http://dx.doi.org/10.4082/kjfm.2016.37.2.97 • Korean J Fam Med 2016;37:97-104

Original Article eISSN: 2092-6715

Relationship between Secondhand Smoking with Depressive Symptom and Suicidal Ideation in Korean Non-Smoker Adults: The Korean National Health and Nutrition Examination Survey 2010–2012 Wook Gim, Jun-Hyun Yoo*, Jin-Young Shin, Ae-Jin Goo Department of Family Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

Background: Research suggests that mental health is affected not only by smoking, but also by secondhand smoking. But the most researches have been conducted in North America and/or Europe. We examined whether this relationship remains evident within the South Korean population. Specifically, we investigated the effect of secondhand smoking on depressive symptoms and suicidal ideation. Methods: We analyzed data from 6,043 non-smoking adults who participated in the 2010–2012 Korea National Health and Nutrition Examination Survey. We compared the presence of depressive symptoms and suicidal ideation in 3,006 participants who were exposed to secondhand smoking in the office or at home with 3,037 non-exposed participants. Results: In unadjusted logistic regression analysis, secondhand smoking exposure group had more suicidal ideations than no secondhand smoking exposure (16.1% vs. 12.2%; odds ratio [OR], 1.50; 95% confidence interval [CI], 1.241–1.804), but risk of depressive symptoms was not significantly different between two groups (15.2% vs. 12.2%; OR, 1.21; 95% CI, 0.997–1.460). In multivariate logistic regression analysis, Among those exposed to secondhand smoking, the OR for depressive symptoms was 1.02 (95% CI, 0.866–1.299) and 1.43 (95% CI, 1.139–1.802) for suicidal ideation. Overall, secondhand smoking at home was significantly related to depressive symptoms and suicidal ideation. Among females, secondhand smoking exposure at home only (not in the office) was related to depressive symptoms and suicidal ideation. Conclusion: Exposure to secondhand smoking, especially at home, may be associated with an increase in especially in female depressive symptoms and suicidal ideation among adults in South Korea. Keywords: Tobacco Smoke Pollution; Depressive Symptoms; Suicidal Ideation

Received: August 13, 2015, Revised: September 4, 2015, Accepted: September 27, 2015 *Corresponding Author: Jun-Hyun Yoo  Tel: +82-2-3410-2440, Fax: +82-2-3410-0388, E-mail: [email protected]

Copyright © 2016 The Korean Academy of Family Medicine This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

98   www.kjfm.or.kr

Wook Gim, et al.  •  Relationship of Secondhand Smoking with Mental Health

INTRODUCTION

2. Data Collection

Smoking is a well-established risk factor for mental illness. Indeed, lit-

KNHANES, the following variables were analyzed for the purpose of

erature suggests that smokers are 1.5 times more likely to experience

this study: (1) physical measurements of weight and height drawn; (2)

major depression than non-smokers.1-3) Psychiatric disorders, like ma-

demographic data, including age, sex, level of education, marital sta-

jor depression disorder, can involve or lead to suicidal attempts and

tus, number of family members, level of income within household,

suicidal ideation,4) and previous studies report that smoking is an in-

type of health insurance, and employment status; (3) health status, in-

Taken from the self-report questionnaires and measurements of the

5-7)

Beyond firsthand

cluding history of gastric cancer, hepatic cancer, colon cancer, breast

smoke, secondhand smoking is related with an increased prevalence

cancer, cervical cancer, lung cancer, thyroid cancer, and other types of

of cardiovascular diseases and lung cancer, and recent research re-

cancer, current morbidity of osteoarthritis, rheumatic arthritis, atopic

veals that secondhand smoking may contribute to mortality.8-11) How-

dermatitis, myocardial infarction, chronic kidney disease, diabetes

ever, it is less clear how secondhand smoking influences psychiatric

mellitus, hypertension, dyslipidemia, stroke, subjective health status,

disorders, even though smoking itself is a well-researched risk factor.

and subjective body perception; (4) healthy behavior, including alco-

Some studies suggest that secondhand smoking is related to depres-

hol intake, passive and active smoking status, participation in vigorous

sion, anxiety, and mental health issues,12-14) but a significant relation-

physical exercise, participation in moderate physical exercise, partici-

ship between secondhand smoking and depression or anxiety was not

pation in at least 30 minutes of walking per day, and average hours of

found in a study that examined serum cotinine levels.15)

sleep; and (5) psychiatric health, including depressive symptoms and

There has been a steady decrease in the rate of smoking in South

suicidal ideation.

fluential factor for suicide-related behaviors.

Korea for the last decade, and the South Korean government has introduced a law to regulate smokers in public areas,16) but the country’s

3. Definitions of Variables

rate of smoking among adults, particularly male, is in the highest

To assess type of secondhand smoking exposure, we categorized sub-

across all Organization for Economic Cooperation and Development

jects into two groups according to the place of exposure: in the work-

(OECD) member countries.17) Further, South Korea’s suicide rates are

place or at home. The non-exposed group consisted of participants

the highest among OECD member countries, according to the statis-

who indicated they had not been exposed to secondhand smoking.

tics of OECD health data, irrespective of gender.17) To our knowledge,

The exposed group consisted of participants who reported that they

research has been conducted on secondhand smoking exposure and

were exposed to secondhand smoking for less than 1 hour or 1 hour or

its associated with depressive symptoms in adolescents in South Ko-

more. These data combined into ‘yes’ on the secondhand smoking

rea, but there is a lack of research investigating adults as well as the re-

status in order to analyze the exposed and non-exposed groups by

lationship between secondhand smoking and suicidal ideation.18) In

workplace and home. With regard to depressive symptoms, we catego-

the present study, we aim to identify the influence of secondhand

rized participants into two groups: the depression group consisted of

smoking exposure on depressive symptoms and suicidal ideation by

individuals who reported depressive symptoms for at least 2 continu-

examining data from the Korea National Health and Nutrition Exami-

ous weeks in the last year and non-depression group consisted of

nation Survey (KNHANES).

those who did not report depressive symptoms. For suicidal ideation, we categorized participants into two groups according to their re-

METHODS

sponse to the question about having suicidal ideas during the past

1. Study Participants

or more. Education level was categorized into 4 groups according to fi-

This study utilized data acquired during the first to third years (2010–

nal graduating status: elementary school or less, middle school, high

2012) of the KNHANES V. The KNHANES is conducted by the South

school, and college or more. Marital status was categorized into 3

Korean Ministry of Health and Welfare, and in the fifth period they ex-

groups as never married, divorced or separated, and married. Number

tracted sample survey group and surveyed total of 25,533 participants

of family members, including the participant, was categorized into 2

aged at least 1 year old from 3,800 families annually. Only data from

groups: 4 or less and more than 5. Household income was categorized

adults who had never smoked was included, thus, data was excluded

into 4 groups as low, middle to low, middle to high, and high. Health

from those who were younger than 19 years of age, or when age was

insurance status was categorized into 2 groups, including those with

missing, those who had ever smoked, or had missing data for smoking

health insurance and those without. Employment status was catego-

experience, and those who did not complete all the questions used in

rized into 2 groups: employed and unemployed. We pooled together

this study or indicated “I don’t know.” Consequently, the total number

history of gastric cancer, hepatic cancer, colon cancer, breast cancer,

of participants included in this study was 6,043. The study was ap-

cervical cancer, lung cancer, thyroid cancer and other types of cancer,

proved by the institutional review board of the Samsung Medical Cen-

and then categorized cancer history into 2 groups: no history of cancer

ter (IRB file no. 2015-02-242).

and history of at least 1 cancer. We calculated current morbidity of any

year. Age was categorized into 3 groups, 19 to 44, 45 to 64, and 65 years

of the following diseases, osteoarthritis, rheumatic arthritis, atopic der-

http://dx.doi.org/10.4082/kjfm.2016.37.2.97

Wook Gim, et al.  •  Relationship of Secondhand Smoking with Mental Health

www.kjfm.or.kr  99

matitis, myocardial infarction, chronic kidney disease, diabetes melli-

(16.2%) indicated that they had experienced such thoughts. For edu-

tus, hypertension, dyslipidemia, and stroke, and subsequently classi-

cation level, 1796 non-exposed participants (59.1%) and 2,057 exposed

fied number of chronic disease into 3 groups: 0, 1, 2, or more. Subjec-

participants (68.4%) recorded that they graduated high school or a

tive body perception was categorized into 5 groups, including very

higher level of education. Finally, alcohol intake data identified 149

thin, thin, normal, fat, and very fat, and subjective health was catego-

non-exposed participants (4.9%) and 469 exposed participants (9.3%)

rized into 5 groups, including very good, good, modest, bad, and very

as problem drinkers (Korean version of AUDIT score of 12 or higher).

bad. To analyze alcohol intake, we categorized subjects into 4 groups 11 (normal), 12 to 24 (problem drink), 15 to 25 (alcohol use disorder),

2. The Relationship between Demographic Characteristics and Depressive Symptoms

and 26 or more (alcohol dependency).19) Regular physical activity was

Table 1 presents the results of univariate logistic regressions for the re-

defined as participating in at least one of the following: vigorous activi-

lationships between sociodemographic characteristics and depressive

ty for more than 20 min/d at least 3 days a week, moderate activity for

symptom and suicidal ideation. In Table 2, we display the univariate

more than 30 min/d at least 5 days a week, or walking more than 30

logistic regression models for the relationships between secondhand

min/d at least 5 day a week. Based on these scores, we categorized

smoking exposure and depressive symptoms or suicidal ideation. The

participants into 2 groups: regular physical activity and non-regular

unadjusted OR for depressive symptoms in the secondhand smoking

physical activity. Average hours of sleep were categorized into 2 groups

group compared to the non-exposed group was 1.21 (95% CI, 0.933–

of less than 6 hours and 6 hours or more.

1.480), but this relationship was not statistically significant (P=0.054).

based on alcohol use disorders identification test (AUDIT) scores: 0 to

Other characteristics that were associated with reporting depressive

4. Statistical Analyses

symptoms include the following: 45 years of age or older, female gen-

We analyzed KNHANES V original data from 2010 to 2012 according to

der, unemployed, low household income, low education level, no

the data analysis guidelines for KNHANES, which, given the complex

medical insurance, divorced or separated, 4 or less family members,

sampling design, included weighting certain variables. We calculated

very thin or very fat subjective body perception, bad or very bad sub-

frequencies and percentages according to participant characteristics,

jective health status, presence of chronic disease, history of cancer, al-

and conducted univariate logistic regression analysis and estimated

cohol use disorder or alcohol dependency, and sleeping less than 6 h/

odds ratio (OR) with 95% confidence intervals (CIs) to examine the

night.

variables that may influence depression symptoms and suicidal idereported depression symptoms and suicidal ideation, we conducted a

3. The Relationship between Demographic Characteristics and the Suicidal Ideation

multivariate logistic regression analysis and adjusted for the variables

The unadjusted OR for suicidal ideation in the secondhand smoking

that were significant in the univariate analyses conducted on both the

group compared to the non-exposed group was 1.50 (95% CI, 1.241–

total sample and by gender. Further, we conducted a sub-group analy-

1.804; P