International Journal of Obesity (2009) 33, 12991308

& 2009 Macmillan Publishers Limited All rights reserved 0307-0565/09 $32.00
www.nature.com/ijo

ORIGINAL ARTICLE
Longitudinal study of the socio-demographic
determinants of changes in body weight and waist
circumference in a multi-ethnic Asian population
SK Ong1, CW Fong1, S Ma1, J Lee2, D Heng1, M Deurenberg-Yap3,4, Y-L Low5, M Tan6, W-Y Lim1
and ES Tai6
1

Epidemiology and Disease Control Division, Ministry of Health, Singapore, Singapore; 2Department of Epidemiology
and Public Health, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore; 3Health
Services Research and Evaluation Division, Ministry of Health Singapore, Singapore, Singapore; 4Duke-Graduate Medical
School, National University of Singapore, Singapore; 5Brenner Centre for Molecular Medicine, Singapore Institute for Clinical
Sciences (SICS), Singapore, Singapore and 6Department of Endocrinology, Singapore Health Services, Singapore, Singapore
Objective: To examine the changes in weight and waist circumference of adult Singaporeans between 1998 and 20052007,
and the associations of these changes with demographic and socio-economic factors.
Methodology: A prospective study, which followed up participants aged 1869 years from the 1998 National Health Survey.
Analysis was performed on data from 2483 individuals (53% of original sample) who returned for follow-up in 20052007. Body
weight and waist circumference were measured both at baseline and follow-up. Logistic regression was used to examine factors
associated with being overweight and obese at baseline. Linear regression was used to examine changes in weight and waist
circumference over time. The variables examined were age, gender, ethnicity, marital status, educational level, housing and
employment status, smoking, alcohol consumption and sports activities.
Results: Mean weight for the population increased over the follow-up period by 1.48 kg (s.d. 4.95) and mean waist
circumference increased by 3.32 cm (s.d. 7.92). Cross-sectionally, those who were overweight or obese were more likely to be
Malays or Indians, married, homemakers and have lower educational level. Prospectively, individuals who gained the most
weight were younger, more likely to be ethnic minority groups and have the lowest body mass index (BMI) at baseline. They
also appeared to be of higher socio-economic status (SES) based on housing type. These associations were statistically significant
even after adjusting for other variables.
Conclusion: Obesity prevention should start early in the younger age. Preventive programs need to reach out to Malay and
Indian ethnic groups and those with higher SES. These findings should be used in designing messaging of preventive strategies.
International Journal of Obesity (2009) 33, 12991308; doi:10.1038/ijo.2009.173; published online 8 September 2009
Keywords: longitudinal; Asian ethnic groups; weight gain; waist circumference; socio-demographic factors

Introduction
The prevalence of overweight and obesity has risen significantly over the past few decades in both developed and
developing countries worldwide, reaching epidemic proportions in many countries.14 In Singapore, cross-sectional
community-based National Health Surveys (NHS) have
found rising trends in obesity and overweight prevalence.
Correspondence: Dr ES Tai, Department of Endocrinology, Singapore General
Hospital, Outram Road, Block 6 level 6, Room B35, Singapore 169608,
Singapore.
E-mail: eshyong@pacific.net.sg
Received 28 April 2009; revised 14 July 2009; accepted 26 July 2009;
published online 8 September 2009

The prevalence of obesity (body mass index, BMI X30 kg m2)

among adult Singapore residents aged 1869 years rose
steadily from 5.1% in 1992 to 6.0% in 1998 and 6.9% in
2004, and the prevalence of overweight (BMI X25 kg m2)
has risen from 26.2% in 1992 to 30.4% in 1998 and 32.5% in
2004.5 With the rising obesity prevalence and the economic
costs associated with obesity,6 policy makers worldwide have
called for effective public health measures to tackle obesity
both at the population level and in high-risk groups.
A recent meta-analysis of 57 prospective studies involving
over 800 000 individuals has confirmed and quantified
the association between obesity and increased mortality,
primarily related to increased rates of diabetes, vascular,
renal and hepatic disease and to a lesser extent from

Longitudinal study on weight and waistline changes

SK Ong et al

1300
neoplastic disease.7 Although the majority of these studies
involved participants in Europe, the USA, Israel and
Australia, large studies from China8 and Korea9 suggest that
the impact of obesity on mortality is similar in populations
of Asian ancestry. Many studies have also documented the
association of obesity with the occurrence of many high
burden chronic diseases.1013
Despite the large number of studies documenting the
negative impact of obesity on morbidity and mortality, few
studies have included repeated measurements of obesity,
which would help to determine whether there are critical
periods in life when individuals become overweight or
obese, and also whether it is possible to identify factors that
would predict weight gain before it occurs. Such studies
would help policy makers develop appropriate preventive
strategies that target the population at greatest risk, at a
time when the risk of weight gain is the greatest. Various
studies have shown that weight gain is a function of baseline
age and BMI, as well as other lifestyle and demographic
factors including dietary intake, physical activity, socioeconomic differences, and to a lesser extent ethnicity and
gender.1417 To our knowledge, few (if any) studies of this
nature have been conducted in Asian populations. Although
the effects of obesity on mortality and chronic disease seem
similar between populations, it is conceivable that the
patterns of weight gain and the factors that predict weight
gain may differ because of differences in patterns of dietary
intake and physical activity between populations or ethnic
groups. It is also possible that different populations may
have different susceptibility to weight gain in an environment of nutrient excess. In addition to weight gain,
increased abdominal fat or waist circumference has been
found to be strongly associated with metabolic and cardiovascular diseases.18 Those with a normal BMI, but large waist
circumference, are also found to have higher mortality
risk.19,20 The aim of this study is to examine the function
of age, baseline BMI and socio-demographic factors in
predicting changes in weight and waist circumference in a
multi-ethnic Asian population living in Singapore.

Materials and methods

Study population
The NHS 1998 methodology has been described in detail
elsewhere.21 Briefly, the survey was a cross-sectional national
survey, which measured the prevalence of major cardiovascular risk factors including obesity, physical inactivity
and smoking among adult Singaporeans aged 1869 years
old. It involved initial identification of individuals by household addresses, followed by stratified sampling of individuals
by age and ethnic group. The two ethnic minority groups,
Malays and Indians, were over-sampled to ensure sufficient
sample size for reliable comparative analysis. A total of 4723
individuals were studied in the NHS 1998.
International Journal of Obesity

The individuals from the NHS 1998 were re-contacted

during the period of 20052007 as part of the Singapore
Prospective Study Program.22,23 Individuals deceased at time
of follow-up (as confirmed by data linkage to the national
registry of births and deaths) were excluded (n 98).
Individuals who had emigrated or had errors in their unique
national identity card numbers, which were used to establish
re-contact, were also excluded (n 33). The rest were
contacted to obtain an appointment for investigators to
administer the questionnaire at the participants home.
Individuals who could not be reached after three home
visits (on separate occasions including one weekend and
weekday) were deemed non-contactable (n 833). Of those
who were contacted, 10 refused to participate. Those who
were contacted and agreed to undergo the follow-up survey
were invited to attend a health examination at a clinic.
A total of 3749 (79%) individuals completed the questionnaire, of which 2483 (53%) also completed both the health
survey and physical examination. In this study, only the
latter were included.

Measurements
Body weight (to the nearest 0.01 kg) was obtained with
participants dressed in light clothing using electronic
weighing scales; height (to the nearest 0.1 cm) was measured
using wall-mounted stadiometer for all the participants
without shoes. Waist circumference (to the nearest 0.1 cm)
was measured at the midpoint between the lower rib margin
and the iliac crest using measuring tape. Participants were
instructed to remove any objects such as keys and mobile
phone before measurement.
The baseline socio-demographic variables included in this
study were gender, ethnic groups, marital status, age, highest
educational level, housing type, employment status and BMI
categories. Highest educational level, housing type and
employment status were used as proxy measures for socioeconomic status (SES). The highest educational level attained
was defined as (i) no formal education received o6 years of
education, (ii) primary education received 67 years
of education, (iii) secondary education received 710 years
of education, (iv) pre-university/diploma received 1113
years of education and (v) degree/professional qualification received X14 years of education. Housing type was
classified as (i) public apartment with 13 room, (ii) public
apartment with four rooms and higher, (iii) private apartment including condominium and (iv) private landed
property. Public apartment refers to subsidized apartment
built by the government. Private apartment refers to apartment or condominium built by private property developers.
A total of 85% of Singapore residents live in public
apartments, 10% in private apartments and 5% in private
landed properties.24 In general, the cost of housing increases
from public apartment to private apartment and landed
property. As 85% of the population lives in public housing,
this variable may not have the same interpretation as in

Longitudinal study on weight and waistline changes

SK Ong et al

1301
other populations. Therefore, the type/size of public apartments, which are associated with cost, is highly relevant
in this context. For employment status, participants were
also grouped into five categories (working, non-working,
homemaker, students and unemployed).
BMI categories were defined based on WHO classification25 of BMI for adultsFunderweight: BMI o18.5 kg m2;
normal weight: BMI 18.5 to o24.9 kg m2; pre-obese or overweight: BMI 25.0 to 29.9 kg m2 and obese: BMI 30.0 kg m2
or more.
Lifestyle factors included in the analyses were baseline
sports activity, smoking and alcohol consumption. Sports
activity was based on the frequency of participation in sports
for at least 20 min per occasion with three categories: regular
exercise (3 days a week), occasional exercise (p3 days a week)
and no exercise (no participation in any form of sports that
lasted for at least 20 min per occasion).26 Smoking status was
obtained using a questionnaire based on the WHO classification criteria for smoking: non-smoker (never smoked before
or smoked too little in the past to be regarded as an exsmoker), ex-smoker (formerly a daily smoker, but currently
does not smoke at all), occasional smoker (smokes cigarettes,
but not everyday) and daily smoker (smokes cigarettes
at least once a day including those who smoke every day,
but have to stop temporarily because of religious fasting
or medical records).27 Alcohol consumption was classified
according to the frequency of alcohol intake: regular drinker
(drinks 44 days a week), frequent drinker (drinks 13 days a
week), occasional drinker (drinks o3 days a month) and
non-drinkers are those who do not drink alcohol.

Data analysis
We carried out linear regression analysis for the change
in weight and waist circumference for all participants.
Change in weight and waist circumference is the difference
between the measurements at baseline and follow-up. The
two measurements were correlated (r 0.526, Po0.001).
The weight, waist circumference and BMI of the cohort were
normally distributed.
We used Students t-test and test of binomial proportions
to examine baseline differences between individuals who
returned for follow-up and were included in the analyses
(respondents) and those who did not (non-respondents).
We examined the association between the risk of being
overweight (BMIX25.0 kg m2) or obese (BMIX30 kg m2)
and the baseline characteristics including age, gender, ethnic
groups, marital status, educational level, housing type and
employment status, using univariate and multiple logistic
regression adjusting for baseline smoking, alcohol consumption and sports activity status.
In the prospective analysis, changes in weight and waist
circumference were calculated by subtracting weight at baseline from weight at follow-up. Repeated measures analysis of
variance were used to examine whether the changes were
different among the categories of variable. We used linear

regression to examine the associations between baseline

factors (age, gender, ethnicity, socio-economic and lifestyle
factors and baseline BMI) and change in weight and waist
circumference.
Earlier studies have reported that the association between
gender and obesity was different between ethnic groups.28
For this reason, we first tested the interaction between
gender and ethnicity in relation to overweight and obesity.
The interaction terms were highly significant (Po0.001).
Therefore, instead of analysing gender and ethnicity separately, we created a new variable (ethnicity/gender), which
included six levels (Chinese man, Chinese woman, Malay
man, Malay woman, Indian man and Indian woman), and
this variable was used to assess the impact of ethnicity and
gender on weight gain. We first performed univariate
analyses (results not shown), second age-adjusted analyses
and then followed by multivariate analyses that adjusted for
all covariates.

Results
Table 1 presents the baseline characteristics of participants
who did (n 2483) and did not (n 2240) return for followup. Those who returned for follow-up were more likely to
be of Chinese ethnicity, be married and working. More
individuals aged 4049 years returned for follow-up than
those aged 1829 years. In addition, they were more likely to
be non-smokers, to have a BMI in the normal range and have
a higher level of education.
The mean duration of follow-up was 7.68 years
(s.d. 0.92). The mean weight and waist circumferences
at baseline were 62.02 kg (s.d. 12.43) and 79.81 cm (s.d.
11.23), respectively. Over the follow-up period, mean weight
increased by 1.48 kg (s.d. 4.95), mean waist circumference
increased by 3.32 cm (s.d. 7.92) and mean BMI increased by
0.18 kg m2 (s.d. 1.84). A total of 73.9% (n 1832) of the
participants had weight change ranging from 5 to 5 kg;
1534 participants (61.9%) had gained weight, whereas
916 participants (36.9%) lost weight. The majority (78.3%,
n 1944) of the participants remained in their baseline BMI
categories. A total of 10.6% (n 159) of participants with
normal baseline BMI (18.524.9 kg m2) became overweight
at follow-up, and 9.4% (n 56) of those with baseline BMI
25.029.9 kg m2 were obese at follow-up (data not shown).
Cross-sectional analyses of baseline factors revealed that
age, marital status, educational level, housing types and
employment status were significantly associated with an
increased risk of being overweight or obese (Table 2). Both
overweight and obesity were associated with age and the risk
increased in a monotonic manner except in the oldest age
group (age 6069) in which the risk declined. Being married,
separated or divorced and from lower educational level were
also associated with increased risk of overweight and obesity.
As discussed in a preceding section, there was a statistically
significant interaction between ethnicity and gender in
International Journal of Obesity

Longitudinal study on weight and waistline changes

SK Ong et al

1302
Table 1

Baseline characteristics of respondents and non-respondents

Respondents
(n 2483) %

Non-respondents
(n 2240) %

Gender
Male
Female

46.4
53.6

45.9
54.1

Ethnic group
Chinese
Malay
Indian

70.2
16.1
13.7

66.3
20.0
13.6

Age (year)
1829
3039
4049
5059
6069

20.7
33.1
29.4
11.0
5.8

30.9
30.6
19.3
10.1
9.1

Marital status
Single
Married
Divorced/widowed/separated

22.8
74.8
2.3

32.1
65.1
2.9

11.5
24.1
33.6
17.2
13.6

17.0
21.3
29.6
17.5
14.6

22.4
61.4

23.8
61.2

7.5
8.7

6.5
8.4

Employment status
Working
Homemaker
Student
Retired
Unemployed

71.1
17.3
5.8
2.8
2.9

66.0
18.9
6.9
2.9
5.3

Smoking status
Non-smoker
Daily smoker
Ex-smoker
Occasional smoker

79.6
13.0
5.4
2.0

76.3
16.7
4.9
2.2

Alcohol consumption
Non-drinker
Occasional drinker
Frequent drinker
Regular drinker

62.1
30.2
5.2
2.5

64.2
28.3
5.7
1.9

Sports activity
No exercise
Occasional exercise
Regular exercise

56.9
26.9
16.2

56.5
27.1
16.4

BMI group (kg m2)

o18.5
18.524.9
25.029.9
X30.0

8.3
60.5
24.0
7.2

9.9
56.3
26.3
7.5

Baseline characteristic

Education level
No formal education
Primary
Secondary
Pre-university/diploma
Degree/professional
qualification
Housing type
Public apartment 13 rooms
Public apartment 4 rooms
and above
Private apartment
Landed property

International Journal of Obesity

P-value

0.753

0.002

o0.001

o0.001

o0.001

0.424

o0.001

0.004

0.220

0.960

0.024

relation to the risk of being overweight or obese. Among the

Chinese, women (as compared with men) were less likely
to be overweight (odds ratio, OR 0.46, 95% confidence
interval, CI 0.35, 0.62) or obese (OR 0.45, 95% CI 0.25, 0.81).
Although the Malays and Indians, irrespective of gender,
were more likely to be overweight or obese than the Chinese
men, the effect of gender in these ethnic groups was the
reverse of that observed in Chinese, with greater risk
observed in women than in men. Compared with Chinese
men, the Malay women had more than seven times risk
of being obese, whereas the Indian women had more than
four times greater risk of being obese. In multivariable
analysis including all the other factors, the adjusted ORs
for ethnicity/gender, marital status, educational level and
employment status remained statistically significant for
being overweight, whereas that of ethnicity/gender and
educational level remained statistically significant for
obesity.
Table 3 presents the mean weight and waist circumference
of the participants by their baseline characteristic groups. In
most groups, the means of weight increased over the followup period. Those aged 6069 at baseline experienced the
most significant loss of weight. For most variables, different
groups showed a statistically significant difference in weight
or waist circumference. For this reason, we have highlighted
those variables that are not associated with significant
difference in weight or waist change between the groups.
The change in weight did not differ in a statistically
significant manner between those living in different housing
types. All groups experienced an increase in waist circumference. The magnitude of change in waist circumference
was not associated with housing type, alcohol consumption
and sports activities.
Table 4 shows the effects of baseline demographic and
socio-economical factors on the change in weight and
waist circumference using linear regression. In general, the
associations observed were similar in relation to changes in
both weight and waist circumference. The greatest increases
for both these parameters were seen in the youngest age
group (age 1829 years) with progressively smaller gains in
weight and waist circumference as age increased. In our
initial univariate analyses (results not shown), statistically
significant associations were observed in relation to marital
status (greater weight gain in those who were single compared with those who were married or divorced/widowed/
separated) and education level (greater weight gain in those
with higher education level), but these were no longer
statistically significant after adjustment for age and other
variables. The greatest increases in weight and waist circumference were also seen in those with BMI in the range
18.524.9 kg m2 at baseline and in those who lived in the
larger public apartments and private homes. In relation to
ethnicity and gender, Malays and Indians experienced
greater increase in weight and waist circumference, as
compare with Chinese. In both these minority ethnic
groups, women experienced higher increases in weight than

Longitudinal study on weight and waistline changes

SK Ong et al

1303
Table 2

Cross-sectional analyses of odds ratio for overweight and obesity at baseline

Baseline characteristic

Overweight
Multivariate a

Age-adjusted
OR

Obesity

95% CI

OR

95% CI

Multivariate a

Age-adjusted
OR

95% CI

OR

95%CI

Age (year)
1829
3039
4049
5059
6069

1.00
2.69
3.45
3.69
3.18

(1.96,
(2.51,
(2.52,
(2.02,

3.68)
4.74)
5.40)
5.02)

1.00
1.44
1.63
1.72
1.30

(0.98,
(1.09,
(1.07,
(0.70,

2.13)
2.43)
2.77)
2.42)

1.00
1.86
2.63
3.32
1.53

(1.09,
(1.56,
(1.84,
(0.66,

3.18)
4.44)
5.99)
3.57)

1.00
0.86
1.05
1.32
0.57

(0.45,
(0.55,
(0.62,
(0.20,

1.63)
2.01)
2.80)
1.63)

Ethnicity/gender
Chinese man
Chinese woman
Malay man
Malay woman
Indian man
Indian woman

1.00
0.58
2.00
2.22
1.49
2.65

(0.46,
(1.41,
(1.54,
(1.02,
(1.84,

0.74)
2.83)
3.20)
2.17)
3.82)

1.00
0.46
1.69
1.59
1.42
1.87

(0.35,
(1.15,
(1.03,
(0.96,
(1.23,

0.62)
2.48)
2.45)
2.09)
2.85)

1.00
0.63
2.78
7.25
2.35
4.34

(0.38, 1.06)
(1.59, 4.86)
(4.54, 11.57)
(1.28, 4.32)
(2.58, 7.30)

1.00
0.45
2.21
4.27
2.28
2.57

(0.25,
(1.19,
(2.33,
(1.22,
(1.38,

0.81)
4.10)
7.84)
4.26)
4.80)

Marital status
Single
Married
Divorced/widowed/separated

1.00
2.45
3.02

(1.80, 3.33)
(1.56, 5.83)

1.00
1.45
1.82

(1.02, 2.06)
(0.90, 3.68)

1.00
2.61
3.65

(1.49, 4.57)
(1.35, 9.84)

1.00
1.42
1.25

(0.75, 2.69)
(0.42, 3.71)

Education
No formal education
Primary
Secondary
Pre-university/diploma
Degree/professional qualification

1.00
1.24
0.90
0.58
0.70

(0.89,
(0.64,
(0.39,
(0.47,

1.72)
1.24)
0.86)
1.04)

1.00
0.98
0.81
0.59
0.61

(0.69, 1.39)
(0.5, 1.16)
(0.38, 0.90)
(0.38, 0.95)

1.00
0.81
0.55
0.20
0.33

(0.52, 1.26)
(0.34, 0.88)
(0.10, 0.41)
0.17, 0.63)

1.00
0.70
0.60
0.30
0.43

(0.43,
(0.36,
(0.14,
(0.20,

Housing type
Public apartment 13 rooms
Public apartment 4 rooms and above
Private apartment
Landed property

1.00
0.97
0.73
0.67

(0.76, 1.22)
(0.48, 1.09)
(0.46, 1.00)

1.00
1.09
1.19
1.19

(0.85, 1.40)
(0.76, 1.87)
(0.77, 1.83)

1.00
0.61
0.32
0.42

(0.44, 0.86)
(0.14, 0.71)
(0.22, 0.81)

1.00
0.78
0.89
1.01

(0.54, 1.12)
(0.37, 2.14)
(0.49, 2.08)

Employment status
Working
Homemaker
Student
Retired
Unemployed

1.00
1.33
0.23
0.63
0.78

(1.04, 1.69)
(0.11, 0.51 )
(0.35, 1.12)
(0.42, 1.45)

1.00
1.31
0.29
0.94
0.82

(0.96,
(0.13,
(0.49,
(0.43,

1.00
2.15
0.29
0.28
0.72

(1.53,
(0.07,
(0.07,
(0.22,

1.00
1.50
0.28
0.49
0.79

(0.95,
(0.06,
(0.10,
(0.23,

1.78)
0.69)
1.83)
1.60)

3.02)
1.24)
1.21)
2.33)

1.14)
1.02)
0.66)
0.93)

2.36)
1.31)
2.26)
2.69)

Abbreviations: OR, odds ratio; CI, confidence interval. aAdjusted for age, gender, ethnicity, marital status, educational level, housing type, employment status and
baseline smoking, alcohol consumption and sports activity status. For age adjustment analyses, the results for the age categories were not adjusted for age. Bold:
statistically significant with P-value o0.05, 95% CI does not include 1.

men and this gender difference was more obvious in relation

to increase in waist circumference.

Discussion
This study has allowed us to correlate the risk of overweight/
obesity and weight gain with socio-demographic factors in
a multi-ethnic Asian population, both cross-sectionally
and prospectively. Cross-sectional analyses revealed that
the prevalence of overweight and obesity increase with age
and are most prevalent among the middle aged individuals
(5059 years). By using a prospective study design, we have

determined that the greatest weight gain occurs in early

adulthood and middle age. We found that over one-third
(54/155 participants) who became overweight during the
period of follow-up were aged 3039 years at baseline (data
not shown). This finding is consistent with the findings
from several other studies.2931 Such weight gain after young
adulthood has also been shown to be associated with
increased risk of cardiovascular diseases and metabolic
syndrome,3234 and adverse effect on quality of life and
long-term mortality.35 In addition, the greatest weight gain
occurred in those with normal or low BMI at baseline, a
finding that also replicates those of earlier studies.7 Taken
together, these data suggest that effective prevention of
International Journal of Obesity

Longitudinal study on weight and waistline changes

SK Ong et al

1304
Table 3

Means of weight and waist circumference at baseline and the changes by baseline characteristics

Baseline characteristics

Weight in kg, means (s.d.)

Baseline

Waist circumference in cm, means (s.d.)

Change a
1.48 (4.95)

Baseline

Change a

79.81 (11.23)

3.32 (7.92)

All

62.02 (12.43)

Age (years)
1829
3039
4049
5059
6069

59.18
62.01
63.46
64.09
60.91

(12.88)
(12.65)
(12.07)
(12.28)
(9.44)

3.58
1.91
0.73
0.42
1.01

(5.80)
(4.50)
(4.52)
(4.33)
(3.94)

73.76
79.07
82.01
84.86
84.87

(10.01)
(10.95)
(10.80)
(10.99)
(9.13)

4.90
3.39
3.15
1.68
1.11

(6.92)
(9.89)
(5.80)
(7.55)
(7.48)

Ethnicity/gender
Chinese man
Chinese woman
Malay man
Malay woman
Indian man
Indian woman

67.11
54.46
69.81
62.55
70.58
62.58

(11.35)
(8.82)
(12.32)
(12.38)
(11.41)
(12.44)

1.54
0.78
2.69
2.56
1.87
2.11

(4.90)
(4.05)
(5.90)
(5.75)
(5.36)
(6.32)

83.98
72.96
86.28
79.55
89.23
82.34

(9.77)
(8.26)
(10.38)
(12.51)
(10.03)
(11.1)

2.70
2.87
4.92
5.25
3.61
4.27

(8.15)
(6.28)
(9.50)
(9.69)
(10.45)
(7.32)

Marital status
Single
Married
Divorced/widowed/separated

58.55 (12.5)
63.08 (12.28)
61.85 (10.53)

Education level
No formal education
Primary
Secondary
Pre-university/diploma
Degree/professional qualification

62.03
63.59
61.25
60.93
62.52

(11.49)
(12.29)
(13.08)
(11.97)
(12.15)

0.10
1.00
2.04
1.88
1.82

Housing typeb,c
Public apartment 13 rooms
Public apartment 4 rooms and above
Private apartment
Landed property

62.18
62.10
60.84
62.08

(12.65)
(12.29)
(13.00)
(12.41)

Employment status
Working
Homemaker
Student
Retired
Unemployed

63.05
59.61
57.11
61.14
61.80

Smoking Status
Non-smoker
Daily smoker
Ex-smoker
Occasional

73.91 (10.08)
81.53 (10.98)
82.63 (10.39)

4.83 (6.37)
2.86 (8.35)
3.40 (5.56)

(4.88)
(5.10)
(5.10)
(4.54)
(4.54)

83.06
82.22
78.42
77.56
79.14

(10.35)
(10.95)
(11.72)
(10.70)
(10.57)

2.01
3.09
3.62
3.41
3.97

(7.86)
(7.05)
(8.40)
(6.40)
(9.69)

1.23
1.71
1.26
0.81

(5.51)
(4.85)
(3.96)
(4.82)

79.93
79.88
78.63
80.08

(11.09)
(11.36)
(11.01)
(10.94)

2.88
3.40
3.99
3.32

(7.42)
(8.10)
(9.84)
(5.84)

(12.49)
(11.45)
(13.05)
(10.62)
(12.85)

1.46
0.64
4.54
0.93
3.30

(4.71)
(4.67)
(6.45)
(4.07)
(6.42)

80.57
79.22
71.22
84.20
78.05

(11.37)
(10.09)
(9.24)
(10.04)
(11.23)

3.13
3.39
5.51
0.21
6.08

(8.11)
(5.81)
(6.83)
(8.65)
(12.65)

60.71
66.50
69.91
63.84

(11.87)
(13.05)
(13.67)
(12.15)

1.44
1.94
0.67
2.47

(4.87)
(5.52)
(4.19)
(5.79)

78.58
83.94
88.00
80.32

(10.89)
(10.97)
(11.37)
(10.92)

3.39
3.35
1.70
4.59

(7.65)
(10.14)
(5.93)
(6.25)

Alcohol consumptionc
Non-drinker
Occasional
Frequent
Regular drinker

61.29
63.00
63.74
64.78

(12.40)
(12.57)
(11.61)
(11.88)

1.55
1.45
0.93
1.54

(5.10)
(4.90)
(3.81)
(4.17)

79.41
80.13
81.37
82.83

(11.21)
(11.48)
(10.07)
(10.62)

3.53
2.91
3.49
2.80

(7.68)
(7.58)
(11.93)
(7.44)

Sports activity statusc

No exercise
Occasional
Regular

61.21 (12.44)
63.07 (13.07)
63.13 (11.08)

1.31 (4.91)
2.05 (4.73)
1.18 (5.41)

79.68 (11.43)
79.77 (11.58)
80.33 (9.91)

3.17 (8.55)
3.51 (6.77)
3.57 (7.43)

BMI group (kg m2)

o18.5
18.524.9
25.029.9
X30.0

46.45
57.84
70.98
84.87

2.85
1.63
1.06
0.15

65.59
76.16
88.18
98.76

4.53
3.39
3.07
2.17

(5.08)
(7.96)
(8.06)
(12.22)

3.18 (5.92)
1.01 (4.52)
0.18 (4.33)

(4.14)
(4.55)
(5.48)
(6.52)

(4.97)
(7.91)
(7.05)
(9.77)

(6.78)
(8.05)
(7.93)
(7.90)

Abbrerviation: SD, standard deviation. aChange in weight and waist circumference is the difference between the measurements at baseline and follow-up. bChanges
in weight were NOT statistically significant different among the categories, with P-value 40.05 using repeated measures of ANOVA. cChanges in waist were NOT
statistically significant different among the categories, with P-value 40.05 using repeated measures of ANOVA.

International Journal of Obesity

Longitudinal study on weight and waistline changes

SK Ong et al

1305
Table 4

Effects of baseline variables on change in weight and waist circumference by linear regression

Baseline characteristic

Change in weight (kg)

Multivariate a

Age-adjusted
b

Change in waist circumference (cm)

95% CI

Multivariate a

Age-adjusted

95% CI

Age (year)
1829
3039
4049
5059
6069

1.67
2.85
4.00
4.59

2.20,
3.39,
4.70,
5.47,

1.15
2.31
3.30
3.71

0.91
2.08
3.03
3.59

1.58,
2.79,
3.92,
4.80,

0.23
1.36
2.15
2.39

1.52
1.76
3.23
3.80

Ethnic/gender
Chinese man
Chinese woman
Malay man
Malay woman
Indian man
Indian woman

0.92
1.20
0.91
0.73
0.78

1.36,
0.46,
0.19,
0.08,
0.01,

0.47
1.94
1.63
1.52
1.54

0.97
1.34
1.51
1.01
1.34

1.49,
0.54,
0.66,
0.21,
0.49,

0.46
2.14
2.35
1.81
2.20

0.06
2.25
2.46
1.23
1.72

Marital status
Single
Married
Divorced/widowed/separated

0.52
0.42

1.05, 0.03
1.78, 0.95

0.60
0.56

1.22, 0.02
1.93, 0.81

0.78
0.44

95% CI

2.38,
2.65,
4.38,
5.26,

0.65
0.87
2.07
2.34

0.66
0.90
2.12
2.23

1.79,
2.09,
3.60,
4.24,

0.47
0.29
0.65
0.22

0.69, 0.79
1.03, 3.48
1.26, 3.66
0.1, 2.54
0.44, 2.99

0.24
2.77
2.91
1.44
1.92

1.10,
1.43,
1.50,
0.10,
0.49,

0.62
4.10
4.32
2.77
3.34

1.66, 0.12
1.83, 2.7

1.11
0.02

2.14, 0.08
2.33, 2.28

Education level
No formal education
Primary
Secondary
Pre-university/diploma
Degree/professional qualification

0.27
0.60
0.20
0.37

0.42,
0.09,
0.56,
0.42,

0.95
1.28
0.96
1.16

0.22
0.39
0.07
0.17

0.48,
0.32,
0.74,
0.71,

0.91
1.10
0.88
1.05

0.43
0.41
0.01
0.75

0.70,
0.72,
1.25,
0.55,

1.56
1.53
1.27
2.05

0.44
0.29
0.05
0.81

Housing type
Public apartment 13 rooms
Public apartment 4 rooms and above
Private apartment
Landed property

0.58
0.34
0.39

0.12, 1.04
0.45, 1.13
0.37, 1.15

0.68
0.95
0.61

0.21, 1.15
0.12, 1.78
0.19, 1.40

0.61
1.36
1.09

0.16, 1.37
0.05, 2.66
0.17, 2.34

0.92
2.06
1.46

Employment status
Working
Homemaker
Student
Retired
Unemployed
Smoking status
Non-smoker
Daily smoker
Ex-smoker
Occasional
Alcohol consumption
Non-drinker
Occasional
Frequent
Regular drinker
Physical activity status
Regular
Occasional
No exercise
BMI group (kg m2)
18.524.9
o18.5
25.029.9
X30.0

0.35
0.89
0.3
0.89

0.86,
0.01,
0.94,
0.25,

95% CI

0.71,
0.89,
1.30,
0.65,

1.60
1.47
1.40
2.27

0.13, 1.69
0.67, 3.44
0.13, 2.78

0.16
1.79
1.54
2.01

0.06
0.86
0.24
0.75

0.65,
0.11,
1.61,
0.39,

0.53
1.82
1.12
1.89

0.6
0.79
0.96
2.25

0.24, 1.44
0.69, 2.27
3.01, 1.09
0.4, 4.11

0.58
0.56
1.41
1.93

0.41,
1.06,
3.69,
0.02,

0.66
0.10
0.67

0.10, 1.22
0.76, 0.94
0.69, 2.02

0.28
0.33
0.42

0.35, 0.91
1.20, 0.55
0.93, 1.76

0.07
1.06
0.93

0.86, 1.00
2.45, 0.34
1.31, 3.15

0.02
1.10
0.67

1.07, 1.03
2.55, 0.36
1.57, 2.91

0.35
0.72
0.73

0.76, 0.07
1.57, 0.13
0.48, 1.94

0.17
0.86
0.50

0.62, 0.28
1.74, 0.02
0.73, 1.73

0.81
0.11
0.15

1.50, 0.12
1.52, 1.29
2.15, 1.85

0.21
0.34
0.33

0.96, 0.54
1.13, 1.80
1.73, 2.38

0.02
0.42

0.55, 0.51
0.03, 0.86

0.30
0.23

0.84, 0.24
0.23, 0.68

0.49
0.08

0.39, 1.36
0.65, 0.81

0.13
0.07

0.78, 1.03
0.82, 0.69

0.57
0.77
1.64

1.27, 0.14
1.53, 0.01
2.6, 0.68

0.75
1.37
2.57

1.45, 0.06
2.15, 0.59
3.57, 1.58

0.64
0.66
1.55

1.8, 0.53
1.93, 0.62
3.14, 0.06

0.90
1.34
2.78

2.07, 0.26
2.64, 0.03
4.44, 1.12

1.57
2.16
0.87
3.83

Abbreviation: CI, confidence interval. aAdjusted for age, gender, ethnic group, marital status, educational level, housing type, baseline BMI, smoking, alcohol
consumption and sports activity status. For age adjustment analyses, the results for the age categories were not adjusted for age. Bold: changes of weight and waist
circumference were found to be statistically significant with P-value o0.05 by linear regression.

International Journal of Obesity

Longitudinal study on weight and waistline changes

SK Ong et al

1306
obesity should begin in early adulthood, before the development of obesity. This is particularly important as the
treatment of long-standing obesity after it has occurred is
often more difficult and the effect on cardiovascular risks
may not be reversible.36
Similar to an earlier study,28 we found that the effect of
gender on obesity is ethnic specific. Malays and Indians had
the highest levels of overweight and obesity at baseline, and
obesity was more pronounced in women from these ethnic
groups. During the period of follow-up, we found that Malay
men gained the most weight, whereas both Malay men and
Malay women exhibited the greatest increase in waist
circumference. In contrast, Chinese women had the lowest
levels of overweight and obesity at baseline and were the
least likely to gain weight or increase their waist circumference. Such ethnic differences in susceptibility to obesity
have been reported earlier.3739 However, the basis of these
ethnic differences is unclear. In some earlier studies, these
differences were attenuated when SES was taken into
account suggesting that these ethnic differences may be
related to differences in SES.37,40 In our study, ethnicity
remains a significant predictor of weight gain and obesity,
even after adjustment for educational level, housing type
and employment status.
Our finding on the association of marital status with
increased risk of being overweight is similar to earlier studies
in which it was found that the risk was attenuated by the
effect of ethnicity and gender.41,42 In our prospective
analyses, being married or having been married is associated
with weight loss during adulthood, but this association was
not statistically significant after adjusting for age and other
covariates. We believe that the effects associated with marital
status were confounded by age, as single individuals were
younger and younger age was associated with greater weight
gain.
The impact of SES on change in weight and waist
circumference in our study is noteworthy. First, in our study,
it appeared that higher SES (higher education level, larger/
more expensive homes) was associated with greater gain in
weight and waist circumference (although the association
with education level was no longer statistically significant
after adjustment for age and other risk factors). This is in
contrast to a study conducted in the United Kingdom
reported that lower SES was associated with greater weight
gain.15 We believe that this may be related to greater time
spent at work, which is often sedentary, among individuals
from higher SES, and hence less time to participate in leisure
time physical activity. Singapore is a small country, which
is completely urbanized and occupational physical activity
largely reflects that of a developed economy (that is,
low levels of occupational physical activity). In this environment, leisure time physical activity becomes a more important determinant of adult weight gain. Even though high
SES was associated with greater weight gain in adulthood,
low SES was associated with higher prevalence of obesity at
baseline. The reason for this apparent contradiction may
International Journal of Obesity

relate to the timing of exposure to high or low SES. At least

one study has shown that SES at particular periods of life has
a greater influence than others. This is particularly true of
childhood SES.43 Taken together, our data suggests that the
risk of overweight and obesity experienced by the lower SES
group may have an earlier origin. As a consequence,
overweight and obesity related to low SES is established
before individual reaches adulthood. In our environment,
subsequent weight gain seems to be more a function of
higher SES.
The prospective nature of the study, with repeated
measurements of obesity, is the main strength of this study.
This is particularly relevant given that limited data of this
form are available in Asia. However, the study does have a
few limitations. First, we recognize that the response rate at
53% was low, which may limit the generalizability of our
findings. When we compared the respondents and nonrespondents, statistically significant differences were noted
for variables that were associated with differences in change
of weight or waist circumference. However, many of our
findings (including those related to age and baseline BMI)
are similar to those observed in other populations and this
gives us confidence that our findings are valid. Second,
obesity was measured at only two time points, which may
lead to some imprecision in the estimates of weight or waist
change. However, we feel that such imprecision would tend
to bias our findings toward the null hypothesis. This means
that, the risk of weight gain identified in association with the
various risk factors studied would have been underestimated
and our conclusions would still be valid. Finally, our analysis
did not include changes in risk factors such as diet and
physical activity, and measurement of overall physical
activity should include activities at work and commuting.
Diet and physical activity are important factors in influencing weight gain.17 Further studies and analyses of the degree
of variability exerted by changes in the key environmental
factors involved in weight gain should be explored as these
may give important clues as to the causes of weight gain. In
addition, future studies on younger population such as
children and adolescents could provide more information on
the effect of environmental factors on obesity development.
Information on other influencing factors such as parental
and childhood obesity,44,45 parity for women,46,47 menopause,48 and mental illnesses49 or physical disabilities50,51
would be relevant as they have been found to be significantly
associated with obesity.
Our study confirms that the secular trends of weight gain,
observed in other countries, are also occurring in Asia. The
majority of this weight gain occurs in early adulthood to
middle age. Overweight and obesity were most common in
Malays and Indians at baseline. This problem is further
compounded by the finding that the minor ethnic groups
experienced the greatest gains in weight and waist
circumference.
Our findings reinforced the point that obesity prevention
strategies should be started early in the younger age groups,

Longitudinal study on weight and waistline changes

SK Ong et al

1307
as it is during this period that more weight gain was
observed. These data may assist in the identification of
opportunities to implement such preventive strategies. For
example, in Singapore, men undergo compulsory military
training with intermittent contact with the armed forces
until the age of 40. Given our findings, this would be an
opportune avenue through which the importance of a
healthy diet and regular physical activity could be communicated. In addition, preventive strategies should also reach
out to the ethnic minority groups and need to be culturally
sensitive and acceptable to ensure that all segments of the
population are reached. Although marital status and education level were not found to be significant predictors of
weight change in multivariate analysis, the knowledge that
individuals who were single had higher levels of education
and larger/more expensive homes tended to gain the most
weight would be useful for designing messaging of preventive strategies.

Conflict of interest
The authors declare no conflict of interest.

Acknowledgements
This study was funded by the Biomedical Research Council
of Singapore grant number 03/1/27/18/216. This study was
approved by the Singapore General Hospital Institutional
Review Board.

References
1 World Health Organization Global InfoBase team. The Surf Report
2. Surveillance of Chronic Disease Risk Factors: Country-Level Data
and Comparable Estimates. World Health Organization: Geneva,
2005.
2 World Health Organization. Obesity and Overweight Fact Sheet
September 2006. Available at http://www.who.int/mediacentre/
factsheets/fs311/en/index.html , accessed on July 11 2008.
3 Wang Y, Lobstein T. Worldwide trends in childhood overweight
and obesity. Int J Pediatr Obes 2006; 1: 1125.
4 World Health Organization. Risk Factor Projects. Overweight and
Obesity, 2005. Available at: http://www.who.int/chp/chronic_
disease_report/part2_ch1/en/index16.html, accessed on July 11,
2007.
5 Epidemiology and Disease Control Division, Ministry of Health,
Singapore. National Health Survey 2004 Singapore. Available at
http://wwwmoh.gov.sg/mohcorp/publicationsreports.aspx?id
2984, accessed on July 11 2008.
6 Butland B, Jebb S, Kopelman P, McPherson K, Thomas S, Mardell J
et al. Foresight. Tackling obesities: future choices - project report. 2nd
edn. Government Office for Science: London, 2007.
7 Prospective Studies Collaboration. Body-Mass Index and CauseSpecific Mortality in 900 000 adults: Collaborative Analyses of
57 Prospective Studies. Lancet 2009; e-pub ahead of print 18
March 2009.
8 Gu D, He J, Duan X, Reynolds K, Wu X, Chen J et al. Body weight
and mortality among men and women in China. JAMA 2006;
295: 776783.

9 Jee SH, Sull JW, Park J, Lee SY, Ohrr H, Guallar E et al. Body-mass
index and mortality in Korean men and women. N Engl J Med
2006; 355: 779787.
10 James WPT, Jackson-Leach R, Mhurchu CN, Kalamara E, Shayeghi
M, Rigby NJ et al. Overweight and obesity (high body mass
index). In: Ezzati M, Lopez AD, Rodgers A, Murray CJL (eds).
Comparative Quantification of Health Risk. Global and Regional
Burden of Disease Attributable to Selected Major Risk Factors. Vol 1.
World Health Organization: Geneva, 2004. pp 497596.
11 Gregg EW, Cheng YJ, Cadwell BL, Imperatore G, Williams DE,
Flegal KM et al. Secular trends in cardiovascular disease risk
factors according to body mass index in US adults. JAMA 2005;
293: 18681874.
12 Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively
studied cohort of U.S. adults. N Engl J Med 2003; 348: 16251638.
13 Flegal KM, Graubard B, Williamson DF, Gail MH. Excess deaths
associated with underweight, overweight, and obesity. JAMA
2005; 293: 18611867.
14 Koh-Banerjee P, Chu NF, Spiegelman D, Rosner B, Colditz G,
Willett W et al. Prospective study of the association of changes in
dietary intake, physical activity, alcohol consumption, and
smoking with 9-y gain in waist circumference among 16 587 US
men. Am J Clin Nutr 2003; 78: 719727.
15 Purslow LR, Young EH, Wareham NJ, Forouhi N, Brunner EJ,
Luben RN et al. Socioeconomic position and risk of short-term
weight gain: prospective study of 14,619 middle-aged men and
women. BMC Public Health 2008; 8: 112.
16 Lewis CE, Smith DE, Wallace DD, Williams OD, Bild DE, Jacobs Jr
DR. Seven-year trends in body weight and associations with
lifestyle and behavioral characteristics in black and white
young adults: the CARDIA study. Am J Public Health 1997; 87:
635642.
17 Sherwood NE, Jeffery RW, French SA, Hannan PJ, Murray DM.
Predictors of weight gain in the pound of prevention study.
Int J Obes 2000; 24: 395403.
18 de Koning L, Merchant AT, Pogue J, Anand SS. Waist circumference and waist-to-hip ratio as predictors of cardiovascular events:
meta-regression analysis of prospective studies. Eur Heart J 2007;
28: 850856.
19 Koster A, Leitzmann MF, Schatzkin A, Mouw T, Adams KF, van
Eijk JT et al. Waist circumference and mortality. Am J Epidemiol
2008; 167: 14651475.
20 Zhang C, Rexrode KM, van Dam RM, Li TY, Hu FB. Abdominal
obesity and the risk of all-cause, cardiovascular, and cancer
mortality: sixteen years of follow-up in US women. Circulation
2008; 117: 16581667.
21 Ministry of Health Epidemiology and Disease Control Department. The National Health Survey 1998 Singapore. Available at
http://www.moh.gov.sg/mohcorp/publicationsreports.aspx?
pubfid 210, accessed on 23 December 2008.
22 Wu Y, Tai ES, Heng D, Tan CE, Low LP, Lee J. Risk factors associated
with hypertension awareness, treatment, and control in a multiethnic Asian population. J Hypertens 2009; 27: 190197.
23 Khaing NEE, Khoo CM, Tai ES, Lim SC, Tavintharan S, Wong TY
et al. Is there a clear threshold for fasting plasma glucose that
differentiates between those with and without neuropathy
and chronic kidney disease? The Singapore Prospective Study
Program. Am J Epidemiol 2009; 169: 14541462.
24 Housing Mobility 1995 and 2005. Singapore Department of
Statistics Mar 2006. Available at http://www.singstat.gov.sg/
pubn/papers/people/op-s10.pdf , accessed on 20 April 2009.
25 World Health Organization. Obesity: Preventing and Managing the
Global Epidemic. Report on a WHO Consultation on Obesity. World
Health Organization: Geneva, 2007. WHO/NUT/NCD/98.1.
26 American College of Sports Medicine. Position stand: the recommended quantity and quality of exercise for developing and
maintaining cardiorespiratory and muscular fitness, and flexibility
in healthy adults. Med Sci Sports Exerc 1998; 30: 975991.

International Journal of Obesity

Longitudinal study on weight and waistline changes

SK Ong et al

1308
27 World Health Organization. Guidelines for Controlling and
Monitoring the Tobacco Epidemic. World Health Organization:
Geneva, 1998.
28 Beydoun MA, Wang Y. Gender-ethnic disparity in BMI and waist
circumference distribution shifts in US adults. Obesity 2009; 17:
169176.
29 Williamson DF. Descriptive epidemiology of body weight and
weight change in U. S. adults. Ann Intern Med 1993; 119 (7 Pt 2):
646649.
30 Wardle J, Boniface D. Changes in the distributions of body mass
index and waist circumference in English adults, 1993/1994 to
2002/2003. Int J Obes (Lond) 2008; 32: 527532.
31 Rissanen A, Heliovarra M, Aromaa A. Overweight and anthropometric changes in adulthood: a prospective study of 17,000
finns. Int J Obes 1988; 12: 391401.
32 Chei CL, Iso H, Yamagishi K, Inoue M, Tsugane S. Body mass
index and weight change since 20 years of age and risk of
coronary heart disease among Japanese: the Japan Public Health
Center-Based Study. Int J Obes (Lond) 2008; 32: 144151.
33 Williams PT. Changes in body weight and waist circumference
affect incident hypercholesterolemia during 7 years of follow-up.
Obesity 2008; 16: 21632168.
34 Williams PT. Increases in weight and body size increase the odds
for hypertension during 7 years of follow-up. Obesity 2008; 16:
25412548.
35 Strandberg TE, Strandberg A, Salomaa VV, Pitkal K, Miettinen TA.
Impact of midlife weight change on mortality and quality of life
in old age. Prospective cohort study. Int J Obes Relat Metab Disord
2003; 27: 950954.
36 Kumanyika SK, Obarzanek E, Stettler N, Bell R, Field AE,
Fortmann SP et al. Population-based prevention of obesity:
the need for comprehensive promotion of healthful eating,
physical activity, and energy balance: a scientific statement
from American Heart Association Council on Epidemiology and
Prevention, Interdisciplinary Committee for Prevention. Circulation 2008; 118: 428464.
37 Chor D, Faerstein E, Kaplan GA, Lynch JW, Lopes CS. Association
of weight change with ethnicity and life course socioeconomic
position among Brazilian civil servants. Int J Epidemiol 2004; 33:
100106.
38 Lewis CE, Jacobs Jr DR, McCreath H, Kiefe CI, Schreiner PJ, Smith
DE et al. Weight gain continues in the 1990s: 10-year trends in
weight and overweight from the CARDIA study. Coronary artery
risk development in young adults. Am J Epidemiol 2000; 151:
11721181.
39 Baltrus PT, Lynch JW, Everson-Rose S, Raghunathan TE, Kaplan
GA. Race/ethnicity, life-course socioeconomic position, and body

International Journal of Obesity

40

41

42
43

44

45

46

47

48

49

50

51

weight trajectories over 34 years: the Alameda County Study.

Am J Public Health 2005; 95: 15951601.
Monteiro CA, Moura EC, Conde WL, Popkin BM. Socioeconomic
status and obesity in adult populations of developing countries: a
review. Bull World Health Organ 2004; 82: 940946.
Sobal J, Hanson KL, Frongillo EA. Gender, Ethnicity, Marital
Status, and Body Weight in the United States. Obesity 2009, e-pub
ahead of print 19 March 2009.
Averett SL, Sikora A, Argys LM. For better or worse: relationship
status and body mass index. Econ Hum Biol 2008; 6: 330349.
Giskes K, van Lenthe FJ, Turrell G, Kamphuis CB, Brug J,
Mackenbach JP. Socioeconomic position at different stages of the
life course and its influence on body weight and weight gain in
adulthood: a longitudinal study with 13-year follow-up. Obesity
2008; 16: 13771381.
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: 869873.
Guo SS, Huang C, Maynard LM, Demerath E, Towne B, Chumlea
WC et al. Body mass index during childhood, adolescence and
young adulthood in relation to adult overweight and adiposity:
the Fels Longitudinal Study. Int J Obes Relat Metab Disord 2000; 24:
16281635.
Gunderson EP, Murtaugh MA, Lewis CE, Quesenberry CP, West
DS, Sidney S. Excess gains in weight and waist circumference
associated with childbearing: the Coronary Artery Risk Development in Young Adults Study (CARDIA). Int J Obes Relat Metab
Disord 2004; 28: 525535.
Wolfe WS, Sobal J, Olson CM, Frongillo Jr EA, Williamson DF.
Parity-associated weight gain and its modification by sociodemographic and behavioral factors: a prospective analysis in US
women. Int J Obes Relat Metab Disord 1997; 21: 802810.
Sternfeld B, Wang H, Quesenberry Jr CP, Abrams B, Everson-Rose
SA, Greendale GA et al. Physical activity and changes in weight
and waist circumference in midlife women: findings from the
study of womens health across the nation. Am J Epidemiol 2004;
160: 912922.
Scott KM, Bruffaerts R, Simon GE, Alonso J, Angermeyer M, de
Girolamo G et al. Obesity and mental disorders in the general
population: results from the world mental health surveys. Int J
Obes (Lond) 2008; 32: 192200.
Lidstone JS, Ells LJ, Finn P, Whittaker VJ, Wilkinson JR, Summerbell CD. Independent associations between weight status and
disability in adults: results from the Health Survey for England.
Public Health 2006; 120: 412417.
Liou TH, Pi-Sunyer FX, Laferre`re B. Physical disability and obesity.
Nutr Rev 2005; 63: 321331.