Content uploaded by Lise Dubois
Author content
All content in this area was uploaded by Lise Dubois on Oct 07, 2014
Content may be subject to copyright.
PEDIATRIC HIGHLIGHT
Early determinants of overweight at 4.5 years in a
population-based longitudinal study
L Dubois
1
and M Girard
2
1
Canada Research Chair in Nutrition and Population Health, Department of Epidemiology and Community Medicine,
University of Ottawa, Ottawa, Ontario, Canada K1N 6N5 and
2
Department of Epidemiology and Community Medicine,
Institute of Population Health, University of Ottawa,Ottawa, Ontario, Canada K1N 6N5
Objectives: The roots of the obesity epidemic need to be traced back as early in life as possible in order to develop effective
means for preventing obesity and its health consequences in the future. The aim of this paper is to examine a broad range of
factors that may simultaneously contribute to childhood overweight in a population-based cohort of children followed from
birth to 4.5 years, to determine which factors exert the most influence in early life.
Design: The analyses were performed using data from the Quebec Longitudinal Study of Child Development 1998–2002
(QLSCD).
Subjects: The study follows a representative sample (n¼2103) of children born in 1998 in the Canadian province of Quebec.
Measure: Measured height and weight were available for 1550 children aged 4.5 years. At 4.5 years, BMI was analyzed using
the US CDC sex- and age-specific growth charts. In order to study children at their highest weights at various ages, odds ratios
were presented for high birth weight, weight-for-stature at or above the 95th percentile at 5 months, and BMI at or above the
95th percentile at 4.5 years. Monthly weight gain between birth and five months has been analyzed. Children were also
evaluated by the Z-score obtained from the standardized weight divided by height. Factors potentially related to children’s
weight include sex, gestational age and birth rank, breastfeeding, mothers’ smoking status during pregnancy, family type at
child’s birth, and family income before pregnancy and when the children were 5 months and 4.5 years old. Other parental
factors such as height and overweight/obesity (based on BMI) and other maternal factors (age, education, immigrant status)
were also part of the analysis.
Results: Being in the highest quintiles of weight gain between birth and 5 months, as well as maternal smoking during
pregnancy, almost double the odds of being overweight at 4.5 years. Parental overweight or obesity also increased the odds of
being overweight at this age, as well as being raised in middle-income or in poor families. A greater proportion of children born
to nonsmoking mothers with higher weights (more than 4000 g) were overweight at 4.5 years, the percentage being greatest
for those in the highest weight-gain categories from birth to 5 months. The pattern was different for children born to smoking
mothers. The greatest proportion of 4.5-year-old overweight children was seen for children born in the normal weight-range
category (3000–4000 g) who were in the highest quintiles of weight gain from birth to 5 months, and for children with high
birth weights (more than 4000 g) who were in the lowest quintiles of birth-to-5-months weight gain. Children who were
overweight at 4.5 years and who had been born to smoking mothers started life with a birth weight around that for the
population means, but they gained more weight in the first 5 months of life than did the children of nonsmoking mothers.
Conclusion: This study indicates that behavioral and social factors exert critical influences on the onset of childhood overweight
in preschool years. From a population–health perspective, interventions aimed at preventing childhood obesity would do well to
target smoking pregnant women, as well as nonsmoking pregnant women at risk for giving birth to high-birth-weight children,
paying particular attention to rapid weight gain in the first months of life.
International Journal of Obesity (2006) 30, 610–617. doi:10.1038/sj.ijo.0803141
Keywords: overweight; preschoolers; determinants; birth cohort; population-based
Introduction
The increase in the prevalence of obesity in recent years has
brought this condition to the forefront of the public health
agenda in many countries. No simple treatment for this
condition is known, nor is it known at what age it is best to
Received 30 November 2004; revised 30 August 2005; accepted 16
September 2005
Correspondence: Professor L Dubois, Department of Epidemiology and
Community Medicine, Faculty of Medicine, Institute of Population Health,
University of Ottawa, 1 Stewart Street, office 303, Ottawa, Ontario, Canada
K1N 6N5.
E-mail: ldubois@uottawa.ca
International Journal of Obesity (2006) 30, 610– 617
&
2006 Nature Publishing Group All rights reserved 0307-0565/06
$
30.00
www.nature.com/ijo
intervene. It is thus important to investigate why adults,
adolescents and children are getting fatter. The roots of the
obesity epidemic need to be traced back as early in life as
possible in order to develop effective means for preventing
this condition and its health consequences in the future.
Various factors at different periods in early life have been
associated with obesity later in life. Birth weight, a crude
indicator of growth in utero, is positively related with adult
fatness and obesity,
1
but it is less clear how birth weight
relates to body weight in childhood. Several confounding
factors influence this relationship. For example, maternal
smoking during pregnancy is related to low birth weight,
which has been associated with catch-up growth in the first
months of life, which in turn has been associated with
overweight and obesity in childhood and later in life.
2–4
Conversely, breast-feeding has been associated with lower
weight gain in infancy and less obesity in childhood and
adolescence,
5–8
but research on the subject remains incon-
clusive.
9,10
Similarly, numerous studies performed in various
developed countries have demonstrated a negative relation-
ship between socioeconomic status and obesity in adults and
adolescents, but a less consistent relationship for children.
1
Maternal smoking during pregnancy is more prevalent
among mothers of lower socioeconomic status, who give
birth to smaller children and breast-feed less. These relation-
ships are complex, and little is known about their overall
influence on body-weight changes and overweight develop-
ment in the preschool years.
The aim of this paper is to examine simultaneously a broad
range of factors related to preschoolers’ overweight using a
population-based cohort of children followed from birth to
4.5 years, to determine which factors exert the most
influence in early life.
Methods
The analyses were performed using data from the Quebec
Longitudinal Study of Child Development 1998–2002
(QLSCD), conducted by Sante
´Que
´bec, a division of the
Institut de la Statistique du Que
´bec (ISQ).
8,11
The study
analyzes the role of familial and social factors on child
health, cognitive and behavioral outcomes. The study
follows a representative sample (n¼2103) of children born
in 1998 in the Canadian province of Quebec (total popula-
tion over 7 million, with approximately 70 000 newborns per
year). The representative sample was chosen by randomly
selecting children born in each public health region in
Quebec during 1998 such that the seasonality effect was
minimized. Twins, children with major diseases or handicaps
at birth, and those who died before the age of 5 months were
not part of the study. The children were first seen at 5
months (gestational age), and once a year thereafter. The
study is based on face-to-face interviews and involves a set of
questionnaires addressed to children’s mothers and fathers.
Children’s questionnaires are answered by the most knowl-
edgeable person, generally the mother. Of the 2103 babies
included in the first round of data collection, 1944 remained
in the sample by age 4 and 5 years (in 2002) (ages of the
children ranged from 44 to 56 months).
Measured height and weight were available for 1550
children aged 4.5 years. Data were weighted by a factor
based on the inverse of the selection probability, the
probability of nonresponse, the poststratification rate, and
the attrition rate, to ensure that the data were longitudinally
representative of the same-age children in the population.
12
Statistical analyses were based on individuals with no
missing values for any of the studied variables. Among the
1550 babies, 1450 (94% of the sample) were included in the
sample. We analyzed the impact of missing data by
conducting with-and-without analyses. Missing data were
excluded from the analyses since they had no impact on the
results.
Birth weight was analyzed in four categories: under 2500 g
(low birth weight), 2500–2999 g, 3000–4000 g, and over
4000 g (high birth weight). At 5 months, weight-for-stature
was defined as weight adjusted for height based on the US
Centre for Disease Control (CDC) sex-specific Growth Chart.
At 4.5 years, BMI was analyzed using the CDC sex- and age-
specific growth charts.
13
Monthly weight gain from birth to
5 months (i.e., the ratio of the weights at both ages divided
by number of months between them) was analyzed in
quintiles. Analysis performed on monthly weight gain
between birth and 5 months, 5 months and 1.5 years, 1.5
and 2.5 years, 2.5 and 3.5 years, and 3.5 and 4.5 years
indicates that there is a significant statistical association
between monthly weight gain between birth and 5 months
and overweight at 4.5 years, but not between monthly
weight gain for each of the other studied ages above 5
months and overweight at 4.5 years. In order to study the
children at the highest weight levels at different ages, odds
ratios (OR) were presented for high birth weight, weight-for-
stature at or above the 95th percentile at 5 months, and BMI
at or above the 95th percentile at 4.5 years. Children were
also evaluated by the Z-score obtained from the standardized
weight divided by height.
Factors potentially related to children’s weight are pre-
sented in Table 1, along with their proportion of the sample
studied. Sex, gestational age and birth rank were collected
from medical records. At 5 months, mothers were asked
about their smoking status during pregnancy, family type at
child’s birth, and family income before pregnancy. Mothers
were also asked for family income when the children were 5
months and 4.5 years old. Parental overweight/obesity was
based on BMI, calculated from reported parental height and
weight (BMI ¼weight in kg/height
2
in meters), and was
defined as having a BMI of ‘25 or above’. The duration of
breast-feeding was obtained from the 5-month questionnaire
and complemented with the 18-month questionnaire, and
was broken into two categories: ‘breast-fed at least 3 months’
and ‘breast-fed less than 3 months or not breast-fed’.
Early determinants of overweight
L Dubois and M Girard
611
International Journal of Obesity
Table 1 Unadjusted odds ratio (with 95% confidence intervals) for birth weight over 4000 g
a
, weight-for-stature at the 95th percentile or above at 5 months
b
, and
BMI at the 95th percentile or above at 4.5 years
c
Description Category % At birth At 5 months At 4.5 years old
Birth weight44000 g Weight-for-StatureX95th percentile
(CDC USA Growth Curves)
BMIX95th percentile (CDC USA
Growth Curves)
P-value
d
OR Confidence intervals P-value
d
OR Confidence
intervals
P-value
d
OR Confidence
intervals
Pregnancy and birth data
Birth weight Less than 2500 g 4.3 FFFF0.0914 1.6 0.8 3.2 0.0005 1.0 0.4 2.6
2500–2999g 15.7 FFF 1.4 1.0 2.2 0.5* 0.2 1.0
3000–4000 g
e
69.3 FFF 1.0 FF 1.0 FF
More than 4000 g 10.7 FFF 1.6 1.0 2.5 2.3* 1.4 3.7
Baby’s sex Girl
e
48.9 0.0008 1.0 FF0.0039 1.0 FF 0.0283 1.0 FF
Boy 51.1 1.8* 1.3 2.6 1.6* 1.2 2.2 1.5* 1.0 2.2
Baby’s birth rank First 46.9 0.0012 0.5* 0.3 0.8 0.3245 1.3 0.8 2.1 0.1896 1.8 1.0 3.4
Second 37.5 1.0 0.6 1.6 1.5 0.9 2.4 1.6 0.9 3.1
Third or more
e
15.6 1.0 FF 1.0 FF 1.0 FF
Maternal smoking
during pregnancy
Not smoking
e
74.8 0.0014 1.0 FF0.2980 1.0 FF 0.0207 1.0 FF
Smoking 25.2 0.5* 0.3 0.7 0.8 0.6 1.2 1.6* 1.1 2.4
Single-parenting at
birth
No
e
91.7 0.9172 1.0 FFo0.0001 1.0 FF 0.4610 1.0 FF
Yes 8.3 1.0 0.6 1.9 3.0* 1.7 4.2 1.3 0.7 2.3
Household income
before pregnancy
Less than $20 000 16.4 0.9112 1.1 0.7 1.9 0.0039 2.3* 1.4 3.7 0.0013 2.2* 1.3 3.6
$20–39 999 28.1 1.0 0.7 1.6 1.6* 1.0 2.5 1.2 0.7 2.0
$40–59 999 26.2 0.9 0.6 1.5 1.2 0.7 1.9 0.7 0.4 1.3
$60 000 or mor e
e
29.3 1.0 FF 1.0 FF 1.0 FF
Mother’s age
group
Under 25 years 22.4 0.1564 0.7 0.4 1.1 0.3760 1.2 0.7 2.1 0.8940 0.8 0.4 1.5
25–29 years 31.0 0.6 0.3 1.0 1.3 0.8 2.2 0.9 0.5 1.6
30–34 years 32.0 0.7 0.5 1.2 1.0 0.6 1.6 0.8 0.5 1.5
35 years or older
e
14.6 1.0 FF 1.0 FF 1.0 FF
Mother’s
education level
No high school
diploma
19.1 0.1305 0.5 0.3 1.0 0.5636 1.3 0.8 2.1 0.4266 1.4 0.8 2.4
High school
diploma
26.9 0.7 0.5 1.1 1.280 0.8 2.0 1.2 0.7 2.1
College degree 27.9 0.8 0.5 1.2 1.1 0.7 1.7 1.5 0.9 2.5
University degree
e
26.1 1.0 FF 1.0 FF 1.0 FF
Mother’s
immigration status
Not immigrant
e
86.9 0.5257 1.0 FF0.5483 1.0 FF 0.1295 1.0 FF
Immigrant 13.1 1.2 0.7 1.9 0.9 0.5 1.4 1.5 0.9 2.4
Data collected when the children were 5 months old
Monthly weight
gain 0–5 months
(in quintiles)
Quintile 1 (Low)
e
19.8 FFFFo0.0001 1.0 FF 0.0261 1.0 FF
Quintile 2 20.1 FFF 0.9 0.4 0.9 1.0 0.5 1.8
Quintile 3 20.2 FFF 2.1* 1.1 3.9 0.7 0.4 1.4
Quintile 4 20.2 FFF 2.5* 1.3 4.7 0.9 0.5 1.6
Quintile 5 (High) 19.8 FFF 7.7* 4.3 13.8 1.6* 1.0 2.7
Breast-fed for 3
months
No 51.0 FFFF0.0156 1.5* 1.1 2.0 0.9074 1.0 0.7 1.5
Yes
e
49.0 FFF 1.0 FF 1.0 FF
Parental characteristics (data collected when the children were 18 months old)
Mother’s weight
(kg)
Mean (s.e.m.) 63
(0.3)
o0.0001 1.023* 1.023 1.048
Mother’s height
(m)
Mean (s.e.m.) 1.63
(0.001)
0.9638 1.068 0.062 18.357
Mother’s BMI Less than 18.5 6.5 o0.0001 0.7 0.3 1.9
18.5–24.9
e
65.2 1.0 FF
25.0–29.9 18.9 2.0* 1.3 3.1
30 or higher 9.4 3.4* 2.1 5.7
Father’s weight
(kg)
Mean (s.e.m.) 80
(0.3)
0.0013 1.025* 1.010 1.040
Early determinants of overweight
L Dubois and M Girard
612
International Journal of Obesity
Statistical analyses were performed with SAS (version 8.2).
All variables were treated as categorical variables. Preliminary
associations between the independent variables and weights
were verified through a w
2
test on contingency tables.
Significant independent variables were included in multi-
variate analyses. However, some variables were removed
because of an inter-correlation with household income.
These variables were no longer associated in multivariate
analyses, and in these cases, intermediate statistical analyses
indicated that the association with household income and
the studied variables was stronger than it was with these
variables. Crude and adjusted OR estimates, as well as their
confidence intervals, were made using logistic regressions.
Weighted data were used in the analysis, and the significance
level was set at 5%.
Results
Table 1 presents the unadjusted OR for the factors examined
according to whether birth weight was over 4000 g, at or
above the 95th percentile for weight-for-stature at 5 months,
and at or above the 95th percentile for BMI at 4.5 years.
At birth, the odds of weighing more than 4000 g were
higher for boys, whereas the odds were lower for first-born
children and for children born to smoking mothers. At 5
months, the odds of being at or over the 95th percentile for
weight-for-stature were higher for boys, for children born to
single mothers, for the poorest children, for children in the
highest quintiles of weight gain between birth and 5
months, and for children who were not breast-fed for at
least 3 months.
At 4.5 years, the odds of being at the 95th percentile or
higher for BMI were lower for children born with weights
between 2500 and 2999 g. The odds were higher for children
with birth weights of more than 4000 g, for boys, for children
born to mothers who smoked during pregnancy, and for
children living in poor families before pregnancy and at 4.5
years. Overweight at 4.5 years was also positively related to
high-weight-gain in the first 5 months of life, and to parental
overweight or obesity. Mother’s age, education and immi-
grant status, as well as mother’s and father’s height, were not
statistically associated with baby weights at the ages studied.
Table 2 presents the adjusted OR for being overweight at
4.5 years. The multivariate analysis was performed for weight
gain between birth and 5 months, maternal smoking during
pregnancy, parental overweight or obesity, and household
income. Data were adjusted for gestational age at birth, child
sex and birth weight.
Being in the highest quintiles of monthly weight gain
between birth and 5 months (OR 1.8 quintile 4; OR 3.9
quintile 5), as well as maternal smoking during pregnancy
(OR 1.8), increased the odds of being overweight at 4.5 years.
Parental overweight or obesity also increased the odds of
being overweight at this age. The odds doubled for children
with one overweight or obese parent, and tripled for children
with two overweight or obese parents. Being raised in
middle-income (OR 1.6) or in poor (OR 2.5) families also
increased the odds of being overweight at 4.5 years.
Table 1 (continued)
Description Category % At birth At 5 months At 4.5 years old
Birth weight44000 g Weight-for-StatureX95th percentile
(CDC USA Growth Curves)
BMIX95th percentile (CDC USA
Growth Curves)
P-value
d
OR Confidence intervals P-value
d
OR Confidence
intervals
P-value
d
OR Confidence
intervals
Father’s height (m)Mean (s.e.m.) 1.76
(0.001)
0.3570 0.247 0.013 4.848
Father’s BMI Less than 18.5 0.5 0.0059 NC FF
18.5–24.9
e
42.9 1.0 FF
25.0–29.9 44.6 1.7* 1.0 2.6
30 or higher 12.0 3.0* 1.7 5.3
No. of overweight
or obese parents
0 parent
e
37.6 FFFFFFFFo0.0001 1.0 FF
1 parent 46.9 F F F FFF 2.1* 1.3 3.4
2 parents 15.5 F F F FFF 3.7* 2.2 6.4
Parental and family characteristics (data collected when the children were 4.5 years old)
Household income Less than $20 000 10.1 FFFFFFFF0.0308 2.0* 1.1 3.6
$20–39 999 20.2 F F F FFF 1.6* 1.0 2.7
$40–59 999 26.4 F F F FFF 1.0 0.6 1.7
$60 000 or more
e
43.2 F F F FFF 1.0 FF
NC ¼not calculable. *Significant at Pp0.05.
a
10.7% of the children had a birth weight greater than 4000 g.
b
12.6% of the children had a weight-for-stature at 5
months at or above the 95th percentile on the CDC USDA Growth curves.
c
8.5% of the children had a BMI at 4.5 years at or above the 95th percentile on the CDC
USDA Growth curves.
d
Overall independent’s P-value from Type III analysis’s w
2
test of the logistic regression
e
Reference category.
Early determinants of overweight
L Dubois and M Girard
613
International Journal of Obesity
We performed further analyses to better understand
the relationship between maternal smoking during
pregnancy, weight gain in the first months of life and
overweight at 4.5 years. Figures 1 and 2 present the
proportion of 4.5-year-old overweight children by birth-
weight and weight-gain categories for those born to
nonsmoking mothers and for those born to smoking
mothers. For children born to nonsmoking mothers
(Figure 1), the data reveal a regular pattern. A greater
proportion of children born with higher weights (more
than 4000 g) were overweight at 4.5 years, the percentage
being greatest for those in the highest weight-gain
categories from birth to 5 months. This represents 31%
of the 4.5-year-old overweight children born to nonsmoking
mothers.
The pattern is different for children born to smoking
mothers (Figure 2). The greatest proportion of 4.5-year-old
overweight children was seen for children born in the
normal weight-range category (3000–4000 g), who were in
the highest quintiles of weight gain from birth to 5 months
(24% of the 4.5-year-old children with smoking mothers),
and for children with high birth weights (more than 4000 g),
who were in the lowest quintiles of birth-to-5-months
weight gain (23%).
Table 2 Adjusted
a
odds ratio (with 95% confidence intervals) for BMI at the 95th percentile or above at 4.5 years
Description Category At 4.5 years BMIX95th percentile (CDC USA growth curves)
P-value
b
OR 95% Confidence intervals
Monthly weight gain from birth to 5 months (in quintiles) Quintile 1 (low)
c
0.0006 1.0 FF
Quintile 2 1.1 0.6 2.2
Quintile 3 1.2 0.6 2.3
Quintile 4 1.8* 1.0 3.5
Quintile 5 (high) 3.9* 1.9 7.9
Mother’s smoking during pregnancy No
c
0.0093 1.0 FF
Yes 1.8* 1.2 2.8
No. of overweight or obese parents at 18 months 0 Parent
c
0.0011 1.0 FF
1 Parent 2.1* 1.3 3.6
2 Parents 3.2* 1.7 5.8
Household income at 4.5 years Less than $20 000 0.0065 2.5* 1.3 4.8
$20–39 999 1.6* 1.0 2.7
$40–59 999 0.9 0.5 1.3
$60 000 or mor e
c
1.0 - -
*Significant at Pp0.05.
a
Adjusted for gestational age and birth weight.
b
Overall independent’s P-value from Type III analysis’s w
2
test of the logistic regression.
c
Reference category.
Less t han
3000g 3000-4000g
More than
4000g
Q1+Q2
Q3
Q4+Q5
0
5
10
15
20
25
30
35
%
Birth weight (in grams)
Weight gai
n
(from 0 to 5
months, in
quintiles)
Figure 1 Proportion (adjusted for household income at 4.5 years) of
children born to non-smoking mothers who were overweight at 4.5 years, by
birth weight (in grams) and weight gain (from 0 to 5 months, in quintiles)
categories.
Less t han
3000g 3000-4000g More than
4000g
Q1+Q2
Q3
Q4+Q5
0
5
10
15
20
25
%
Birth weight (in grams)
Weight gain
(from 0 to 5
months, in
quintiles)
Figure 2 Proportion (adjusted for household income at 4.5 years) of
children born to smoking mothers who were overweight at 4.5 years, by birth
weight (in grams) and weight gain (from 0 to 5 months, in quintiles)
categories.
Early determinants of overweight
L Dubois and M Girard
614
International Journal of Obesity
Figure 3 presents the weight-for-height functions between
birth and 4.5 years for children at different percentiles of
BMI at 4.5 years, according to the population means for
these values at each age. For each 4.5-year BMI category,
children born to smoking mothers had birth weights lower
than the mean, and lower than the one for children born to
non-smoking mothers. After birth, in each BMI category,
children of smoking mothers gained more weight than
children of nonsmoking mothers. Children who were over-
weight at 4.5 years and who were born to smoking mothers
started their life with a birth weight around that for the
population means, but they gained more weight than
children from nonsmoking mothers in the first 5 months
of life. By comparison, children who were overweight at 4.5
years but who were born to nonsmoking mothers began life
with higher weights than the population mean, did not gain
more weight than the population mean between birth and 5
months, and had lower weights at 4.5 years than did those
born to smoking mothers.
Discussion
This study indicates that when a broad range of factors were
taken into consideration simultaneously, maternal smoking
during pregnancy, low socioeconomic status at 4. 5 years,
and parental overweight or obesity had the largest effect on
childhood overweight in the studied population. Parental
overweight or obesity, which may indicate a genetic
predisposition toward overweight in preschool years, could
act in concert with behavioral and social factors to exert a
critical influence on the onset of childhood overweight at
various stages of child development.
Parental height has previously been reported as being an
important factor in offspring overweight.
1
This factor,
however, was not related to childhood overweight in our
sample. It may be that the parents of our children had
reached their genetic potential for height because of better
socioeconomic conditions while they were growing up,
compared with conditions for parents from studies per-
formed in other countries or at other times.
Breast-feeding has also been reported by different authors
as being a protective factor against overweight and obesity.
5–7
We did not find this relationship in our study, however.
Univariate analysis indicates that breast-feeding reduces the
odds of having high weight-for-stature at 5 months, but not
the odds of being overweight at 4.5 years. These results are in
accordance with those of other authors who have reported
no relationship between breast-feeding and overweight in
childhood or adolescence.
9,10
Our results are similar to those
of Parsons et al.
9
which were also based on data from a birth
cohort using multiple confounding factors such as maternal
smoking during pregnancy, maternal BMI, social class, etc.,
and which followed children from pregnancy on. These
results may be more accurate than those from studies that
employed cross-sectional studies, from research that used
retrospectively reported data on breast-feeding and on
weight and height, and from studies that did not take into
consideration maternal smoking during pregnancy or ma-
ternal BMI. Breast-feeding clearly should be promoted
because it has many benefits for children and their mothers,
but from a public health perspective, it is less evident that it
can be cited as a protective factor for obesity. Mothers who
breast-fed should thus not avoid taking other measures to
prevent obesity as their children grow up.
Low birth weight followed by a rapid weight gain in the
first months of life, high birth weight, and high weight gain
in the first months of life have been associated with obesity
later in life.
2,3,14
Research on low birth weight and high birth
weight is generally performed separately, rather than on the
whole spectrum of birth weight in a population. In our
study, a higher rate of weight gain in the first 5 months of life
was related to overweight at 4.5 years, independently of
other factors in our multivariate analysis. This result is
similar to that of Stettler et al.
2
Our study goes further by
analyzing in more detail the interrelationships of all these
elements at the population level, and it adds the dimension
of maternal smoking to the analysis.
Maternal smoking during pregnancy, which is related to
lower birth weights, is associated with overweight at 4.5
years. This relationship between maternal smoking during
pregnancy and childhood overweight or obesity has been
reported in other studies performed in various countries.
15–17
Other research has indicated that maternal smoking during
pregnancy may influence obesity through its contribution to
lower birth weights, which are followed by periods of rapid
weight gain – or catch-up growth – in the first months of
life.
18
The rate of weight gain between birth and 4 months in
the USA has also been associated with overweight at 7 years,
2
-1.2
-1.0
-0.7
-0.5
-0.2
0.1
0.3
0.6
0.8
1.1
1.3
1.6
1.8
2.1
2.3
Child's age (months)
Mean z-score
Born from non-smoking mothers
Born from smoking mothers
050
≥95th
85-94th
10-84th
<10th
BMI according
to percentile on
CDC curves
5
Figure 3 Z-scores for weight, adjusted for height, for 4.5-year-old children
in four BMI percentile groups (CDC USA growth curves) by maternal smoking
during pregnancy.
Early determinants of overweight
L Dubois and M Girard
615
International Journal of Obesity
but this study did not take into account maternal smoking
during pregnancy.
In our study, univariate analysis indicates that low birth
weight is not per se a risk factor for overweight, since low-
birth-weight children generally remain smaller in preschool
years. Only a small proportion of these children – the
proportion being similar for children born to smoking and
non-smoking pregnant women – are overweight at 4.5 years,
even when they are in the highest quintiles of weight gain in
their first months of life. This finding accords with other
studies indicating that children born with lower weights
tend to stay smaller in early childhood,
19,20
given that high
birth weights are more consistently associated with later
obesity than low birth weights.
18
With reference to maternal smoking during pregnancy,
birth weight, and weight gain in the first months of life, our
analysis indicates that three groups of children were the
most prone to being overweight at 4.5 years. The first group is
made up of high-birth-weight children with nonsmoking
mothers, especially children with the greatest weight gain in
the first months of life. This group could have been subject
to conditions favouring high weight gain during pregnancy
that also continue to exert their influence after birth. The
second group is made up of normal birth weight children with
smoking mothers who are in the highest categories of weight
gain in the first months of life. It may be that even though
these children belong to the normal-birth-weight category,
their growth was initially restricted because of maternal
smoking but then caught up after birth. The third group is
made up of high-birth-weight children of smoking mothers
with the least weight gain in the first months of life. This
growth pattern could indicate that even in the presence of
maternal smoking during pregnancy, other factors were
strong enough to counteract the effect of tobacco, such that
body weight increased rapidly even in the absence of catch-
up growth. As nearly half of the overweight 4.5-year-old
children were in the first group, from a population health
perspective, interventions aimed at preventing childhood
obesity would do well to target smoking pregnant women, as
well as nonsmoking pregnant women at risk for giving birth
to high-birth-weight children, paying particular attention to
weight gain in the first months of life.
When all BMI categories are considered at 4.5 years, the
data show that compared with children born to nonsmoking
mothers, those born to smoking mothers gain more weight
than the mean in their first months of life. This could be the
result of catch-up-growth episodes for these children, even
when they are born at weights near the mean of the
population. Depending on birth weight, these accelerated-
growth episodes will not necessarily predispose them to
childhood overweight, which is most often the case for
children born at near-the-mean weights than for those born
at lower-than-the-mean weights. Children born to non-
smoking mothers did not gain more weight than the means,
which could indicate that their weight gains were not the
result of episodes of catch-up growth. This finding is
important given that the 1958 British Birth Cohort Study
concluded that the relationship between maternal smoking
and obesity could hold until adulthood.
17
Moreover, in the
USA, 60% of overweight 5–10-year-old children already have
associated cardiovascular disease risk factors.
21
In rich
countries like Canada, where mean birth weight is increas-
ing,
22–24
we should pay special attention to high-birth-
weight children in general and to those who gain more
weight after birth in particular.
With reference to the ‘foetal origin hypothesis’, even if a
greater proportion of children with high birth weights are at
risk of being overweight at 4.5 years, special attention should
be devoted to low-birth-weight/high-postnatal-weight-gain
children because they may be at risk for various clinical
conditions as children and adults.
25,26
Our study included children who could well become,
depending on their circumstances, obese in the future. They
were born and will be raised in a materially well-off society
with a universal health care system. Public health interven-
tions for preventing childhood overweight should prioritize
the reduction of maternal smoking during pregnancy, which
is related to lower birth weights but also increases the odds of
being overweight at 4.5 years. Public health policy should
also be directed at preventing high birth weights. After birth,
children from families of lower socioeconomic status and
those with overweight or obese parents should receive
particular attention. Children with these risk factors may
be predisposed to lifestyles such as high-caloric diets and
physical inactivity.
Acknowledgements
L Dubois’s research work is supported by the Canada
Research Chair Program. This study has been partly financed
by the Canadian Institutes of Health Research, Population
Health Initiative. This study is funded in part by the
Ministe
`re de la Sante
´et des Services sociaux du Que
´bec
(Quebec Ministry of Health and Social Services).
References
1 Parsons TJ, Power C, Logan S, Summerbell CD. Childhood
predictors of adult obesity: a systematic review. Int J Obes Relat
Metab Disord 1999; 23 (suppl 8): S1–107.
2 Stettler N, Zemel BS, Kumanyika S, Stallings VA. Infant weight
gain and childhood overweight status in a multicentre cohort
study. Pediatrics 2002; 109 (2): 194–199.
3 Ong KKL, Ahmed ML, Emmett PM, Preece MA, Dunger DB,
ALSPAC Study Team. Association between postnatal catch-up
growth and obesity in childhood: prospective cohort study. BMJ
2000; 320: 967–971.
4 Rasmussen KM. The ‘‘Fetal Origins’’ Hypothesis: challenges and
opportunities for maternal and child nutrition. Ann Rev Nutr
2001; 21: 73–95.
5 Gillman MW, Rifas-Shiman SL, Camargo CA, Berkey CS, Frazier
AL, Rockett HRH, Field AE, Colditz GA. Risk of overweight among
Early determinants of overweight
L Dubois and M Girard
616
International Journal of Obesity
adolescents who were breastfed as infants. J Am Med Assoc 2001;
285: 2461–2467.
6 Von Kries R, Koletzko B, Sauerwald T, von Mutius E, Barnert D,
Grunert V, von Voss H. Breast feeding and obesity: cross-sectional
study. BMJ 1999; 319: 147–150.
7 Armstrong J, Reilly JJ, The Child Health Information Team.
Breastfeeding and lowering the risk of childhood obesity. Lancet
2002; 359: 2003–2004.
8 Dubois L, Be
´dard B, Girard M, Beauchesne 2000 L’alimentation.
E
´tude longitudinale du de
´veloppement des enfants du Que
´bec (E
´LDEQ
1998–2002). Les nourrissons de 5 mois. Institut de la statistique du
Que
´bec: Que
´bec.
9 Parsons TJ, Power C, Manor O. Infant feeding and obesity
through the lifecourse. Arch Dis Childhood 2003; 88: 793–794.
10 Hediger ML, Overpeck MD, Kuezmarski RJ, Ruan WJ. Association
between infant breastfeeding and overweight in young children.
J Am Med Assoc 2001; 285 (19): 2453–2460.
11 Dubois L, Girard M 2002 L’alimentation des enfants d’a
ˆge
pre
´scolaire, E
´tude longitudinale du de
´veloppement des enfants du
Que
´bec (E
´LDEQ 1998-2002). Institut de la statistique du Que
´bec:
Que
´bec.
12 Cox GG, Cohen SB 1985 Methodological Issues for Health Care
Surveys. Marcel Dekker: New York, 446 pages.
13 National Center for Health Statistics and the National Center for
Chronic Disease Prevention and Health Promotion. 2000 CDC
Growth Charts. http://www.cdc.gov/growthcharts.
14 Tanaka T, Matsuzaki A, Kuromaru R, Kinukawa N, Nose Y,
Matsumoto T, Hara T. Association between birthweight and body
mass index at 3 years of age. Pediatr Int 2001; 43: 641–646.
15 Von Kries R, Toschke AM, Koletzko B, Slikker W. Maternal
smoking during pregnancy and childhood obesity. Am J Epidemiol
2002; 156: 954–961.
16 Wideroe M, Vik T, Jacobsen G, Bakketeig LS. Does maternal
smoking during pregnancy cause childhood overweight? Paediatr
Perinatal Epidemiol 2003; 17: 171–179.
17 Power C, Jefferis JMH. Fetal environment and subsequent
obesity: a study of maternal smoking. Int J Epidemiol 2002; 31:
413–419.
18 Dietz WH, Gortmaker SL. Preventing obesity in children and
adolescents. Ann Rev Public Health 2001; 22: 337–353.
19 Strauss RS, Dietz WH. Effect of intrauterine growth retardation in
premature infants on early childhood growth. J Pediatr 1997; 130:
95–102.
20 Hediger ML, Overpeck MD, Maurer KR, Kuczmarski RJ, McGlynn
A, Davis WW. Growth of infants and young children born small
or large for gestational age. Arch Pediatr Adolesc Med 1998; 152:
1225–1231.
21 Freedman DS, Dietz WH, Srinivasan SR, Berenson GS. The
relation of overweight to cardiovascular risk factors among
children and adolescents: The Bogalusa Heart Study. Pediatrics
1999; 103: 1175–1182.
22 Kramer MS, Morin I, Yang H, Platt RG, Usher R, McNamara H,
Joseph KS, Wen SH. Why are babies getting bigger? Temporal
trends in fetal growth and its determinants. J Pediatr 2002; 141:
538–542.
23 Bergmann KE, Bergmann RL, von Kries R, Bohm O, Richter R,
Dudenhausen JW, Wahn U. Early determinants of childhood
overweight and adiposity in a birth cohort study: role of breast-
feeding. Int J Obes Relat Metab Disord 2003a; 27: 162–172.
24 Bergmann RL, Ritcher R, Bergmann KE, Plagemann A, Brauer M,
Dudenhausen JW. Secular tarends in neonatal macrosomia in
Berlin: influences of potential determinants. Paediatr Perinatal
Epidemiol 2003b; 17 (3): 244–249.
25 Forsen T, Eriksson JG, Tuomilehto J, Osmond C, Barker DJP.
Growth in utero and during childhood among women who
develop coronary heart disease: longitudinal study. BMJ 1999;
319: 1403–1407.
26 Eriksson J, Forsen T, Tuomilehto J, Osmond C, Barker D. Fetal
childhood growth and hypertension in adult life. Hypertension
2000; 36: 790–794.
Early determinants of overweight
L Dubois and M Girard
617
International Journal of Obesity
