Second-hand smoke or passive smoke is the involuntary inhalation of cigarette smoke from lit cigarettes and the exhaled mainstream from smokers. Second-hand smoke has been recognized as a risk factor for lung cancer in nonsmokers and classified as a Group I carcinogen by the International Agency for Research on Cancer.1 The association between second-hand smoke and cardiovascular diseases was described as causal by the UK Scientific Committee on Tobacco and Health2 in 2004 and by a Report of the US Surgeon General3 in 2006, by applying the Bradford-Hill criteria4 to the available body of evidence.5 Although the number of published studies is considerable, a small proportion of these were cohort studies, and the great majority were published 10 or more years ago. Prospective cohort studies are more reliable than case-control studies, particularly for weak associations; this is because outcomes can be ascertained after the collection of exposure information, thereby ruling out recall bias and reverse causality. Over the last decade, only 4 cohort studies investigating second-hand smoke and cardiovascular diseases have been published. In 3 of these,6–8 a positive association between second-hand smoke and cardiovascular diseases was found; the fourth found no association.9
No study has jointly investigated the effect of self-reported second-hand smoke exposure together with cotinine measurement in relation to cardiovascular mortality in never-smokers. Plasma cotinine (a metabolite of cotinine) is a reliable, although short-term, marker of second-hand smoke exposure10 and has been successfully validated against questionnaires addressing second-hand smoke exposure in various settings.11 Our study incorporated the aforementioned advantages (cohort design, cotinine measurement), thus substantially improving the overall study design and exposure assessment. The aim of our analysis was to investigate the risk of cardiovascular, cerebrovascular, and all circulatory deaths (cardiovascular, cerebrovascular, and hypertensive diseases, and diseases of peripheral arteries and veins), as well as total mortality, in relation to self-reported exposure to second-hand smoke in a large cohort of Europeans. The reliability of self-reported questionnaire information was evaluated by examining plasma cotinine in a subsample of subjects.
The EPIC Cohort
The European Prospective Investigation into Cancer and Nutrition (EPIC) is an ongoing multicenter European cohort study that recruited more than 500,000 healthy volunteers from 10 countries (Sweden, Denmark, Norway, the Netherlands, United Kingdom, France, Germany, Spain, Italy, and Greece) between 1993 and 1998. The study was originally designed for investigating the relationship between cancer and nutrition, but provides the opportunity to study many other outcomes as well. Dietary and nondietary (lifestyle, anthropometric, occupational) information12 was collected through 2 questionnaires at enrollment. Follow-up is carried out at regional level via National and Regional Cancer and Mortality Registries, and it is virtually complete.13 Currently, the median follow-up time of the study is 9.9 years.
The present analysis considered only 135,233 never-smokers (26% of the whole cohort) for whom information on second-hand smoke was available (after exclusion of 4,486 subjects without information on school level) from France, Italy, Netherlands, Germany, Sweden (Malmö only), Denmark, and Norway (except for second-hand smoke at home in Naples and Bilthoven; second-hand smoke at work in Bilthoven, and second-hand smoke in childhood in Utrecht) (eTable 1, http://links.lww.com/EDE/A358). Second-hand smoke exposure was assessed through the enrollment questionnaire with the following questions: “Does someone regularly smoke at home?” and “On work, are there people smoking in your presence?” Furthermore, subjects were asked if any of their parents smoked during their childhood (providing a proxy for second-hand smoke exposure during childhood). In a subset of centers, subjects were also asked hours/d of second-hand smoke exposure (centers in France and Italy except Naples) or number of cigarettes/d smoked by their partner (centers in Norway and Italy except Naples). The definition of never-smoker refers to any type of smoking.
Plasma cotinine was measured in 969 nonsmoking subjects identified as controls in the context of a nested case-control study.12 Plasma cotinine measurements were performed at the Mario Negri Institute (Milan, Italy), using ion-exchange chromatography and liquid chromatography-atmospheric pressure ionization–tandem mass spectrometry.3
For each subject who died during the follow-up period, the underlying cause of death was identified according to the International Classification of Diseases, 10th revision (ICD-10). Subjects were defined as dying from a circulatory disease when their underlying cause of death was coded I00 to I99. A cerebrovascular cause of death was coded I60 to I69 and a cardiovascular cause of death was defined as any circulatory death excluding cerebrovascular causes. In addition, for a subgroup analysis of circulatory causes of death, the following categories were considered: chronic rheumatic fever (I05-I09), hypertensive diseases (I10-I15), ischemic heart diseases (I20-I25), pulmonary heart diseases and diseases of the pulmonary circulation (I26-I28), other forms of heart diseases (I30-I52), diseases of the arteries, arterioles, and capillaries (I70-I79), and diseases of the veins, lymphatic vessels, and lymph-nodes (I80-I89).
To evaluate the reliability of questionnaire information on second-hand smoke, subjects were classified according to quartiles of the plasma cotinine distribution, after exclusion of subjects whose measurements were below the detection threshold (≤0.02 μg/L [≤0.11 nmol/L]) (classified as nondetectable). The association between cotinine level and self-reported second-hand smoke exposure at home and in the workplace was assessed using a t test. Distributions of cotinine level were compared between second-hand smoke-exposed and -unexposed groups both at home and in the workplace. Also, hazard ratios (HRs) and 95% confidence intervals (CIs) were computed for quartiles of cotinine in relation to mortality.
Cox proportional hazard models, with age as the primary time variable, were used to investigate the hazard associated with second-hand smoke exposure in relation to cardiovascular, cerebrovascular, and circulatory mortality and total mortality.
Each model was adjusted for potential confounders: sex, educational level (non/primary, technical secondary, university), physical activity (number of hours in leisure physical activity below the median, above the median and undetermined), and BMI (underweight <18.5 kg/m2, normal weight 18.5–24.9 kg/m2, overweight 25.0–29.9 kg/m2, obese >30.0 kg/m2); analyses were stratified by recruitment center. Interactions between second-hand smoke exposure at home, at work, and in childhood and BMI, physical activity, and education were tested with the likelihood ratio test (differences between models were considered statistically significant for P values <0.10). Further models were used to adjust separately for fruit and vegetable intake, fat intake (including energy intake), and alcohol consumption. Data on self-reported past medical history (previous myocardial infarction, previous stroke, and diagnosis of hypertension and hyperlipidemia) were available for a subset of persons (n = 53,041); separate models including these covariates as further potential confounders were run separately. The consistency of results was explored by running the same models on country-specific subcohorts, and testing differences of the crude association between second-hand smoke exposure and cardiovascular mortality with a Mantel-Haenszel exact test for homogeneity. Dose-response relationships between second-hand smoke and mortality outcomes were assessed using nonexposed subjects as the reference group, and both number of hours/d of exposure (<1 hours/d; 1–2 hours/d, ≥3 hours/d) and number of packs per day smoked by the partner (0.5, 1, ≥1.5) as categories of exposure. All statistical analyses were carried out using STATA 8.2 software (StataCorp, College Station, TX).
A sensitivity analysis exploring the potential for residual confounding due to socioeconomic status (SES) on the association between second-hand smoke exposure and cardiovascular mortality was carried out according to the methods described by Greenland.14 Estimates of the relative risk (RR) associated with SES were used to assess the potential effect of confounding between cardiovascular mortality and second-hand smoke. Based on the literature reporting higher cardiovascular mortality among those belonging to a lower socio-economic status,15 we hypothesized a relative risk for the association between SES and cardiovascular mortality ranging from 1.5 to 2.5. Using the distribution of subjects with no or primary school level in the current cohort (19% among the unexposed at home and 28% among the exposed, Table 1), plausible estimates of the prevalence of socioeconomic disadvantage among those unexposed and exposed to second-hand smoke were estimated. Estimates of the socioeconomic disadvantage in the unexposed group ranged from 5% to 30%, and in the exposed group from 20% to 40%.
Information on second-hand smoke at home, in the workplace, or in childhood was collected in 16, 15, and 17 of the 23 EPIC centers, respectively. Among the 251,266 never-smokers in the entire cohort, 135,233 (54%) were included in the present analysis. Overall in the study sample there were proportionally more women (n = 115,311; 85%) than men. Three partially overlapping subsets of available data were considered: 69,870 to analyze second-hand smoke exposure at home; 57,534 to analyze workplace exposure; and 111,698 to analyze exposure during childhood. The proportions of men and women exposed to second-hand smoke at home were 33% for each group. Exposure to second-hand smoke in the workplace was 90% in men and 84% in women. Exposure during childhood was 73% in men and 67% in women. Among the population included in this analysis, there were 4056 deaths: 399 were due to cardiovascular diseases, 146 to cerebrovascular diseases, and 545 to circulatory diseases (as well as 90 due to lung cancer).
Tables 1 and 2 show the demographic characteristics and distribution by exposure status. A steadily decreasing gradient of both total mortality and cause-specific mortality was observed with increasing educational level.
Cotinine measurements were available in a subset of 969 subjects. Levels were undetectable in 678 (70%). The remaining subjects had a median cotinine concentration of 0.66 μg/L (range, 0.03–8.87). We divided these subjects into quartiles of cotinine level. Cotinine measurements were strongly associated with self-reported exposure to second-hand smoke at home. The median cotinine concentration of detectable values among those exposed to second-hand smoke at home was 0.82 μg/L (n = 56; mean, 5.5 μg/L; skewness, 6.6; kurtosis, 48.4), while among the unexposed the median was 0.02 μg/L (n = 110; mean, 2.9 μg/L; skewness, 10.5; kurtosis, 115.2) (t test after log-transformation, P < 0.001). In contrast, cotinine was not associated with workplace second-hand smoke. The median was 0.02 μg/L in both categories of exposed and unexposed, although the means differed (4.1 μg/L among the exposed (n = 154) and 0.9 μg/L among the unexposed (n = 11)). The Figure 1 shows the distribution of cotinine measurements according to self-reported exposure to second-hand smoke at home and in the workplace.
We found no clear trend for cotinine in relation to total mortality, circulatory, or cardiovascular deaths; this is probably due to small numbers of outcomes (data not shown). Only subjects in the third quartile of distribution of plasma cotinine were at substantially increased risk of dying from a cardiovascular disease, compared with the subjects with undetectable levels (HR = 4.5 [95% CI = 1.3–15]).
The Cox regression analyses did not show any relationship between self-reported second-hand smoke exposure and total mortality (Table 3). For specific causes of death, second-hand smoke exposure at home was associated with an increased risk of dying from cardiovascular diseases (1.38 [1.01–1.90]) and with circulatory diseases (1.28 [0.98–1.69]). These results were consistent in men and women. No interaction was observed between second-hand smoke exposure at home, at work, and in childhood and BMI, physical activity, and education level (likelihood ratio test P > 0.15 for all comparisons). Neither second-hand smoke exposure at work nor during childhood was associated with total mortality or with specific causes of death. These results did not change after adjustment for fruit and vegetable intake, fat intake, and alcohol consumption (results not shown). Data on past medical history were available for 53,041 (39%) subjects. After adjustment for hypertension and hyperlipidemia, the modest association between cardiovascular mortality and second-hand smoke exposure at home remained (1.31 [0.86–1.98]) but with a wider confidence interval due to a reduction of the sample size. The association between second-hand smoke exposure at home and cardiovascular mortality remained positive across countries when analyzed separately. The Mantel-Haenszel exact test for homogeneity showed no evidence of differences among countries (P = 0.93).
The analysis of circulatory disease subtypes suggests a stronger association of second-hand smoke with death from hypertensive disease, ischemic heart diseases, other forms of heart disease, and diseases of veins, lymphatic vessels and lymph nodes, though comparisons are based on small numbers (Table 4). In contrast, second-hand smoke was not associated with mortality from pulmonary heart diseases and diseases of the pulmonary circulation, or with cerebrovascular diseases and diseases of the arteries, arterioles, and capillaries.
When dose-response relationships were examined, we found trends between the number of reported hours of exposure to second-hand smoke at home and risk of cardiovascular death (HR per each additional hour/d = 1.28 [95% CI = 1.02–1.60]), as well as risk of circulatory death (1.25 [1.04–1.50]) (Table 5). In particular, subjects exposed to at least 3 hours a day of passive smoking at home had an almost 2-fold increased risk of dying from a circulatory disease (HR = 1.98 [1.11–3.53]) or a cardiovascular disease (1.91 [0.93–4.04]) compared with the unexposed group. The dose-response relationship with the number of cigarettes smoked by the partner at home was suggestive of a trend. The largest HRs were for study participants whose partner smoked on average more than 30 cigarettes/d (for all circulatory deaths, HR = 2.94 [C.I. 1.11–7.78]).
It is well known that SES is a strong independent risk factor for all circulatory diseases.16 We assessed the potential for residual confounding due to SES on the association between second-hand smoke exposure and cardiovascular mortality (due to misclassification of SES by educational level). Results of the sensitivity analysis are shown in eTable 2, http://links.lww.com/EDE/A358. We calculated new RRs for cardiovascular mortality in relation to exposure to second-hand smoke at home for both sets of estimated prevalences. Even presuming high levels of residual confounding by SES, the association between second-hand smoke and cardiovascular mortality would persist, with relative-risk estimates between 1.47 and 1.73.
Our findings suggest that being exposed to second-hand smoke at home increases the risk of dying from cardiovascular diseases. Plasma cotinine level was strongly associated with self-reported exposure to second-hand smoke at home. These findings are consistent with the majority of results coming from recent studies investigating all-cause and cardiovascular disease mortality and second-hand smoke.5,8,17,18 A prospective cohort study among women in Shanghai6 found a positive association between second-hand smoke exposure at home and vascular diseases (in particular, stroke, a leading cause of death in China), and between second-hand smoke exposure at work and lung cancer. In New Zealand, despite a short follow-up period, there was a positive association between second-hand smoke exposure at home and all-cause and cardiovascular disease mortality in both men and women, although no association with lung cancer.7 McGhee et al17 also found a positive association between circulatory mortality and home-second-hand smoke but this study was somewhat limited because passive smoking status of the deceased was determined retrospectively through family interviews (with questions relating to second-hand smoke exposure at least 10 years prior to death). Finally, the British regional heart study8 (a smaller study of 4729 men) found an association between cotinine measurement and ischemic heart diseases, but not with cerebrovascular diseases.
The point estimate of the main finding of our study (the association between second-hand smoke exposure and cardiovascular diseases) is higher than those observed in recent meta-analyses,5 but consistent with the conclusions drawn by the UK Scientific Committee on Tobacco and Health2 and the Report of the US Surgeon General.3 Also, these results are supported by a dose-response relationship between the levels of second-hand smoke exposure at home and risk of dying from a cardiovascular disease and by consistency across countries and when subtypes of underlying circulatory causes of death are considered. Second-hand smoke did not seem to be associated with cerebrovascular disease mortality in our study, which might be due to the fact that this category included diseases with very different etiological mechanisms (ranging from intracerebral hemorrhage to thromboembolic or atherosclerotic ischemic stroke), thus attenuating any association.
A possible limitation of our analysis is the potentially poor adjustment for socioeconomic status, a variable that is well-known to be associated with cardiovascular outcomes.18,19 However, our sensitivity analysis showed that an excess risk of cardiovascular mortality persisted after adjustment for SES based on moderate to extreme assumptions. Further adjustment for past medical history carried out on a subsample of subjects tended to weaken slightly the association between second-hand smoke exposure and cardiovascular mortality, but a dose-response relationship was still apparent.
The discrepancy of results when analyzing second-hand smoke exposure at home and at work is more difficult to interpret; it may be due to lower recent exposure levels at work (indicated by lower cotinine concentrations in those exposed in the workplace). Questionnaire information on exposure was collected at the time of recruitment; since 1997, however, many European countries have restricted smoking in public places. Smoking bans were put in place in the Netherlands in January 2004 followed by Norway (June 2004), Italy (January 2005), Sweden (May 2005), Spain (January 2006), and the United Kingdom (July 2007).20 A proportion of those who originally reported exposure to second-hand smoke in the workplace probably became less exposed or unexposed during the follow-up period, leading to misclassification of exposure.
The effect of second-hand smoke on the cardiovascular system is likely to be acute (short-term),21 whereas lung cancer is related to long term exposure. As recently reported in a literature review on second-hand smoke and cardiovascular diseases, second-hand smoke is apparently capable of precipitating the acute manifestations of cardiovascular diseases via the activation of pro-atherosclerotic mechanisms, such as endothelial dysfunction, platelet activation, oxidative stress, and inflammatory reactions.22 Such mechanisms could help explain the nonlinear relationship between second-hand smoke exposure and circulatory diseases and thus explain the lack of association with recent (and decreasing) exposure at work. In summary, the present prospective study shows that exposure to second-hand smoke at home increases the risk of cardiovascular mortality. These results strengthen the evidence supporting an increased risk of cardiovascular mortality in relation to second-hand smoke exposure.
Europe Against cancer Program of the European Commission (SANCO); ISCIII, Red de Centros RCESP, C03/09; Deutsche Krebshilfe; Deutsches Krebsforschungszentrum; German Federal Ministry of Education and Research; Danish Cancer Society; Health Research Fund (FIS) of the Spanish Ministry of Health; Spanish Regional Governments of Andalucia, Asturias, Basque Country, Murcia and Navarra; Cancer Research, United Kingdom; Medical Research Council, United Kingdom; Stroke Association, United Kingdom; British Heart Foundation; Department of Health, United Kingdom; Food Standards Agency, United Kingdom; Wellcome Trust, United Kingdom Greek Ministry of Health; Greek Ministry of Education; Italian Association for Research on Cancer (AIRC); Italian National Research Council; Dutch Ministry of Public Health, Welfare and Sports; World Cancer Research Fund; Swedish Cancer; Swedish Scientific Council; Regional Government of Skåne, Sweden; Norwegian Cancer Society; Research Council of Norway; French Ministry of Health and French Health Products Safety Agency. All authors are independent from funders. Mortality data from the Netherlands are obtained from “Statistics Netherlands.”
1. IARC. Tobacco Smoke and Involuntary Smoking. Monograph on the Evaluation of Carcinogenic Risk to Humans
. Lyon, France: IARC; 2004.
2. UK Department of Health. Second-Hand Smoke: Review of Evidence Since 1998
. London: UK Department of Health; 2004.
3. US Department of Health and Human Services. The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General—Executive Summary
. Atlanta: US Department of Health and Human Services, Centres for Disease Control and Prevention Coordinating Centre for Health Promotion National Centre for Chronic Disease Prevention and Health Promotion and Office on Smoking and Health; 2006.
4. Bradford-Hill AB. The environment and disease: association or causation? Proc R Soc Med
5. Enstrom JE, Kabat GC. Environmental tobacco smoke and coronary heart disease mortality in the United States—a meta-analysis and critique. Inhal Toxicol
6. Wen W, Shu XO, Gao YT, et al. Environmental tobacco smoke and mortality in Chinese women who have never smoked: prospective cohort study. BMJ
7. Hill SE, Blakely T, Kawachi I, Woodward A. Mortality among lifelong nonsmokers exposed to secondhand smoke at home: cohort data and sensitivity analyses. Am J Epidemiol
8. Whincup PH, Gilg JA, Emberson JR, et al. Passive smoking and risk of coronary heart disease and stroke: prospective study with cotinine measurement. BMJ
9. Enstrom JE, Kabat GC. Environmental tobacco smoke and tobacco related mortality in a prospective study of Californians, 1960–98. BMJ
10. Seccareccia F, Zuccaro P, Pacifici R, et al. Serum cotinine as a marker of environmental tobacco smoke exposure in epidemiological studies: the experience of the MATISS project. Eur J Epidemiol
11. Nondahl DM, Cruickshanks KJ, Schubert CR. A questionnaire for assessing environmental tobacco smoke exposure. Environ Res
12. Vineis P, Hoek G, Krzyzanowski M, et al. Air pollution and risk of lung cancer in a prospective study in Europe. Int J Cancer
13. Riboli E, Hunt KJ, Slimani N, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr
14. Greenland S. Basic methods for sensitivity analysis of biases. Int J Epidemiol
15. Marmot MG. Socio-economic factors in cardiovascular disease. J Hypertens Suppl
16. Tyroler HA. The influence of socioeconomic factors on cardiovascular disease risk factor development. Prev Med
. 1999;29(6 pt 2):S36–S40.
17. McGhee SM, Ho SY, Schooling M, et al. Mortality associated with passive smoking in Hong Kong. BMJ
18. Marmot MG. Socio-economic factors in cardiovascular disease. J Hypertens Suppl
19. Tyroler HA. The influence of socioeconomic factors on cardiovascular disease risk factor development. Prev Med
. 1999;29(6 pt 2):S36–S40.
21. Raupach T, Schafer K, Konstantinides S, Andreas S. Secondhand smoke as an acute threat for the cardiovascular system: a change in paradigm. Eur Heart J
22. Ruckerl R, Ibald-Mulli A, Koenig W, et al. Air pollution and markers of inflammation and coagulation in patients with coronary heart disease. Am J Respir Crit Care Med