Part Two
Environmental Tobacco Smoke
Background
2.1 The Third Report38 of the ISCSH (1983) recorded a tentative link between exposure to environmental tobacco smoke (ETS) and lung cancer. This topic was pursued in more detail in the Fourth Report39 of the ISCSH (1988) in the light of information published since the Third Report and in response to increasing public concern about the postulated link between ETS and a number of adverse health effects.
2.2 After making allowances for misclassification of smokers and other confounding factors, the Fourth Report39 of the ISCSH concluded that there was an increase in the risk of lung cancer from exposure to environmental tobacco smoke in the range of 10% to 30%. This meant that people who had never smoked, but who had been exposed to environmental tobacco smoke through most of their lives, had a 10% to 30% higher risk of lung cancer than non-smokers who were not exposed to tobacco smoke. The Fourth Report39 also concluded that ETS might have other effects on health, and recommended that continued attention should be given to the role of ETS in the occurrence of respiratory illnesses in children.
2.3 In 1992 the United States Environmental Protection Agency published a report entitled “Respiratory Health Effects of Passive Smoking: lung cancer and other disorders”.40 It confirmed the findings published in the Fourth Report39on exposure to environmental tobacco smoke and lung cancer risk and also identified additional links between passive smoking and certain childhood illnesses.
2.4 At the inaugural meeting of SCOTH (1994), Committee members decided to undertake a comprehensive review of the health effects of exposure to ETS. It was agreed to update the assessment published in the Fourth Report39 on the effect of ETS exposure in relation to lung cancer and to respiratory diseases in children, to consider the effect of ETS exposure on the development of ischaemic heart disease, and to examine whether ETS had deleterious effects on fetal growth and preterm delivery.
Environmental Tobacco Smoke and Lung Cancer
2.5 Since earlier assessments by the National Research Council,41 the US Department of Health and Human Services,42 the fourth report of the ISCSH,39 the Australian report of the National Health and Medical Research Council (NHMRC) Working Party,43 and the US Environmental Protection Agency report (EPA)40 the number of epidemiological studies of ETS and lung cancer has more than doubled and there are additional data on the effect of biases, dietary confounding and the use of biomarkers to measure exposure to ETS.
2.6 An updated assessment on lung cancer and ETS44 was commissioned by the Department of Health. This consisted of a review of the relevant literature with a meta-analysis of the epidemiological studies. It was prepared by Mr A Hackshaw, Dr M Law and Professor N Wald and was considered by the Committee in 1997. The results of this review confirmed earlier conclusions that exposure to ETS is a cause of lung cancer.
2.7 The Committee also had sight of the of the California Environmental Protection Agency 1997 report “Health Effects of Exposure to Environmental Tobacco Smoke”45 and the Commonwealth of Australia NHMRC scientific information paper on “The Health Effects of Passive Smoking”, November 1997.46 Both these reports concluded that passive smoking caused lung cancer.
DH Commissioned Report on ETS and Lung Cancer44
2.8 This report, which was based on an analysis of 37 epidemiological studies of lung cancer in women who were life-long non-smokers living with smokers, showed that women had a statistically significant excess risk of lung cancer of 24% (95% confidence intervals, 13 - 36%). The analysis also showed that there was a dose response relationship between the risk of lung cancer and the number of cigarettes smoked by a person's partner, as well as the duration over which they had been exposed to their smoke. After adjusting for potential biases (misclassification and underestimation) and dietary confounding, the authors concluded that the underestimation of risk because of exposure to ETS in the reference group tended to cancel out the effects of misclassification bias and dietary confounding, so that the unadjusted (that is the observed) pooled relative risk from epidemiological studies provided a valid estimate of the true risk of lung cancer due to passive smoking. Overall the authors concluded that the evidence (epidemiological, dosimetric and toxicological) leads to the conclusion that breathing other people's tobacco smoke causes lung cancer in non-smokers. After adjustment for biases and confounding, the estimated excess risk is 26% (95% confidence interval 7-47%) which equates to several hundred deaths per year in Britain. The accumulation of evidence did not alter the earlier scientific assessment, and the present estimate is unlikely to be significantly altered by the accumulation of additional information.
47 the Congressional Research Service Report: ETS and lung cancer risk ( November 1995 )48 and a paper by Armitage A K et al (1997) entitled “Environmental Tobacco Smoke - is it really a carcinogen?”.45
2.11 The TMA responded to the publication of the paper by Hackshaw and colleagues44 by submitting a further commentary to the Committee. SCOTH members considered this supplementary commentary alongside all the other data but did not accept the TMA's position. (See also para. 2.13 below).
Review by the Committee on Carcinogenicity
2.12 The TMA submission was also evaluated separately by the Department of Health's Committee on Carcinogenicity of Chemicals in Food, Consumer Products and the Environment (CoC). The CoC not only assessed the paper by Hackshaw and colleagues44 and the TMA submission but also considered additional evidence such as the chemical composition of ETS and the presence of genotoxic carcinogens in ETS, exposure to ETS and deposition of genotoxic carcinogens in the respiratory tract. The CoC also considered the mutagenicity data on ETS particulates and the available animal inhalation studies with side stream smoke. Particular attention was paid to the available publications dealing with the investigation of carcinogen-adducts with blood proteins and the excretion of tobacco specific carcinogens in the urine of non-smokers exposed to ETS. Taking all the supporting data into consideration the COC concluded that passive smoking in non-smokers exposed over a substantial part of their life is associated with a 10-30% increase in the risk of lung cancer. The detailed conclusions of the CoC are attached at Annex H.
2.13 The CoC also considered the supplementary TMA commentary and concluded that there are no new data in the TMA commentary to change the conclusion of the Committee.
Report of a “European Working Group” on ETS and Lung Cancer46
2.14 The “European Working Group”, which was supported by the tobacco industry, concluded that there is no elevated risk of lung cancer from ETS exposure. The Group decided that, although meta-analysis showed a weak association of lung cancer risk with having a husband who smokes and some indication of a dose-response, this could be explained by misclassification and confounding. SCOTH members noted that the report had not been independently peer reviewed. There was no narrative account of the methods used and only a brief description of the meta-analysis. The report omitted relevant published genotoxicity and adduct studies. The inclusion criteria for studies were not adequately explained. It was thought that the division of the data into small subsets was inappropriate. The failure to investigate and take account of heterogeneity was noted. The omission of results on male never smokers, and failure to consider the increase in lung cancer risk associated with underestimation bias while allowing for the reduction in the risk due to misclassification bias was also noted. The Committee agreed that this report was an unsatisfactory examination of the scientific position and led to incorrect conclusions.
2.15 The “European Working Group” Report was assessed in 1996 in an article by Davey Smith and Phillips50 which also considered the Philip Morris media campaign (see para. 2.31 below). The authors concluded that, “...the partial and biased nature [of the advertisements] and “expert” report at the heart of the latest industry campaign represents a continuation of its characteristic behaviour”. The article drew attention to a report from the industry to the US Tobacco Institute as long ago as 1978 which stated that public worries about smoking were “the most dangerous development to the long term viability of the tobacco industry that has yet occurred” and that “the strategic and long run antidote to the passive smoking issue is...developing and widely publicising clear-cut, credible medical evidence that passive smoking is not harmful to the non-smoker's health”. Twenty years on the Committee has not seen such evidence.
Report of the United States Congressional Research Service48
2.16 The Congressional Research Service (CRS) provides a service to Congress and is funded by the US Government. The CRS report was not peer reviewed and the authors themselves comment that it was produced under resource constraints which precluded detailed review of all relevant studies. Committee members noted that the Report gave undue prominence to studies from within the industry or from its consultants. Problems connected with misclassification of smokers and the question of threshold effects were presented as seriously threatening the conclusion that ETS causes lung cancer, but much of the discussion was hypothetical and speculative. Members agreed that the Report did not critically challenge the detailed reviews by independent scientists, concluding that ETS causes lung cancer in non smokers.
Summary and Conclusions from SCOTH's Review of ETS and Lung Cancer
2.17 Members reviewed all the evidence on ETS and lung cancer discussed in the preceding paragraphs. They noted the extensive review provided by the CoC and accepted that Committee's overall conclusions: in particular it was noted that exposure to ETS leads to the delivery of genotoxic carcinogens to all parts of the respiratory tract. Reservations were expressed over the methodology of the TMA submission: it contained only a limited narrative and appropriate sensitivity analyses were not undertaken. Members noted that the results obtained for smoking by husband or by spouse were consistent with the findings of the US Environmental Protection Agency in 1992.40 Geographical heterogeneity had been introduced by a paper from China which yielded a relative risk of 0.79 which, if real, would indicate a protective effect of passive smoking. Although adjustment was made for confounding by misclassification bias, no adjustment was made for other potential sources of confounding, such as underestimation bias. No meta-analysis of dose response was undertaken although individual studies of limited statistical power were assessed for dose response. The Committee concluded that the TMA submission failed to examine the available evidence as a whole, notably that inhaling tobacco smoke from active smoking was a potent cause of lung cancer, (the TMA declined to express an opinion on this issue when asked by the Committee); that genotoxic carcinogens present in ETS were inhaled and absorbed by non-smokers and that the level of risk of lung cancer due to ETS exposure was consistent with the expected risk estimated from the effect of active smoking taking into account the lower exposure. The TMA did not give reasons why, in the light of this evidence, they reached their negative conclusion.
2.18 Members accepted the methods of analysis used by Mr Hackshaw and colleagues. The paper included a narrative review. Exclusion of one particular study (the one from China referred to above) removed any evidence of heterogeneity. Members accepted the exclusion of this study which gave implausible results and also noted the authors' own comment that the effect of ETS could have been obscured by another cause of lung cancer: exposure to open coal fires with little ventilation. Appropriate adjustments were made for misclassification bias by current and former smokers and for dietary confounding. A meta-analysis of dose responses was undertaken. A sensitivity analysis was carried out for misclassification which supported the risk estimation. SCOTH members noted that both submitted meta-analytical reports (i.e. that of the TMA and that of Mr Hackshaw and colleagues) gave similar risk estimates to that of the EPA,40 but differing conclusions had been drawn on the relevance of confounding and biases. The conclusions of the paper prepared by Mr Hackshaw and colleagues were judged to have been based on the totality of the evidence and an appropriate consideration of the epidemiological data in the context of other available evidence, including that from active smoking.
2.19 SCOTH members concluded that long term exposure of non-smokers to ETS caused an increased risk of lung cancer which, in those living with smokers, is in the region of 20-30%.
2.20 In the Fourth Report of the ISCSH39 it was thought helpful to explain exactly what the increased risk meant. For clarity that explanation is repeated here: If the risk of lung cancer in non-exposed non-smokers is 10 per 100,000, based on rates in non-smokers in the 35+ age group,51 a 20-30% increased risk in exposed non-smokers would be a rate of 12-13 per 100,000 per year. Thus we would expect an additional 2-3 lung cancer cases a year per 100,000 non-smokers regularly exposed to ETS. The numbers of people so exposed are not known precisely but an estimate would suggest about several hundred extra lung cancer deaths a year are caused by exposure to passive smoking.44 There are about 35,000 lung cancer deaths in the United Kingdom per year: it is estimated that 30,000 of these are directly attributable to active smoking.
Environmental Tobacco Smoke and Respiratory Diseases in Children
2.21 The U.S. Surgeon General's 1986 report42 and that of the Environmental Protection Agency (1992)40 considered the evidence associating parental smoking and respiratory diseases in childhood. A systematic review of research studies from which, where possible, summary estimates of the relative risks or odds ratios were produced, was carried out for the Department of Health by Dr Derek Cook, Professor Ross Anderson and Professor David Strachan.52 An Executive Summary of this review is to be found at Annex I. The authors were particularly concerned to consider the importance of residual confounding from other environmental factors, to assess the importance of exposure at different ages, and to distinguish between pre- and post-natal exposure. The following topics were considered: sudden infant death; lower respiratory tract illness in pre-school children; prevalence of asthma and respiratory symptoms in schoolchildren; incidence, severity and prognosis of asthma; bronchial reactivity; allergic sensitisation and ear disease and adenotonsillectomy.
2.22 The authors reviewed the evidence on parental smoking and sudden infant death syndrome (SIDS) and acute lower respiratory illness (LRI) in infancy and concluded that the relationship was causal. The elevated risks associated with smoking by other household members indicated that post-natal exposure was the predominant cause. For SIDS the adjusted pooled odds ratio for maternal smoking was 2.08 (95% CI, 1.90-2.21) ie. a doubling of risk. For LRI in infancy the pooled odds ratio for either parent smoking was 1.48 (95% CI, 1.40-1.57) and for maternal smoking was 1.64 (95% CI, 1.54-1.73). The associations with lower respiratory illness remained after adjustment for confounding factors and showed evidence of dose response.
2.23 There was also convincing evidence that parental smoking increased the risk of asthma and respiratory symptoms in schoolchildren, although at a lower risk than for infants. Maternal smoking had a greater effect than paternal smoking. There was evidence of a dose response relationship between risk and the number of smokers in the home for all symptoms (asthma, wheeze, cough, phlegm and breathlessness). The pooled odds ratios, where children were exposed to two parents smoking, were 1.52 (95% CI, 1.34-1.72) for asthma, 1.40 (95% CI, 1.29-1.51) for wheeze and 1.61 (95% CI, 1.50-1.73) for cough.
2.24 The report on bronchial hyper-reactivity summarised effect measures from eight studies with a pooled estimate of relative odds of 1.28 (95% CI, 1.08-1.52). Six other investigations did not show statistically significant effects and results from a further four studies were unpublished. It was concluded that the relationship between bronchial hyperactivity and ETS could have been overestimated by positive publication bias. No consistent association was found between parental smoking and allergic sensitisation. Significant and unexplained heterogeneity of odds ratios between studies created difficulties for interpretation. It was concluded that ETS exposure was not consistently related to allergic sensitisation and the relationship with bronchial hyper-reactivity had not been established.
2.25 It was also concluded that parental smoking caused acute and chronic middle ear disease in children, the pooled odds ratios for recurrent otitis media if either parent smoked being 1.41 (1.19-1.65), and for middle ear effusion, 1.38 (1.23-1.55).
Environmental Tobacco Smoke and Ischaemic Heart Disease
2.26 Meta-analyses of the epidemiological studies indicated a relative risk of ischaemic heart disease (IHD) of about 1.3 (an excess risk of 30%) in non-smokers exposed to environmental tobacco smoke compared to those not exposed. This effect appears implausibly large, since the relative risk of IHD with active smoking is about two-fold and exposure, based on urinary cotinine studies, is only equivalent to about 1% of that from actively smoking. With a linear dose-response relationship the expected relative risk would be 1.01 (1% of the excess risk of about 100% from active smoking). An explanation for this substantial difference was needed.
2.27 The evidence on active and passive smoking and IHD was reviewed in a paper prepared for the Committee53 by Dr Malcolm Law, Dr J Morris and Professor Nicholas Wald which suggested an explanation. Part of the association was due to dietary difference between non-smokers who live with smokers and those who do not, and part of the association was judged to be causal. The best estimate of the reversible (cause and effect) component of the association is a relative risk of 1.23 or an excess risk of 23% in non-smokers exposed to ETS compared to those not exposed. The main causal factor appears to be an increase in platelet aggregation, a major step in the formation of thrombi, which may occlude the coronary arteries. The dose-response relationship between ETS exposure and platelet aggregation is non-linear and is consistant with results from other studies of the effects of tobacco smoke on platelet aggregation. The Committee concluded that there was a cause and effect relationship between passive smoking and IHD and that enhanced platelet aggregation was a plausible mechanism. It was also noted that there were other possible mechanisms by which ETS exposure could have an adverse effect on the cardiovascular system, including reduction in oxygen transportation to the heart, acceleration of atherosclerosis53 and increases in plasma fibrinogen. If the size of the effect was as great as Law et al estimated, it would represent a significant public health problem. The Committee, in drawing their conclusions, considered a commentary on the Law et al paper44 which was received from the TMA.
2.28 The Committee noted the US prospective study of passive smoking among 32,000 nurses which reported an increased IHD risk, judged to be largely causal. (Kawachi et al.)54
Environmental Tobacco Smoke and Pregnancy Outcome
2.29 The Committee noted the Avon Longitudinal Survey of Pregnancy and Childbirth (ALSPAC) which is a prospective study and includes 14,100 live births. They were grateful to receive a report55 by Professor Jean Golding on preliminary analyses related to “Passive Smoking and Outcome of Pregnancy”.
The Confidential Enquiry into Stillbirths and Deaths in Infancy
2.30 The Committee considered the Report of the National Advisory Board of the Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI)56 and the subsequent publication in the British Medical Journal in July 1996.57 This case - control study shows that after controlling for maternal smoking during pregnancy, ETS was significantly associated with the sudden infant death syndrome. The adjusted odds ratio for paternal smoking was 2.50 and if both parents smoked, 3.79. There was a dose-response effect. These odds ratios are consistent with previously published studies. This study was included in the commissioned systematic review.48
Philip Morris “Passive Smoking Campaign”
2.31 In June 1996, Philip Morris launched an advertisment campaign in newspapers in a number of European countries. The aim of the campaign was to undermine public health messages about the health risks of passive smoking. The Committee were aware that the Advertising Standards Authority (ASA) received complaints about the Philip Morris campaign which compared the risk of passive smoking with a variety of other everyday activities like eating a biscuit, eating pepper frequently, drinking milk or chlorinated water and cooking frequently with rapeseed oil. The complaint that the advertisements misrepresented the findings of the studies (quoted in the advertisments) was upheld and, in a report published in October 1996, the Authority said “...it considered that the advertisement gave the misleading impression that passive smoking had been conclusively proved to pose less danger to the health of UK consumers than the five activities placed above it in the table in the advertisement. The Authority asked the advertisers to withdraw the advertisement”.
Conclusions
2.32 Exposure to environmental tobacco smoke is a cause of lung cancer and, in those with long term exposure, the increased risk is in the order of 20-30%.
2.33 Exposure to environmental tobacco smoke is a cause of ischaemic heart diseases and if current published estimates of magnitude of relative risk are validated, such exposure represents a substantial public health hazard.
2.34 Smoking in the presence of infants and children is a cause of serious respiratory illness and asthmatic attacks.
2.35 Sudden infant death syndrome, the main cause of post-neonatal death in the first year of life, is associated with exposure to environmental tobacco smoke. The association is judged to be one of cause and effect.
2.36 Middle ear disease in children is linked with parental smoking and this association is likely to be causal.
Recommendations
2.37 Smoking in public places should be restricted on the grounds of public health. The level of restriction should vary according to the different categories of public place but smoking should not be allowed in public service buildings or on public transport, other than in specially designated and isolated areas. Wherever possible, smoking should not be allowed in the work place.
2.38 There is a need for public education about the risks of smoking in the home particularly in relation to respiratory diseases in children.
2.39 Health education programmes should focus on the dangers of ETS in fetal development and, postnatally, in the sudden infant death syndrome.
|
