Original Article Obesity

EPIDEMIOLOGY/GENETICS

BMI-for-Age Z-Score Distribution Shifts

Among Chinese Children: Gender Disparity
Yi Song1, Hai-Jun Wang1*, Jun Ma1*, Patrick W.C. Lau2, Peijin Hu1, Bing Zhang1 and Zhiqiang Wang1,3

Objective: To identify gender differences among Chinese school-aged children from 1995 to 2010, and
to project the future BMI-for-age Z-score distribution and prevalence of obesity.
Methods: The data were from four cross-sectional surveys (1995, 2000, 2005, and 2010) of Chinese
National Surveys on Students Constitution and Health (CNSSCH) with a national representative sample of
Chinese children, involving more than 200,000 participants at each survey. BMI-for-age Z-score distribu-
tional shifts overall and in percentiles were compared by gender. Average shift was calculated for four
survey periods and used for projecting future distributions and obesity prevalence.
Results: BMI-for-age Z-score increased more in their upper percentile distribution, indicating that Chi-
nese children have become heavier over the past 15 years. Gender disparity in BMI-for-age Z-score has
become wider during the period. Over a 15-year period, BMI-for-age Z-score shift among girls has been
stable, while boy’s BMI-for-age Z-score shifts has increased linearly. By 2020, the obesity prevalence is
predicted to be 10.18% and 4.99% for boys and girls, respectively.
Conclusions: The wider gender disparity suggested a larger proportion of obesity in boys than in girls.
Therefore, gender-specific preventive guidelines and public health policies for childhood obesity and
cardiovascular diseases are urgently needed in China.
Obesity (2014) 22, 1187-1193. doi:10.1002/oby.20676

Introduction findings in local areas of China and other countries have also dem-
onstrated that the distribution of BMI has increased at every percen-
The increasing prevalence of childhood obesity constitutes a serious
tile in both genders, and in particular, there were greater shifts in
public health problem in both developed and developing countries
the upper part of the distribution between the 85th and 95th percen-
(1-3). China has the world’s largest population, in which the preva-
tiles (11-18).
lence of childhood overweight has markedly increased from 1.11%
in 1985 to 9.62% in 2010, while the prevalence of childhood obesity
It has been reported that the changing pace of obesity prevalence in
increased from 0.13% in 1985 to 4.95% in 2010 (4). BMI is the
boys was faster than that in girls (19). However, the question still
common index to define overweight and obesity (5,6). Prior studies remains unanswered as to whether shifts in the whole spectrum of
have shown that mean BMI values, similar to obesity prevalence, BMI distribution varied by gender over time during the past 15
have continued to increase among children in the past decades in years in China. Analyzing the BMI distribution shifts may help us
many countries (7-9), for example, the increases in mean BMI to clarify whether the increase in obesity was the result of changes
among Japanese school children from 1976 to 2000 were 0.32 and among the entire population, or whether these changes were concen-
0.24 kg/m2 per decade for boys and girls, respectively (10). Previous trated in a subgroup of more susceptible individuals, such as a

1
Institute of Child and Adolescent Health, School of Public Health, Peking University, Beijing, China. Correspondence: Hai-Jun Wang (whjun1@bjmu.
edu.cn) or Jun Ma ([email protected]) 2 Department of Physical Education, Hong Kong Baptist University, Kowloon Tong, Hong Kong, China 3 Centre
for Chronic Disease, School of Medicine, University of Queensland, Health Sciences Building, Royal Brisbane & Women’s Hospital, Herston, Queensland
4029, Australia

Funding agencies: The preparation for publication of the present study were supported by grant from National Natural Science Foundation of China (81172683), and
review and data analysis were supported by new teacher fund for doctor station, the Ministry of Education of China (20110001120102) and NHMRC of Australia
(APP1045000).
Disclosure: The authors declared no competing interests.
Author contributions: YS, HJW, and JM conceived and carried out experiments, YS, PJH, BZ, HJW, JM conceived experiments and analyzed data. YS, PJH, BZ, JM
carried out experiments, YS, HJW, JM, PJH, BZ, ZQW contributed reagents/materials/analysis tools. All authors were involved in writing the paper and had final approval
of the submitted and published versions.
Additional Supporting Information may be found in the online version of this article.
Received: 17 September 2013; Accepted: 3 December 2013; Published online 9 December 2013. doi:10.1002/oby.20676

www.obesityjournal.org Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 1187

 1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Obesity Gender Disparity on BMI Z-Score Distribution Song et al.

TABLE 1 Sample sizes in CNSSCH of different years

Boys Girls Total

Age group
(yrs) 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010

7- 8188 8746 9697 8973 8186 8695 9623 8974 16,374 17,441 19,320 17,947
8- 8183 8771 9725 8970 8203 8701 9608 8964 16,386 17,472 19,333 17,934
9- 8185 8704 9759 8971 8189 8695 9643 8985 16,374 17,399 19,402 17,956
10- 8745 8813 9874 8980 8758 8837 9766 8980 17,503 17,650 19,640 17,960
11- 8749 8691 9869 8990 8656 8664 9665 8992 17,405 17,355 19,534 17,982
12- 8740 8700 9717 8979 8599 8724 9537 8983 17,339 17,424 19,254 17,962
13- 8740 8591 9677 8978 8553 8703 9749 8969 17,293 17,294 19,426 17,947
14- 8747 8680 9606 8985 8577 8665 9622 8975 17,324 17,345 19,228 17,960
15- 8745 8710 9930 8979 8583 8651 9818 8977 17,328 17,361 19,748 17,956
16- 8681 8694 9800 8962 8518 8666 9744 8939 17,199 17,360 19,544 17,901
17- 8728 8590 9749 8954 8531 8568 9695 8971 17,259 17,158 19,444 17,925
18- 8653 8114 9273 8930 8457 8024 8992 8920 17,110 16,138 18,265 17,850
Total 103,084 103,804 116,676 107,651 101,810 103,593 115,462 107,629 204,894 207,397 232,138 215,280

gender-specific subgroup. Furthermore, it is also important for us to province from 1995 to 2010. More than 85% of the sampled schools
discover whether there was the gender disparity of BMI distribution remained the same in all surveys. This study only included subjects
shift, and whether this difference between the genders changed with of Han ethnicity, who accounted for 92% of the total Chinese popu-
time, which may help researchers and health policy makers to con- lations, from 26 mainland provinces and four municipalities of
sider targeted intervention strategies. mainland China, excluding Tibet (where Han ethnicity is minority).
All eligible participants had lived in the same area for at least one
This study used the recent data of Chinese National Survey on Stu- year and received a medical examination before measurement, to
dents’ Constitution and Health (CNSSCH, 2010), and earlier data ensure that they had no physical or mental disorders. All subjects
(CNSSCH, 1995, 2000, 2005), which are a series of nationally rep- were selected by stratified cluster sampling, from some classes as
resentative surveys among school-aged children in China. Consider- clusters randomly selected from each grade in the selected
ing the BMI values of children change with age, and BMI-for-age school. Each province had an equal size of sample from three
Z-scores were commonly used (20), we calculated BMI-for-age Z- socioeconomic classes (“upper”, “moderate”, and “low”) at the
scores based on the distribution of the reference population (19-21). region level. Five aspects were taken into consideration in defining
The objectives of the present study included: (1) to assess the the socioeconomic status (SES) of the region level: regional gross
trend in BMI-for-age Z-score among a representative sample of domestic product, total yearly income per capita, average food con-
Chinese children over a 15 year period, (2) to identify the gender sumption per capita, natural growth rate of population, and the
differences in BMI-for-age Z-score shifts, and (3) to project the regional social welfare index (19,22). In each gender-age subgroup,
future BMI-for-age Z-score distribution and prevalence of obesity there were 8024 to 9869 participants, and the ratio of boy/girl or
based on the past shifts and several assumptions, such as no change urban/rural approximately equaled to 1:1 in each survey (Table 1).
in demographic composition of Chinese children during the next The project was approved by the Medical Research Ethics Commit-
10 years. tee of the University of Queensland (#2011001199).

Measures
Methods Height (cm) and weight (kg) were measured according to the fol-
Subjects lowing standard procedures in all survey sites (18,21). Participants
The CNSSCH has been conducted every five years since 1985, were required to wear light clothes and stand straight, barefoot and
jointly launched by the Ministry of Education, the Ministry of at ease when being measured. Weight was measured to the nearest
Health, the Ministry of Science and Technology, the State of Nation 0.1 kg with a standardized lever scale and height to the nearest 0.1
Affairs, and the State Sports General Administration of People’s cm with a portable stadiometer. Both the scales and stadiometers
Republic of China. We used four surveys (1995, 2000, 2005, and were calibrated before use. BMI was calculated as body weight (kg)
2010) of CNSSCH, as our goal was to examine the shifts over the divided by height (m) squared (kg/m2). BMI-for-age Z-score is a
recent 15 years, and these more recent data were better for predict- quantitative measure of the deviation of a specific BMI value from
ing future trends. The sampling procedure previously described in the mean of that population, and was calculated with the WHO
detail (3) was used to ensure sufficient numbers of participants for 2007 references (20). The reference population used here is the
all geographic and economic regions in all CNSSCH at different fixed population recommended by WHO. Obesity was defined by
time points. The participants were primary and high school students using the references developed by Working Group on Obesity in
aged 7-18 years, who were selected from the same areas in each China (WGOC) (23). The children and adolescents with observed

1188 Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 www.obesityjournal.org

 1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article Obesity
EPIDEMIOLOGY/GENETICS

BMI95th age- and gender-specific BMI percentile value were Distribution shift patterns in BMI-for-age Z-score
defined as obese. For both males and females aged 18 years, those by gender
with observed BMI  28 kg/m2 were considered as obese (23). All Figure 1 showed BMI-for-age Z-score distribution curves and their
measurements were conducted by a team of public health professio- shift patterns over time by gender. For both genders, the curves of
nals in each survey site and they were required to pass a training BMI-for-age Z-score at different years were presented as bell-shaped
course for anthropometric measurements. distributions experiencing a lowering of the peaks with time. Addi-
tionally we observed the curve of either boys or girls shifted to the
Statistical analyses right with time. The shifts among girls were less obvious than that
Differences in BMI and BMI-for-age Z-score means, between two among boys. Among girls, the BMI-for-age Z-score distribution
adjacent years, were tested using ANOVA test. P-value <0.05 was appeared to shift at a relatively equal pace across all percentiles.
considered as statistically significant. The distributions of BMI-for- However, unlike girls, boys experienced shifts in BMI-for-age Z-
age Z-scores for 1995, 2000, 2005, and 2010 CNSSCHs were rep- score being relatively faster at percentiles above the median, com-
resented using kernel densities, which are nonparametric smoothed pared to those below it.
graphs independent of bin width when compared to histograms.
The changes in percentiles of BMI-for-age Z-scores distribution at During the study period, the median BMI Z-score increased 0.17 for
the four surveys were analyzed graphically using the mean–differ- boys and 0.10 for girls per 5 years (Table S1). The gender differ-
ence plot (m–d) curves. In the curves, the actual percentile was ence was 0.07 (95% CI: 0.06-0.07) which was statistically signifi-
plotted on the x-axis against the absolute shifts in the mean of cant P < 0.001. The increments in 75th percentile were significantly
BMI-for-age Z-scores between two adjacent surveys on the y-axis. higher than in 25th percentile in both genders (P < 0.001). The 95th
We used the quantile regression models to assess the trends of dif- percentile increased 0.34 per five years in boys and 0.20 in girls
ferent BMI percentiles and to test the differences between genders with a difference of 0.14 (95% CI: 0.13-0.14) while the gender dif-
(24). ference in the increasment in 5th percentile was 0.00 (95% CI:
20.01-0.01).
Predictions of future BMI-for-age Z-score distribution and the preva-
lence of obesity were conducted by using time series analyses— To study time shifts in BMI-for-age Z-score distribution more
Holt’s exponential smoothing method (25). Assuming no changes in closely, changes per 5 years in mean BMI-for-age Z-score in each
age and gender distributions among the Chinese children, means of percentile group were plotted in Figure 2. BMI-for-age Z-score
BMI-for-age Z-scores and prevalence of obesity were projected. The increased more at the upper percentiles, while the increase was very
analysis was conducted based on the equations below: limited at the lower percentiles. Gender differences in shift patterns
were noted. Among boys, 5-year shifts varied from <0.1 at the
St 5aXt 1ð12aÞðSt21 1bt21 Þ lower percentiles, to around 0.5 at the upper percentiles. While the
changes from 1995 to 2000 (I1) showed the steepest shifts from the
5th to 95th percentile, the shifts from 2000 to 2005 (I2) displayed a
bt 5bðSt 2St21 Þ1ð12bÞbt21
moderate increasing pattern. Changes from 2005 to 2010 (I3) were
Ft1m 5St 1bt m less steep than that from 1995 to 2000 between 5th and 85th percen-
tile. Among girls, shifts varied from <0.1 at the lower percentiles to
St is the new smoothed mean value that drifts over time; bt the coef- around 0.3 at the upper percentiles. However, each shift curve for
ficient on time (trend estimate) that also drifts over time, Xt the new boys was steeper than those for girls. In addition, the greatest shift
actual value at time t, Ft1m the forecast for m periods into the in BMI-for-age Z-score was observed at the 90th percentile both in
future, a and b are smoothing parameters that minimize the in- boys and girls, noticeably greater among boys than that among girls.
sample sum-of-squared forecast error. The analyses were conducted Furthermore, between 1995 and 2010 years, the greatest differences
using SPSS 20.0 and Stata 12.1 (24). in BMI-for-age Z-scores at the 10th, 50th, and 90th percentiles
occurred during puberty ages (boys: 10-13 years; girls: 9-12 years,
data not shown).

Results The students lived in the areas with higher SES has the higher BMI-
for-age Z-score according to the percentile increasing (Table S2).
Fifteen-year shifts in mean BMI-for-age Z-score
Moreover, there were the larger shifts in upper percentiles of BMI
by gender and age in the high SES areas than those in moderate or low SES areas
Mean BMI-for-age Z-score by gender and age, as well as increments (Figure S1).
per 5 years, were presented in Table 2. The BMI-for-age Z-scores
increased by 0.56 units for boys, and 0.32 units for girls over the
past 15 years, while it increased by 0.18-0.20 units per 5 years for
boys, and 0.10-0.11 units per 5 years for girls. The increments were
greater in boys than in girls, and the gender disparity has become Projected future BMI-for-age Z-score
wider with time, i.e. BMI-for-age Z-score difference between boys and prevalence of obesity by gender
and girls in 1995, 2000, 2005, and 2010 was 0, 0.1, 0.1, and 0.2 Figure 3 shows our future projections for 2015 and 2020. By 2020,
units, respectively. During the 15-year period, mean BMI has mean BMI-for-age Z-scores and prevalence of obesity will be signif-
increased by 1.24 kg/m2 for boys, and 0.71 kg/m2 for girls, while it icantly higher among boys than girls. The gap in BMI-for-age Z-
increased by 0.40-0.42 kg/m2 per 5 years for boys, and 0.23-0.25 score between boys and girls is predicted to become wider with
kg/m2 per 5 years for girls (data not shown). time. A similar pattern is observed for the prevalence of obesity,

www.obesityjournal.org Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 1189

 1190
Obesity

TABLE 2 Shifts in mean BMI-for-age Z-score from 1995 to 2010, by gender and age

Boys Girls
Age
group
(yrs) 1995 2000 2005 2010 I1 I2 I3 It 1995 2000 2005 2010 I1 I2 I3 It
a a a a a a
7- 20.661.0 20.361.2 20.161.3 0.161.4 0.17 0.11 0.08 0.35 20.760.9 20.561.0 20.361.0 20.261.1 0.21 0.15 0.09 0.45

Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014

8- 20.661.0 20.361.2a 0.061.3a 0.161.3a 0.05 0.11 0.12 0.29 20.860.9 20.661.0a 20.461.1a 20.361.1a 0.19 0.19 0.10 0.48
9- 20.661.1 20.361.2a 0.061.3a 0.261.4a 0.31 0.23 0.17 0.72 20.860.9 20.661.1a 20.461.1a 20.361.1a 0.24 0.17 0.13 0.54
10- 20.661.1 20.361.3a 0.061.3a 0.261.4a 0.31 0.29 0.16 0.76 20.861.0 20.661.1a 20.461.1a 20.361.2a 0.21 0.19 0.15 0.55
11- 20.661.2 20.361.3a 20.161.3a 0.261.4a 0.28 0.31 0.18 0.77 20.861.1 20.661.1a 20.461.2a 20.361.2a 0.15 0.20 0.14 0.49
12- 20.761.1 20.461.3a 20.261.3a 0.161.3a 0.32 0.29 0.17 0.78 20.761.1 20.661.1a 20.561.1a 20.361.1a 0.11 0.14 0.18 0.44
13- 20.661.1 20.561.2a 20.361.3a 20.161.3a 0.26 0.27 0.25 0.79 20.661.0 20.661.1 20.461.1a 20.361.1a 0.00 0.14 0.15 0.29
14- 20.661.1 20.561.2a 20.461.2a 20.261.2a 0.24 0.27 0.22 0.73 20.560.9 20.561.0 20.461.0a 20.361.0a 0.02 0.11 0.12 0.24
15- 20.761.0 20.561.1a 20.561.2a 20.361.2a 0.10 0.18 0.26 0.54 20.560.9 20.560.9a 20.461.0a 20.361.0a 0.04 0.06 0.06 0.16
16- 20.760.9 20.661.0a 20.561.1a 20.461.1a 0.09 0.15 0.21 0.45 20.460.8 20.460.8a 20.460.9 20.360.9a 0.05 0.00 0.04 0.08
17- 20.760.9 20.661.0a 20.661.1 20.461.1a 0.11 0.07 0.18 0.36 20.460.8 20.460.8a 20.460.9 20.460.9a 0.06 20.02 0.04 0.07
18- 20.860.8 20.661.0a 20.761.1 20.661.1a 0.09 0.06 0.11 0.26 20.560.8 20.460.8a 20.460.8a 20.460.9 0.08 20.05 0.00 0.03
Total 20.661.0 20.461.2a 20.361.3a 20.161.3a 0.20 0.18 0.18 0.56 20.660.9 20.561.0a 20.461.0a 20.361.0a 0.11 0.11 0.10 0.32

Values were presented as mean 6 SD if not indicated otherwise.

I1: increments of BMI-for-age Z-score from 1995 to 2000.
I2: increments of BMI-for-age Z-score from 2000 to 2005.
I3: increments of BMI-for-age Z-score from 2005 to 2010.
It: total increments of BMI-for-age Z-score from 1995 to 2010.
a
P < 0.05, indicating difference in mean BMI-for-age Z-score values between two adjacent times was statistically significant.

www.obesityjournal.org
Gender Disparity on BMI Z-Score Distribution Song et al.

1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article Obesity
EPIDEMIOLOGY/GENETICS

Figure 1 BMI-for-age Z-score distribution shifts among Chinese children from 1995 to 2010 by gender.

which will reach 8.45% in 2015 and 10.18% in 2020 among boys, genders, i.e. heavy Chinese children have become heavier over time.
more than 2 folds compared with those among girls. Changes of BMI percentile values and obesity prevalence in Chinese
children have been reported (3,11,12,18). However, no reports have
documented the gender disparity in the changes in the distribution
of BMI-for-age Z-scores and obesity prevalence together among
Discussion Chinese children. To our knowledge, for the first time, we reported
This study indicated that BMI-for-age Z-scores and BMI values a significant gender disparity in BMI trend among Chinese children.
have increased at almost all age groups over the past 15 years, Specifically, the shifts in BMI-for-age Z-scores were relatively faster
which is consistent with previous reports in China. One study at percentiles above the median than below, and faster in boys than
reported 6.6% of boys and 3.3% of girls aged 7-18 years were obese in girls. Although the BMI-for-age Z-scores were almost identical
in 2010 (4), compared to only 1.4% and 0.9% in 1995 (3). Another for boys and girls in 1995, more boys have become obese than girls
study of Chinese children in 16 major cities (12) found that overall over the past 15 years. The average increments in BMI-for-age Z-
increments of BMI values from 1950 to 2005 were 2.6 kg/m2 for scores also showed a gender disparity, in which the increments were
males and 1.8 kg/m2 for females, with increasing rates of 0.8 and steeper for boys than for girls at most of percentiles.
0.6 per decade, respectively. Zhang et al. (18) also demonstrated
that in Shandong province of China from 1985 to 2010, the average Gender differences in trends in BMI and obesity have also been
increments of BMI were 2.18 kg/m2 for boys, and 1.21 kg/m2 for reported in other populations, including children and adults. Ogden
girls. These findings are in alignment with the previous studies in et al. found that there was a significant increase in obesity preva-
other countries (8-10,14-17). lence between 1999-2000 and 2009-2010 in males aged 2-19 years
(odds ratio, 1.05; 95% CI, 1.01-1.10) but not in females (odds ratio,
We confirmed that the shifts across the spectrum of BMI-for-age Z- 1.02; 95% CI, 0.98-1.07) in US and there was a significant increase
scores experienced faster increases at the upper percentiles in both in BMI among adolescent males aged 12 through 19 years (P 5

Figure 2 Average increments in BMI-for-age Z-score across their distribution by gender.

www.obesityjournal.org Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 1191

 1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Obesity Gender Disparity on BMI Z-Score Distribution Song et al.

Figure 3 Projected future mean BMI-for-age Z-score and obesity prevalence for 2015 and 2020 among Chinese students aged 7-18 years old
by gender.

0.04) but not among any other age group or among females (7). In boys (32-34). The Chinese 2005 NYRBS showed that 23.6% of
Jena, East Germany, Zellner et al. found that the long term of aver- girls and 9.1% of boys tried to lose weight by restricting their diets
age BMI increase of the 7- to 14-year-old children was 1.8 kg/m2 in (29).
boys and 2.1 kg/m2 in girls between the first investigation in 1880
and the last study in 2005/2006 (16). Beydoun and Wang reported Based on the shifts observed over the past 15 years in this study,
that shifts across the spectrum of BMI and waist circumference var- assuming little change in demographic composition for China over
ied between race/ethnicity and gender groups in US adults. They the next 10 years, we projected the future BMI-for-age Z-score dis-
found that non-Hispanic black women had the largest shifts and tribution for each gender group, and the prevalence of obesity in
would have central obesity and obesity prevalence of 90.8% and China for the first time. Although the predicted obesity prevalence
70.7% by 2020 (26). Based on these previous studies and our study, in 2020 in China (10.18% among boys and 4.99% among girls) are
we found the gender disparity of BMI had different characteristics lower than those in many other countries, such as United States
among different populations. Our study showed specific gender dis- (19.6% of males and 17.1% of females aged 12-19 during 2009-
parity of BMI shifts in Chinese children, which provided important 2010) (7), and England (17.1% among boys and 14.8% among girls
evidence for developing public health strategy based on the context. from 2010 Health Survey for England) (8), the gender difference
will be wider with a more rapid increase than in other countries.
As genetic drifts are unlikely to occur within 15 years, there are Furthermore, the absolute number of obese children will be huge, as
several possible reasons for gender differences in BMI-for-age Z- more than 9.6 million Chinese children was obese in 2010, and
score distribution shifts. Firstly, according to the traditional Chi- approximately 14.8 million will be obese in 2020, based on the total
nese culture, parents prefer sons to daughters. The “one child poli- number of children aged 7-18 (194,599,052) in the 2010 population
cy” in China was initiated formally in 1979, which had some bene- census report of the People’s Republic of China (35). Moreover,
fits for children in terms of more food supply, better housing, and obesity is an independent risk factor of cardiovascular diseases
health subsidies (27). However, parents gave more food to boys (CVD), which is the leading cause of death in China (36). The num-
when the food supply became sufficient along with the secular ber of patients with CVD was 230 million in 2009 in China (36),
changes in society and economic development during the period of and the prevalence of coronary heart disease, stroke, and CVDs
1995-2000 (i.e. The Engel coefficients of urban and rural residents were 0.74%, 1.07%, and 1.78% in males; and 0.51%, 0.60%, and
were 35.7% and 41.1% in 2010 compared with 50.1% and 58.6% 1.10% in females, respectively (37). If the obesity prevalence of the
in 1995) (28). Our study also identified the larger shifts in upper gender disparity has been continuously widening, the gender dispar-
percentiles of BMI in high SES areas, indicating that the SES con- ity of CVDs will be exacerbated in the future. The government of
tributed to the BMI shifts. Secondly, the dietary and physical activ- the People’s Republic of China announced the regulation
ity behavioral difference between boys and girls may partly explain “Strengthening sports to enhance young people’s health” in 2007
the gender differences in BMI-for-age Z-score distribution shifts. (38). The aim was to refine and implement “Global strategy for diet,
For example, the Chinese 2005 NYRBS (National Youth Risk physical activity, and health” of the World Health Organization
Behavior Surveillance) reported that 4.3% of boys and 2.7% of (39). Based on our study results, i.e. the BMI-for-age Z-score and
girls frequently had soft-drinks, while 12.7% of boys and 4.3% of prevalence of obesity will increase and the gap between boys and
girls spent 2 h per day on playing computer games (29). Thirdly, girls will become wider with time, we suggested that gender-specific
as in other countries, body image dissatisfaction is more prevalent interventions are needed, such as correct body weight measure, body
among girls than among boys in China (30,31). It was reported image education in schools, etc. We advocate that the parents and
that 28.1% female teenagers were unhappy with their body image society should improve their cognition that obesity is a major health
compared with 16.1% males (31). In Asia, the desire to be thin is concern in children, especially among boys, and suggest separating
common among young girls in the culture context in which thin is the physical education class by gender that is not popular in many
beauty, and more girls regard thinness as an ideal body shape than schools but would increase the boys’ exercise (40).

1192 Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 www.obesityjournal.org

 1930739x, 2014, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/oby.20676 by University Of British Columbia, Wiley Online Library on [25/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article Obesity
EPIDEMIOLOGY/GENETICS

Our study has a few limitations. First, it was not a prospective 12. Ji CY, Chen TJ. Secular changes in stature and body mass index for Chinese youth
in sixteen major cities, 1950s-2005. Am J Hum Biol 2008;20:530-537.
cohort study, as each CNSSCH was cross-sectional survey con-
13. Wang H, Du S, Zhai F, Popkin BM. Trends in the distribution of body mass index
ducted in different subjects. The trends of BMI-for-age Z-score may among Chinese adults, aged 20-45 years (1989-2000). Int J Obes (Lond) 2007;31:
not reflect the exact increments along with increasing age. Second, 272-278.
the predictions of BMI-for-age Z-score distribution and prevalence 14. Chrzanowska M, Koziel S, Ulijaszek SJ. Changes in BMI and the prevalence of
overweight and obesity in children and adolescents in Cracow, Poland, 1971-2000.
of obesity were based on the trends observed during the past 15 Econ Hum Biol 2007;5:370-378.
years, and on an assumption of limited future population changes. It 15. Kurokawa N, Satoh H. Recent trends of body mass index distribution among school
is possible that change in population composition might affect the children in Sendai, Japan: Decrease of the prevalence of overweight and obesity,
predicted values. 2003-2009. Obes Res Clin Pract 2011;5:e1-e8.
16. Cardoso HFV, Caninas M. Secular trends in social class differences of height,
weight and BMI of boys from two schools in Lisbon, Portugal (1910-2000). Econ
The present study has several strengths. First, it was based on the Hum Biol 2010;8:111-120.
largest samples of Chinese children with national representation 17. Zellner K, Ulbricht G, Kromeyer-Hauschild K. Long-term trends in body mass
overall and for different age and gender subgroups. Second, it was index of children in Jena, Eastern Germany. Econ Hum Biol 2007;5:426-434.
the first study that examined the shifts in BMI-for-age Z-score in 18. Zhang YX, Wang SR. Secular trends in body mass index and the prevalence of
overweight and obesity among children and adolescents in Shandong, China, from
China and compared to the shifts across gender, and used past trends 1985 to 2010. J Pub Health 2011;34:131-137.
in distributional shifts to predict future BMI-for-age Z-score and 19. Song Y, Wang H-J, Ma J, Wang Z. Secular trends of obesity prevalence in urban
obesity prevalence. Chinese children from 1985 to 2010: Gender disparity. PLoS ONE 2013;8:e53069.
doi:10.1371/journal.pone.0053069.
20. World Health Organization (2007).WHO Reference 2007 SPSS macro package.
In conclusion, there were significant differences in BMI-for-age Z- [WWW document] http://www.who.int/entity/growthref/tools/readme_spss.pdf.
score distribution shifts between boys and girls in China. Gender- 21. Preedy VR (eds). Handbook of Anthropometry: Physical Measures of Human Form
specific preventive guidelines and public health policies for child- in Health and Disease. New York: Springer Science 1 Business Media, LLC; 2012.
hood obesity and cardiovascular diseases are urgently needed in 22. Ji C-Y, Cheng TO. Prevalence and geographic distribution of childhood obesity in
China in 2005. Int J Cardiol 2008;131:1-8.
China. O 23. Ji CY, WGOC. Body mass index reference for screening overweight and obesity in
Chinese school-age children. Biomed Environ Sci 2005;18:390-400.
24. StataCorp. Stata Statistical Software: Release 12. College Station, Texas: StataCorp
LP; 2011.
Acknowledgments 25. Rasmussen R. On time series data and optimal parameters. Omega. 2004;32:111-
120.
We thank WK Liao, WH Xing, and X Zhang for their permission 26. Beydoun MA, Wang Y. Gender–ethnic Disparity in BMI and Waist Circumference
on accessing the 1995, 2000, 2005, and 2010 CNSSCH data. We Distribution Shifts in US Adults. Obesity 2009;17:169-176.
also appreciate the students who participated in the surveys for their 27. Bredenkamp C (2008). Health reform, population policy and child nutritional status
in China. [WWW document] http://elibrary.workbank.org/content/workingpaper/
cooperation. 10.1596/1813-9405-4587.
28. National Bureau of Statistics of China. The income of urban and rural households
and Engel coefficients Available: http://data.stats.gov.cn/workspace/index?a5q
C 2013 The Obesity Society
V &type5global&dbcode5hgnd&m5hgnd&dimension5zb&code5A0A0105&region50
00000&time52011,2011. Accessed 2013 October 31.
References 29. Ji CY (eds). Report on the 2005 Chinese National Survey on Youth’s Health Risk
Behavior. Beijing: Peking University Medical Science Press; 2007. (In Chinese)
1. Ebbeling CB, Pawlak DB, Ludwig DS. Childhood obesity: Public-health crisis,
common sense cure. Lancet 2002;360:473-482. 30. Brennan M A., Lalonde C E. Bain JL. Body image perceptions: Do gender
differences exist? Psi Chi J Undergrad Res 2010; 15:130-138.
2. Lobstein T, Baur L, Uauy R. Obesity in children and young people: a crisis in
public health. Obes Rev 2004;5(s1):4-85. 31. Luo BW, Gao YB, Ye LH, Zhou LH, Peng WB. The study of the actuality of
teenagers’depression of body image. Psychol Dev Educ 2005;21:89-93. (In
3. Ji C-Y, Cheng TO. Epidemic increase in overweight and obesity in Chinese
Chinese).
children from 1985 to 2005. Int J Cardiol 2009;132:1-10.
4. Ma J, Cai CH, Wang HJ, et al. The trend analysis of overweight and obesity in 32. Chisuwa N, O’Dea JA.Body image and eating disorders amongst Japanese
Chinese students during 1985-2010. Chin J Prev Med 2012;46:781-789. (In adolescents. A review of the literature. Appetite 2010;54:5-15.
Chinese). 33. Hayashi F, Takimoto H, Yoshita K, Yoshiike N. Perceived body size and desire for
5. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for thinness of young Japanese women: a population-based survey. Br J Nutr 2006;96:
child overweight and obesity worldwide: International Survey. BMJ 2000;320:1240- 1154-1162.
1243. 34. Freedom Leung , Sharon Lam, Sherrien Sze. Cultural Expectations of Thinness in
6. Ma J, Wang Z, Song Y, Hu P, Zhang B. BMI percentile curves for Chinese Chinese Women. Eat Disord 2001;9:339-350.
children aged 7-18 years, in comparison with the WHO and the US Centers for 35. National bureau of statistics of China (2011). Tabulation on the 2010 population
Disease Control and Prevention references. Pub Health Nutr 2010;13:1990-1996. census of the People’s Republic of China. [WWW document] http://
7. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in www.stats.gov.cn/tjsj/pcsj/rkpc/6rp/indexch.htm.
body mass index among US children, 1999-2010. JAMA 2012;307:483-490. 36. Hu SS, Kong LZ, Gao RL, et al. Editorial board. Outline of the report on
8. National obesity observatory (2011). National Child Measurement Programme cardiovascular disease in China, 2010. Biomed Environ Sci 2012;25:251-256.
Changes in children’s body mass index between 2006/07 and 2009/10. 37. Yang ZJ, Liu J, Ge JP, Chen L, Zhao ZG, Yang WY. China National Diabetes and
[WWW document] http://www.noo.org.uk/uploads/doc/vid_13261_NCMP_Changes_ Metabolic Disorders Study Group. Prevalence of cardiovascular disease risk factor
in_childrens_BMI.pdf. in the Chinese population: The 2007-2008 China National Diabetes and Metabolic
9. Vignerova J, Humenikova L, Brabec M, Riedlova J, Blaha P. Long-term changes in Disorders Study. Eur Heart J 2012;33:213-220.
body weight, BMI, and adiposity rebound among children in the Czech republic. 38. The central people’s government of the People’ Republic of China (2005). Views of
Econ Hum Biol 2007;5:409-425. strengthening of youth sports to enhance young people’s health. 2007.5.7. http://
10. Matsushita Y, Yoshiike N, Kaneda F, Yoshita K, Takimoto H. Trends in childhood www.gov.cn/gongbao/content/2007/content_663655.htm.
obesity in Japan over the last 25 years from the national nutrition survey. Obes Res 39. World Health Organization (2004). Global strategy for diet, physical activity, and
2004;12:205-214. health. http://www.who.int/dietphysicalactivity/strategy/eb11344/en/index.html.
11. Zhang J, Wang H, Wang Z, He YN. Changes in the distribution of body mass 40. Song Y, Ma J (eds). Activity instruction on physical education and extra-curricular
index among aged children 6-17 in China from 1991 to 2006. Wei Sheng Yan Jiu sport among children and adolescents-collection of aerobic exercise prescription.
2008;37:728-732. (In Chinese) Beijing: Peking University Medical Science Press; 2013. (In Chinese)

www.obesityjournal.org Obesity | VOLUME 22 | NUMBER 4 | APRIL 2014 1193