© 2010 The Korean Academy of Medical Sciences.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0)
which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
pISSN 1011-8934
eISSN 1598-6357
Relationship between Obesity and Depression in the Korean
Working Population
This study was conducted to investigate potential relationship between obesity and
depression in Korean working population. A total of 8,121 workers, aged 21-75 yr,
participated in this nationwide cross-section research. The participants asked to complete a
questionnaire regarding the socio-demographic factors and health-related behaviors,
height, and weight. To measure degree of depression, the Korean version of the Center for
Epidemiologic Studies-Depression scale (CES-D) was used. Chi-square test and multivariate
logistic regression analyses were performed to examine the degree of association between
obesity and depression. Compared to normal-weight women, overweight women had a
lower adjusted odds ratio (OR) for depression (OR=0.65, 95% confdence interval 0.46 to
0.97). Underweight (1.42, 0.83 to 2.44) and obese women (BMI ≥30) had higher ORs for
depression (1.47, 0.64 to 3.36), but these were insignifcant. Compared to normal-weight
men, obese men had higher crude ORs (1.94, 1.03 to 3.62) and adjusted ORs (1.77, 0.89 to
3.53) for depression, while underweight and overweight men showed no signifcant
differences. These fndings suggest that being overweight may protect Korean female
worker against depression.
Key Words: Obesity; Depression; Body Mass Index; Workers
Ji-Yong Kim1, Hye-Mi Chang2,
Jung-Jin Cho3, Sang-Ho Yoo3,
and Soo-Young Kim4
Seocho SamSung Med Clinic1, Kangbuk Samsung
Hospital, Seoul; Department of Family Medicine2,
CHA Hosipital, Seoul; Department of Family
Medicine3, Hallym University Sacred Heart Hospital,
Hallym University College of Medicine, Anyang;
Department of Preventive Medicine4, Eulji University
College of Medicine, Daejeon, Korea
Received: 26 February 2010
Accepted: 30 June 2010
Address for Correspondence:
Jung-Jin Cho, M.D.
Department of Family Medicine, Hallym University Sacred Heart
Hospital, Hallym University College of Medicine, 170
Gwanpyeong-ro, Dongan-gu, Anyang 431-070, Korea
Tel: +82.31-380-1783, Fax: +82.31-380-1782
E-mail: threej@hallym.or.kr
This research was supported by the Occupational Safety and
Health Research Institute (2005).
DOI: 10.3346/jkms.2010.25.11.1560 • J Korean Med Sci 2010; 25: 1560-1567
ORIGINAL ARTICLE
Medicine General & Social Medicine
INTRODUCTION
Obesity and depression are common disorders among Korean
adults (1). Tey contribute to substantial morbidity and mortality (2, 3), and the management of these disorders has been a
critical challenge for individual patients, physicians and the
community at large (4). Tey can lead to many complications
including cardiovascular diseases and type 2 diabetes mellitus
(5, 6).
Because of their close comorbid occurrence and mutual influences, recent studies have been giving attention to the topic
how obesity can be associated with depression (7). However,
fndings on the probable connection between obesity and depression have been inconsistent. Atlantis and Baker reported
a systematic review of studies on the seeming relationship between obesity and depression with prospective cohort studies
and cross-sectional studies (8). Based on their review, results of
all prospective cohort studies were consistent and suggested that
obesity was associated with a possible occurrence of depression.
Most cross-sectional studies from the United States supported a
positive association between obesity and depression in women
but not in men. In contrast, most cross-sectional studies outside
the United States failed to find such associations (8). In Asia,
studies revealed similar results that obesity had a positive association with depression (8). However, those studies were performed only on the elderly (9-11) and they used their own criteria in diagnosing obesity and depression (10).
Several factors have been listed to explain the inconsistencies
in those fnding of the cross-sectional studies. Te methodological differences in the diagnosis of obesity and depression among
different populations need to be standardized (12). Another reason is some possible ethnic differences of the relationship between obesity and depression (9). However most existing studies on this topic were done on Western populations, and only a
few studies have been conducted in Asians (10, 11).
Hence, we investigated the relationship between obesity and
depression in the Korean working population in order to determine whether the association in Korean context differs from
that had been conducted in previous studies.
MATERIALS AND METHODS
Study participants
Tis is a cross-sectional study among workers (21-75 yr of age)
recruited from a nationwide sample from April to October 2005
in Korea. Tere were 8,121 participants from 329 companies. A
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
DOI: 10.3346/jkms.2010.25.11.1560 http://jkms.org 1561
random (cluster) sample of companies was taken in two stages.
Te frst stage was a proportional selection of sample companies based on the Korean Standard Industrial Classification,
while the second selection was based on company size according to three groups: companies with below 99 employees, 100
to 299, and over 300. After selecting the companies, the survey
participants were subjected under the annual medical screenings until fnally coming up with an adequate number of subjects proportional to the total distribution of workers in Korea.
Anthropometric measures
Just before the survey, height and weight of participants were
measured during the annual medical check-up provided by
their employers. Survey participants were asked to give their
height and weight in the questionnaire. Body mass index (BMI
kg/m2) was calculated using these two values. Te subjects were
then categorized into four groups according to BMI: underweight
(BMI<18.5), normal (18.5≤BMI<25), overweight (25≤BMI<30),
and obese (BMI>30). In the Asia-Pacifc context, BMI over 25 is
considered obese, but not for Europe or the Unites States, according to the World Health Organization (WHO) 2002 Criteria (13).
In the present research, we used the WHO criteria in determine
the relationship between obesity and depression. The WHO
Expert Consultation recommended that the current WHO BMI
cut-off points should be retained as the international classifcation because the observed risk varied according to the BMI (14).
Socio-demographic and occupational factors
Data including occupational status, monthly household income,
educational level, marital status, past medical history, family history of cardiovascular disease, smoking habit, alcohol drinking,
and exercise were obtained using the self-reported questionnaire.
Participants were grouped according to the following: age
was categorized as below 30, 31-40, 41-50, and above 50 yr old.
Educational level was classifed into four groups: below middle
school, below high school, below college, and above the college.
Monthly household income was categorized into four groups:
less than 1.25 million Korean won, 1.25 to 2.0 million, 2.0 to 3.8
million, and over 3.8 million. Marital status was divided into three
groups: married, never married, and divorced, widowed or separated. Smoking habit was classified into current smoker, exsmoker and never smoked. Alcohol drinking habit was grouped
as more than 5 times a week, 3-4 times, 1-2 times, less than 1 time
a week, and non-drinker. Exercise status was classifed as daily,
5-6 times a week, 3-4 times, less than 3 times, and never. Alcohol drinking habit and exercise status were fnally dichotomized
as ever (at least once a week) versus never.
Depression measure
Te Korean version of the Center for Epidemiologic Studies-Depression scale (CES-D) was used to measure depression. The
CES-D is a self-reported questionnaire which consists of 20 questions. Two optimal cutoff points were suggested by Cho et al. :
24/25 as the point which corresponds best to the clinical diagnosis of depression and 20/21 as the point which most effectively detects and covers depressive symptoms during screening
(15, 16). A cutoff score of 21 or above is appropriate for epidemiologic purpose in a community setting and a cutoff score of
25 or above is appropriate for clinical diagnosis for depression.
With a cutoff score of 25 or above, relationships with the diagnosis of major depression in a community study, which used a
structured interview and the Diagnostic and Statistical Manual
of Mental Disorders (DSM) criteria, led to a highest kappa value
and a sensitivity of 91% (15, 16).
Survey participants with score of 25 or higher on the CES-D
were classifed as depressed.
Statistical analysis
Age and CES-D scores were presented as mean and standard
deviation. To evaluate the relationship between various factors
and depression, chi square test was used. Multivariate logistic
regression analysis was used, with the BMI groups as independent variable and depression as dependant variable, to evaluate the relationship between obesity and depression. For multivariate logistic regression analysis, BMI groups were compared
with normal weight group as reference. To determine probable
interactions between sex, BMI groups, and depression (7, 17), a
general linear model was established. We found an interaction
between sex and BMI groups (P=0.017), and therefore we analyzed our data separately according to gender in multivariate
logistic regression analysis. Covariates were age, educational
level, monthly household income, marital status, smoking habit, alcohol drinking habit, and exercise status. Te crude odds
ratios (ORs) were calculated to investigate the unadjusted relationship between obesity and depression, and then two sets of
models were ftted. Te frst model included age, monthly household income, marital status, and educational level to adjust to
the well-known risk factors of depression. The second model
included alcohol drinking habit and exercise status to adjust to
how lifestyle factors affect depression.
Statement of ethics
We received IRB approval for our study after we started to collect the accomplished self-reported questionnaires (Hallym research number 2010-1028). In the IRB approval, informed consent from the survey participants was waived.
RESULTS
Socio-demographic and baseline characteristics of the
subjects
A total of 5,231 male and 2,890 female workers participated in
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
1562 http://jkms.org DOI: 10.3346/jkms.2010.25.11.1560
the survey. Te mean age was 38.6±9.6 yr. Te mean value of the
CES-D scores was 11.0±8.87. Among the total number of participants, 23.2% were overweight and 1.6% was obese. About half
or 56.8% regularly exercised at least once a week. Tere were signifcant distribution difference in all the variables between male
and female, except exercise status (Table 1).
Table 2 shows the socio-demographic and baseline characteristics of the male participants. Only two variables, smoking
habit and alcohol drinking habit, were statistically different between the BMI groups: there were more current smokers in the
obese group and more alcohol drinkers in the overweight group.
Te prevalence of depression assessed by CES-D was not different between BMI groups.
Table 3 shows that there were no statistically different variable
among the female BMI groups, except for the smoking habit and
the prevalence of depression.
Bivariate analyses
Based on the CES-D assessment (≥25), the prevalence of depression was 10.6% of the total, 9.7% in male and 12.3% in female.
Te prevalence of depression was statistically different between
male and female.
Tere was lower prevalence of depression in 31-40 yr old age
group and those who regularly exercised. On the other hand,
there was high prevalence of depression in the female, divorced,
widowed or separated, and low educational level groups, and
Table 1. Baseline characteristics with full samples
Parameters
Male Female Total
P value
No. (%) No. (%) No. (%)
No. 5,231 (64.4) 2,890 (35.6) 8,121 (100)
Age (yr) 41.35±8.84 33.73±9.02 38.64±9.62 <0.001
BMI (mean±SD) 22.89±2.99 23.20±3.01 23.00±3.00 <0.001
CES-D 10.46±8.49 11.99±9.44 11.0±8.87 <0.001
Age (yr)
≤30
31-40
41-50
51-
607 (11.5)
1,921 (36.3)
1,861 (35.2)
897 (17.0)
1,371 (47.2)
882 (30.4)
453 (15.6)
199 (6.9)
1,978 (24.1)
2,803 (34.2)
2,314 (28.3)
1,096 (13.4)
<0.001
Educational level
Middle school
High school
College
Over college
212 (4.1)
1,958 (37.6)
2,617 (50.3)
415 (8.0)
158 (5.6)
793 (27.8)
1,787 (62.7)
112 (3.9)
370 (4.6)
2,751 (34.2)
4,404 (54.7)
527 (6.5)
<0.001
Marital status
Married
Single
Divorce
4,149 (79.6)
981 (18.8)
80 (1.5)
1,404 (48.8)
1,401 (48.7)
73 (2.5)
5,553 (68.7)
2,382 (29.5)
153 (1.9)
<0.001
Monthly income (million won)
<1.25
1.25-<2.0
2.0-<3.8
≥3.8
562 (11.1)
1,226 (24.3)
987 (19.6)
2,273 (45.0)
1,275 (45.9)
766 (27.6)
490 (17.7)
245 (8.8)
1,837 (23.5)
1,992 (25.5)
1,477 (18.9)
2,518 (32.2)
<0.001
Smoking
Never
Ex-smoker
Current
1,298 (23.0)
1,590 (28.1)
2,764 (48.9)
2,294 (93.1)
82 (3.3)
89 (3.6)
3,592 (44.3)
1,672 (20.6)
2,853 (35.1)
0.003
Alcohol drinking
None
At least once/wk
1,150 (21.9)
4,105 (78.1)
1,465 (51.0)
1,407 (49.0)
2,615 (32.2)
5,512 (67.8)
<0.001
Exercise
None
At least once/wk
2,299 (43.8)
2,952 (56.2)
1,211 (42.0)
1,671 (58.0)
3,510 (43.2)
4,623 (56.8)
0.125
BMI
Underweight
Average
Overweight
Obese
310 (5.9)
3,699 (70.3)
1,176 (22.3)
78 (1.5)
131 (4.5)
1,996 (68.9)
715 (24.7)
55 (1.9)
441 (5.4)
5,695 (69.8)
1,891 (23.2)
133 (1.6)
0.004
CES-D
<25
≥25
4,794 (90.3)
466 (9.7)
2,577 (87.7)
316 (12.3)
7,371 (89.4)
782 (10.6)
0.002
Data are No. (%) or mean±SD. BMI denotes body mass index, calculated as weight (kg) divided by square height (m2). CED-D denotes the Center for Epidemiologic StudiesDepression scale. Underweight group has BMI less than 18.5. Average group has BMI between 18.5 and 24.9. Overweight group has BMI between 25.0 and 30.0. Obese
group has BMI over 30.0. Difference between means by Student t-test or difference between proportions by chi square test.
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
DOI: 10.3346/jkms.2010.25.11.1560 http://jkms.org 1563
was even higher in the underweight and obese groups. Tose
who smoke or drink alcohol had no association with depression
(Table 4).
Multivariate analyses
Compared to the normal weight women, overweight women
had lower OR for depression (OR=0.62, 95% confdence interval
0.45 to 0.84), while underweight women (1.57, 0.97 to 2.54) and
obese women (1.29, 0.60 to 2.77) had an insignifcantly higher
OR for depression. Tese relationships remained even after covariates were adjusted in Models 1 and 2.
Compared to normal weight men, overweight men (1.04, 0.83
to 1.31), underweight (1.24, 0.84 to 1.82) and obese (1.94, 1.03
to 3.62) had insignifcantly higher ORs for depression. However,
after covariates in Models 1 and 2 were adjusted, the OR of overweight men turned negative (0.93, 0.72 to 1.20 in Model 1, 0.90,
0.70 to 1.17 in Model 2), but still were not signifcant. Te odd
ratios of underweight and obese men for depression remained
the same after the adjustments (Table 5).
DISCUSSION
In this study, we investigated the cross-sectional relationship
between obesity and depression in Korea. Based on the Korean
CES-D, the prevalence of depression in men was 8.9% and 10.9%
in women. Tese values were lower compared to previous survey results. In the sample collected during the National Health
and Health Behavior Examination Survey in Korea, the prevalence of depression as measured by the Korean CES-D (above
the cutoff of 25) was 6.8% in men and 10.4% in women (18). Te
prevalence of depression in a study conducted in a non-institutionalized general population using DSM-IV criteria was 20.4%
in men and 21.5% in women (19). Such discrepancies between
our results and the previous ones might be due to methodological differences in diagnosis and the healthy worker effect.
As our multivariate analyses showed, the relationship between
one’s body weight and depression depended on BMI status. Te
overweight group revealed a negative association with depression while the obesity group showed positive association especially in women, although a similar pattern was also identifed
in men. Te overweight women group showed an OR of 0.62,
Table 2. Baseline characteristics in male study population
Parameters Age (yr) |
Underweight | Normal | Overweight | Obese | P value 0.190 |
≤30
31-40
41-50
51-
38 (12.4)
106 (34.5)
105 (34.2)
58 (18.9)
433 (11.7)
1,374 (37.1)
1,303 (35.2)
589 (15.9)
126 (10.7)
405 (34.4)
415 (35.3)
230 (19.6)
6 (7.7)
29 (37.2)
27 (34.6)
16 (20.5)
Educational level
Middle school
High school
College
Over College
12 (4.0)
131 (43.2)
143 (47.2)
17 (5.6)
139 (3.8)
1,346 (36.9)
1,871 (51.3)
290 (8.0)
57 (5.0)
437 (38.0)
560 (48.7)
97 (8.4)
3 (3.9)
34 (44.7)
31 (40.8)
8 (10.5)
0.152
Marital status
Married
Single
Divorce
236 (77.9)
60 (19.8)
7 (2.3)
2,895 (79.4)
696 (19.1)
56 (1.5)
938 (81.0)
203 (17.5)
17 (1.5)
62 (81.6)
14 (18.4)
0 (0.0)
0.658
Monthly Income (million won)
<1.25
1.25-<2.0
2.0-<3.8
≥3.8
29 (9.9)
67 (22.8)
55 (18.7)
143 (48.7)
382 (10.8)
894 (25.3)
706 (20.0)
1,548 (43.8)
141 (12.5)
244 (21.7)
203 (18.1)
536 (47.7)
7 (9.3)
17 (22.7)
15 (20.0)
36 (48.0)
0.713
Smoking
Never-smoker
Ex-smoker
Current smoker
17 (24.6)
16 (23.2)
36 (52.2)
704 (23.5)
830 (27.7)
1,457 (48.7)
328 (22.8)
452 (31.5)
657 (45.7)
23 (24.2)
20 (21.1)
52 (54.7)
0.015
Exercise
None
Yes
148 (48.4)
158 (51.6)
1,601 (43.6)
2,071 (56.4)
501 (42.8)
669 (57.2)
36 (46.8)
41 (53.2)
0.337
Alcohol Drinking
None
At least once/wk
45 (14.7)
264 (85.2)
810 (22.0)
2,869 (78.0)
273 (23.4)
892 (76.6)
17 (21.8)
61 (78.2)
0.014
CES-D
<25
≥25
275 (89.6)
32 (10.4)
3,382 (91.4)
317 (8.6)
1,071 (91.1)
105 (8.9)
66 (84.6)
12 (15.4)
0.143
Data are No. (%). BMI denotes body mass index, calculated as weight (kg) divided by square height (m2). CED-D denotes the Center for Epidemiologic Studies-Depression scale.
Underweight group has a BMI less than 18.5. Average group has a BMI between 18.5 and 24.9. Overweight group has a BMI between 25.0 and 29.9. Obese group has a BMI
over 30.0. Difference between proportions by chi square test.
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
1564 http://jkms.org DOI: 10.3346/jkms.2010.25.11.1560
which was substantial without the adjustments. Tis negative
association with depression remained the same after the covariates, such as age, monthly household income, educational level,
smoking, alcohol drinking habits, and exercise status, were controlled. Te overweight men revealed a crude OR of 1.04. For the
adjusted OR in Model 2, however, this value turned into 0.90,
which showed similar results with those of the women, while
still an insignifcant value. Tese results from our study were similar to those of other studies conducted in Asia (10, 11). Te study
conducted in China (11) revealed that the overweight women
and men had a negative association with depression. In Japan,
the results were quite similar with China, but the OR for women
decreased with a BMI of over 26 (10). According to a recent systematic review, most cross-sectional studies conducted in the
United States support a positive association between obesity
(BMI over 30) and depression in women. However, most crosssectional studies conducted outside the United States did not
support such associations (8).
Tere are several possible explanations that support our results.
First, environmental factors such as cultural background and
dietary habits may have been contributing factors. East Asian
countries such as Korea, China and Japan share similar cultural
backgrounds and dietary habits, and their staple food is mostly
rice. Carbohydrate, which is the main component of rice, is
believed to be a remedy in relieving symptoms of depression
through the serotonergic system and as a reason of weight gain
when overeaten. Serotonin release is also involved in various
functions such as in sleep onset, pain sensitivity, blood pressure
regulation, and mood control. Hence, people in these countries
have learned to overeat carbohydrates (particularly snack foods,
like potato chips or pastries) to make themselves feel better. Second, we may also consider biological mechanisms, such as the
jolly fat hypothesis. Tat hypothesis proposes that overweight
people, both male and female, have lower risk for depression
and show reduced depressive symptoms due to several possible mechanisms that may include higher consumption of certain nutrients that are helpful in reducing or preventing depressive symptoms (20, 21). They suggested the possible negative
emotional effect of food deprivation. On the contrary, dietary
habits such as periodic overeating of obese people may sometimes serve as protective mechanism against the experience and
display of anxiety and depression (20, 21).
Table 3. Baseline characteristics in female study population
Parameters Age (yr) |
Underweight | Normal | Overweight | Obese | P value 0.976 |
≤30
31-40
41-50
51-
63 (49.2)
39 (30.5)
17 (13.3)
9 (7.0)
947 (47.5)
598 (30.0)
311 (15.6)
139 (7.0)
328 (45.9)
225 (31.5)
116 (16.2)
46 (6.4)
28 (50.9)
15 (27.3)
7 (12.7)
5 (9.1)
Educational level
Middle school
High school
College
Over College
10 (8.0)
36 (28.8)
74 (59.2)
5 (4.0)
104 (5.3)
535 (27.2)
1,256 (64.0)
69 (3.5)
41 (5.9)
207 (29.7)
413 (59.3)
35 (5.0)
3 (5.7)
10 (18.9)
38 (71.7)
2 (3.8)
0.339
Marital status
Married
Single
Divorce
51 (41.1)
68 (54.8)
5 (4.0)
957 (48.3)
980 (49.4)
45 (2.3)
364 (51.5)
321 (45.4)
22 (3.1)
26 (49.1)
26 (49.1)
1 (1.9)
0.234
Monthly Income (million won)
<1.25
1.25-<2.0
2.0-<3.8
≥3.8
57 (47.9)
33 (27.7)
25 (21.0)
4 (3.4)
887 (46.5)
521 (27.3)
321 (16.8)
178 (9.3)
300 (43.8)
199 (29.1)
129 (18.8)
57 (8.3)
26 (49.1)
11 (20.8)
11 (20.8)
5 (9.5)
0.294
Smoking
Never-smoked
Ex-smoker
Current smoker
273 (91.9)
9 (3.0)
15 (5.1)
1,517 (94.3)
50 (3.1)
42 (2.6)
115 (97.5)
2 (1.7)
1 (1.6)
10 (66.7)
2 (13.3)
3 (20.0)
<0.001
Exercise
None
At least once/wk
60 (47.2)
67 (52.8)
822 (41.6)
1,156 (58.4)
302 (42.5)
408 (57.5)
22 (40.0)
33 (60.0)
0.622
Alcohol drinking
None
At least once/wk
56 (43.8)
73 (55.7)
1,014 (51.4)
960 (48.6)
363 (51.6)
341 (48.4)
26 (48.1)
28 (51.9)
0.431
CES-D
<25
≥25
106 (82.8)
22 (17.2)
1,763 (88.4)
232 (11.6)
661 (92.4)
54 (7.6)
47 (85.5)
8 (14.5)
0.002
Data are No. (%). BMI denotes body mass index, calculated as weight (kg) divided by square height (m2). CED-D denotes the Center for Epidemiologic Studies-Depression scale.
Underweight group has a BMI less than 18.5. Average group has a BMI between 18.5 and 24.9. Overweight group has a BMI between 25.0 and 29.9. Obese group has BMI
over 30.0. Difference between proportions by chi square test or Fisher’s exact test.
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
DOI: 10.3346/jkms.2010.25.11.1560 http://jkms.org 1565
Although we could not draw statistical signifcance from our
data, our results prove that the obese group is positively associated with depression in both men and women. Tis result is similar with the result of most cross-sectional studies from the United States, which supports a positive association between obesity and depression in women. Tis, however, is different from the
results of most cross-sectional studies outside the United States.
Also, the result of most cross-sectional studies from the United
States in men failed to find such associations (8). The studies
conducted in Asia, they did not separately assess a group with
BMI over 30. Tis means, therefore, that it is difcult to compare
our study with previous Asian studies (10, 11). With our analysis
using the above overweight group (BMI over 25.0) (Table 3), the
OR for depression in women (0.70, 0.51 to 0.97) showed similar
results with the results of most cross-sectional studies outside the
United States, which showed negative association with depression. Our fndings are unique primarily because we separately
assessed the association of overweight (25≤BMI<30) and obese
(BMI>30) groups, which were defned by the WHO criteria, with
depression (14).
One’s genetic predisposition was also considered as an important factor that determines how obesity may be related with
depression, especially in the obese group with BMI over 30 (6,
22). However, our fndings suggest that environmental factors
may also contribute to a positive association between obesity
and depression. After adjustments were made to factor in the
environment, the signifcance of the association between obesity and depression disappeared. Te male obese group showed
a statistically signifcant association with depression (OR=1.94),
but after factoring in the monthly household income and educational level, the association became statistically insignifcant.
Although results from the bivariate analyses of monthly household income and educational level of obese men were not statistically different from those of other BMI groups, it becomes
possible that socioeconomic factors serve as confounders. Also,
Table 4. Relationship of various characteristics and depression
Parameters
CES-D<25 CES-D≥25
P value
No. (%) No. (%)
Sex
Male
Female
4,819 (91.1)
2,588 (89.1)
467 (8.8)
317 (10.9)
0.002
Age
≤30
31-40
41-50
51-
1,784 (90.2)
2,582 (92.1)
2,098 (90.7)
943 (86.0)
194 (9.8)
221 (7.9)
216 (9.3)
153 (14.0)
<0.001
Educational level
Middle school
High school
College
Over college
286 (77.3)
2,412 (87.7)
4,099 (93.1)
506 (96.0)
84 (22.7)
339 (12.3)
305 (6.9)
21 (4.0)
<0.001
Marital status
Married
Single
Divorce
5,033 (90.6)
2,164 (90.8)
126 (82.4)
520 (9.4)
218 (9.2)
27 (17.6)
<0.001
Monthly Income (million won)
<1.25
1.25-<2.0
2.0-<3.8
≥3.8
1,601 (87.2)
1,799 (90.3)
1,364 (92.3)
2,354 (93.5)
236 (12.8)
193 (9.7)
113 (7.7)
164 (6.5)
<0.001
Smoking
Never
Ex-smoker
Current
3,091 (89.5)
1,437 (90.7)
2,426 (90.0)
362 (10.5)
147 (9.3)
269 (10.0)
0.413
Exercise
None
At least once/wk
3,137 (89.4)
4,217 (91.2)
373 (10.6)
406 (8.8)
0.005
Alcohol drinking
None
At least once/wk
2,367 (90.5)
4,990 (90.5)
248 (9.5)
522 (9.5)
0.985
BMI
Underweight
Average
Overweight
Obese
386 (87.5)
5,146 (90.4)
1,732 (91.6)
113 (85.0)
55 (12.5)
549 (9.6)
159 (8.4)
20 (15.0)
0.008
CED-D denotes the Center for Epidemiologic Studies-Depression scale. Underweight
group has a BMI less than 18.5. Average group has BMI between 18.5 and 24.9.
Overweight group has a BMI between 25.0 and 29.9. Obese group has BMI over 30.0.
P value by chi square test.
Table 5. Logistic regression analysis for obesity and depression
Crude OR P value Adjusted OR* P value Adjusted OR† P value
Normal weight 1.00 1.00 1.00
Underweight
Men
Women
1.24 (0.84-1.82)
1.57 (0.97-2.54)
0.26
0.06
1.27 (0.85-1.90)
1.47 (0.86-2.50)
0.23
0.15
1.30 (0.86-1.95)
1.42 (0.83-2.44)
0.20
0.19
Overweight
Men
Women
1.04 (0.83-1.31)
0.62 (0.45-0.84)
0.70
0.003
0.93 (0.72-1.20)
0.64 (0.46-0.89)
0.62
0.009
0.90 (0.70-1.17)
0.65 (0.46-0.97)
0.46
0.01
Obese
Men
Women
1.94 (1.03-3.62)
1.29 (0.60-2.77)
0.03
0.50
1.77 (0.89-3.53)
1.48 (0.65-3.36)
0.10
0.34
1.79 (0.89-3.59)
1.47 (0.64-3.36)
0.09
0.36
Above overweight
Men
Women
1.09 (0.87-1.37)
0.66 (0.49-0.89)
0.41
0.007
0.98 (0.77-1.25)
0.69 (0.50-0.95)
0.91
0.02
0.95 (0.74-1.22)
0.70 (0.51-0.97)
0.73
0.03
Depression: Score of the Center for Epidemiologic Studies-Depression scale ≥25. Underweight group has a BMI less than 18.5. Average group has a BMI between 18.5 and
24.9. Overweight group has a BMI between 25.0 and 29.9. Obese group has a BMI over 30.0. Above overweight group has BMI over 25.0.
*Adjusted for age, monthly household income, marital status, and educational level; †Adjusted for age, monthly household income, marital status, educational level, smoking,
exercise, and alcohol drinking.
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
1566 http://jkms.org DOI: 10.3346/jkms.2010.25.11.1560
body image might have negative emotional influence on our
participants, because the obese population with BMI over 30 is
not as prevalent (1.6%) as in other countries.
Our study has several limitations. First, the cross-sectional
nature of the study limited our capacity to demonstrate a causal
relationship between obesity and depression. However, any reciprocal relationships have not been previously proven; depression did not increase the risk of future obesity (23). Second, most
exposures and outcome measures of the study were based on a
subjective report and the self-reported questionnaire. Even
though the depression was not diagnosed by a physician using
DSM-IV criteria, we used a validated measure and defned depression with the cut-off point of 25, which is known to have
high specifcity for depression (15, 16). Height and weight were
also self-reported by the participants and the BMI was calculated based on these values. Just before the participants proceeded
to answer the questionnaire, their height and weight were measured objectively as their annual medical screening provided by
their employers for legal purpose. Tis process could decrease
any discrepancy between the self-reported data and the measured BMI. Tird, there is a possibility of selection bias. To minimize selection bias, we recruited our subjects proportional to
industrial classifcation and company size. However, upon including participants from their work places (24), it is hard to
avoid the ‘healthy worker effect’. However, as over one-third of
the current adult population in Korea belongs to the working
group; it is difcult to recruit subjects without the healthy worker effect. Fourth, the small sample size could lead to statistical
insignifcance. Te proportion of obese subjects (BMI over 30)
was only 1.6% of the total subjects, and this could lead to a broad
confdence interval. If our study included more participants belonging to the obese group, the association of that group with
depression might have resulted with much more signifcance.
Tere is a possibility that being obese (BMI over 30) could also
increase the risk of depression for both male and female in Korea.
Further prospective studies will be needed to confrm whether
indeed being overweight has a protective effect on depression
in women and that being obese (BMI over 30) increases the risk
of depression in Korea, as well as in other areas in Asia.
Our study also has several strengths. First, the study was based
on a wide-scale, nationwide survey, which included respondents belonging within the 21-75 yr old age range. Second, we
conducted multivariate analyses according to gender, which allowed us to identify gender-specifc patterns of association between obesity and depression (25), and we separately assessed
depression is associated with the overweight (25≤BMI<30) and
obese (BMI>30) groups, which were defned by the WHO criteria. This analysis, which was based on different body weight
groups, provided evidence that the association between obesity
and depression may depend on the severity of obesity in Asia.
In conclusion, recently conducted studies demonstrate a significant inverse relationship between depression and being
overweight in women.
ACKNOWLEDGMENTS
We are grateful to our co-investigator for this project and the
workers who participated in our survey.
REFERENCES
1. Korean Ministry of Health and Welfare. Korean national health and
nutritional examination survey. Seoul, 2005.
2. Kim NS, Moon OR, Kang JH, Lee SY, Jeong BG, Lee SJ, Yoon TH, Hwang
KH, Kim N, Moon O, Kang J, Lee S, Jeong B, Lee S, Yoon T, Hwang K. Increasing prevalence of obesity related disease for koreans associated with
overweight and obesity. Korean J Prev Med 2001; 34: 309-15.
3. Park JH, Yoon SJ, Lee HY, Cho HS, Lee JY, Eun SJ, Park JH, Kim Y, Kim
YI, Shin YS. [Estimating the burden of psychiatric disorder in Korea]. J
Prev Med Public Health 2006; 39: 39-45.
4. Jeong BG, Moon OR, Kim NS, Kang JH, Yoon TH, Lee SY, Lee SJ, Yoon T,
Lee S, Lee S. Socioeconomic costs of obesity for korean adults. Korean J
Prev Med 2002; 35: 1-12.
5. Van der Kooy K, van Hout H, Marwijk H, Marten H, Stehouwer C, Beekman A. Depression and the risk for cardiovascular diseases: systematic
review and meta analysis. Int J Geriatr Psychiatry 2007; 22: 613-26.
6. Faith MS, Matz PE, Jorge MA. Obesity-depression associations in the
population. J Psychosom Res 2002; 53: 935-42.
7. Stunkard AJ, Faith MS, Allison KC. Depression and obesity. Biol Psychiatry 2003; 54: 330-7.
8. Atlantis E, Baker M. Obesity effects on depression: systematic review of
epidemiological studies. Int J Obes (Lond) 2008; 32: 881-91.
9. Sachs-Ericsson N, Burns AB, Gordon KH, Eckel LA, Wonderlich SA,
Crosby RD, Blazer DG. Body mass index and depressive symptoms in
older adults: the moderating roles of race, sex, and socioeconomic status.
Am J Geriatr Psychiatry 2007; 15: 815-25.
10. Kuriyama S, Koizumi Y, Matsuda-Ohmori K, Seki T, Shimazu T, Hozawa
A, Awata S, Tsuji I. Obesity and depressive symptoms in elderly Japanese:
the Tsurugaya Project. J Psychosom Res 2006; 60: 229-35.
11. Li ZB, Ho SY, Chan WM, Ho KS, Li MP, Leung GM, Lam TH. Obesity and
depressive symptoms in Chinese elderly. Int J Geriatr Psychiatry 2004;
19: 68-74.
12. Dragan A, Akhtar-Danesh N. Relation between body mass index and
depression: a structural equation modeling approach. BMC Med Res
Methodol 2007; 7: 17.
13. World Health Organisation, International Association for the Study of
Obesity, International Obesity TaskForce. Te Asia-Pacifc Perspective:
Redefning obesity and its treatment. Sydney: Health Communications
Melbourne; 2000.
14. WHO Expert Consultaion. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363: 157-63.
15. Cho MJ, Kim KH. Diagnostic validity of the CES-D (Korean version) in
the assessment of DSM-III-R major depression. J Korean Neuropsychiatr
Assoc 1993; 32: 381-99.
Kim J-Y, et al. • Relationship between Obesity and Depression in the Korean Working Population
DOI: 10.3346/jkms.2010.25.11.1560 http://jkms.org 1567
16. Cho MJ, Kim KH. Use of the Center for Epidemiologic Studies Depression
(CES-D) Scale in Korea. J Nerv Ment Dis 1998; 186: 304-10.
17. Dong C, Sanchez LE, Price RA. Relationship of obesity to depression: a
family-based study. Int J Obes Relat Metab Disord 2004; 28: 790-5.
18. Cho MJ, Nam JJ, Suh GH. Prevalence of symptoms of depression in a nationwide sample of Korean adults. Psychiatry Res 1998; 81: 341-52.
19. Ohayon MM, Hong SC. Prevalence of major depressive disorder in the
general population of South Korea. J Psychiatr Res 2006; 40: 30-6.
20. Crisp AH, McGuiness B. Jolly fat: relation between obesity and psychoneurosis in general population. Br Med J 1976; 1: 7-9.
21. Crisp AH, Queenan M, Sittampaln Y, Harris G. ‘Jolly fat’ revisited. J Psychosom Res 1980; 24: 233-41.
22. Onyike CU, Crum RM, Lee HB, Lyketsos CG, Eaton WW. Is obesity associated with major depression? Results from the Tird National Health
and Nutrition Examination Survey. Am J Epidemiol 2003; 158: 1139-47.
23. Roberts RE, Deleger S, Strawbridge WJ, Kaplan GA. Prospective association between obesity and depression: evidence from the Alameda County
Study. Int J Obes Relat Metab Disord 2003; 27: 514-21.
24. Cho JJ, Kim JY, Chang SJ, Fiedler N, Koh SB, Crabtree BF, Kang DM, Kim
YK, Choi YH. Occupational stress and depression in Korean employees.
Int Arch Occup Environ Health 2008; 82: 47-57.
25. Cole MG, Dendukuri N. Risk factors for depression among elderly community subjects: a systematic review and meta-analysis. Am J Psychiatry
2003; 160: 1147-56.