Validating the waist-height ratio and developing centiles for use amongst children and adolescents.

Abstract

To assess the statistical validity of the waist-height ratio (WHtR) as an appropriate method of adjusting waist circumference (WC) for height in children and adolescents.
Recently, WHtR has been proposed to be of greater value than body mass index (BMI), in predicting obesity-related cardiovascular co-morbidities in children. This index, however, is yet to be extensively validated within the paediatric population.
Height and WC in centimetres, were measured in 3597 children from grades 1 (5-7 years), 5 (9-11 years) and 10 (15-17 years). Log regression analyses using WC and height were performed to determine appropriate powers (p) to raise height, to completely adjust the index for height, by sex and grade. Correlations between WHtR and height were assessed.
Statistically, the WHtR is only valid for use among grade 1 boys and girls (p = 1.09 [95%CI 0.95-1.23] and p = 1.07 [95%CI 0.92-1.22], respectively) and grade 10 girls p = 0.85 (95%CI 0.62-1.08). However, the error (0.25%-1.85%), associated with the use of this index, in all ages and both sexes is clinically and biologically acceptable.
The WHtR is a clinically and biologically valid index to use among Australian children and adolescents.

Full text

Page 1

Acta Pædiatrica ISSN 0803–5253
REGULAR ARTICLE
Validating the waist-height ratio and developing centiles for use amongst
children and adolescents
Smita Nambiar (s.nambiar@uq.edu.au)1, Helen Truby1, Rebecca A Abbott2, Peter S.W. Davies1
1.The University of Queensland, Children’s Nutrition Research Centre, School of Medicine, Royal Children’s Hospital, Herston QLD, Australia
2.The University of Queensland, School of Human Movement Studies, St Lucia, QLD, Australia
Keywords
Body mass index, Cardiovascular co-morbidities,
Childhood obesity, Validation, Waist-height ratio
Correspondence
Smita Nambiar, The University of Queensland,
Children’s Nutrition Research Centre,
School of Medicine, Royal Children’s Hospital,
Herston QLD 4029, Australia.
Tel: + 61 7 3365 5567 |
Fax +61 7 3346 4684 |
Email: s.nambiar@uq.edu.au
Received
9 May 2008; revised 30 July 2008;
accepted 1 September 2008.
DOI:10.1111/j.1651-2227.2008.01050.x
Abstract
Aim: To assess the statistical validity of the waist-height ratio (WHtR) as an appropriate method of
adjusting waist circumference (WC) for height in children and adolescents.
Introduction: Recently, WHtR has been proposed to be of greater value than body mass index (BMI),
in predicting obesity-related cardiovascular co-morbidities in children. This index, however, is yet to be
extensively validated within the paediatric population.
Methods: Height and WC in centimetres, were measured in 3597 children from grades 1 (5–7
years), 5 (9–11 years) and 10 (15–17 years). Log regression analyses using WC and height were
performed to determine appropriate powers (p) to raise height, to completely adjust the index for
height, by sex and grade. Correlations between WHtR and height were assessed.
Results: Statistically, the WHtR is only valid for use among grade 1 boys and girls (p = 1.09 [95%CI
0.95–1.23] and p = 1.07 [95%CI 0.92–1.22], respectively) and grade 10 girls p = 0.85 (95%CI
0.62–1.08). However, the error (0.25%–1.85%), associated with the use of this index, in all ages
and both sexes is clinically and biologically acceptable.
Conclusion: The WHtR is a clinically and biologically valid index to use among Australian children and
adolescents.
INTRODUCTION
One of the most commonly used anthropometric indices for
assessingoverweightandobesityinchildrenisthebodymass
index (BMI). This index is essentially an attempt to adjust
body weight for height, using the equation weight/heightp,
where (p) is a power chosen to remove the influence of
height on weight. In the case of BMI, a (p) value of 2 is used
to adjust for this influence (1). The BMI has been used to
determine whether and individual is overweight or obese,
as well as to predict body composition and health risk (2).
It has, however, been acknowledged, that in all of these
cases, the BMI is less than ideal and has some significant
drawbacks (3). Recently, a new index, the waist–height ratio
(WHtR) has been proposed to be of greater value as a simple
anthropometric index, for predicting abdominal obesity and
related cardiovascular co-morbidities in adults and children
(4).
This index may be superior to the BMI, because of the
use of waist circumference (WC) in its calculation (5–7).
The WHtR is calculated by dividing the WC (commonly
Abbreviations
BMI, body mass index; cm, centimetres; CI, confidence interval;
DBP, diastolic blood pressure; HDL-C, high-density lipoprotein;
HKQ, Healthy Kids Queensland: Physical Activity and Nutri-
tion Survey 2006; LDL-C, low-density lipoprotein; T-Chol, total
cholesterol; TG, triglycerides; SBP, systolic blood pressure; WC,
waist circumference; WHtR, waist–height ratio
at the level of the umbilicus) by height, both measured
in centimetres (8–10). In this case, the generalised equa-
tion waist/heightpis effectively waist/height1. The WHtR
has been significantly associated with cardiovascular risk
factors, because of abdominal obesity both in adults and
children and the metabolic syndrome in adults (4,8–15). In
adults, a WHtR of 0.5 or less is considered to be normal
and values greater than 0.5 classifies an adult (who may or
may not be overweight as classified by weight or BMI) to
be at greater risk of the metabolic syndrome and cardio-
vascular risk factors (9,10). The 0.5 cut-off is used for both
adult males and females and it has been suggested that it
is suitable to apply this cut-off in children and adolescents
(4,8). The WHtR cut-off was established based on studies
that assessed strengths of associations between body com-
position, body fat distribution and measures to assess car-
diovascular health, such as blood pressure and blood lipid
profiles (4,8–10,12–14). Statistical analyses using correla-
tions and receiver operating characteristics (ROC), found
that cut-offs close to 0.5 to be most sensitive in correctly
identifying persons who with elevated cardiovascular health
risks (4,8–10,12–14).
The WHtR has also been shown to be well associated with
cardiovascular risk factors in children (4,8,11,16–20). In a
study by Hara et al. in 2002, it was found that among 9–10
and 12–13-year-old boys and girls (n = 880), WHtR was cor-
relatedpositivelywiththelifestyle relateddisease prevention
score, low-density lipoprotein (LDL-C), triglycerides (TG),
systolic blood pressure (SBP) and diastolic blood pressure
148
C ?2008 The Author(s)/ Journal Compilation C ?2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 148–152

Page 2

Nambiar et al.
Validating the WHtR and centiles in children
(DBP), whilst high-density lipoprotein (HDL-C) was nega-
tively correlated with WHtR. These correlations were statis-
tically significant at p = <0.001. The strongest correlation of
r = 0.6 was observed between the WHtR and the lifestyle re-
lated disease prevention score, suggesting that this index was
the best at predicting cardiovascular risk in children com-
pared to BMI (11). Further studies have also shown that
WHtR is superior to BMI in its ability to predict childhood
morbidity (16,20), and that the WHtR is simpler to under-
stand, as it is potentially one cut-off value dependant of age,
sex and ethnicity (4,8,11,16–20).
While the WHtR is an index that is well correlated with
abdominal obesity and cardiovascular risk factors in chil-
dren, it is yet to be determined among children and adoles-
cents as to whether this index is the best method of adjusting
WC for height. There is no prior reason why in the gener-
alized equation waist/heightp, (p) should be equal to 1, that
is, simply, WC divided by height. Generally, dividing A by
B does not completely adjust for the denominator (21). For
example, in regards to BMI, height (in metres) is raised to
a (p) of 2, that in order to adjust weight for the influence
of height. Among children, it has been found via log–log re-
gression, that in order to remove the influence of height, (p)
should be raised to the power of 2 in pre-school children,
gradually increasing to 3 at age 11, then falling back to 2
after puberty (1). Therefore, raising (p) to the value of 2 is
a numerically convenient and to a large extent, statistically
appropriate compromise.
The aim of this study, therefore, is to test the statistical
validity of the WHtR, in the form waist/height1. That is, to
ascertain whether dividing the WC by height, completely
adjusts the index for height or, if height should be raised
to another suitable p-value, that may vary with and be de-
pendant on the child’s age and sex. As far as we are aware,
such an analysis has not been undertaken on data arising
from children and adolescents. Such analysis is vital, as the
index is gaining popularity in the literature in paediatric-
based studies. We have therefore conducted this analysis in
a large sample of children and adolescents over a wide age
range.
SUBJECTS AND METHODS
Nutrition, physical activity and anthropometric data were
collected from 3691 Australian children in Queensland, as
part of the Healthy Kids Queensland: Physical Activity and
Nutrition Survey 2006 (HKQ) (22). In this randomized clus-
ter design study, children were approached in a school set-
ting, from both metropolitan and non-metropolitan regions.
The students were from grades 1 (ages 5–7 years), 5 (9–
11 years) and 10 (15–17 years) (22). Children from grades
1 and 5 were selected, as these were critical times in the
development of overweight and obesity, whilst avoiding ef-
fects of early menarche amongst the grade 5 females. Grade
10 students were selected because it was the last year of
compulsory schooling, and therefore an opportunity to ob-
tain a substantial number of children in their adolescence.
While the sample of children in this study were predomi-
nantly Caucasian, approximately 19% of children identified
themselves as either of Aboriginal and Torres Strait Islander
origin; born in a country other than Australia; or spoke a
language other than English at home (22).
Anthropometric data were successfully obtained from
3597 consenting students. Data from children with a noted
growth disorder likely to affect stature were excluded. Mea-
surements included weight (kg), height (cm) and WC (in
centimetres, measured at the level of the umbilicus). Weight
was recorded using digital scales to the nearest 0.1kg. Stu-
dents were in school uniforms, but removed shoes and
any excess clothing. Height was measured with a portable
stadiometer. Two measurements were taken. A third was
recordedifthefirsttwodifferedbymorethan0.5cm.Theav-
erage of the closest two measures was used in analyses. Stu-
dents were required to remove shoes and hair ties. WC was
measured using a retractable steel measuring tape, against
the skin. Two measurements were taken. A third measure
was recorded if the first two deferred by more than 0.5 cm.
The average of the closest two WC measures was used in
analyses.
WHtRwascalculatedbydividingWC(cm)byheight(cm).
Natural logarithms of WC (cm) and height (cm) were used
in linear regression analyses to determine appropriate pow-
ers (p) to raise height, to completely adjust the index for
height, by sex and age (21,23). Pearson’s correlations be-
tween WC and height (Ht) were performed to determine
the relationship between WC and height. Pearson’s correla-
tions between WHtR and Ht were calculated. If the index
(waist/height1) is entirely appropriate, then the correlation
between these variables should be zero or close to zero.
Independent samples t-tests and one-sample Kolmogorov–
Smirnov tests were performed to observe if males and
females differed significantly for various anthropometric
variables and to assess for sample skewness within each
variable. Using the LMS (where L = Box-Cox power, M =
generalised mean of data and S = generalised coefficient
of variation) method (24), centiles for WHtR were calcu-
lated for each sex within each grade. There was no attempt
to smooth centiles across age groups because of significant
missing data for ages 7.00–9.00 years and 12.00–14.00 years.
Statistical analysis was performed using SPSS Version
14.0 (SPSS Inc., Chicago, IL, USA).
The Healthy Kids Queensland 2006 survey was approved
byTheUniversityofQueenslandandEducationQueensland
Ethics Committees.
RESULTS
Table 1 shows some basic physical characteristics of the
sample. In grade 1, the mean age, weight and height, were
significantly greater in males compared to females. WC was
also higher among the males, while WHtR was significantly
higher in females. In grade 5, on average, males were sig-
nificantly older; however, females were significantly heav-
ier with larger WC than males. Females were also taller,
although this was not statistically significant. In grade 10,
males were on average, older and were significantly heavier,
taller and had higher WC and than females. Across all age
groups, the mean WHtR was higher in females than males;
C ?2008 The Author(s)/ Journal Compilation C ?2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 148–152
149

Page 3

Validating the WHtR and centiles in children
Nambiar et al.
Table 1 Descriptives for anthropometric variables by grade and sex
Males Females
GradeVariablesnMeanSDMinMaxn MeanSDMin Max
1Age (years)∗
Weight (kg)∗
WC (cm)
Ht (cm)∗
WHtR∗
Age (years)∗
Weight (kg)∗
WC (cm)∗
Ht (cm)
WHtR
Age (years)
Weight (kg)∗
WC (cm)∗
Ht (cm)∗
WHtR∗
5576.18
23.1
56.2
118.8
0.47
10.20
36.4
65.7
140.9
0.46
15.20
64.4
77.2
172.7
0.45
0.38
3.9
5.4
5.4
0.04
0.42
8.2
9.0
6.5
0.06
0.41
12.2
8.7
7.6
0.05
5.35
14.2
44.5
98.3
0.40
9.07
19.9
43.7
113.9
0.31
14.21
34.9
58.0
150.2
0.36
7.57
42.7
86.2
134.1
0.69
12.20
69.9
102.5
160.4
0.86
16.63
118.9
123.4
198.6
0.70
542 6.07
22.1
56.0
117.0
0.48
10.12
37.4
66.8
141.4
0.47
15.16
57.4
75.3
163.4
0.46
0.36
3.8
5.4
5.3
0.04
0.39
9.1
9.8
6.9
0.06
0.40
9.9
8.1
5.9
0.05
4.98
14.7
46.0
102.0
0.40
9.37
21.9
47.0
104.1
0.34
14.18
31.1
56.2
143.2
0.36
7.42
47.9
83.3
137.2
0.67
11.9
88.6
108.5
167.9
0.79
16.87
110.0
112.3
181.3
0.66
5
706 782
10 472538
∗Independent samples t-test; significant differences between males and females within each grade p < 0.05.
Table 2 Calculated WHtR centiles, using the LMS method, by sex and grade
GenderGradeCentiles
5th
10th
25th
50th
75th
90th
95th
Males1
5
10
1
5
10
0.423
0.391
0.385
0.426
0.389
0.395
0.432
0.403
0.395
0.435
0.402
0.406
0.449
0.427
0.413
0.451
0.427
0.427
0.471
0.458
0.439
0.473
0.461
0.445
0.499
0.497
0.473
0.500
0.505
0.488
0.531
0.540
0.514
0.532
0.557
0.525
0.554
0.571
0.546
0.556
0.597
0.551
Females
Table 3 Pearson’s correlation between WC and height by grade and sex
GradeMalesFemales
1
5
10
0.54
0.47
0.29
0.51
0.44
0.29
All correlations significant at p < 0.01.
this difference was statistically significant in the grade 1 and
10 cohort.
Analysis also revealed that age was normally distributed
across all ages and both sexes. With the exception of grade
10 males, this was also true for height. WC and WHtR were
also skewed across all ages and both sexes. Again, with the
exception of grade 10 males, weight was also skewed across
the other groups.
Calculated centiles for WHtR for each sex within the
grades are shown in Table 2.
The results in Table 3 clearly show the correlation be-
tween WC and height was statistically significant in all cases.
Indeed, this correlation is the premise for a need to adjust
WC for height. As an illustration, the relationship between
WC and height for grade 1 boys (r = 0.54, p < 0.01) is shown
in Figure 1.
Figure 1 Pearson’s correlation between WC and height in grade 1 males Pear-
son’s correlation (r = 0.537, p < 0.01).
Table 4 presents the power (p) to which height must be
raised, to completely adjust WC for height by grade and
sex. The 95% confidence intervals for (p) are also shown.
Pearson’s correlations between WHtR and height are illus-
trated in Table 5. Again as an example, the relationship be-
tween waist/height1and height for grade 1 males is shown
in Figure 2. The correlation is very close to zero (r = 0.07),
as would be expected if the adjustment of waist/height1is
appropriate.
DISCUSSION
The WHtR is a newly developed index, proposed to be su-
perior to the BMI, because of its relationship to cardiovas-
cular risks, as a consequence of abdominal obesity. It has
been shown to be positively correlated with cardiovascular
150
C ?2008 The Author(s)/ Journal Compilation C ?2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 148–152

Page 4

Nambiar et al.
Validating the WHtR and centiles in children
Table 4 Actual power (p) height must be raised to completely adjust WC for
height, by grade and sex
Grade SexPower (p)95% Confidence interval (CI)
1M
F
M
F
M
F
1.09
1.07
1.33
1.24
0.75
0.85
(0.95–1.23)∗
(0.92–1.22)∗
(1.15–1.51)
(1.06–1.42)
(0.54–0.96)
(0.62–1.08)∗
5
10
∗Value of 1 falls within the boundaries of the confidence interval.
Table 5 Pearson’s correlations for WHtR and height; percentage error (r2×
100) represents the variation accounted for in WHtR, by the differences in
height, after adjustment for WC for height, using the generalized equation
waist/height
GradeSexrr2
% error
1M
F
M
F
M
F
0.07
0.05
0.14
0.10
−0.11
−0.05
0.00
0.00
0.02
0.01
0.01
0.00
0.49
0.29
1.85
1.00
1.12
0.25
5
10
Figure 2 Pearson’s correlation between WHtR and Ht in grade 1 males Pear-
son’s correlation (r = 0.07) (ns).
risks factors such as SBP and DBP, total cholesterol (T-
Chol), TG, LDL-C and negatively correlated with HDL-C
(9–14,16,17,19,25,26). It is also positively correlated with
body fat variables such as percent trunk fat and percent body
fat, measured by dual energy x-ray absorptiometry (DEXA),
in both adults and children (25,26). This index, however, it
is yet to be validated in childhood and adolescents, in terms
of whether it is an appropriate method, of best adjusting WC
for height. This paper aimed to test this in a cohort of 3597
Australian children aged between 5–17 years of age.
The initial stages of analysis found all differences between
males and females for anthropometric variables were as ex-
pected. Generally in grade 1, the males were significantly
older and had larger measurements than the females. By
grade 5, it was the females who were heavier, taller and
with wider WC, an indication that some may have entered
puberty. This commonly occurs, on average approximately
two years earlier in females than males (27). By grade 10
the pattern had reversed again, this time showing that most
males will have now passed peak growth velocity, whilst fe-
male subjects by this stage, had reached a plateau in growth,
particularly in terms of height (27). The centiles calculated
for WHtR, shown in Table 2, might be useful in determining
extremes in the age groups studied here, in other studies.
Presently, WC is divided by height to adjust for the influ-
ence of height. Therefore, height is being raised to the power
of one (waist/height1). Figure 1, as an example, clearly in-
dicates a moderately strong association was found between
WC and height, showing that generally, the taller individ-
ual has the larger WC. Therefore, the influence of height
must be removed so that an unbiased estimate of WC can be
obtained. As the results indicate, the current method of de-
riving this index is only statistically valid for use among boys
and girls aged 5–7years, where p = 1.09 (95% CI 0.95–1.23)
and p = 1.07 (95% CI 0.90–1.22), respectively, and girls aged
15–17years p = 0.85 (95% CI 0.62–1.08). In these cases, the
confidence intervals found for these sexes and grades, in-
clude one, which is waist/height1or waist/height. For the
males and females in grade 5 and the males in grade 10,
however, the confidence intervals did not include 1, mean-
ing that statistically, the WHtR is not an entirely appropri-
ate method for adjusting WC for height in all ages, in this
study.
While this may be the case statistically, results of Pear-
son’s correlations between WHtR and height indicate that
the error induced in using a power of 1 may be biologically
acceptable. The correlations between these variables are ex-
tremely weak for all ages and both sexes, especially among
the grade 1 females, where the correlation is closest to zero.
The r2values indicate that the error associated with the use
of this index is small. For example, for the groups for which
the index was statistically inappropriate, the error associated
with the correlations range from 0.25% to 1.85%. Thus, with
less than 2% of the variation in WHtR explained by height,
a p-value of 1 is a numerically convenient number to raise
height, to adjust for its influence. In using this approach,
an analogy can be drawn with studies that investigated the
most appropriate power to which height should be raised to
remove its influence within the BMI. As Cole states (1), if
in this expression weight/height2, were an adequate index, it
would be uncorrelated with height at all ages and both sexes.
Rollard–Cacheraetal.(1982)showedthatthisistrue,except
during puberty in both sexes where the correlation between
BMI and height is statistically significant (28). This variation
from the weight/height2model did not however sufficiently
detract from the overall view that the BMI is an acceptable
index across childhood (28). The same logic applies here, in
the use of waist/height1. The WHtR is therefore clinically
and biologically validated as an acceptable index to be used
among children and adolescents.
C ?2008 The Author(s)/ Journal Compilation C ?2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 148–152
151

Page 5

Validating the WHtR and centiles in children
Nambiar et al.
CONCLUSION
The WHtR can be used among Australian children and
adolescents. It may be more appropriate to use than BMI
because of its ability to explain body fat distribution and
associated cardiovascular health risks. It is simpler to cal-
culate than the BMI, and it may prove to be useful in child
population health studies, as it is more feasible to take an-
thropometric measurements than collect blood samples as
a method for screening those with increased risk of obesity
related cardiovascular co-morbidities. The WHtR is a clin-
ically and biologically accepted index that may be used to
screen for abdominally obese children and adolescents, at
risk of cardiovascular health problems.
ACKNOWLEDGEMENTS
Queensland Health commissioned and funded the Healthy
Kids Queensland Survey. A Steering Committee includ-
ing representation from Queensland Health, Education
Queensland,IndependentSchoolsQueensland,theQueens-
land Catholic Education Commission and the Queensland
Department of Local Government Sport and Recreation
provided advice, guidance and support regarding the sur-
vey.
References
1. Cole T. Weight/heightpcompared to weight/height2for
assessing adiposity in childhood: influence of age and bone
age on p during puberty. Ann Hum Biol 1986; 13: 433–51.
2. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a
standard definition for child overweight and obesity
worldwide: international survey. BMJ 2000; 320: 1240–3.
3. Prentice A, Jebb S. Beyond body mass index. Obes Rev 2001;
2: 141–7.
4. Ashwell M, Hsieh SD. Six reasons why the waist-to-height
ratio is a rapid and effective global indicator for the health
risks of obesity and how its use could simplify the
international public health message on obesity. Int J of Food
Sci and Nutr 2005; 56: 303–7.
5. Despres J, Morrjani S, Lupien P, Tremblay A, Nadeau A,
Bouchard C. Regional distribution of body fat, plasma
lipoproteins and cardiovascular disease Arterioscler. Throm
and Vasc Biol. 1990; 10: 497–511.
6. Janssen I, Katzmarzyk PT. Waist circumference and not body
mass index explains obesity-related health risk. Am J of Clin
Nutr 2004; 79: 379–84.
7. Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM,
Bouchard C, Berenson GS. Body mass index, waist
circumference and clustering of cardiovascular risk factors in
bi-racial sample of children and adolescents. Pediatrics 2004;
114: 198–205.
8. Ashwell M. Waist to height ratio and the Ashwell shape chart
could predict risks of obesity in adults and children in all
ethnic groups. Nutr and Food Sci 2005; 36: 359–64.
9. Hsieh S, Yoshinaga H. Abdominal fat distribution and
coronary heart disease risk factors in men- waist/height ratio
as a simple and useful predictor. Int J of Obes Relat Metab
Disord 1995; 19: 585–9.
10. Hsieh SD, Yoshinaga H. Waist/height ratio as a simple and
useful predictor of coronary heart disease risk factors in
women. Int Med 1995; 34: 1147–52.
11. Hara M, Saitou E, Iwata F, Okada T, Harada K.
Waist-to-height ratio is the best predictor of cardiovascular
disease risk factors in Japanese school children. J Athero
Throm 2002; 9: 127–32.
12. Hsieh S, Yoshinaga H, Muto T. Waist to height ratio, a simple
and simple and practical index for assessing central fat
distribution and metabolic risk in Japanese men and women.
Int J of Obes 2003; 27: 610–6.
13. Hsieh SD, Moto T. Metabolic syndrome in Japanese men and
women with special reference to the anthropometric criteria
for the assessment of obesity: proposal to use the waist to
height ratio. Prev Med 2006; 42: 135–9.
14. Hsieh SD, Muto T. The superiority of waist-to-height ratio as
an anthropometric index to evaluate clustering of coronary
risk factors among non-obese men and women. Prev Med
2005; 40: 216–20.
15. Hsieh SD, Yoshinaga H, Sakurai TMY. Regular physical
activity and coronary risk factors in Japanese men. Circulation
1998; 97: 661–5.
16. Freedman DS, Kahn HS, Mei Z, -Strawn LMG, Dietz WH,
Srinivasan SR. Relation of body mass index and
waist-to-height ratio to cardiovascular disease risk factors in
children and adolescents: the Bogalusa Heart Study. Am J of
Clin Nutr 2007; 86: 33–40.
17. Ho SY, Lam TH, Janus ED. Waist to stature ratio is more
strongly associated with cardiovascular risk factors than other
simple anthropometric indices. Ann Epidemiol 2003; 13:
683–91.
18. Kahn HS, Imperatore G, Cheng YJ. A population- based
comparison of BMI percentiles and waist to height ratio for
identifying cardiovascular risk in youth. Pediatrics 2005; 146:
482–8.
19. Savva S, Tornaritis M, Savva M, Kourides Y, Panagi A,
Silikiotou N, et al. Waist circumference and waist-to-height
ratio are better predictors of cardiovascular disease risk
factors in children than body mass index. Int J of Obes 2000;
24: 1453–8.
20. Yan W, Bingxian H, Hua Y, Jianghong D, Jun C, Dongliang G,
et al. Waist-to-height ratio is an accurate and easier index for
evaluating obesity in children and adolescents. Obesity 2007;
15: 748–52.
21. Davies P, Cole T. The adjustment of measures of energy
expenditure for body weight and body composition. Int J of
Body Comp Res 2003; 1: 45–50.
22. Abbott R, MacDonald D, Makinnon L, Stubbs C, Lee A,
Harper C, et al. Healthy Kids Queensland Survey 2006-
Summary Report. Brisbane: Queensland Health. 2007.
23. Cole T. Sympercents: symmetric percentage differences on the
100 logE scale simply the presentation of log transformed
data. Stat Med 2000; 19: 3109–25.
24. Cole TJ. The LMS method for constructing normalized growth
standards. Eur J Clin Nutr 1990; 44: 45–60.
25. Daniels SR, Khoury P, Morrison JA. Utility of different
measures of body fat distribution in children and adolescents.
Am J of Epidemiol 2000; 152: 1179–83.
26. Kagawa M, Hills AP, Binns CW. Usefulness of the
waist-to-height ratio to predict trunk fat accumulation in
young Japanese and Australian Caucasian males living in
Australia. Int J of Body Comp Res 2007; 5: 57–63.
27. Tanner J, Davies PS. Clinical Longitudinal standards for
height and height velocity for North American Children. J of
Paeds 1985; 107: 317–29.
28. Rolland-Cachera M, Sempe M, Guilloud-Bataille M, Patois E,
Pequignant-Guggenbuhl F, Fautrad V. Adiposity indices in
children. Am J of Clin Nutr 1982; 36: 178–84.
152
C ?2008 The Author(s)/ Journal Compilation C ?2008 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 148–152