Social determinants of smoking behavior: the Healthy Twin Study, Korea.

Abstract

The purpose of this study was to identify any influence of socioeconomic status on smoking and smoking cessation in a situation where genetic factors are controlled.
The sample for this study was 2502 members of the twins and families cohort who participated in the Korean Healthy Twins Study from 2005 to 2009. Groups of brothers or sisters, including twins and fraternal twins, were compared in terms of smoking and smoking cessation behaviors according to differences in socioeconomic status and gender.
In a situation with complete control of genetic factors, results showed that the daily smoking amount, cumulative smoking amount, and dependence on nicotine decreased with higher-status occupations, and the rate of smoking and amount of cumulative smoking decreased with higher levels of education. Regarding smoking cessation behavior, a higher level of education was associated with a lower smoking cessation rate, and no significant gender differences were found.
Environmental factors had a stronger influence on smoking behavior than did genetic factors. Genetic factors had greater influence on smoking cessation than did environmental factors; however, this requires verification in further studies.

Full text

Page 1

INTRODUCTION
According to current theories on health, factors that
can affect human health include socioeconomic level,
environmental causes, biological causes including
genetic causes, individual health-related behaviors, and
the medical system of a country [1]. Individual health-
related factors include smoking, drinking, exercise,
eating habits, and sleeping habits [2]; of these, smoking
is especially important. According to the World Health
Organization (WHO), in the year 2009, 5 million people
worldwide die from smoking-related causes every year,
and smoking is one of the most important preventable
risks to human health [3-5].
One environmental influence on human health is
socioeconomic status (education level, income level,
occupation), as has been reported in previous studies [6-
8]. Differing socioeconomic levels are associated with
different smoking behaviors; in lower socioeconomic
levels, people have a higher tendency to smoke [9-14].
In Korea, low socioeconomic status is associated with
risky smoking-related behaviors [15-18] and with a low
rate of success rate in quitting smoking [15,17].
Genetic factors can also affect smoking behaviors, and
previous studies have shown that not only nicotine
addiction but also behaviors related to smoking have a
genetic component [19-23]. These findings show that
socioeconomic status, other environmental causes, and
genetic causes are all associated with smoking, but most
research conducted in Korea has focused on the
relationship between socioeconomic factors and
smoking. There is no previous report on smoking-related
environmental and genetic causes in twins. The effects
of genetic factors on an individual’ s health differ
according to population group [24]. Thus, it is important
to examine how socioeconomic level is related to
smoking among Korean subjects when genetic factors
are controlled.
Based on Korea’s twin cohort data, in this study, we
examined how socioeconomic factors affect smoking
behavior (smoking rate, daily smoking amount,
cumulative smoking amount), efforts to stop smoking
and the success of those efforts, and the degree to which
a smoker is dependent on nicotine.
http://dx.doi.org/10.3961/jpmph.2012.45.1.29
pISSN 1975-8375
eISSN 2233-4521
Journal of Preventive Medicine and Public Health
January 2012, Vol. 45 No. 1, 29-36
This article is available at http://jpmph.org/.
Social Determinants of Smoking Behavior:
The Healthy Twin Study, Korea
Youn Sik Kim, Hansoo Ko, Changgyo Yoon, Dong-Hun Lee, Joohon Sung
Graduate School of Public Health, Seoul National University, Seoul, Korea
O Or ri ig gi in na al l A Ar rt ti ic cl le e
Corresponding author : Joohon Sung, MD, PhD
559 Gwanak-ro, Gwanak-gu, Seoul 151-742, Korea
Tel: +82-2-880-2828, Fax: +82-2-880-2787, E-mail: jsung@snu.ac.kr
Received: 8 September 2011, Accepted: 22 October 2011
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.
Objectives: The purpose of this study was to identify any influence of socioeconomic status on smoking and smoking
cessation in a situation where genetic factors are controlled.
Methods: The sample for this study was 2502 members of the twins and families cohort who participated in the Korean
Healthy Twins Study from 2005 to 2009. Groups of brothers or sisters, including twins and fraternal twins, were compared
in terms of smoking and smoking cessation behaviors according to differences in socioeconomic status and gender.
Results: In a situation with complete control of genetic factors, results showed that the daily smoking amount, cumulative
smoking amount, and dependence on nicotine decreased with higher-status occupations, and the rate of smoking and
amount of cumulative smoking decreased with higher levels of education. Regarding smoking cessation behavior, a higher
level of education was associated with a lower smoking cessation rate, and no significant gender differences were found.
Conclusions: Environmental factors had a stronger influence on smoking behavior than did genetic factors. Genetic
factors had greater influence on smoking cessation than did environmental factors; however, this requires verification in
further studies.
Key words: Smoking, Smoking cessation, Social class, Twins
J Prev Med Public Health 2012;45(1):29-36
cc ○

Page 2

METHODS
I. Participants
This research was conducted from April 2005 to
December 2009 with twins and their family members
(n=2860) from Samsung Medical Center and Inje
University Busan Paik Hospital who volunteered to
participate in this study. Identical twins over the age of 30
were selected, and these twins’first-degree relatives,
non-identical twins with or without family members, and
4th-degree blood relationships were included if the
participants agreed; family members based on marriage
relationships (except for parents) were not included in
this research. A questionnaire was used to collect
information on diet and exercise, family history of
disease, smoking, drinking, changes in the environment,
and socioeconomic status. In this study, we used only the
questionnaire data related to smoking and socioeconomic
level. Data were collected from 2502 people (358 people
refused to complete the questionnaire). These subjects
included twin brother and sister pairs (males 332 [166
pairs], females 556 [278 pairs]) and siblings pairs
including a twin (males 732 [366 pairs], and females
1152 [576 pairs]).
II. Measures
Zygosity was confirmed using 16 short tandem repeat
(STR) markers (15 autosomal STR markers and 1 sex-
determining marker) in 67% of the twins, and zygosity
data from previous research were available for the rest
[25]. Variables showing socioeconomic level, income
level, educational level, and occupation used existing
literature as a reference [6,7,15]. For income, equivalized
household income was calculated as the square root of
the household income divided by the number of people
in the household. Based on this figure, subjects were
divided into quartiles, and four income groups were
established: low income, ≤25%; middle income, the
two quartiles comprising 25% to 75%; and high income,
>75% [26-28]. Education level was classified in there
levels: middle school graduation or less, high school
graduation, and college graduation and above. For
occupation, Statistics Korea’s standard occupation
classification table was used based on data that subjects
provided, and the occupations were classified into non-
manual and manual workers. People whose occupations
were manager (public and high-ranking officer in a
company; administration and management support
manager; expert service management position and its
manager; construction, electricity, and production-related
management position; sale and customer service
management position), specialist and related profession,
and white-collar workers were classified as non-manual
workers. People whose occupations were service
worker, sales-related worker, and simple labor workers
were characterized as manual workers [26,28,29].
People who did not fit in either category were classified
as “other” (housekeepers, students, and fully retired); data
from these subjects were excluded from further analyses.
Variables related to smoking behavior were smoking
status (present smoker: someone who is smoking at the
present and who has smoked more than 400 cigarettes;
past smoker: someone who does not smoke now but
who has smoked more than 400 cigarettes; non-smoker:
someone who has never smoked before or a smoker who
has smoked fewer than 400 cigarettes) [30], daily
smoking amount (cigarettes/d), cumulative smoking
amount (packs/y), and nicotine dependence (assessed
using the Fagerstrom test for nicotine dependence). To
find the extent of nicotine dependence, we used data that
the participants provided in the Korea Twin Study using
a questionnaire from previous research [31]. Variables
related to smoking cessation were attempts to quit
smoking (tried or did not try) and success in quitting
(continuous non-smoking for 6 months or longer).
III. Statistical Analyses
Based on previous studies [32,33], data were divided
into a sibling group and a monozygotic twin group
(MZ). All subject variables are presented as frequencies
and percentages; daily smoking amount, cumulative
smoking amount, and nicotine dependence are presented
as means with standard deviations. In the remaining
analyses, the analysis was done with common targets
only, with exclusion of missing values (the process
excluded missing-value targeted subjects with siblings or
twins if one twin was missing, and those data were not
included in the analysis). For smoking-related variables,
data were available for 937 pairs of siblings including
twins (male, 365 pairs; females, 572 pairs) and 442 pairs
of MZ twins (male, 165 pairs; female, 277 pairs). For
non-smoking-related variables, data were available for
909 pairs of siblings including twins (male, 341 pairs;
females, 568 pairs) and 436 pairs of MZ twins (males,
161 pairs; females, 275 pairs). In sibling pairs with twins
included and in every pair of MZ twins (sibling pairs
were selected in the same sex siblings), to assess
differences in smoking behavior according to
socioeconomic level, differences in smoking rate were
30 Youn Sik Kim et al.
J Prev Med Public Health 2012;45(1):29-36

Page 3

Social Determinants and Smoking Behavior 31
J Prev Med Public Health 2012;45(1):29-36
calculated using conditional logistics, and odds ratios
(95% confidence interval [CI]) were used to indicate
differences in associations between smoking rate and
socioeconomic levels between the pairs. Differences in
daily smoking amount, cumulative smoking amount, and
nicotine dependence were observed, and differences in
smoking behavior values were calculated by regression
analysis and are expressed in terms of beta values (95%
CI) when socioeconomic status varied by one grade.
To observe changes in smoking cessation according to
socioeconomic level, we controlled for elements such as
age, gender, and drinking habits (yes or no, because
smoking cessation variables were binary) and then
assessed differences in the rate of smoking cessation
attempts and successes according to socioeconomic
status. In the siblings group that includes twins and in the
MZ twin group, we used conditional logistic regression
to compare the smoking-cessation attempt rate (failure to
quit smoking for 6 months) and smoking-cessation
success rate (success in quitting smoking for 6 months)
to compare sibling pairs (female siblings and male
siblings paired together) according to socioeconomic
level. Because it was possible that the smoking rate
would differ according to gender, the analysis was done
separately by gender. All analyses were done with SAS
version 9.2 (SAS Inc., Cary, NC, USA).
RESUL TS
Table 1 shows the data on socioeconomic level and
smoking-related variables. The sibling group including
Table 1. Characteristics of study population from the Healthy Twin Study
Subject characteristics
All pairs (Sibling+twins)1
MaleFemale
Monozygotic twins
Male Female
No. of participants
Age (y)
Education level
Middle school or lower
High school
College or higher
Equivalized household income
≤252
25 - 753
>754
Occupation
Manual
Non-manual
Others5
Smoking status
Non-smoker
Ex-smoker
Current smoker
Smoking cessation attempted
No
Yes
Successful smoking cessation
No (not or less than 6 mo)
Yes (more than 6 mo)
Alcohol status
Non-drinker
Past-drinker
Current-drinker
Cigarettes/d6
Packs year6
FTND6
732 (38.85)
39.72 (8.21)0
96 (13.11)
338 (46.17)
298 (40.71)
177 (24.18)
434 (59.29)
120 (16.39)
278 (37.98)
400 (54.64)
54 (7.38)0
172 (23.50)
179 (24.45)
381 (52.05)
103 (19.22)
433 (80.78)
374 (67.63)
179 (32.37)
82 (11.20)
34 (4.64)0
616 (84.15)
16.95 (7.99)0
16.42 (12.16)
3.52 (2.39)0
1152 (61.15)
39.04 (7.83)0
155 (13.45)
583 (50.61)
414 (35.94)
272 (23.61)
665 (57.73)
215 (18.66)
200 (17.36)
414 (35.94)
538 (46.70)
1025 (88.98)
40 (3.47)0
83 (7.20)0
37 (31.90)
79 (68.10)
80 (66.67)
40 (33.33)
357 (30.99)
116 (10.07)
679 (58.94)
12.51 (38.04)
6.08 (6.33)0
2.29 (2.28)0
332 (37.39)
39.15 (7.50)0
46 (13.86)
146 (43.98)
140 (42.17)
55 (16.57)
210 (63.25)
67 (20.18)
118 (35.54)
193 (58.13)
21 (6.33)0
84 (25.38)
86 (25.98)
161 (48.64)
34 (14.29)
204 (85.71)
159 (64.90)
86 (35.10)
52 (15.66)
13 (3.92)0
267 (80.42)
16.17 (6.99)0
14.89 (9.75)0
3.33 (2.18)0
556 (62.61)
37.78 (7.05)0
53 (9.53)0
290 (52.16)
213 (38.31)
85 (15.29)
324 (58.27)
147 (26.44)
98 (17.63)
218 (39.21)
240 (43.17)
498 (89.73)
16 (2.88)0
41 (7.39)0
15 (27.78)
39 (72.22)
40 (71.43)
16 (28.57)
165 (29.73)
59 (81.94)
331 (59.64)
9.57 (6.05)0
6.08 (5.18)0
2.59 (2.35)0
Values are presented as number (%).
FTND, Fagerstrom test for nicotine dependence.
1Dizygotic twins and monozygotic twins.
2<860000 won/mo.
3860000-1730000 won/mo.
4>1730000 won/mo.
5Included housewives, students, unemployed, retired.
6Current smokers only.

Page 4

twins contained 732 males (366 pairs; mean age, 39.72
±8.21 years) and 1152 females (576 pairs; mean age,
39.04±7.83 years), and the monozygotic twin group had
322 males (166 pairs; mean age, 39.15±7.50 years) and
556 females (278 pairs; mean age 37.78±7.05 years). In
both the siblings group and the monozygotic twin group,
males occupied majorities in the higher levels of
education and income. In terms of occupation, more
females than males were classified in the “others” group
including many women who were housewives. More
males than females reported occupations categorized as
non-manual work. In the sibling group (twins included)
52.05% were current smokers, and 48.64% of subjects in
the monozygotic twin group were current smokers. The
daily smoking amount, the cumulative smoking amount,
and nicotine dependence were higher in males than in
females.
Table 2 shows changes in smoking behavior according
to socioeconomic level with genetic causes controlled.
First, overall, single-grade increases in education level,
income level, and occupation level resulted in no
statistical difference in rate of smoking, daily smoking
amount, or cumulative smoking amount. However, in
the case of monozygotic twins, whose genetic factors
were 100% controlled, when education level increased
by one grade, smoking decreased by 0.67 times (95%
CI, 0.51 to 0.93), and cumulative smoking decreased by
2.9 fold (95% CI, -4.77 to -1.95; this difference was
statistically significant), but income level was not
significantly associated with smoking-related behaviors.
With respect to occupation level, an increase of one
grade in occupation level was associated with a 2.48-
fold decrease in daily smoking amount (95% CI, -3.52 to
-1.44), and cumulative smoking amount decreased by
8.7 fold (95% CI, -10.29 to -7.11); these differences
were statistically significant.
32 Youn Sik Kim et al.
J Prev Med Public Health 2012;45(1):29-36
Table 2. Smoking indices and social determinants in sibling and monozygotic twin pairs (analyses between within-
pair differences)
Difference (95% CI) per one grade increase in socio-economic status
Sibling+twin1(n=937 pairs)
Smoking
amount/d2
(β ±95% CI)
0.79 (0.62, 1.01)
0.97 (0.77, 1.23)
0.97 (0.80, 1.19)-0.91 (-2.36, 0.54)
Smoking rate
OR (95% CI)
Pack year3
(β ±95% CI)
Smoking rate
OR (95% CI)
Smoking amount/d
(β ±95% CI)
Pack year
(β ±95% CI)
Monozygotic twin (n=442 pairs)
Education4
Income5
Occupation6
-0.87 (-2.69, 0.95)
0.65 (-0.84, 2.14)
-2.05 (-4.74, 0.64)
-0.38 (-2.72, 1.97)
0.74 (-1.44, 2.93)
0.67* (0.51, 0.93)
1.30 (0.96, 1.75)
0.89 (0.69, 1.15)
-1.13 (-2.52, -0.42)
-0.63 (-2.10, 0.12)-
-2.48** (-3.52, -1.44)
-2.90** (-4.77, -1.95)0
-0.08 (-2.12, 0.96)-0
-8.70** (-10.29, -7.11)
Adjusted for age, gender, alcohol status.
OR, odds ratio; CI, confidence interval.
1Dizygotic twins and monozygotic twins.
2Cigarettes/d (current smokers only).
3Current smokers only.
4Per one-grade difference where grade I: middle school or lower, grade II: high school, grade III: college or higher.
5Per one-grade difference where grade I: ≤25% (<860000 won/mo), grade II: 25%-75% (860000-1730000 won/mo), grade III: >75% (>1730000 won/mo).
6Compared with manual labor.
*p<0.05, **p<0.01.
Table 3. Smoking cessation behaviors according to social determinants in sibling and monozygotic twin pairs
(analyses by pair-wise differences)
Siblings+twins1(n=909 pairs)
Attempted
smoking cessation
OR (95% CI)
Successful smoking
cessation (≥6 mo)
OR (95% CI)
FTND2
(β ±95% CI)
Attempted
smoking cessation
OR (95% CI)
Successful smoking
cessation (≥6 mo)
OR (95% CI)
FTND
(β ±95% CI)
Monozygotic twins (n=436 pairs)
Education3
Income4
Occupation5
1.68 (0.79, 3.56)
0.80 (0.44, 1.45)
1.35 (0.73, 2.48)
1.29 (0.68, 2.44)
1.21 (0.70, 2.10)
2.55** (1.50, 4.31)
-0.70* (-1.28, -0.11)
0.65* (0.15, 0.96)0
-0.28 (-0.77, 0.22)-
1.16 (0.42, 0.97)
0.70 (0.25, 1.96)
0.53 (0.15, 1.92)
0.45* (0.21, 0.97)
0.60 (0.28, 1.31)
1.17 (0.35, 3.84)
-0.18 (-0.25, 0.40)-
0.52* (0.09, 0.74)--
-1.29** (-1.80, -1.03)
Adjusted for age, gender, alcohol status.
OR, odds ratio; CI, confidence interval; FTND, Fagerstrom test for nicotine dependence.
1Dizygotic twins and monozygotic twins.
2Measured for current smokers only.
3Per one-grade difference where grade I: middle school or lower, grade II: high school, grade III: college or higher.
4Per one-grade difference where grade I: ≤25% (<860000 won/mo), grade II: 25%-75% (860000-1730000 won/mo), grade III: >75% (>1730000 won/mo).
5Compared with manual labor.
*p<0.05, **p<0.01.

Page 5

Social Determinants and Smoking Behavior 33
J Prev Med Public Health 2012;45(1):29-36
Table 3 shows changes in smoking-cessation behavior
according to socioeconomic level, controlling for genetic
factors. Within each gender, when education level and
income level increased by one grade, the odds ratio for
nicotine dependence dropped by 0.70 times (95% CI, -
1.28 to -0.11) and increased 0.65 times (95% CI, 0.15 to
0.96), respectively (statistically significant differences),
but no similar result was found for stop-smoking
attempts or stop-smoking success. When occupation
level increased by one grade, the odds ratio for stop-
smoking success was 0.45 times higher (95% CI, 0.21 to
0.97); this was a statistically significant difference.
However, no statistically significant difference was
found in the stop-smoking attempt rate or in the nicotine
dependence rate. When income level and occupation
level increased by one grade, the odds ratios for nicotine
dependence were 0.52 times higher (95% CI, 0.09 to
0.74) and 1.29 times lower (95% CI, -1.80 to -1.03),
respectively (statistically significant differences), but the
stop-smoking attempt rate and success rate were not
significantly different.
Table 4 shows changes in smoking behavior by gender
Table 4. Gender-specific smoking behavior patterns and social determinants in sibling pairs and twin pairs
(analyses by within-pair differences in the two factors)
Siblings+twins1
Smoking
amount/d2
(β ±95% CI)
n=730 (365 pairs)
-1.21 (-3.91, 1.49)0
-1.35 (-4.56, 1.86)0
-2.47 (-4.72, 0.22)0
n=1144 (572 pairs)
-0.18 (-7.37, 7.23)0
-6.73 (-13.77, 0.31)
-0.08 (-5.14, 4.98)0
Smoking rate
OR (95% CI)
Pack year3
(β ±95% CI)
Smoking rate
OR (95% CI)
Smoking amount/d
(β ±95% CI)
Pack year
(β ±95% CI)
Monozygotic twins
Male
Education4
Income5
Occupation6
Female
Education
Income
Occupation
0.62** (0.45, 0.84)
1.03 (0.77, 1.37)
0.90 (0.69, 1.19)
0.95 (0.59, 1.53)
1.01 (0.70, 1.47)
1.24 (0.90, 1.73)
-2.00 (-6.78, 2.78)0
-6.55 (-12.35, 0.75)
-2.66 (-7.19, 1.87)0
1.81 (-6.03, 9.65)0
5.05 (-2.75, 12.85)
1.07 (-4.20, 6.34)0
0.28** (0.18, 0.48)
0.67 (0.37, 1.19)
0.88 (0.43, 1.82)
0.99 (0.63, 1.59)
1.03 (0.70, 1.54)
1.11 (0.79, 1.56)
n=331 (165 pairs)
-1.03 (-2.54, 0.48)-
-0.28 (-1.93, 1.37)-
-2.63** (-4.43, -0.83)
n=554 (277 pairs)
-1.51 (-5.86, 2.84)-
-3.11 (-6.50, 0.28)-
-1.15 (-3.89, 1.59)-
-2.99** (-4.81, -1.17)
-0.85 (-2.87, 1.17)-
-6.75** (-8.96, -4.54)
0.07 (-6.04, 6.19)
-3.19 (-6.91, 0.53)
2.79 (-1.40, 6.98)
Adjusted for age, alcohol status.
OR, odds ratio; CI, confidence interval.
1Dizygotic twins and monozygotic twins.
2Cigarettes/d (current smokers only).
3Current smokers only.
4Per one-grade difference where grade I: middle school or lower, grade II: high school, grade III: college or higher.
5Per one-grade difference where grade I: ≤25% (<860000 won/mo), grade II: 25%-75% (860000-1730000 won/mo), grade III: >75% (>1730000 won/mo).
6Compared with manual labor.
*p<0.05, **p<0.01.
Table 5. Gender-specific smoking-cessation behavior and social determinants in sibling and twin pairs (using
within-pair differences)
Sibling+twin1
Successful
smoking cessation
(≥6 mo)
OR (95% CI)
Attempted
smoking cessation
OR (95% CI)
FTND2
(β ±95% CI)
Attempted
smoking cessation
OR (95% CI)
Successful
smoking cessation
(≥6 mo)
OR (95% CI)
FTND
(β ±95% CI)
Monozygotic twin
Male
Education3
Income4
Occupation5
Female
Education
Income
Occupation
1.07 (0.63, 1.81)
0.85 (0.54, 1.35)
1.00 (0.66, 1.53)
1.75 (0.40, 7.69)
1.36 (0.41, 4.54)
0.67 (0.26, 1.69)
n=682 (341 pairs)
1.51 (0.93, 2.46)
0.85 (0.56, 1.30)
1.62* (1.09, 2.42)
n=1136 (568 pairs)
1.07 (0.25, 4.51)
1.61 (0.35, 7.41)
0.57 (0.18, 1.80)
-0.44 (-1.20, 0.32)
0.04 (-0.52, 0.61)
-0.80 (-2.09, 0.49)
-0.64 (-3.07, 1.79)
1.55 (-0.96, 4.06)
1.15 (-0.67, 2.97)
1.10 (0.63, 1.94)
0.89 (0.52, 1.41)
1.18 (0.72, 1.94)
0.99 (0.30, 3.27)
0.98 (0.37, 2.59)
1.29 (0.55, 3.02)
n = 322 (161 pairs)
0.77 (0.49, 1.22)
0.83 (0.55, 1.24)
1.13 (0.76, 1.67)
n=550 (275 pairs)
0.78 (0.49, 1.22)
0.83 (0.56, 1.24)
1.13 (0.77, 1.67)
-0.37 (-0.82, 0.08)-
0.62 (-0.26, 1.50)-
-0.72* (-1.31, -0.13)
-0.13 (-1.11, 0.85)-
-0.46 (-1.07, 0.15)-
-0.44 (-1.09, 0.21)-
Adjusted for age, alcohol status.
OR, odds ratio; CI, confidence interval; FTND, Fagerstrom test for nicotine dependence.
1Dizygotic twins and monozygotic twins.
2Current smokers only.
3Per one-grade difference where grade I: middle school or lower, grade II: high school, grade III: college or higher.
4Per one-grade difference where grade I: ≤25% (<860000 won/mo), grade II: 25%-75% (860000-1730000 won/mo), grade III: >75% (>1730000 won/mo).
5Compared with manual labor.
*p<0.05, **p<0.01.