METABOLIC SYNDROME AND RELATED DISORDERS

Volume XX, Number XX, 2018

 Mary Ann Liebert, Inc.
Pp. 1–7
DOI: 10.1089/met.2017.0150

The Metabolic Syndrome and Health Behaviors

in School Children Aged 13–16 Years in Ubon Ratchathani:
UMeSIA Project

Phalakorn Suebsamran, MS,1 Taksin Pimpak, MPH,1 Patnapa Thani, MS,1 and Parinya Chamnan, MD, PhD2,3
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.

Abstract
Background: Although metabolic syndrome (MetS) has become a significant public health problem worldwide,
little evidence exists to describe the prevalence of MetS in school children using MetS different classifications,
and its association with health-related behaviors.
Methods: Using data on 393 children and adolescents aged 13–16 years participating in the Ubon Ratchathani
Metabolic Syndrome in Adolescent (UMeSIA) Project, the prevalence of MetS was determined using the
International Diabetes Federation (IDF) 2007 definition and other three classifications reported in the previous
literature. The prevalence of the MetS was compared across sex, a family history of diabetes, type of schools,
and health behaviors using chi-square test. The prevalence of MetS and proportion of abnormalities in MetS
components were compared across different MetS classifications using Cochran Q test.
Results: The prevalence of MetS in Thai school children was 3.1%, 5.8%, 6.9%, and 11.2% when using IDF,
Cook’s, Ford’s, and De Ferranti’s classifications respectively (Difference in MetS prevalence across the four
MetS classifications, P < 0.001). Using IDF classification, prevalence was higher in male than female students
(5.9% and 1.2%, respectively, P < 0.001) and higher in those with a family history of diabetes than those
without (8.5% and 2.3%, respectively, P < 0.001). Students from a sports school had considerably lower MetS
prevalence than those from conventional schools (1.9% and 3.5% respectively, P < 0.001). Those with MetS
spent significantly longer time watching TV than those without (median (interquartile range) time to watch TV
180.0 (120.0, 240.0) and 120.0 (60.0, 180.0) min per day respectively, P = 0.002).
Conclusions: The prevalence of MetS in Thai school children was modest and varied greatly when different
MetS definitions were applied. Interventions to optimize time spent watching TV and increase physical activity
may be beneficial in reducing the risk of the MetS in children and adolescents.

Keywords: prevalence, metabolic syndrome, health behaviors, children and adolescents

Introduction across countries and ethnicity.3 Previous studies have re-

ported that the prevalence of MetS in children and adoles-

T he metabolic syndrome (MetS) has become a sig-

nificant public health problem worldwide. Since first
described in the 1970 s, MetS has been used as a risk factor
for diabetes and cardiovascular disease in adult populations
in many countries.1 Individuals with MetS have a three-fold
higher risk of developing heart attack or stroke than those
without, and the syndrome doubles the risk of death from
such diseases.2
It is estimated that approximately 20%–25% of adult
population worldwide have MetS, with varying prevalence
cents ranged 17%–50%; however, these studies were mainly
done in obese and overweight children.4–8 A number of
studies have examined the prevalence of MetS in children
and adolescents within normal weight ranges, and they re-
ported a prevalence of 1%–11%.6,9–13
A systematic review of 85 studies10 revealed that the
overall prevalence of the MetS in children varied consider-
ably, ranging from 0% to 19.2%, and the prevalence in
overweight and obese children was 11.9% and 29.2% re-
spectively. However, the review does not address the impact

1
College of Medicine and Public Health, Ubon Ratchathani University, Ubon Ratchathani, Thailand.
2
Ubon Ratchathani Provincial Public Health Office, Ubon Ratchathani, Thailand.
3
Cardiometabolic Research Group, Department of Social Medicine, Sanpasithipasong Hospital, Ubon Ratchathani, Thailand.

1
 2 SUEBSAMRAN ET AL.
of using different MetS classification on the burden of MetS.
Different classifications to define MetS may result in signif-
icant discrepancies in the burden of MetS. To our knowledge,
no studies so far have compared the prevalence of MetS in
Asian children when different classifications were applied.
Therefore, this study was aimed to describe the preva-
lence of MetS in school children in Ubon Ratchathani,
Thailand, using four different classifications. We also
compared the prevalence of MetS across sex, a family his-
tory of diabetes and type of schools. Lastly, we compared
the levels of health-related behaviors between those with
and without the to identify target behaviors for prevention.
Methods
This study was based on data from the Ubon Ratchathani
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.
Metabolic Syndrome in Adolescent Project (UMeSIA),
which is a school-based survey aimed at investigating the
burden and risk factors of the MetS in school children. The
UMeSIA project was done in five different secondary
schools in Ubon Ratchathani, Thailand. Four secondary
schools were randomly selected from a total of 25 schools in
district and provincial areas of Ubon Ratchathani and the
only sports school in the province was included. The sports
school is a school for children aged between 13 and 16 years
old who have athletic talents. In addition to general educa-
tion, great emphasis is given to physical education and
physical fitness to nurture its students to reach their poten-
tials in sports excellence. In each school, one classroom was
randomly selected for each year of study (Grades 7, 8, and
9). All students in each class were invited to participate in
this survey. Between October 2013 and September 2014, a
total of 393 students were included in the study. Study size
was determined based on previous data suggesting MetS
prevalence of 4.0% in Thai children in Ongkhaluck District,
Nakhonnayok, Thailand.6 The UMeSIA project was ap-
proved by Ubon Ratchathani University Ethics Committee
(Project Number UBU-EC-4/2557). All children and their
parents gave written informed consents.
Students were questioned by researchers about their
personal and medical history along with health behaviors,
including diet, physical activity, and leisure-time activity.
Data on participant’s age, sex, time spent watching TV,
tutorial time at weekend, a history of breast feeding, birth
weight, a family history of diabetes mellitus were collected
using interviewer-administered questionnaire. We also de-
veloped a new simple dietary scores questionnaire to assess
dietary behavior. The questionnaire was consisted of 10
simple questions about the frequency of dietary behaviors,
for example, ‘‘How often do you drink sugary beverages? ()
never () occasionally () regularly.’’ The content validity of
the questionnaire was determined by obtaining the item-
Table 1. Classifications Used to Define Metabolic Syndrome in Children
Definition WC TG (mg/dL)
IDF 2007 (3) ‡90th percentile ‡150
Cook et al.2003 (22) ‡90th percentile ‡110
De Ferranti et al.2004 (23) ‡75th percentile ‡100
Ford et al.2005 (21) ‡90th percentile ‡110
TG, triglycerides; FBS, fasting blood sugar; WC, waist circumference (age and sex-specific); BP, blood pressure; SBP/DBP, systolic/
diastolic blood pressure.
objective congruence value for each questionnaire item,
which ranged from 0.7 to 1.0. Its reliability was tested,
with the Cronbach Alpha coefficient of 0.571. Physical
activity was assessed using the Global Physical Activity
Questionnaire (GPAQ) and reported as MET-minutes per
week.14,15
Physical examination was taken and fasting blood sam-
ples were also obtained. Blood pressure was taken after 15-
min rest in a sitting position using a standard mercury
sphygmomanometer. Blood pressure was measured twice
1 min apart and the average of the two readings was used as
an individual’s blood pressure.16 Waist circumference was
measured using a nonstretch tape measure at the midpoint
between the lowest rib and iliac crest at the end of exhala-
tion.17 Body mass index (BMI) was calculated by dividing
an individual’s weight in kilograms by the square of height
in meters. Overweight was defined as a BMI at or above the
85th percentile and below the 95th percentile for children
and teens of the same age and sex. Obesity was defined as a
BMI at or above the 95th percentile for children and teens of
the same age and sex.18 Blood samples were drawn after 12-
hr overnight fasting. Approximately 30–45 min after the
samples were obtained, they were centrifuged at 2500–
3000 rpm for 10 min, and the serum was stored in a refrig-
erator at the temperature of below 8C before being sent to
the laboratory. Fasting blood sugar (FBS), triglycerides
(TG), total cholesterol and high-density lipoprotein (HDL)
were measured using Beckman Coulter AU680 automate.
Classifications of the MetS
Four different classifications were used to define MetS in
this study (Table 1). Subjects who had central obesity as
defined by each classification and the presence of two or
more abnormalities of other clinical features (increased
plasma glucose, elevated TG, low HDL-cholesterol, and
high blood pressure) were categorized as having MetS. High
blood pressure was defined individually after ranking blood
pressure of students with the same age and sex.
Statistical analyses
Participant characteristics were presented as mean (stan-
dard deviation: SD) and number (%) for continuous and
categorical variables, respectively. Chi-square test was used
to compare categorical variables between individuals with
and without MetS. Student t-test and Mann–Whitney-U test
were used to compare normally and non-normally distrib-
uted continuous variables between the two groups, respec-
tively. Cochran Q test was used to compare the prevalence
of MetS and abnormalities in each MetS component across
four different classifications. We also carried out a concor-
dance test to assess the agreement between two different
HDL (mg/dL) FBS (mg/dL) BP (mmHg)
<40 ‡100 SBP/DBP ‡130/85
‡40 ‡110 ‡90th percentile
<50 ‡110 ‡90th percentile
‡40 ‡100 ‡90th percentile
 THE METABOLIC SYNDROME IN SCHOOL CHILDREN 3

MetS classifications and reported Cohen’s kappa. We also
examined the prevalence of having high cardiometabolic
risk, according to the concept of metabolic health,19 in all
students and those with normal weight, overweight, and
obesity and also across tertiles of BMI and waist circum-
ference. Metabolically unhealthy was defined as individuals
who had one or more abnormalities of the following metabolic
parameters: blood glucose, triglyceride, HDL-cholesterol, and
blood pressure.16 A P-value of <0.05 was considered as sta-
tistically significant.
Results
Characteristics of 393 UMeSIA participants are presented
in Table 2. The mean (SD) age of study participants was
14.9 (0.98) years, with 38.7% being male. Individuals with
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.
ably higher in boys than girls (5.9% vs. 1.2% respectively,
P < 0.001). Those with a family history of diabetes had a higher
MetS prevalence than those without (8.5% vs. 2.3%,
P < 0.001). Regarding types of schools, students from a sports
school had lower MetS prevalence than those from conven-
tional schools (1.9% vs. 3.5% respectively, P < 0.001). MetS
prevalence was 45.0%, 5.3%, and 0.3% in students with obe-
sity, overweight, and normal weight, respectively (P < 0.001).
Those within the 3rd tertile of BMI had higher MetS prevalence
than those in the 2nd and 1st tertiles of BMI (8.4%, 0%, and
0.8% respectively, P < 0.001). Similar results were observed
regardless of the MetS classification used.
When using the IDF classification, the prevalence of ab-
normality in waist circumference, triglyceride, HDL-
cholesterol, systolic and diastolic blood pressure, and FBS
was 15.6%, 3.3%, 25.6%, 12.2%, 4.6%, and 0.8% respec-

MetS had higher BMI, waist circumference, systolic blood
pressure, and triglyceride than individuals without MetS
(P < 0.05). Those with MetS were more likely to have a
family history of diabetes and spent more hours watching
TV than those without MetS (P = 0.043 and P = 0.002 re-
spectively). Individuals with and without MetS were similar
regarding FBS levels, history of breastfeeding, birth weight,
simple dietary score, and tutorial time at weekend. Table 3
shows the prevalence of MetS by sex, family history of
diabetes, type of school, and the abnormalities of MetS com-
ponents, using different MetS classifications. The prevalence of
MetS in these school children was 3.1%, 5.8%, 6.9%, and
11.2% when using the International Diabetes Federation (IDF),
Cook’s, Ford’s and De Ferranti’s classifications respectively
(P-for-trend <0.001). The highest concordance was found be-
tween the definitions by Cook et al. and Ford et al. (kappa =
0.842), while the lowest concordance was between the de
Ferranti et al. and IDF definitions (kappa = 0.029). The preva-
lence of MetS defined by the IDF classification was consider-
tively (Table 3). De Ferranti’s classification almost doubled
the prevalence of abnormality in TG, HDL-cholesterol, and
waist circumference compared with using other classifica-
tions. IDF classification resulted in a significantly lower
prevalence of high diastolic blood pressure and triglyceride
than other classifications. There was no difference in the
prevalence of abnormalities in FBS and systolic blood
pressure when different MetS classifications were applied.
In an additional analysis, students from the sports school
had lower BMI and waist circumference than those from
conventional schools (Appendix Table A1). Sports school
students had substantially higher levels of physical activity
than the students from conventional schools [median (IQR)
7800 (4920–12,160) and 2400 (1060–4860) MET-min/week
respectively, P < 0.001]. Those from the sports school spent
lesser time watching TV than those from conventional
schools [median (IQR) 60.0 (60.0–112.5) and 120.0 (60.0–
180.0) min/day respectively, P < 0.001]. They had lower
FBS and more favorable blood lipids than those from

Table 2. Characteristics of 393 School Children Participating in the UMeSIA Project

All participants With MetS** Without MetS**
Characteristics (n = 393) (n = 12) (n = 381) P-value
Male sex* 152 (38.7) 9 (5.9) 143 (94.1) 0.020
Age{, years 14.9 (0.9) 15.1 (0.9) 14.9 (0.9) 0.488
BMI{, kg/m2 20.5 (3.6) 29.8 (5.5) 20.2 (3.1) <0.001
BMI category* <0.001
Obese 20 (5.1) 9 (75.0) 11 (2.9)
Overweight 38 (9.6) 2 (16.6) 36 (9.5)
Normal weight 335 (85.3) 1 (8.4) 334 (87.6)
Waist circumference{, cm 70.6 (9.4) 97.0 (13.3) 69.8 (8.0) <0.001
HDL cholesterol{, mg/dL 48.6 (12.0) 35.0 (11.0) 48.8 (11.7) <0.001
Triglyceride{, mg/dL 71.5 (36.0) 117.5 (47.7) 70.0 (30.0) 0.002
Systolic blood pressure{, mmHg 114.6 (13.5) 131.4 (13.0) 114.1 (13.3) <0.001
Diastolic blood pressure{, mmHg 68.0 (9.7) 73.9 (7.5) 67.7 (9.6) 0.030
Fasting blood sugar{, mg/dL 84.1 (6.1) 86.0 (6.5) 83.8 (5.7) 0.212
Breast feeding for 6 months* 221 (57.3) 7 (3.2) 214 (96.8) 0.826
Birth weight{, grams 2935.4 (512.4) 3233.3 (924.6) 2924.7 (493.7) 0.348
Family history of diabetes mellitus* 47 (12.0) 4 (8.5) 43 (91.5) 0.043
metabolic equivalent{, MET minutes/week 4857.6 (4542.8) 3724.8 (2808.0) 4925.1 (4562.2) 0.339
Dietary Score{ 22.1 (2.4) 24.6 (4.0) 22.0 (2.3) 0.081
Time spent watching TV{, minutes/day 120.0 (60.0, 180.0) 180.0 (120.0, 240.0) 120.0 (60.0, 180.0) 0.002
Tutorial time{, minutes/weekend 240.0 (225.0) 380.0 (91.6) 315.0 (159.3) 0.286
Data in the table are presented as number (%), mean (SD) and median (IQR), and comparison across groups was performed using Chi-
square test, Student t-test, and Mann–Whitney-U test for categorical (*), normally ({), and non-normally ({) distributed continuous variables
respectively. ** MetS = metabolic syndrome defined using the IDF definition. Obesity, overweight, and normal weight are defined as BMI
of ‡95th, ‡85th–<95th and <85th percentiles respectively.
 4 SUEBSAMRAN ET AL.

Table 3. Prevalence of the Metabolic Syndrome According to Different Classifications in Adolescents

Aged 13–16 Years by Sex, Family History of Diabetes, School Type, Categories of BMI
and Waist Circumference, Obesity, and Abnormalities in MetS Components (n = 393)
IDF Cook et al. De Ferranti Ford
2007 (3) 2003 (22) et al. 2004 (23) et al. 2005 (21) P-valuea
Number of students having MetS (percentage) 12 (3.1) 23 (5.8) 44 (11.2) 27 (6.9) <0.001
Sex <0.001
Male 9 (5.9) 16 (10.5) 24 (15.8) 18 (11.8)
Female 3 (1.2) 7 (2.9) 20 (8.3) 9 (3.7)
Family history of diabetes <0.001
Yes 4 (8.5) 5 (10.6) 10 (21.3) 7 (14.6)
No 8 (2.3) 18 (5.2) 34 (9.8) 20 (5.8)
Type of school <0.001
Sports school 2 (1.9) 3 (2.8) 4 (3.8) 3 (2.8)
Conventional school 10 (3.5) 20 (7.0) 40 (14.0) 24 (8.4)
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.

BMI tertiles <0.001

3rd tertile 11 (8.4) 22 (16.7) 40 (30.5) 24 (18.2)
2nd tertile 0 0 3 (2.2) 2 (1.5)
1st tertile 1 (0.8) 1 (0.8) 1 (0.8) 1 (0.8)
BMI categoryb <0.001
Obesity 9 (45.0) 12 (60.0) 17 (85.0) 12 (60.0)
Overweight 2 (5.3) 7 (17.9) 15 (39.5) 9 (23.1)
Normal weight 1 (0.3) 4 (1.2) 12 (3.6) 6 (1.8)
WC <0.001
3rd tertile 12 (8.1) 23 (15.5) 43 (29.1) 27 (18.2)
2nd tertile 0 0 1 (0.7) 0
1st tertile 0 0 0 0
Abnormalities in MetS components
TG 13 (3.3) 56 (13.7) 73 (17.8) 56 (13.7) <0.001
HDL-c 102 (25.6) 113 (28.4) 204 (51.3) 113 (28.4) <0.001
SBP 50 (12.2) 50 (12.2) 50 (12.2) 50 (12.2) 0.999
DBP 19 (4.6) 48 (11.7) 48 (11.7) 48 (11.7) <0.001
FBS 3 (0.8) 1 (0.2) 1 (0.2) 3 (0.8) 0.446
WC 64 (15.6) 64 (15.6) 140 (34.6) 64 (15.6) <0.001
Data in the table are presented as number (percentage).
a
P-value for comparison across different MetS definitions using Cochran Q test.
b
Obesity, overweight, and normal weight are defined as BMI of ‡95th, ‡85th–<95th and <85th percentiles respectively.
Mets, metabolic syndrome; HDL-c, high density lipoprotein cholesterol; BMI, body mass index.

conventional schools. Students from both types of schools Discussion

were similar regarding systolic blood pressure, breast
feeding for 6 months, and birth weight.
Regarding metabolic health, 35% of the school students
were at cardiometabolic risk (Table 4). Male students had a
higher prevalence of being metabolically unhealthy than their
female counterparts (prevalence of 41% and 30% respectively,
P = 0.034). The prevalence of being metabolically unhealthy
was 29%, 62%, and 85% in students with normal weight,
overweight, and obesity respectively (P-for-trend <0.001). The
prevalence of this metabolically unhealthy condition was 22%,
31%, and 51% in those in the lowest, middle, and highest
tertiles of BMI (P-for-trend <0.001).
Figure 1A–D show levels of health behaviors in students
with and without the MetS. Students with the MetS had a
similar level of physical activity to those without (P = 0.339)
(Fig. 1A). Those with MetS spent significantly longer time
watching TV than those without [median (IQR) time
watching TV 180.0 (120.0, 240.0) and 120.0 (60.0, 120.0)
min per day respectively, P = 0.002] (Fig. 1B). Those with
and without MetS were similar regarding dietary score and
tutorial time at a weekend (P = 0.081 and P = 0.286 respec-
tively) (Fig. 1C, D).
Based on this school-based survey, the prevalence of MetS
in school children of Ubon Ratchathani was modest and
varied greatly when different MetS classifications were ap-
plied. Male students studying in conventional schools who
had a family history of diabetes were at greater risk of the
MetS. The most prevalent components of MetS were low
levels of HDL cholesterol and increased waist circumference,
while the least prevalent component was hyperglycemia.
High prevalence of adolescents being at cardiometabolic risk
was observed and metabolically unhealthy normal weight was
prevalent in these school students. The only health-related
behavior that was positively associated with the risk of having
MetS in these school children was time spent watching TV.
Using different classifications is likely to result in varying
burdens of MetS. In this study, great variation in MetS
prevalence was observed when different classifications were
used and the lowest prevalence was observed when the IDF
classification was used. This is consistent with other previous
studies in children,4,12,13 which showed more than 10-time
difference in MetS prevalence when different classifications
were applied. Our study suggests generally low concordance
 THE METABOLIC SYNDROME IN SCHOOL CHILDREN 5

Table 4. Prevalence of Being Metabolically was 4.0%, which was far lower than our findings. However,
Unhealthy in Adolescents Aged 13–16 Years the investigators did not use all components of the classi-
by Sex, Family History of Diabetes, School Type, fication to define MetS. This might have led to underesti-
and Categories of BMI mation of MetS prevalence in their study and might explain
and Waist Circumference (n = 393) the difference in MetS prevalence between their study and
Metabolically unhealthy P-value our own.
Using the IDF 2007 classification, the prevalence of MetS
Overall 137 (34.9) in Vietnamese children aged 10–18 years was 4.6%,11
Sex 0.034 slightly higher than our findings when using the same
Male 63 (41.4) classification. This might be explained by different age
Female 74 (30.7) groups of study participants in the two studies. Also, in-
Family history of diabetes 0.003 clusion of sports school students in our study may lead to a
Yes 26 (55.3) low overall MetS prevalence. Of note, the difference in
No 110 (32.0) MetS prevalence observed between Thai and Vietnamese
Type of school <0.001 children and adolescents seemed to disappear when they
Sports school 16 (15.1) grew up. The prevalence of MetS in Thai adults was higher
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.

Conventional school 121 (42.2)
BMI <0.001
3rd tertiles 67 (50.8)
2nd tertiles 42 (31.1)
1st tertiles 28 (22.0)
BMI categorya <0.001
Obesity 17 (85.0)
Overweight 24 (61.5)
Normal weight 96 (28.7)
Waist circumference <0.001
3rd tertile 72 (48.6)
2nd tertile 44 (32.6)
1st tertile 21 (19.2)
Data in the table are presented as number (percentage), and
comparison across groups was performed using Chi-square test.
a
Obesity, overweight and normal weight are defined as BMI of
‡95th, ‡85th–<95th and <85th percentiles respectively.
between the de Ferranti et al. and IDF definitions, while the
highest concordance was found between the definitions by
Cook et al. and Ford et al. This makes it difficult to compare
MetS prevalence across populations and studies. Similar to
our study, the previous studies also found that high fasting
glucose was the least prevalent MetS component and in-
creased HDL cholesterol was the most prevalent component
regardless of the MetS definitions used. Such findings have
also been observed in adult populations.20–22
Previous studies largely examined the prevalence of MetS
in obese and overweight children,4–8 with a few studies ex-
amining MetS prevalence in normal weight children. As ex-
pected, studies in Thai obese children and adolescents showed
considerable burden of MetS, although the prevalence varied
greatly from 17%5,6 to 50%.7 A systematic review by Friend
et al.10 similarly showed that the overall prevalence of MetS
in children varied considerably, ranging from 0% to 19.2%
with a median of 3.3%, while the prevalence in overweight
and obese children was 11.9% (range 2.8%–29.3%) and
29.2% (range 10%–66%) respectively. However, the review
does not address the impact of using different MetS classifi-
cations on the burden of MetS. To our knowledge, no studies
have compared the prevalence of MetS in Asian children
when different classifications were applied.
Previous studies in Thai children within all weight ranges
reported relatively low MetS prevalence compared to our
study. Using De Ferranti’s classification, Rerksuppaphol
and Rerksuppaphol6 found that the prevalence of the MetS
than that of Vietnamese adults (23.2% vs. 18.5% respec-
tively).23,24
Generally, MetS and abnormalities in its components are
more prevalent in obese and overweight than normal weight
individuals and this is true for both general and central
obesity. Although school children in this study were fairly
lean, the trend toward higher MetS prevalence in those
within higher tertiles of BMI and waist circumference was
clearly observed. This is consistent with previous studies in
populations with higher average adiposity.4,12,13
Increasing interest has focused on normal weight indi-
viduals who have high cardiometabolic risk. Although BMI
within a normal range has been reported to be associated
with a reduced risk of cardiometabolic diseases and all-
cause mortality, not all individuals in this BMI range have
similarly low risk and metabolically unhealthy phenotypes
may be already present in normal weight people.19 Previous
studies suggest that normal weight but metabolically un-
healthy individuals were at a three-fold higher risk of all-
cause mortality and/or cardiovascular events than normal
weight individuals who are metabolically healthy.19 Meta-
bolically unhealthy normal weight is prevalent. It has been
reported that approximately 20% of normal weight adults
were metabolic unhealthy. Our study found that normal
weight children and adolescents had an even higher preva-
lence of being metabolically unhealthy (29%).
Sedentary behaviors and low levels of physical activity
have reportedly been associated with an increased risk of the
MetS.25 In our study, although no significant difference in
physical activity levels between those with and without
MetS, the sports school students were at a lower risk of
MetS than those in a conventional school. This may be
explained by that the sports school’s curriculum requires
many more hours of physical education than that of con-
ventional schools, or simply the explicit difference in stu-
dent recruitment. No previous studies have compared MetS
prevalence between different types of schools. Moreover,
longer time spent watching TV was associated with an in-
creased risk of having MetS. This is similar to previous
studies in adult populations26,27 and children.28,29
There remains uncertainty about which MetS classifica-
tion is preferred for determining and monitoring trends in
MetS over time. This may not be solved by evidence on
MetS prevalence. Choices of classifications should rather be
based on the ability of MetS in children using different
classifications to predict diabetes and cardiovascular disease
in later life.3 As time spent watching TV was associated
 6 SUEBSAMRAN ET AL.
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.

FIG. 1. Boxplots comparing levels of health behaviors (A) physical activity, (B) time spent watching TV, (C) dietary
score, and (D) tutorial time at weekend between school children with and without metabolic syndrome in the UMeSIA
project, P-values from Student t-test or Mann–Whitney-U test.

with MetS in many studies, including our own, public health
interventions to optimize TV watching hours should be
encouraged. Furthermore, longitudinal follow-up and ef-
fective surveillance of the MetS and related health behaviors
should also be put in place to better monitor impact of
public health interventions given.
This study was among a few studies that compared MetS
prevalence in Asian children and adolescents when using
different MetS classifications, and it was the first to report
the MetS prevalence in students from different types of
schools, for example, sports versus conventional secondary
schools. All MetS components were assessed using standard
equipment and procedures. However, our study had a
number of limitations. Health-related behaviors, particularly
diet and physical activity, were assessed using self-report
questionnaire. This might have led to over- or underesti-
mation of their behaviors; for example, people are likely to
report their health behaviors better than they actually are.
Consequently, this might have altered their association with
MetS. However, GPAQ is widely used and may be suitable
for large-scale epidemiologic studies. In addition, a simple
diet questionnaire was tested before use in this study and its
validity and reliability were acceptable. Lastly, further
studies are needed to account for several other factors that
may be associated with the presence or development of
MetS in children and adolescents, such as early life expo-
sures and psychosocial factors.
The overall prevalence of MetS in Thai school children
was modest and varied greatly when different classifications
were applied. MetS prevalence was different by type of
schools and time spent watching TV. This warrants further
research on interventions addressing sedentary behaviors in
school children. Longitudinal studies to examine which
classification best predicts the future risk of diabetes and
cardiovascular disease in midlife are also needed
Acknowledgments
The authors thank Narinukun School, Lukhamhan School,
Ubon Ratchathani Sports School, and Avemaria School for
their participation in this study. We are grateful to the lab
technician team of the College of Medicine and Public
Health, Ubon Ratchathani University for their assistance in
laboratory analysis.
 THE METABOLIC SYNDROME IN SCHOOL CHILDREN
Author Disclosure Statement
No competing financial interests exist.
References
1. Oda E. Metabolic syndrome: Its history, mechanisms, and
limitations. Acta Diabetol 2012;49:89–95.
2. Mulè G, Calcaterra I, Nardi E, et al. Metabolic syndrome in
hypertensive patients: An unholy alliance. World J Cardiol
2014;6:890–907.
3. Alberti KG, Zimmet P, Shaw J. The metabolic syndrome-a
new worldwide definition. Lancet (London, England) 2005;
366:1059–1062.
4. Braga-Tavares H, Fonseca H. Prevalence of metabolic
syndrome in a Portuguese obese adolescent population
according to three different definitions. Eur J Pediatr 2010;
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.
169:935–940.
5. Iamopas O, Chongviriyaphan N, Suthutvoravut U. Meta-
bolic syndrome in obese Thai children and adolescents. J
Med Assoc Thailand 2011;94(Suppl 3):S126–S132.
6. Rerksuppaphol L, Rerksuppaphol S. Prevalence of Meta-
bolic Syndrome in Thai Children: A Cross-sectional Study.
J Clin Diagn Res 2014;8:PC04-PC7.
7. Rerksuppaphol S, Rerksuppaphol L. Metabolic Syndrome
in Obese Thai Children: Defined Using Modified ‘The
National Cholesterol Education Program/Adult Treatment
Panel III’ Criteria. J Med Assoc Thai 2015;98(Suppl 10):
S88–S95.
8. Panamonta O, Thamsiri N, Panamonta M. Prevalence of
type II diabetes and metabolic syndrome among overweight
school children in Khon Kaen, Thailand. J Med Assoc Thai
2010;93:56–60.
9. Chung JY, Kang HT, Shin YH, et al. Prevalence of meta-
bolic syndrome in children and adolescents—the recent
trends in South Korea. J Pediatr Endocrinol Metab 2013;
26:105–110.
10. Friend A, Craig L, Turner S. The prevalence of metabolic
syndrome in children: A systematic review of the literature.
Metab Syndr Relat Disord 2012;11:71–80.
11. Nguyen TH, Tang HK, Kelly P, et al. Association between
physical activity and metabolic syndrome: A cross sec-
tional survey in adolescents in Ho Chi Minh City, Vietnam.
BMC Public Health 2010;10:141.
12. Agudelo GM, Bedoya G, Estrada A, et al. Variations in the
prevalence of metabolic syndrome in adolescents according
to different criteria used for diagnosis: Which definition
should be chosen for this age group? Metab Syndr Relat
Disord 2014;12:202–209.
13. Suárez-Ortegón MF, Ramı́rez-Vélez R, Mosquera M, et al.
Prevalence of metabolic syndrome in urban colombian
adolescents aged 10–16 years using three different pediatric
definitions. J Trop Pediatr 2013;59:145–149.
14. Bull FC, Maslin TS, Armstrong T. Global physical activity
questionnaire (GPAQ): Nine country reliability and validity
study. J Phys Activity Health 2009;6:790–804.
15. Surveillance and Population-Based Prevention, Prevention
of Noncommunicable Diseases Department, World Health
Organization. Global Physical Activity Questionnaire
(GPAQ) Analysis Guide. Geneva, World Health Organi-
zation.
16. Buranakitjaroen P, Sitthisook S, Wataganara T, Opha-
scharoensuk V, et al. Thai Guidelines on The Treatment of
7
Hypertension 2012 Update 2015. Bangkok: Thai Hy-
pertension Society; 2015: 1–3.
17. Ross R, Berentzen T, Bradshaw AJ, Janssen I, et al. Does
the relationship between waist circumference, morbidity
and mortality depend on measurement protocol for waist
circumference? Obes Rev 2008;9:312–325.
18. Barlow SE. Expert committee recommendations regarding
the prevention, assessment, and treatment of child and ad-
olescent overweight and obesity: Summary report. Pedia-
trics 2007;120(Suppl 4):S164–S192.
19. Stefan N, Schick F, Haring HU. Causes, characteristics, and
consequences of metabolically unhealthy normal weight in
humans. Cell Metab 2017;26:292–300.
20. Ford ES, Giles WH. A comparison of the prevalence of the
metabolic syndrome using two proposed definitions. Dia-
betes Care 2003;26:575–581.
21. Bhowmik B, Afsana F, Siddiquee T, et al. Comparison of
the prevalence of metabolic syndrome and its association
with diabetes and cardiovascular disease in the rural pop-
ulation of Bangladesh using the modified National Cho-
lesterol Education Program Expert Panel Adult Treatment
Panel III and International Diabetes Federation definitions.
J Diabetes Invest 2015;6:280–288.
22. Qiao Q. Comparison of three different definitions for the
metabolic syndrome in non-diabetic Europeans. Br J Dia-
betes Vasc Dis 2005;5:161–168.
23. Aekplakorn W, Kessomboon P, Sangthong R, et al. Urban
and rural variation in clustering of metabolic syndrome
components in the Thai population: Results from the fourth
National Health Examination Survey 2009. BMC Public
Health 2011;11:854.
24. Son le NT, Kunii D, Hung NT, et al. The metabolic syn-
drome: Prevalence and risk factors in the urban population
of Ho Chi Minh City. Diabetes Res Clin Pract 2005;67:
243–250.
25. Oliveira RG, Guedes DP. Physical Activity, Sedentary
Behavior, Cardiorespiratory Fitness and Metabolic Syn-
drome in Adolescents: Systematic Review and Meta-
Analysis of Observational Evidence. PLoS One 2016;11:
e0168503.
26. Vioque J, Torres A, Quiles J. Time spent watching televi-
sion, sleep duration and obesity in adults living in Valencia,
Spain. Int J Obes Relat Metab Disord 2000;24:1683–1688.
27. Aadahl M, Kjær M, Jørgensen T. Influence of time spent
on TV viewing and vigorous intensity physical activity
on cardiovascular biomarkers. The Inter 99 study. Eur J
Cardiovasc Prev Rehabil 2007;14:660–665.
28. Mark AE, Janssen I. Relationship between screen time and
metabolic syndrome in adolescents. J Public Health 2008;
30:153–160.
29. Ford ES, Ajani UA, Mokdad AH. The Metabolic Syndrome
and Concentrations of C-Reactive Protein Among U.S.
Youth. Diabetes Care 2005;28:878–881.
Address correspondence to:
Phalakorn Suebsamran, MS
College of Medicine and Public Health
Ubon Ratchathani University,
Ubon Ratchathai 34190
Thailand
E-mail: palakorn.s@ubu.ac.th
(Appendix follows/)
 8 SUEBSAMRAN ET AL.

Appendix Table 1. Comparison of Characteristics of Students from the Sports

and Conventional Schools in the UMeSIA Project
Characteristics Sports school (n = 107) Conventional school (n = 286) P-value
Male sex* 79 (73.8) 79 (26.2) <0.001
Age{, years 14.1 (0.5) 15.1 (0.9) <0.001
BMI{, kg/m2 19.2 (2.3) 20.9 (3.9) <0.001
Waist circumference{, cm 69.5 (12.1) 74.3 (15.2) <0.001
HDL cholesterol{, mg/dL 54.3 (8.7) 38.7 (8.5) <0.001
Triglyceride{, mg/dL 69.6 (32.5) 78.2 (26.8) <0.001
Systolic blood pressure{, mmHg 108.3 (14.5) 119.9 (23.5) 0.202
Diastolic blood pressure{, mmHg 55.7 (14.5) 69.0 (11.7) 0.014
Fasting blood sugar{, mg/dL 83.6 (2.5) 84.5 (4.6) 0.002
Breast feeding for 6 months* 62 (58.5) 167 (56.6) 0.098
Birth weight{, grams 2826.2 (380.7) 2948.8 (525.2) 0.177
Family history of diabetes mellitus* 8 (7.5) 39 (13.7) 0.097
metabolic equivalent{, MET minutes/week 7800 (4920–12,160) 2400 (1060–4860) <0.001
Downloaded by Karolinska Institutet University Library from www.liebertpub.com at 08/27/18. For personal use only.

Dietary Score{ 20.5 (2.5) 21.7 (2.4) <0.001

Time spent watching TV{, minutes/day 60.0 (60.0–112.5) 120.0 (60.0–180.0) <0.001
Data in the table are presented as number (%), mean (SD) and median (IQR), and comparison across groups was performed using Chi-
square test, Student t-test, and Mann–Whitney-U test for categorical (*), normally({), and non-normally({) distributed continuous variables
respectively.