|Home | About | Journals | Submit | Contact Us | Français|
To evaluate the association between secondhand smoke (SHS) exposure and depression. Tobacco smoking and depression are strongly associated, but the possible effects of SHS have not been evaluated.
The 2005 to 2006 National Health and Nutrition Examination Survey (NHANES) is a cross-sectional sample of the noninstitutionalized civilian U.S. population. SHS exposure was measured in adults aged ≥20 years by serum cotinine and depressive symptoms by the Patient Health Questionnaire. Zero-inflated Poisson regression analyses were completed with adjustment for survey design and potential confounders.
Serum cotinine-documented SHS exposure was positively associated with depressive symptoms in never-smokers, even after adjustment for age, race/ethnicity, gender, education, alcohol consumption, and medical comorbidities. The association between SHS exposure and depressive symptoms did not vary by gender, nor was there any association between SHS smoke exposure and depressive symptoms in former smokers.
Findings from the present study suggest that SHS exposure is positively associated with depressive symptoms in never-smokers and highlight the need for further research to establish the mechanisms of association.
The U.S. Surgeon General has concluded that secondhand smoke (SHS) causes premature death and disease in persons who do not smoke and that no scientific evidence establishes a risk-free level of exposure (1). Currently, only 41.2% of the U.S. population is covered by comprehensive clean indoor air legislation, which prohibits smoking in all workplaces, bars, and restaurants (2). In the National Comorbidity Survey-Replication Study (3), the prevalence of lifetime major depressive disorder was 16.2% (32.6–35.1 million U.S. adults), and it was 6.6% within the last 12 months (13.1–14.2 million U.S. adults). Depression has been related to an increased risk of mortality (4,5) and morbidity (6). The burden of depression worldwide is a major public health issue. In countries with high per capita income, depression is the third leading cause of disability adjusted life years (8–41 million years) (7).
Although smokers are clearly at greater risk for depression (8–12), the nature of the relationship between smoking and depression is uncertain. Smokers may smoke because they are depressed and therefore “self-medicate” (11), or depression could be the result of smoking (13). Alternatively, a third factor, such as a shared genetic predisposition, may explain this relationship (8). A recent study by Nakata and colleagues (14) found that the self-report of SHS exposure was related to depression among a group of Japanese never-smoking workers. Although the study by Nakata et al. (14) was the first to find an association between SHS and depression, it was limited by the self-report measure of SHS exposure. Arheart and colleagues (15) recently found that self-report of SHS exposure can be an imprecise measure of biologically confirmed SHS exposure.
The purpose of the present study was to explore the possible association between SHS exposure and depressive symptoms in a population-based sample of the U.S. adult population, using an objective, biological measure of SHS exposure. Growing evidence suggests that SHS exposure is associated with chronic diseases, such as cardiovascular disease (16), respiratory disease (17), diabetes mellitus (18), hypertension (19), thyroid disease (20), and cancers (21). All of these chronic conditions are also associated with depression (22–26). Thus, chronic diseases may explain the association between SHS exposure and depression. Further, other psychological or biological mechanisms may explain the association between SHS exposure and depression. SHS exposure may be a proxy to stressful living and working environments, and stress has been associated with depression (27,28). In addition, perhaps similar neurobiological mechanisms observed in firsthand smokers may be observed in persons who have SHS exposure. That is, chronic exposure to cigarette smoke may lead to lower levels of dopamine and γ-aminobutyric acid (GABA), which are associated with mood disorders (29,30). Women are also at greater risk for depression than men (3), and the association between SHS exposure and depression may be more apparent for women than men. Therefore, we hypothesized that SHS exposure would be positively associated with depressive symptoms even after adjustment for potential confounders and that the strength of this association differs in men versus women.
The National Health and Nutrition Examination Survey (NHANES) (http://www.cdc.gov/nchs/nhanes.htm), a household survey combined with a medical examination, is conducted every year by the National Center for Health Statistics in the Centers for Disease Control and Prevention. The survey is designed to be representative of the noninstitutionalized civilian U.S. population. Complete data from nonsmokers aged ≥20 years were obtained from the 2005 to 2006 NHANES (n = 2965). The response rate for the interview component was 74.4%, and the response rate for the clinical examination component was 71.5%. This study was approved by the University of Miami Institutional Review Board for human subjects research.
For the 2005 to 2006 NHANES, the detectable limit for serum cotinine (a metabolite of nicotine) was 0.011 ng/mL. As described previously (31,32), the value for data below the detectable limits was the limit divided by the square root of 2. SHS exposure was included in the analyses as the natural log transformation of serum cotinine.
The Patient Health Questionnaire-9 (33) was used to measure depressive symptoms. Participants responded to nine items measuring depressive symptoms in the last 2 weeks by choosing either: Not at all (0); Several days (1); More than half the days (2); or Nearly every day (3). A summed score from 0 to 27 was used to classify “Depressive Symptoms” (33). Participants were excluded if they were missing data on ≥3 of the nine items (n = 3). If participants were missing one or two items, scores were imputed substituting the mean of an individual’s nonmissing items for the missing items.
Covariates were chosen a priori on the basis of literature review. Covariates were defined as potential confounders if they were associated with both the predictor and outcome and could possibly account for the association between SHS and depressive symptoms. Age, race/ethnicity, gender, education, alcohol consumption, and medical comorbidities were measured by participant self-report. The medical comorbidities in this study were cardiovascular disease (i.e., either angina, congestive heart failure, coronary heart disease, heart attack, or stroke), respiratory disease (i.e., either asthma, chronic bronchitis, or emphysema), diabetes mellitus, hypertension, thyroid disease, and history of cancer. Alcohol consumption was measured as the average number of drinks per day in the last year in a subset of our analytic sample aged 20 years to 59 years (n = 1320).
Participants were asked if they had smoked ≥100 cigarettes in their lifetime. Those responding “yes” were asked if they currently smoked cigarettes. Self-identified current smokers were excluded in the analysis (n = 943). We also considered those with serum cotinine values of >3 ng/mL as current smokers (n = 124) (34). In addition, participants with incomplete covariate information (age, race/ethnicity, gender, education, and comorbidities) were excluded from the analysis (n = 2). Stratified analyses were performed in “never-smokers” (defined as those who reported not smoking ≥100 cigarettes in their lifetime [n = 2026]) and “former smokers” (defined as those who reported smoking ≥100 cigarettes in their lifetime, but did not currently smoke and who had cotinine values of ≤3 ng/mL [n = 938]).
STATA version 10.0 was used to analyze a continuous measure of depressive symptoms with SHS exposure as the main predictor when adjusting for potential confounders. Because the distribution of depressive symptoms was positively skewed and had an excess of zeros (37.7%), linear regression analyses were not appropriate. Furthermore, the summed score of depressive symptoms is similar to count data (35). Other epidemiological studies have used the same measure of depressive symptoms as a count variable (36,37). We applied both zero-inflated Poisson and zero-inflated negative binomial regressions. The likelihood ratio test was used to determine which method best fit the data. In the present study, zero-inflated Poisson regression fit the data better than zero-inflated negative binomial regression. We also took into account sample weights and the effects of the complex sampling design of the NHANES. Model 1 was adjusted for age, gender, race/ethnicity, and education. Model 2 was further adjusted for medical comorbidities. Finally, subgroup analyses among persons aged 20 years to 59 years were conducted, adjusting for age, gender, race/ethnicity, education, and alcohol consumption to evaluate the significance of this latter covariate as a possible confounder in the relationship between SHS and depressive symptoms.
Table 1 presents the sample characteristics. The mean of serum cotinine (log) was −3.15 (standard error [SE] = 0.05), depression was 2.31 (SE = 0.08), and age was 48.60 (SE = 0.95). Most of the sample was non-Hispanic white (72.9%, SE = 3.07), 12.4% was Hispanic (SE = 1.55), 10.1% was non-Hispanic black (SE = 1.97), and 4.6% was non-Hispanic other (SE = 0.60); 43.6% was male (SE = 1.00) and 56.4% was female (SE = 1.00); 63.6% had < high school education (SE = 1.81), 22.2% had the equivalent of high school education (SE = 1.13), and 14.2% had > high school education (SE = .53).
The interaction between gender and SHS exposure was not significant in never-smokers and in former smokers, when adjusting for age, gender, race/ethnicity, and education (t = 1.13, p= .27 and t = −0.21, p = .83). The association between SHS exposure and depressive symptoms among former smokers was not significant when adjusting for age, race/ethnicity, gender, and education (β = −0.01, SE = 0.04, p = .68). Among never-smokers, SHS exposure was positively associated with depression after adjusting for age, gender, race/ethnicity, and education (β = 0.09, SE = 0.03, p = .02) (Table 2). SHS exposure remained positively associated with depression after further adjustment for cardiovascular disease, respiratory disease, diabetes mellitus, hypertension, thyroid disease, and cancers (β = 0.09, SE = 0.03, p = .03). SHS exposure was positively associated with depression in subgroup analyses (persons aged 20–59 years) for age, gender, race/ethnicity, education, and alcohol consumption (β = 0.10, SE = 0.03, p = .01).
Sixty-five percent of the working U.S. population is potentially exposed to SHS in the workplace (21), and >35 million U.S. adults may suffer at least one lifetime episode of major depression (3). Thus, our finding that SHS exposure is associated with the level of depressive symptoms among never-smokers is of clear public health significance. The association between SHS exposure and depressive symptoms was not explained by demographics, alcohol consumption, or the presence of chronic conditions.
Several possible explanations may account for our finding that SHS exposure is associated with the level of depressive symptoms in never-smokers, even after controlling for comorbidities and socioeconomic status. SHS exposure may be a proxy for stressful home and occupational environments, given that stress has been related to depression (27,28). Similar to smokers, prolonged SHS exposure among never-smokers may lead to lower levels of dopamine and lower levels of GABA, which have been related to an increased risk for depression (29,30). Recent studies have shown that the dopamine system genotype variability may explain a significant proportion of the interindividual variability in smoking-induced dopamine release, and indicate that smoking-induced dopamine release has a genetic predisposition (38–40). Former smokers may have been less subject to nicotine dependence when they smoked (relative to those who continue to smoke) and exhibit homeostatic levels of dopamine and GABA. In addition, by virtue of their current status, former smokers were likely more motivated to quit smoking, consistent with a desire to live a healthy lifestyle, and less likely to be depressed by self-selection (41,42). Of interest, one study found a significantly decreased risk of depression among former smokers, who had quit over 10 years prior compared with an increased risk for nonsmokers and for smokers (28). Therefore, it may be possible that the particular characteristics of the individual (such as their nicotine dependence, personality, and genetic make-up) may play a protective role, as observed in the former smokers in our study.
Findings from this study must be interpreted with caution. If first and secondhand smokers are more likely to suffer from depression because of their exposure to cigarettes and the same psychological or biological mechanisms are shared, then it is likely that the temporal association between smoking and depression is that smoking precedes the onset of depression. For firsthand smokers, it is still possible for them to “self-medicate” and the association between firsthand smoking and depression may be bidirectional. However, it is highly unlikely that never-smokers exposed to SHS are self-medicating. This is only the second study to find an association between SHS exposure and depression (14) and, thus, future studies are needed to establish consistent results. The mechanisms of association are also important to establish causality, and further research is needed to examine whether first- and secondhand smokers share similar mechanisms of association.
An advantage of the present analysis is that it addressed the hypothesis that SHS exposure among nonsmokers is associated with increased levels of depressive symptoms, using a population-based U.S. sample. Furthermore, SHS exposure was measured, using an objective biological measure, i.e., serum cotinine, and depressive symptoms was measured by the well-validated Patient Health Questionnaire (33). However, this study is cross-sectional, so temporal associations cannot be established. A small number of intermittent or light smokers, who denied tobacco use, were possibly retained in our analysis, despite excluding participants with cotinine values of >3 ng/mL. However, if these intermittent smokers are present, then they would have not unduly biased our results. The exact mechanism by which SHS exposure is associated with depression is not clear. Research employing longitudinal study designs, which include serial assessment of SHS exposure, is needed to examine causality between SHS and depression. Finally, detailed information on the source(s) of exposure (e.g., at home, work, and/or elsewhere) and a past history of depression are also unknown.
Depression is a major cause of morbidity (6,7,12), mortality (4,5), and reduced worker productivity (43). Studies have shown that banning smoking in workplaces and other public settings leads to immediate reductions in hospitalizations and disease-specific symptoms (44–47). However, no previous studies have looked specifically at the impact smoking bans might have on depression and levels of depressive symptomatology. In addition, these data suggest that interventions designed to eliminate smoking in the home are also needed. Modest financial incentives to prevent smoking at home already exist, most directly through owner and renter insurance rates, which can be higher for smokers (48). However, public health campaigns and similar interventions that promote greater awareness of the negative consequences associated with both smoking and SHS exposure are also warranted to encourage individuals to stop smoking inside the home. Thus, policy makers and employers have an important role in developing programs to ban and/or prevent SHS that can be applied in a variety of settings including the workplace, the community, and individual homes.
This work was funded, in part, by grants from the Flight Attendant Medical Research Institute (Dr. Lee, Principal Investigator) and Grant R01 OH03915 from the National Institute of Occupational Safety and Health (Dr. Fleming, Principal Investigator).
This study was completed while Frank C. Bandiera was a doctoral student in epidemiology at the University of Miami. Mr. Bandiera is a recipient of an individual pre-doctoral grant from the National Institute of Mental Health to study the association between asthma and mental health outcomes and receive advanced training in statistical methods, genetic epidemiology, behavioral medicine, and psychoneuroimmunology (1F31MH084567).
Because the NHANES is a publicly available database, all authors had access to the data.
Frank C. Bandiera, MPH, conceived the study, conducted preliminary statistical analyses, and drafted an initial version of the manuscript. David J. Lee, PhD, Lora Fleming, MD, PhD, and Noella Dietz, PhD, assisted with the conceptualization of the study and provided expertise on SHS exposure and health outcomes. They also reviewed and added sections to the manuscript. Kristopher Arheart, EdD, and William G. LeBlanc, PhD, served as statisticians for the study, providing statistical expertise, data analysis, and assisted with revision of the manuscript. Evelyn P. Davila, MPH, Alberto J. Caban-Martinez, Kathryn McCollister, PhD, MPH, Berrin Serdar, MD, PhD, and John E. Lewis, PhD, added sections and assisted in reviewing versions of the manuscript.