·6
8·4
5·8
3·0
7·2
6·9
0·1
16·6
9·6
14·0
16·3
56·5
6·3
4·5
5·6
7·6
20·8
11·0
7·0
13·5
14·8
46·4
5·1
3·7
5·7
7·6
19·7
18·9
10·7
14·2
16·9
60·6
5·3
4·4
5·6
7·5
19·8
WHR, waist:hip ratio.
Table 2. Assessment of agreement between the Siri (1961) equation and the other methods used
to estimate body fat percentage
Limits of agree-
ment*
Comparison Mean difference*95% CI
SD*† CI*LowerUpper
Siri–Brozek‡
Siri–Deurenberg§
Siri–Lean{
Siri–IMP
21·4
21·5
20·5
24·5
21·7, 21·2
21·8, 21·2
20·9, 20·1
24·8, 24·1
2·9
4·2
4·8
4·3
11·4
16·8
19·3
17·2
27·2
29·9
210·2
213·1
4·3
6·9
9·1
4·2
IMP, impedance analysis.
*Bland–Altman method.
† SD of difference.
‡Brozek et al. (1963).
§Deurenberg et al. (1991).
{Lean et al. (1996).
Body fat percentage in university students 829

Page 4

20·5 (95% CI 20·9, 20·1). However, the largest mean
difference for BF% was between Siri and impedance
(24·5). The Bland–Altman plots for BF% illustrate the
mean differences and the fairly large limits of agreement
by different methods (Fig. 1).
Discussion
Despite a relatively low mean BMI (22·7 (SD 3·1) kg/m2),
body fat levels determined by the Siri equation were classi-
fied in 6·1% of the subjects as obese and in 9·7% as over-
weight. The percentages classified as overweight and obese
were lower than those reported in previous studies (Lowry
et al. 2000; Huang et al. 2003).
The results showed that the BMI is a poor predictor of
body fatness, since the sensitivity was low in comparison
with the reference method (Siri equation). It was shown
in earlier studies (Garn et al. 1986; Smalley et al. 1990;
Hannan et al. 1995; Deurenberg et al. 2001) that BMI
has considerable limitations in predicting an individual’s
BF%. This is why the body composition measurement is
necessary for the individual evaluation of fatness focused
on preventive medicine.
In females, 5·6% would be falsely classified as non-
obese according to their BMI and in males 1·1% would
be falsely classified as obese using this weight–height
index. Underprediction of obesity might be considered as
a greater error than an equal-magnitude overprediction
would be. Classifying an individual as lean, when in fact
the individual is truly obese regarding his or her body fat-
ness, may put this individual at risk from diseases associ-
ated with obesity and, potentially, delay any possible
beneficial therapy.
The different forms of obesity (android and gynoid) and
the different health risks associated with them are other
considerations to take into account. Recent studies indicate
that abdominal obesity is more strongly associated with
obesity-related health problems than is adiposity measured
by BMI (Booth et al. 2000). In the present study, 9·8 and
31·2% of the students were classified at risk according to
the waist circumference and WHR, respectively. We
have observed similar results with the Siri equation,
waist circumference and WHR. Concerning these results,
Fig. 1. Bland–Altman plots to compare body fat percentage (BF%) measured by different methods. (a) Comparison of Siri (1961) and Brozek
et al. (1963) (Siri–Brozek); (b) comparison of Siri (1961) and Deurenberg et al. (1991) (Siri–Deurenberg); (c) comparison of Siri (1961) and
Lean et al. (1996) (Siri–Lean); (d) comparison of Siri (1961) and bioelectrical impedance analysis (Siri–impedance).
M. Arroyo et al.830

Page 5

Taylor et al. (1998) observed that waist circumference was
better than the WHR when screening for regional fat distri-
bution using dual-energy X-ray absorptiometry as the
reference.
All methods were significantly different for BF%. When
considering the bias among methods for BF%, the Siri and
Lean equations were the most similar. Not surprisingly, the
equation that included the waist circumference (one of the
most labile sites of fat deposition) displayed the highest
agreement with Siri. However, this situation is somewhat
confusing since we registered the highest CI when we com-
pared Siri and Lean. Brozek had the narrowest limits of
agreement relative to the reference (27·2, 4·3%) of the
BF%. It should be noted that Siri and Brozek are the two
equations that include log S 4 skinfolds. Measurements
of skinfold thickness are an easy method of assessing
BF% and believed to be reasonably precise (Hannan et al.
1995; Sarria et al. 1998).
If an error of 4 percentage points BF% is considered as
reasonable (Lohman, 1992), in line with the standard error
of estimation of most prediction equations (Durnin &
Womersley, 1974; Deurenberg et al. 1991; Gallagher
et al. 1996), the present results of mean differences
between Siri and impedance are acceptable. However, the
CI between Siri and impedance was larger. This is to be
expected as the impedance formula uses additional infor-
mation, which, theoretically, enables us to distinguish
between fat and fat-free mass. Bioimpedance analysis
would overestimate BF%, so should be used with caution
in the diagnosis of obesity in this population.
Different findings have been observed by McNeill et al.
(1991) in adults and by Deurenberg et al. (1989) in chil-
dren. McNeill et al. (1991) observed the skinfold thickness
method to be as good as bioelectrical impedance in lean
and overweight groups of women. In children, Deurenberg
et al. (1989) published a study in which prediction for-
mulas for body composition from body impedance were
presented; in pre-pubescent boys and girls BF% could be
predicted with an error of about 4·2%, which is compar-
able with the prediction error for the assessment of BF%
from skinfold thickness (Deurenberg et al. 1990).
The prediction of BF% from BMI, age and sex (Deuren-
berg equation) assumes that, when BMI increases over a
certain threshold, the excess value is due to body fat in a
fixed part. This assumption certainly has its flaws and it
explains why the prediction formula generally underesti-
mates BF% at high values of body fat.
Nevertheless, the four methods were different from the
Siri equation and, in many cases, the limits of agreement
may be considered as high. Similar findings have been
observed in children (Parker et al. 2003) and in older
adults (Aghdassi et al. 2001; Barbosa et al. 2001). We
found that measurements of body fat may depend on
many factors, and the different methods studied are gener-
ally not directly interchangeable.
Although the methods and/or equations used in the pre-
sent study have commonly been used to estimate BF% in
young adults, they must not be used as a standard
method. Each method has limitations and the comparison
can be useful for an interpretation of results. More com-
parative studies should be conducted to get a better insight
into the generalisation of prediction methods and formulas.
Individual results and classifications have to be interpreted
with caution in the diagnosis and monitoring of overweight
and obesity.
Acknowledgements
The present study was supported by the University of the
Basque Country(UPV 00154.310-E-13972/2001
UPV 00101.125-15283/2003). We would like to thank
the students who participated in the present study.
and
References
Aghdassi E, Tam C, Liu B, McArthur M, McGeer A, Simor A &
Allard JP (2001) Body fat of older adult subjects calculated
from bioelectrical impedance versus anthropometry correlated
but did not agree. J Am Diet Assoc 101, 1209–1212.
Barbosa AR, Santarem JM, Jacob Filho W, Meirelles ES &
Marucci JM (2001) Comparison of body fat using anthropome-
try, bioelectrical impedance and DEXA in elderly women. Arch
Latinoam Nutr 51, 49–56.
Bland JM & Altman DG (1986) Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet i, 307–310.
Booth ML, Hunter C, Gore CJ, Bauman A & Owen N (2000) The
relationship between body mass index and waist circumference:
implications for estimates of the population prevalence of over-
weight. Int J Obes Relat Metab Disord 24, 1058–1061.
Bray G, Bouchard C & James P (1998) Definitions and proposed
current classifications of obesity. In Handbook of Obesity,
pp. 31–40 [G Bray, C Bouchard and P James, editors].
New York: Marcel Dekker.
Brozek J, Grande F, Anderson JT & Keys A (1963) Densitometric
analysis of body composition: revision of some quantitative
assumptions. Ann N Y Acad Sci 110, 113–140.
Deurenberg P, Andreoli A, Borg P, Kukkonen-Harjula K,
de Lorenzo A, van Marken Lichtenbelt WD, Testolin G,
Vigano R & Vollaard N (2001) The validity of predicted
body fat percentage from body mass index and from impedance
in samples of five European populations. Eur J Clin Nutr 55,
973–979.
Deurenberg P, Pieters JJ & Hautvast JG (1990) The assessment of
the body fat percentage by skinfold thickness measurements in
childhood and young adolescence. Br J Nutr 63, 293–303.
Deurenberg P, van-der-Kooy K, Paling A & Withagen P (1989)
The assessment of the body composition in 8–11 year-old chil-
dren by bioelectrical impedance. Eur J Clin Nutr 43, 623–629.
Deurenberg P, Weststrate JA & Seidell JC (1991) Body mass
index as a measure of body fatness: age and sex specific predic-
tion formulas. Br J Nutr 65, 105–114.
Deurenberg P, Yap M & van Staveren WA (1998) Body mass
index and percent body fat: a meta analysis among different
ethnic groups. Int J Obes Relat Metab Disord 22, 1164–1171.
Deurenberg-Yap M, Schmidt G, van Staveren WA & Deurenberg
P (2000) The paradox of low body mass index and high body
fat percent among Chinese, Malays and Indians in Singapore.
Int J Obes Relat Metab Disord 24, 1011–1017.
Durnin JVGA & Womersley J (1974) Body fat assessed from
total body density and its estimation from skinfold thickness:
measurements on 481 men and women aged from 16 to 72
years. Br J Nutr 32, 77–97.
Ellis KJ (1996) Measuring body fatness in children and
young adults: comparison of bioelectrical impedance analysis,
Body fat percentage in university students 831

Page 6

total body electrical conductivity, and dual-energy X-ray
absorptiometry. Int J Obes Relat Metab Disord 20, 866–873.
Emslander HC, Sinaki M, Muhs JM, Chao EY, Wahner HW,
Bryant SC, Riggs BL & Eastell R (1998) Bone mass and
muscle strength in female college athletes (runners and swim-
mers). Mayo Clin Proc 73, 1151–1160.
Fors H, Gelander L, Bjarnason R, Albertsson-Wikland K &
Bosaeus I (2002) Body composition, as assessed by bioelectrical
impedance spectroscopy and dual-energy X-ray absorptiometry,
in a healthy pediatric population. Acta Paediatr 91, 755–760.
Gallagher D, Visser M, Sepulveda D, Pierson RN, Harris T &
Heymsfield SB (1996) How useful is body mass index for com-
parison of body fatness across age, sex and ethnic groups? Am J
Epidemiol 143, 228–239.
Garn SM, Leonard WR & Hawthorne VM (1986) Three limi-
tations of the body mass index. Am J Clin Nutr 44, 996–997.
Gruber AJ, Pope HG Jr, Lalonde JK & Hudson JI (2001) Why do
young women diet? The roles of body fat, body perception, and
body ideal. J Clin Psychiatry 62, 609–611.
Gutin B, Litaker M, Islam S, Manos T, Smith C & Treiber F
(1996) Body-composition measurement in 9–11y old children
by dual-energy X-rayabsorptiometry,
measurements, and bioimpedance analysis. Am J Clin Nutr
63, 287–292.
Hannan WJ, Wrate RM, Cowen SJ & Freeman CPL (1995) Body
massindexasanestimateofbodyfat.IntJEatDisord18,91–97.
Heymsfield SB, Allison DB, Wang ZM, Baumgartner RN &
Ross R (1998) Evaluation of total and regional body
composition. In Handbook of Obesity, pp. 41–78 [G Bray,
C Bouchard and P James, editors]. New York: Marcel Dekker.
Heymsfield SB, Lichtman S, Baumgartner RN, Wang J, Kamen Y,
Aliprantis A & Pierson RN Jr (1990) Body composition of
humans: comparisonoftwo
models that differ in expense, technical complexity, and
radiation exposure. Am J Clin Nutr 52, 52–58.
Huang TT, Harris KJ, Lee RE, Nazir N, Born W & Kaur H (2003)
Assessing overweight, obesity, diet, and physical activity in
college students. J Am Coll Health 52, 83–86.
Kitano T, Kitano N, Inomoto T & Futatsuka M (2001) Evaluation
of body composition using dual-energy X-ray absorptiometry,
skinfold thickness and bioelectrical impedance analysis in Japa-
nese female college students. J Nutr Sci Vitaminol (Tokyo) 47,
122–125.
Lean ME, Han TS & Deurenberg P (1996) Predicting body
composition by densitometry from simple anthropometric
measurements. Am J Clin Nutr 63, 4–14.
Lohman TG (1992) Advances in Body Composition Assessment.
Champaign, IL: Human Kinetics Publisher.
Lowry R, Galuska DA, Fulton JE, Wechsler H, Kann L & Collins
JL (2000) Physical activity, food choice, and weight manage-
ment goals and practices among US college students. Am J
Prev Med 18, 18–27.
Luke A, Durazo-Arvizu R, Rotimi C, Prewitt TE, Forrester T,
WilksR,OgunbiyiOJ,Schoeller
Cooper RS (1997) Relation between BMI and body fat in
black population samples from Nigeria, Jamaica and the
United States. Am J Epidemiol 145, 620–628.
skinfold-thickness
improvedfour-component
DA, McGeeD&
McNeill G, Fowler PA, Maughan RJ, McGaw BA, Fuller MF,
Gvozdanovic D & Gvozdanovic S (1991) Body fat in lean
and overweight women estimated by six methods. Br J Nutr
65, 95–103.
Martin A & De Dios J (1993) Bioestadı ´stica para las Ciencias de
la Salud (Biostatistic in Health Science). Madrid, Spain: Norma.
Mitsuzono R & Komiya S (1991) Comparison of methods for
estimating percent body fat in distance-runners. Ann Physiol
Anthropol 10, 219–227.
National Institutes of Health, National Heart, Lung and Blood
Institute (1999) Clinical Guidelines on the Identification,
Evaluation, and Treatment of Overweight and Obesity in
Adults. The Evidence Report. Bethesda, MD: National Insti-
tutes of Health.
Parker L, Reilly JJ, Slater C, Wells JC & Pitsiladis Y (2003) Val-
idity of six field and laboratory methods for measurement of
body composition in boys. Obes Res 11, 852–858.
Sarria A, Garcia-Llop LA, Moreno LA, Fleta J, Morellon MP &
Bueno M (1998) Skinfold thickness measurements are better
predictors of body fat percentage than body mass index in
male Spanish children and adolescents. Eur J Clin Nutr 52,
573–576.
SEEDO (1996) Consenso espan ˜ol 1995 para la evaluacio ´n de la
obesidad y para la realizacio ´n de estudios epidemiolo ´gicos
(1995 Spanish consensus for the evaluation of obesity and to
carry out epidemiologic studies). Med Clin (Barc) 107,
782–787.
Servicio Central de Publicaciones del Gobierno Vasco (1994)
Encuesta de Nutricio ´n de la Comunidad Auto ´noma del Paı ´s
Vasco (Nutrition Survey in Autonomous Region of Basque
Country). Vitoria-Gasteiz, Spain: Dpto. de Salud.
Siri WE (1961) Body composition from fluid spaces and density:
analysis of methods. In Techniques for Measuring Body Com-
position, pp. 223–244 [J Brozeck and A Henschel, editors].
Washington, DC: National Academy of Sciences.
Smalley KJ, Knerr AN, Kendrick ZV, Colliver JA & Owen OE
(1990) Reassessment of body mass indices. Am J Clin Nutr
52, 405–408.
Taylor RW, Keil D, Gold EJ, Williams SM & Goulding A (1998)
Body mass index, waist girth, and waist-to hip ratio as indexes
of total and regional adiposity in women: evaluation using
receiver operating characteristic curves. Am J Clin Nutr 67,
44–49.
Treuth MS, Butte NJ, Wong WW & Ellis KJ (2001) Body com-
position in prepubertal girls: comparison of six methods. Int J
Obes Relat Metab Disord 25, 1352–1359.
Wellens R, Chumlea WC, Guo S, Roche AF, Reo NV &
Siervogel RM (1994) Body composition in white adults by
dual-energy X-ray absorptiometry, densitometry, and total
body water. Am J Clin Nutr 59, 547–555.
World Health Organization (1995) Physical Status: the Use and
Interpretation of Anthropometry. Technical Report Series no.
854. Geneva: WHO.
World Health Organization (1998) Programme of Nutrition,
Family and Reproductive Health. Obesity. Preventing and
Managing the Global Epidemic. Report of a WHO Consultation
on Obesity. Geneva, 3–5 June 1997, Geneva: WHO.
M. Arroyo et al.832