Preventive Medicine 67 (2014) 100–105

Contents lists available at ScienceDirect

Preventive Medicine

j ourna l homepage: www.e lsev ie r .com/ locate /ypmed

Association between self-reported physical activity and indicators ofbody composition in Malaysian adolescents

Tin Tin Su a,⁎, Pei Ying Sim a, Azmi Mohamed Nahar b, Hazreen Abd Majid a, Liam J. Murray c,Marie M. Cantwell c, Nabilla Al-Sadat a, Muhammad Yazid Jalaludin d

a Centre for Population Health (CePH), Department of Social and Preventive Medicine, Faculty of Medicine, University of Malaya, Malaysiab Department of Sports Medicine, Faculty of Medicine, University of Malaya, Malaysiac Centre for Public Health, School of Medicine, Dentistry and Biomedical Sciences, Queen's University of Belfast, Belfast, UKd Department of Paediatrics, Faculty of Medicine, University of Malaya, Malaysia

⁎ Corresponding author. Fax: +60 3 7967 4975.E-mail address: tintinsu03@yahoo.com (T.T. Su).

http://dx.doi.org/10.1016/j.ypmed.2014.07.0010091-7435/© 2014 Published by Elsevier Inc.

a b s t r a c t

a r t i c l e i n f o

Available online 10 July 2014

Keywords:ObesityAdolescentsWaist circumferenceBody fatnessPhysical activityPAQ-CMalaysia

Background: Obesity and lack of physical activity are fast becoming a concern among Malaysian adolescents.Objective: This study aims to assess physical activity levels among Malaysian adolescents and investigate the

association between physical activity levels and body composition such as bodymass index (BMI), waist circum-ference (WC) and percentage of body fat.

Subjects and methods: 1361 school-going 13 year old multi-ethnic adolescents from population representa-tive samples in Malaysia were involved in our study. Self-reported physical activity levels were assessed usingthe validated Malay version of the Physical Activity Questionnaire for Older Children (PAQ-C). Height, weight,body fat composition and waist circumference (WC) were measured. Data collection period was from Marchto May 2012.

Results: 10.8% of the males and 7.4% of the females were obese according to the International Obesity TaskForce standards. A majority of the adolescents (63.9%) were physically inactive. There is a weak but significantcorrelation between physical activity scores and the indicators of obesity. The adjusted coefficient for bodyfatness was relatively more closely correlated to physical activity scores followed by waist circumference andlastly BMI.

Conclusion: This study demonstrates that high physical activity scores were associated with the decreasedprecursor risk factors of obesity.

© 2014 Published by Elsevier Inc.

Introduction

The prevalence of cardiovascular disease (CVD) risk in theMalaysianpopulation is alarmingly high andobesity is a known risk factor (Instituteof Public Health, 2008). Physical activity has been shown to play an im-portant role in preventing cardiovascular diseases among Malaysianadults (Dhanjal et al., 2001).

Childhood obesity is now becoming a growing concern in Malaysia(Bauman et al., 2011; Rampal et al., 2007). Previous studies haveshown an increasing trend of overweight from 20.7% in 2002 to 26.5%in 2008 among 6–12 year-old children in Peninsular Malaysia (Ismailet al., 2009) and a higher prevalence (34.2%) of overweight and obesechildren in metropolitan Kuala Lumpur (Wee et al., 2011). Body fatnessis linked to the increased prevalence of overweight and obesity in ado-lescents contributing to the risk of metabolic syndrome in childhood

which may exacerbate and cause cardiovascular diseases in adulthood(Ali et al., 2014).

In addition, 21% of secondary school adolescents, aged 14 to16 years, of the Petalingdistrict in Selangor statewere found to bephys-ically inactive (Aniza and Fairuz, 2009). Easy access to the internet anduse of digital gadgets at an early age have resulted in a sedentary life-style which in turn has decreased physical activity levels even furtheramong adolescents (Lau et al., 2013). This information highlights theneed to determine the prevalence of obesity and to explore the relation-ship between reported physical activity and body fatness amongMalaysian adolescents.

Previous studies which have investigated the association betweenphysical activity and obesity in Malaysia focused on one geographicalarea, either one school or district and used Body Mass Index (BMI) asa parameter of obesity (Dan et al., 2011; Rezali et al., 2012). However,recent studies showed that Waist Circumference (WC) (McCarthyet al., 2003) and percentage of body fat are more sensitive parametersof obesity compared to BMI (Deurenberg‐Yap et al., 2000, 2002). Asianpopulations are more likely to have a higher percentage of abdominal

Table 1Socio-demographic characteristics of Malaysian adolescents (N = 1327).

Socio-demographic characteristics Mean ± SD or N (%)

Age (years) 12.9 ± 0.3Gender Male 508 (38.3)

Female 819 (61.7)Ethnicity Malay 1086 (81.8)

Chinese 100 (7.5)Indians 100 (7.5)Others 41 (3.1)

Place of residence Urban 702 (52.9)Rural 625 (47.1)

101T.T. Su et al. / Preventive Medicine 67 (2014) 100–105

fat for any given BMI compared to Caucasians (Boyko et al., 2000). Fur-thermore, Malaysian children and adolescents have also been shown tohave the largestWC compared to their peers in theUK, Australia, Turkeyand Hong Kong (Poh et al., 2011).

In order to fill the current knowledge gap, this study aims to assessself-reported physical activity levels among Malaysian adolescents andalso investigate the possible association between self-reported physicalactivity levels and indicators of body composition such as BMI, WC, andpercentage of body fat using data from a population representativesample.

Methods

Study design

Data used in this studywas from thefirstwave of an adolescent cohort studyconducted in Peninsular Malaysia. The cohort included 1361 adolescents andthe study was supported by the University Malaya Research Grant. The detailof the study was reported previously (Abu Hanifah et al., 2013).

Study area and duration of the study

The study was conducted in two states and onemetropolitan area of Penin-sular Malaysia. The state of Perak from the northern region and Selangor statefrom the central region of Peninsular Malaysia were chosen purposively basedon discussion with the Ministry of Education. The Federal territory of KualaLumpur was included as the metropolitan area. The data collection for thefirst wave cohort study was carried out between March and May 2012.Malaysia has an equatorial climate and temperature is more or less consistentthroughout the year. The choice of data collection time does not affect thestudy findings.

Study population

The study populationwere school children, aged 13 yearswhowere attend-ing the first year of governments' Secondary school (Secondary One). The inclu-sion criteria were that students must be able to read and write in Malay (thenational language of Malaysia).

Sampling procedure

A two stage cluster sampling method was applied. In the first stage, 15schools were randomly selected from all government secondary schoolsbelonging to two states and the federal territory of Kuala Lumpur. The samplingframe was constructed according to the list of secondary schools provided bythe Ministry of Education. Schools were selected using a computer generatedrandomnumber table and each school had the sameprobability of being selected.At the second stage, all students attending Secondary One from selectedschools were invited to participate in the study. Out of 2694 students whoreceived the invitation letter, 1361 of themparticipated in the study, an averageresponse rate of 51%. There were no significant differences in terms of socio-demographic characteristics (age, gender and ethnicity) of respondents andnon-respondents.

Measurements

Anthropometric measurementHeight was measured without socks and shoes using a calibrated vertical

Seca Portable 217 Stadiometer, to the nearestmillimetre.Weightwasmeasuredwith light clothing using a Seca 813 digital electronic weighing scale, to thenearest decimal fraction of a kilogramme. BMI was calculated as weight inkilogrammes divided by the square of height in metres. Body fat compositionwas estimated with bioelectrical impedance using a Tanita portable scale (SC-240, Body Composition Analyser, Tanita Europe B.V., The Netherlands). TheBody Composition Analyser SC-240 has acceptable accuracy compared to thedual-energy X-ray absorptiometry in white and African-American adolescents(Barreira et al., 2013). WC was measured with a non-elastic Seca measuringtape (Seca 201, Seca, UK), to the nearest millimetre. WC measurement wasdone at the natural waist which is the midway between the lowest rib margin(tenth rib) and highest point of the iliac crest with the tape around the body

in horizontal position (World Health Organization, 2008). All measurementswere done by trained research assistants.

Assessment of self-reported physical activitySelf-reported physical activity levels were assessed using the validated

Malay version of the Physical Activity Questionnaire for Older Children (PAQ-C) which has good internal consistency and acceptable validity (Crocker et al.,1997; Dan et al., 2007; Dan et al., 2011). There were 10 items in the PAQ-Cwhich captured the level of physical activity in the last 7 days. The first item in-cluded the type and frequency of sports or/and dance the adolescents did duringthe past 7 days. The 2nd to 8th items in the questionnaire assessed the activityof the adolescents during physical education (PE) classes, recess, lunch time,right after school, evenings, weekend and leisure time. The answers to items 2to 8 used a five-point Likert scale [1 (lowest) to 5 (highest)]. Item 9 includedthe frequency of participating in daily physical activity in the previous week.Item 10 asked the adolescents to report any unusual activities during the previ-ous week which have not already been recorded. The mean self-reported phys-ical activity score was further categorised into low (score b 2.33), moderate(2.33–3.66) and high (score N 3.66). We followed the categorization of a previ-ous study which used the PAQ-C (Crocker et al., 1997).

Ethical approval

This study was approved by the Ethics committee, University Malaya Medi-cal Centre (Ref. no. 896.34). Participation in the studywas voluntary andwritteninformed consent was obtained from the parents or guardian as well as theparticipants.

Statistical analysis

All data were analysed using IBM SPSS Statistics version 21. Descriptive andbivariate analyseswere done as preliminary data analysis. Continuous variableswere presented as means with 95% confidence intervals. The associationsbetween self-reported physical activity scores and gender, ethnicity and placeof residence was analysed using ANOVA. The indicators of obesity such asBMI, WC and body fatness were also examined in terms of their associationwith gender, ethnicity and place of residence using ANOVA. Lastly, multiplelinear regressions were applied to determine the association between self-reported physical activity scores and BMI, WC and body fatness. Crude andadjusted analyses (age, gender, ethnicity and place of residence) were carriedout for each indicator of obesity and presented as standardized regressioncoefficients.

Results

Among 1361 respondents, 1327 adolescents were included in thefinal analysis. The socio-demographic characteristics of adolescents arepresented in Table 1. Of the sample, 508 were males and 819 femaleswith the mean age of 12.9 ± 0.3 years. Using IOTF standards to catego-rise adolescents according to BMI, results (Table 2) showed that in total,8.7% of the adolescents were obese (10.8% of males and 7.4% of females)and a further 15.9% (16.1% of males and 15.8% of females) were over-weight. 53% of males and 54.5% of females were within the normalBMI range.

Results showed that 17% of males and 15.6% of females have a WCgreater than the 90th percentile.

Table 2BMI of adolescents using the International Obesity Task Force (IOTF) standard classification (N = 1327).

BMI classification BMI range for

Males Females Total N (%)

BMI range N (%) BMI range N (%)

Thinness grade 3 BMI b13.59 7 (1.4) b13.92 16 (2.0) 23 (1.7)Thinness grade 2 BMI 13.59 to 14.47 23 (4.5) 13.92 to 14.84 41 (5.0) 64 (4.8)Thinness grade 1 BMI 14.48 to 15.83 72 (14.2) 14.85 to 16.25 127 (15.5) 199 (15.0)Normal 15.84 to 21.90 269 (53.0) 16.26 to 22.57 446 (54.5) 715 (53.9)Overweight 21.91 to 26.83 82 (16.1) 22.58 to 27.75 128 (15.6) 210 (15.8)Obesity ≥26.84 55 (10.8) ≥27.76 61 (7.4) 116 (8.7)

BMI = Body mass index.

102 T.T. Su et al. / Preventive Medicine 67 (2014) 100–105

Most adolescents were in the low physical activity category (63.9%),followed by the moderate (34.0%) and high physical activity levels(1.9%). In school, during physical education (PE) classes, 37.8% some-times and 21.2% always participated in physical activities. Howeverduring recess and lunch time, most adolescents had low physical activ-ity levels (e.g.: sitting down, talking and reading). In the evenings, 36.5%were active for 2–3 evenings/week, 26.4% were active on only one eve-ning while 16.8% were not involved in any activity in the evening. Sim-ilarly on weekends, 33.2% were active 1–3 times/week, 16.4% were notactive at all and only 8.2% and 9% were active for 4 and 5 times/weekrespectively (Table 3).

Gender differences were observed in physical activity scores withmales scoring 18% higher at 2.46 (95% CI: 2.29, 2.64) compared tofemales 2.02 (95% CI: 1.91, 2.12). Self-reported physical activity scoreswere also statistically significantly different among different ethnicpopulations (P b 0.05). Themost active ethnic groupwas others (a com-bination group of all minority ethnicities such as indigenous populationgroups and non-Malaysians) who scored 2.50 (95% CI: 2.31, 2.68). TheIndian adolescent scores were slightly lower at 2.31 (95% CI: 2.03,2.59) followed by the Malays at 2.21 (95% CI: 2.18, 2.24). The Chineseadolescents were the least active at 1.92 (95% CI: 1.72, 2.17). No

Table 3Level of physical activity in adolescents (N = 1327).

Physical activity (PA) N (%)

Last 7 days during physicaleducation class

Never 49 (3.7)Hardly ever 238 (17.9)Sometimes 501 (37.8)Quite often 258 (19.4)Always 281 (21.2)

Last 7 days during recess Sat down (talking, reading) 651 (49.1)Stood around or walked around 476 (35.9)Ran or played a little bit 119 (9.0)Ran around or played quite a bit 43 (3.2)Ran or played hard most of thetime

38 (2.9)

Last 7 days during lunch timebesides eating lunch

Sat down (talking, reading) 783 (59.1)Stood around or walked around 356 (26.8)Ran or played a little bit 131 (9.9)Ran around or played quite a bit 23 (1.7)Ran or played hard most of thetime

34 (2.6)

Last 7 days, number of evenings None 223 (16.8)1 time last week 350 (26.4)2 or 3 times last week 484 (36.5)4 times last week 119 (9.0)5 times or more last week 151 (11.4)

Last weekend number of times None 217 (16.4)1 time last week 441 (33.2)2 or 3 times last week 441 (33.2)4 times last week 109 (8.2)5 times or more last week 119 (9.0)

Level of physical activity Low PA (score b 2.33) 858 (63.9)Moderate PA (score 2.33–3.66) 456 (34.0)High PA (score N 3.66) 26 (1.9)

significant differences in physical activity scores were observed be-tween adolescents who were in rural or urban residences (Table 4).

Gender differences were seen for both WC and percentage of bodyfat but not for BMI. Body fatness in females was 27% significantly higherthan males (P b 0.001) whereas WC was 3% higher in males comparedto females (P b 0.05). WC was also significantly higher in Chinese(71.4 cm) and Indians (71.2 cm) compared to Malays (68.5 cm) (Pb 0.05). Furthermore, adolescents living in urban areas had higher WCand body fatness by 2% and 7% respectively, compared to those livingin rural areas. Differences in BMI were not significant between gender,ethnicity and place of residence (Table 5). Abdominal obesity wasassessed using the cut off points, N90th percentile, 83.8 cm for boysand 78.8 cm for girls (Poh et al., 2011). The prevalence of abdominalobesity among male and female adolescents was 17% and 15.6%respectively.

Crude and adjusted regression coefficients were calculated for theassociation between physical activity scores and indicators of obesity.The adjusted values were controlled for age, gender, ethnicity andplace of residence. Therewas aweakbut significant correlation betweenself-reported physical activity scores and indicators of obesity (P b 0.05).Table 6 includes the adjusted coefficients and shows that body fatnesswas most closely correlated with self-reported physical activity scoresfollowed by WC and lastly BMI.

Discussion

This paper focuses on the prevalence of obesity among Malaysianadolescents and whether their self-reported levels of physical activityare reflected in their body compositions.

Physical activity is a complex construct to assess and is influenced byvarious factors. Previous studies have demonstrated the positive effectsof physical activity intervention programmes in schools resulting in thedecreased prevalence of childhood obesity (Alberga et al., 2013; Fedewaet al., 2013; Morrow et al., 2013; Sigmund and Sigmundova, 2013).Furthermore, effective strategies to increase moderate-to-vigorousphysical activity on a daily basis in schools during recess time havebeen linked to reduced rates of obesity and enhanced student's healthand academic results (Efrat, 2013; Parrish et al., 2013). Such anapproach may be useful to encourage the 85% of adolescents in this

Table 4Socio-demographic differences in physical activity scores in adolescents according to gen-der, ethnicity and place of residence (N = 1327).

Physical activity scores ANOVA

Gender Female 2.02 (1.91,2.12) P b 0.001Male 2.46 (2.29,2.64)

Ethnicity Malay 2.21 (2.18,2.24) P b 0.05Chinese 1.92 (1.72,2.17)Indian 2.31 (2.03,2.59)Others 2.50 (2.31,2.68)

Place of residence Rural 2.14 (1.95,2.32) P = 0.32Urban 2.34 (2.25,2.43)

Table 5Socio-demographic differences in BMI, waist circumference and body fat in adolescents according to gender, ethnicity and place of residence (N = 1327).

BMI Waist circumference Body fat

(kg/m2) (cm) (%)

Gender Female 20.0 (19.0,21.0) 67.8 (67.1,68.5) 25.8 (25.1,26.6)Male 18.7 (17.1,20.2) 70.9 (69.7,72.0) 18.9 (17.7,20.2)P value P = 0.15 P b 0.001 P b 0.001

Ethnicity Malay 19.9 (19.6,20.3) 68.5 (67.8,69.2) 23.1 (22.3,23.9)Chinese 20.2 (18.1,22.2) 71.4 (69.4,73.4) 24.1 (21.8, 26.4)Indian 18.1 (15.6,20.7) 71.2 (69.1,73.3) 24.3 (21.7, 26.8)Others 19.1 (17.4,20.8) 69.6 (66.5,72.7) 21.0 (17.2, 24.7)P value P = 0.40 P b 0.05 P = 0.93

Place of residence Rural 18.7 (17.0,20.3) 68.2 (67.3, 69.2) 22.3 (21.3,23.3)Urban 20.0 (19.2,20.8) 69.7 (68.8, 70.5) 24.0 (23.1,24.9)P value P = 0.16 P b 0.05 P b 0.05

103T.T. Su et al. / Preventive Medicine 67 (2014) 100–105

study who were inactive during recess and lunch time to participateand improve their physical activity levels. Our findings have shownthat only 2% of Malaysian adolescents are involved in high levels ofphysical activity and 64% have low physical activity levels.

Over the past decade, rapid urbanization has caused lifestyle chang-es in adolescents, as they adapt to a more sedentary lifestyle and inevi-tably low physical activity (Dan et al., 2011; Lau et al., 2013). Sedentarybehaviours, such as watching television and playing electronic gamestogether with low physical activity, in adolescents are associatedwith overweight and obesity (Lau et al., 2013). In countries such asAustralia (Eime et al., 2013), New Zealand (Stoner et al., 2013), Iran(Hajian-Tilaki and Heidari, 2012), China (Ying-Xiu et al., 2013) andSweden (Teder et al., 2013), steps have been taken to modify the life-style of adolescents by introducing structured programmes to increasephysical activity levels and reduce sedentary activities not only inschools but also at home. This would help to prevent long term compli-cations of overweight and obesity in adolescents.

It was observed that, adolescents living in rural areas are less likelyto be obese compared to those living in urban areas due to environmen-tal differences (Ghosh, 2011; Hodgkin et al., 2010; Mohan et al., 2004).Similarly, our findings also showed a significantly higher WC and bodyfat percentage in adolescents living in urban areas compared to ruraladolescents. On the other hand, most studies carried out in Westerncountries showed that adolescents in rural areas are at higher risk ofoverweight and obesity compared to urban areas (Bertoncello et al.,2008; Davis et al., 2011) attributed to differences in social and economicenvironments across countries (Wang et al., 2002). Further studies areneeded to determine the differences in physical activity between ruraland urban Malaysian adolescents and how the level of urbanization af-fects their level of physical activity.

The overall outcome of this study showed that self-reported physicalactivity scores of adolescents were inversely correlated with indicatorsof obesity such as BMI, WC and body fatness. The higher the selfreported physical activity scores, the lower the risk factors but dueto confounding factors such as lifestyle behaviours, dietary intakesand genetic influences, the effects are significant but have weak cor-relation. Studies have revealed inconsistent results across ethnic

Table 6Standardized regression coefficients of the association between physical activity and riskfactors of obesity in adolescents (N = 1327).

Physical activity scores

Crude Adjusted

BMI (kg/m2) −0.055⁎ −0.058⁎

Waist circumference (cm) −0.018 −0.069⁎

Body fat (%) −0.164⁎⁎ −0.088⁎

Adjusted values are controlled for age, gender, ethnicity and place of residence.⁎ P b 0.05.⁎⁎ P b 0.001.

groups in Malaysia with regard to body composition such as WC(Poh et al., 2011) and BMI (Dunn et al., 2012; Tan et al., 2012). Inour findings, the Chinese and Indian populations are more likely tobe overweight and obese compared to the majority Malay popula-tion and concurrently the Malays had higher self-reported physicalactivity scores compared to the Chinese. This was in line with astudy by Tan et al. (2012) but in contrast to a study by Rezali et al.(2012) showing that the prevalence of obesity is highest amongMalays (56.0%) followed by Chinese (30.1%) and Indian (13.5%) ado-lescents. Nevertheless, body fatness has been reported to play amoreimportant role in differentiating normal from overweight and obeseadolescents compared to BMI (Goonasegaran et al., 2012). However,it is important to note that body fatness is strongly confounded by ethnic-ity, particularly among the Chinese,Malays and Indians (Deurenberg-Yapet al., 2000).

In addition, WC in adolescents is well-defined as one of the riskfactors formetabolic syndromeby the International Diabetes Federation(IDF-reference), and aWC N90th percentile indicates abdominal obesityin adolescents. According to Poh et al. (2011) a WC at or greater thanthe 90th percentile based on the WHO BMI-for-age-reference (WHO-reference) defines abdominal obesity in adolescents of Malaysia; thesecut-offs corresponding to 83.8 cm for males and 78.8 cm for females.Our findings showed that 17% of males and 15.6% of females have aWC greater than the 90th percentile. The WC of overweight obese ado-lescents was recorded on average as 93.4±7.7 cm formales and 86.9±6.4 cm for females. Interestingly, males showed a significantly higherWC despite being more active than the females and this may indicatean over-reporting of physical activity in males.

Furthermore, the definition of childhood obesity derived from theInternational Obesity Task Force (IOTF) reference (Cole and Lobstein,2012) in association with various factors was used because it wasshown to be most appropriate in the Asian population among adoles-cents (Viner et al., 2010). However, there are still conflicting results inthe assessment of BMI in the international system such as The Interna-tional Obesity Task Force (Cole and Lobstein, 2012), the World HealthOrganization (De Onis et al., 2007) and the Centres for Disease Controland Prevention (Kuczmarski et al., 2002)which, requires further valida-tion for a more consistent and effective approach in public health andclinical applications (Gonzalez-Casanova et al., 2013).

In conclusion, percentage body fatnesswas themost important indi-cator of obesity followed by WC and lastly BMI in Malaysian adoles-cents. This study also demonstrates that self-reported high physicalactivity scores were inversely associated with precursor risk factors ofobesity. These results may imply a need for the public health sector todevelop more effective measures in schools to counteract the growingepidemic of overweight and obesity in Malaysia. Physical activity rec-ommendations should be implemented in schools to encourage adoles-cents to lead a healthier lifestyle and possibly develop into a lifetimehabit. This message should also be translated to parents in order to pro-mote physical activity at homes as a public health key message.

104 T.T. Su et al. / Preventive Medicine 67 (2014) 100–105

Prospective studies are needed to investigate towhat extent; high levelsof physical activity can reverse the conditions of overweight and obesityin adolescents.

Limitations

Amajor limitation of this studywas the assessment of 7-dayphysicalactivity derived from self-reports using the standardized Physical Activ-ity Questionnaire for Adolescents. Although this may affect the validityof the physical activity levels and scores, a study has shown that thePAQ-C has good internal consistency, and has acceptable validity (Janzet al., 2008). In addition, PAQ-C cannotmeasure the energy expenditureby physical activity. However the focus of our research was mainly onphysical activity level, thus this limitation could not affect our studyfindings. The Tanita portable bioelectrical impedance device, which isvery applicable for mass population screening, was not validated forMalaysian children.We recommend a future study to validate the Tanitaportable scale with the gold standard one.

Furthermore, the cut-off points used to define overweight and obe-sity using WC and BMI measures may underestimate the prevalencein this population, as observed when comparing with previous studieson the estimation of prevalence. It is clear that the variety of baselinemeasurements is not only affected by the diverse population but alsothe environment the adolescents live in, resulting in the weak correla-tion. Despite these limitations, our findings still showed significantassociations between self-reported physical activity scores and the indi-cators of body composition in the predicted directions.

Author's contribution

Thefirst and second authors carried out the data analysis. All authorscontributed in the design of the study, data collection and write-up ofmanuscript.

Conflict of interests

The authors declare that there are no conflicts of interests. Theauthors alone are responsible for the content and writing of the paper.

Acknowledgments

The study was funded by the University of Malaya Research Grant:RG299-11HTM and Vice Chancellor Research Grant: UMQUB3D-2011.The post-doctoral research fellow position for this project was jointlyfunded by the University of Malaya and Queen's University of Belfast.

References

Abu Hanifah, R., Mohamed, M.N.A., Jaafar, Z., et al., 2013. The correlates of body composi-tion with heart rate recovery after step test: an exploratory study of Malaysian ado-lescents. PLoS ONE 8 (12), e82893. http://dx.doi.org/10.1371/journal.pone.0082893.

Alberga, A.S., Medd, E.R., Adamo, K.B., et al., 2013. Top 10 practical lessons learned fromphysical activity interventions in overweight and obese children and adolescents.Appl. Physiol. Nutr. Metab. 38, 249–258.

Ali, O., Cerjak, D., Kent Jr., J.W., et al., 2014. Obesity, central adiposity and cardiometabolicrisk factors in children and adolescents: a family-based study. Pediatr. Obes. 9,e58–e62.

Aniza, I., Fairuz, M.R., 2009. Factors influencing physical activity level among secondaryschool adolescents in Petaling District, Selangor. Med. J. Malays. 64, 228–232.

Barreira, T.V., Staiano, A.E., Katzmarzyk, P.T., 2013. Validity assessment of a portablebioimpedance scale to estimate body fat percentage in white and African-Americanchildren and adolescents. Pediatr. Obes. 8, e29–e32.

Bauman, A., Ma, G., Cuevas, F., et al., 2011. Cross-national comparisons of socioeconomicdifferences in the prevalence of leisure-time and occupational physical activity, andactive commuting in six Asia-Pacific countries. J. Epidemiol. Community Health 65,35–43.

Bertoncello, C., Cazzaro, R., Ferraresso, A., et al., 2008. Prevalence of overweight and obe-sity among school-aged children in urban, rural and mountain areas of the VenetoRegion, Italy. Public Health Nutr. 11, 887–890.

Boyko, E.J., Fujimoto, W.Y., Leonetti, D.L., et al., 2000. Visceral adiposity and risk oftype 2 diabetes: a prospective study among Japanese Americans. Diabetes Care 23,465–471.

Cole, T.J., Lobstein, T., 2012. Extended international (IOTF) body mass index cut-offs forthinness, overweight and obesity. Pediatr. Obes. 7, 284–294.

Crocker, P.R., Bailey, D.A., Faulkner, R.A., et al., 1997. Measuring general levels of physicalactivity: preliminary evidence for the Physical Activity Questionnaire for Older Chil-dren. Med. Sci. Sports Exerc. 29, 1344–1349.

Dan Jr., S.P., Mohd Nasir, M.T., Zalilah, M.S., 2007. Sex and ethnic differentials in physicalactivity levels of adolescents in Kuantan. Malays. J. Nutr. 13, 109–120.

Dan, S.P., Mohd, N.M., Zalilah, M.S., 2011. Determination of factors associated with phys-ical activity levels among adolescents attending school in Kuantan, Malaysia. Malays.J. Nutr. 17, 175–187.

Davis, A.M., Bennett, K.J., Befort, C., et al., 2011. Obesity and related health behaviorsamong urban and rural children in the United States: data from the National Healthand Nutrition Examination Survey 2003–2004 and 2005–2006. J. Pediatr. Psychol.36, 669–676.

De Onis, M., Onyango, A.W., Borghi, E., et al., 2007. Development of a WHO growth refer-ence for school-aged children and adolescents. BullWorld Health Organ 85, 660–667.

Deurenberg-Yap, M., Schmidt, G., van Staveren, W.A., et al., 2000. The paradox of lowbody mass index and high body fat percentage among Chinese, Malays and Indiansin Singapore. Int. J. Obes. Relat. Metab. Disord. 24, 1011–1017.

Deurenberg‐Yap, M., Chew, S., Deurenberg, P., 2002. Elevated body fat percentage andcardiovascular risks at low body mass index levels among Singaporean Chinese,Malays and Indians. Obes. Rev. 3, 209–215.

Dhanjal, T.S., Lal, M., Haynes, R., et al., 2001. A comparison of cardiovascular risk factorsamong Indo-Asian and Caucasian patients admitted with acute myocardial infarctionin Kuala Lumpur, Malaysia and Birmingham, England. Int. J. Clin. Pract. 55, 665–668.

Dunn, R.A., Tan, A.K., Nayga, R.M., 2012. Obesity inequality in Malaysia: decomposingdifferences by gender and ethnicity using quantile regression. Ethn. Health 17,493–511.

Efrat, M.W., 2013. Exploring effective strategies for increasing the amount of moderate-to-vigorous physical activity children accumulate during recess: a quasi-experimentalintervention study. J. Sch. Health 83, 265–272.

Eime, R.M., Young, J.A., Harvey, J.T., et al., 2013. A systematic review of the psychologicaland social benefits of participation in sport for children and adolescents: informingdevelopment of a conceptual model of health through sport. Int. J. Behav. Nutr.Phys. Act. 10, 98.

Fedewa, A.L., Candelaria, A., Erwin, H.E., et al., 2013. Incorporating physical activity intothe schools using a 3-tiered approach. J. Sch. Health 83, 290–297.

Ghosh, A., 2011. Rural–urban comparison in prevalence of overweight and obesity amongchildren and adolescents of Asian Indian origin. Asia Pac. J. Public Health 23, 928–935.

Gonzalez-Casanova, I., Sarmiento, O.L., Gazmararian, J.A., et al., 2013. Comparing threebody mass index classification systems to assess overweight and obesity in childrenand adolescents. Rev. Panam. Salud Publica 33, 349–355.

Goonasegaran, A.R., Nabila, F.N., Shuhada, N.S., 2012. Comparison of the effectiveness ofbody mass index and body fat percentage in defining body composition. Singap.Med. J. 53, 403–408.

Hajian-Tilaki, K., Heidari, B., 2012. Prevalences of overweight and obesity and their asso-ciation with physical activity pattern among Iranian adolescents aged 12–17 years.Public Health Nutr. 15, 2246–2252.

Hodgkin, E., Hamlin, M.J., Ross, J.J., et al., 2010. Obesity, energy intake and physical activityin rural and urban New Zealand children. Rural Remote Health 10, 1336.

Institute of Public Health, 2008. The Third National Health and Morbidity Survey (NHMSIII) 2006. Ministry of Health, Malaysia.

Ismail, M.N., Ruzita, A.N., Raduan, S., et al., 2009. Prevalence and trends of overweight andobesity in two cross-sectional studies of Malaysian children, 2002–2008. MASO 2009Scientific Conference on Obesity: Obesity and Our Environment. Kuala Lumpur.

Janz, K.F., Lutuchy, E.M., Wenthe, P., et al., 2008. Measuring activity in children and ado-lescents using self-report: PAQ-C and PAQ-A. Med. Sci. Sports Exerc. 40, 767–772.

Kuczmarski, R.J., Ogden, C.L., Guo, S.S., et al., 2000. CDC Growth Charts for the UnitedStates: methods and development. Vital Health Stat. 11 (2002), 1–190.

Lau, X.C., Chong, K.H., Poh, B.K., et al., 2013. Physical activity, fitness and the energy cost ofactivities: implications for obesity in children and adolescents in the tropics. Adv.Food Nutr. Res. 70, 49–101.

McCarthy, H.D., Ellis, S.M., Cole, T.J., 2003. Central overweight and obesity in British youthaged 11–16 years: cross sectional surveys of waist circumference. BMJ 326, 624.

Mohan, B., Kumar, N., Aslam, N., et al., 2004. Prevalence of sustained hypertension andobesity in urban and rural school going children in Ludhiana. Indian Heart J. 56,310–314.

Morrow Jr., J.R., Tucker, J.S., Jackson, A.W., et al., 2013. Meeting physical activity guidelinesand health-related fitness in youth. Am. J. Prev. Med. 44, 439–444.

Parrish, A.M., Okely, A.D., Stanley, R.M., et al., 2013. The effect of school recess interven-tions on physical activity: a systematic review. Sports Med. 43, 287–299.

Poh, B.K., Jannah, A.N., Chong, L.K., et al., 2011. Waist circumference percentile curves forMalaysian children and adolescents aged 6.0–16.9 years. Int. J. Pediatr. Obes. 6,229–235.

Rampal, L., Rampal, S., Khor, G.L., et al., 2007. A national study on the prevalence of obesityamong 16,127 Malaysians. Asia Pac. J. Clin. Nutr. 16, 561–566.

Rezali, F.W., Chin, Y.S., Mohd Yusof, B.N., 2012. Obesity-related behaviors of Malaysianadolescents: a sample from Kajang district of Selangor state. Nutr. Res. Pract. 6,458–465.

Sigmund, E., Sigmundova, D., 2013. Longitudinal 2-year follow-up on the effect of a non-randomised school-based physical activity intervention on reducing overweight andobesity of Czech children aged 10–12 years. Int. J. Environ. Res. Public Health 10,3667–3683.

105T.T. Su et al. / Preventive Medicine 67 (2014) 100–105

Stoner, L., Shultz, S.P., Lambrick, D.M., et al., 2013. The combating obesity in Maori andPasifika adolescent school-children study: COMPASS methodology and study proto-col. Int. J. Prev. Med. 4, 565–579.

Tan, A.K., Yen, S.T., Feisul, M.I., 2012. Determinants of body weight status in Malaysia: anethnic comparison. Int. J. Public Health 57, 279–288.

Teder, M., Morelius, E., Nordwall, M., et al., 2013. Family-based behavioural interventionprogram for obese children: an observational study of child and parent lifestyle inter-pretations. PLoS One 8, e71482.

Viner, R.M., Cole, T.J., Fry, T., et al., 2010. Insufficient evidence to support separate BMIdefinitions for obesity in children and adolescents from south Asian ethnic groupsin the UK. Int. J. Obes. (Lond.) 34, 656–658.

Wang, Y., Monteiro, C., Popkin, B.M., 2002. Trends of obesity and underweight in olderchildren and adolescents in the United States, Brazil, China, and Russia. Am. J. Clin.Nutr. 75, 971–977.

Wee, B.S., Poh, B.K., Bulgiba, A., et al., 2011. Risk of metabolic syndrome among chil-dren living in metropolitan Kuala Lumpur: a case control study. BMC Public Health11, 333.

World Health Organization, 2008. Waist Circumference and Waist–Hip Ratio Report of aWHO Expert Consultation (Geneva, 8–11 December 2008).

Ying-Xiu, Z., Jing-Yang, Z., Jin-Shan, Z., et al., 2013. The role of 1-h physical activity everyday in preventing obesity in adolescents in Shandong, China. Eur. J. Pediatr. 172,325–330.