- Email: [email protected]
obesity in Australian migrant children
Obesity Research & Clinical Practice (2008) 2, 179—187
ORIGINAL ARTICLE
Parental obesity as a predictor of childhood overweight/obesity in Australian migrant children Yang Li a, Kaye Brock a,∗, Rosemary Cant a, Liang Ke a, Stephen Morrell b a
Department of Behavioural and Community Health Sciences, Faculty of Heath Sciences, University of Sydney, East St, Lidcombe, Sydney, NSW 2141, Australia b School of Public Health, Faculty of Medicine, University of Sydney, NSW, Australia Received 27 February 2008 ; received in revised form 28 April 2008; accepted 30 April 2008
KEYWORDS Childhood overweight/obesity; Immigrant; Australia; Parental obesity; Exercise
Summary Obesity levels are increasing disproportionately in immigrant children worldwide. We investigated predictors of immigrant children’s obesity status in a well-documented 3-year follow-up study of children at 9 years (n = 1232) and 12 years (n = 628) of age living in inner city areas of Sydney (Australia). The major immigrant groups in this sample were from Europe, the Middle East and SE Asia. Having an obese parent and having either parent or child not being born in Australia and not playing organised sport were predictors of childhood obesity. If either parent was obese compared to non-obese, then the odds ratio for a 9-year-old child of being obese compared to normal weight was 4.9 (95% CI: 3.0—8.0); for 12-year olds the odds ratio was 8.0 (95% CI: 3.6—18). For the survey of 9-years old, a parent born outside Australia or if the child himself was born outside Australia was associated with an almost twofold chance of being obese (OR = 1.8, 95% CI: 1.1—3.0), and not participating in organised sports was significantly associated with childhood obesity. Nine-year olds who did not participate in organised sports outside of school hours were almost twice as likely to be obese (OR = 1.9, 95% CI: 1.2—2.9). When stratified by ethnicity, participation in sports was not related to lower obesity status in children of SE Asian origin. One reason for this could be because SE Asian had very limited participation in sports compared to their counterparts. Accordingly, we believe that health promotion messages to avoid obesity need to be targeted ethnic-specifically. © 2008 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
Introduction
∗
Corresponding author. Tel.: +61 2 9351 9124. E-mail address: [email protected] (K. Brock).
Obesity levels are increasing disproportionately in immigrant children worldwide [1], and childhood obesity is now endemic in Australia [2]. Two NSW surveys (1997 and 2004) reported children of
1871-403X/$ — see front matter © 2008 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.orcp.2008.04.007
180 European and Middle Eastern children background to have disproportionately high rates of obesity [2,3]. In the 1997 survey (n = 5518), those of European (30%) and Middle Eastern origin (28%) had a higher proportion of overweight than their Asian and Australian counterparts (19% and 20%, respectively) [2]. By the 2004 survey (n = 5407), 38% of children from a Middle Eastern background, 26% from European, 26% from Australian and 24% of Asian background children were overweight [3]. Differences in risk factors for higher, and increasing, obesity prevalence by ethnicity were not investigated in these studies nor did other large studies of childhood obesity in NSW [4], Victoria [5] and South Australia [6] stratify by ethnic background. Accordingly, we took the opportunity to investigate risk factors associated with obesity in a longitudinal study of 9-year-old Sydney school children from Australian, European, Middle Eastern and South East Asian (SE Asian) ethnic origins, first surveyed in 1994/1995 and subsequently followed up in 1997 at age 12 years. Factors known to be associated with obesity status such as parental obesity, child physical activity levels both inside and outside school and socio-economic status (SES) were investigated.
Methods Study participants Data analysed for this paper originally were collected as part of the ‘‘Sydney aircraft noise & child blood pressure study’’ [7], a longitudinal study that initially surveyed and measured a cohort of 1232 year 3/4 school children attending 75 primary schools randomly sampled from within 20 km of Sydney airport in 1994/1995. At the followup survey 3 years later (1997), measurements were taken from 628 of the baseline participants when they were in year 6 of primary school. The mean ages for the 1994/1995 and 1997 study cohorts were 9.2 ± 0.4 and 11.9 ± 0.3 years, respectively.
Y. Li et al. In line with the current international standards, the study subjects’ obesity status was defined based on their BMI [9]. There has been debate over how to define children’s obesity status but a recent consensus [9] has established criteria which use worldwide data tracked back from adult overweight and obesity standards (9-year-old children: overweight BMI: 19.09—22.78 kg/m2 , obese: BMI ≥ 22.79 kg/m2 ; 12-year-old children: overweight BMI: 21.45—26.34 kg/m2 , obese: BMI ≥ 26.35 kg/m2 ). A similar procedure was followed for children of SE Asian origin. For all adults excepting those of Asian origin, obesity was defined as BMI ≥ 30 kg/m2 and overweight as 25 kg/m2 ≤ BMI < 30 kg/m2 [10]. For Asian subjects a lower cut off BMI for obesity (BMI ≥ 25 kg/m2 ) was used [11], and overweight was defined by a BMI range of 23—24.9 kg/m2 . Longitudinal BMI change from 9 to 12 years was defined as a ‘‘high’’ increase if greater than +2 kg/m2 .
Ethnicity classifications Children’s ethnic origins were defined according to the child’s or parents’ country of birth, in line with the standard classification of countries for social statistics (ASCCSS), issued by Australian Bureau of Statistics (ABS) [12]. A child was classified as having an ‘immigrant background’ if either the child or one or more parent was born outside of Australia. Where a birth country of the child or a parent was in Europe, the Middle East, South East Asia (SE Asia) or any country other than Australia, the child’s ethnic origin was allocated to that ethnic sub-group. The sub-group ‘Australian’ then was composed of children who themselves and both parents were Australian born.
Physical activity measurements Two measures of child physical activity were adopted as indicators of levels of physical activity. First, each study participant was asked if they played actively during their lunch or recess breaks at school; they also were asked if they participated in organised sport outside of school hours.
Overweight/obesity measurements
Socio-economic status (SES)
Child anthropometric measures, including weight and height, were conducted at baseline and follow-up. Heights were measured with socks on by a Harpenden pocket stadiometer [8], and weights were measured with a Tanita System 502 digital bathroom scale accurate to ±100 g [7].
As SES is thought to be associated with obesity, we measured it as both a risk factor and a confounder for ethnicity in this study, with the postcode of each child’s home address used as the SES marker. For the present study, the general index of deprivation, developed by the ABS from a principal
Obesity in Australian migrant children
181
components analysis of key postcode level demographic, wealth and income indicators from the Australian census, was used as the SES measure [13]. ‘‘Low’’ SES level was defined as below the median index score. Individual SES information was not collected from parents themselves in the original study in order to maximise the survey participation rate.
Statistical analysis Descriptive statistics (mean, standard deviation, p-value) and inferential statistics (t-test, oneway ANOVA, chi-square and logistic regression) were performed using SPSS version 13 [14]. Risk estimates with 95% confidence intervals were calculated, reported as odds ratios for cross-sectional data and as relative risks for changes occurring between the initial and 3-year follow-up surveys, using logistic regression [15]. Logistic regression was used to establish the degree of association between the various risk factors and different levels of childhood obesity and obesity/overweight status adjusted for SES and age. For predictors of longitudinal changes in BMI exceeding 2 kg/m2 relative risk estimates were calculated and presented with and without baseline BMI levels included in the regression models.
Results Ethnic differences by overweight/obesity The ethnic composition of the child cohort at age 9 years was 31% Australian born, 20% European, 11% Middle Eastern and 18% of SE Asian origin, with 10% comprising a mixture of various countries of birth, including New Zealand, Pacific Islands and North Asian. A further 10% had no report country of birth information recorded. Similar distributions were found in the children at age 12 (data not shown). As can be seen from Fig. 1, a significantly higher percentage of 9-year-old Middle Eastern and SE Asian children came from low-SES areas compared to Australians (47% and 71%, respectively, vs. 34%; OR = 1.7, 95% CI: 1.2—2.6; OR = 4.7, 95% CI: 3.2—6.7, respectively); accordingly, all subsequent comparisons were adjusted for SES. In addition Fig. 1 shows that SE Asian and Middle Eastern children have significantly less participation in sports compared to Australians or Europeans. At age 9 years children of European and Middle Eastern origin had higher mean BMI measurements
Figure 1 Prevalence of overweight and distribution of risk factors for overweight by ethnic origin, 9-year-old children living in Sydney, Australia.
(mean ± S.D.) compared to Australian children (European: 18.1 ± 3.0 kg/m2 ; Middle Eastern: 18.9 ± 3.4 kg/m2 ; Australian 17.6 ± 3.3 kg/m2 ). Children from Australian backgrounds also had the lowest prevalence of overweight (25%) compared to those from European (31%), SE Asian (33%) and Middle Eastern backgrounds (40%). The prevalence differences were similar in boys and girls, and reflected their mean BMI measurements (boys: Australian 17.7 ± 2.9 kg/m2 , European 18.1 ± 3.0 kg/m2 , Middle Eastern 19.0 ± 3.6 kg/m2 and SE Asian 17.7 ± 3.3 kg/m2 ; girls: Australian 17.7 ± 2.8 kg/m2 , European 18.0 ± 3.1 kg/m2 , 2 Middle Eastern 18.8 ± 3.3 kg/m and SE Asian 17.5 ± 3.4 kg/m2 ).
Predictors of obesity status in children Having an obese parent, being not ethnically Australian, being of low SES and not participating in organised sports were significant predictors of both childhood obesity and overweight/obesity (Table 1). If either parent was obese compared to non-obese, then the odds ratio for a 9-yearold child of being obese compared to normal weight was 4.9 (95% CI: 3.0—8.0); for 12-year olds the odds ratio was 8.0 (95% CI: 3.6—18). Similar odds ratios were seen for overweight and obese (Table 1). For the survey of 9-years old, a parent born outside Australia or if the child himself was born outside Australia was associated with an almost twofold chance of being obese (OR = 1.8, 95% CI: 1.1—3.0), and not participating in organised sports was significantly associated with childhood obesity. Nine-year olds who did not
182
Table 1
Risk factors for childhood obesity status by ethnic origin in 9-year-old children living in Sydney, Australia (OR, 95% CI)
Risk factors
Obesity and overweight risk for children aged 9 years Total (n = 1232), ORa (95% CI)
Australian (n = 381), ORa (95% CI)
European (n = 251), ORa (95% CI)
Middle Eastern (n = 132), ORa (95% CI)
SE Asian (n = 216), ORa (95% CI)
Non-Australian born No Yesb Yesc
1.0 1.3 (1.0—1.7) 1.8 (1.1—3.0)
— — —
— — —
— — —
— — —
Maternal obesity No Yesb Yesc
1.0 2.9 (1.8—4.8) 4.6 (2.5—8.6)
1.0 2.8 (1.3—6.1) 5.0 (1.6—16)
1.0 4.2 (1.4—12.4) 6.1 (1.6—24)
1.0 1.7 (0.5—5.6) 2.5 (0.6—10.5)
1.0 5.8 (2.4—14.1) 8.2 (3.0—23)
Paternal obesity No Yesb Yesc
1.0 2.6 (1.7—4.1) 4.4 (2.5—7.8)
1.0 2.6 (1.2—5.8) 12.7 (4.2—39)
1.0 2.5 (1.0—5.8) 3.5 (1.0—12.6)
1.0 3.0 (1.0—9.6) 4.3 (1.1—17)
1.0 3.9 (2.0—7.4) 6.7 (2.9—15)
Parental obesity No Yesb Yesc
1.0 2.9 (2.1—4.2) 4.9 (3.0—8.0)
1.0 3.2 (1.7—5.9) 13.6 (5.0—37)
1.0 2.9 (1.4—5.9) 4.6 (1.7—12.6)
1.0 3.0 (1.2—7.5) 4.1 (1.2—14)
1.0 3.7 (2.0—6.8) 5.7 (2.6—13)
Low SES No Yesb Yesc
1.0 1.1 (0.8—1.4) 1.2 (0.8—1.9)
1.0 1.0 (0.6—1.7) 0.7 (0.3—1.8)
1.0 0.9 (0.5—1.6) 1.0 (0.4—2.8)
1.0 2.2 (1.0—4.5) 1.8 (0.7—5.0)
1.0 0.8 (0.5—1.6) 1.2 (0.5—2.8)
In-school activity Yes Nob Noc
1.0 1.5 (1.1—2.1) 1.6 (1.0—2.8)
1.0 1.1 (0.6—2.1) 1.1 (0.4—3.5)
1.0 2.3 (1.1—4.7) 1.6 (0.5—5.2)
1.0 1.0 (0.4—2.6) 1.3 (0.4—4.3)
1.0 1.6 (0.7—3.6) 1.9 (0.7—5.2)
Organised sports Yes Nob Noc
1.0 1.2 (0.9—1.6) 1.9 (1.2—2.9)
1.0 1.1 (0.6—1.8) 2.3 (1.0—5.7)
1.0 1.4 (0.8—2.5) 1.4 (0.6—3.6)
1.0 1.3 (0.6—2.6) 2.6 (0.9—7.6)
1.0 0.9 (0.5—1.5) 0.8 (0.4—1.8)
a b c
Adjusted for age and SES. OR (95% CI) for overweight. OR (95% CI) for obesity.
Y. Li et al.
Obesity in Australian migrant children participate in organised sports outside of school hours were almost twice as likely to be obese (OR = 1.9, 95% CI: 1.2—2.9); this association with obesity was particularly evident in Australian and Middle Eastern children. Notable was that lack of in-school activity was associated with risks for overweight and obesity in SE Asian and European children. When adjusted for age and SES, children from Middle Eastern background had double the odds of being overweight compared to their Australian peers (OR = 2.0, 95% CI: 1.3—3.0), and SE Asian children 1.6 the odds (95% CI: 1.1—2.3). The odds were over threefold for obesity (Middle Eastern OR = 3.2, 95% CI: 1.7—6.3; SE Asian OR = 3.2, 95% CI: 1.8—5.9) (Table 2). These differences were also evident when stratified by gender except there was no significant difference in odds of being overweight for SE Asian girls compared to their Australian counterparts. Middle Eastern children had twice the adjusted odds of having an obese parent in comparison with their Australian counterparts (26% vs. 15%; OR = 2.1, 95% CI: 1.3—3.6), and also had nearly triple the odds of not attending organised sports activity outside school hours (52% vs. 26%; OR = 2.8, 95% CI: 1.8—4.3) than Australian children. In comparison, SE Asian children had three times the odds (adjusted for SES and age) of having an obese parent compared to their Australian peers (37% vs. 15%; OR = 3.4, 95% CI: 2.2—5.2). In addition, there was a significantly greater percentage of 9-yearold SE Asian children not participating in organised sports activities outside school than that of the Australian children (26% vs. 61%; OR = 3.9, 95% CI: 2.7—5.7).
Associations between ethnic background and risk factors for childhood obesity status Parental obesity at baseline was consistently the strongest predictor of obesity/overweight in children across all ethnic groups, after adjusting for SES and age (Table 1). The association with obesity ranged from fourfold in Middle Eastern children to fourteen fold in Australian children, however the associations with overweight and obesity were consistently threefold over all ethnic groups. It appears that maternal obesity was a stronger risk factor in SE Asian children obesity status whereas paternal obesity affected Middle Eastern and Australian children’s obesity status. The other predictors of interest were a twofold difference in risk of obesity with lack of in-school activity in total population and SE Asian children (OR = 1.6, 95% CI: 1.0—2.8) for total population and (OR = 1.9, 95% CI:
183 0.7—5.2) for SE Asians. There were similar twofold risks for no participation in organised sports for Australian and Middle Eastern children’s obesity status (OR = 2.3, 95% CI: 1.0—5.7) for Australians and (OR = 2.6, 95% CI: 1.0—7.6) for Middle Eastern.
Longitudinal BMI change Similar risk factors were associated with high-BMI increase (BMI increase >2 kg/m2 ) independent of age and SES over the 3-year period from age 9 to 12 years (Table 3 ). Parental obesity was the major risk factor for high-BMI increase across all ethnic groups except for Middle Eastern children (RRs: all 2.2, 95% CI: 1.3—3.6; Australian 3.0, 95% CI: 1.2—7.2; European 2.7, 95% CI: 1.0—7.3; SE Asian 3.3, 95% CI: 1.2—9.1). The only other significant predictors for high increase in BMI over 3 years was being from a low-socio-economic area or not attending organised sports after school hours, in Australian children only (RRs: 1.9, 95% CI: 1.0—3.6; 2.4, 95% CI: 1.2—4.8, respectively). However, as can be seen in Table 3, if baseline BMI was also accounted for, parental obesity status no longer remained as a predictor of 3-year longitudinal BMI change from age 9 to 12 in our study cohort. However, baseline BMI was significantly associated with high-BMI change (Table 3), and parental obesity significantly associated with baseline BMI, suggesting a conferring of high BMI on the child in the first place. Being from a low-socio-economic area or not attending organised sports outside school hours, nevertheless, were still significantly associated with BMI increase over 3 years in the Australian children only (RR: 1.9, 95% CI: 1.0—3.9; 2.7, 95% CI: 1.3—5.7, respectively).
Discussion This study has demonstrated that cross-sectionally having an obese parent, not being born in Australia and having either parent not being born in Australia, and not being involved in physical activity were strong predictors of childhood obesity at 9 years old, independent of SES and age. However parental obesity was not related to longitudinal change in obesity after adjustment for initial BMI. This indicates that BMI tracking is strong, since baseline BMI was related to parental BMI, and, established quite early, appears as the mechanism for further BMI increase. Both the prevalence of and factors associated with obesity varied by ethnic background. Higher participation in organised
184
Y. Li et al.
Table 2 Prevalence of overweight/obesity (BMI) by ethnicity, gender and their comparison with Australian counterparts in 9-year-old children living in Sydney, Australia (OR, 95% CI) Ethnic origin by country of birth Australian European Middle Eastern SE Asian a
Overweight
Obesity a
Boys + girls, ORa (95% CI)
Boys, OR (95% CI)
1.0 1.3 (0.9—1.9) 2.0 (1.3—3.0) 1.6 (1.1—2.3)
1.0 1.2 (0.8—2.1) 2.2 (1.1—4.0) 1.7 (1.0—3.0)
Girls, OR (95% CI)
a
1.0 1.5 (0.9—2.4) 1.8 (1.0—3.3) 1.4 (0.8—2.4)
Boys + girls, ORa (95% CI)
Boys, ORa (95% CI)
Girls, ORa (95% CI)
1.0 1.6 (0.9—3.1) 3.2 (1.7—6.3) 3.2 (1.8—5.9)
1.0 1.5 (0.7—3.5) 3.2 (1.3—7.9) 2.9 (1.3—6.6)
1.0 1.9 (0.7—4.9) 3.3 (1.2—9.0) 3.2 (1.4—7.7)
Adjusted for age and SES.
sport was related more strongly to lower obesity status in children of Australian, Middle Eastern or European backgrounds, which may be because SE Asian children had very limited participation in organised sport compared to their counterparts. The predictors of childhood obesity found here parallel the findings of three other Australian cohort studies [4—6] which also found parental obesity status to be a risk factor for children’s obesity but without breakdowns by ethnicity. In the absence of studies on immigrant parental obesity, it is helpful to compare the present findings with those that have been conducted within the countries of origin, particularly studies of European [16,17] and SE Asian [18] populations, that have linked child obesity status to parental obesity status. To our knowledge one Korean study [18] has reported the risk of being obese in children aged 7—12 years (n = 3059), as 1.9 (95% CI: 1.3—2.8) for those whose mothers were obese, 1.5 (95% CI: 1.3—1.7) for those whose fathers were obese, and 2.4 (95% CI: 1.8—3.1) if either parent was obese. Somewhat unexpectedly in the present study, children’s socio-economic level was a significant risk factor for obesity only in Middle Eastern children, despite consistent reports from the literature showing SES to be related to obesity [19]. It was because of this, a priori, that all the data in this paper were adjusted for socio-economic status. Physical activity level of children, represented in our study by a question on playing activity in the child’s recess or lunch breaks in school, and by participation in sport outside school, also proved to be significantly associated with obesity in our study cohort. This is supported by previous research [19,20]. SE Asian children living in Australia in our study tended to have relatively high-participation rates in in-school activity (boys: 96%; girls: 79%) but not after-school organised sports (boys: 42%; girls: 36%) compared to those living in China [21], as well
as in Chinese immigrants living in US [22]. However these rates are low compared to Australian children in our study (in-school activity: boys 93%, girls 74%; after-school organised sports: boys 77%, girls 72%). In summary, despite there being some 12 cross-sectional [2,3,23—30] and 6 cohort studies [4—6,31—33] investigating childhood obesity in Australia over the past three decades indicating a rise in obesity rates, to our knowledge the only studies that have analysed overweight or obesity prevalence in terms of ethnicity were the 1997 NSW Schools Physical Activity and Nutrition Survey [2] and the 2004 NSW Schools Physical Activity and Nutrition Survey (SPANS) [3]. Neither investigated risk factors associated with different ethnic backgrounds. Also, in these studies, study participants were only categorized according to language spoken most at home, in contrast, to the present study where ethnicity was defined by both parental and children’s country of birth. The strengths of this study are that the children were taken from a very well-designed random sample of inner city Sydney schools and there was accurate data on both their country of birth and their parents’ country of birth. The total sample had sufficient power to be able to stratify by ethnic group to obtain reasonable statistical estimates. And child weights and heights were measured according to a consistent protocol and were not self reported. The main study limitations were that there were very limited nutritional and physical activity data. Also, there was a substantial drop out on follow-up (49%) which possibly could lead to bias in analysis of obesity change. This study was voluntary and thus it is not surprising that although the study children were from randomly selected schools, the response rate was only 40% and there was almost 50% attrition between the baseline and follow-up surveys. However, as study children were compared with themselves in the cohort analy-
Risk factors
Risk factors for increase in BMI (>2 kg/m2 ) over 3 years by ethnic origin in 9—12-year children (RR, 95% CI) BMI increase (>2 kg/m2 from age 9 to 12) Total (n = 586) RRa (95% CI)
Australian (n = 187)
1.3 (1.2—1.4)
RRa (95% CI) 1.4 (1.2—1.6)
Overweight/obese (age 9) No 1.0 Yes 2.7 (1.8—4.2)
1.0 3.5 (1.7—7.4)
BMI (age 9)
RRb (95% CI)
European (n = 132) RRb (95% CI)
RRa (95% CI) 1.2 (1.0—1.4)
Middle Eastern (n = 68) RRb (95% CI)
1.0 2.5 (1.1—5.9)
RRa (95% CI) 1.2 (1.0—1.4)
SE Asian (n = 85) RRb (95% CI)
1.0 1.0 (0.3—2.9)
RRa (95% CI) 1.8 (1.3—2.3)
RRb (95% CI)
Obesity in Australian migrant children
Table 3
1.0 13.4 (3.8—47.8)
Maternal obesity No 1.0 Yes 2.6 (1.3—5.2)
1.0 1.5 (0.7—3.2)
1.0 3.0 (1.0—8.5)
1.0 1.4 (0.4—4.7)
1.0 5.0 (0.9—26.4)
1.0 5.0 (0.9—28.2)
1.0 0.4 (0.04—4.1)
1.0 0.2 (0.02—2.9)
1.0 2.3 (0.6—9.0)
1.0 0.5 (0.1—2.9)
Paternal obesity No 1.0 Yes 1.9 (1.0—3.5)
1.0 1.3 (0.7—2.4)
1.0 4.1 (1.2—13.8)
1.0 2.7 (0.7—9.9)
1.0 1.6 (0.5—5.3)
1.0 1.2 (0.3—4.4)
1.0 1.2 (0.3—4.7)
1.0 1.0 (0.2—4.2)
1.0 3.3 (1.1—9.8)
1.0 1.7 (0.5—6.1)
Parental obesity No 1.0 Yes 2.2 (1.3—3.6)
1.0 1.4 (0.8—2.4)
1.0 3.0 (1.2—7.2)
1.0 1.6 (0.6—4.3)
1.0 2.7 (1.0—7.3)
1.0 2.4 (0.9—6.7)
1.0 1.1 (0.3—3.8)
1.0 0.8 (0.2—3.2)
1.0 3.3 (1.2—9.1)
1.0 1.9 (0.6—6.3)
Low SES No Yes
1.0 1.3 (0.9—1.8)
1.0 1.4 (0.9—2.0)
1.0 1.9 (1.0—3.6)
1.0 1.9 (1.0—3.9)
1.0 0.9 (0.4—2.0)
1.0 1.0 (0.5—2.2)
1.0 0.5 (0.2—1.5)
1.0 0.5 (0.2—1.4)
1.0 1.1 (0.4—2.7)
1.0 1.9 (0.6—6.0)
No in-school activity No 1.0 Yes 1.4 (0.9—2.3)
1.0 0.6 (0.3—1.1)
1.0 1.5 (0.7—3.3)
1.0 0.9 (0.3—2.9)
1.0 0.6 (0.2—1.9)
1.0 0.3 (0.1—1.3)
1.0 2.0 (0.6—6.8)
1.0 0.1 (0.01—1.4)
1.0 1.9 (0.6—6.2)
1.0 0.9 (0.2—4.7)
No organised sports No 1.0 Yes 1.3 (0.9—1.9)
1.0 1.4 (0.9—2.0)
1.0 2.4 (1.2—4.8)
1.0 2.7 (1.3—5.7)
1.0 0.8 (0.4—1.8)
1.0 0.7 (0.3—1.6)
1.0 0.8 (0.3—2.1)
1.0 0.8 (0.3—2.2)
1.0 1.2 (0.5—3.2)
1.0 1.4 (0.5—4.4)
a and SES. b Adjusted for age, SES and baseline BMI.
185
186
Y. Li et al.
sis, factors associated with progression to obesity would still be revealed. In addition, the differences between those lost to follow-up vs. those followed up pertaining to the study factors was in ethnic distribution (lower proportion of SE Asians at follow-up) higher mean of BMI, waist circumference and body fat in the children (but not in obesity in their parents) and less sports (PA) activity outside school (Appendix A). If we predicted any bias from these differences we would expect the association with parental obesity to be even stronger if there was no attrition, i.e., a bias toward the null with respect to parental obesity. It also should be noted that, all our findings at follow-up were consistent with the baseline findings.
in these different groups. As evidence from our data support a relationship between obesity levels and in-school activity in those from SE Asian and Middle Eastern backgrounds, we recommend that special programs be established in all schools with high-minority group populations, to involve children in school-break physical activity. We also believe that as there is large ethnic variation in participation in sports outside schools that ethnospecific educational programs should be established and sponsored by community leaders to underline for parents the importance of physical activity outside school for their children. In addition, ethnospecific events both in and outside of school should be encouraged to incorporate physical activity components.
Conclusion and recommendations Conflicts of interest statement In conclusion, more research into ethnic and migrant variation in obesity in larger cohorts in both Australia and other countries with large immigrant populations such as Great Britain, Canada and the USA. In addition and in parallel there needs to be in depth studies into ethnic variation of metabolic responses to exercise and diet
Risk factor for obesity in children (%)
Appendix A. Risk factor data on those who did and did not respond to 3-year follow-up
Total population (n = 1232) Follow-up (n = 628) (51%)
Parental obesity (%) Australian (%) Middle Eastern (%) SE Asian (%) European (%) PA outside school (%) PA in school (%) Low SES (%) Female (%) Age (years), mean ± S.D. Birth weight, mean ± S.D. SBP, mean ± S.D. DBP, mean ± S.D. BMI, mean ± S.D. WC, mean ± S.D. %TBF, mean ± S.D.
There are no financial and conflict of interest for any of the authors.
54 54 54 44 55 64 84 55 49 9.2 ± 0.4 3.3 ± 0.6 111 ± 10 59 ± 8 17.7 ± 2.9 61 ± 8 10 ± 6
Lost to follow-up (n = 604) (49%) 46 46 46 56 46 56 84 61 50 9.2 ± 0.4 3.3 ± 0.6 112 ± 10 60 ± 8 18.1 ± 3.2 62 ± 9 11 ± 6
Significance (p) ns ns ns 0.06 ns <0.05 ns 0.06 ns ns ns ns ns <0.05 <0.05 <0.05
Obesity in Australian migrant children
References [1] Crawford PB, Story M, Wang MC, Ritchie LD, Sabry ZI. Ethnic issues in the epidemiology of childhood obesity. Pediatr Clin North Am 2001;48:855—78. [2] Booth ML, Wake M, Armstrong T, Chey T, Hesketh K, Mathur S. The epidemiology of overweight and obesity among Australian children and adolescents, 1995—97. Aust N Z J Public Health 2001;25:162—9. [3] Booth ML, Okely A, Wilson ED, Hardy L. NSW Schools Physical Activity and Nutrition Survey 2004 (SPANS). Sydney: NSW Health; 2006. [4] Garnett SP, Cowell CT, Baur LA, Shrewsbury VA, Chan A, Crawford D, et al. Increasing central adiposity: the Nepean longitudinal study of young people aged 7—8 to 12—13 y. Int J Obes (Lond) 2005;29:1353—60. [5] Hesketh K, Wake M, Waters E, Carlin J, Crawford D. Stability of body mass index in Australian children: a prospective cohort study across the middle childhood years. Public Health Nutr 2004;7:303—9. [6] Magarey AM, Daniels LA, Boulton TJ, Cockington RA. Predicting obesity in early adulthood from childhood and parental obesity. Int J Obes (Lond) 2003;27:505—13. [7] Morrell S. Aircraft noise and child blood pressure. Sydney: School of Public Health, University of Sydney; 2003. p. 568. [8] Holtain Ltd., Crosswell, UK. [9] Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. Br Med J 2000;320:1240—3. [10] Himes JH, Dietz WH. Guidelines for overweight in adolescent preventive services: recommendations from an expert committee. The Expert Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services. Am J Clin Nutr 1994;59:307—16. [11] Inou S, Zimmet P, Shah B, Marmot MG. The Asia-Pacific perspective: redefining obesity and its treatment. World Health Organisation, International Association for the Study of Obesity, International Obesity Task Force; 2000. [12] Castles I. Australian standard classification of countries for social statistics. Australian Bureau of Statistics; 1990. p. 143. [13] Castle I. Information paper: 1991 census—–socio-economic indexes for areas. Canberra: Australian Bureau of Statistics; 1993. p. 38. [14] Coakes SJ, Steed LG, Dzidic P. SPSS version 13.0 for windows: analysis without anguish. Milton, Qld: John Wiley & Sons Australia; 2005. [15] Menard SW. Applied logistic regression analysis. Thousand Oaks, CA: Sage Publications; 1995. [16] Savva SC, Kourides Y, Tornaritis M, Epiphaniou-Savva M, Chadjigeorgiou C, Kafatos A. Obesity in children and adolescents in Cyprus. Prevalence and predisposing factors. Int J Obes (Lond) 2002;26:1036—45. [17] De Vito E, La Torre G, Langiano E, Berardi D, Ricciardi G. Overweight and obesity among secondary school children in Central Italy. Eur J Epidemiol 1999;15:649—54. [18] Lee K, Kwon ER, Park TJ, Park MS, Lenders CM. Parental overweight as an indicator of childhood overweight: how sensitive? Asia Pac J Clin Nutr 2006;15:196—200.
187 [19] Moussa MA, Skaik MB, Selwanes SB, Yaghy OY, Bin-Othman SA. Factors associated with obesity in school children. Int J Obes (Lond) 1994;18:513—5. [20] McMurray RG, Harrell JS, Deng S, Bradley CB, Cox LM, Bangdiwala SI. The influence of physical activity, socioeconomic status, and ethnicity on the weight status of adolescents. Obes Res 2000;8:130—9. [21] Tudor-Locke C, Ainsworth BE, Adair LS, Du S, Popkin BM. Physical activity and inactivity in Chinese school-aged youth: the China Health and Nutrition Survey. Int J Obes (Lond) 2003;27:1093—9. [22] Gordon-Larsen P, McMurray RG, Popkin BM. Adolescent physical activity and inactivity vary by ethnicity: the National Longitudinal Study of Adolescent Health. J Pediatr 1999;135:301—6. [23] Timperio A, Salmon J, Telford A, Crawford D. Perceptions of local neighbourhood environments and their relationship to childhood overweight and obesity. Int J Obes (Lond) 2005;29:170—5. [24] Rehor PR, Brownsey E. A preliminary investigation into the prevalence and trends of overweight and obesity of northern Tasmanian primary school children. Aust N Z J Public Health 2002;26:479—80. [25] O’Dea JA. Differences in overweight and obesity among Australian schoolchildren of low and middle/high socioeconomic status. Med J Aust 2003;179:63. [26] Wilcken DE, Lynch JF, Marshall MD, Scott RL, Wang XL. Relevance of body weight to apolipoprotein levels in Australian children. Med J Aust 1996;164:22—5. [27] Harvey PWJ, Althaus M-M. The distribution of body mass index in Australian children aged 7—15 years. Aust J Nutr Diet 1993;50:151—3. [28] Wickramasinghe VP, Cleghorn GJ, Edmiston KA, Murphy AJ, Abbott RA, Davies PSW. Validity of BMI as a measure of obesity in Australian White Caucasian and Australian Sri Lankan children. Ann Hum Biol 2005;32:60—71. [29] Vincent SD, Pangrazi RP, Raustorp A, Tomson LM, Cuddihy TF. Activity levels and body mass index of children in the United States, Sweden, and Australia. Med Sci Sports Exerc 2003;35:1367—73. [30] Kremer PJ, Bell AC, Sanigorski AM, Swinburn BA. Overweight and obesity prevalence in children based on 6- or 12-month IOTF cut-points: does interval size matter? Int J Obes (Lond) 2006;30:603—5. [31] Burke V, Beilin LJ, Dunbar D. Family lifestyle and parental body mass index as predictors of body mass index in Australian children: a longitudinal study. Int J Obes (Lond) 2001;25:147—57. [32] Burke V, Beilin LJ, Simmer K, Oddy WH, Blake KV, Doherty D, et al. Predictors of body mass index and associations with cardiovascular risk factors in Australian children: a prospective cohort study. Int J Obes (Lond) 2005;29:15— 23. [33] Mamun AA, Lawlor DA, O’Callaghan MJ, Williams GM, Najman JM. Family and early life factors associated with changes in overweight status between ages 5 and 14 years: findings from the Mater University Study of Pregnancy and its outcomes. Int J Obes (Lond) 2005;29:475— 82.
Available online at www.sciencedirect.com
ORIGINAL ARTICLE
Parental obesity as a predictor of childhood overweight/obesity in Australian migrant children Yang Li a, Kaye Brock a,∗, Rosemary Cant a, Liang Ke a, Stephen Morrell b a
Department of Behavioural and Community Health Sciences, Faculty of Heath Sciences, University of Sydney, East St, Lidcombe, Sydney, NSW 2141, Australia b School of Public Health, Faculty of Medicine, University of Sydney, NSW, Australia Received 27 February 2008 ; received in revised form 28 April 2008; accepted 30 April 2008
KEYWORDS Childhood overweight/obesity; Immigrant; Australia; Parental obesity; Exercise
Summary Obesity levels are increasing disproportionately in immigrant children worldwide. We investigated predictors of immigrant children’s obesity status in a well-documented 3-year follow-up study of children at 9 years (n = 1232) and 12 years (n = 628) of age living in inner city areas of Sydney (Australia). The major immigrant groups in this sample were from Europe, the Middle East and SE Asia. Having an obese parent and having either parent or child not being born in Australia and not playing organised sport were predictors of childhood obesity. If either parent was obese compared to non-obese, then the odds ratio for a 9-year-old child of being obese compared to normal weight was 4.9 (95% CI: 3.0—8.0); for 12-year olds the odds ratio was 8.0 (95% CI: 3.6—18). For the survey of 9-years old, a parent born outside Australia or if the child himself was born outside Australia was associated with an almost twofold chance of being obese (OR = 1.8, 95% CI: 1.1—3.0), and not participating in organised sports was significantly associated with childhood obesity. Nine-year olds who did not participate in organised sports outside of school hours were almost twice as likely to be obese (OR = 1.9, 95% CI: 1.2—2.9). When stratified by ethnicity, participation in sports was not related to lower obesity status in children of SE Asian origin. One reason for this could be because SE Asian had very limited participation in sports compared to their counterparts. Accordingly, we believe that health promotion messages to avoid obesity need to be targeted ethnic-specifically. © 2008 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
Introduction
∗
Corresponding author. Tel.: +61 2 9351 9124. E-mail address: [email protected] (K. Brock).
Obesity levels are increasing disproportionately in immigrant children worldwide [1], and childhood obesity is now endemic in Australia [2]. Two NSW surveys (1997 and 2004) reported children of
1871-403X/$ — see front matter © 2008 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.orcp.2008.04.007
180 European and Middle Eastern children background to have disproportionately high rates of obesity [2,3]. In the 1997 survey (n = 5518), those of European (30%) and Middle Eastern origin (28%) had a higher proportion of overweight than their Asian and Australian counterparts (19% and 20%, respectively) [2]. By the 2004 survey (n = 5407), 38% of children from a Middle Eastern background, 26% from European, 26% from Australian and 24% of Asian background children were overweight [3]. Differences in risk factors for higher, and increasing, obesity prevalence by ethnicity were not investigated in these studies nor did other large studies of childhood obesity in NSW [4], Victoria [5] and South Australia [6] stratify by ethnic background. Accordingly, we took the opportunity to investigate risk factors associated with obesity in a longitudinal study of 9-year-old Sydney school children from Australian, European, Middle Eastern and South East Asian (SE Asian) ethnic origins, first surveyed in 1994/1995 and subsequently followed up in 1997 at age 12 years. Factors known to be associated with obesity status such as parental obesity, child physical activity levels both inside and outside school and socio-economic status (SES) were investigated.
Methods Study participants Data analysed for this paper originally were collected as part of the ‘‘Sydney aircraft noise & child blood pressure study’’ [7], a longitudinal study that initially surveyed and measured a cohort of 1232 year 3/4 school children attending 75 primary schools randomly sampled from within 20 km of Sydney airport in 1994/1995. At the followup survey 3 years later (1997), measurements were taken from 628 of the baseline participants when they were in year 6 of primary school. The mean ages for the 1994/1995 and 1997 study cohorts were 9.2 ± 0.4 and 11.9 ± 0.3 years, respectively.
Y. Li et al. In line with the current international standards, the study subjects’ obesity status was defined based on their BMI [9]. There has been debate over how to define children’s obesity status but a recent consensus [9] has established criteria which use worldwide data tracked back from adult overweight and obesity standards (9-year-old children: overweight BMI: 19.09—22.78 kg/m2 , obese: BMI ≥ 22.79 kg/m2 ; 12-year-old children: overweight BMI: 21.45—26.34 kg/m2 , obese: BMI ≥ 26.35 kg/m2 ). A similar procedure was followed for children of SE Asian origin. For all adults excepting those of Asian origin, obesity was defined as BMI ≥ 30 kg/m2 and overweight as 25 kg/m2 ≤ BMI < 30 kg/m2 [10]. For Asian subjects a lower cut off BMI for obesity (BMI ≥ 25 kg/m2 ) was used [11], and overweight was defined by a BMI range of 23—24.9 kg/m2 . Longitudinal BMI change from 9 to 12 years was defined as a ‘‘high’’ increase if greater than +2 kg/m2 .
Ethnicity classifications Children’s ethnic origins were defined according to the child’s or parents’ country of birth, in line with the standard classification of countries for social statistics (ASCCSS), issued by Australian Bureau of Statistics (ABS) [12]. A child was classified as having an ‘immigrant background’ if either the child or one or more parent was born outside of Australia. Where a birth country of the child or a parent was in Europe, the Middle East, South East Asia (SE Asia) or any country other than Australia, the child’s ethnic origin was allocated to that ethnic sub-group. The sub-group ‘Australian’ then was composed of children who themselves and both parents were Australian born.
Physical activity measurements Two measures of child physical activity were adopted as indicators of levels of physical activity. First, each study participant was asked if they played actively during their lunch or recess breaks at school; they also were asked if they participated in organised sport outside of school hours.
Overweight/obesity measurements
Socio-economic status (SES)
Child anthropometric measures, including weight and height, were conducted at baseline and follow-up. Heights were measured with socks on by a Harpenden pocket stadiometer [8], and weights were measured with a Tanita System 502 digital bathroom scale accurate to ±100 g [7].
As SES is thought to be associated with obesity, we measured it as both a risk factor and a confounder for ethnicity in this study, with the postcode of each child’s home address used as the SES marker. For the present study, the general index of deprivation, developed by the ABS from a principal
Obesity in Australian migrant children
181
components analysis of key postcode level demographic, wealth and income indicators from the Australian census, was used as the SES measure [13]. ‘‘Low’’ SES level was defined as below the median index score. Individual SES information was not collected from parents themselves in the original study in order to maximise the survey participation rate.
Statistical analysis Descriptive statistics (mean, standard deviation, p-value) and inferential statistics (t-test, oneway ANOVA, chi-square and logistic regression) were performed using SPSS version 13 [14]. Risk estimates with 95% confidence intervals were calculated, reported as odds ratios for cross-sectional data and as relative risks for changes occurring between the initial and 3-year follow-up surveys, using logistic regression [15]. Logistic regression was used to establish the degree of association between the various risk factors and different levels of childhood obesity and obesity/overweight status adjusted for SES and age. For predictors of longitudinal changes in BMI exceeding 2 kg/m2 relative risk estimates were calculated and presented with and without baseline BMI levels included in the regression models.
Results Ethnic differences by overweight/obesity The ethnic composition of the child cohort at age 9 years was 31% Australian born, 20% European, 11% Middle Eastern and 18% of SE Asian origin, with 10% comprising a mixture of various countries of birth, including New Zealand, Pacific Islands and North Asian. A further 10% had no report country of birth information recorded. Similar distributions were found in the children at age 12 (data not shown). As can be seen from Fig. 1, a significantly higher percentage of 9-year-old Middle Eastern and SE Asian children came from low-SES areas compared to Australians (47% and 71%, respectively, vs. 34%; OR = 1.7, 95% CI: 1.2—2.6; OR = 4.7, 95% CI: 3.2—6.7, respectively); accordingly, all subsequent comparisons were adjusted for SES. In addition Fig. 1 shows that SE Asian and Middle Eastern children have significantly less participation in sports compared to Australians or Europeans. At age 9 years children of European and Middle Eastern origin had higher mean BMI measurements
Figure 1 Prevalence of overweight and distribution of risk factors for overweight by ethnic origin, 9-year-old children living in Sydney, Australia.
(mean ± S.D.) compared to Australian children (European: 18.1 ± 3.0 kg/m2 ; Middle Eastern: 18.9 ± 3.4 kg/m2 ; Australian 17.6 ± 3.3 kg/m2 ). Children from Australian backgrounds also had the lowest prevalence of overweight (25%) compared to those from European (31%), SE Asian (33%) and Middle Eastern backgrounds (40%). The prevalence differences were similar in boys and girls, and reflected their mean BMI measurements (boys: Australian 17.7 ± 2.9 kg/m2 , European 18.1 ± 3.0 kg/m2 , Middle Eastern 19.0 ± 3.6 kg/m2 and SE Asian 17.7 ± 3.3 kg/m2 ; girls: Australian 17.7 ± 2.8 kg/m2 , European 18.0 ± 3.1 kg/m2 , 2 Middle Eastern 18.8 ± 3.3 kg/m and SE Asian 17.5 ± 3.4 kg/m2 ).
Predictors of obesity status in children Having an obese parent, being not ethnically Australian, being of low SES and not participating in organised sports were significant predictors of both childhood obesity and overweight/obesity (Table 1). If either parent was obese compared to non-obese, then the odds ratio for a 9-yearold child of being obese compared to normal weight was 4.9 (95% CI: 3.0—8.0); for 12-year olds the odds ratio was 8.0 (95% CI: 3.6—18). Similar odds ratios were seen for overweight and obese (Table 1). For the survey of 9-years old, a parent born outside Australia or if the child himself was born outside Australia was associated with an almost twofold chance of being obese (OR = 1.8, 95% CI: 1.1—3.0), and not participating in organised sports was significantly associated with childhood obesity. Nine-year olds who did not
182
Table 1
Risk factors for childhood obesity status by ethnic origin in 9-year-old children living in Sydney, Australia (OR, 95% CI)
Risk factors
Obesity and overweight risk for children aged 9 years Total (n = 1232), ORa (95% CI)
Australian (n = 381), ORa (95% CI)
European (n = 251), ORa (95% CI)
Middle Eastern (n = 132), ORa (95% CI)
SE Asian (n = 216), ORa (95% CI)
Non-Australian born No Yesb Yesc
1.0 1.3 (1.0—1.7) 1.8 (1.1—3.0)
— — —
— — —
— — —
— — —
Maternal obesity No Yesb Yesc
1.0 2.9 (1.8—4.8) 4.6 (2.5—8.6)
1.0 2.8 (1.3—6.1) 5.0 (1.6—16)
1.0 4.2 (1.4—12.4) 6.1 (1.6—24)
1.0 1.7 (0.5—5.6) 2.5 (0.6—10.5)
1.0 5.8 (2.4—14.1) 8.2 (3.0—23)
Paternal obesity No Yesb Yesc
1.0 2.6 (1.7—4.1) 4.4 (2.5—7.8)
1.0 2.6 (1.2—5.8) 12.7 (4.2—39)
1.0 2.5 (1.0—5.8) 3.5 (1.0—12.6)
1.0 3.0 (1.0—9.6) 4.3 (1.1—17)
1.0 3.9 (2.0—7.4) 6.7 (2.9—15)
Parental obesity No Yesb Yesc
1.0 2.9 (2.1—4.2) 4.9 (3.0—8.0)
1.0 3.2 (1.7—5.9) 13.6 (5.0—37)
1.0 2.9 (1.4—5.9) 4.6 (1.7—12.6)
1.0 3.0 (1.2—7.5) 4.1 (1.2—14)
1.0 3.7 (2.0—6.8) 5.7 (2.6—13)
Low SES No Yesb Yesc
1.0 1.1 (0.8—1.4) 1.2 (0.8—1.9)
1.0 1.0 (0.6—1.7) 0.7 (0.3—1.8)
1.0 0.9 (0.5—1.6) 1.0 (0.4—2.8)
1.0 2.2 (1.0—4.5) 1.8 (0.7—5.0)
1.0 0.8 (0.5—1.6) 1.2 (0.5—2.8)
In-school activity Yes Nob Noc
1.0 1.5 (1.1—2.1) 1.6 (1.0—2.8)
1.0 1.1 (0.6—2.1) 1.1 (0.4—3.5)
1.0 2.3 (1.1—4.7) 1.6 (0.5—5.2)
1.0 1.0 (0.4—2.6) 1.3 (0.4—4.3)
1.0 1.6 (0.7—3.6) 1.9 (0.7—5.2)
Organised sports Yes Nob Noc
1.0 1.2 (0.9—1.6) 1.9 (1.2—2.9)
1.0 1.1 (0.6—1.8) 2.3 (1.0—5.7)
1.0 1.4 (0.8—2.5) 1.4 (0.6—3.6)
1.0 1.3 (0.6—2.6) 2.6 (0.9—7.6)
1.0 0.9 (0.5—1.5) 0.8 (0.4—1.8)
a b c
Adjusted for age and SES. OR (95% CI) for overweight. OR (95% CI) for obesity.
Y. Li et al.
Obesity in Australian migrant children participate in organised sports outside of school hours were almost twice as likely to be obese (OR = 1.9, 95% CI: 1.2—2.9); this association with obesity was particularly evident in Australian and Middle Eastern children. Notable was that lack of in-school activity was associated with risks for overweight and obesity in SE Asian and European children. When adjusted for age and SES, children from Middle Eastern background had double the odds of being overweight compared to their Australian peers (OR = 2.0, 95% CI: 1.3—3.0), and SE Asian children 1.6 the odds (95% CI: 1.1—2.3). The odds were over threefold for obesity (Middle Eastern OR = 3.2, 95% CI: 1.7—6.3; SE Asian OR = 3.2, 95% CI: 1.8—5.9) (Table 2). These differences were also evident when stratified by gender except there was no significant difference in odds of being overweight for SE Asian girls compared to their Australian counterparts. Middle Eastern children had twice the adjusted odds of having an obese parent in comparison with their Australian counterparts (26% vs. 15%; OR = 2.1, 95% CI: 1.3—3.6), and also had nearly triple the odds of not attending organised sports activity outside school hours (52% vs. 26%; OR = 2.8, 95% CI: 1.8—4.3) than Australian children. In comparison, SE Asian children had three times the odds (adjusted for SES and age) of having an obese parent compared to their Australian peers (37% vs. 15%; OR = 3.4, 95% CI: 2.2—5.2). In addition, there was a significantly greater percentage of 9-yearold SE Asian children not participating in organised sports activities outside school than that of the Australian children (26% vs. 61%; OR = 3.9, 95% CI: 2.7—5.7).
Associations between ethnic background and risk factors for childhood obesity status Parental obesity at baseline was consistently the strongest predictor of obesity/overweight in children across all ethnic groups, after adjusting for SES and age (Table 1). The association with obesity ranged from fourfold in Middle Eastern children to fourteen fold in Australian children, however the associations with overweight and obesity were consistently threefold over all ethnic groups. It appears that maternal obesity was a stronger risk factor in SE Asian children obesity status whereas paternal obesity affected Middle Eastern and Australian children’s obesity status. The other predictors of interest were a twofold difference in risk of obesity with lack of in-school activity in total population and SE Asian children (OR = 1.6, 95% CI: 1.0—2.8) for total population and (OR = 1.9, 95% CI:
183 0.7—5.2) for SE Asians. There were similar twofold risks for no participation in organised sports for Australian and Middle Eastern children’s obesity status (OR = 2.3, 95% CI: 1.0—5.7) for Australians and (OR = 2.6, 95% CI: 1.0—7.6) for Middle Eastern.
Longitudinal BMI change Similar risk factors were associated with high-BMI increase (BMI increase >2 kg/m2 ) independent of age and SES over the 3-year period from age 9 to 12 years (Table 3 ). Parental obesity was the major risk factor for high-BMI increase across all ethnic groups except for Middle Eastern children (RRs: all 2.2, 95% CI: 1.3—3.6; Australian 3.0, 95% CI: 1.2—7.2; European 2.7, 95% CI: 1.0—7.3; SE Asian 3.3, 95% CI: 1.2—9.1). The only other significant predictors for high increase in BMI over 3 years was being from a low-socio-economic area or not attending organised sports after school hours, in Australian children only (RRs: 1.9, 95% CI: 1.0—3.6; 2.4, 95% CI: 1.2—4.8, respectively). However, as can be seen in Table 3, if baseline BMI was also accounted for, parental obesity status no longer remained as a predictor of 3-year longitudinal BMI change from age 9 to 12 in our study cohort. However, baseline BMI was significantly associated with high-BMI change (Table 3), and parental obesity significantly associated with baseline BMI, suggesting a conferring of high BMI on the child in the first place. Being from a low-socio-economic area or not attending organised sports outside school hours, nevertheless, were still significantly associated with BMI increase over 3 years in the Australian children only (RR: 1.9, 95% CI: 1.0—3.9; 2.7, 95% CI: 1.3—5.7, respectively).
Discussion This study has demonstrated that cross-sectionally having an obese parent, not being born in Australia and having either parent not being born in Australia, and not being involved in physical activity were strong predictors of childhood obesity at 9 years old, independent of SES and age. However parental obesity was not related to longitudinal change in obesity after adjustment for initial BMI. This indicates that BMI tracking is strong, since baseline BMI was related to parental BMI, and, established quite early, appears as the mechanism for further BMI increase. Both the prevalence of and factors associated with obesity varied by ethnic background. Higher participation in organised
184
Y. Li et al.
Table 2 Prevalence of overweight/obesity (BMI) by ethnicity, gender and their comparison with Australian counterparts in 9-year-old children living in Sydney, Australia (OR, 95% CI) Ethnic origin by country of birth Australian European Middle Eastern SE Asian a
Overweight
Obesity a
Boys + girls, ORa (95% CI)
Boys, OR (95% CI)
1.0 1.3 (0.9—1.9) 2.0 (1.3—3.0) 1.6 (1.1—2.3)
1.0 1.2 (0.8—2.1) 2.2 (1.1—4.0) 1.7 (1.0—3.0)
Girls, OR (95% CI)
a
1.0 1.5 (0.9—2.4) 1.8 (1.0—3.3) 1.4 (0.8—2.4)
Boys + girls, ORa (95% CI)
Boys, ORa (95% CI)
Girls, ORa (95% CI)
1.0 1.6 (0.9—3.1) 3.2 (1.7—6.3) 3.2 (1.8—5.9)
1.0 1.5 (0.7—3.5) 3.2 (1.3—7.9) 2.9 (1.3—6.6)
1.0 1.9 (0.7—4.9) 3.3 (1.2—9.0) 3.2 (1.4—7.7)
Adjusted for age and SES.
sport was related more strongly to lower obesity status in children of Australian, Middle Eastern or European backgrounds, which may be because SE Asian children had very limited participation in organised sport compared to their counterparts. The predictors of childhood obesity found here parallel the findings of three other Australian cohort studies [4—6] which also found parental obesity status to be a risk factor for children’s obesity but without breakdowns by ethnicity. In the absence of studies on immigrant parental obesity, it is helpful to compare the present findings with those that have been conducted within the countries of origin, particularly studies of European [16,17] and SE Asian [18] populations, that have linked child obesity status to parental obesity status. To our knowledge one Korean study [18] has reported the risk of being obese in children aged 7—12 years (n = 3059), as 1.9 (95% CI: 1.3—2.8) for those whose mothers were obese, 1.5 (95% CI: 1.3—1.7) for those whose fathers were obese, and 2.4 (95% CI: 1.8—3.1) if either parent was obese. Somewhat unexpectedly in the present study, children’s socio-economic level was a significant risk factor for obesity only in Middle Eastern children, despite consistent reports from the literature showing SES to be related to obesity [19]. It was because of this, a priori, that all the data in this paper were adjusted for socio-economic status. Physical activity level of children, represented in our study by a question on playing activity in the child’s recess or lunch breaks in school, and by participation in sport outside school, also proved to be significantly associated with obesity in our study cohort. This is supported by previous research [19,20]. SE Asian children living in Australia in our study tended to have relatively high-participation rates in in-school activity (boys: 96%; girls: 79%) but not after-school organised sports (boys: 42%; girls: 36%) compared to those living in China [21], as well
as in Chinese immigrants living in US [22]. However these rates are low compared to Australian children in our study (in-school activity: boys 93%, girls 74%; after-school organised sports: boys 77%, girls 72%). In summary, despite there being some 12 cross-sectional [2,3,23—30] and 6 cohort studies [4—6,31—33] investigating childhood obesity in Australia over the past three decades indicating a rise in obesity rates, to our knowledge the only studies that have analysed overweight or obesity prevalence in terms of ethnicity were the 1997 NSW Schools Physical Activity and Nutrition Survey [2] and the 2004 NSW Schools Physical Activity and Nutrition Survey (SPANS) [3]. Neither investigated risk factors associated with different ethnic backgrounds. Also, in these studies, study participants were only categorized according to language spoken most at home, in contrast, to the present study where ethnicity was defined by both parental and children’s country of birth. The strengths of this study are that the children were taken from a very well-designed random sample of inner city Sydney schools and there was accurate data on both their country of birth and their parents’ country of birth. The total sample had sufficient power to be able to stratify by ethnic group to obtain reasonable statistical estimates. And child weights and heights were measured according to a consistent protocol and were not self reported. The main study limitations were that there were very limited nutritional and physical activity data. Also, there was a substantial drop out on follow-up (49%) which possibly could lead to bias in analysis of obesity change. This study was voluntary and thus it is not surprising that although the study children were from randomly selected schools, the response rate was only 40% and there was almost 50% attrition between the baseline and follow-up surveys. However, as study children were compared with themselves in the cohort analy-
Risk factors
Risk factors for increase in BMI (>2 kg/m2 ) over 3 years by ethnic origin in 9—12-year children (RR, 95% CI) BMI increase (>2 kg/m2 from age 9 to 12) Total (n = 586) RRa (95% CI)
Australian (n = 187)
1.3 (1.2—1.4)
RRa (95% CI) 1.4 (1.2—1.6)
Overweight/obese (age 9) No 1.0 Yes 2.7 (1.8—4.2)
1.0 3.5 (1.7—7.4)
BMI (age 9)
RRb (95% CI)
European (n = 132) RRb (95% CI)
RRa (95% CI) 1.2 (1.0—1.4)
Middle Eastern (n = 68) RRb (95% CI)
1.0 2.5 (1.1—5.9)
RRa (95% CI) 1.2 (1.0—1.4)
SE Asian (n = 85) RRb (95% CI)
1.0 1.0 (0.3—2.9)
RRa (95% CI) 1.8 (1.3—2.3)
RRb (95% CI)
Obesity in Australian migrant children
Table 3
1.0 13.4 (3.8—47.8)
Maternal obesity No 1.0 Yes 2.6 (1.3—5.2)
1.0 1.5 (0.7—3.2)
1.0 3.0 (1.0—8.5)
1.0 1.4 (0.4—4.7)
1.0 5.0 (0.9—26.4)
1.0 5.0 (0.9—28.2)
1.0 0.4 (0.04—4.1)
1.0 0.2 (0.02—2.9)
1.0 2.3 (0.6—9.0)
1.0 0.5 (0.1—2.9)
Paternal obesity No 1.0 Yes 1.9 (1.0—3.5)
1.0 1.3 (0.7—2.4)
1.0 4.1 (1.2—13.8)
1.0 2.7 (0.7—9.9)
1.0 1.6 (0.5—5.3)
1.0 1.2 (0.3—4.4)
1.0 1.2 (0.3—4.7)
1.0 1.0 (0.2—4.2)
1.0 3.3 (1.1—9.8)
1.0 1.7 (0.5—6.1)
Parental obesity No 1.0 Yes 2.2 (1.3—3.6)
1.0 1.4 (0.8—2.4)
1.0 3.0 (1.2—7.2)
1.0 1.6 (0.6—4.3)
1.0 2.7 (1.0—7.3)
1.0 2.4 (0.9—6.7)
1.0 1.1 (0.3—3.8)
1.0 0.8 (0.2—3.2)
1.0 3.3 (1.2—9.1)
1.0 1.9 (0.6—6.3)
Low SES No Yes
1.0 1.3 (0.9—1.8)
1.0 1.4 (0.9—2.0)
1.0 1.9 (1.0—3.6)
1.0 1.9 (1.0—3.9)
1.0 0.9 (0.4—2.0)
1.0 1.0 (0.5—2.2)
1.0 0.5 (0.2—1.5)
1.0 0.5 (0.2—1.4)
1.0 1.1 (0.4—2.7)
1.0 1.9 (0.6—6.0)
No in-school activity No 1.0 Yes 1.4 (0.9—2.3)
1.0 0.6 (0.3—1.1)
1.0 1.5 (0.7—3.3)
1.0 0.9 (0.3—2.9)
1.0 0.6 (0.2—1.9)
1.0 0.3 (0.1—1.3)
1.0 2.0 (0.6—6.8)
1.0 0.1 (0.01—1.4)
1.0 1.9 (0.6—6.2)
1.0 0.9 (0.2—4.7)
No organised sports No 1.0 Yes 1.3 (0.9—1.9)
1.0 1.4 (0.9—2.0)
1.0 2.4 (1.2—4.8)
1.0 2.7 (1.3—5.7)
1.0 0.8 (0.4—1.8)
1.0 0.7 (0.3—1.6)
1.0 0.8 (0.3—2.1)
1.0 0.8 (0.3—2.2)
1.0 1.2 (0.5—3.2)
1.0 1.4 (0.5—4.4)
a and SES. b Adjusted for age, SES and baseline BMI.
185
186
Y. Li et al.
sis, factors associated with progression to obesity would still be revealed. In addition, the differences between those lost to follow-up vs. those followed up pertaining to the study factors was in ethnic distribution (lower proportion of SE Asians at follow-up) higher mean of BMI, waist circumference and body fat in the children (but not in obesity in their parents) and less sports (PA) activity outside school (Appendix A). If we predicted any bias from these differences we would expect the association with parental obesity to be even stronger if there was no attrition, i.e., a bias toward the null with respect to parental obesity. It also should be noted that, all our findings at follow-up were consistent with the baseline findings.
in these different groups. As evidence from our data support a relationship between obesity levels and in-school activity in those from SE Asian and Middle Eastern backgrounds, we recommend that special programs be established in all schools with high-minority group populations, to involve children in school-break physical activity. We also believe that as there is large ethnic variation in participation in sports outside schools that ethnospecific educational programs should be established and sponsored by community leaders to underline for parents the importance of physical activity outside school for their children. In addition, ethnospecific events both in and outside of school should be encouraged to incorporate physical activity components.
Conclusion and recommendations Conflicts of interest statement In conclusion, more research into ethnic and migrant variation in obesity in larger cohorts in both Australia and other countries with large immigrant populations such as Great Britain, Canada and the USA. In addition and in parallel there needs to be in depth studies into ethnic variation of metabolic responses to exercise and diet
Risk factor for obesity in children (%)
Appendix A. Risk factor data on those who did and did not respond to 3-year follow-up
Total population (n = 1232) Follow-up (n = 628) (51%)
Parental obesity (%) Australian (%) Middle Eastern (%) SE Asian (%) European (%) PA outside school (%) PA in school (%) Low SES (%) Female (%) Age (years), mean ± S.D. Birth weight, mean ± S.D. SBP, mean ± S.D. DBP, mean ± S.D. BMI, mean ± S.D. WC, mean ± S.D. %TBF, mean ± S.D.
There are no financial and conflict of interest for any of the authors.
54 54 54 44 55 64 84 55 49 9.2 ± 0.4 3.3 ± 0.6 111 ± 10 59 ± 8 17.7 ± 2.9 61 ± 8 10 ± 6
Lost to follow-up (n = 604) (49%) 46 46 46 56 46 56 84 61 50 9.2 ± 0.4 3.3 ± 0.6 112 ± 10 60 ± 8 18.1 ± 3.2 62 ± 9 11 ± 6
Significance (p) ns ns ns 0.06 ns <0.05 ns 0.06 ns ns ns ns ns <0.05 <0.05 <0.05
Obesity in Australian migrant children
References [1] Crawford PB, Story M, Wang MC, Ritchie LD, Sabry ZI. Ethnic issues in the epidemiology of childhood obesity. Pediatr Clin North Am 2001;48:855—78. [2] Booth ML, Wake M, Armstrong T, Chey T, Hesketh K, Mathur S. The epidemiology of overweight and obesity among Australian children and adolescents, 1995—97. Aust N Z J Public Health 2001;25:162—9. [3] Booth ML, Okely A, Wilson ED, Hardy L. NSW Schools Physical Activity and Nutrition Survey 2004 (SPANS). Sydney: NSW Health; 2006. [4] Garnett SP, Cowell CT, Baur LA, Shrewsbury VA, Chan A, Crawford D, et al. Increasing central adiposity: the Nepean longitudinal study of young people aged 7—8 to 12—13 y. Int J Obes (Lond) 2005;29:1353—60. [5] Hesketh K, Wake M, Waters E, Carlin J, Crawford D. Stability of body mass index in Australian children: a prospective cohort study across the middle childhood years. Public Health Nutr 2004;7:303—9. [6] Magarey AM, Daniels LA, Boulton TJ, Cockington RA. Predicting obesity in early adulthood from childhood and parental obesity. Int J Obes (Lond) 2003;27:505—13. [7] Morrell S. Aircraft noise and child blood pressure. Sydney: School of Public Health, University of Sydney; 2003. p. 568. [8] Holtain Ltd., Crosswell, UK. [9] Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. Br Med J 2000;320:1240—3. [10] Himes JH, Dietz WH. Guidelines for overweight in adolescent preventive services: recommendations from an expert committee. The Expert Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services. Am J Clin Nutr 1994;59:307—16. [11] Inou S, Zimmet P, Shah B, Marmot MG. The Asia-Pacific perspective: redefining obesity and its treatment. World Health Organisation, International Association for the Study of Obesity, International Obesity Task Force; 2000. [12] Castles I. Australian standard classification of countries for social statistics. Australian Bureau of Statistics; 1990. p. 143. [13] Castle I. Information paper: 1991 census—–socio-economic indexes for areas. Canberra: Australian Bureau of Statistics; 1993. p. 38. [14] Coakes SJ, Steed LG, Dzidic P. SPSS version 13.0 for windows: analysis without anguish. Milton, Qld: John Wiley & Sons Australia; 2005. [15] Menard SW. Applied logistic regression analysis. Thousand Oaks, CA: Sage Publications; 1995. [16] Savva SC, Kourides Y, Tornaritis M, Epiphaniou-Savva M, Chadjigeorgiou C, Kafatos A. Obesity in children and adolescents in Cyprus. Prevalence and predisposing factors. Int J Obes (Lond) 2002;26:1036—45. [17] De Vito E, La Torre G, Langiano E, Berardi D, Ricciardi G. Overweight and obesity among secondary school children in Central Italy. Eur J Epidemiol 1999;15:649—54. [18] Lee K, Kwon ER, Park TJ, Park MS, Lenders CM. Parental overweight as an indicator of childhood overweight: how sensitive? Asia Pac J Clin Nutr 2006;15:196—200.
187 [19] Moussa MA, Skaik MB, Selwanes SB, Yaghy OY, Bin-Othman SA. Factors associated with obesity in school children. Int J Obes (Lond) 1994;18:513—5. [20] McMurray RG, Harrell JS, Deng S, Bradley CB, Cox LM, Bangdiwala SI. The influence of physical activity, socioeconomic status, and ethnicity on the weight status of adolescents. Obes Res 2000;8:130—9. [21] Tudor-Locke C, Ainsworth BE, Adair LS, Du S, Popkin BM. Physical activity and inactivity in Chinese school-aged youth: the China Health and Nutrition Survey. Int J Obes (Lond) 2003;27:1093—9. [22] Gordon-Larsen P, McMurray RG, Popkin BM. Adolescent physical activity and inactivity vary by ethnicity: the National Longitudinal Study of Adolescent Health. J Pediatr 1999;135:301—6. [23] Timperio A, Salmon J, Telford A, Crawford D. Perceptions of local neighbourhood environments and their relationship to childhood overweight and obesity. Int J Obes (Lond) 2005;29:170—5. [24] Rehor PR, Brownsey E. A preliminary investigation into the prevalence and trends of overweight and obesity of northern Tasmanian primary school children. Aust N Z J Public Health 2002;26:479—80. [25] O’Dea JA. Differences in overweight and obesity among Australian schoolchildren of low and middle/high socioeconomic status. Med J Aust 2003;179:63. [26] Wilcken DE, Lynch JF, Marshall MD, Scott RL, Wang XL. Relevance of body weight to apolipoprotein levels in Australian children. Med J Aust 1996;164:22—5. [27] Harvey PWJ, Althaus M-M. The distribution of body mass index in Australian children aged 7—15 years. Aust J Nutr Diet 1993;50:151—3. [28] Wickramasinghe VP, Cleghorn GJ, Edmiston KA, Murphy AJ, Abbott RA, Davies PSW. Validity of BMI as a measure of obesity in Australian White Caucasian and Australian Sri Lankan children. Ann Hum Biol 2005;32:60—71. [29] Vincent SD, Pangrazi RP, Raustorp A, Tomson LM, Cuddihy TF. Activity levels and body mass index of children in the United States, Sweden, and Australia. Med Sci Sports Exerc 2003;35:1367—73. [30] Kremer PJ, Bell AC, Sanigorski AM, Swinburn BA. Overweight and obesity prevalence in children based on 6- or 12-month IOTF cut-points: does interval size matter? Int J Obes (Lond) 2006;30:603—5. [31] Burke V, Beilin LJ, Dunbar D. Family lifestyle and parental body mass index as predictors of body mass index in Australian children: a longitudinal study. Int J Obes (Lond) 2001;25:147—57. [32] Burke V, Beilin LJ, Simmer K, Oddy WH, Blake KV, Doherty D, et al. Predictors of body mass index and associations with cardiovascular risk factors in Australian children: a prospective cohort study. Int J Obes (Lond) 2005;29:15— 23. [33] Mamun AA, Lawlor DA, O’Callaghan MJ, Williams GM, Najman JM. Family and early life factors associated with changes in overweight status between ages 5 and 14 years: findings from the Mater University Study of Pregnancy and its outcomes. Int J Obes (Lond) 2005;29:475— 82.
Available online at www.sciencedirect.com