Second-hand smoke exposure in Canada : Prevalence , risk factors , and association with respiratory and cardiovascular diseases

1Department of Medicine, Queen’s University; 2Clinical Research Centre, Kingston General Hospital, Kingston, Ontario Correspondence and reprints: Dr Nicholas Vozoris, Division of Respirology, Department of Medicine, University of Toronto, Medical Sciences Building, Room 6263, 1 King’s College Circle, Toronto, Ontario M5S 1A8. Telephone 416-978-1562, fax 416-971-2112, e-mail nick.vozoris@utoronto.ca N Vozoris, MD Lougheed. Second-hand smoke exposure in Canada: Prevalence, risk factors, and association with respiratory and cardiovascular diseases. Can Respir J 2008;15(5):263-269.

There are limited data on the prevalence of second-hand smoke exposure among adults in Canada (44) and internationally (18,(45)(46)(47)(48)(49)(50), and which sociodemographic groups are at greatest risk for exposure (7,45,46,(51)(52)(53).Using data from the 2003 Canadian Community Health Survey (CCHS), Pérez (44) reported a national prevalence of second-hand smoke exposure in Canada of 33%.This prevalence estimate and the reported associations with limited sociodemographic variables are potentially problematic, because the 2003 CCHS did not take into consideration second-hand smoke exposure in private places such as friends' or relatives' homes, and did not reliably ascertain personal exposure in homes or workplaces.To date, associations with chronic respiratory disease and cardiovascular disease in Canada have been minimally explored (54).
The objectives of the present study were to estimate the prevalence of second-hand smoke exposure in Canada while controlling for previous smoking status, to identify sociodemographic characteristics of individuals most likely to lead to exposure, and to examine the relationship between secondhand smoke exposure and disease (particularly respiratory diseases) in never-and ex-smokers, via an analysis of data from the 2000/2001 CCHS.

Data
The present study was conducted using data from the public use microdata file for the 2000/2001 CCHS Cycle 1.1.The CCHS is an ongoing survey conducted by Statistics Canada that collects cross-sectional, self-reported, sociodemographic and health data on Canadians aged 12 years and older.A detailed description of the survey design and methodology has been previously published (44,(55)(56)(57).The response rate was 84.7% (44), yielding a total sample of 130,880 respondents.
Second-hand smoke exposure Cycle 1.1 of the CCHS was selected for analysis because of the strength of its assessment question for second-hand smoke exposure.Unlike the subsequent cycle, the 2000/2001 CCHS Cycle 1.1 included the following question on second-hand smoke exposure that was asked only of current nonsmokers: "In the past month, were you exposed to second-hand smoke on most days?"The strengths of this question are that it directly asks about regular, recent second-hand smoke exposure, it is not limited by location of exposure, and it attempts to capture significant exposure history by asking whether it occurs "on most days".One limitation of this survey question is that it limits exposure to only the preceding month.Those who responded affirmatively to the question were classified as being exposed to second-hand smoke, while those who responded negatively were classified as not being exposed to second-hand smoke.
Because ex-smokers would a priori be expected to have poorer health than never-smokers, all of the analyses were stratified by previous smoking status, as defined by the response to the following survey question: "Have you ever smoked a whole cigarette?"Those who responded affirmatively to this question were classified as ex-smokers, while those who responded negatively were classified as never-smokers.Although a conventional definition of never-smoking is smoking 100 cigarettes or fewer in one's life, a strict definition of never-smoking (ie, never having smoked a whole cigarette) was specifically selected.The consumption of 100 cigarettes or fewer to denote a never-smoker is an arbitrary cut-off number, and there is no evidence to date to suggest what threshold of cigarette smoking is required to cause disease (42).
The survey also asked questions about places of secondhand smoke exposure among those exposed: "In the past month, were you exposed to second-hand smoke at home; in a car or other vehicle; in public places; while visiting friends or relatives?"Second-hand smoke exposure at work was explored by the following survey question: "At your place of work, what are the restrictions on smoking?" Possible responses included the following: "restricted completely", "allowed in designated areas", "restricted only in certain places", "not restricted at all", "don't know", refused or not stated.Those who responded that smoking at work was completely restricted were classified as not having been exposed to second-hand smoke at their workplace, while all other respondents were classified as having been exposed to second-hand smoke at their workplace.

Sociodemographic and health variables
All sociodemographic and health variables were based on selfreport.The following sociodemographic variables were examined: sex, province, age, race, immigration status, education level, income source, income level and whether children younger than 12 years of age resided in the household.Sex, province and age were chosen because they have been previously shown to be associated with second-hand smoke exposure (44).Education level, income source, income level, race and immigration status were chosen to examine whether exposure to second-hand smoke was associated with indexes of socioeconomic status.The presence of children younger than 12 years of age in the household was selected to determine whether it was a protective factor with respect to second-hand smoke exposure risk.
The following chronic health conditions were examined: asthma, chronic bronchitis, emphysema, heart disease and hypertension.These health conditions were chosen because previous epidemiological research have found them to be associated with second-hand smoke exposure .Respondents were classified as having a particular condition only if they reported that it lasted or was expected to last for six months or longer, and that it had been diagnosed by a health care professional.

Statistical methods
All analyses were performed using SAS version 9.1 (SAS Institute, USA).Cross-tabulations were calculated to estimate the prevalence of second-hand smoke exposure.Crosstabulations and multiple logistic regression were used to examine the association between second-hand smoke exposure and the selected sociodemographic and health variables.Unadjusted odds of an individual exposed to second-hand smoke reporting the selected health conditions were calculated.Sex, province, age group, immigration status, education level, income source and level, and the presence of children younger than 12 years of age in the household were then included in a single regression model to control for their potentially confounding effects on the odds of reporting the selected health conditions, because all of these variables were found to be independent risk factors for second-hand smoke exposure.
The CCHS uses a complex sampling design, involving stratification and multistage clustering techniques.To account for the effects of stratification, all results were weighted using the sample weights provided in the public use microdata file.All sample weights were rescaled before each analysis by dividing the original weight by the average weight of respondents included in the specific analysis, according to Statistics Canada guidelines (55).To account for the effects of clustering, all CIs were calculated using bootstrap resampling techniques, with a set of 500 bootstrap weights created by Statistics Canada to reflect the sampling design used in the survey.

RESULTS
Of the 130,880 survey respondents, 96,717 (73.9%) were current nonsmokers.This sample of current nonsmokers included 48,540 never-smokers and 48,177 ex-smokers.Approximately 25% of never-smokers and 30% of ex-smokers self-reported recent second-hand smoke exposure, representing a total of 5.3 million Canadians (Table 1).
The most frequently self-reported sites of second-hand smoke exposure were public places and the workplace (60% to 75%), but exposure was commonly reported in private locations such as homes of respondents, homes of relatives or friends, and vehicles (Table 2).Sex, province, age, immigration status, education level, income source, income level and whether young children resided in the household were all independent risk factors for self-reported second-hand smoke exposure regardless of previous smoking status (Tables 3 and 4).
Among never-smokers, those exposed to second-hand smoke were more likely to report asthma and less likely to report heart disease and hypertension.There were no associations observed with chronic bronchitis and emphysema.After adjusting for potential confounders, however, second-hand smoke exposure was positively associated with chronic bronchitis, but there was now no association with heart disease and hypertension.The remainder of the disease associations remained unchanged (Table 5).Among ex-smokers, those exposed to second-hand smoke were more likely to report asthma and chronic bronchitis, less likely to report heart disease and hypertension, and there were no associations with emphysema.After adjusting for potential confounders, however, second-hand smoke exposure was positively associated with hypertension and there was now no association with heart disease.The remainder of the disease associations remained unchanged (Table 6).

DISCUSSION
The prevalence of self-reported recent second-hand smoke exposure on most days among current nonsmokers in Canada in 2000/2001 was high (27.7%).The most frequently selfreported sites of second-hand smoke exposure were public places and the workplace, but exposure was commonly reported in private locations as well.The following factors were identified as risk factors for second-hand smoke exposure: men; residences in Quebec, Atlantic Canada and the Territories; younger ages; nonimmigrant status; low education and income levels; social assistance receipt; and households without children younger than 12 years of age.After controlling for potential confounders, both never-and ex-smokers exposed to second-hand smoke had significantly higher odds of selfreporting asthma (20% to 30%) and chronic bronchitis (50%) than those not exposed.Among ex-smokers, those exposed to second-hand smoke also had significantly greater odds of selfreporting hypertension (20%).No associations were observed between second-hand smoke exposure and emphysema or heart disease.
The present study has several limitations that must be acknowledged.The cross-sectional nature of the data does not support causal inferences.Findings were based entirely on selfreported measures, thereby potentially introducing recall and social desirability biases.The fact that the survey question limited exposure to within the preceding month and that we are trying to relate exposure to chronic disease outcomes is a limitation of the analysis.Our method of stratification by previous smoking status may result in some misclassification, potentially underestimating the risk of disease among never-smokers if a significant proportion of individuals with trivial smoking history were classified as ex-smokers.The risk of disease among those exposed to second-hand smoke may also be underestimated because second-hand smoke exposure was limited to 'most days during the past month'.Subjects exposed to second-hand smoke in one-half or fewer days per month would have been classified as 'not exposed', although exposure to second-hand smoke could     ; † Refers to the highest level of education that the respondents attained; ‡ Refers to the main source of household income; § Income level was defined as a five-level categorical variable describing income adequacy, and was based on information about gross total household income in the past 12 months and family size still have been substantial during those times.The fact that the odds of reporting the selected chronic health conditions are generally consistent between never-smokers and ex-smokers may be a reflection of misclassification.Longitudinal data, including a more objective and comprehensive measure of second-hand smoke exposure, would be required to confirm the associations reported here.
Given that the CCHS Cycle 2.1 did not reliably ascertain personal second-hand smoke exposure in homes, and that selfreporting was not time-limited (Cycle 1.1 limited exposure to 'the past month'), it is not surprising that Pérez (44) found a higher prevalence estimate (33%) than that reported in the present study.Our overall prevalence estimate for Canada is similar to or lower than that reported for many European countries (49)(50), but higher than that reported for the United States (48)(49)(50).Our prevalence estimate is likely lower than the true prevalence of second-hand smoke exposure because survey-based exposure assessments have been reported to significantly underestimate second-hand smoke exposure when compared with more objective measurements (6,12).The difference in assessment of second-hand smoke exposure between the survey cycles makes it difficult to draw conclusions about changes over time.
Like Pérez (44), we found that second-hand smoke exposure risk is increased for men, younger ages and residents of Quebec, but we have identified that indexes of low socioeconomic status, such as low income, low education and social assistance receipt, are also risk factors.Studies outside Canada have made similar observations between some indexes of socioeconomic status and second-hand smoke exposure (7,46,(51)(52)(53).The fact that the most frequent settings of secondhand smoke exposure were public places and workplaces supports the use of public health legislation to help limit second-hand smoke exposure.At the same time, a significant proportion of second-hand smoke exposure continues to occur in settings that fall outside of the realm of conventional public health regulation (eg, homes and personal vehicles).
Although firm conclusions on potential causal associations between second-hand smoke exposure and adult respiratory diseases have not yet been reached by scientific experts (36)(37)(38)(39)(40)(41)(42), our results contribute to a growing body of literature indicating that second-hand smoke exposure is associated with adult asthma (1-10) and COPD (1)(2)9,(11)(12)(13)(14)(15)(16)(17)(18).In a sample of 4197 never-smoking Swiss adults, self-reported second-hand smoke exposure was associated with greater odds of chronic bronchitis symptoms (OR=1.65)and physician-diagnosed asthma (OR=1.39)(1).A recent cohort study of approximately 3000 never-smoking Italian women found that those exposed to second-hand smoke were 2.3 times more likely to have asthma, chronic bronchitis or emphysema (9).Eisner et al have reported that second-hand smoke exposure is directly related in a dose-dependent fashion to increasing asthma (6) and COPD (12) severity, as indicated by respiratory symptoms, corticosteroid and other medication use, hospital admissions and home oxygen use.
The present study is one of the few studies (24) to have reported positive associations between second-hand smoke exposure and hypertension.The absence of significant positive associations in the present study between second-hand smoke exposure and heart disease adds to the controversy on the topic.A meta-analysis of 18 published studies (28) involving more than 650,000 individuals reported a dose-response relationship with coronary artery disease, with RRs of 1.23 for nonsmokers exposed to one to 19 cigarettes/day and 1.31 for nonsmokers exposed to 20 or more cigarettes/day, compared with nonexposed nonsmokers.However, a more recent meta-analysis (35) concluded that when all relevant studies are included and appropriately combined, the risk of death from coronary artery   disease in never-smokers exposed to second-hand smoke is roughly 5% and had previously been overestimated.A cohort study (25) though subsequent to this meta-analysis found an excess risk of cardiovascular mortality of approximately 40% among exposed never-smokers with no baseline cardiovascular disease.Differences in study population, sample size, measurement of second-hand smoke exposure and health outcome, and the control of potentially confounding variables may contribute to discrepancy in results among studies.

CONCLUSIONS
Approximately 28% of Canadian current nonsmokers reported regular, recent second-hand smoke exposure in 2000/2001.Risk factors for second-hand smoke exposure were identified.Exposure to second-hand smoke was associated with chronic respiratory and cardiovascular diseases in a very large sample of never-and ex-smokers using a very strict definition of previous smoking.Although the cross-sectional nature of the present data precludes causal inference, given the high prevalence of second-hand smoke exposure in Canada, the associations identified merit further study to help clarify the role of second-hand smoke exposure in healthy individuals, and in those with already established respiratory and cardiovascular disease.Hundreds of municipalities across Canada have already established regulations that limit smoking in public places and the workplace, although they vary in scope and enforceability (44).The Ontario provincial government recently introduced the Smoke-Free Ontario Act, which, as of May 31, 2006, comprehensively banned smoking in all workplaces and public places throughout the province, with heavy financial penalties for offenders (58).
Other governments in Canada and around the world may need to follow Ontario's lead, and even expand public health efforts to target presently unregulated exposure sites such as homes and personal vehicles, to help protect the health of their residents if the potential deleterious respiratory and other health effects of second-hand smoke exposure are confirmed.
Second-hand smoke exposure and health Can Respir J Vol 15 No 5 July/August 2008 267

TABLE 3
*Adjusted for other variables listed in table; † Refers to the highest level of education that the respondents attained; ‡ Refers to the main source of household income; § Income level was defined as a five-level categorical variable describing income adequacy, and was based on information about gross total household income in the past 12 months and family size

TABLE 4
*Adjusted for other variables listed in table

TABLE 5
*Numbers represent respondents who reported the specific chronic condition.Percentages represent the per cent of individuals by SHS exposure status who reported the specific chronic conditions; † Adjusted for sex, age group, province, education level, income source, income adequacy, immigration status and presence of children younger than 12 years of age in household

TABLE 6
*Numbers represent respondents who reported the specific chronic condition.Percentages represent the per cent of individuals by SHS exposure status who reported the specific chronic conditions; † Adjusted for sex, age group, province, education level, income source, income adequacy, immigration status and presence of children younger than 12 years of age in household