International Journal of Obesity (2005) 29, 15–23

& 2005 Nature Publishing Group All rights reserved 0307-0565/05 $30.00
www.nature.com/ijo

PAPER
Predictors of body mass index and associations with
cardiovascular risk factors in Australian children:
a prospective cohort study
V Burke1*, LJ Beilin1, K Simmer2,3, WH Oddy4, KV Blake2, D Doherty2,3, GE Kendall4,
JP Newnham2,3, LI Landau5 and FJ Stanley4
1
School of Medicine and Pharmacology, The University of Western Australia, Royal Perth Hospital and Western Australian
Institute for Medical Research, Perth, Australia; 2The University of Western Australia, School of Women’s and Infants’
Health, Australia; 3Women’s and Infants’ Research Foundation, Australia; 4Telethon Institute for Child Health Research
and Centre for Child Health Research, School of Paediatrics and Child Health, University of Western Australia; and 5The
University of Western Australia, Faculty of Medicine and Dentistry, Australia

OBJECTIVE: To examine predictors of body mass index (BMI) at the age of 8 y in a prospective study of Australian children.
DESIGN: Longitudinal survey of a cohort of Australian children followed from the 16th week of gestation to 8 y.
SUBJECTS: In total, 741 boys and 689 girls who attended the survey as 8 y olds.
MEASUREMENTS: Weight and height, blood pressure measured by automated oscillometry, fasting blood lipids and glucose.
Questionnaire assessment of activity and diet.
RESULTS: Proportions of overweight including obesity in boys and girls were, respectively, 22 and 25% at 1 y, 14 and 14% at
3 y, 13 and 18% at 5 y and 15 and 20% at 8 y. At the age of 1, 3, 6 and 8 y, children with overweight including obesity showed
significantly more adverse cardiovascular risk factors. Blood pressure (BP) was significantly higher by 2/3 mmHg (systolic/
diastolic) at 1 y, 3/2 mmHg at 3 y, 4/2 mmHg at 5 y and 6/2 mmHg at 8 y; HDL was significantly lower (P ¼ 0.002) by 8% and
triglycerides were significantly higher by 27% (Po0.001). In multivariate regression, BMI at the age of 8 y was significantly
predicted positively by birth weight, mother’s BMI and hours spent in watching television at the time of the survey of 6 y olds.
Mothers being ex-smokers or non smokers and children being ‘slightly active’ and ‘active’ negatively predicted BMI in 8 y olds.
In a subset of 298 children with information about fathers, paternal BMI was an additional independent predictor. Maternal or
paternal overweight including obesity each independently increased risk of overweight including obesity at the age of 8 y three-
fold. A food factor with consumption of cereals and breads as the major components derived from a Food Frequency
Questionnaire in a subset of 340 children was also an independent negative predictor of BMI in multivariate models.
CONCLUSION: The increasing rate of overweight including obesity, particularly in girls, is associated with an increase in
cardiovascular risk factors very early in life. Improvement of health-related behaviours within the family and a focus on
promotion of activity in children should be priorities in achieving weight control.
International Journal of Obesity (2005) 29, 15–23. doi:10.1038/sj.ijo.0802750
Published online 17 August 2004

Keywords: body mass index; weight; blood pressure; blood lipids; birth weight; children

Introduction life. 2–4 The worldwide increase in the prevalence of obesity

Obesity in childhood adversely affects children’s health and
well-being,1 and leads to serious consequences in the longer
term because of tracking of overweight and obesity to adult
*Correspondence: Dr V Burke, School of Medicine and Pharmacology,
The University of Western Australia, Royal Perth Hospital Unit, Box X2213
GPO, Perth 6847, Australia. E-mail: vburke@cyllene.uwa.edu.au
Received 22 February 2004; revised 1 June 2004; accepted 11 June 2004;
published online 17 August 2004
and overweight in children and adults5 implies a substantial
increase in life style diseases such as Type II diabetes,
hypertension and cardiovascular disease.6 Addressing the
factors leading to obesity and its associated morbidities has
become a major public health concern.1
Temporal trends for an increase in the proportion of
overweight and obese children and adolescents have been
reported widely,1,7–9 affecting even 4–5 y olds.10 Population
surveys among Australians aged 7–15 y showed that the
 BMI and cardiovascular risk factors in children
V Burke et al
16
prevalence of overweight including obesity doubled between
1985 and 1997;11 Lazarus et al12 estimated that body mass
index (BMI) had increased by more than 1 kg/m2 in 7 to
12-y-old Australians in that interval. Magarey et al13 applied
the new international standards for overweight and obesity
in children,14 using data from the cross-sectional Australian
Health and Fitness Survey from 1985 and the National
Nutrition Survey carried out in 1995. These authors reported
that about 11% of boys and 12% of girls were overweight or
obese in 1985, but the proportion had increased to about
19% of boys and 21% of girls in the 1995 survey.
Longitudinal studies have also shown increasing
proportions of overweight and obesity with age in chil-
dren.15–18 In a longitudinal study of a cohort of children
in Perth, Western Australia, recruited at the age of 9 y
and followed to 25 y of age, rates of overweight including
obesity increased in males from 9% in 9-y-old boys to 42%
in 25-year-old men and, in females, from 8 to 32%.19 We
also found that, even in 9 y olds, blood pressure (BP) was
significantly higher in the overweight or obese group and,
by the age of 25 y, 53% of those men and 32% of the
women were either hypertensive or were at risk of hyperten-
sion. Our findings in that prospective study, with tracking
of BMI from childhood through young adult life and clear
associations with cardiovascular risk as indicated by BP,
highlight the importance of addressing weight control early
in life.
We now report findings in a cohort of younger Australian
children, followed from the 16th week of gestation to 8 y of
age, using prospectively collected data to examine predictors
of overweight and obesity, associations with risk factors for
cardiovascular disease and tracking of BMI from birth.
Methods
The Western Australian Pregnancy Cohort Study subjects
(‘Raine cohort’) were serially recruited from the public
antenatal clinic at King Edward Memorial Hospital (KEMH),
or nearby private practices in Perth, Western Australia
between 1989 and 1992.20 The study started as a pregnancy
cohort in which 2979 women were enrolled at or before the
18th week of gestation from the antenatal booking clinics at
KEMH. Approximately 100 women per month were enrolled
for a total of 30 months commencing in May 1989 and
finishing in November 1991. The criteria for enrolment were
gestational age between 16 and 20 weeks, sufficient profi-
ciency in English to understand the implications of partici-
pation, an expectation to deliver at KEMH and an intention
to remain in Western Australia so that follow-up through
childhood would be possible. All mothers gave written
informed consent and the study was approved by the
institutional ethics committees. The initial cohort comprised
2860 live births. The present analysis included the 2087
children remaining after exclusion of multiple births, those
born before 37 weeks of gestation and those with congenital
abnormalities.
International Journal of Obesity
Children were resurveyed at the age of 1, 3, 6 and 8 y.
Weight was measured to the nearest 100 g using Wedderburn
Digital Chair Scales with children wearing only under-
clothes, and height to the nearest 0.1 cm with a Holtain
Stadiometer. At 45 min after entering the room, the child
had BP measured, seated, using the appropriate cuff size with
an oscillometric method (Dinamap 8100, Critikon, USA).
Two readings were taken 1 min apart and the average used in
analysis.
Total LDL and HDL cholesterol and glucose were measured
at the age of 8 y, after an overnight fast, by the Core Laboratory
at Royal Perth Hospital using a Hitachi 917 autoanalyser.
At the age of 8 y, a short questionnaire was used as an
indicator of fat consumption in the whole cohort.21 In a
subset of 340 children, food frequency questionnaires (FFQ)
that comprised 222 items were completed by parents in
association with the child (CSIRO, Adelaide, South Australia)
and provided data used for calculation of the child’s intake of
nutrients.
In the survey of 6 y olds, questions addressed the
following: spare time activities categorised as outdoor play,
reading, building blocks, television watching, video games,
dolls, pretending, drawing/painting and others; time spent
in watching TV; and involvement in organised sport (yes/
no). In 8 y olds, questions addressed qualitative assessment
of activity as sedentary, slightly active or active; involvement
in organised sport (yes/no); and time spent in watching TV.
Birth weight was obtained from midwives’ records and the
occurrence of gestational diabetes from antenatal records.
The duration of breastfeeding and the age at which other
milk was introduced were determined from responses at the
time of the surveys of 1 and 3 y olds. Mothers were also asked
‘at what age did you stop breastfeeding?’ (in months) and ‘at
what age did you first give your child any milk other than
breast milk?’ (in months) when children were surveyed as 1
and 3 y olds.
At the time of the survey of 8 y olds, mothers had weight
measured using electronic scales and height by a stadi-
ometer. They also provided information about current
smoking status (never smoker, ex-smoker, current smoker)
and medical history, including whether they had been
diagnosed as diabetic. Mothers’ educational levels were
obtained by questionnaire at the time of delivery and again
when children were 8 y old. Educational level was cate-
gorised into three groups: education not higher than
secondary schooling; and technical qualification or
diploma; and tertiary qualification. The questionnaire also
recorded family annual income in four categories appro-
priate to the year of the survey. When the children were
8 y old, for example, the categories were: o$25 000; $25 001–
$40 000; $40 001–$60 000; and 4$60 000. Income was
then grouped into two categories with the two upper
and two lower ranges aggregated. Data were sought
from fathers, but paternal weight and height were obtained
from only 298 fathers whose children attended the survey
as 8 y olds.
 BMI and cardiovascular risk factors in children
V Burke et al
17
Statistical methods Table 1 Characteristics of boys and girls attending the survey as 8 y olds
SPSS 11.5. (SPSS, Chicago, IL, USA) was used for analyses.
Boys, N ¼ 741 Girls, N ¼ 689 P-value
Descriptive statistics are presented as mean and s.d. for
continuous variables. In 1 y olds, overweight or obesity was Height (cm) 129.6 (5.9) 128.4 (5.9) o0.001
defined according to the 85th and 95th centiles of the Weight (kg) 28.4 (5.7) 28.0 (5.7) 0.173
Systolic BP 104 (10) 104 (10) 0.384
NCHS/CDC growth charts.22 For 3, 6 and 8 y olds, the Diastolic BP 56 (6.6) 56 (6.6) 0.583
international standards for overweight and obesity in Blood glucose 4.6 (0.4) 4.4 (0.4) o0.001
children derived by Cole et al14 were used with the cutpoint Total cholesterol 4.4 (0.7) 4.6 (0.8) o0.001
nearest to the child’s age. Repeated measures analysis of LDL 2.5 (0.6) 2.8 (0.8) 0.001
HDL 1.54 (0.30) 1.53 (0.35) 0.775
variance with adjustment for sex was used to examine TG 0.59 (0.56, 0.62) 0.64 (0.60, 0.67) 0.065
associations between BMI and socioeconomic status (SES) as Activity
indicated by family income or maternal education level. Sedentary 68 (9%) 93 (14%) o0.001
Principal components’ analysis with varimax rotation23 was Slightly active 332 (45%) 369 (54%)
Active 334 (46%) 214 (32%)
used to identify factors that described dietary patterns
derived from the FFQ. Factor scores were subsequently used Hours/day watching television 2.8 (0.9) 2.7 (1.0) 0.099
as independent variables in regression models. Regression Family income
analysis was used for modelling with BMI at the age of 8 y as r$40 000 303 (43%) 315 (48%) 0.032
4$40 000 409 (57%) 342 (52%)
the dependent variable. Initial models were examined with Maternal education
adjustment only for sex and all variables that attained a P- Secondary 310 (43%) 325 (50%) 0.090
value o0.1 were included in subsequent multivariable Diploma/other 251 (35%) 204 (31%)
analysis, with additional adjustment for maternal education, Tertiary 158 (22%) 127 (19%)

using backward selection of variables. Binary logistic regres-

sion was used to determine predictors of overweight
including obesity in 8 y olds in models adjusted for sex and
maternal education. Pearson’s correlations were used to
examine associations between weight at birth and at 1, 3, 6
and 8 y as an indication of tracking. Results were considered
significant if Po0.05.
Results
After exclusion of multiple pregnancies, congenital abnorm-
alities and children born before 37 weeks of gestation, the
present analysis included 1430 children (741 boys, 689 girls)
in whom weight and height were measured at the age of 8 y.
Earlier surveys with weight and height available comprised
1710 children at 1 y, 1184 at 3 y and 1480 at 6 y. The
characteristics of the 8 y olds are shown in Table 1.
Statistically significant differences were seen in a higher rate
of sedentary behaviour among girls (Po0.001) who also
showed higher BMI, total cholesterol and LDL cholesterol
(Po0.001). Time spent watching television did not differ
significantly between boys and girls at the age of 8 y.
However, there were significant differences associated with
classification of activity levels with a mean of 3.2 (0.1) h in
those considered to be sedentary, 2.8 (0.1) h in the slightly
active group and 2.6 (0.1) h in those regarded as active
(Po0.001).
Associations with SES
Figure 1 shows the BMI at the ages of 1, 3, 6 and 8 y
according to family income. BMI was similar in both income
groups to the age of 3 y, but in 6 and 8 y olds, BMI was higher
in families with lower income (RANOVA P ¼ 0.004). At 1 y,
22.2% were overweight including obese (10.5% obese) in the
lower income group and 24.6% (10.3% obese) in the upper
income group (P ¼ 0.323). Respective proportions were
13.5% (1.4% obese) and 14.1% (1.4% obese) at 3 y
(P ¼ 0.955), 18.0% (5.4% obese) and 12.8% (3.0% obese) at
6 y (P ¼ 0.020) and 20.6% (5.6% obese) and 15.2% (2.6%
obese) at 8 y (P ¼ 0.007). Findings were similar in relation to
maternal education (Figure 1), with lower BMI by the age of
6 y in children of mothers with tertiary education (RANOVA
P ¼ 0.001).
Tracking of BMI, overweight and obesity from infancy
In boys, the overall rate of overweight including obesity was
22.1% at 1 y with 13.4% overweight nonobese and 8.7%
obese. At 3 y, 13.5% of boys were classified as being
overweight including obesity (1.0% obese) as were 12.6%
at 6 y (4.0% obese) and 15.4% at 8 y (4.5% obese). Among 1-
y-old girls, 25.1% were categorised as being overweight
including obesity (11.8% obese). Respective proportions
were 14.0% at 3 y (1.6% obese), 17.9% at 6 y (3.7% obese)
and 19.6% (3.2% obese) at 8 y (Figure 2). Table 2 shows
partial correlation coefficients with adjustment for sex
between birthweight and weights at 1, 3, 6, and 8 y. All
coefficients were statistically significant (Po0.001) but
coefficients were lower with a greater interval between
measurements.
Among children with weight recorded at 8 y, of the 310
children classified as overweight or obese at the age of 1 y,
103 (33%) were overweight or obese at 8 y of age. Of 128
overweight or obese 3 y olds, 68 (53%) were overweight or
obese at the age of 8 y; among 196 overweight or obese 6 y

International Journal of Obesity

 BMI and cardiovascular risk factors in children
V Burke et al
18

Figure 2 Proportion of overweight, nonobese or obese boys and girls at the

ages of 1, 3, 6 and 8 y. At the age of 1 y, overweight and obesity were defined
with reference to CDC growth charts.22 For 3, 6 and 8 y olds, overweight and
obesity were defined by the standards of Cole et al.14

statistically significant odds ratio (OR) for being overweight

or obese at the age of 1 y (OR 3.38, 95% CL 2.50, 4.55), 3 y
(OR 8.54, 95% CL 5.69, 12.81) and 6 y (OR 51.13, 95% CL
33.71, 77.52) with Po0.001 for each OR.

Associations between overweight including obesity and

cardiovascular risk factors
BP was significantly higher in children who were classified as
overweight and obese with a difference of 2/3 mmHg
(systolic/diastolic) at 1 y, 3/2 mmHg at 3 y, 4/2 mmHg at 6 y
and 6/2 mmHg at 8 y (Table 3). Blood lipids and blood
glucose were measured at the age of 8 y but not in earlier
surveys. HDL was significantly lower (P ¼ 0.002) in the group
that was overweight including obesity and triglycerides were
significantly higher (Po0.001) as shown in Table 3.
Although LDL was also higher in that group, this not reach
Figure 1 BMI in children at the ages of 1, 3, 6 and 8 y related to upper and statistical significance (P ¼ 0.078) nor did the between-group
lower family income groups or to maternal education classified as secondary, difference for total cholesterol (P ¼ 0.162) or blood glucose
tertiary or other. (P ¼ 0.182). Interpretation was unchanged after adjustment
for sex.
Table 2 Partial correlation coefficients relating birthweight, and weight at
years 1, 3, 6 and 8 with adjustment for sex showing number and P-value
Factor analysis of FFQ categories
Weight year 1 Weight 3 y Weight 6 y Weight 8 y Five factors with eigenvalues greater than 1 and explaining
53% of the variance were extracted using categories of foods
Birthweight 0.3551 (1707) 0.3623 (1181) 0.2681 (1477) 0.2016 (1427)
derived from the FFQ. These factors comprised bread, cereals
Po0.001 Po0.001 Po0.001 Po0.001
Weight year 1 0.7839 (1109) 0.6016 (1380) 0.4446 (1332) and spreads such as jam (‘cereals’); takeaway foods, processed
Po0.001 Po0.001 Po0.001 meats and desserts (‘takeaways’); fruit, vegetables and dairy
Weight 3 y 0.8051 (1016) 0.7154 (985) foods (‘fruitveg’); chicken, meat dishes, eggs and fish
Po0.001 Po0.001
(‘meats’); and snacks, confectionary, cakes and sweet bev-
Weight 6 y 0.8722 (1295)
Po0.001 erages (‘sweets’). Factor scores were saved and used as
independent variables in regression models.

olds, 154 (79%) were overweight or obese as 8 y olds. In Predictors of BMI in 8 y olds
separate logistic regression models, adjusted for sex, being Table 4 shows variables examined as predictors of BMI at the
overweight or obese at the age of 8 y was associated with a age of 8 y in models adjusted only for sex. Statistically

International Journal of Obesity

 BMI and cardiovascular risk factors in children
V Burke et al
19
Table 3 Associations between cardiovascular risk factors and overweight Table 4 Regression models adjusted only for sex with BMI at 8 y of age as the
including obesity at the ages of 1, 3, 6 and 8 y dependent variable

Nonoverweight Overweight including obesity P-value Variable Coefficient 95% CL P-value R2

1y Birth weight (kg) 0.669 0.382, 0.951 o0.001 0.013

SBP 99.6 (0.4) 101.8 (0.8) 0.013 Duration of 0.030 0.053, 0.007 0.012 0.004
DBP 58.2 (0.4) 61.0 (0.8) 0.001 breastfeeding
(months)
3y Age when milk was first 0.038 0.076, 0.001 0.048 0.002
SBP 97.9 (0.3) 100.4 (0.8) 0.004 introduced (months)
DBP 52.6 (0.3) 54.2 (0.8) 0.052 Mothers’ BMI (kg/m2) 0.120 0.096, 0.144 o0.001 0.071
Fathers’ BMI (kg/m2) 0.140 0.088, 0.193 o0.001 0.062
6y (n ¼ 383)
SBP 102.4 (0.3) 106.6 (0.6) o0.001 Mother smoking status
DBP 54.3 (0.2) 56.2 (0.5) 0.001 Never smoker 0.732 1.059,0.405, o0.001 0.012
Ex-smoker 0.513 0.850, 0.175 0.003
8y Mother diabetes 0.850 0.276, 1.424 0.004 0.005
SBP 103.1 (0.3) 109.0 (0.7) o0.001 (children 8 y old)
DBP 55.8 (0.2) 58.2 (0.4) o0.001 Mother gestational 1.485 0.451, 2.250 0.005 0.004
Blood glucose 4.52 (0.02) 4.60 (0.04) 0.182 diabetes
Total cholesterol 4.47 (0.04) 4.60 (0.09) 0.162 Time watching TV (h/
HDL 1.56 (0.02) 1.43 (0.04) 0.002 day)
LDL 2.62 (0.04) 2.78 (0.08) 0.078 6y 0.297 0.163, 0.431 o0.001 0.013
Triglycerides 0.59 (0.56, 0.61) 0.75 (0.66, 0.84) o0.001 8y 0.185 0.076, 0.294 0.001 0.007
Activity
Slightly active vs 0.732 1.159, 0.305 0.001 0.019
sedentary
Active vs sedentary 1.181 1.622, 0.741 o0.001

significant relationships were seen with the following Fat intake (% energy)a 0.069 0.136, 0.002 0.044 0.019
Saturated fat intake 0.141 0.266, 0.015 0.025 0.022
variables: birth weight, duration of breastfeeding, age at (% energy)a
which other milk was introduced, gestational diabetes, Takeaways factorb 0.399 0.056, 0,742 0.023 0.024
maternal diabetes at the time of the survey of 8 y olds, Bread, cereals factorb 0.343 0.679, 0.007 0.046 0.019
mothers’ and fathers’ BMI, maternal smoking status, hours a
Intake calculated from FFQ. b
Factors obtained by principal components
spent watching TV at the age of 6 and at 8 y, subjective analysis using FFQ.
assessment of activity as ‘slightly active’ or ‘active’ at the age
of 8 y and family income. There was also a statistically
significant inverse association with the ‘cereals’ factor and
with intake of total and saturated fat (% energy) and a The inclusion of father’s BMI produced a similar model
positive association with the ‘takeaways’ factor. All P-values (Table 5), using data from 298 children, with the addition of
for associations with intake of energy, protein, carbohydrate, a significant positive relationship with paternal BMI. The
refined sugars, polyunsaturated or monounsaturated fat, relationship with hours spent watching television at the age
calcium, magnesium, potassium or sodium were 40.10, as of 6 y was similar to the model that included only maternal
was the association with playing organised sport at the age of data, but this was no longer statistically significant
6 or 8 y. (P ¼ 0.052).
All variables with P-valueso0.1 in regression adjusted only In the subset of children who completed the FFQ, mother’s
for sex were entered into backwards selection models, BMI and children’s activity category were again predictive of
adjusted for maternal education, using three sets of variables. BMI at the age of 8 y, while the cereals factor was an
This approach was necessary because of missing values for independent inverse predictor (Table 5). No other dietary
variables relating to fathers and the use of the FFQ only in a variable was an independent predictor of BMI at the age of
subset of 340 children. 8 y in multivariate models.
In multivariate models that excluded data from fathers
and FFQ (Table 5), with adjustment for sex and maternal
education, BMI at the age of 8 y was independently predicted Predictors of overweight including obesity at the age
positively by mother’s BMI at the time of the survey of 8 y of 8 y
olds, by birthweight and by time spent in watching Table 6 shows logistic regression models related to over-
television at the age of 6 y. There were significant inverse weight including obesity at the age of 8 y. In the model
relationships with mothers never having smoked or being excluding paternal data, maternal overweight including
ex-smokers, relative to those who were current smokers, and obesity increased the OR for overweight including obesity
to children being classified as ‘slightly active’ and ‘active’ in 8 y olds 2.6-fold, and for each hour per day of watching
relative to being sedentary at the age of 8 y. television at the age of 6 y, the risk of overweight including

International Journal of Obesity

 BMI and cardiovascular risk factors in children
V Burke et al
20
Table 5 Backwards selection models with BMI at the age of 8 y as the Table 6 Logistic regression models with overweight including obesity at the
dependent variable including variables with Po0.10 from Table 2 with age of 8 y as the dependent variable
adjustment for sex and maternal education (three groups)
Odds Ratio 95% Confidence limits P-value
95% Confidence
Coefficient limits P-value Model excluding paternal data
Mother overweight or obese 2.588 1.838, 3.645 o0.001
Model excluding paternal data Mother ex-smoker 0.573 0.378, 0.868 0.009
(Adjusted R2 ¼ 0.131, Mother never smoked 0.346 0.224, 0.534 o0.001
N ¼ 1133) TV watching at 6 y (hours) 1.422 1.260, 1.605 o0.001
Birth weight 0.573 0.259, 0.886 0.001 Slightly active 8 y 0.439 0.275, 0.699 0.001
Mother’s BMI (kg/m2) 0.119 0.094, 0.145 o0.001 Active 8 y 0.227 0.135, 0.381 o0.001
Mother ex-smoker 0.663 1.032,0.295 o0.001
Mother never smoked 0.996 1.360, 0.632 o0.001 Model including paternal data
TV watching at 6 y (h) 0.153 0.008, 0.297 0.039 Mother overweight or obese 3.181 1.393, 7.265 0.006
Slightly active 8 y 1.002 1.445, 0.548 o0.001 Father overweight or obese 3.100 1.173, 8.197 0.023
Active 8 y 1.268 1.739, 0.295 o0.001 Mother ex-smoker 0.220 0.078, 0.619 0.004
Mother never smoked 0.128 0.037, 0.444 0.001
Model including paternal data TV watching at 6 y (h) 1.534 1.164, 2.021 0.002
(Adjusted R2 ¼ 0.199, N ¼ 298) Slightly active 8 y 0.418 0.149, 1.168 0.096
Mother’s BMI (kg/m2) 0.076 0.032, 0.119 0.001 Active 8 y 0.150 0.043, 0.520 0.003
Father’s BMI (kg/m2) 0.129 0.071, 0.186 o0.001
Birth weight (kg) 1.118 0.540, 1.697 o0.001 Models are shown with and without paternal variables and are adjusted for sex
Mother ex-smoker 0.973 1.704, 0.242 0.009 and maternal education (three groups).
Mother never smoked 1.537 2.303, 0.771 o.001
TV watching at 6 y (h) 0.260 0.003, 0.522, 0.052
Slightly active 0.796 1.592, 0.001 0.050
Active 1.185 2.016, 0.353 0.005
including obesity between surveys, with 80% of overweight
Model including maternal data or obese 6 y olds still overweight or obese at the age of 8 y,
and FFQ (Adjusted R2 ¼ 0.182, demonstrate tracking. An increase in cardiovascular risk
N ¼ 214)
Mother’s BMI (kg/m2) 0.184 0.125, 0.243 0.000
factors associated with overweight including obesity is
Slightly active 0.819 1.959, 0.320 0.158 already apparent in infancy and, by the age of 8 y, these
Active 1.207 2.380, 0.033 0.044 children have higher BP and more adverse levels of blood
Cereals factor 0.395 0.725, 0.066 0.019 lipids.
The association of these risk factors with overweight
including obesity indicate that features of the metabolic
syndrome are already present in these children. The
obesity at the age of 8 y increased by about 40%. Risk Pathobiological Determinants of Atherosclerosis in Youth
decreased by 40% in the children of ex-smokers and by 60% (PDAY) group have documented the associations between
where mothers had never smoked. There was a 60% decrease cardiovascular risk factors and the presence of coronary
in risk among children who were slightly active and a atherosclerosis at post mortem in adolescents and young
decrease of 70% associated with being classified as active. adults who have died from external causes.24 The pattern of
In the model that included paternal data, there were cardiovascular risk associated with overweight including
independent effects of maternal and paternal overweight obesity in children in the present study implies elevated
including obesity, each increasing the risk by more than long-term risk and emphasises the potential importance of
three-fold. As in the previous logistic model, nonsmoking or factors leading to weight control in childhood.
ex-smoking mothers and activity in the children decreased We found that lower BMI in 6 and 8 y olds was associated
the risk of overweight including obesity at the age of 8 y, with higher family income and with tertiary education in
while risk increased with hours of watching television at the mothers, consistent with the inverse association reported
age of 6 y. Slight activity was not a statistically significant between SES and BMI in Australian children more than 6 y
predictor in this model. old.25 In the present study, associations between SES and
BMI changed with age and lower BMI in the higher SES
group emerged only in the surveys of 6 and 8 y olds.
BMI at the age of 8 y is strongly predicted by maternal and
Discussion paternal BMI, by mothers’ smoking habits and by children’s
This prospective study in Australian children has shown that activity level at 8 y, as well as by earlier habits of television
the proportion of overweight and obesity increases between watching. Dietary variables and energy intake were not
the ages of 3 and 8 y, particularly in girls, with 15% of 8-year- predictive of BMI at the age of 8 y in models adjusted for
old boys and 20% of girls being overweight or obese. confounders, except for an inverse relationship with con-
Significant correlations between weight at 8 y of age and sumption of bread and cereals. Maternal or paternal over-
weight at birth, 3 and 6 y and persistence of overweight weight and a pattern of television watching in earlier years

International Journal of Obesity


substantially increases risk of overweight including obesity
at the age of 8 y.
Overweight and obesity were classified according to the
international standards suggested by Cole et al14 in 3, 6, and
8 y olds. These standards do not apply to infants, and
overweight and obesity were defined with reference to the
NCHS growth charts in 1 y olds. It is therefore not possible to
compare rates of overweight and obesity in 1 y olds and
those in older children. However, it is clear from the rates in
3, 6 and 8 y olds that there is an increase in the prevalence of
overweight including obesity from the age of 3–8 y, particu-
larly in girls. This increase in overweight and obesity seen in
a prospective study is in agreement with reports from other
cross-sectional1,7–12 and longitudinal studies in Australia and
elsewhere.15–19
A greater rate of sedentary behaviour in 8-year-old girls
than in boys is consistent with differences in the rates of
overweight and obesity between boys and girls. It is also
consistent with Australian population data showing that
girls between the ages of 6 and 14 y are less physically active
than boys.26 Parents’ subjective assessment of their child’s
activity level as sedentary, slightly active or active was an
independent predictor of BMI in 8 y olds and of the risk of
overweight including obesity at the age of 8 y. Longitudinal
studies have shown a decline in levels of physical activity,27
and secular trends to a decrease in activity reflect societal
change.1 Sedentary pastimes (television, computer games)
are easily accessible, while fears about safety have limited
unsupervised activities such as walking to school and playing
in the local neighbourhood. The importance of activity in
early childhood in limiting deposition of body fat is high-
lighted by data from the longitudinal Framingham children’s
study that showed higher activity levels between the ages of
4 and 11 y to be associated with smaller gains in body fat in
adolescence.28 The Third National Health and Nutrition
Examination in the United States reported not more than
two episodes per week of vigorous activity in 20% of children
over the age of 8 y.29 Such statistics emphasise the urgency of
encouraging the habit of physical activity from an early age,
with a focus on increasing activity in girls.
Greater BMI in childhood is associated with watching
television in cross-sectional29 and in longitudinal studies,30
possibly related both to sedentary behaviour and high-
energy foods that are advertised during children’s viewing
times.31 However, not all surveys have agreed with these
findings.32 We found that greater BMI at the age of 8 y was
predicted by greater time spent watching television at the
age of both 6 and 8 y in models adjusted only for sex. In
models that included confounders, only television watching
at the age of 6 y predicted BMI in 8 y olds, independent of
categories of physical activity. The statistically significant
association between hours of watching television at 8 y and
the category of activity level may have led to exclusion of
that variable from multivariate models. Television watching
at the age of 6 y was predictive of BMI at 8 y even after
inclusion of potential confounders, possibly mediated by
BMI and cardiovascular risk factors in children
V Burke et al
21
lack of physical activity and its effect on accumulation of
body fat.28 Programmes that focus on reducing time spent
watching television, rather than encouraging physical
activity, have been successful in weight control33 and weight
reduction in children.34 However, the association between
television viewing and BMI may be more complex, as
Burdette et al35 have shown that maternal obesity and
depression are associated with greater viewing time in
preschool children.
Dietary data from FFQ were available only from a subset of
the cohort at the age of 8 y. In models unadjusted for a range
of confounders, a food factor that encompassed takeaway
foods, desserts and processed meats was a positive predictor
of BMI at the age of 8 y, while a factor that measured
consumption of breads and cereals was inversely associated
with BMI. However, only the cereals factor remained
statistically significant in models that included mother’s
BMI and children’s activity. Low glycaemic index foods have
been promoted in the management of obesity, possibly
acting both through effects on fat oxidation and satiety.36
Children who ate low glycaemic index foods at breakfast ate
less at lunchtime than those having a high glycaemic index
breakfast,37 and greater satiety may mediate the association
we found between greater cereal consumption and lower
BMI.
Intake of total fat and saturated fat derived from the FFQ
showed an inverse association with BMI in 8 y olds in our
study. Although this association is not intuitive, it was
consistent with the direction of the nonsignificant relation-
ship between BMI and fat intake derived from the short fat
questionnaire21 that was not limited to children who
completed the FFQ. One possible explanation is the use of
weight reduction diets in overweight or obese children and/
or increased consumption of high glycaemic index sugars.
Parental BMI is recognised to be an important predictor of
BMI in children. The risk of obesity in adult life is more than
doubled in children with one obese parent and the risk
increases to 10-fold if both parents are obese.2 While genetic
factors are important determinants of BMI, with moderate
estimated heritability,38,39 interactions between genes and
environment substantially influence adiposity. Obesity is not
an inevitable result of a genetic predisposition.39,40 Health-
related behaviours aggregate within families.41–43 In families
identified as ‘obesogenic’ on the basis of parental physical
activity and diet, 7-year-old girls had higher BMI and
skinfold thickness than those from families where parents
practised more healthy behaviours.44 In a study that used
supermarket receipts to identify patterns of food purchasing,
households with overweight individuals bought foods high-
er in fat and energy than households where participants were
leaner.45 Dietary fat preferences in children are also asso-
ciated with overweight in parents.46
Smoking shows clustering with adverse life style choices47
such as sedentary behaviour, higher fat consumption and
greater alcohol intake.48,49 In Australian 10–12 y olds, we
have shown that children of parents who smoke have higher
International Journal of Obesity
 BMI and cardiovascular risk factors in children
V Burke et al
22
BMI with a central distribution of body fat, have a greater
intake of dietary fat, exercise less and watch more televi-
sion.50 Maternal smoking during pregnancy has also been
reported to increase the risk of later obesity51 and may be
mediated by effects on intrauterine growth in addition to
these lifestyle factors.
While genetic factors and their interaction with the
environment influence BMI, it is important to recognise
the role of adverse health-related behaviours within families,
particularly families where there is overweight or obesity.
Tracking of overweight and obesity and of cardiovascular risk
factors to adult life suggest that control of excessive weight
gain in childhood is likely to have long-term benefits.
Parental obesity should not be regarded as necessarily
leading to overweight or obesity in children, but should be
considered as an indication for evaluation of family life style.
Better health-related behaviours within these families have
the potential to improve health substantially in parents and
children.
Acknowledgements
This study was supported by the Raine Foundation and
Healthway.
References
1 American Academy of Pediatrics Policy Statement. Prevention of
pediatric overweight and obesity. Pediatrics 2003; 112:
424–430.
2 Guo SS, Huang C, Maynard LM, Demerath E, Towne B, Chumlea
WC, Siervogel RM. Body mass index during childhood, adoles-
cence and young adulthood in relation to adult overweight and
adiposity: the Fels Longitudinal Study. Int J Obes 2000; 24:
1628–1635.
3 Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH.
Predicting obesity in young adulthood from childhood and
parental obesity. N Engl J Med 1997; 337: 869–873.
4 Power C, Lake JK, Cole TJ. Measurements of long-term health
risks of child and adolescent fatness. Int J Obes Relat Metab Disord
1997; 21: 507–526.
5 World Health Organization. Obesity: Preventing and Managing the
Global Epidemic. World Health Organization: Geneva; 1998.
6 Cameron AJ, Welborn TA, Zimmet PZ, Dunstan DW, Owen N,
Salmon J, Dalton M, Jolley D, Shaw JE. Overweight and obesity in
Australia: the 1999–2000 Australian diabetes, obesity and lifestyle
Study (AusDiab). Med J Aust 2003; 178: 427–432.
7 Tremblay MS, Katzmarzyk PT, Willms JD. Temporal trends in
overweight and obesity in Canada, 1981–1996. Int J Obes Relat
Metab Disord 2002; 26: 538–543.
8 Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence
and trends in overweight among US children and adolescents,
1999–2000. JAMA 2002; 288: 1728–1732.
9 Frye C, Heinrich J. Trends and predictors of overweight and
obesity in East Germany. Int J Obese Relat Metab Disrod 2003; 27:
963–969.
10 Ogden CL, Troiano RP, Breifel RR, Kuczmarski RJ, Flegal KM,
Johnson CL. Prevalence of overweight among preschool children
in the United States, 1971 through 1994. Pediatrics 1997; 99 (4)
http://www.pediatrics.org/cgi/content/full/99/4/e.
11 Booth ML, Chey T, Wake M, Norton K, Hesketh K, Dollman J,
Robertson I. Change in the prevalence of overweight and obesity
International Journal of Obesity
among young Australians, 1969–1997. Am J Clin Nutr 2003; 77:
29–36.
12 Lazarus R, Wake M, Hesketh K, Waters E. Change in body mass
index in Australian primary school children 1985–1997. Int J Obes
Relat Metab Disord 2000; 24: 679–684.
13 Magarey AM, Daniels LA, Boulton TJC. Prevalence of overweight
and obesity in Australian children and adolescents: reassessment
of 1985 and 1995 data against new international definitions. Med
J Aust 2001; 174: 561–564.
14 Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a
standard definition for child overweight and obesity worldwide:
international survey. BMJ 2000; 320: 1240–1243.
15 Strauss RS, Pollack HA. Epidemic increase in childhood over-
weight, 1986–1998. JAMA 2001; 286: 2845–2848.
16 Kimm SY, Barton BA, Obarzenek E, McMahon RP,
Kronsberg SS, Waclawiw MA, Morrison JA, Schreiber GB, Sabry
ZI, Daniels SR. Obesity development during adolescence in a
biracial cohort: NHLBI Growth and Health Study. Pediatrics 2002;
110: e54.
17 Luo J, Hu FB. Time trends of obesity in pre-school children in
China from 1989–1997. Int J Obes Relat Metab Disord 2002; 26:
553–558.
18 Magarey AM, Daniels LA, Boulton TJ, Cockington RA. Predicting
obesity in early adulthood from childhood and parental obesity.
Int J Obes Relat Metab Disord 2003; 27: 505–513.
19 Burke V, Beilin LJ, Dunbar D. Associations between blood pressure
and overweight defined by new standards for body mass index in
childhood. Prev Med 2004; 38: 558–564.
20 Newnham JP, Evans SF, Michael CA, Stanley FJ, Landau LI. Effects
of frequent ultrasound during pregnancy: a randomised con-
trolled trial. Lancet 1993; 342: 887–891.
21 Kinlay S, Heller RF, Halliday JA. A simple score and questionnaire
to measure group changes in dietary-fat intake. Prev Med 1991;
20: 378–388.
22 National Center for Health Statistics. CDC growth charts: United
States, August 2002 http://www.cdc.gov/growth-charts/.
23 Kim J-O, Mueller CW. Factor Analysis. Statistical Methods and
Practical Issues. Sage Publications: Newbury Park; 1978.
24 McGill HC, McMahan CA, Herderick EE, Zieske AW, Malcolm G,
Tracy RE, Strong JP. Pathological Determinants of Atherosclerosis
in Youth (PDAY) Research Group. Obesity accelerates the
progression of coronary atherosclerosis in young men. Circulation
2002; 105: 2712–2718.
25 O’Dea JA. Differences in overweight and obesity among Austra-
lian schoolchildren of low and middle/high socioeconomic
status. Med J Aust 2003; 179: 63.
26 Australian Bureau of Statistics 2003. http://www.abs.gov.au/
ausstats/abs@.nsf/0/3C268646099C00C1CA256CAE0010852E?
Open&Highlight ¼ 0,children,sport.
27 Berkey CS, Rockett HR, Field AE, Gillman MW, Frazier L, Camargo
CA, Colditz GA. Activity dietary intake and weight changes in a
longitudinal study of preadolescent and adolescent boys and
girls. Pediatrics 2000; 105 (4): E56.
28 Moore LL, Gao D, Bradlee ML, Cupples LA, Sundarajan-Rama-
murti A, Proctor MH, Hood MY, Singer MR, Ellison RC. Does early
childhood physical activity predict body fat change through
childhood? Prev Med 2003; 37: 10–17.
29 Anderson RE, Crespo CJ, Bartlett SJ, Cheskin LJ, Pratt M.
Relationship of physical activity and television watching with
body weight and level of fatness among children. Results from
the Third National Health and Nutrition Examination Survey.
JAMA 1998; 279: 938–942.
30 Proctor MH, Moore LL, Gao D, Cupples LA, Bradlee ML, Hood
MY, Ellison RC. Television viewing and change in body fat from
preschool to early adolescence: The Framingham Children’s
Study. Int J Obes Relat Metab Disord 2003; 27: 827–833.
31 Hill J, Radimer K. A content analysis of food advertisements in
television for Australian children. Aust J Nutr Diet 1997; 54:
174–180.

32 Durant R, Baranowski T, Johnson M, Thompson WO. The
relationship among television watching, physical activity and
body composition of young children. Pediatrics 1994; 94:
449–455.
33 Robinson T. Reducing children’s television to prevent obesity: a
randomized controlled trial. JAMA 1999; 282: 1561–1567.
34 Epstein LH, Myers MD, Raynor HA, Saelens BE. Treatment of
pediatric obesity. Pediatrics 1998; 101 (Suppl): 554–570.
35 Burdette HL, Whitaker RC, Kahn RS, Harvey-Berino J. Association
of maternal obesity and depressive symptoms with television-
viewing time in low-income preschool children. Arch Pediatr
Adolesc Med 2003; 157: 894–899.
36 Brand-Miller JC, Holt SH, Pawlak DB, McMillan J. Glycemic index
and obesity. Am J Clin Nutr 2002; 76 (Suppl 2): 281S–285S.
37 Warren JM, Henry JK, Simonite V. Low glycemic index breakfasts
and reduced food intake in preadolescent children. Pediatrics
2003; 112: e414–e419.
38 Coady SA, Jaquish CE, Fabsitz RR, Larson MG, Cupples LA,
Myers RH. Genetic variability of adult body mass index: a
longitudinal assessment in Framingham families. Obes Res.
2002; 10: 675–681.
39 Hunt MS, Katzmarzyk PT, Perusse L, Rice T, Rao DC, Bouchard C.
Familial resemblance of 7-year changes in body mass and
adiposity. Obes Res 2002; 10: 507–517.
40 Guillaume M, Bjorntorp P. Obesity in children: environmental
and genetic aspects. Horm Metab Res 1996; 28: 573–581.
41 Cunnane SC. Childhood origins of lifestyle-related risk factors
for coronary heart disease in adulthood. Nutr Health 1993; 56:
107–115.
42 Wild RA, Taylor EL, Knehans A, Cleaver V. Matriarchal model for
cardiovascular prevention. Obstet Gynecol Survey 1994; 49:
147–152.
BMI and cardiovascular risk factors in children
V Burke et al
23
43 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 2001; 25: 147–157.
44 Krahnstoever Davison K, Lipps Birch L. Obesigenic families:
parents’ physical activity and dietary intake patterns predict
girls’ risk of overweight. Int J Obes Relat Metab Disord 2002; 26:
1186–1193.
45 Ransley JK, Donnelly JK, Botham H, Khara TN, Greenwood DC,
Cade JE. Use of supermarket receipts to estimate energy and fat
content of food purchased by lean and overweight families.
Appetite 2003; 41: 141–148.
46 Fisher JO, Birch LL. Fat preferences and fat consumption of 3- to
5-year-olds children are related to parental adiposity. J Am Diet
Assoc 1995; 95: 759–764.
47 Burke V, Milligan RA, Beilin LJ, Dunbar D, Spencer M, Balde E,
Gracey MP. Clustering of health-related behaviors among
18-year-old Australians. Prev Med 1997; 26: 724–733.
48 Raitakari OT, Porrka KV, Taimela S, Telama R, Rasanen L, Viikari
JS. Effects of persistent physical activity and inactivity on
coronary risk factors in children and young adults. The
Cardiovascular Risk in Young Finns Study. Am J Epidemiol 1994;
140: 195–205.
49 Matanoski G, Kanchanaraska S, Lantry D, Chang Y. Character-
istics of non-smoking women in the NHANES I and NHANES I
epidemiologic follow-up study with exposure to spouses who
smoke. Am J Epidemiol 1995; 142: 149–157.
50 Burke V, Gracey MP, Milligan RA, Thompson C, Taggart AC, Beilin
LJ. Parental smoking and risk factors for cardiovascular disease in
10- to 12-year-old children. J Pediatr 1998; 133: 206–213.
51 Power C, Jefferis BJMH. Fetal environment and subsequent
obesity: a study of maternal smoking. Int J Epidemiol 2002; 31:
413–419.
International Journal of Obesity