Deakin Research Online
This is the published version:
Cameron, A. J., Magliano, D. J., Dunstan, D. W., Zimmet, P. Z., Hesketh,
K., Peeters, A. and Shaw, J. E. 2012, A bi-directional relationship between
obesity and health-related quality of life : evidence from the longitudinal
AusDiab study, International journal of obesity, vol. 36, no. 2, pp. 295303.
Available from Deakin Research Online:
http://hdl.handle.net/10536/DRO/DU:30046272
Reproduced with the kind permission of the copyright owner.
Copyright : 2012, Macmillan Publishers
International Journal of Obesity (2012) 36, 295–303
& 2012 Macmillan Publishers Limited All rights reserved 0307-0565/12
www.nature.com/ijo
ORIGINAL ARTICLE
A bi-directional relationship between obesity and
health-related quality of life: evidence from the
longitudinal AusDiab study
AJ Cameron1,2, DJ Magliano1,3, DW Dunstan1, PZ Zimmet1, K Hesketh2, A Peeters3 and JE Shaw1
1
Clinical Diabetes and Epidemiology Department, Baker IDI Heart and Diabetes Institute, Melbourne, Victoria, Australia;
Centre for Physical Activity and Nutrition Research, Deakin University, Burwood, Victoria, Australia and 3Department of
Epidemiology and Preventive Medicine, Monash University, Melbourne, Victoria, Australia
2
Objective: To assess the prospective relationship between obesity and health-related quality of life, including a novel assessment
of the impact of health-related quality of life on weight gain.
Design and setting: Longitudinal, national, population-based Australian Diabetes, Obesity and Lifestyle (AusDiab) study, with
surveys conducted in 1999/2000 and 2004/2005.
Participants: A total of 5985 men and women aged X25 years at study entry.
Main outcome measure(s): At both time points, height, weight and waist circumference were measured and self-report data
on health-related quality of life from the SF-36 questionnaire were obtained. Cross-sectional and bi-directional, prospective
associations between obesity categories and health-related quality of life were assessed.
Results: Higher body mass index (BMI) at baseline was associated with deterioration in health-related quality of life over 5 years
for seven of the eight health-related quality of life domains in women (all Pp0.01, with the exception of mental health,
P40.05), and six out of eight in men (all Po0.05, with the exception of role-emotional, P ¼ 0.055, and mental health, P40.05).
Each of the quality-of-life domains related to mental health as well as the mental component summary were inversely associated
with BMI change (all Po0.0001 for women and Pp0.01 for men), with the exception of vitality, which was significant in women
only (P ¼ 0.008). For the physical domains, change in BMI was inversely associated with baseline general health in women only
(P ¼ 0.023).
Conclusions: Obesity was associated with a deterioration in health-related quality of life (including both physical and mental
health domains) in this cohort of Australian adults followed over 5 years. Health-related quality of life was also a predictor of
weight gain over 5 years, indicating a bi-directional association between obesity and health-related quality of life. The
identification of those with poor health-related quality of life may be important in assessing the risk of future weight gain, and a
focus on health-related quality of life may be beneficial in weight management strategies.
International Journal of Obesity (2012) 36, 295–303; doi:10.1038/ijo.2011.103; published online 10 May 2011
Keywords: quality of life; longitudinal; waist circumference; body mass index
Introduction
An obesity epidemic is present in most developed as well as
developing nations. The consequences of this for individuals
include a significantly higher risk of obesity-related disease,
and a resulting economic burden related to both higher
healthcare costs and decreased productivity.1–4 The impact
Correspondence: Dr AJ Cameron, Centre for Physical Activity and Nutrition
Research, Deakin University, 221 Burwood Highway, Burwood 3125,
Australia.
E-mail: [email protected]
Received 17 December 2010; revised 27 February 2011; accepted 13 April
2011; published online 10 May 2011
of obesity on overall physical and mental well-being has also
been clearly established, with consistently lower scores on
scales of health-related quality of life in obese individuals.5–9
This has been evident even among obese individuals without
chronic disease.5
Health-related quality of life may also plausibly impact
weight status, with a reduction in the physical quality of life
domains likely to affect energy expenditure through physical
activity levels and time spent sedentary. Mental health
domains may impact energy intake, with links between
eating behaviour and emotional state, body dissatisfaction
and self-esteem having been established previously.10–12 In
related fields, strong evidence exists linking depression and
weight gain (and vice versa),13 and some research has
Obesity and health-related quality of life
AJ Cameron et al
296
suggested a link between health-related quality of life and
development of both stroke and diabetes.14,15 Despite this,
we are unaware of any evidence linking health-related
quality of life and future weight gain. The possibility that
health-related quality of life might be associated with weight
gain is of obvious importance given both the high and rising
numbers of overweight and obese individuals across the
world, and the relevance of health-related quality of life to
all individuals.
The national, longitudinal Australian Diabetes, Obesity and
Lifestyle Study (AusDiab) included assessments of both healthrelated quality of life and obesity (using measured height,
weight and waist circumference) at two time points. Using data
from this large study of the adult Australian population,
we examine here the potential bi-directional relationships
between health-related quality of life and obesity.
Methods
Survey procedures
The study methods and response for the AusDiab study have
been described in detail elsewhere.1,16,17 Briefly, the study
was a population-based, national survey of 11 247 adults
(5049 men and 6198 women), aged X25 years, conducted
from 1999 to 2000 (involving 55.3% of those completing a
household interview). The sample was drawn from 42
randomly selected clusters across Australia based on Census
Collector Districts and stratified by state. Five years later
between 2004 and 2005, 6537 participants (58%) were reexamined. A total of 2789 men and 3350 non-pregnant
women had data on anthropometric measures and healthrelated quality of life at both time points. Of these, 154
individuals reported a cancer other than skin cancer in
annual follow-up surveys between the baseline and followup surveys, and were excluded, leaving a total of 5985
AusDiab participants available for these analyses. A comparison of the profile of responders and non-responders to the
AusDiab follow-up survey has been published previously,
with differences in several physiological and socio-demographic characteristics, but not body mass index (BMI),
observed between responders and non-responders.17 After
adjustment for age and sex, the responders also had higher
quality-of-life scores (both mental and physical component
summary, Po0.0001).
Height was measured to the nearest 0.5 cm without shoes,
using a stadiometer. Weight was measured to the nearest
0.1 kg using a mechanical beam balance. Waist circumference was measured halfway between the lower border of the
ribs and the iliac crest, on a horizontal plane.18 Waist
circumference and BMI cut-off points for overweight (male
94 cm and 25 kg m2; female 80 cm and 25 kg m2) and
obesity (male 102 cm and 30 kg m2; female 88 cm and
30 kg m2) were based on the standard definition published
by the World Health Organization.19 A 0.7% portion of the
cohort had a baseline BMI o18.5 kg m2 and could be
International Journal of Obesity
classified as underweight. Health-related quality of life was
assessed using the self-administered SF-36 (version 1) questionnaire, used with permission from the Medical Outcomes
Trust (Boston, MA, USA), and reported using Australian
norm-based scores according to previously published guidelines.20,21 The use of norm-based weights gives each domain
score a mean of 50 and an s.d. of 10, allowing change in
scores to be assessed on a comparable scale. Using Australian
data to calculate norm-based scores (from the 1995
Australian National Health Survey (n ¼ 18 468))21 helps to
account for cultural differences in the way in which health
and health-related quality of life are viewed between
populations.22
Based on the 36 included items, scores were calculated for
each of eight domains (physical functioning, role-physical,
bodily pain, general health, vitality, social functioning, roleemotional and mental health).20 ‘Role-physical’ and ‘roleemotional’ refer to the presence of problems with work or
other daily activities as a result of physical health and mental
health, respectively. Summary measures of the physical and
mental components of the survey were also calculated based
on a factor analysis of the eight domains among participants
in the 1995 National Health Survey, as previously reported.21
For participants living in family units, individual income
was calculated using a modified version of the Organisation
for Economic Cooperation and Development (OECD)
equivalence scale to adjust for the number of adults and
children in the household.23 The highest level of education
attained was measured by a self-report questionnaire (categories included ‘never attended school’, ‘primary school’,
‘some high school’, ‘completed high school’ and ‘University/
TAFE’). Alcohol consumption (g day1) was estimated from a
self-administered, validated food frequency questionnaire
developed by the Anti-Cancer Council of Victoria,24 as
reported previously.25 A 75-g oral glucose-tolerance test was
completed at baseline and follow-up for assessment of
incident diabetes, with diabetes classified as previously
described by Magliano et al.26 Incident cardiovascular disease
(heart attack, stroke, coronary artery bypass graft or
angioplasty) was assessed by physician adjudication of
medical records as previously described).27 Self-reported
television viewing and physical activity time (leisure time
physical activity measured using the Active Australia survey)28 over the last week were assessed as previously
described using validated tools.29 The AusDiab survey
protocols were approved by the ethics committee of the
International Diabetes Institute and Monash University’s
Standing Committee on Ethics in Research involving
Humans (SCERH). Informed consent was obtained from all
participants at both baseline and follow-up surveys.
Statistical methods
Scores for each of the health-related quality of life domains
were divided into tertiles (for sex and 10-year age groups),
with the exception of role-emotional, role-physical and
Obesity and health-related quality of life
AJ Cameron et al
297
social functioning, which, owing to ceiling effects and a
limited number of items used to create the scores, were
dichotomized into all those who scored the highest possible
score, and the remainder. The decision to divide continuous
health-related quality-of-life domains into categories was
based on the limited variability of these items, which would
have made an analysis of them in a continuous manner
problematic, and the fact that analysis of categorical healthrelated quality-of-life variables would be analogous to that of
BMI, which was divided into normal, overweight and obese
categories based on established cut-off points. Cross-sectional associations between BMI categories of normal, overweight and obesity and continuous variables were tested by
one-way analysis of variance. For dichotomous variables,
w2-statistics were used. Mean changes in BMI, waist circumference and in each of the health-related quality-of-life
domains were calculated. The effect of BMI on change
in health-related quality-of-life measures was estimated in
sex-specific models containing linear terms for both variables. Similarly, the effect of health-related quality of life on
change in BMI was also estimated in sex-specific models
containing linear terms for both variables. In all models
assessing change in either BMI or health-related quality of
life, the respective baseline variable was included in order to
model the change, as suggested by Vickers and Altman.30
The estimated marginal mean change in BMI or healthrelated quality-of-life domains, stratified by health-related
quality-of-life categories (tertiles or dichotomous categories)
and the normal, overweight and obesity categories of BMI,
respectively, was reported. All analyses were repeated using
waist circumference instead of BMI (reported in an onlineonly appendix). The sample was divided into those whose
BMI increased and those whose BMI was stable or decreased
(change in BMIp0 kg m2) between baseline and follow-up.
For these two categories, change in the physical component
summary for different categories of BMI at baseline was
calculated.
Variables that may affect both weight and quality of life
were considered for inclusion in the models tested. Significant predictors of change in either BMI or health-related
quality of life that were included in all regression models
included baseline age, sex, smoking status, education level
and income. Chronic diseases that may affect both quality of
life and weight include diabetes and cardiovascular disease.
We repeated the analyses after the exclusion of those with
these incident conditions; however, the reported findings
were not materially different. An indicator of incident cases
of both conditions was, however, a significant predictor of
change in both BMI and health-related quality of life, and
was also included in all models. Alcohol consumption was
not a significant predictor of change in either health-related
quality of life (any of the eight domains, or the component
summaries) or BMI, either entered as categorical or linear
terms, in sex-specific multivariable models and was therefore
not included in any analyses presented here. The effect of
television viewing time and physical activity time on the
relationship between baseline quality of life and change in
BMI was tested with the addition of these variables to the
multivariate model.
Results
Cross-sectional associations between obesity and health-related
quality of life
The descriptive statistics of the cohort and cross-sectional
associations of demographic and health-related quality-oflife characteristics with BMI categories are presented in
Table 1. Strong cross-sectional associations (Po0.0001) were
evident between BMI categories and the physical functioning, bodily pain, general health and vitality domains, as well
as the physical component summary, in both men and
women. Associations were also seen in women only between
BMI categories and the role-physical, social functioning and
role-emotional domains (Po0.0001). Mental health and the
mental component summary showed no cross-sectional
relationship with BMI categories.
Baseline BMI and change in health-related quality of life
Higher BMI at baseline (modelled as a linear effect) was
associated with a greater reduction in health-related quality
of life over 5 years for seven of the eight health-related
quality-of-life domains in women (all Pp0.01), and for six
out of eight domains in men (all Po0.05). The exceptions
were change in the role-emotional domain in men, for
which a borderline association with baseline BMI was
present (P ¼ 0.055), and change in the mental health
domain, which was not associated with baseline BMI in
either men or women. Change in the physical component
summary (Po0.0001 in both sexes), but not the mental
component summary, was also associated with baseline BMI.
Figure 1 shows that a gradient in change in health-related
quality of life was seen from normal to overweight and
obesity for most domains, with the difference between the
normal and overweight categories being significant for the
bodily pain domain (men P ¼ 0.002, women P ¼ 0.018) and
the physical component summary (men, P ¼ 0.003; women,
P ¼ 0.038) in both men and women.
Baseline health-related quality of life and change in BMI
Over 5 years, BMI increased on average by 0.88 kg m2 in
men and by 0.70 kg m2 in women. Each of the quality-oflife domains related to mental health (modelled as linear
effects) as well as the mental component summary were
inversely associated with change in BMI over 5 years (all
Po0.0001 for women and Pp0.01 for men), with the
exception of vitality, which was significant in women only
(P ¼ 0.008). For the physical domains, change in BMI was
inversely associated with baseline general health in women
only (P ¼ 0.023). After additional adjustment for variables
International Journal of Obesity
International Journal of Obesity
55.2)
55.7)
59.3)
57.6)
52.8)
56.7)
55.3)
57.1)
55.8)
55.7)
Abbreviation: BMI, body mass index. wPo0.0001; zPo0.05.
(46.8,
(48.6,
(44.1,
(45.3,
(37.6,
(45.6,
(45.0,
(45.4,
(45.4,
(43.9,
49.7
50.3
48.7
50.3
43.4
51.2
50.2
49.9
49.7
48.7
55.2)
55.7)
59.3)
57.6)
52.8)
56.7)
55.3)
57.1)
55.9)
55.9)
50.5
50.6
49.3
50.0
44.3
51.5
50.6
50.5
50.0
49.2
Quality-of-life domain
Physical functioning
Role-physical
Bodily pain
General health
Vitality
Social functioning
Role-emotional
Mental health
Physical component summary
Mental component summary
(46.8,
(48.6,
(44.1,
(45.3,
(37.6,
(45.6,
(55.3,
(45.4,
(46.2,
(44.8,
3275
(42, 60)
(22.7, 29.4)
(60.1, 77.7)
(75.0, 92.5)
36.5
467 (238, 700)
8.3 (0.4, 12.3)
4.1
10.2
51
26.6
70.3
84.8
5985
(42, 60)
(23.5, 29.4)
(64.9, 86.0)
(80.5, 99.5)
40.9
500 (280, 714)
13.5 (0.8, 19.4)
5.4
11.8
N
Age (years)
BMI (kg m2)
Weight (kg)
Waist circumference (cm)
University/TAFE education (%)
Weekly income ($AUD)
Alcohol consumption (g day1)
Incident diabetes or CVD (%)
Current smoker
51.3
26.9
76.6
90.5
Women
All
51.4
50.9
49.9
49.6
45.3
51.8
51.2
51.3
50.4
49.8
(48.9,
(48.6,
(44.1,
(44.3,
(40.2,
(51.1,
(55.3,
(45.4,
(47.1,
(46.2,
57.3)
55.7)
59.3)
55.1)
52.8)
56.7)
55.3)
57.1)
56.1)
56.0)
2710
(43, 60)
(24.5, 29.3)
(74.8, 91.7)
(89.8, 103.9)
46.1
538 (333, 766)
19.7 (3.0, 29.1)
6.9
13.6
51.5
27.2
84.1
97.2
Men
52.2
51.5
50.5
52.1
44.6
52.2
51.0
50.2
52.1
48.9
(51.0,
(55.7,
(44.1,
(47.7,
(37.6,
(51.1,
(55.3,
(45.4,
(48.7,
(44.6,
57.3)
55.7)
59.3)
57.6)
52.8)
56.7)
55.3)
57.1)
57.0)
55.3)
1469
(40, 57)
(21.0, 23.7)
(54.5, 64.0)
(70.0, 79.0)
42.5
497 (280, 714)
8.8 (0.6, 13.0)
2.2
10.6
48.6
22.2
59.4
74.8
Normal
(o25 kg m2)
49.3
50.6
48.3
50.0
43.6
51.1
50.0
49.8
49.4
48.7
(46.8,
(48.6,
(43.7,
(45.3,
(37.6,
(45.6,
(45.0,
(43.0,
(45.0,
(44.2,
55.2)
55.7)
52.9)
57.6)
50.3)
56.7)
55.3)
57.1)
55.2)
55.9)
1088
(44, 62)
(26, 28.4)
(67.0, 76.3)
(82.3, 91.1)
32.9
456 (208, 638)
8.9 (0.5. 13.0)
4.3
10.5
53.2
27.2
71.8
86.9
Overweight
(25–30 kg m2)
Women
45.2
47.5
45.8
47.2
40.6
49.3
48.8
49.2
45.3
48.3
(40.0,
(40.3,
(39.7,
(40.3,
(32.6,
(45.6,
(45.0,
(43.0,
(38.6,
(42.3,
53.1)w
55.7)w
52.9)w
55.1)w
50.3)w
56.7)w
55.3)w
57.1)
52.6)w
56.3)
52.7
50.9
50.6
50.8
46.0
51.8
50.9
51.3
51.5
49.5
(51.0,
(48.6,
(44.1,
(45.3,
(40.2,
(51.1,
(55.3,
(47.7,
(48.7,
(45.8,
57.3)
55.7)
59.3)
57.6)
52.8)
56.7)
55.3)
57.1)
56.8)
55.8)
803
(40, 59)
(22.0, 24.2)
(66.8, 76.9)
(82.8, 91.5)
50.9
539 (333, 766)
18.6 (3.0, 27.7)
3.4
16.3
50
23.0
71.5
86.9
718
(44, 61)w
(31.4, 36.7)w
(80.8, 97.6)w
(94.8, 108.5)w
29.8w
422 (200, 575)w
6.5 (0.2, 7.2)w
7.9
9.2
52.6
34.7
90.4
102.2
Normal
(o25 kg m2)
BMI category
Obese
(430 kg m2)
Cross-sectional associations of demographic and health-related quality-of-life characteristics with BMI categories at baseline
Data are arithmetic mean
(25th, 75th percentile) or percent
Table 1
51.6
51.1
50.0
50.0
45.6
52.0
51.4
51.4
50.7
50.0
(48.9,
(48.6,
(44.1,
(45.3,
(40.2,
(51.1,
(55.3,
(45.4,
(47.6,
(46.3,
57.3)
55.7)
59.3)
55.1)
52.8)
56.7)
55.3)
57.1)
56)
56)
1347
(43, 61)
(26.1, 28.4)
(78.7, 89.2)
(93.3, 102.1)
46.7
533 (333, 766)
20.3 (3.2, 30.2)
7.3
12.6
52.4
27.2
84.2
97.5
Overweight
(25–30 kg m2)
Men
49.1
50.3
48.6
47.1
43.9
51.4
51.0
51.0
48.3
49.8
(45.3,
(48.6,
(43.7,
(40.3,
(37.6,
(51.1,
(55.3,
(45.4,
(43.8,
(45.9,
55.2)w
55.7)
59.3)w
55.1)w
52.8)w
56.7)
55.3)
57.1)
54.7)w
56.3)
560
(44, 60)z
(30.8, 34.2)w
(93.3, 107.8)w
(105.5, 116.8)w
38.0w
549 (333, 766)
19.8 (2.7, 29.6)
11.2
12.4z
51.6
33.0
101.9
111.5
Obese
(430 kg m2)
Obesity and health-related quality of life
AJ Cameron et al
298
reflecting television viewing and physical activity time (to
test for any mediating effect), this relationship was not
attenuated. In men, a borderline inverse relationship was
apparent between both baseline bodily pain and general
health, and change in BMI (P ¼ 0.056 and P ¼ 0.080,
respectively).
Figure 2 shows that the relationship between healthrelated quality of life at baseline and change in BMI was not
as clearly ‘dose-dependent’ as that between obesity at
baseline and decreasing health-related quality of life. Only
those in the lowest health-related quality-of-life category at
baseline had significantly greater increases in BMI over the
following 5 years.
The sex-specific results for the associations between
categories of health-related quality of life and categories of
both BMI and waist circumference are presented graphically
in an online-only appendix (Supplementary eFigures 1–4).
Broadly similar results were seen for both waist circumference and BMI in men and women.
Effect of weight gain or loss on the relationship between baseline
BMI category and change in health-related quality of life
The estimated marginal mean change in both the physical
and mental component summary over 5 years was assessed
separately for those whose BMI was stable or decreased, and
those whose BMI increased. A greater decrease in the
physical component summary was seen among those whose
weight increased over the follow-up period compared with
those whose weight was stable or decreased (1.32 (95%
confidence interval: 1.52 to 1.10) versus 0.48 (0.83 to
0.14), Po0.0001). The change in the physical component
summary for those who were normal, overweight and obese
at baseline and stratified by weight change status (increased
or decreased/stable) is presented in Table 2. An association
between BMI category and estimated marginal mean change
in the physical component of health-related quality of life
was only seen for those whose BMI increased between
baseline and follow-up, suggesting that the relationship
between baseline BMI and decreased health-related quality
of life over the course of the follow-up was largely seen
among those whose BMI increased in that time period
(69.8% of the cohort). The estimated change in the mental
component summary was not different among those whose
weight increased or decreased over the 5 years of follow-up.
Effect of age on the relationship between BMI and health-related
quality of life
After dividing the cohort into those aged less than 65 years
and those aged 65 years or over, the bi-directional associations between body weight and health-related quality of life
were assessed.
Comparing obese individuals to those with normal weight,
the greatest differences in change in each of the physical
functioning (PF), role-physical (RP), vitality (VT),
Obesity and health-related quality of life
AJ Cameron et al
299
Figure 1 Change in health-related quality of life over 5 years, by baseline BMI categories. Models included baseline health-related quality-of-life domain (in order
to model change), age, sex, income, education, smoking status and incident diabetes or cardiovascular disease.
role-emotional (RE) and mental health (MH) domains over
5 years were seen in those aged 65 years or more at baseline
(PF: o65 ¼ 1.44 (Po0.0001), 65 þ ¼ 1.79 (P ¼ 0.019);
RP: o65 ¼ 1.27 (Po0.0001), 65 þ ¼ 2.15 (P ¼ 0.030);
VT: o65 ¼ 0.84 (P ¼ 0.009), 65 þ ¼ 1.81 (P ¼ 0.013); RE:
o65 ¼ 1.02 (P ¼ 0.003), 65 þ ¼ 2.12 (P ¼ 0.025); MH:
o65 ¼ 0.21 (P40.05), 65 þ ¼ 1.28 (P ¼ 0.044)). For social
functioning (SF), the greatest difference was seen in those aged
o65 years (o65 ¼ 1.50 (Po0.0001), 65 þ ¼ 0.39 (P40.05)).
Obesity status had a similar effect on change in bodily pain
(BP) and general health (GH) scores in both age groups (BP:
o65 ¼ 1.95 (Po0.0001), 65 þ ¼ 1.96 (P ¼ 0.016); GH:
o65 ¼ 1.27 (Po0.0001), 65 þ ¼ 1.15 (P ¼ 0.106)).
Examining changes in BMI by tertiles (or halves) of
health-related quality-of-life domains, age-related differences were observed for the bodily pain, social functioning,
role-emotional and mental health domains.
Comparing the top and bottom tertiles of bodily pain, the
greatest differences in BMI change were observed in those
aged 65 years or over (o65 ¼ 0.037 (P ¼ 0.638),
65 þ ¼ 0.313 (P ¼ 0.045)). For each of social functioning,
role-emotional and mental health, the greatest differences
were observed among those aged less than 65 years
(SF:
o65 ¼ 0.17
(P ¼ 0.004),
65 þ ¼ 0.11
(P ¼ 0.35);
RE: o65 ¼ 0.29 (Po0.0001), 65 þ ¼ 0.23 (P ¼ 0.067); MH:
o65 ¼ 0.19 (P ¼ 0.009), 65 þ ¼ 0.13 (P ¼ 0.37)).
Discussion
In this longitudinal analysis of a large Australian cohort of
adults, we present evidence of the association between
obesity and reduced health-related quality of life. Importantly, we found reduced health-related quality of life to be
not only a consequence of obesity as has been observed
previously, but also a predictor of weight gain. This
bi-directional association, with each predicting deterioration
in the other, highlights the complexity of the relationship
between weight and health-related quality of life. Further
complexity is evident from our observation of a much
stronger relationship between baseline BMI and deterioration in health-related quality of life among those who gained
weight over the follow-up period (over two-thirds of the
cohort). The baseline BMI of those whose weight was stable
or decreased over 5 years was not associated with deterioration in health-related quality of life.
Several previous cross-sectional and prospective studies
have shown obesity and/or weight change to be related to
deterioration in health-related quality of life in both men
and women, with many of these studies using the SF-36
questionnaire, as used here.9,31–37 Most studies have concluded that obesity is related more strongly to physical
rather than mental health domains,37 including a recent
large, cross-sectional Australian study that found little
International Journal of Obesity
Obesity and health-related quality of life
AJ Cameron et al
300
Figure 2 Change in BMI over 5 years, by tertiles (or dichotomous categories) of health-related quality-of-life domains at baseline. Data adjusted for baseline BMI
(in order to model change), age, sex, income, education, smoking status and incident diabetes or cardiovascular disease. BMI, body mass index.
Table 2 Change in the physical component summary by baseline BMI group
for those whose weight was stable or decreased, and those whose weight
increased between baseline and follow-upa
Baseline BMI
group
Change in the physical component summary
BMI stable or decreased between
baseline and follow-up
BMI increased between
baseline and follow-up
Men
Normal
Overweight
Obese
1.35 (2.4, 0.3)
0.98 (1.7, 0.2)
0.17 (1, 1.3)
0.06 (0.5, 0.6)
1.02 (1.5, 0.6)
2.96 (3.7, 2.2)
Women
Normal
Overweight
Obese
0.4 (0.4, 1.2)
0.16 (0.9, 0.6)
1.8 (2.7, 0.8)
0.72 (1.2, 0.3)
1.87 (2.4, 1.3)
2.68 (3.4, 2.0)
Abbreviation: BMI, body mass index. aModels included baseline health-related
quality-of-life domain (in order to model change), age, sex, income,
education, smoking status and incident diabetes or cardiovascular disease.
difference in the mental component summary between
obesity categories after adjustment for demographic and
socio-economic factors.7 We also found a strong and graded
relationship between obesity and the physical components
of health-related quality of life. However, unlike many
International Journal of Obesity
previous studies, a clear relationship was also seen between
obesity and some components of mental health.
The influence of health-related quality of life on subsequent weight gain has to our knowledge not been reported
previously. For six of the eight health-related quality-of-life
domains, those in the lowest category at baseline had a
significantly greater increase in weight over the ensuing 5
years. In contrast to the association between obesity and
deterioration in health-related quality of life (most evident
in the physical domains), increase in weight was greater for
those in the lowest category of the mental health domains
(role-emotional, mental health, social functioning and the
mental health summary score).
Plausible mechanisms exist linking health-related quality
of life and weight gain. Reduced physical (or indeed social)
functioning may be linked to increased BMI through a
reduced capacity or willingness to engage in physical activity
and increases in sedentary time. Our observation of no
change in the association between either baseline general
health or vitality with change in BMI (in women) upon
inclusion of variables for television viewing and physical
activity time is not supportive of this hypothesis; however,
the measures of sedentary behaviour and physical activity
used here do not capture workplace sedentary behaviours or
transportation. Components of quality of life aligned to
Obesity and health-related quality of life
AJ Cameron et al
301
mental health may be linked to weight gain through
psychological pathways related to body dissatisfaction and
self-esteem,11,12 and/or through eating behaviour related to
emotional state.10 Physiological links between weight gain
and both the hypothalamic–pituitary–adrenal axis and the
sympathetic nervous system (‘stress centres’) have been
suggested as being responsible for the link between depression and weight gain,11,13,38 with a similar process being
potentially responsible for the strong relationship between
the mental health components of health-related quality of
life and weight gain seen here.
The complex interactions of behavioural and physiological, as well as genetic, factors that are responsible for weight
gain may differ for particular health-related quality-of-life
domains. A particularly interesting observation from our
analysis was that baseline BMI was associated with decreases
in the physical components of health-related quality of life,
whereas it was the mental health-related components at
baseline that were associated with weight gain over the
follow-up period. This suggests that the bi-directional
association observed was not simply a circle of the same
two factors, further strengthening the validity of the finding
relating quality of life at baseline to weight gain. We also
observed that the relationship between health-related quality
of life and body weight was at least partially agedependent. An effect of age on the association between
obesity and various other health outcomes (the metabolic
syndrome and its components) has also been observed
previously.39 Using an instrument designed to assess
weight-specific effects on quality of life (Impact of Weight
on Quality of Life-Lite (IWQOL-Lite)) among an overweight
and obese population, the association between obesity and
quality of life was also shown to vary with age.40 In a study of
those over 65 years, both underweight and obesity were
found to be associated with impaired health-related quality
of life.41 Hopman et al.42 showed in a population-based adult
Canadian sample that, whereas mean changes in healthrelated quality of life over a 5-year period were small,
changes were larger in older age groups, and for physically
oriented domains. In our study, the effect of baseline BMI on
change in health-related quality-of-life domains was generally stronger in those aged over 65 years, whereas the effect
of baseline health-related quality of life on change in BMI
was generally stronger in those aged less than 65 years. Two
notable exceptions were the observations that obese individuals had greater reductions in social functioning than those
with normal weight only in those aged less than 65 years,
and that bodily pain at baseline was only associated with
increased BMI in those aged over 65 years.
The clinical relevance of the findings presented here bears
some scrutiny. Due to the large sample size of the study,
small changes in either health-related quality of life or BMI
may be statistically significant. The magnitude of the
observed changes in each of the health-related quality-oflife domains and BMI are not extremely large (approximately
0.3 BMI units and 2 points for health-related quality of life).
The changes reported are over 5 years and may be clinically
relevant when the number of years in a lifetime for which an
individual is overweight or obese, and over which individuals gain weight, are considered. Furthermore, as the
health-related quality-of-life scores were based on Australian
norms, it is difficult to compare these results with the
published minimal clinically important difference levels.
Study limitations and considerations
As the AusDiab study was large, national and populationbased, the results presented here are likely to broadly reflect
the association between obesity and health-related quality of
life in the adult Australian population. Non-response and
loss to follow-up, however, mean that we cannot assume that
the results are completely representative of all Australians.
Measured height, weight and waist circumference at both
time points in this study means that bias related to obesity
measurement is likely to be minimal. We used version 1 of
the SF-36 questionnaire here, which differs somewhat from
version 2 that was designed to be more easily understood,
result in less missing data and improve the sensitivity of the
role-function scales, as well as improving cross-cultural
validity.22 Some studies show meaningful differences
between the versions whereas others do not (reviewed in
reference Hawthorne et al.22).
The similar results seen for both BMI and waist circumference suggests that the relationship between health-related
quality of life and both abdominal and general obesity is
similar. Although we included gold-standard measures of
incident diabetes and cardiovascular disease, a limitation of
this study is that we did not assess the presence of other
relevant chronic diseases that may have affected both quality
of life and weight change, (such as HIV, thyroid conditions,
osteoarthritis and depression). Differences between men and
women in responses to questions addressing mental health
status may also have impacted the findings, although similar
trends were observed for both sexes.
Although the SF-36 questionnaire is designed for use in
any population, other health-related quality of life tools
exist. Use of the IWQOL-Lite questionnaire, which was
developed specifically to measure the impact of weight on
health-related quality of life among persons being treated for
obesity,43 has shown that other elements of the broader
concept of quality of life are also related to obesity.
Specifically, mean scores for self-esteem, sexual life, public
distress and work domains have been shown to vary
according to BMI level.44 Whether such measures are also
predictive of changes in BMI, and among non-obese
individuals, will be an interesting topic for future studies.
Conclusion
Together with the widely recognized health and mortality
consequences of obesity,1 the negative consequences of
International Journal of Obesity
Obesity and health-related quality of life
AJ Cameron et al
302
obesity are further highlighted in this report of an association
between BMI and deterioration in both physical and mental
health components of health-related quality of life. With
approximately 60% of adult Australians either overweight or
obese,45 and similar figures in many other developed
countries,46 it is clear that the obesity-related burden of
diminished health-related quality of life is substantial.
Our observation of a link between health-related quality of
life and weight gain adds an important new insight to this
complex relationship. Poor health-related quality of life
(particularly related to mental health) can be considered a
risk factor for future weight gain. The identification of those
with poor health-related quality of life may be important in
the assessment of the risk of future weight gain, and a focus
on health-related quality of life may be beneficial in weight
management strategies.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
We are grateful to the many people involved in organizing
and conducting the AusDiab study, and especially the
participants for volunteering their valuable time. AJC had
full access to all of the data of the study and takes
responsibility for the integrity of the data and the accuracy
of the data analysis. AJC is supported by a capacity building
grant from the Australian NHMRC (Grant 425845). JES is
supported by an NHMRC fellowship (586623). KH is
supported by a National Heart Foundation of Australia
Career Development Award. AP and DWD are supported by
Victorian Health Promotion Foundation Public Health
Research Fellowships. AusDiab was supported by a grant
from the National Health and Medical Research Council of
Australia (grant number 233200). The AusDiab study,
co-coordinated by the Baker IDI Heart and Diabetes Institute,
acknowledges the support by the following: Australian
Government Department of Health and Ageing, Abbott
Australasia, Alphapharm, AstraZeneca, Aventis Pharma,
Bristol-Myers Squibb, City Health Centre Diabetes ServiceF
Canberra, Diabetes Australia, Healthy Living NT, Eli Lilly
Australia, Estate of the late Edward Wilson, GlaxoSmithKline, Jack Brockhoff Foundation, Janssen-Cilag, Kidney
Health Australia, Marian & EH Flack Trust, Menzies Research
Institute, Merck Sharp & Dohme, New South Wales Health,
Northern Territory Department of Health and Families,
Novartis, Novo Nordisk, Pfizer, Pratt Foundation, Queensland Health, Roche Diagnostics Australia, Royal Prince Alfred
Hospital (Sydney), Sanofi-Synthelabo, South Australia
Health, Tasmanian Department of Health and Human
Services, Victorian Department of Human Services, and
Western Australia Health.
International Journal of Obesity
References
1 Cameron AJ, Dunstan DW, Owen N, Zimmet PZ, Barr ELM,
Tonkin AM et al. Health and mortality consequences of
abdominal obesity: evidence from the AusDiab study. Med J Aust
2009; 191: 202–208.
2 World Health Organization. Obesity: preventing and managing the
global epidemic. Report of a WHO consultation. World Health
Organization Technical Report Series 894 2000; i–xii, 1–253.
3 Access Economics Pty Ltd. The Economic Costs of Obesity.
Commissioned by Diabetes Australia, 2006.
4 Colagiuri S, Lee CM, Colagiuri R, Magliano D, Shaw JE, Zimmet
PZ. et al. The cost of overweight and obesity in Australia. Med J
Aust 2010; 192: 260–264.
5 Jia H, Lubetkin EI. The impact of obesity on health-related
quality-of-life in the general adult US population. J Public Health
(Oxf) 2005; 27: 156–164.
6 Soltoft F, Hammer M, Kragh N. The association of body mass
index and health-related quality of life in the general population:
data from the 2003 Health Survey of England. Qual Life Res 2009;
18: 1293–1299.
7 Renzaho A, Wooden M, Houng B. Associations between body
mass index and health-related quality of life among Australian
adults. Qual Life Res 2010; 19: 515–520.
8 Serrano-Aguilar P, Munoz-Navarro SR, Ramallo-Farina Y, TrujilloMartin MM. Obesity and health related quality of life in the
general adult population of the Canary Islands. Qual Life Res
2009; 18: 171–177.
9 Daviglus ML, Liu K, Yan LL, Pirzada A, Garside DB, Schiffer L et al.
Body mass index in middle age and health-related quality of life
in older age: the Chicago heart association detection project in
industry study. Arch Intern Med 2003; 163: 2448–2455.
10 Allison DB, Heshka S. Emotion and eating in obesity? A critical
analysis. Int J Eat Disord 1993; 13: 289–295.
11 Bjorntorp P, Rossner S, Udden J. ‘Consolatory eating’ is not a
myth. Stress-induced increased cortisol levels result in leptinresistant obesity. Lakartidningen 2001; 98: 5458–5461.
12 Atlantis E, Ball K. Association between weight perception and
psychological distress. Int J Obes (Lond) 2008; 32: 715–721.
13 Luppino FS, de Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx
BW et al. Overweight, obesity, and depression: a systematic review
and meta-analysis of longitudinal studies. Arch Gen Psychiatry
2010; 67: 220–229.
14 Myint PK, Surtees PG, Wainwright NWJ, Luben RN, Welch AA,
Bingham SA et al. Physical health-related quality of life predicts
stroke in the EPIC-Norfolk. Neurology 2007; 69: 2243–2248.
15 Tapp RJ, O0 Neil A, Shaw JE, Zimmet PZ, Oldenburg BF. Is there a
link between components of health-related functioning and
incident impaired glucose metabolism and type 2 diabetes? The
Australian Diabetes Obesity and Lifestyle (AusDiab) study.
Diabetes Care 2010; 33: 757–762.
16 Dunstan DW, Zimmet PZ, Welborn TA, Cameron AJ, Shaw J,
de Courten M et al. The Australian Diabetes, Obesity and Lifestyle
Study (AusDiab)Fmethods and response rates. Diab Res Clin
Pract 2002; 57: 119–129.
17 Barr L, Magliano D, Zimmet P, Polkinghorne K, Atkins R, Dunstan
D et al. AusDiab 2005. The Australian Diabetes, Obesity and Lifestyle
Study. Tracking the Accelerating Epidemic: its Causes and Outcomes.
International Diabetes Institute: Melbourne, 2006.
18 Dalton M, Cameron AJ, Zimmet PZ, Shaw JE, Jolley D, Dunstan
DW et al. Waist circumference, waist–hip ratio and body mass
index and their correlation with cardiovascular disease risk
factors in Australian adults. J Intern Med 2003; 254: 1–9.
19 World Health Organization. ObesityFPreventing and Managing the
Global Epidemic: Report of a WHO Consultation on Obesity. World
Health Organization: Geneva, 1998.
20 Ware JE, Kosinski M. SF-36 Physical & Mental Health Summary
Scales: a Manual For Users of Version 1, 2nd edn. QualityMetric
Inc.: Lincoln, RI, 2001.
Obesity and health-related quality of life
AJ Cameron et al
303
21 Australian Bureau of Statistics. National Health Survey: SF36
Population Norms, Australia, 1995. Australian Bureau of Statistics:
Canberra, 1997, p 13.
22 Hawthorne G, Osborne RH, Taylor A, Sansoni J. The SF36 version
2: critical analyses of population weights, scoring algorithms and
population norms. Qual Life Res 2007; 16: 661–673.
23 De Vos K, Zaidi M. Equivalence scale sensitivity of poverty
statistics for the member states of the European Community.
Rev Income Wealth 1997; 43: 319–333.
24 Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti
Cancer Council of Victoria FFQ: relative validity of nutrient
intakes compared with weighed food records in young to middleaged women in a study of iron supplementation. Aust NZ J Publ
Health 2000; 24: 576–583.
25 White SL, Polkinghorne KR, Cass A, Shaw JE, Atkins RC, Chadban
SJ. Alcohol consumption and 5-year onset of chronic kidney
disease: the AusDiab study. Nephrol Dial Transplant 2009; 24:
2464–2472.
26 Magliano DJ, Barr EL, Zimmet PZ, Cameron AJ, Dunstan DW,
Colagiuri S et al. Glucose indices, health behaviors, and incidence
of diabetes in Australia: the Australian Diabetes, Obesity and
Lifestyle Study. Diabetes Care 2008; 31: 267–272.
27 Barr EL, Zimmet PZ, Welborn TA, Jolley D, Magliano DJ, Dunstan
DW et al. Risk of cardiovascular and all-cause mortality in
individuals with diabetes mellitus, impaired fasting glucose,
and impaired glucose tolerance: the Australian Diabetes, Obesity,
and Lifestyle Study (AusDiab). Circulation 2007; 116: 151–157.
28 Australian Institute of Health and Welfare (AIHW). The Active
Australia Survey. A Guide and Manual for Implementation, Analysis
and Reporting. Australian Institute of Health and Welfare (AIHW):
Canberra, 2003.
29 Dunstan DW, Barr EL, Healy GN, Salmon J, Shaw JE, Balkau B
et al. Television viewing time and mortality: the Australian
Diabetes, Obesity and Lifestyle Study (AusDiab). Circulation 2010;
121: 384–391.
30 Vickers AJ, Altman DG. Statistics notes: analysing controlled
trials with baseline and follow up measurements. BMJ 2001; 323:
1123–1124.
31 Yancy Jr WS, Olsen MK, Westman EC, Bosworth HB, Edelman D.
Relationship between obesity and health-related quality of life in
men. Obes Res 2002; 10: 1057–1064.
32 Stafford M, Hemingway H, Marmot M. Current obesity, steady
weight change and weight fluctuation as predictors of physical
functioning in middle aged office workers: the Whitehall II
Study. Int J Obes Relat Metab Disord 1998; 22: 23–31.
33 Brown WJ, Mishra G, Kenardy J, Dobson A. Relationships
between body mass index and well-being in young Australian
women. Int J Obes Relat Metab Disord 2000; 24: 1360–1368.
34 Coakley EH, Kawachi I, Manson JE, Speizer FE, Willet WC,
Colditz GA. Lower levels of physical functioning are associated
with higher body weight among middle-aged and older women.
Int J Obes Relat Metab Disord 1998; 22: 958–965.
35 Lean ME, Han TS, Seidell JC. Impairment of health and quality of
life in people with large waist circumference. Lancet 1998; 351:
853–856.
36 Fine JT, Colditz GA, Coakley EH, Moseley G, Manson JE, Willett
WC et al. A prospective study of weight change and health-related
quality of life in women. J Am Med Assoc 1999; 282: 2136–2142.
37 Laaksonen M, Sarlio-Lahteenkorva S, Leino-Arjas P, Martikainen P,
Lahelma E. Body weight and health status: importance of
socioeconomic position and working conditions. Obes Res 2005;
13: 2169–2177.
38 Bjorntorp P. Do stress reactions cause abdominal obesity and
comorbidities? Obes Rev 2001; 2: 73–86.
39 Alexander CM, Landsman PB, Grundy SM. The influence of age
and body mass index on the metabolic syndrome and its
components. Diabetes Obes Metab 2008; 10: 246–250.
40 Zabelina DL, Erickson AL, Kolotkin RL, Crosby RD. The effect of
age on weight-related quality of life in overweight and obese
individuals. Obesity (Silver Spring) 2009; 17: 1410–1413.
41 Yan LL, Daviglus ML, Liu K, Pirzada A, Garside DB, Schiffer L et al.
BMI and health-related quality of life in adults 65 years and older.
Obes Res 2004; 12: 69–76.
42 Hopman WM, Berger C, Joseph L, Towheed T, VandenKerkhof E,
Anastassiades T et al. The natural progression of health-related
quality of life: results of a five-year prospective study of
SF-36 scores in a normative population. Qual Life Res 2006; 15:
527–536.
43 Kolotkin RL, Meter K, Williams GR. Quality of life and obesity.
Obes Rev 2001; 2: 219–229.
44 Forhan M, Vrkljan B, MacDermid J. A systematic review of the
quality of psychometric evidence supporting the use of an
obesity-specific quality of life measure for use with persons who
have class III obesity. Obes Rev 2009; 11: 222–228.
45 Cameron AJ, Welborn TA, Zimmet PZ, Dunstan DW, Owen N,
Salmon J et al. Overweight and obesity in Australia: the 1999–
2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab).
Med J Aust 2003; 178: 427–432.
46 OECD. Health at a Glance 2007: OECD Indicators. Organisation for
Economic Cooperation and Development: Paris, France, 2007.
Supplementary Information accompanies the paper on International Journal of Obesity website (http://www.nature.com/ijo)
International Journal of Obesity