Association of exposure to environmental tobacco smoke in childhood with chronic obstructive pulmonary disease and respiratory symptoms in adults

Authors


Ane Johannessen, Centre for Clinical Research, Haukeland University Hospital, 5021 Bergen, Norway. Email: ane.johannessen@helse-bergen.no

ABSTRACT

Background and objective:  Exposure to environmental tobacco smoke (ETS) is associated with impaired lung function in childhood, which in turn, is associated with chronic obstructive pulmonary disease (COPD) in adulthood. However, little is known regarding the direct association between childhood exposure to ETS and the development of COPD. The main objective of the present study was to examine the associations between childhood ETS exposure and adult COPD and respiratory symptoms.

Methods:  Patients with COPD (n = 433) and control subjects (n = 325) participated in the Bergen COPD Cohort Study during 2006–2009. Participants performed spirometry and answered extensive questionnaires. The risk factors for COPD, morning cough, cough with phlegm, chronic cough and dyspnoea were examined using logistic regression analysis. Analyses were stratified by gender.

Results:  The prevalence of childhood exposure to ETS was 61%. After adjustment, women who were exposed to ETS during childhood had a higher risk of COPD than those who were not exposed: odds ratio 1.9, 95% confidence interval 1.0, 3.7. Other important predictors for COPD and respiratory symptoms among women were occupational dust exposure (COPD), family history of COPD (COPD, all symptoms), current exposure to ETS in the home (morning cough) and education (COPD, dyspnoea). ETS exposure during childhood was associated with respiratory symptoms among males (odds ratios 1.5–1.7). Risk factors for COPD among men were occupational dust exposure, family history of COPD and level of education. Occupational dust exposure and family history of COPD also predicted dyspnoea among males.

Conclusions:  Exposure to ETS during childhood was associated with COPD and respiratory symptoms in adulthood. Although active smoking is still the most important risk factor for COPD, reduction of childhood ETS exposure could contribute to the prevention of COPD and respiratory symptoms.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is characterized by airflow limitation that is not fully reversible.1 It is a disease with substantial consequences in terms of morbidity and mortality.2 The symptoms typically associated with COPD are dyspnoea, cough and sputum production. The risk factors for COPD may be endogenous or exogenous, with personal cigarette smoking being by far the most important.1,3,4 Other environmental risk factors include occupational dusts and chemicals,3,5 smoke from home cooking and heating fuels,3,6 and exposure of adults to environmental tobacco smoke (ETS).7–9

COPD is caused by exposure to harmful agents over a long period of time. Therefore, in order to assess individual risk, lifetime exposures and lung function development should ideally be taken into consideration. A recent study by Stern and co-workers showed that poor lung function in infancy was associated with reduced lung function in early adulthood.10 This is important in the context of COPD because lung function in early adulthood is associated with COPD later in life.11,12 Individuals attain their pulmonary plateau phase in their early 20s, at which time, lung function is at its peak. As lung function declines with age, individuals with the lowest initial lung function are likely to be the first to reach the threshold for diagnosis of COPD.13

Following this line of reasoning, exposures that affect lung function during childhood are likely to be associated with the development of COPD in adulthood. Numerous studies have shown that exposure to ETS during fetal life and childhood is associated with reduced lung function and asthmatic symptoms in children and adolescents.14–19 A few studies have also shown that exposure to ETS in childhood is a predictor for reduced lung function and respiratory symptoms in adults.20–22 However, there is a gap in our knowledge regarding the direct association between childhood ETS exposure and COPD. To our knowledge, only one previous study has addressed this issue: A recent cross-sectional study from China showed an association between self-reported lifetime exposure to ETS and COPD, but not between COPD and childhood or adult exposure to ETS, when analyzed separately.23

Therefore, there is an explicit need for more research regarding the association between childhood exposures and COPD. The aim of the present study was to assess whether exposure to ETS during childhood is associated with the development of COPD and COPD-related symptoms in adults.

METHODS

Subjects

The present study was based on baseline visits for the Bergen COPD Cohort Study 2006–2009, which has been described in detail previously.24 Briefly, a total of 758 subjects (433 COPD patients and 325 control subjects without COPD), aged 40–79 years, were included at baseline. COPD patients were recruited from the outpatient clinic at the Department of Thoracic Medicine, Haukeland University Hospital, Bergen, Norway, and from participants in a previous large observational COPD study at the same centre. Control subjects from a previous randomly sampled population from the same geographical area were invited to participate. COPD patients had a clinical diagnosis of COPD, a smoking history of more than 10 pack-years and a post-bronchodilator forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio of <0.7. COPD patients also had a post-bronchodilator FEV1 of <80% of predicted, based on Norwegian reference values.25 One pack-year was defined as smoking 20 cigarettes per day for one year. Control subjects had a baseline post-bronchodilator FEV1/FVC ≥ 0.7 and FEV1 > 80% of predicted. Informed consent was obtained from each participant prior to the baseline visit, and the study was approved by the Regional Committee of Medical Research Ethics.

All participants were examined by a study physician, who also conducted structured interviews to ascertain medical histories. In addition, the participants completed self-administered questionnaires at baseline, regarding respiratory symptoms and exposures to risk factors for COPD. Spirometry was performed on a Viasys Masterscope (Viasys, Hoechberg, Germany), after inhalation of 0.4 mg of salbutamol. The highest FEV1 value from three attempts was used in the analyses.

Outcome variables

The main outcome variable in the analyses was the presence of COPD (yes/no). In addition, associations between exposure variables and COPD-related symptoms were analyzed: (i) morning cough; (ii) cough with phlegm; (iii) chronic cough; and (iv) dyspnoea grade 2 or worse. These symptoms were defined by affirmative answers to the questions: (i) ‘Do you usually cough in the morning?’; (ii) ‘Do you usually have phlegm when coughing?’; (iii) ‘Do you cough each day for three months or more during a year?’; and (iv) ‘Do you get short of breath when climbing two flights of stairs?’ or ‘Do you get short of breath when walking on level ground at your own pace?’ or ‘Do you get short of breath when at rest?’

Exposure variables

The main exposure variable of interest in the present study was exposure to ETS during childhood. This was assessed by the question ‘For how many years of your childhood would you estimate that you were exposed to environmental smoke?’. If the participants answered ≥1 year to this question, they were categorized as having been exposed to ETS during childhood. Other relevant exposure variables were current exposure to ETS in the home (‘Are you exposed to environmental tobacco smoke at home?’), current exposure to ETS at work (‘Are you exposed to environmental tobacco smoke at work?’), exposure to occupational dust or gas (‘Have you ever had a workplace with much dust or gas in the air?’), and a family history of COPD. The latter was defined as an affirmative answer to at least one of the following questions: ‘Does (did) your father have chronic bronchitis?’, ‘Does (did) your father have emphysema?’, ‘Does (did) your mother have chronic bronchitis?’, ‘Does (did) your mother have emphysema?’, ‘Do (did) any of your brothers or sisters have chronic bronchitis?’, ‘Do (did) any of your brothers or sisters have emphysema?’.

Statistical analyses

Analyses were performed using Stata 10.0 software (StataCorp. Stata Statistical Software: release 10. 2007 TX: StataCorp LP). Differences in baseline characteristics between COPD patients and control subjects were analyzed using unpaired t-tests for continuous variables and chi square tests for categorical variables.

Unadjusted and adjusted logistic regression analyses were performed, to examine the effect of exposure variables on COPD status, and to examine associations between exposure variables and the COPD-related symptoms of morning cough, cough with phlegm, chronic cough and dyspnoea. All analyses were performed separately for men and women, and all multivariate analyses were adjusted for age, smoking status and pack-years smoked. In addition, the multivariate analyses included all exposure variables that were significantly associated with the outcome in univariate analyses. All P values were two-sided, and values <0.05 were considered statistically significant.

RESULTS

Characteristics of the study population

In the present study, 57% of the participants were COPD patients (Table 1). There were slightly more men among the cases than among the controls, and the COPD patients were on average five years older than the control subjects. More control subjects were current smokers, whereas more COPD patients were ex-smokers. The control subjects had smoked fewer pack-years and had a slightly higher mean body mass index than the patients. More control subjects had received a higher education, whereas more COPD patients had completed compulsory education only. Exposures to ETS, both in childhood and adulthood, were slightly more prevalent among COPD patients than control subjects, although the differences were not statistically significant. However, substantially more patients than control subjects had been exposed to dust or gas at work and reported a family history of COPD.

Table 1.  Characteristics of the 758 chronic obstructive pulmonary disease (COPD) patients and control subjects, aged 40–79 years, in the Bergen COPD Cohort Study
 COPD patientsControl subjectsPTotal
  • † 

    Chi-square test for categorical variables, unpaired t-test for continuous variables.

  • ETS, environmental tobacco smoke; SD, standard deviation.

Women, n (%)175 (40)149 (46)0.13324 (43)
Men, n (%)258 (60)176 (54)434 (57)
Age, years, mean (SD)64 (7)59 (10)<0.00161 (9)
Never-smokers, n (%)046 (14)<0.00146 (6)
Ex-smokers, n (%)243 (56)103 (32)346 (46)
Current smokers, n (%)190 (44)176 (54)366 (48)
Pack-years, mean (SD)40 (23)27 (16)<0.00135 (21)
Body mass index, kg/m2, mean (SD)25.3 (5.5)26.4 (3.7)<0.00125.8 (4.8)
Higher education, n (%)41 (10)75 (25)<0.001116 (16)
Intermediate education, n (%)263 (65)185 (62)448 (64)
Lower education, n (%)102 (25)39 (13)141 (20)
Childhood ETS exposure, n (%)274 (63)188 (58)0.13462 (61)
Current ETS exposure at home, n (%)130 (32)87 (29)0.49217 (31)
Current ETS exposure at work, n (%)100 (25)61 (21)0.14161 (23)
Exposure to dust/gas at work, n (%)284 (70)147 (50)<0.001431 (62)
Family history of COPD, n (%)154 (36)51 (16)<0.001205 (27)
Total, n (%)433 (100)325 (100) 758 (100)

Predictors for COPD

For women, childhood exposure to ETS was a significant risk factor for COPD in both unadjusted and adjusted analyses (adjusted odds ratio (OR) 1.9, 95% confidence interval (CI) 1.0, 3.7; Table 2 and Fig. 1). Occupational exposure to dust or gas, and a family history of COPD were also significant predictors (OR 1.9, 95% CI 1.0, 3.4 and OR 3.7, 95% CI 1.9, 7.2, respectively). In addition, increasing age, being an ex-smoker, increasing number of pack-years smoked and lower educational achievement were associated with COPD among women. For men, neither childhood nor adult ETS exposures were important predictors for COPD, after adjustment for potential confounders (Table 3 and Fig. 1). However, occupational exposure to dust or gas and a family history of COPD were associated with an increased risk of COPD (OR 2.2, 95% CI 1.8, 4.0 and OR 3.6, 95% CI 1.9, 7.0, respectively).

Table 2.  Unadjusted and adjusted logistic regression analyses of the risk of chronic obstructive pulmonary disease (COPD) according to childhood exposure to environmental tobacco smoke (ETS), current exposure to ETS at home, current exposure to ETS at work, exposure to dust or gas at work, family history of COPD, BMI and level of education, for 324 women in the Bergen COPD Cohort Study
 UnadjustedAdjusted
OR95% CIOR95% CI
  • The multivariate model was also adjusted for age, smoking status and pack-years smoked, in addition to significant predictors from the univariate analyses. The reference categories were no childhood exposure to ETS, no current exposure to ETS at home, no current exposure to ETS at work, no exposure to dust/gas at work, no family history of COPD and higher education.

  • P < 0.05.

  • BMI, body mass index; CI, confidence interval; OR, odds ratio.

Childhood exposure to ETS1.56*1.00, 2.451.91*1.00, 3.67
Current exposure to ETS at home1.380.84, 2.25
Current exposure to ETS at work2.54*1.25, 5.171.390.57, 3.37
Exposure to dust/gas at work2.02*1.26, 3.231.86*1.01, 3.42
Family history of COPD3.03*1.83, 5.023.71*1.91, 7.22
BMI1.380.84, 2.25
Intermediate education2.54*1.25, 5.171.390.57, 3.37
Lower education2.02*1.26, 3.231.86*1.01, 3.42
Figure 1.

Adjusted odds ratios, with 95% confidence intervals, for the risk of chronic obstructive pulmonary disease (COPD) among 758 subjects in the Bergen COPD Cohort Study. Black lines and circles represent data for men, and grey lines and squares represent data for women. All estimates were also adjusted for age, smoking status, pack-years smoked, body mass index and level of education. The reference categories were no childhood exposure to environmental tobacco smoke (ETS), no current exposure to ETS at home, no current exposure to ETS at work, no occupational exposure to dust or gas and no family history of COPD.

Table 3.  Unadjusted and adjusted logistic regression analyses of the risk of chronic obstructive pulmonary disease (COPD) according to childhood exposure to environmental tobacco smoke (ETS), current exposure to ETS at home, current exposure to ETS at work, exposure to dust or gas at work, family history of COPD, BMI and level of education, for 434 men in the Bergen COPD Cohort Study
 UnadjustedAdjusted
OR95% CIOR95% CI
  • The multivariate model was also adjusted for age, smoking status and pack-years smoked, in addition to significant predictors from the univariate analyses. The reference categories were no childhood exposure to ETS, no current exposure to ETS at home, no current exposure to ETS at work, no exposure to dust/gas at work, no family history of COPD and higher education.

  • P < 0.05.

  • † 

    In the multivariate model, an interaction term between BMI and age was included due to the significant interaction between these two variables in relation to the risk of COPD.

  • BMI, body mass index; CI, confidence interval; OR, odds ratio.

Childhood exposure to ETS1.080.73, 1.591.120.66, 1.89
Current exposure to ETS at home0.950.62, 1.47
Current exposure to ETS at work0.920.60, 1.44
Exposure to dust/gas at work2.86*1.77, 4.632.16*1.81, 3.95
Family history of COPD3.15*1.87, 5.293.61*1.86, 6.98
BMI0.93*0.89, 0.981.580.95, 2.61
Intermediate education1.470.86, 2.522.68*1.10, 6.53
Lower education3.42*1.68, 6.961.030.51, 2.07

Predictors for COPD-related symptoms

For women, childhood ETS exposure was not significantly associated with any of the respiratory symptoms analyzed, although there was a tendency for increased risk (ORs ranging from 1.1 to 1.7, Table 4 and Fig. 2). Increased risk of morning cough was associated with current exposure to ETS at home and a family history of COPD. A family history of COPD was a significant predictor for cough with phlegm, chronic cough and dyspnoea. For women, dyspnoea was also associated with a lower level of educational achievement.

Table 4.  Adjusted logistic regression analyses of the risk of respiratory symptoms (morning cough, cough with phlegm, chronic cough and dyspnoea) according to childhood exposure to environmental tobacco smoke (ETS), current exposure to ETS at home, exposure to dust or gas at work, family history of chronic obstructive pulmonary disease (COPD) and level of education, for 324 women and 434 men in the Bergen COPD Cohort Study
 Morning coughCough with phlegmChronic coughDyspnoea
OR (95% CI)OR (95% CI)OR (95% CI)OR (95% CI)
  • The multivariate models included exposures that were significant in the univariate analyses, and were also adjusted for age, smoking status and pack-years smoked.

  • P < 0.05. Current exposure to ETS at work was not significant in the univariate analyses and was therefore not included in any of the multivariate models. Analyses of the risk of morning cough for women and dyspnoea for men were also adjusted for body mass index, based on the results of the univariate analyses. The reference categories were no childhood exposure to ETS, no current exposure to ETS at home, no exposure to dust at work, no family history of COPD and higher education.

  • CI, confidence interval; OR, odds ratio.

Women    
 Childhood exposure to ETS1.37 (0.8,2.4)1.34 (0.8, 2.3)1.09 (0.6, 2.0)1.65 (0.9, 3.0)
 Current exposure to ETS at home1.65* (1.0, 2.8)1.20 (0.7, 2.0)1.44 (0.8, 2.6)
 Exposure to dust at work1.51 (0.9, 2.7)1.57 (0.9, 2.7)
 Family history of COPD1.92* (1.1, 3.3)1.69* (1.0, 2.9)2.29* (1.3, 4.2)2.91* (1.6, 5.3)
 Intermediate education1.69 (0.8, 3.6)1.58 (0.7, 3.4)1.22 (0.5, 3.0)2.78* (1.1, 6.9)
 Lower education1.09 (0.4, 2.7)1.57 (0.6, 4.0)1.79 (0.6, 5.0)3.03* (1.1, 8.7)
Men    
 Childhood exposure to ETS1.68* (1.1, 2.6)1.55* (1.0, 2.4)1.61* (1.0, 2.5)1.52 (0.9, 2.5)
 Current exposure to ETS at home1.51 (0.9, 2.4)
 Exposure to dust at work2.17* (1.2, 3.9)
 Family history of COPD1.57 (0.9, 2.6)2.43* (1.4, 4.3)
 Intermediate education1.81 (0.9, 3.6)
 Lower education2.13 (0.9, 4.8)
Figure 2.

Adjusted odds ratios, with 95% confidence intervals, for childhood environmental tobacco smoke (ETS) exposure on chronic obstructive pulmonary disease (COPD)-related symptoms among 758 subjects in the Bergen COPD Cohort Study. Black lines and circles represent data for men, and grey lines and squares represent data for women. All estimates were also adjusted for age, smoking status and pack-years smoked. The reference category was no childhood exposure to ETS.

With regard to adjustment variables, active smoking was associated with all respiratory symptoms among women, and in addition the risk of dyspnoea was associated with increasing age.

For men, childhood ETS exposure was associated with increased risk for all symptoms, with OR (95% CI) ranging from 1.5 (0.9, 2.5) for dyspnoea to 1.7 (1.1, 2.6) for morning cough (Table 4 and Fig. 2). In addition, dyspnoea was associated with occupational exposure to dust or gas, and with a family history of COPD. Among men, the risk for all symptoms increased with active smoking and with increasing age.

DISCUSSION

The present study showed that women who had been exposed to ETS during childhood had a 1.9-fold greater risk of COPD than women who had not been exposed. Men who had been exposed to ETS during childhood had a 1.5- to 1.7-fold greater risk of COPD-related symptoms than men who had not been exposed. In this study population, exposure to ETS during childhood was overall a much stronger risk factor than exposure to ETS in adulthood.

Although it is widely acknowledged that active smoking is by far the most important independent risk factor for the development of COPD and COPD-related symptoms, childhood exposures have recently been shown to be important determinants of adult lung function. Recent results from the European Community Respiratory Health Survey showed that adult lung function and susceptibility to COPD were partly determined early in life, and that maternal smoking significantly affected FEV1 in adulthood.26 The results from the present study are in accordance with these observations. Children who are exposed to cigarette smoke in their homes are likely to have lower lung function at their peak than non-exposed children, and lung function in early adulthood is associated with COPD later in life.11,12 In the present analyses, both the magnitude and precision of the risk estimates were modest. However, this is to be expected for a risk factor study of a complex chronic disease such as COPD, particularly when looking at childhood exposures and a long latency of effect.

Recall bias is a general problem when seeking exposure information from far back in time, although the Nurses Health Study recently showed that daughters' reports of mothers' smoking habits were quite reliable.27 Patients in general tend to under-report harmful exposures that are within their control (for example active smoking) and over report harmful exposures that are beyond their control (for example ETS exposure).28 However, in this study population, the prevalence of reported childhood ETS exposure was almost evenly distributed among COPD patients and control subjects (63% and 58%, respectively).

There is always a possibility that active smoking at some point may confound the effect of ETS exposure. In the present study, all analyses were adjusted for pack-years smoked and current smoking status. In this way, an attempt was made to reduce confounding of the effect of ETS exposure by active smoking. A future study with an adequate sample size to enable stratification by active smoking status and active smoking history, without loss of statistical power, may shed light on this complex issue.

The present study showed that whereas childhood ETS exposure was associated with COPD in women, it was associated with respiratory symptoms in men. However, a recent report from the European Community Respiratory Health Survey showed that childhood disadvantages, such as maternal smoking, affected the risk of COPD in the same way for both men and women.26 The risk estimates reported from the European Community Respiratory Health Survey were comparable with those calculated in the present study—one childhood disadvantage yielded ORs for COPD of 1.6 for women and 1.7 for men. Furthermore, in the present study, the observed trends for the effects of childhood ETS exposure on COPD and symptoms were similar for men and women, although the effect on COPD was not statistically significant for men and the effect on symptoms was not statistically significant for women.

To our knowledge, this is the first study that has explicitly examined the association between ETS exposure during childhood and the risk of COPD in adults, except for the recent Chinese Guangzhou Biobank Cohort Study.23 The Guangzhou study did not find an association between childhood ETS exposure and COPD. This contrasts with the present analyses, which showed that women exposed to ETS in childhood had almost twice the risk for COPD, as compared with non-exposed women. There may be several reasons for the discrepancy between these results and the Chinese results.

First, according to the World Health Organization fact sheet on smoking, the prevalence smoking in China is 67% among men and only 4% among women. According to Statistics Norway, 19% of Norwegian men and women were smokers in 2010. In the Guangzhou study, the vast majority of the study population was female (88%). Traditionally, girls have spent a larger proportion of their childhood with their mothers than with their fathers. As the Chinese study population consisted mainly of women, and as Chinese women smoke much less than Norwegian women, the exposure level of the Chinese study population was probably low compared with that of the present population. In addition, the skewed gender distribution of the Guangzhou study population may have resulted in that study being vulnerable to selection bias.

Secondly, the Guangzhou study population consisted of never-smokers, whereas the present study population consisted mainly of ever-smokers. The Scottish Renfrew-Paisley (Midspan) Family Study showed that childhood ETS exposure affected lung function differentially in never-smokers and ever-smokers.22

Other studies that have examined childhood ETS exposure and adult respiratory outcomes, including asthma and respiratory symptoms, have shown similar findings to the present study. Skorge and co-workers found that 17% of the incidence of adult asthma could be attributed to maternal smoking.20 The same study also showed that there were significant associations between ETS exposure during childhood and respiratory symptoms in adulthood. In addition, a recent report on the Multi-Ethnic Study of Atherosclerosis (MESA) indicated that childhood exposure to ETS was associated with emphysema among adult non-smokers.29

In summary, exposure to ETS during childhood was associated with approximately a twofold greater risk of COPD in women, and was a significant risk factor for COPD-related symptoms in men. These results suggest that the burden of COPD could be reduced if children were not exposed to cigarette smoke. In addition to the known short-term effects of ETS exposure, there is increasing evidence of long term effects, which need to be investigated further through longitudinal population surveys. Further studies are also needed to elucidate the possibility that there are different phenotypes for men and women, with respect to their COPD risk profile.

ACKNOWLEDGEMENTS

This work was supported by the Norwegian Research Council and the Western Norway Regional Health Authority, and work on this paper was supported by a postdoctoral grant from the Western Norway Regional Health Authority. The authors thank statistician Roy Miodini Nilsen for valuable methodological advice, and the staff at the Centre for Clinical Research at Haukeland University Hospital for useful discussions.

Ancillary