Tobacco Documents Online

Smoke and the Non-Smoker Scientific Aspects of Environmental Tobacco Smoke

Smoke And The Non-Smoker Summary Many claims have been made about the alleged~ health effects of ETS - environ- mental tobacco smoke. Among the most frequent are that smoking inthe home and public places, especially the workplace, causes lung cancer and diseases of the lung and heart in non-smokers, and it is particularly harmful to those who already have respiratory and heart diseases. The claim that exposure to ETS is as- sociated with an increased risk of lung cancer is based almost solely upon epi- demiologic studies of non-smoking wives of smokers. Although a number of these studies report small increases in risk for non-smokers married to smokers com- pared to those married to non-smokers, most of the reported associations are not statistically significant. This means that the data in those studies do not support rejection& the hypothesis that there is no association between ETS and lung cancer in non-smokers. Moreover, such small, positive associations are likely the result of bias and~ confounding factors. For example, actual ETS exposures were not measured in any of the studies: instead, exposures were estimated by recall, an inaccurate method of estimating expo- sures. In addition, most of the studies failed to consider important potential con- founding factors such as diet, occupation, exposure to outdoor air pollution4 and ex- posure to combustion products of various cooking and heating fuels. The available data on other respiratory effects in adults are inconclusive and do not adequately support the claim that ETS causes pulmonary disease in non-smo- kers. There is no reliable evidence suppor- ting the judgement that exposure to ETS causes asthma. In addition, studies which have assessed the influence of ETS expo- sures on asthmatics are not conclusive one way or the other. While two studies sug- gest a response to ETS, four studies report no objective changes in asthmatics even after prolonged and heavy exposure to ETS. In the studies reporting an effect among asthmatics, researchers were un- able to rule out the influence of emotional and psychological stress on the reactions of their patients. One of the most emotional claims re- lating to ETS is that parental smoking adversely affects the respiratory health of children. While some studies suggest the existence of an association, others do not. Moreover, recent studies which have examined dampness and overcrowding in the home, poor diet, daycare attendance, access to medical care, etc., suggest that such factors are associated with childhood respratory health independent of whether or not the parents smoke. Similarly, published studies on ETS and the risk of cardiovascular disease sug- gest no clear trends. The studies typically report associations which are not statisti- cally significant, and important con- founders such as lifestyle, diet, exercise and heredity are rarely considered. Given the known limitations of epi- demiology as a reliable means of detect- ing low-risk effects, the lack of actual ETS exposure measurements, the effect that misclassification of a small proportion of smokers as non-smokers can have in rela- tion to bias, and the fact that many studies have obvious weaknesses, the only logicaU conclusion at the present time is that ETS has not been established scientifically as a cause of disease in non-smokers. Introduction Several studies are frequently cited as "proof ' that tobacco smoke is harmful' to non-smokers. The most widely known studies are those of Hirayamal and Tri- chopoulos,2 who reported that non-smo- kers exposed to tobacco smoke have a higher risk of developing lung cancer, and, White and Froeb,3 who asserted that non- smokers with long-term exposure to to- bacco smoke in the workplace have significantly reduced small-airways func- tion. Other studies have suggested that children, adults with pre-existing heart disease and asthmatic individuals are ad- versely affected by exposure to environ- mental tobacco smoke. In 1986, two reports in the U.S., one issued by a National Academy of Sciences Committee to the U.S. Environmental Protection Agency4 and the other by the Office of the Surgeon General,5 which reviewed the evidence on ETS, concluded that ETS exposure is harmful to non-smo- kers. More recently, the Independent Scien- tific Committee on Smoking and Health Page 1

Smoke And The Non-Smoker in the United Kingdom issued its report with conclusions similar to those of the U.S. Surgeon General.6 These reports have been, used by anti-smokers to press- ure legislative bodies, unions, office and restaurant owners to adopt non-smoking policies in an attempt to promote a 'smoke-free' society. Critical analyses of these studies and reports by physicians and scientists, together with other studies which reported finding no association between ETS and human disease, suggest that these adverse health claims are highly tenuous and with- out a$ound scientific basis: For example, in October 1986, an international gather- i~ng~ of scientists in Essen, Germany, con- sidered the experimentaI and toxicological findings regarding ETS7. Gostomzyk concluded in his review of the proceedings that "so far, even toxicology has not been able to ascertain with any greater degree of probability than did epi- demiology that there exists a1ink between damage to health and,passive smoking".S Two international conferences on in- door air quality, one in London in 1988,, the other in Brussels in 1989, featured' many important presentations which chal- lenge the reported association of ETS with chronic disease in non-smokers. Other re- search papers presented data which indi', cate that non-smoker exposure to ETS in public places is negligible. 9,10 In May 1988; organisers of a sympo- sium on ETS in Austria entitled "Illness Due to Passive Smoking?" issued a press release in which they concluded that "a causal relationship between ETS and i1l- ness cannot be established" and that "there is no positive evidence that cancer and other such illnesses are caused by passive smoking".11 In November 1989, eighty scientists and physicians from around the world attended an international symposium on ETS at McGill University in Montreal,. Canada. In the summary of the proceed- ings, co-organiser Joseph Wu concluded that the published data on ETS, "when critically examined and evaluated, do not provide scientific justification for the claim that ETS is a cause of disease in non-smokers":12 In the proceedings from the Interna- tional Conference on Indoor Air Quality in Tokyo, published in 1990, Dr. Hitoshi Kasuga, conference organiser, concluded that "mostparticipants" concurred that the data on ETS and lung cancer in non-smo- kers suggests thatthe "relationship, if one exists at all, is very slight":13 Witorsch suggested that ''a thorough and critical examination of the relevant literature fails to provide compelling evi- dence that exposure to ambient tobacco smoke produces adverse chronic health effects"'.14 Aviado wrote that his review of the literature "indicates that there is no substantial evidence to support the view that exposure to environmental, tobacco smoke presents a significant health hazard to the non-smoker".15 Uberla cautioned that the argument that ETS causes lung cancer in non-smo- kers may "have serious negative effects on the credibility of epidemiology" and may become an "obstacle"'to the anti-smoking lobby's " oai"' of the reduction of active smoking.~6 Uberla based his observation upon a comprehensive review of the studies to date on ETS and lung cancer. He concluded that, at best, the data for establishing an association are "inadequ- ate" and that, consequently, "a causal con- nection cannot be inferred". In June 1990, the U.S. Environmental Protection Agency produced a draft report with respect to lung cancer andrespiratory effects in children for public comment. Although the EPA report concludes that there is an increased risk of lung cancer im non-smokers, due ostensibly to ETS ex- posure, the report remains in draft form and is currently undergoing public review and review by the EPA's Science Advi- sory Board. The principal studies on which the EPA draft report is based are reviewed within this document. In addi- tion this document reviews many studies that were not considered in the EPA draft report.i7 The claim that ETS is a proven cause of disease in non-smokers carries a great deal of emotional weight even without appropriate scientific support. Therefore,, the charges are serious enough to warrant close examination. Lung Cancer Two articles published in 1981 are among the most, frequently cited to sup- port the claim~ that ETS exposure in- creases the non-smoker's risk of lung cancer. A study of Japanese women by Hirayama reported that non-smoking Page 2

Smoke And The Non-Smoker wives of heavy smokers have a greater risk of developing lung cancer than non- smoking wives of non-smokers.1 In a study of Greek women, Trichopoulos concluded that a non-smoking woman whose husband~ smokes has twice the risk of developing lung cancer as a non-smok- ing woman married to a non-smoker.2 Both studies have been widely criticised~ in the scientific literature. Numerous inadequacies and inconsisten- cies in the Hirayama study have been pointed out by noted scientists and physi- cians who have questioned the methodo- logy used, confounding factors, statistical analysis and the validity of its conclu- sions.18'34 Criticisms of the Trichopouloss study related to the small number of sub- jects, because several!tumours lacked his- tologicali confirmation, and because controls and cases were from different hospitals.35 Some more recently published studies are claimed to provide support to the re- ported findings of Hirayama and Tricho- poulos,36'41 but a close examination of the methods used, results and conclusions of these studies reveals that they do not amount to scientific proof that ETS causes ~ disease to non-smokers. For example, a report by Knoth and co-workers which, identified 39 non-smoker lung cancer cases, asserted that exposure to ETS was the "most plausible ex lanation" for the reported lung cancers.i Apparently the study failed~ to provide a comparison group as a control population and one reviewer's assessment was that the study provided "only tentative conclusions based on poor data analysed by unaccept- able methods".42 In a 1983 study, Correa and co-workers reported an increased risk of lung cancer for non-smokers married to smokers in a southern U.S. state.43 The results were based on only 30 cases and did not take into accountoccupational, indoor and out- door exposures or other confounding vari- ables. One of the co-authors of this study subsequently reported no association for a similar population in a neighbouring state.44 In 1985, a published article by Repace and Lowrey estimated that exposure to ETS is responsible for 500 to 5,0001ung cancer deaths a year among non-smokers in the United States.45 Two theoretical models were used to estimate risks of lung cancer from ETS exposures. The first model relied upon a"reinterpretation"' of the epidemiological studies of lung cancer in non-smokers, and the second provided an estimate for lung cancer mortality from a study of the Seventh Day Adventists. Critics have pointed out that the estimates are based on errors and "unrealistic as- sumptions"' which result in overestima- dons of exposure.46,47 For example, reanalysis of the model showed that, depending upon the assump- tions and input data used, such model estimates are inherently inaccurate and may vary as much as 300-fold.48 One critic noted that the exposure and dose levels used were not based on actual measurements, and that actual measure- ments reported~by other researchers range from "ten-to-one-hundred'-fold less than that in the Repace and Lowrey model".49 Others have suggested that they failed to control for other confounding factors, and questioned the methods of analysis used. ~0'52 Moreover the estimate of non- smoker lung cancer risks in the workplace has been criticised by another group of scientists who pointed out that none of the epidemiological studies which examine the relationship between ETS exposure and disease in non-smoking working women reports a statistically significant increase in risk.46 Another study which suggested that the risk for all cancers is increased among non-smokers exposed to ETS was based on samples selected from death notices in a northeastern U.S. newspaper.53 Ana- lysis of the data revealed that the reported risk disappeared when age comparisons of cases and controls were considered.54 Sandler and co-workem reported in- creased risks for lung cancer and cancer of other tissues in non-smokers exposed to ETS.55'57 One critic challenged the authors' selectionof controls and their use of questionnaire information,58 another observed that under one interpretation, the data suggest that exposure to ETS had no effect on overall cancer incidence.59,60 A third critic called the studies "seriously flawed"' and questioned the authors' methods of data collection.61 Others have challenged Sandler's conclusions because the potential confounding variables such as diet, h1liene and lifestyle were ex- cluded.62, Page 3

Smoke And The Non-Smoker The complex nature of the scientific data is further illustrated by the mixed results of some studies; that is, certain data in a study may suggest a relationship be- tween ETS and lung cancer while other data from the same study may not. For example: - In a: first report on a health survey con- ducted in two urban communities in Scotland, Gillis and co-workers re- ported an increased lung cancer risk in ETS-exposed male non-smokers buu found no such increase in females.64 Five years later, a further follow-up of the same populationi cohort reported a positive association between ETS ex- posure and lung cancer in non-smo- kers.6S This association was not statistically significant and the re- searchers acknowledged that, the num- ber of lung cancer deaths among self reported non-smokers, a total of only 9 for men and, women combined, did not provide sufficient, statistical power to detect a risk of the magnitude expected by the researchers. - Kabat andlWynder, in a 1984 study of 25 male and, 53 female non-smoker lung cancer cases, reported a greater amount of ETS exposure at work for male non-smoking cases.b6 No signifi- cant effect was reported for males ex- posed to ETS at home or for female cases exposed eitherat, work orat home. - Garfinkel reported on data from an American Cancer Society follow-up study from 1960-72, involving nearly 180,000 non-srnoking women.67 By comparing the lung cancer mortality rates of women who were reportedly exposed to different levels of tobacco smoke, Garfinkel determined that none of the observed differences were statis- tically significant and that "compared, to non-smoking women married to non- smoking husbands, non-smokers mar- ried to smoking husbands showed very little, if any, increased risk of lung cancer"'. - In 1985, Garfinkel and co-workers pub- lished a study of 134 non-smoking women with lung cancer who were se- lected from hospital records from 1971 to 1981.68 Two types of analyses of the data produced apparently contradictory results. While a statistically significant dose-response relationship was re- ported between non-smoking female lung cancer patients and the number of cigarettes smoked per day by their spouses, no significant relationship was observed between the occurrence of lung cancer in non-stnokidg women~ and the total number of hours per day of exposure to ETS over the last f ve or 25 years. Researchers from Japan reported~ elev- ated risks of lung cancer for non-smok- ing women if the mother or the father-in-law, wwere smokers. However, they reported no significant associ- ations between lung cancer incidence and either spousal smokin or ETS ex- posure in the workplace 6~ In a 1988 case control study of Chinese women, researchers reported a signifi- cant association between lung cancer in non-smoking women and spousal smoking. No significant associations were reported~between lung cancer and exposure to ETS from parents or col- leagues.70 Inoue and Hirayama reported the results of a case-control study on lung cancer in non-smoking women fromitwo cities in Japan. The authors reported a margi- nally significant increased lung cancer risk among non-smoking women whose husbands smoked more than 20 cigarettes per day. However, the data revealed no significant associations for lung cancer in relation to spousal smok- ing of less than 19 cigarettes per day, or for the overall category of "being mar- ried~ to a smoker" (~both exposure ea- tegories combined.) 1 In a highly publicised article, re- searchers from the U.S. reported that household exposure of 25 or more smoker years during childhood and adolescence could double the risk of lung cancer. However, the study noted that exposures to spousal smoking, smoking in the workplace, or smoking in social settings were not significantly associated with non-smoker lung cancer incidence, nor were cumulative exposures during adulthood or over a lifetune.7Z Other researchers have observed no sig- nificant association between ETS and lung cancer. These include, for example, a 1982 study from Hong Kong by Chan and Fung which re- ported fewer 'passive smokers' among Page 4

Smoke And The Non-Smoker lung cancer patients than among con- trols.73 A study of 200 female lung cancer cases among Chinese women by Koo and co-workers concluded that ETS expo- sure had no statistically significant im- pact on lung cancer incidence. No significant relationships were seen for any of a variet of indicators of expo- sure to ETS.74+~5 In a re-analysis of the data on non-smok- ing wives of smokers,76,77 Koo and co-workers indicated that wives of non- smoking husbands had 'healthier' life- styles than wives with smoking husbands; they exhibited higher socio- economic status and had better indices of family cohesiveness and lower fre- quencies of selected health problems and complaints. An important and stat- istically significant difference was found in the diets of the two groups. Wives with smoking husbands con, sumed more processed and spicy foods and ate fewer fresh fruits and veget- ables than did wives of non-smoking menL They concluded that such corre- lates of smoking status "act as import- ant confounders when evaluating health risks among families with smok- ing husbands". Wu, et a] in a study of female lung cancer cases in a large metropolitan area in the U.S., reported in 1985 that the lung cancer risk among the 31 non- smokers in the study population was not affected by ETS exposure.78 In a study undertaken to investigate the occurrence of lung cancer in an indus- trial area in Scotland,79 the researchers found that ETS exposure did not signi- ficantly differ between the cases and controls in the study. In 1986, researchers investigated the potential role of ETS on the incidence of non-smoker lung cancer in Eng- land.80 They observed no significant trends in increased risk among the pa- tients studied. Preliminary results from the on-going American Health Foundation study of 90 non-smoker lung cancer cases indi- cate no statistically significant associ- ations for lung cancer among males or females 4vho were exposed to ETS dur- ing childhood or adulthood (in the home or the workplace).81 - Sobue and colleagues, in a study of lung cancer among non-smoking women in Osaka, Japan, reported no statistically significant associations between ETS exposure during childhood or adult- hood and~ lung cancer among non-smo- kers. The researchers noted, however, a statistically significant increased risk of lung cancer among women who had used wood or straw as cooking fue1s.82 - In 1989, Swedish researchers assessed the relationship between ETS exposure and lung cancer among 38 never- smok- ing women. They reported no statisti- cally significant associations between lung cancer incidence and ETS expo- sure in childhood, in adulthood or over a lifetime. 83 - The largest case-control study to date (439 non-smoking lung cancer pa- tients) was published as a doctoral dis- sertation~ at Yale University in 1!987. The author, Varela, reported no statisti'- callysignificant increased risks for lung cancer among non-smokers exposed to ETS from, spousal smoking, at the workplace or in social settings.84 Perhaps as a result of the failure of individual studies to report consistently, significant results, certain researchers have recently begun using another method of statistical analysis which com- bines the reported results from numerous studies. With this method, called meta- analysis, they claim to have calculated estimated excess risks for non-smokers exposed to ETS that are 10 to 50 per cent greater than those for non-exposed non- smokers. In basic terms, they have reached these broad: or `generalisations' by calculating an aver- age of the relative risks for non-smokers with lung cancer reported by those studies.6X5-87 This method of generalisa- tion has been criticised for its question- able application to those studies, due to the wide variety in their study designs, population selections, techniques of ana- lysis and their results."-90 This method ignores basic methodological weaknesses which, according to one recent U.S. gov- ernmental report, are characteristic of each published study on ETS and lung cancer.91 The majority of studies con- sidered in the various 'meta-analyses' do not report statistically significant in- creases in lung cancer risks for non-smo- kers. For example, only two of the thirteen~ Page 5

I Smoke And The Non-Smoker studies considered by the 1986 National Academy of Sciences Report achieved statistical significance.92 In a recent analysis, Letzel and co-wor- kers considere&only those epidemiologi- cal' studies on ETS which met minimum requirements of scientific quality. The re- searchers performed meta-analysis on the data pooled from the 1,023 possible com- binations of the case-control studies examined.90 Only twenty-four of the combinations gave rise to statistically sig- nificant results. The statistical signific- ance in the twenty-four combinations was dominated by the presence of three studies of questionable scientific merit 93 There- fore the authors concluded that any excess risk had not' been demonstrated and that the studies could' not be used to support the claim that ETS exposure increases the risk of lung cancer in non-smokers. Lee has argued thar the increased risks re- ported in various meta-analyses are the result of an inherent bias in study design rather than the result of an~enuine effect from exposure to ETS.94' Lee suggests that the reported risks cannot be explained on the basis of ETS exposure or dose for the non-smoker and are the result of bias caused by a small number of smokers who are misreported in the studies as non-smo- kers. Other kinds of misclassification may contribute to the reported ~ increase in lung cancer risks among non-smokers, accord- ing to several scientists. For example, none of the studies on ETS and lung cancer provides direct observational in- formation on ETS exposures. In many of the studies the subjects themselves pro- vided the information ~ The questions were rarely about the extent of ETS exposure and normally related to whether or not the wife or another householder or family member smoked. Such estimates may lead to a kind of misclassification, called exposure misclassification 100 which hass been shown by Garfinkel,6$Friedman,101 and others1'Ot-105 to lead to improper in- dices of exposure and incorrect estima- tions of risk. In~ Garfinkel's study, for example, relative risks varied from 0.83 to 0.77 when the women with lung cancer or the husband was the respondent, to a risk of 3.57 when a son or daughter re- sponded.89 Kilpatrick has analysed another form of misclassification, called differential misclassification, which results, in his view, "from the tendency of respondents to inflate the amount of ETS exposure for lung cancer cases and deflate the report of exposure for controls''.100 Wynder notes that "relatives of a non-smoking lung cancer patient are more likely to report passive inhalation exposure on the part of their relative than are the relatives of a control patient":106 Another potential bias is known as "publicadon bias", which is generated by the failure to publish studies which re rt negative or weakly positive results.1008 Scientists have recently expressed con- cern over the growing trend in the lit- erature which tends to overemphasise (and hence publish) only those studies which report increases in risk.109-111 Pub- lished studies which are thus combined for meta-analyses may not truly represent all investigations on the issue of ETS ex- posures and lung cancer. Most of the epidemiological studies on ETS and lung cancer have failed to con- sider diet, occupation and exposures to indoor or outdoor pollution as potential confounding elements. The importance of such factors is underscored by five recent- ly published reports, one each from Japan and the Netherlands, and three from China.112-116 The Chinese and Japanese reports suggest that indoor pollution generated by kerosene heaters, coal, stoves, liquefied petroleum gas and expo- sures to cooking oil vapours may be re- sponsible for the increased risk of lung cancer among oriental women. In addi- tion, relatively few studies have con- sidered ETS exposures outside the home; only four studies sought information about total ex osure to ETS from various sources 68,7a,~5,80 None of these studies reported a significant association between total ETS exposure and lung cancer. In the Netherlands study it was reported that birdkeeping was a significant, inde- pendent risk factor for lung cancer.116 Modelling studies have revealed a dis- crepancy between the relative risk of about 1.5 calculated from epidenziologi- cal studies of lung cancer and the low levels of exposure indicated by cotinine measurements in those exposed to ETS.117 According to the model, a relative risk of 1.5 results from the effec- tive exposure to abouthalf of one cigarette per day. But cotinine measurements indi- PaBe 6

Smoke And The Non-Smoker cate a level of exposure that is between one seventeenth and one fiftK of the amount indicated in the model. Th.e auth- ors conclude that "the relative risks of lung cancer due to passive smoking ... seem to be at variance with the numbers of cigarettes per day equivalent estimated from cotinine measurements".117 Similarly, attempts to estimate risks to non-smokers exposed to ETS which are based upon reported risks associated with active smoking are not warranted due to the vast quantitative and qualitative dif- ferences among mainstream smoke, side- stream smoke, and ETS. The potential toxicity of ETS has been assessed in a number of studies. 118-122 These studies typically tested the body fluids of non-smokers exposed to ETS for mutagens, or substances capable of alter- ing the genetic structure of cells. Al- though it can be argued that the scientific literature on this subject is very limited, and that the laboratory tests employed~are crude, the studies reported no mutagenic activity attributable to ETS exposure inn the body fluids of non-smokers. Animal inhalation experiments using sidestream smoke or constituents of side- stream smoke are also inconclusive. 123,124 German scientists exposed rats and hamsters to very high levels of side- stream smoke during a 90-day inhalation experiment. 123 The researchers reported no significant physiological effects on the tissues of the animals. In his comprehens- ive review of the literature on suspected pulmonary carcinogens, Dr Domingo Aviado observed~ that none of the consti- tuents in sidestream smoke whi& have been identified as potentially carci- nogenic has induced pulmonary cancer in a12namals under experimental conditions. Given these difficulties in interpreta- tion, it is therefore not suprising that a statistician~ should conclude that "it is un- likely that any epidemiological study has been, or can be, conducted which could permit establishing that the risk of lung cancer has been raised by passive smok- ing. Whether or not the risk is raised re- mains to be taken as a matter of faith according to one's choice".107 Other Respiratory Effects 1. `Healthy' Adults Most of the data available show that the effects of ETS on air flow function of exposed adult non-smokers vary from negligible to quite small and overall' the data fail to support the claim that exposure to ETS causes pulmonary disease in non- smokers. The suggestion of a link be- tween ETS and respiratory, disease in non-smokers was raised in an American study by White and Froeb in 1980.3 They measured the small airways function of smokers and non-smokers and concluded that non-smokers exposed to tobacco smoke at work for 20 or more years had reduced function of the small airways compared to non-smokers not so exposed. Similarly, in 1983, French researchers re- ported that non-smoking spouses over 40 years of age who were married to smokers exhibited decreases in pulmonary func- tion compared to non-smoking spouses of non-smokers.125 In 1989, a study of a Scottish population reported significantly lower pulmonary function among non- smokers who lived with smokers com- pared with non-smokers who were not exposed to ETS.65 The American study has been criticised for a number of reasons. 126-133 For example, Lebowitz concluded that the study was "improperly designed"'from an epidemiological point of view:131 There were problems "inherent in the study", including the selection of the study group, the proper measurement of smoke in the workplace and the statistical significance of the data which appeared to depend on the unexplained omission of data for 3,000 people who were originally in- cluded in the study.131-133 Lebowitz wrote: "Even with a biased population, poor study design, and incorrect statistical evaluation, there were no clear-cut, con- sistent, medically meaningful differences between passive smokers and groups of non-smokers; a corrected statistical ana- lysis strengthened this conclusion."133 Likewise the French study, which re- portedly showed significant differences in lung function in one part of the study population, was criticised because there was no significant difference between ex- posed and non-exposed non-smokers in the population as a whole.133 As the Page 7 O N ~ O J~LL

Smoke And The Non-Smoker "healthiest" population in the study appar- ently lived~ in the most polluted areas, it was suggested~ that the study, may have been flawed due to biased population se- lection~ and testing or other confounding factors: Neither study is consistent with other research on lung disease and lung function in adult non-smokers. For example, a 1984 study of 1,351 German office workers reportedly found no effect on pulmona function among exposed non-smokers34 In a 1988 update of the study, the investigators noted that "there is no evidence that average everyday passive.smoke exposure in the office or at home leads to an essential reduction of lung function in~ healthy adults".135 A group of epidemiologists (at the Johns Hopkins University School of Medicine) reponed~ thao in a sample populatnon of 1,724 adults; the frequency of respiratoryy symptoms in non-smokers was not asso- ciated with the number of smokers in the household.136 Nor was the frequency of impaired ventilatory function significant- ly higher if there were stnokers in the home. However, their analysis did show, that "among men who never smoked ciga- rettes, gas cooking was definitely associ- ated with impaired ventilatory function, even when corrected for multiple com- parisons". Other studies on pulmonary function, respiratory disease and environmental to- bacco smoke are pertinent. For example: - Jones and co-workers reported that in a case-control study of several hundred non-smoking women from a U:S. study populations there was no significant as- sociation between~ decreases in lung function and exposures to ETS in the home.137 - Co-ordinators of an epidemiological study of obstructive lvngdisease in Ari- zona reported finding no effects from ETS ex osure in their adult study popu- lation.1~8 - In a study of 376 families, Yale Univer- sity scientists reportedly found no evi- dence that environmental tobacco smoke affected either lung function or symptoms in adults.139 - In a Canadian study on the effect of substances in indoor and outdoor air on the lung function of housewives, re- searchers concluded that ETS exposure did not influence lung function but that gas stove usage played a role in average lung function decline.140 These studies and others were re- viewed at a U.S. National Institutes of Health workshop on ETS expooure where it was concluded that a possible effect from environmental tobacco smoke "varies from negligible to quite small".141 2. Asthmatics There is no convincing evidence thati inhalation~ of ETS causes the disease called asthma. Exposure to ETS may pro- voke a reaction within the airways in some individuals who are already asthmatic or who may have other variants of reactive airways disease, but even that is uncertain. When such reactions appear to occur, it is unclear whether this is due to an irritating effect of smoke on the airways, a non-air- way triggering effect uponi the airways or is psychological in origin. In some people withi asthma, even the sight of tobacco smoke may produce a reaction, in~ the same way as a reaction can be produced in some asthmatics by particular colours. However, there is no evidence which clearly indicates that a true allergic reac- tion occurs as a result of the inhalation of ETS. A 1981 study by Dahms and co-wor- kers reported decreases in~the pulmonary function of several asthmatics ex osed to environmental tobacco smoke.12 How- ever, the subjects were challenged under the unrealistic experimental conditions of exposure to hi'gK levels of smoke in an enclosed chamber. The authors noted their experiment lacked proper controls and the changes observed may have been due to psychological, not physical factors: A Canadian group reported that respir- atory data collected from a group of as- thmatic volunteers exposed to tobacco smoke "do not suggest that asthmatic sub- jects have an unusual sensitivity" to ETS exposure.143 Although several volunteers claimed that they experienced wheezing and tightness in the chest due to the expo- sure, the researchers found that the "physiological data gave little support to the concept of a subgroup with particular sensitivity". They noted that these reac- tions probably were due to the "suggesti- bility" of the subjects. The difficulty in determining the im- pact of psychological responses in such Page 8

Smoke And The Non-Smoker studies is clearly demonstzated by the re- sults of two Australian studies. A 1985 report suggests that short-term ETS expo- sures are ca able of inducing reactions in asthmatics,~" but a 1983 study tends to support the theory that psychological re- actions may partially explain symptoms observed during such exposures. In the 1983 study, Ing and Breslin reported that while asthmatics exposed to tobacco smoke complained of subjective symp- toms, no significant measurement of air- ways obstruction was observed.145 A report on results from a Iarge-scale epidemiological study in the U.S. sug- gests that ETS in the home does not affect symptoms or pulmonary function in either child or adult asthmatics, although dust and pollen in the home apparently can provoke such effects.146 A group of researchers from Yale University studied the short-term effects of ETS exposure on~ a group of young asthmatic patients.147 They observed no changes in lung flow rates and concluded that such exposures present "no acute res- piratory risk" to asthmatics. 3. Children No health claim regarding ETS is ca- pable of provoking stronger feelings than the charge that parents who smoke harm~ the health of their children. The issue of children and parental smoking is laden~ with emotion, and the scientific basis for the claim is difficult to interpret. Some studies examined respiratory symptoms or illness such as coughs and colds b questionnaire responses from parents,l~ others measured lung function with spe- cial equipment at a school or health fa- cility~49 Such studies, each with a different sample size, data collection method and analysis, have led to varying conclusions.141 For instance, although certai>~ 54 66 have reported adverse find- ings 4 others have observed no significant relationship between parental smokin and respiratory illness in child- ren.139,~48,167-17'9 For example, after a five year study of over 400 children, Dutch researchers concluded that there was "no evidence" that parental smoking had an appreciable effect on respiratory symptoms in school children.172 The authors of studies reporting ad- verse effects from ETS exposures among children concede that their conclusions must be viewed with caution because of numerous confounding factors. For example, one group of British researchers acknowledged the possible influence of factors such as cross-infection in the home and genetic susceptibility to childhood respiratory illness and syrnptoms.154,155 Researchers in Hong Kong reported "a highly, ssignificant correlation between the frequency of respiratory illnesses" of mo- thers and their children.180 Others have conceded that the reliance on question- naires for information about respiratory symptoms casts doubt on the findings. In one study which reported a significant association between parental smoking and respiratory symptoms, for example, it was noted thateven "slight changes"'in the way the questions were phrased~could re- sult "in substantial differences in the type of responses one obtains":156 Similarly, another study observed that there was a significant difference in the respiratory symptoms reported depending on which parent completed the questionnaire. 164 The contradictory nature of reported~ findings on the issue of parental smoking is also apparent in the growing number of studies examining the relationship be- tween parental smoking and children's respiratory or lung function. Although several reports have claimed that parental smoking results in decreased pulmonary lung function in children 152,156,166,18I- 187 others have not.139,Y88-191 In 1982, for example, Lebowitz and co-workers suggested that a comparison of body size with lung function eliminated any re- ported correlation between parental smoking and children's lung function,149 Two years later, a reanalysis of families from their study group again showed that "parental smoking did not have a signifi- cant effect on children's pulmonary func- tion; smoking habits of others in the household (predominantly siblings) did not have any effect either".138 In 1988, investigators re-examined thirty studies on ETS exposures among children and evaluated the studies for their scientific validity.192 They noted that while several studies of adequate scien- tific design had reported a statistically significant relationship between ETS ex- posures and respiratory illness in children,. "most studies had significant design prob- lems that prevent reliance on their conclu- sions". The authors conclude that "many Page 9