European Journal of Clinical Nutrition (2009) 63, 1305–1312

& 2009 Macmillan Publishers Limited All rights reserved 0954-3007/09 $32.00
www.nature.com/ejcn

ORIGINAL ARTICLE
Age, sex and ethnic differences in the prevalence
of underweight and overweight, defined by using
the CDC and IOTF cut points in Asian children
NT Tuan and TA Nicklas

USDA/ARS, Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA

Background: No nationally representative data from middle- and low-income countries have been analyzed to compare the
prevalence of underweight and overweight, defined by using the Centers for Disease Control and Prevention (CDC) and the
International Obesity TaskForce (IOTF) body mass index cut points.
Objective: To examine the consistency in the prevalence of underweight and overweight, defined by using the CDC and IOTF
cut points in Chinese, Indonesian and Vietnamese children.
Methods: We used data from 1600 Chinese, 11 756 Indonesian and 53 826 Vietnamese children aged 2–18 years, who
participated in three recent, representative surveys in China, Indonesia and Vietnam. A smaller difference between prevalence
and a higher k-statistic indicated a higher consistency level.
Results: The prevalence of underweight was higher with the IOTF than the CDC cut points; absolute differences in the Chinese,
Indonesian and Vietnamese were 6, 10 and 13% (boys), and 10, 13 and 19% (girls), respectively. The prevalence of overweight
was more consistent (absolute differences were o2%, except for the 2–5.9-year-old Chinese and Indonesian children (from 2 to
o5%)). Values of k-statistic (from 0.55 to 0.88) varied by age, sex and ethnicity. The consistency was gradually improving from
the Vietnamese to Indonesians and to Chinese boys and girls, from girls to boys, from the younger to older boys and from the
older to younger girls.
Conclusions: The age, sex and ethnic differences in the prevalence of underweight and overweight suggest a systematic
evaluation of the cut points.
European Journal of Clinical Nutrition (2009) 63, 1305–1312; doi:10.1038/ejcn.2009.90; published online 19 August 2009

Keywords: Asian children; CDC growth charts; IOTF cut points; overweight; underweight

Introduction acceptance as a good screening tool for childhood obesity

Underweight, overweight and related diseases, which lead to
elevated health and economic burdens, have become emer-
ging problems worldwide (Murray and Lopez, 1997; WHO,
2005; Popkin, 2006). These trends have led to a demand of
having accurate definitions for underweight and overweight
in children, which facilitate timely detections, treatments
and preventions, and help to avoid unnecessary treatments
of false-positive cases (Guillaume, 1999; Neovius et al., 2004).
Sex- and age-specific body mass index (BMI) is gaining
Correspondence: Dr NT Tuan, USDA/ARS, Children’s Nutrition Research
Center, Department of Pediatrics, Baylor College of Medicine, 1100 Bates
Avenue, Houston, TX 77030, USA.
E-mail: ttn1@bcm.edu
Received 13 January 2009; revised 5 May 2009; accepted 7 July 2009;
published online 19 August 2009
(Guillaume, 1999; Must and Anderson, 2006; Barlow, 2007).
Several BMI classification systems have been widely used
(Must et al., 1991; WHO expert committee, 1995; Cole et al.,
2000, 2007; Kuczmarski et al., 2002).
Controversy remains regarding the selection of inter-
national cut points for underweight and overweight in
children. Being a more comprehensive reference population
and cut points, the Centers for Disease Control and
Prevention (CDC) (2000) (Kuczmarski et al., 2002) sex-
specific BMI-for-age growth charts (CDC cut points) have
been recommended to replace the earlier definition that was
developed by Must et al. and adopted by the World Health
Organization (Kuczmarski et al., 2002; Barlow, 2007). How-
ever, it is still uncertain whether the BMI growth charts
developed for American children could be used internation-
ally (Cole et al., 2000, 2007). The International Obesity
 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1306
TaskForce (IOTF) has developed and proposed international
cut points for overweight and underweight in children (Cole
et al., 2000, 2007). With the inclusion of children from six
countries in three continents, the IOTF reference population
might represent better growth patterns of international
children compared with that of the CDC growth charts.
However, it is still uncertain whether the IOTF cut points are
appropriate to be used in non-representative countries
(Hesketh and Ding, 2000; Neovius et al., 2004; Cole et al.,
2007). As the CDC and IOTF cut points are comprehensive,
available and easy to be used, health care providers and
researchers have a choice between the two classification
systems. However, differences in findings, which are attrib-
uted to the use of different cut points, make it difficult to
monitor global and national trends, to perform comparison
between studies and to employ in public health and clinical
settings (Neovius et al., 2004). Thus, research is needed to
compare and contrast the findings derived from the CDC
and IOTF cut points in different populations.
Studies from high-income countries showed that the use
of the CDC and IOTF cut points provide comparable
prevalence of overweight and obesity (O’Neill et al., 2006;
Vidal et al., 2006; Ko et al., 2008). However, no nationally
representative data from middle- and low-income countries
have been analyzed to compare the prevalence of under-
weight and overweight, defined by using the CDC and IOTF
cut points. We conducted the study to examine the
consistency in the prevalence of underweight and over-
weight, defined by using the CDC and IOTF cut points, in
2–18-year-old Chinese, Indonesian and Vietnamese children.
Subjects and methods
Study population
We used data from 1600 Chinese, 11 756 Indonesian and
53 826 Vietnamese children aged 2–18 years, who partici-
pated in three recent, representative surveys: the China
Health and Nutrition Survey conducted in 2006 (CHNS), the
Indonesia Family Life Survey conducted in 2000 (IFLS) and
the Vietnam National Health Survey conducted in 2002
(VNHS). The CHNS conducted in 2006 was the most recent
wave of an ongoing study established in the late 1980s in
nine Chinese provinces that vary substantially in geography,
economic development, public resources and health indica-
tors. The sample represented B50% of the Chinese popula-
tion (Popkin et al., 1995). The IFLS conducted in 2000 was a
part of an ongoing longitudinal survey established in the
early 1990s in 13 Indonesian provinces. The sample
represented B83% of the Indonesian population (Strauss
et al., 2004). The VNHS conducted in 2002 was the largest
nationally representative health survey ever conducted in all
61 Vietnam provinces (Ministry of Health—General Statis-
tical Office, 2003; Tuan et al., 2008). Survey instructions, data
sets and questionnaires may be downloaded from the web
sites of the CHNS (http://www.cpc.unc.edu/china), the IFLS
European Journal of Clinical Nutrition
(http://www.rand.org/labor/FLS/IFLS) and the Vietnam Min-
istry of Health (http://www.moh.gov.vn). On the basis of
the survey conducted by the World Bank in 2008, Vietnam
was listed as having a low-income economy, whereas China
and Indonesia were listed as having lower-middle-income
economies (World Bank, 2008).
We included 2–18-year-old boys and girls who were not
pregnant or lactating at the time of survey, and for whom
measurements of weight and height were complete and
plausible (for example, BMI z-score of 5 to 5 and BMI value
of 10 to 50 kg/m2). The number of individuals with
implausible values was 34, 154 and 280 in the CHNS, IFLS
and VNHS, respectively, which represented B0.5–2% of the
eligible participants with measured weight and height.
Removal of biologically implausible outliers did not affect
point estimates, but did increase the estimate precision. BMI
(kg/m2) was calculated on the basis of weight and height,
which were measured by trained health workers who
followed standardized procedures and used regularly cali-
brated equipment. The CHNS used SECA 880 scales and
SECA 206 wall-mounted metal tapes (SECA); the IFLS and
VNHS used SECA 890 scales and Shorr measuring boards
(Popkin et al., 1995; Ministry of Health—General Statistical
Office, 2003; Strauss et al., 2004).
Nutrition status assessment
The CDC 2000 sex-specific BMI-for-age growth charts and
the IOTF age- and sex-specific BMI cut points were used to
evaluate the nutrition status of the 2–18-year-old children.
The CDC growth charts were developed on the basis of data
from five nationally representative surveys in the US from
1963 to 1994 (B30 000 children). Smoothed age- and sex-
specific 5th and 85th percentiles for BMI were used as criteria
for underweight and overweight, respectively (Kuczmarski
et al., 2002). The IOTF cut points for overweight and
underweight were developed using the data from nationally
representative surveys that were implemented from 1963 to
1993 in six high- and upper-middle-income countries
(B200 000 children in Brazil, Great Britain, Hong Kong,
the Netherlands, Singapore and the United States). Regres-
sion procedures were used to create smoothed BMI curves
that generated adult cut points of 18.5 (underweight) and
25 kg/m2 (overweight) at the age of 18 years (Cole et al.,
2000, 2007).
Each child was defined to be underweight, normal weight
or overweight by using the CDC and IOTF cut points. When
the CDC growth charts were used, a child with a BMI o5th
percentile (BMI z-score o1.645) was classified to be
underweight, whereas a child with a BMIX85th percentile
(BMI z-scoreX1.036) was classified to be overweight or obese
(Kuczmarski et al., 2002). A child with a BMI less than age-
and sex-specific IOTF cut points for underweight (Cole et al.,
2007) was classified to be underweight, whereas a child with
a BMI greater than or equal to age- and sex-specific IOTF cut
points for overweight (Cole et al., 2000) was classified to be

overweight or obese. In this paper, except when specified,
we used the overweight term to infer both overweight
and obese status. The nutrition status classification based on
the IOTF cut points and the estimation of BMI z-scores
compared with the CDC reference population were deter-
mined by using the LMSgrowth 2.64, an add-in program
written for Microsoft Excel by Pan and Cole (Pan and Cole,
2008).
Statistical methods
To examine the consistency in nutrition status classifica-
tions, defined by using the CDC and IOTF cut points, we
estimated and compared the prevalence of underweight and
the prevalence of overweight classified by using two sets of
cut points (two-sided w2-tests; P-valueo0.05). A k-statistic,
which indicates the consistency between the two rating
Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1307
systems, was estimated by dividing the difference between
the observed and the expected probabilities of concordance
(for example, between two sets of cut points) with the
expected probability of non-concordance. A k-statistic
40.75 denotes excellent reproducibility, 0.4–0.75 denotes
good reproducibility and 0–3.9 denotes marginal reproduci-
bility (Rosner, 2006). Kappa statistics and 95% confidence
interval were estimated using the kapci program written for
Stata (Stata Inc., College Station, TX, USA) (Reichenheim,
2004). Two-sided independent t-tests (P-valueo0.05) were
used to compare two k-statistics.
Given a large sample size in each ethnicity, we performed
stratified analyses by age and sex groups. Survey and
weighted commands were used when appropriate to take
into account sampling weights and design effects. We
conducted all analyses using the Stata software version
10.1 (Stata Inc.).

25 CDC Growth Chart 25 CDC Growth Chart

Chinese
Chinese
Indonesian

percentiles
Indonesian
Vietnamese

85th
Vietnamese
Body Mass Index (kg/m2)
Body Mass Index (kg/m2)

20
20

percentiles
5th
t
15
15

10
10
2-3.9 4-5.9 6-7.9 8-9.9 10-11.9 12-13.9 14-15.9 16-18 2-3.9 4-5.9 6-7.9 8-9.9 10-11.9 12-13.9 14-15.9 16-18
Age (y) Age (y)

25 CDC Growth Chart 25 CDC Growth Chart

percentiles
Chinese Chinese

85th
Indonesian Indonesian
Vietnamese Vietnamese
Body Mass Index (kg/m2)

Body Mass Index (kg/m2)

20 20
percentiles
5th

15 15

10 10
2-3.9 4-5.9 6-7.9 8-9.9 10-11.9 12-13.9 14-15.9 16-18 2-3.9 4-5.9 6-7.9 8-9.9 10-11.9 12-13.9 14-15.9 16-18
Age (y) Age (y)

Figure 1 Median body mass index (kg/m2) by age in 2–18-year-old Figure 2 The 5th and 85th percentiles for body mass index (kg/m2)
boys (a) and girls (b). Sampling weight was taken into account in the by age in 2–18-year-old boys (a) and girls (b). Sampling weight was
estimation of the median. CDC, Centers for Disease Control and taken into account in the estimation of the 5th and 85th percentiles.
Prevention. CDC, Centers for Disease Control and Prevention.

European Journal of Clinical Nutrition

 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1308
The role of the funding sources and ethical consideration
The authors had full access to all of the data in the study, and
take responsibility for the integrity of the data and the
accuracy of the data analysis. The sponsors were not
involved in the study design, the collection, analysis or
interpretation of data, and in the writing or submission of
the manuscript for publication. The Institutional Review
Board of the Baylor College of Medicine reviewed and
approved the study.
Results
In all age groups, the median BMI in Chinese boys was B2
and 2.5-units higher compared with that in Indonesian and
Vietnamese boys, respectively. The median BMI in 2–14-year-
old Chinese boys was similar to that in the CDC BMI-for-age
growth chart for boys (Figure 1a). The median BMI in
Chinese girls was B1–2-units and 1–2.5-units higher com-
pared with that in Indonesian and Vietnamese girls,
respectively (Figure 1b). Seven-year-old Indonesian and
Vietnamese children, and 6-year-old Chinese children had
the lowest median BMI compared with other age groups.
The 5th percentiles for BMI were comparable among the
Chinese, Indonesian and Vietnamese boys (Figure 2a) and
girls (Figure 2b). Chinese boys and girls had the highest 85th
percentiles for BMI compared with those of Indonesian and
Vietnamese boys and girls. The 85th percentiles for BMI in
Chinese boys and girls were closest to those in respective
CDC growth charts.
The prevalence of underweight, defined by using the IOTF
cut points, was higher compared with that obtained from the
CDC BMI-for-age growth charts (absolute differences were 6,
10 and 13% in Chinese, Indonesian and Vietnamese boys,
and 10, 13 and 19% in Chinese, Indonesian and Vietnamese
girls, respectively) (Table 1). The consistency in the pre-
valence of underweight was gradually improving from the
Vietnamese to Indonesians and to Chinese boys and girls,
from girls to boys, from the younger to older boys and from
the older to younger girls (Figure 3).
The prevalence of overweight, defined by using the IOTF
cut points, was slightly different compared with that
obtained from the CDC BMI-for-age growth charts in all

Table 1 Prevalence (%) of underweight defined by the CDC 2000 BMI-for-age growth charts and the IOTF cut points in 2–18-year-old Chinese,
Indonesian and Vietnamese childrena

Chinese (n ¼ 1600) Indonesian (n ¼ 11 756) Vietnamese (n ¼ 53 826)

Prevalence CDC b Prevalence IOTF c Diff.d Prevalence CDC Prevalence IOTF Diff. Prevalence CDC Prevalence IOTF Diff.

Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI

Boys
All age 12.9 (10.2, 15.5) 18.3 (15.3, 21.4) 5.5 24.5 (23.3, 25.8) 34.4* (33.1, 35.7) 9.9 36.1 (35.3, 36.9) 49.0* (48.1, 49.8) 12.9
2–5.9 years 14.5 (9.2, 19.8) 21.0 (14.4, 27.5) 6.5 21.7 (19.4, 24.1) 33.2* (30.5, 35.8) 11.4 26.1 (24.6, 27.5) 40.0* (38.3, 41.6) 13.9
6–9.9 years 13.0 (8.4, 17.6) 18.9 (13.4, 24.4) 5.9 25.0 (22.5, 27.5) 35.6* (32.9, 38.3) 10.6 35.1 (33.7, 36.5) 49.2* (47.8, 50.6) 14.1
10–13.9 years 9.3 (5.3, 13.2) 15.2 (10.3, 20.1) 5.9 27.1 (24.6, 29.5) 35.8* (33.2, 38.5) 8.8 43.4 (42.1, 44.7) 56.9* (55.6, 58.2) 13.6
14–18 years 15.4 (10.5, 20.3) 19.0 (13.8, 24.1) 3.6 24.0 (21.7, 26.3) 32.8* (30.2, 35.4) 8.8 35.6 (34.3, 36.9) 45.9* (44.5, 47.3) 10.4

Girls
All age 13.7 (10.9, 16.5) 24.1* (20.3, 27.8) 10.3 18.6 (17.5, 19.7) 31.9* (30.6, 33.2) 13.3 30.3 (29.6, 31.1) 49.3* (48.5, 50.1) 19.0
2–5.9 years 11.3 (6.1, 16.4) 19.2 (13.0, 25.4) 7.9 19.8 (17.6, 22.1) 29.4* (26.8, 32.0) 9.5 26.8 (25.3, 28.3) 42.3* (40.7, 44.0) 15.5
6–9.9 years 14.5 (9.1, 19.9) 23.8 (17.5, 30.2) 9.3 22.5 (20.1, 24.9) 35.5* (32.7, 38.2) 13.0 37.8 (36.4, 39.2) 56.8* (55.3, 58.2) 19.0
10–13.9 years 16.0 (10.7, 21.2) 26.8 (19.8, 33.8) 10.8 21.4 (19.0, 23.7) 39.1* (36.3, 41.9) 17.7 36.7 (35.4, 38.1) 57.4* (56.0, 58.7) 20.6
14–18 years 12.3 (7.1, 17.5) 25.1* (18.3, 31.9) 12.8 10.2 (8.6, 11.9) 22.8* (20.5, 25.0) 12.5 18.1 (17.1, 19.2) 37.6* (36.3, 38.9) 19.5

Both sexes
All age 13.3 (11.2, 15.3) 21.0 (18.3, 23.7) 7.8 21.7 (20.8, 22.5) 33.2* (32.3, 34.1) 11.5 33.3 (32.7, 33.9) 49.1* (48.4, 49.8) 15.8
2–5.9 years 13.1 (9.2, 16.9) 20.2 (15.4, 24.9) 7.1 20.8 (19.2, 22.4) 31.3* (29.4, 33.2) 10.5 26.4 (25.3, 27.5) 41.1* (39.9, 42.4) 14.7
6–9.9 years 13.7 (10.1, 17.3) 21.1 (16.9, 25.3) 7.4 23.8 (22.1, 25.5) 35.5* (33.6, 37.5) 11.7 36.4 (35.4, 37.4) 52.9* (51.8, 54.0) 16.5
10–13.9 years 12.4 (8.9, 16.0) 20.7 (16.0, 25.4) 8.2 24.3 (22.6, 26.0) 37.4* (35.5, 39.3) 13.1 40.2 (39.2, 41.1) 57.1* (56.1, 58.2) 17.0
14–18 years 13.9 (10.2, 17.5) 22.0* (17.7, 26.3) 8.1 17.4 (15.9, 18.9) 28.0* (26.3, 29.7) 10.6 27.1 (26.3, 28.0) 41.9* (40.9, 42.9) 14.8

Abbreviations: BMI, body mass index; CDC, Centers for Disease Control and Prevention; CI, confidence interval; Diff., difference; Est., estimate; IOTF, International
Obesity TaskForce.
a
Point estimates (Est.) for the prevalence (%) and 95% CI were derived from survey and weighted commands to represent about 50% Chinese, 83% Indonesian and
100% Vietnamese populations.
b
Participants with a BMI o5th percentile for BMI (BMI z-score o1.645) of the CDC 2000 sex-specific BMI-for-age growth charts were classified to be underweight.
These underweight participants were the numerator for the estimation of the Prevalence CDC.
c
Participants with a BMI less than age- and sex-specific IOTF cut points for underweight were classified to be underweight. These underweight participants were the
numerator for the estimation of the Prevalence IOTF.
d
Absolute difference in prevalence (Diff., %) ¼ Prevalence IOTFPrevalence CDC.
*Different from Prevalence CDC (P-valueo0.05, w2-test).

European Journal of Clinical Nutrition

 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1309
25 Chinese Vietnamese (0.69–0.79) boys (Figure 4a). The ranges of
Absolute Difference (%)

Indonesian k-statistics in Chinese, Indonesian and Vietnamese girls

20 Vietnamese were 0.65–0.80, 0.63–0.76 and 0.55–0.68, respectively
in Prevalence

(Figure 4b). With an increase in age, k-statistics tended to

15 increase in boys and decrease in girls. The estimation of
k-statistics was precise (Supplementary Table 1).
10

5 Discussion
0
This study showed age, sex and ethnic differences in the
2-5.9 6-9.9 10-13.9 14-18
prevalence of underweight and overweight, defined by using
the CDC and IOTF cut points, in Chinese, Indonesian and
25 Chinese Vietnamese children. In addition, it showed a larger
Indonesian difference between the prevalence of underweight than that
Absolute Difference (%)

20 Vietnamese of overweight.
in prevalence

We found a small difference in the prevalence of over-

15
weight, defined by using the CDC and IOTF cut points,
which is similar to the findings from some study populations
10
in high-income countries. An absolute difference of o5%
was found in 596 Irish children aged from 5 to 12 years
5
(O’Neill et al., 2006), in 2098 Hong Kong Chinese children
aged from 11 to 18 years (Ko et al., 2008), in 258 Italian
0
children aged from 3 to 6 years (Vidal et al., 2006) and in
2-5.9 6-9.9 10-13.9 14-18
2431 Swiss children aged from 6 to 12 years (Zimmermann
et al., 2004). In addition, less than 2.5% difference was
25 Chinese observed in the prevalence of overweight and obesity in a
Indonesian
sample of 1709 American children aged from 4 to 15 years at
Absolute Difference (%)

20 Vietnamese
baseline in a Bogalusa Heart Study cohort (Janssen et al.,
in prevalence

15
10
5 prevalence of underweight than that of overweight. First,
0 higher compared with those of the CDC growth charts
2-5.9 6-9.9 10-13.9
Age (y)
Figure 3 The absolute difference (%) in the prevalence of under-
weight estimated by using the Centers for Disease Control and
Prevention (CDC) and International Obesity TaskForce (IOTF) cut
points in 2–18-year-old boys (a), girls (b) and in both sexes (c).
Absolute difference in prevalence (%) ¼ Prevalence IOTFPrevalence
CDC. Participants with a body mass index (BMI) o5th percentile for
BMI (BMI z-score o1.645) of the CDC 2000 sex-specific BMI-for-
age growth charts were classified to be underweight. These
underweight participants were the numerator for the estimation of
the Prevalence CDC. Participants with a BMI less than age- and
sex-specific IOTF cut points for underweight were classified to be
underweight. These underweight participants were the numerator
for the estimation of the Prevalence IOTF.
ethnic, sex and age groups (o2% in almost all groups, except
for the 2–5.9-year-old Chinese and Indonesian children
(from 2 to o5%)) (Table 2).
Values of k-statistic were highest in Chinese boys
(0.81–0.88) compared with the Indonesian (0.74–0.82) and
2005). With large representative samples of 2–18-year-old
children, our study may provide a more complete evaluation
of the consistency between the CDC and IOTF cut points.
Our study showed a larger difference between the
the IOTF cut points for underweight are systematically
14-18 (Kuczmarski et al., 2002; Cole et al., 2007), which lead to a
systematic increase in the prevalence of underweight defined
by using the IOTF cut points compared with those defined by
using the CDC growth charts. Second, because there is a
smaller difference and a crossover between the two sets of
cut points for overweight (Cole et al., 2000; Kuczmarski et al.,
2002), the inconsistency in the prevalence of overweight
may be neutralized, which lead to an overall increase in
consistency. Moreover, because the BMI distributions in the
Asian populations (especially in Indonesian and Vietnamese
children) were shifted to the left compared with those of the
American reference population, the absolute difference in
the prevalence of underweight was larger (because a lot of
children had low BMI), whereas the difference in the
prevalence of overweight was smaller (because only a small
number of children fell in the higher range of BMI).
This study showed an ethnic variation in the consistency
in findings from the CDC and IOTF cut points. There are
several potential explanations for the variation. First,
compared with the CDC growth charts, BMI distributions

European Journal of Clinical Nutrition

 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1310
Table 2 Prevalence (%) of overweight defined by the CDC 2000 BMI-for-age growth charts and the IOTF cut points in 2–18-year-old Chinese,
Indonesian and Vietnamese childrena

Chinese (n ¼ 1600) Indonesian (n ¼ 11 756) Vietnamese (n ¼ 53 826)

Prevalence CDC b Prevalence IOTF c Diff.d Prevalence CDC Prevalence IOTF Diff Prevalence CDC Prevalence IOTF Diff.

Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI

Boys
All age 18.6 (15.5, 21.7) 17.4 (14.4, 20.4) 1.2 5.2 (4.6, 5.8) 4.2 (3.7, 4.8) 1.0 2.1 (1.9, 2.3) 1.7 (1.5, 1.9) 0.4
2–5.9 years 25.8 (19.1, 32.5) 21.5 (14.9, 28.1) 4.3 10.7 (9.0, 12.4) 8.0 (6.5, 9.4) 2.8 4.4 (3.7, 5.1) 2.9 (2.3, 3.5) 1.5
6–9.9 years 21.0 (15.4, 26.6) 19.3 (14.2, 24.4) 1.7 4.3 (3.2, 5.4) 3.0 (2.0, 3.9) 1.3 2.4 (2.0, 2.8) 1.8 (1.4, 2.1) 0.6
10–13.9 years 16.0 (11.1, 21.0) 15.2 (10.5, 19.8) 0.8 4.0 (2.9, 5.1) 3.7 (2.6, 4.7) 0.3 1.7 (1.3, 2.0) 1.6 (1.2, 1.9) 0.1
14–18 years 11.8 (7.1, 16.5) 13.8 (8.7, 19.0) 2.1 2.3 (1.5, 3.1) 2.7 (1.8, 3.5) 0.3 0.8 (0.5, 1.0) 0.8 (0.6, 1.1) 0.1

Girls
All age 13.8 (11.1, 16.6) 12.8 (10.0, 15.5) 1.1 5.9 (5.2, 6.5) 5.5 (4.8, 6.1) 0.4 1.7 (1.4, 1.9) 1.5 (1.3, 1.7) 0.2
2–5.9 years 26.5 (18.9, 34.1) 21.9 (14.4, 29.3) 4.6 9.9 (8.3, 11.5) 7.7 (6.3, 9.1) 2.2 4.3 (3.5, 5.0) 3.2 (2.5, 3.9) 1.1
6–9.9 years 16.6 (11.0, 22.1) 15.5 (10.1, 21.0) 1.0 4.4 (3.2, 5.6) 4.5 (3.3, 5.7) 0.1 1.5 (1.2, 1.8) 1.4 (1.1, 1.7) 0.1
10–13.9 years 10.8 (7.0, 14.6) 10.8 (7.0, 14.6) 0.0 3.8 (2.7, 4.9) 3.7 (2.6, 4.8) 0.1 1.2 (0.9, 1.5) 1.2 (0.9, 1.5) 0.0
14–18 years 4.3 (1.1, 7.5) 4.8 (1.5, 8.1) 0.5 5.8 (4.5, 7.0) 6.3 (5.0, 7.5) 0.5 0.6 (0.4, 0.8) 0.7 (0.5, 1.0) 0.2

Both sexes
All age 16.4 (14.0, 18.7) 15.3 (13.0, 17.5) 1.1 5.5 (5.1, 6.0) 4.8 (4.4, 5.2) 0.7 1.9 (1.7, 2.1) 1.6 (1.4, 1.7) 0.3
2–5.9 years 26.1 (21.1, 31.1) 21.7 (16.8, 26.5) 4.5 10.3 (9.1, 11.5) 7.8* (6.8, 8.9) 2.5 4.3 (3.8, 4.9) 3.0* (2.5, 3.5) 1.3
6–9.9 years 19.0 (14.7, 23.4) 17.6 (13.6, 21.7) 1.4 4.3 (3.5, 5.2) 3.7 (2.9, 4.4) 0.6 1.9 (1.7, 2.2) 1.6 (1.3, 1.8) 0.4
10–13.9 years 13.6 (10.2, 16.9) 13.1 (9.9, 16.3) 0.4 3.9 (3.1, 4.7) 3.7 (2.9, 4.4) 0.2 1.4 (1.2, 1.7) 1.4 (1.1, 1.6) 0.1
14–18 years 8.1 (5.1, 11.1) 9.4 (6.2, 12.6) 1.3 4.0 (3.3, 4.7) 4.4 (3.6, 5.1) 0.4 0.7 (0.5, 0.8) 0.8 (0.6, 1.0) 0.1

Abbreviations: BMI, body mass index; CDC, Centers for Disease Control and Prevention; CI, confidence interval; Diff., difference; Est., estimate; IOTF, International
Obesity TaskForce.
a
Point estimates (Est.) for the prevalence (%) and 95% CI were derived from survey and weighted commands to represent about 50% Chinese, 83% Indonesian and
100% Vietnamese populations.
b
Participants with a BMI X85th percentile for BMI (BMI z-score X1.036) of the CDC 2000 sex-specific BMI-for-age growth charts were classified to be overweight.
These overweight participants were the numerator for the estimation of the Prevalence CDC.
c
Participants with a BMI greater than or equal to age- and sex-specific IOTF cut points for overweight were classified to be overweight or obesity. These overweight
participants were the numerator for the estimation of the Prevalence IOTF.
d
Absolute difference in prevalence (Diff., %) ¼ Prevalence IOTFPrevalence CDC.
*Different from Prevalence CDC (P-valueo0.05, w2-test).

were shifted to the left with the order Chinese, Indonesian
and Vietnamese children in all age and sex groups, which
lead to the decrease in the level of consistency from Chinese,
Indonesian to Vietnamese children. Second, differences in
maturational timing across populations might contribute to
the ethnic difference. As there is a trend of decreased
maturational age with an increase in BMI, body fat or
socioeconomic status (Wyshak and Frisch, 1982; Kaplowitz,
2006; Euling et al., 2008a, 2008b), which advances from
Vietnam, Indonesia, China, to the US (World Bank, 2008),
Chinese adolescents might reach maturation age earlier
compared with the Vietnamese and Indonesians, but latter
compared with Americans. An adjustment for maturation
age in these populations would lead to an increase in age-
and sex-specific mean BMI (Daniels et al., 1997; Wang and
Adair, 2001), and thus an increase in the consistency
between the CDC and IOTF cut points. As CDC and IOTF
cut points do not take into account maturation status (Cole
et al., 2000, 2007; Kuczmarski et al., 2002), it is suspected that
the more similar in maturational timing and BMI distribu-
tion may yield the more valid and consistent estimates of
the overweight and underweight prevalence. In addition, the
variations in genetic factors, dietary intakes, pre- and
postnatal growths, socioeconomic status and other environ-
mental factors might lead to different growth patterns
(Parent et al., 2003; Wehkalampi et al., 2008; Euling et al.,
2008a, b), and thus to the differences in BMI distributions
and in the consistent levels.
It is unexpected to observe a much higher prevalence of
underweight defined by international cut points (the IOTF)
compared with those defined by the CDC growth charts (for
American) in these Asian populations. The IOTF reference
population represents children from some high- and upper-
middle-income countries (Cole et al., 2000, 2007). A larger
number of children who might share similar growth patterns
in the IOTF reference population (B200 000 in the IOTF and
30 000 in CDC reference populations) would lead to a
decrease in values of standard deviation. As z-score (which
is based on standard deviation) was used as one of the
components for the development of IOTF cut points (Cole
et al., 2007), the IOTF cut points for underweight may
systematically be higher compared with those developed by
using the same method, but with a smaller and/or more
heterogeneous reference population.

European Journal of Clinical Nutrition

 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1311
1 not able to identify the more valid system. Further study is
needed to validate and/or define BMI cut points on the basis
* of current or future disease risk and total body fat in children
Kappa Statistic

* *† (Zimmermann et al., 2004; Janssen et al., 2005; O’Neill et al.,

*†
0.5 2006; Vidal et al., 2006; Voss et al., 2006; Ko et al., 2008)
using a more diverse reference population. Other concerns
Chinese relating to the use of secondary data are the control over
Indonesian
sample sizes and the comparability of data collection
Vietnamese
protocols. In this study, the sample sizes were large enough
0
2-5.9 6-9.9 10-13.9 14-18 and the estimates were precise in all stratified analyses. In
addition, data collection protocols for all studied variables
1 (for example, sex, weight, height and age) are comparable
across surveys.
In conclusion, this study showed age, sex and ethnic
Kappa Statistic

* differences in the prevalence of underweight and over-

*† *†
0.5 *† weight, defined by using the CDC and IOTF cut points, in
selective low- and lower-middle-income countries. The
Chinese
findings suggest that each country should carefully evaluate
Indonesian
Vietnamese any external cut points before using it in their populations.
0
Although country-specific cut points show advantages in
2-5.9 6-9.9 10-13.9 14-18 tracking nutrition status in a given country (Guillaume,
1999; Neovius et al., 2004; Must and Anderson, 2006), effort
1 should be made to create universally accepted cut points for
underweight and overweight in children to facilitate
Kappa Statistic

* * between-country comparisons.
*† *† *† *†
0.5
Conflict of interest
Chinese
Indonesian The authors declare no conflict of interest.
Vietnamese
0
2-5.9 6-9.9 10-13.9 14-18 Acknowledgements
Age (y)

Figure 4 k-statistic values in the classifications of underweight,
normal weight and overweight in 2–18-year-old boys (a), girls (b)
and in both sexes (c). Nutrition status of the children was
categorized into three groups: underweight, normal weight and
overweight by using the Centers for Disease Control and Prevention
(CDC) body mass index (BMI)-for-age growth charts and the
International Obesity TaskForce (IOTF) cut points. A k-statistic
indicates the consistency between the two rating systems (for
example, the CDC and the IOTF cut points). k-statistic 40.75
denotes excellent reproducibility, 0.4–0.75 denotes good reprodu-
cibility and 0–3.9 denotes marginal reproducibility. *Different from
Chinese; wDifferent from Indonesian (Po0.05; independent t-test).
Our findings were based on large, recent, representative
samples from 2- to 18-year-old children in China, Indonesia
and Vietnam, which (a) allowed us to explore magnitudes
and trends of consistency in various age, sex and ethnic
groups and (b) ensured the generalizability of the findings to
respective populations. In addition, this study contributes to
the current knowledge by examining the consistency
between CDC and IOTF cut points in low- and lower-
middle-income countries, and in both overweight and
underweight categories. Although we could point out some
differences between the two classification systems, we were
The study was finally supported by the Vietnam Educational
Foundation (VEF) and the USDA ARS CRIS 6250-51000.
References
Barlow SE (2007). Expert committee recommendations regarding the
prevention, assessment, and treatment of child and adolescent
overweight and obesity: summary report. Pediatrics 120, S164–S192.
Cole TJ, Bellizzi MC, Flegal KM, Dietz WH (2000). Establishing a
standard definition for child overweight and obesity worldwide:
international survey. BMJ 320, 1240–1243.
Cole TJ, Flegal KM, Nicholls D, Jackson AA (2007). Body mass index
cut offs to define thinness in children and adolescents: interna-
tional survey. BMJ 335, 194.
Daniels SR, Khoury PR, Morrison JA (1997). The utility of body mass
index as a measure of body fatness in children and adolescents:
differences by race and gender. Pediatrics 99, 804–807.
Euling SY, Herman-Giddens ME, Lee PA, Selevan SG, Juul A, Sorensen
TI et al. (2008a). Examination of US puberty-timing data from
1940 to 1994 for secular trends: panel findings. Pediatrics 121
(Suppl 3), S172–191.
Euling SY, Selevan SG, Pescovitz OH, Skakkebaek NE (2008b). Role of
environmental factors in the timing of puberty. Pediatrics 121
(Suppl 3), S167–S171.
Guillaume M (1999). Defining obesity in childhood: current practice.
Am J Clin Nutr 70, 126S–1130.

European Journal of Clinical Nutrition

 Nutrition status classifications in Asian children
NT Tuan and TA Nicklas
1312
Hesketh T, Ding QJ (2000). Standard definition of child overweight
and obesity worldwide. Young Chinese people in Hong Kong are
not representative of those in China. BMJ 321, 1158–1159.
Janssen I, Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM,
Bouchard C et al. (2005). Utility of childhood BMI in the
prediction of adulthood disease: comparison of national and
international references. Obesity 13, 1106–1115.
Kaplowitz P (2006). Pubertal development in girls: secular trends.
Curr Opin Obstet Gynecol 18, 487–491.
Ko G, Ozaki R, Wong G, Kong A, So W-Y, Tong P et al. (2008). The
problem of obesity among adolescents in Hong Kong: a compa-
rison using various diagnostic criteria. BMC Pediatr 8, 10.
Kuczmarski RJ, Ogden CL, Guo SS (2002). CDC growth charts for the
United States: Methods and development: National Center for
Health Statistics. Vital Health Stat 11 (246): 1–190.
Ministry of Health—General Statistical Office (2003). The Report:
Results of the Vietnam National Health Survey 2001-2002. Medical
Publishing House: Hanoi.
Murray CJ, Lopez AD (1997). Global mortality, disability, and the
contribution of risk factors: Global Burden of Disease Study. Lancet
349, 1436–1442.
Must A, Anderson SE (2006). Body mass index in children and
adolescents: considerations for population-based applications.
Int J Obes 30, 590–594.
Must A, Dallal GE, Dietz WH (1991). Reference data for obesity: 85th
and 95th percentiles of body mass index (wt/ht2) and triceps
skinfold thickness. Am J Clin Nutr 53, 839–846.
Neovius M, Linne Y, Barkeling B, Rossner S (2004). Discrepancies
between classification systems of childhood obesity. Obes Rev 5,
105–114.
O’Neill JL, McCarthy SN, Burke SJ, Hannon EM, Kiely M, Flynn A et al.
(2006). Prevalence of overweight and obesity in Irish school children,
using four different definitions. Eur J Clin Nutr 61, 743–751.
Pan H, Cole TJ (2008). lmsGrowth, a Microsoft Excel add-in to access
growth references based on the LMS method. Version 2.64.
Parent AS, Teilmann G, Juul A, Skakkebaek NE, Toppari J,
Bourguignon JP (2003). The timing of normal puberty and the
age limits of sexual precocity: variations around the world, secular
trends, and changes after migration. Endocr Rev 24, 668–693.
Popkin BM (2006). Global nutrition dynamics: the world is shifting
rapidly toward a diet linked with noncommunicable diseases.
Am J Clin Nutr 84, 289–298.
Supplementary Information accompanies the paper on European Journal of Clinical Nutrition website (http://www.nature.com/ejcn)
European Journal of Clinical Nutrition
Popkin BM, Paeratakul S, Zhai F, Ge K (1995). Dietary and
environmental correlates of obesity in a population study in
China. Obes Res 3 (Suppl 2), 135s–143s.
Reichenheim ME (2004). Confidence intervals for the kappa statistic.
Stata J 4, 421–428.
Rosner B (2006). Hypothesis testing: categorical data. In: Rosner B
(ed). Fundamentals of Biostatistics. Thomson-Brooks/Cole: Belmont,
CA. pp. 385–463.
Strauss J, Beegle K, Sikoki B, Dwiyanto A, Herawati Y, Witoelar F
(2004). The third wave of the Indonesia Family Life Survey (IFLS3):
overview and field report: RAND Labor and Population WR-144/
1-NIA/NICHD.
Tuan NT, Tuong PD, Popkin BM (2008). Body mass index (BMI)
dynamics in Vietnam. Eur J Clin Nutr 62, 78–86.
Vidal E, Carlin E, Driul D, Tomat M, Tenore A (2006). A comparison
study of the prevalence of overweight and obese Italian preschool
children using different reference standards. Eur J Pediatr 165,
696–700.
Voss LD, Metcalf BS, Jeffery AN, Wilkin TJ (2006). IOTF thresholds for
overweight and obesity and their relation to metabolic risk in
children (EarlyBird 20). Int J Obes (Lond) 30, 606–609.
Wang Y, Adair L (2001). How does maturity adjustment influence the
estimates of overweight prevalence in adolescents from different
countries using an international reference? Int J Obes Relat Metab
Disord 25, 550–558.
Wehkalampi K, Silventoinen K, Kaprio J, Dick DM, Rose RJ,
Pulkkinen L et al. (2008). Genetic and environmental influences
on pubertal timing assessed by height growth. Am J Hum Biol 20,
417–423.
WHO (2005). Preventing Chronic Diseases: A Vital Investment: WHO
Global Report. WHO: Geneva.
WHO Expert Committee (1995). Physical Status: The Use and
Interpretation of Anthropometry. WHO: Geneva.
World Bank (2008). World Development Indicators 2008. World Bank
Publications: Washington, DC.
Wyshak G, Frisch RE (1982). Evidence for a secular trend in age of
menarche. N Engl J Med 306, 1033–1035.
Zimmermann MB, Gubeli C, Puntener C, Molinari L (2004). Detection
of overweight and obesity in a national sample of 6-12-y-old Swiss
children: accuracy and validity of reference values for body mass
index from the US Centers for Disease Control and Prevention and
the International Obesity Task Force. Am J Clin Nutr 79, 838–843.