Abstract
The aims of this article are to synthesize the evidence on health effects of environmental tobacco smoke (ETS) in the elderly and to discuss questions for future research. Health effects are divided into aetiological and prognostic studies.
There is convincing evidence that ETS causes lung cancer and coronary heart disease, both of which are diseases of the elderly. Several cross-sectional studies show increased occurrence of chronic respiratory symptoms and deficits in ventilatory lung function in relation to ETS exposure at home and/or at work.
A limited number of studies have found significant relations between ETS exposure and asthma, chronic obstructive pulmonary disease (COPD), pneumococcal infections and stroke in the elderly. Longitudinal studies are needed before any definite conclusions can be made concerning ETS and noncarcinogenic respiratory diseases in the elderly.
The potential role of environmental tobacco smoke exposure as a prognostic factor determining development of a pre-existing respiratory or heart disease is an important new area for research.
Personal smoking is a well-established cause of lung cancer, chronic obstructive pulmonary disease (COPD) and coronary heart disease 1. The first reports associating spouse's smoking to lung cancer in nonsmoking females were published in 1981 2, 3. Since then the research on health effects of environmental tobacco smoke (ETS) has expanded, especially on respiratory effects in children and lung cancer in adults 4.
The aims of this article are to synthesize the evidence on health effects of ETS in the elderly and to discuss questions for future research. The health effects are separated into two categories: aetiological studies, i.e. studies on development of a new disease, and prognostic studies, i.e. studies on development over time (or natural history) of a pre-existing disease. The incidence of many diseases that have been related to ETS exposure, such as COPD and lung cancer, increases with age 1, 5. Table 1⇓ summarizes the aetiological studies of ETS in the elderly. It also presents the author's judgement on causality of the relations, based on the amount of studies, their validity, evidence of dose-response relations, and biological plausibility. Surprisingly few studies have evaluated the role of ETS in determining the prognosis of a pre-existing disease, such as asthma or COPD, although this may be of great importance for the elderly in their every day life. In this age period, small changes in respiratory function can have a critical impact on their overall functional capacity. In addition, elderly people with an underlying disease may be exposed to very high levels of ETS, if their disease restricts them to indoor environments.
Environmental tobacco smoke exposure
Tobacco smoke contains >4,000 compounds, including >50 human and/or animal carcinogens and many irritant and toxic agents 6, 7. Mainstream smoke and sidestream smoke are qualitatively very similar, but the quantities of constituents differ due to different burning conditions 6. Sidestream smoke, emitted directly into the air during burning of a tobacco product between puffs, contains considerably higher concentrations of many carcinogens and toxic substances (table 2⇓), but is diluted into a larger volume of air. Detectable levels of cotinine, a biomarker of ETS exposure in serum, urine and saliva, have been measured in ≥80% of the nonsmoking populations 8, 9. In a survey representing a US civilian noninstitutionalized population, the reported ETS exposure at home or at work was 18% among males and 11% among females in the age group ≥60 yrs 9. Assessment of exposure to ETS is an essential part when studying health effects, carrying out risk assessment, and developing preventive strategies, as discussed in previous articles 10, 11.
Aetiological studies of environmental tobacco smoke exposure and health in the elderly
Lung cancer
The induction period of lung cancer is long, so its risk is likely to be related to the cumulative ETS exposure over entire adulthood, while the diagnosis is often confirmed in the older age. Altogether 38 case-control studies 12–49 and five prospective studies 50–54 from 12 countries have been published on ETS and lung cancer, and several meta-analyses of these have been performed 6, 55–60 (figs. 1a and 1b⇓). In a recent meta-analysis of studies on spousal smoking, the pooled odds ratio (OR) was 1.24 (95% confidence interval (CI) 1.13–1.36) for females and 1.34 (95% CI 0.97–1.84) for males 58. The estimated OR for females and males combined was 1.23 (95% CI 1.13–1.34) (table 1⇓). Adjustment for potential misclassification of some active smokers as never-smokers, misclassification of the reference group due to ETS exposure from sources other than the spouse, and confounding by dietary habits did not change these estimates. Several studies have shown a dose-dependent increase in the risk of lung cancer by the number of cigarettes smoked daily by the spouse, by the number of years the subject lived with a smoker, and by cumulative exposure in pack-yrs 4, 58.
Workplace is another major source of ETS exposure in adulthood 6, 8, 9. Altogether 20 studies have assessed the risk of lung cancer in relation to workplace ETS exposure 4, 16, 18, 19, 21, 29, 31–33, 35, 37, 38, 41, 42, 45–49, 52, 61, 62. The studies including quantitative assessment of occupational ETS exposure have usually indicated a relation with the risk of lung cancer, and in general, the risk estimates for workplace ETS exposure have been consistent with the estimates for spousal smoking. In a recent meta-analysis of 14 workplace ETS studies, the pooled OR for lung cancer was 1.25 (95% CI 1.08–1.41) (table 1⇓) 60.
In conclusion, a causal relation between ETS and lung cancer is supported by a large amount of studies from different geographical locations, genetic populations and cultural environments. ETS contains several known carcinogenic substances and its carcinogenic effect is biologically plausible 4, 6. Abundant evidence exists on exposure-response relation between ETS and the risk of lung cancer. Selection or information biases or confounding may have affected the risk estimates of lung cancer to some extent, but they do not explain the observed relations completely, and their effects tend to cancel out when adjustments are made for them.
Chronic respiratory symptoms
A total of 12 cross-sectional 63–74 and two longitudinal studies 75, 76 addressed the role of ETS exposure in the development of chronic respiratory symptoms in adults, but many of the studies focussed on younger age groups. Only seven cross-sectional studies included subjects aged >60 yrs 65–69, 71, 73, while none were designed to address the effects of ETS in particular in the elderly. Most of the studies assessed household ETS exposure, while two assessed workplace exposures 69, 73. All seven studies, except one 65, showed an increased risk of chronic respiratory symptoms, including cough, phlegm production, wheezing, and breathlessness, in relation to ETS exposure at home and/or in the workplace, although not all reached statistical significance 67, 68. The excess risk of symptoms related to ETS varied between 35% and >300% (table 1⇓).
A study of females from Singapore showed dose-response relations between the number of household smokers and the risk of cough, wheezing and dyspnoea 71. A study from Sweden included 205 never-smokers with severe α1-antitrypsin deficiency (with PiZZ genotype), who were suggested to be especially susceptible to the adverse effects of ETS 73. Significantly increased risk was observed for chronic bronchitis in relation to ETS exposure.
In conclusion, based on seven cross-sectional studies from different countries, ETS exposure appears to cause chronic respiratory symptoms in the elderly 4. This is biologically plausible, since tobacco smoke is known to contain several irritative compounds, such as ammonia, sulphur dioxide, nitrogen oxides, acrolein, and formaldehyde 6, 7. In addition, active smoking is known to cause chronic respiratory symptoms 5. All of the studies adjusted for confounders, although for a variable set of them, including age, sex, socioeconomic status, occupational exposures, and other indoor sources of pollution. However, longitudinal studies focussing on the older age groups are needed.
Lung function
Altogether 20 cross-sectional studies 64, 65, 67, 68, 71, 77–91, one case-control study 92, and four longitudinal studies 81, 90, 93, 94 have addressed the relations between ETS exposure and ventilatory lung function in adulthood. A meta-analysis including nine of the cross-sectional studies estimated the effect of ETS on forced expiratory volume in one second (FEV1) to be −2.7% (95% CI −4.1–−1.2%) (table 1⇓) 90. However, most of the adult studies focussed on younger age groups. Only 10 cross-sectional studies 65, 67, 68, 71, 84–87, 89, 90 and one longitudinal study 90 included subjects aged >60 yrs.
The results on lung function have been somewhat inconsistent. The majority of the studies detected significant adverse effects on FEV1, forced vital capacity (FVC), and/or indices of small airways function in relation to home and/or workplace ETS exposure 67, 71, 84, 86, 90. The harmful effect on FEV1 was from −50 mL in the Singapore study 71 to −257 mL in the China study 86. All but one cross-sectional study assessed exposure from household sources, while three studies assessed workplace ETS exposure 84, 86, 89. Two of the cross-sectional studies focussed especially on the elderly and showed no significant effects of household smoking on lung function, but both are likely to have been hampered by survival bias 85, 87. Two studies showed a significant dose-response relations between the amount of household smoking and reduction in FEV1 86, 90. The only longitudinal study including the elderly did not find any significant effect of ETS on the change in lung function in a 7-yr follow-up (table 1⇓) 90.
A study from California addressed the effects of ETS on peak expiratory flow (PEF) variability 89. This variability was significantly increased among males who had worked for 10 yrs with a smoker.
In conclusion, several cross-sectional studies including the elderly indicate small, but significant, reductions in ventilatory function parameters among subjects exposed to ETS at home and/or in the workplace. This effect seems to be dose-dependent, and is observed mainly in countries and in occupations with high ETS exposure levels. Exposure misclassification and confounding do not explain the observed relations. Most of the studies controlled for sex, age, height, socioeconomic status and/or education. Some also adjusted for housing conditions and other indoor air pollutants, occupational exposures, and outdoor air pollutants. The well-known effects of active smoking on lung-function impairment support the biological plausibility of the adverse effects of ETS 5. It is possible that the small effects on lung function detected in several cross-sectional studies are due to a susceptible group experiencing a more pronounced lung-function deficit in relation to ETS exposure. Potential determinants of susceptibility should be studied in the future. A limited number of studies focussing on elderly populations have been carried out, and there is a need for more longitudinal studies.
Asthma
Six epidemiological studies have addressed the role of ETS in induction of asthma in adulthood 71, 72, 95–98. Four of them 71, 95, 96, 98 included subjects >60 yrs. All four studies found an increased risk of asthma in relation to ETS exposure, although the effect reached statistical significance in only two studies 95, 98. All three studies that included assessment of workplace ETS exposure found an increased risk of asthma related to workplace exposure 95, 96, 98. The excess risk of asthma related to ETS was estimated to be 45–100% (table 1⇓). Two studies based the definition of asthma on objective measurements 96, 98. In the Swedish study, a potential problem was inclusion of active smokers in the study population 96.
A 10-yr longitudinal study among the Seventh-Day Adventists from USA reported an OR of 1.45 (95% CI 1.21–1.80) for asthma in relation to 10 yrs of workplace ETS exposure 95. In a 15-yr follow-up of this cohort, the OR was 1.21 per 7 yrs of workplace ETS exposure (95% CI 1.04–1.39) among females 99. The slight reduction in the risk estimate probably reflects selective loss of subjects who were more likely to have asthma and high levels of ETS. This type of survival bias is of concern especially in studies of the elderly.
In conclusion, a limited number of studies have been published on ETS and development of asthma in the elderly, but they consistently indicate an increased risk of asthma among those exposed to ETS at home or at work. All studies controlled for confounding, although the set of confounders was variable. Factors that were controlled included age, sex, education, occupational exposures, atopy, and outdoor air pollution. Only two studies based the definition of asthma on objective measurements. Potential mechanisms by which ETS could induce asthma include an inflammatory reaction in the airways related to the irritative substances contained by ETS, which would be a similar type of mechanism as has been shown in the case of irritant-induced occupational asthma 100. Skin-test positivity to tobacco is rare 101–103, and the effect of ETS on asthma seems to occur via irritative rather than allergic mechanisms. Tobacco smoke may also increase epithelial permeability to environmental allergens, thus enhancing allergic reactions to other inhalable allergens 104, 105. Before making any definite conclusions, more studies among adults, and especially among the elderly, are needed.
Chronic obstructive pulmonary disease
Chronic obstructive pulmonary disease develops slowly over the years and is often diagnosed in older age. Three case-control 106–108 and three longitudinal studies 2, 109, 110 have investigated the effects of ETS on development of or mortality from COPD. All of them included subjects aged ≥60 yrs. COPD was defined in variable ways, based on symptoms and/or diagnoses made by a doctor reported in questionnaires, lung function measurements, mortality registers, or a combination of these. An increased risk of COPD was found in all six studies, at least in the high ETS exposure categories. The excess risk related to ETS exposure was estimated to be from 60% to >400% (table 1⇓). Workplace ETS exposure was assessed in the longitudinal study of Seventh-Day Adventists from USA 110, 111, while the others limited exposure assessment to household only.
In conclusion, a limited number of studies have addressed the relation between ETS exposure and development of COPD, but all of them showed an increased risk. Dose-response relation was suggested by four studies 2, 107, 108, 110. Age and sex were taken into account as potential confounders in all of the studies, the other factors adjusted for included the participant's own or spouse's occupation, housing quality, other indoor pollutants, and outdoor pollution. Tobacco smoke has been demonstrated to induce inflammatory reactions in the airways and lung parenchyma of active smokers 112–114 and the biological effects of ETS are likely to be similar to those of the mainstream smoke. More studies with better outcome and exposure assessment are needed before any definite conclusions can be made.
Respiratory infections
There is convincing evidence of the role of passive smoking in enhancing susceptibility to respiratory infections in children 4, 115, but only one study addressed infections in adults 116. This population-based case-control study from USA investigated the effects of tobacco smoke on pneumococcal bacteraemia and meningitis in adults aged 18–64 yrs. The estimated OR for ETS exposure was 2.5 (95% CI 1.2–5.1) (table 1⇓), adjusted for sex, race, presence of chronic illnesses, education, and living with children going to day care. The population attributable risk for passive smoking was 17%. A dose-response relation was observed between the hours of daily ETS exposure and the risk of invasive pneumococcal disease.
There may be several mechanisms by which ETS increases susceptibility to infections: tobacco smoke impairs host defence mechanisms through weakening of immunological responses and mucociliary clearance 5, 117, 118 and tobacco smoke has been shown to enhance bacterial adherence and to disrupt respiratory epithelium 119, 120. More studies are needed addressing the relations between ETS exposure and infections in adulthood, especially among the elderly.
Coronary heart disease
A recent meta-analysis including nine prospective and 10 case-control studies estimated the risk of coronary heart diseases (CHD) in relation to ETS exposure at home 121. The OR was 1.30 (95% CI 1.22–1.38) at the age of 65 yrs (table 1⇓). Another meta-analysis, based on eight case-control and nine cohort studies, yielded a risk estimate of 1.25 (1.17–1.33) related to spouse's smoking 122. Most of the studies included the elderly or had a long follow-up extending into the older ages. The effect across the studies appeared to be consistent and was not weakened by adjusting for potential confounders in these studies 122. Several studies found an increasing risk of CHD in relation to the number of cigarettes smoked by the spouse and the duration of exposure 122. Steenland 123 carried out a meta-analysis of three cohort and two case-controls studies reporting ORs for ETS exposure at work. The estimated pooled risk ratio was 1.21 (95% CI 1.04–1.41) (table 1⇓).
In conclusion, there is accumulating evidence suggesting that ETS exposure at home 121, 122 and in the workplace 123, 124 causes coronary heart disease. The size of the risk estimates for ETS seem quite high compared to the estimates for active smoking, but there are several biologically plausible explanations for this. Many of the substances linked to CHD are in the vapour phase in ETS, but in the particle phase in mainstream smoke, so they are deposited more completely in the lungs and are harder to clear than particle deposits 125. Experimental studies have shown that ETS causes platelet activation and endothelial injury and suggest that passive smokers are less able to adapt to the adverse effects of toxins of tobacco smoke than active smokers are 126–128.
Stroke
The association of active smoking with cerebrovascular stroke has been demonstrated in many studies and is likely to be causal 129, 130. Two case-control studies, both including the elderly, have addressed the relations of ETS exposure with the risk of cerebral ischaemia 130, 131. In a hospital-based case-control of subjects aged 20–87 yrs, exposure to a smoking spouse was an independent risk factor for cerebral ischaemia, with an OR of 1.7 (95% CI 1.1–2.6) 130. The Auckland Stroke Study was a population-based case-control study among the age group 35–74 yrs 131. Nonsmokers exposed to ETS at home and/or at work had a significantly increased stroke incidence and fatality in males (OR 2.10, 95% CI 1.33–3.32) and in females (OR 1.66, 95% CI 1.07–2.57), after adjusting for known risk factors of stroke, such as hypertension, diabetes, and heart disease (table 1⇓). In conclusion, these two studies suggest an independent effect of ETS exposure on the risk of stroke.
Prognostic studies of environmental tobacco smoke exposure in the elderly
Chronic respiratory symptoms
No study was identified on the role of ETS in determining persistence or remission of chronic respiratory symptoms in the elderly. A cross-sectional study of 2,992 adults from the UK assessed respiratory symptom severity across smoking categories, combining several symptoms and their frequency in the past month 132. The median score was 2.8 for never-smokers and 4.2 for passive smokers. The adjusted OR of severe symptoms was 1.4 (95% CI 1.0–1.8) in relation to current household ETS exposure.
Lung function
A study from California examined a relatively young population of bartenders before and after prohibition of smoking in bars 133. The mean age of this population was 43 yrs. Cessation or reduction of ETS exposure at work led to an improved FEV1 and FVC. No study has addressed the potential prognostic effects on lung function development of reducing ETS exposure in the elderly.
Asthma
In community- and hospital-based surveys, 69–78% of asthma patients report that cigarette smoke aggravates their asthmatic symptoms 134–136. Six epidemiological studies 91, 137–141 have addressed the effects of ETS on pre-existing asthma in adults, but only one of these included subjects aged >60 yrs 137. The studies on younger age groups detected increased need of asthma medications, increased number of emergency department visits, urgent physician visits, and hospitalizations, increased work-related asthmatic symptoms, and lower FEV1 and forced mid-expiratory flow levels in relation to ETS exposure 138–140. In a panel study of 164 asthmatic nonsmokers aged 18–70 yrs from USA, daily ETS exposure was significantly related to an increased risk of moderate or severe cough (OR 1.21, 95% 1.01–1.46), moderate or severe breathlessness (OR 1.85, 95% CI 1.57–2.18), and nocturnal asthma symptoms (OR 1.24, 95% CI 1.00–1.53) 137. In addition to the epidemiological studies, several experimental studies, usually focussing on younger age groups, have suggested that there is a subpopulation of asthmatics who are sensitive to ETS 4, 102. These asthmatics experience increased respiratory symptoms, decreased lung function and increased bronchial hyperresponsiveness in response to ETS exposure. The determinants of such susceptibility are not well understood.
In conclusion, a limited number of epidemiological studies and several experimental studies, almost all of which have focussed on younger age groups, suggest that ETS exposure contributes to severity and exacerbations of asthma among adult asthmatics. The elderly were included in only one study. Thus, prognostic studies on effects of ETS on pre-existing asthma in the elderly are needed.
Chronic obstructive pulmonary disease
No study was identified on the effects of ETS on the long-term prognosis of COPD. Two studies assessed the impact of ETS exposure on respiratory-related activity restrictions 142, 143. Both of these studies used data collected in the US National Health Interview Surveys conducted by the Census Bureau for the National Center for Health Statistics. In the first study, respiratory-related restricted activity in the past 2 weeks increased among nonsmokers on average by 1% per exposure to one cigarette per day at home 142. In the other study, respiratory disease exacerbation was defined as activity limitation or physician visit in the preceding 2 weeks because of chronic bronchitis, asthma, emphysema, or chronic sinusitis 143. It increased significantly in relation to ETS exposure at home and/or work, with an OR of 1.44 (95% CI 1.07–1.95) adjusted for age, sex, season, socioeconomic status, and race. In conclusion, these two studies suggest that ETS exposure contributes to the adverse consequences of COPD, but there is a lack of prognostic studies on COPD.
Coronary heart disease
Experimental studies have demonstrated among patients with stable angina that ETS exposure increases heart rate, blood pressure, and blood carboxyhaemoglobin, and reduces exercise ability 127, 128. Again, there is lack of epidemiological studies assessing the role of ETS as a long-term prognostic factor in subjects with pre-existing heart disease.
Summary and questions for future research
Table 1⇓ summarizes the evidence from aetiological studies on ETS in adults. The risk estimates are from published meta-analyses when available or from individual studies including the elderly. Abundant studies have been published on lung cancer and CHD, while aetiological research on the other health effects in adults has only recently received attention. Few studies on the noncarcinogenic respiratory health effects have focussed on the elderly.
Surprisingly, few studies have evaluated the role of ETS in determining the prognosis of pre-existing diseases. The effects of ETS on prognosis of diseases may have a critical impact on the functional capacity and quality of life among the elderly, but there is a lack of such studies in elderly populations.
Questions for future research
The evidence on the aetiological role of ETS in asthma, lung function impairment, COPD, respiratory infections, and stroke is suggestive but limited, and there is a clear need for more aetiological studies on these diseases and conditions in the elderly. Since selective survival of healthy (and unexposed) subjects is of special concern in studies of older age groups, longitudinal studies initiated in younger age periods and followed through the older age periods are needed. Focussing on high-quality assessment of both ETS exposure and disease condition in question is essential for achieving valid and precise risk estimates. Exposure assessment should take the relevant exposure period into account, which is often different in aetiological and in prognostic studies 11. For all health effects characteristics that determine susceptibility to the adverse effects of ETS should be studied. As a part of identifying especially sensitive subgroups, potential interactions between ETS and other environmental exposures should be addressed in the future. For example, previous occupational exposures may increase susceptibility of the elderly to adverse effects of ETS.
The potential role of environmental tobacco smoke exposure as a prognostic factor determining the development of a pre-existing respiratory or ischaemic heart disease is an important new area for research. Such research is likely to “throw light” on the harmful impact of environmental tobacco smoke especially among the elderly, who often have chronic diseases and may be exposed to high environmental tobacco smoke levels because of activity limitations related to these diseases.
Footnotes
- Received August 10, 2001.
- Accepted August 10, 2001.
- © ERS Journals Ltd