Tobacco Documents Online

REGULATORY T9xICOLO(Il ,YND PHARMACOLUGY`L1, 2$1-29.3 .(199.5) .. ~hi~ f°rotec ed ~lal n~, be aw rtitle 17 ~.S~~e) t Environmental Tobacco Smoke and Coronary Heart Syndromes: Absence of an Association Gio BATTA GORI Thr.Health Pnhc..C"erzter: fi7r7-t Barr Roaq'. Bethesda. .'4farvland.'20816-t016 Received September 14, 1994 the tobacco industry, this present review strives for a Concerns about possible cardiovascular and espe- comprehensive evaluation of available knowledge by cially coronary effects of environmental tobacco smoke avoiding assumptions and standing by the evidence. (ETS) derive from the reported effects of active smok- In considering ETS as a possible CHD risk factor it ing. Despite similarities, however, ETS._has composi- .... will. be useful to add.r..ess what is known of the chemical, tion and physical characteristics different from the physical, and.bioiogical comparability of ETS and the mainstream smoke (MS) that active smokers inhale and smoke that...smokers inhale.. Exposures and doses will appears relatively more chemically inert and less bio- then be compared in light of the:no-observable adverse logically active. ETS doses to nonsmokers are small and .- effect levels for active smoking and CHD, as reported in often below the sensitivity of detectlon technologies. the literature.. Finally, the epidemiologic studies of ETS They are several orders of magnitude less than MS e.xposure and possible CHD risk are evaluated against doses in active smokers. Numerous epidemiologic stud- the background of numerous confounders, difficulties in ies report that the active smoking of less than 10 clga- ,.. establishing and measuring exposures, classification and rettes/day is not associated with measurable risk of cor- other loglstic biases, statistical criteria of significance, onary heart disease (.CHD)....Thus, even assuming that _ and inferential conjectures from in vitro and in viuo ex- ETS and MS have equivalent biologic activities, con- ceivable ETS doses to nonsmokers are far below appar- periments and clinical effects in humans. In a public health context this review should be seen ent no-effect thresholds for active smoking. Hence, it is no surprise that epidemiologic reports are inconclusive as purely a - risk assessment exercise. The National Acad- about a possible association of ETS.exposure and CHD, emy of Sciences determined that risk assessment should some suggesting a slight elevation, others a reduction of be an objective scientific analysis distinct from the judg- risk. Often, the elevations reported are higher than the mental risk management policies that may follow (NAS, 1983). Accordingly, this review follows standard Prlnci- CHD risk values associated with active smoking. ~uch equivocations likely result from the presence of con- ples of the scientific method. trasting protective or aggravating confounders, of . Most epidemiologic° studies of chronic multifactorial which more than 200 have been reported in the litera- ' diseases are not adaptable to the scientific method. Ob- ture-confounders ture-confounders that were not and could not be ade- . servational studies-the basis of ETS epidemiologic quately controlled by any epidemiologic study. By sci- studies-encounter especially vexing logical difficulties entific standards, the weight of evidence continues to in their interpretation, fully recognized by epidemiologic falsify the hypothesis that ETS exposure might be a theory. For instance, in his authoritative analysis Roth- CHD CHD risk factor. c 1995 Academic Press. Inc. man writes: Desaite philosophic injunctions concerning inductive inference, - INTRODUCTION r~(ena have commonly been used to make such inferences. The 3ustification offered has been that the exigencies of public health problems demand action and that despite imperfect knowledge ; . Several reviews have attempted to appralse.the litera- causal inferences must be made. (Rothman, 1986, p. 17). _ ture on environmental tobacco' smoke (ETS) and coro nary heart disease (CHD) (Glanz and Parmley, 1991; Clearly, the exigencies of public health represent im- Taylor and Johnson, 1992; Steenland, 1992; Wells, ' peratives other than scientific, raising the question of 1994). In general these reviews have heen se)ecti~e and what rules of evidence to adopt if the scientific method conjectural and have failed to account for the many per- does not apply. The question has been resolved in a set tinent considerations that a scientific evaluation re- of judgmental guidelines-the criteria mentioned by quires. Although writte..n, by a long-time consultant to Rothman-initially proposed in the first Surgeon Gen- 281 0273-2300/95 $6.00 Copyright G 1995 by Academic Press. Inc. All rights ofreproduction in any form reserved.

1 I I I I I I I I I I 282 GIO BATTA . GORI eral`s report on smoking and expanded shortly after by .. a.bsorptions to environm.e. ntal. surfaces, and , other Sir A. Bradford Hill (USSG, 1964; Hill, 1965). They changes (NAS, 1986; USSG, 1986; USEPA, 1992c: Gue- _ are now familiar considerations used to extend judg- rin et al., 1987; Baker and Proctor, 1990). The reducing mental inferences of causal associations in a metas- capacity and free radicals of MS are lost within minutes cience context: strength, consistency, specificity, tempo- ':(Schmeltz et al., 1977, Tanigawa et al., 1994), and ETS rality, response gradient, plausibility, coherence, anal- is considerably less cytotoxic than inhaled MS (Sonnen- ogy, and experimental evidence. feld and Wilson, 1987). On this basis, judgmental inferences of causality have Of the several thousand components identified in ...... _ ..... been advanced in situations that substantially met these mainstream smoke, only i00 or so have been detected in qualifiers. However, other difficulties in the epidemiol- sidestream smoke under field conditions, due to extreme ogy of multifactorial diseases have identified additional dilutions. Because of even greater dilution, only about ...:. . . . obstacles to causal inferences, also described by Roth- 20 ETS components have been identified directly under man and well known to epidemiologists. These are struc- field conditions. In natural settings, most ETS compo- tural problems deriving from the inevitable biases in nents are below the sensitivitv of current analytical ca- _ .. data collection and the logistic impossibility of control- pabilities (Guerin et al., 1987; Baker and Proctor, 1990). ling for multiple confounding in multifactorial diseases. Compilers of ETS reports from the National Academy The resulting uncertainties in tiiany studies often pre- of Sciences (NAS, 1986), the U.S. Surgeon General vent even the application of the "causality criteria and, (USSG, .1986), and the Environmental Protection therefore, of causality inferences. Thus, even more elas- Agency (USEPA, 1992c) have been forced to infer the tic ways of appraisal have been sought in what is known presence of ETS components by proxy, based on the as the "weight of evidence" approach (NAS, 1983). In composition of the sidestream smoke from which ETS principle, this approach entails a lose qualitative inte- primarily derives. gration of what evidence may be available, sometimes Nominally, then, ETS and mainstream smoke may sufficient to justify holding onto a research hypothesis -'share some components, but their chemical and physical but clearly inadequate for its verification. On the other differences are substantial. Moreover, the presence of hand, contrary weight of evidence may. be sufficient .to": most ETS components can only be postulated because falsify and discard a research hypothesis on the basis of they are beyond material detection. Also, the chemical reliable indirect evidence. The latter is precisely the si..t ... an.d biologic reactivity_of ETS is less than that for the uation in the matter of ETS and cardiovascular diseases. MS that active smokers inhale, because of the loss of free radicals, other quenchings, and absorption losses duririg ETS AND ACTIVE SMOKING dilution and aging. ETS comes from the dilution of sidestream smoke ESTIMATING ETS EXPOSURE produced by smoldering cigarettes an:d from the small residues of mainstream smoke (MS) exhaled by active A major limitation of epidemiologic studies on ETS smokers. Generated and existing under much different has been the unreliable estimates of dose, which com- conditions, these different smokes have some similari- pound the uncertainties of personal or proxy recall of ties but marked differences in chemical and physical the intensity, frequency, and duration of exposures over composition and behavior. All comprise aa gas phase and individual lifetimes. Even the simple dichotomous clas- small respirable suspended particles (RSP). These par- sification of exposed and nonexposed subjects presents ticles in turn may contain at various times different recognized uncertrtaintie.s, such as those deriving from the self-classification of some smokers as nonsmokers or the propensity to exaggere ate exposure estimates by public- ity-.sensitized patients (USEPA, 1992c; Lee, 1992, 1993). On comparatively more solid grounds, a range of proba- ble momentary exposures to ETS can be inferred from physical and chemicaI. derivations. These inferences are also insufficient to determine cumulative exposures, but raise compelling doubts about the reliability and mean- ing of epidemiologic estimates. On the basis of extrapolations from sidestream and mainstream smoke data, the National Academy of Sci- ences calculated that for nicotine alone the difference in peak inhalation concentrations between smokers and dergoes oxidative and photochemical transformations, ETS-exposed nonsmokers varies between 57,000- and ... . . polymerizations from loss of water and volatiles, reac- 7,000,000-fold (NAS, 1986). Dose estimates based on tions with other environmental components, differential body fluid concentrations of nicotine or cotinine yield amounts of water and other volatile components that may exchange with the gas phase. Mainstream smoke-inhaled directly by smokers-is concentrated and confined to the moist environment of mouth, throat, and lung. Its higher gas-phase concentra- tions favor larger respirable particles that condense and retain more water and volatiles. By contrast, ordinary ETS is over 100,000 times more diluted, with much lower humidity and extremely low concentrations of volatiles. Evaporation is faster from ETS particles, which- within fractions of a second from their generation-at- tain sizes 50 to 100 times smaller in. mass and volume than their mainstream counterparts. As ETS ages, it un-

PASSIVE SMOKING.AtiD CORONARY HEART SYNDROMES. 'T83 higher values, but depend on environmental and phar- TABLE 1 _ _ _ _ macokinetic assumptions of unlikely validity (USEPA, Relative Dose Estimate of Respirable Suspended _ _ . :...... _. .... ... _ ........ _ _ 1992c; USOSHA, 1994). They also depend on a set of - Particulates (RSP) in Typical Active Smokers and equally dubious syste..ms atic .a.ssumptions as listed by the ETS-Exposed Nonsmokers I I I 11 I I I I I National Research Council: "Current smoking patterns Active smoker reflect past patterns. Cotinine"or nicotine concentra- tions . . . are linearly correlated to recent exposures to ETS and to the carcinogens in ETS_among no.ns:moke.rs 30 cigarettes per day 15 mg RSP inhaled per cigarette 90% lung retention efficiency Daily dose about 400 mg All subjects in the various studies began to be exposed to ETS-Exposed 0.05 mg RSP/cubic meter of air ETS at the same time and have continued to be exposed Nonsmoker ia hr per day exposure° at the same rate throughout the follow-up period. ..." 0.7 cubic meters per hour inhaled (NAS, 1986, p. 290). 10% lung retention efficiency Estimates of exposure to other ETS components are ~a11V UUSC QUVUI V.VVilGO In~,' even more problematic because of numerous sources e.x- Crudedose._ratio 0,00525:400 about 1:75,000 ternal to ETS. For instance, plasma concentrations of Lung surface permeability some three times greater in smokers volatile organics in nonsmokers appear to be as much as Lung clearance some three times more efficient in nonsmokers ETS dose distributed over greater surface deeper in lungs two-thirds of the corresponding levels in active smokers Net dose ratio at target tissue <1:500,000 IAngerer et at., 1992; Brugnone et al, 1992; Perbellini et al., 1988)-an indication of significant sources other ° USOSHA 1994; Emmons et at.. 1992. than tobacco combustion (Richter et al.,. 1994; Ong et ..a.l., 1994). By utilizing surrogate..sidestre.am bmoke values, con- Overall, these considerations lead to the conclusion ceivable ETS exposure has been compared with current -.:that the prevalent ETS-RSP dose is minuscule. Al- federal standards of permissible occupational exposure though difficult to define, Table 1 shows that it is likely to several smoke components. Considering an unventil- at least 100,000 tinies smaller than the mainstream ated room of 100 m3 (3533 cubic feet)., the.. number of : smoke dose in active smokers, as official EPA reports cigarettes that would have to be burned before reaching acknowledge (USEPA, 1992a) For the.average ETS-ex- otFicial threshold limit values varies among 1170 for posed individual, this estimate translates into an annual methylchloride to 13,300 for benzene to 222,000 for ben- dose equivalent to far less than the mainstream RSP of zo(a)pyrene to 1,000,000 for toluene. (Gori and Mantel, 1.. cigar..ette.. evenly dispersed over a 12-month period 1991). (Gori an.d. .M. .a.ntel, 1991). In any event, the measurement of ETS respirable par- Documented observations-also plainly perceived in ticles (ETS-RSP) has been more fruitful than the mea- everyone's life and experience-indicate that people surement of single chemical species. Methods have been have innate capacities to cope with multiple low-level devised that separate particles that may derive from exposures. The question, then, is whether very small ETS and other sources. Use of these methods yields the, doses of ETS pose plausible risks to nonsmokers. current consensus=supported by EPA's reports on ETS-that prevailing concentrations of ETS-RSP are COULD MINUTE ETS EXPOSURES below 50 µg/m3 in households with smokers, the envi- POSE A HEALTH RISK? ronments studied in most epidemiologic studies that have suggested elevated risks (USEPA, 1992c; Gori and Because direct measurements of the biologic activi- Mantel, iVIantel, 1991; Samet, 1992; Steenland, 1992), ties, exposures, and doses of ETS are so problematic, ini- Because of aerodynamic size and other differences, tial attempts have inferred ETS-linked health risks by EPA recognizes that only about 10% of inhaled ETS- arithmetic derivation from the apparent risks associated _.:::: , RSP may be retained by nonsmokers, compared to with active smoking. However, this approach has been nearly 90% for mainstream smoke RSP in active smok- controversial (USEPA, 1992b). A review by the EPA ers (USEPA, 1992c). Furthermore, lung cl..e. arance... is (USEPA,.1992b) dedicates a chapter to the proposition faster and more efficient in n.onsmok.es rs_than in_smokers that. ". ., due to the similarity in chemical composition (Kennedy et al., 1984:; VAstag et al., 1985; Foster et al., between [mainstream smoke] and ETS and the known 1985; Zayas et al., 1990; Gerde et al., 1991), and target human exposure to ETS. .., ETS would also be classi- cell doses are far smaller because of greater lung surface fied as a . . . human carcinogen" (pp. 4-10). Elsewhere, at the greater depths reached by ETS-RSP (Mercer and the review lists the many differences of MS and ETS, Crapo, 1993). Recent studles (Emmons et al., 1992) en- isuggesting that risk extrapolation from active smoking dorsed by the Occupational Safety and Health Admin- may not be feasible (pp. 2-7). A direct comparison also ~ istration indicate that the average duration of daily has been questioned in a review of the association of (= exposure to ETS is around 1.5 hr, with an upper confi- ETS exposure and CHDs, citing the obvious contrasts CA dence interval of 2 hr (USOSHA, 1994). (Steenland, 1992). In reality, the only reasonable infer- ftJ ~ . . ~~ ~ ~ ~

284 GIO BATTA GORI I I I I I I I I ences would come from the foregoing estimates. Accord..were the same in nonsmokers and in smokers of 1-14 ingly, the health risks of ETS-if any would have to be . cigarettes/day. My ortality for all diseases was actually so much smaller as to be unmeasurable, when compared slightly lower in smokers of 1-14 cigarettes/day than in to MS-associated risks. nonsmokers. Moreover, epidemiologic studies of active smoking For male British doctors, mortality rates in the group give evidence of, no.-observable-adverse„-effect-levels smoking 1...-14 cigar.e. ttes1...d. ay were slightly elevated after (NOAELs), namely that at low daily consumption of cig- 20 vears of observation (Doll and Peto, 1976). However, arettes the epidemiologic risks associated. with certain ' extrapolation from doseJresponse functions indicates diseases become nonsignificant (Gori, 1976; Gori and that mortality for cardiovascular diseases would not be Mantel, 1991). No-effect observations at coinparatively elevated for smokers inhaling an evenly dispersed daily high doses are also routinely reported in experimental dose equivalent to 4-5 pre=1960 cigarettes (Gori; 1976; _ ,:.. . animal exposure to whole smoke or its fractions. In a Gori-Mantel, 1991). recent evaluation of smoking and health issues, the Con- - A large cohort study of white women in Maryland re- gress.ionaI Research Service of the Library of Congress ported threshold effects (Bush and Comstock, 1983). ...... ........... ...... ... ... . ........ _ stated: The Western Collaborative Group Study also reported The existence of an exposure threshold for disease onset.below which many passive smokers fall is not implausible. Most organ- isms have the capacity to cleanse themselves of some level_of con- taminants. It is for this reason that public polrcv usually does not clear threshold values, especially after adjustment for type A personality confounding (Jenkins et al., 1968). - Moreover, many studies would have likely reported threshold values if their lowest exposure. category had been 10 0 or less cigarettesJday (Oliver, 1974; PPRG,G . . insist that everv unit of air o.rr water pollution be removed from 1978; Rosenberg et al., 1985). the environment. . . . In fact, strongly nonlinear relationships m which health effects rise with the square of exposure, and more, . In a more circumstantial context, a study by the Na- have have been found with respect to active smoking (see Surgeon Gen tional Center for Health Statistics found that patterns eral's Report, 1989, p. 44). Were these ielationships projected of heart diseases and ischemic heart diseases did not par- backwards to construct the lower (unknown) portion of the health . allel patterns of smoking_habits in different U.S. regions effect/physical damage function, the observed relationship might lead researchers a priori to expect no empirical relationship. Thus, (Gillum, 1994). Another recent study reports that ciga- the issue raised by this potential break in the causative chain is whether researchers should expect to find a significant relation- ship between passive smoking and health effects. iGravelle and Zimmermann, 1994, p. 45). rette smoking fails to explain international differences in mortality_for chronic obstructive pulmonary diseases (Brown et al., 1994). The worldwide and massive WHO- MONICA study recently concluded that cigarette srffok- ing, hypertension, and total cholesterol "measured The presence of NOAELs for active smokingg would cross-sectionally at the population level do not reflect have a disposing relevance in the evaluation of claimed well the variation in mortality between populations" CHD risks of ETS exposure. A compendium of 34-year (MONICA, 1994). Much earlier, the large Seven Coun- follow-up data was recently published for the prospec- tries study determined in the '60s that smoking was not tive Framingham study, the longest and most closely associated with excess CHD mortality (Keys, 1980). In monitored epidemiologic study of its kind in the United the Finnish North Karelia-Kuopio study, female CHD States (Freund et al.; 1993). The report states: "For all mortality declined 68% despite an 80% increase in CHD a clear relationship exists only for younger men. smoking prevalence between 1972 and 1992 (Vartiainen Women may have a slightly increased risk below age 65.,. et al., 1994). while no relatioliship is seen for men or women past age To these reports one should add the generally ac- 65." However, for any smokers of 1-10 cigarettes/day cepted evidence that moderate pipe and cigar smoking the age-adjusted rates are generally below the rates of. are not associated with increased risks of lung cancer nonsmokers. Age-adjusted rates are also unchanged for and cardiovascular and respiratory illnesses. On this ba- lung cancer in any smokers of 1-10 cigarettes/day below sis, the first Surgeon General's report on smoking and age 65. These reports are consonant with previous re- health, and others, concluded that nicotine and carbon ports from the Framingham trial (Kannel et at., 1987). monoxide are unlikely to have adverse cardiovascular : A large Swedish study also reported no correlation be- effects, since their blood levels are quite similar in ciga- tween smoking and cardiovascular illness (Lapidus, 1985). The British Doctors study in England is widely regarded as the best continuing study outside the United States and perhaps the best in the world. A 1980 paper reported on mortality rates in female British doctors af- ter 22 years of observation (Doll et al., 1980). There was no increase in mortality rates at any level of daily. ciga- rette consumption for pulmonary heart disease. For ischemic heart disease and lung cancer, mortality rates rette, pipe, and cigar smokers (USSG, 1964; Wald et al., 1981). In regard to nicotine, this conclusion was also echoed officially by the Independent Scientific Commit- tee on Smoking and Health of the United Kingdom (Froggatt, .1988). Such considerations directly oppose the conjecture that ETS is a CHD risk factor, given that pipe and cigar smokers must be among the subjects most substantially exposed to ETS. The combined evidence of epidemiologically derived r ~

PASSIVE SMOKING AND CORONARY HEART SYNDRONIES 285 NOAEL thresholds for"active smoking and CHD sup- vascular risks and threshold levels, as reviewed above. ports the a priori inference that typical ETS exposures The paradox has not gone unnoticed, the reported risk could not represent a CHD risk. in active smokers being only slightly higher than the ap- parent risk froni ETS studies, despite vastly smaller EPIDEMIOLOGIC STUDIES ETS doses (Steenland, 1992). The closeness of the reported risk values for ETS and Of the published studies on the possible association of active smoking. indicates either the unlikely hypothesis ETS exposure and CHD risk, most are not statistically of extreme differences..in the specific biologic potencies _ _ significant. As could be expected, Fig. I shows that the of ETS and mainstream smoke or the more plausible instabilitv of results is more pronounced in studies of likelihood of interferences from biases'and confounding small sample size. On the other hand, the larger stud , risk factors other than ETS (Steenland, 1992). In fact, ies-derived from the databases of the American Cancer the role of confounders is certain, given that only a few Societv and of the National Center for" Health .Statis- of the nine studies of ETS and cardiovascular diseases tics-have massive power and do not sustain an associa- have controlled for at most 2 or 3 of the over 250 CHD tion of ETS exposure and CHD risk. risk factors reported in the literature (Hopkins and Wil- In the smaller studies, the sample deficiencies amplify .,liams, 1991). the effects of biases and confounders and thus the vari- A defensible metaanalysis exercise requires a reason- ance of the results. CHDs are of multifactorial origin and able degree of homogeneity in study design, subject entry unless all significant factors can be accurately con- criteria, interview questionnaire and procedures, bias trolled, the noise-to-signal ratiot would:be large and re- and confounder controls, disease and mortality diagnos- sults suggesting slightly increased or decreased risk tic certification criteria, statistical methods and rules of would be uninterpretable. In fact, the.weakness of the interpretation, and other variables. Without some rea- risk signals precludes not only causal conclusions, but sonable homogeneity, metaanalysis estimates impose even the formulation of reasonable research hypotheses. too many conjectural assumptions and become question- ........ Taking clues from the first Surgeon Gene.r.al's report able exercises in numerology devoid of scientific content (USSG, 1964), several studies..have shown that smokers .. and justification. In fact, even a cursory analysis shows in general display lifestyles that include peculiar risk that it is virtually impossible to certify minimal stan- factors _ factors other than smoking; for instance, they'_may exer .....dards_ of homogeneity for the available ETS-CHD cise less, consume more alcohol, have less healthy diets, studies.: ; ,. and so on (Margetts and Jackson, 1993; Cress et al., - , 1994). Also, the literature shows that these less healthy Hirayama (1981, 1984, 1990). This cohort study-of habits and risks eventually extend to nonsmoking mem- nonsmoking Japanese women suffers from critical flaws. bers of a household (Gori and Mantel, ..1991). The rela-_ . It did not control for classical CHD risk factors and de- tively low elevation of CHD risk for active smokers be- termined smoking status only at enrollment time in 1965 littles the impact of smoker/nonsmoker misclassifica- but not during the 15-year follow-up. The author admit- tion, but the instability of results is still enhanced by ted having incorrectly calculated relative risk values in interview, exposure assessment, publication; and other different papers (Hirayama, .1990), leaving unresolved biases present in virtually all of the epidemiologic re- discrepancies between the 1981 and subsequent reports. ports cited. For all such and other reasons, epidemiologic The overall results did not achieve statistical signifi- studies-especially small ones-suggesting ari associa- cance. Follow-up losses are likely to have been excessive tion of ETS exposure and CHDs are doomed to produce (Layard and Viren, 1989). the equivocal results reported in Fig. 1. Published metaanalysis estimates of the combined ex- Garland et al. (1985). This study of a middle-class cess risk from selected ETS-CHD studies have produced white cohort from Rancho Bernardo, California, deter- typical estimates of risk around 1.3. Authoritative text- mined smoking status, plasma cholesterol, obesity in- books books and commentators have warned that relative risk dex, and°systolic blood pressure only at enrollment time values less than 2 or 3 may not be interpretable, espe but not.during the 10-year follow-up. The reported age- cially when studies are affected by multiple uncertain- adjud sted mortality rate for never-smoking wives of cur- ties, and depend on simplistic working hypotheses rently smoking husbands was based on 2 deaths only and (Wynder, 1987; Rothman, 1982; Breslow and Day, was smaller than the corresponding rate for wives of for- 1980). In fact, some individual ETS-CHD studies report mer smoking husbands (2.7 vs 3.6). no change in risk or reduced risk, inverse`dose/response gradients, or relative risks for ETS.exposures that are of Lee et al. (19$6). A case-control study from England the same or higher magnitude as risks. for active smok- generally reporting no association of CHD and ETS ing. The apparent risk of cardiovascular diseases for' ac exposure, is described. The study suffers from exposure tive smokers is only 1.7 (Steenland, 1992), and even for characterization imprecision and does not control for active smokers there are solid reports of negative cardio CHD confounders.

286 Author SEX GIQ BATTA GORI ......... . ..::... ........... Decreased Risk Increased Risk CASES 4 LvVois, CPS-I H 7775 LeVois, CPS-X P 7233 LeVois, CPS-II Y 1966. LaVois, CPS-II F 1099 Helsinq F 988 Layard F 914 Hirayama F ~19{ Layard M 475 HelsinQ M 371 Dobson M 183 Dobson F 160 Hole MP 84 Butler t,ee Humble P F 80 7 76 LaVecchia M 69 He, 1994 Jackson P 59 61 49 Lavecchia F 44 Lee H 41 He, 1989 P 34 2.50 1 1..11 -t 1.05 1.00 t- 1.08 1.15 -j- 1.08 1.14 -t- 1.14 1.04-~-+-- 1.36 1.19 TT - 1.16 1.06 - +--- i.42 1.37 --- -- 1.28 1.05 --•--- 1.64 2.00 --------+- --------- 1.86 1.47 ----- I ----'---------------- 4.13 1.2 ----- -r------------- 3.35 i.61. ------ ------- 1.72 . 1.78 -------- t------- 1.69 0.99.{----- * ---------- 2.57 ............. ..:................... .................:.:..... .....::::...:. . ... .. . ------------------- 2.97 .... ..... ...... . . . 1.79 -------- ------------------ 2.77 ---- ------- -------------------- 3.00 2.00 ---------- -•----------------- 2.87 1.56 ------ -' -------------- 2.08 6 Martin F 23 1.2 -------------- •------------------------- //- 5.70 Jackson P 20 i. 5--------------------------- --------------- //- 13.1 Garland F 19 1.25 -- ------------------------•----------------------//-15.3 svendseII H 13 1.04 ------ `"------------ 2.711 Palmer P r FIG. 1. Reported increased and decreased risks of heart diseases in nonsmokers exposed to environmental tobacco smoke listed in order of decreasing sample size. Cases are exposed and unexposed combined. The logarithmic analysis leading to relativee risks (RR) or odd ratios (OR) confines reduced risk values to the interval from 0' to l, while increased risk values span the interval from 1 to infinity. The correct visual representation requires that reduced and increased risk values be displayed on equivalent scales. The symmetry is achieved by giving reduced risk values as the reciprocals of the less than 1 values obtained logarithmically and.as such reported in published studies. The central vertical line represents the null risk value (value 1). Decreased risk values are to its left, increased risk values to its right. For each study cited, the asterisk marks the midpoint estimate and the dotted lines span the domain of the 95% confidence interval, if reported in the cited publications.

I I I I ~ I I I I r I 1 I I PASSIVE SMOKING AND CbRONARY HEART SYNDROMES ... ... `S0 7 ... .. ..... Martin et al. (1986) (abstract). An abstract of a small _ cohort showed a nonsignificant CHD relati . ve risk of 1.4 case-control study reported at the 114th meeting of the for nonsmoking wives of smokers, although the.value de- American Public Health Association. Based on 23 cases, . rived from only,four recorded CHD deaths. For the AH- it it reported a CHD relative risk (RR) of 2.6. No informa- SMOG cohort, CHD relative risks for those married to tion was given about data collection and analysis meth- smokers were decreased in nonsmoking men and in- _ odologies, and no confounding adjustments were made creased in nonsmoking women. The same inverse asso- ciation Svendsen et al. (1987). This cohort study followed held for nonsmoking men and women working ---- with smokers. There were significant discrepancies in for about 7 year.s a su..bsample of nonsmoking husbands of smoking wives from the MRFIT intervention trial. Its results are of questionable relevance because the exposure classification listed in both cohorts. regarding many individuals MRFIT trial recruited only men encumbered with Gillis et al. (1984); Hole et al. (1989). These two stud- known CHD risks and therefore a sample not represen- ies refer to an urban Scottish cohort of men and women tative of an open population. The nonsmoker category combined The latter is a 11.5-year follow-up report, giv- included exsmokers. Nonsmoking husbands of smoking ing CHD mortality relative risks of 2.01 and 2.27 for pas- wives were heavier and consumed more alcohol than . sive and active. s, moking, respectively,. The authors com- nonsmoking husbands of nonsmoking wives. Moreover, mented that the RR value for passive smoking "seems none of the reported relative risk values are statistically large in comparison with that found for active smoking, significant, and the sample size is one of the smallest. - and the possibility that chance has inflated the risk can- --- not be excluded. ..." They also noted that such a small Helsin et al. (1988); Sandler et al (1989). These two difference precluded_ the inference that the passive studies reported on one cohort of nev.er-sm.o.king men smoking association was biologically plausible. A deci- and one of never-smoking women from.the Washin. ~on sive flaw of this study is that the nonsmoker category County of Maryland, followed from 19633 to 1975. The included exsmokers, with no indication of their preva- study did not control for many known CHD risk factors l.ence.in:the cohort nor of the length of time of.their ab- and did not report on follow-up losses. Only deaths oc- - stinence. curring inside but not outside Washington County Were reported. Dose-effect gradients were not apparent Rel Humble et at. (1990). This study concerns a small ru- ative risks for females differ in the two papers (1.24 vs ral cohort of nonsmoking women from Evans County, 1.19), although the same database and adjustment were Georgia. None of the reported relative risk values at- apparently used. Despite an elaborate scoring procedure tained statistical significance. For white women, ETS for estimating ETS exposure, no data were reported re- lating to smoking spouses alone, thus precluding a rea- sonable comparison of this with other studies. Smoking data were collected only for 1963 at entry, but not for the follow-up period. The authors admitted inadequate control of confounders, among other things because exposure was inversely associated with socioeconomic status, showing reduced risk values for the less affluent group. Smoking status was determined at entry only but not during the 20 years follow-up. The authors also noted that no information was available of exposure changes due to remarriage. "[t]he general increase in mortality [..d..uring the study] Jackson (1989). This unpublished dissertation is leaves open the possibility that the lifestyles of people difficult to interpret because of unclear determination of who live with smokers differ from those who.do.not live ' exposures, lack of control for confounders and biases, with smokers. Factors such as alcohol consumption and and the discrepancy between male and female reports. dietary habits which are correlated with both smoking and risks for some diseases [and which were not col- He et al. (1989). A case-control study from China lected in the 1963 census] seem especially likely to be based on 34 cases and 68 controls reported a CHD risk alternative explanations for our findings, to the extent of 1.5 for nonsmoking women married to smokers. Con- that diets and alcohol use are similar among household trary to the authors' claim, this value cannot be statisti- members." These studies were considered inadequate by cally significant on account of its small value, the small the ETS reviews of the National Academy of Sciences, number of cases, and the wide variance reported for the the U.S. Surgeon General, and EPA (NAS, 1986; USSG, unadjusted estimate. The study is questionable in that 1986; USEPA, 1992c). no details of data collection and subject selection proce- Butler (1988). This study concerns a cohort of Cali- fornia, Seven Day Adventists. The sample may be incon= 11 sistent, as only 58% of registered Adventist households responded to questionnaires. Two cohorts were ex- tracted from the responses: One spouse-pair cohort and the other a cohort of subjects participating in an air pol- lution survey (the AHS.IVIOG cohort). The spouse-pair dures are given. Apparently no corrections for confound- ers were introduced, even though the study states that 11 "[w]omen exposed to passive smoke also showed abnor- mal levels of serum LDL-C, HDL-C, apoAl, and apoB." Dobson et al. (1991). An Australian case-control study reports a decreased risk for nonsmoking men, but an increased CHD risk (RR = 2.46) for nonsmoking I

... . .... ......... 288 GIO BATTA i'iORI t ~ I i I I I i I I . women married to smokers. Based on a small sample, no ies highlight the futility of further studies to measure excess myocardial infarction (MI) risk was reported for what appears to be an absent effect, in light of the vast workplace exposure in men or women. The study states MS/ETS dose differentials and the CHD-NOAELs for that no results were statistically significant._No controls ....active smoking. for CHD risk factors were provided. Serious method- ... With this in mind, conjectures of possible causal ological problems are apparent, as information for cases pathogenic mechanisms become a superfluous exercise. and controls was obtained by different procedures. La Vecchia et al. (1993). A small case-control study from northern Italy reported an overall male/female ad- justed CHD RR of l.`?1: Responsibly, the authors con- cluded: "The interpretation of this studv remains incon- However, investigations about mechanistic hypotheses have produced additional evidence of no-effect thresh- olds for active smoking and CHD and further contribute to negate the hypothesis of CHD risks from ETS. clusive because of the lack of statistical significance, ETS. .AND POSSIBLE CHD RISK FACTORS limited exposure assessment, and potential misclassifi `` "` IN HUMANS cation of smoking status of subjects tnterviewed_and their spouses." The evidence of NOAEL thresholds for active smok- He et al. (1994). A study of nonsmoking Chinese women exposed to ETS both at home.arid at the work.- place is described. There is no control for workplace con- founders and the sample size is very small. ing and ETS is not confined to epidemiologtc studies: it is also generally apparent in studies of the biomarkers and risk factors that are thought to have a pathogno- monic role. _Thres.hoids sh.ou..ld.. alsoo be viewed in the context of the obvious resistance and defenses that hu- mans must develop ag..ainst xenobiotics from a variety of LaYard ('1995). A case-control study extracted from sources and against the inevitable background of DNA the massive database of the National Mortality Follow- damage from natural metabolic functions that cause an back Survev (NMSF), a national probability sample of average of 10,000 measurable DNA modification events U.S. adult deaths in 1986, produced by the National per hour in each mammalian cell (Billen, 1990; Ames .. .. .... .... Center for ea t Statistics. From a total of 18,7133 and Gold, 1991). deaths, the sample provided 475 male and 914 female deaths for ischemic heart disease (IHD). It also provided ETS and xenobiotics, It has been suggested that mu- 988 male and 1930 female controls from death causes tagens s. an .d...carcinogens may have a role in the initiation unrelated to smoking. Users of less than 100 cigarettes.~ and pathogenesis of CHDs. In this context, the meaning lifetime were classified as nonsmokers. Multiple regres- of the mutagenicity associated with smoking remains sion over several variables of interest yielded IHD odds speculative. Extensive studies by the National Toxicol- ratios of 0.9 7 (0. f3-1.28 95`.~'~ CI) for nonsmoking males ogy Program and other researchers have failed to vali- and 0.99 (0.84-1.16 95`f'e CI) for nonsmoking females date mutagenicity as a predictor of carcinogenicity in married to smokers. No dose/response gradient was ob- . animals bioas.says or in man (Ashby and Tennant, 1991; served in relation to spouse's smoking habits. Zeiger et al., 1990; Tennant, 1988; IARC, 1987; USEPA, 1986; OECD, 1984). Studies have also shown that the LeVois and Layard (1995). An analvsis of data from mutagenicity of smokers' urine is elevated only in smok- the two large Cancer Prevention Studys I and II (CPS-I, ers of over 10 cigarettes/day, offering other evidence of a CPS-II), two national surveys conducted by the Ameri- can Cancer Society (ACS). These two.surveys provide a total of 18,073 CHD deaths, a number that far outstrips all combined cases ever published on the subject. These threshold effect (van Doorn et aL, 1979; Mohtashamipur et al., 1985). More recent studies report that-although mutagenic=cigarette smoking .may' actually protect against additional genotoxic insults (Oesch et al., 1994). and the data from the National Mortality Followback.. .. A review by the International Agency for Research on Survey provide a record of over 19,462 CHD deaths, Cancer listed 10 studies that could not find differences compared with less than 2400 from all reports published in sister chromatid exchange (SCE) frequencies in pe- to date. The CPS-I and CPS-II studies report risks of ' ripheral lymphocytes of smokers and nonsmokers, while CHD for ETS exposure ranging from 0.97 to 1.01. studies reporting a dose/response gradient were consis- . Pooled data from CPS-I, CPS-II, and the NMFS give a tent with a NOAEL threshold for smoking less than 10. CHD risk estimate of 1.00 (0.97-1.04 95% CI). cigarettes/day (IARC, 19.86, pp. 191-192). The situation The absence of homogeneous characteristics among ETS-CHD studies precludes a metaanalysis numerical summation that could be credible and scientifically jus- tified. Especially the data from the stronger studies show that epidemiologic studies are unable to conclude that ETS exposures represent a CHD risk. Together, all stud- has been confirmed by more recent studies of SCE fre- quencies in ETS-exposed subjects (Sorsa et al., 1989; Gorgels et al., 1992). Significantly, studies report that aborted fetuses from smoking mothers have 40% less chromosomal abnor- malities than fetuses from nonsmoking mothers (Kline. et al., 1993), while other studies report that maternal I

PASSIVE SMOKING AND CORONARY HEART SYNDROMES "89 smolcing ~s assoc~ated'with a much decreased risk of used to induce platelet aggregation were 10- to 20-fold mongoloid retardation or Down syndrome (Kline et al., higher than under normal physiologic conditions. The 1993; Cuckle et al., 1990). _ study also found implausible changes in platelet sensi- In general, no association was found between 4-ami- tivity in nonsmokers but not in active smokers. Equally, nobiphenyl-hemoglobin adduct.s.., pack.years of smoking, .;"studies by Davis et al. (1989) .used a nonstandard platelet and cancer diagnosis (Weston et a1., 1991). Following aggregation assay, which is influenced by the procedure nuclease P_1 enrichment techniques, DNA adducts were to draw blood, the presence and kind of anticoagulant, ......: .. ........ detected in oral tissues, but the adduct burden in smok- centrifugation parameters, and other variables..Also, the ers of 1-10 cigarettes/day was not different from the bur- meaning of endothelial cell carcasses in circulation is den in nonsmokers, giving another confirmation of not apparent: In any event the study reports that "After threshold (Jones et al., 1993). Studies of human lung passive smoking, the percentage carboxyhemoglobin . , cancer tissues found that serum cotinine levels and ad- level did not correlate significantly (P > 0.60) with duct levels were not correlated and could detect adducts the platelet aggregate ratio or the endothelial cell only in 7 of 38 individual tumor. samples (Shields et al count. . Neither the plasma nicotine concentration ........ .. . 1993). DNA adducts of aromatic hydrocarbons in lym after passive smoking nor its change from before to after phocytes were not found to correlate with smoking hab- passive smoking was significantly (P > 0.20) correlated 11 its (van Schooten et al., 1992; Grzybowska et al., 1993) with the corresponding values of the platelet aggregate DNA adducts were at similar levels.in sperm cells of ratio or the endothelial cell count." Further, the paper ; smokers and nonsmokers, a finding of interest given the states that "The significance of the platelet aggregate intense DNA replication in spermatogenesis (Gallagher formation and an increased concentration of anuclear et al., 1993). Equal similarities were reported for DNA carcasses of endotheiial cells in blood after passive adducts of cervix tissues (King et aG., 1994), smoking is not known." I I I I I ETS, thrombus formation. Hypotheses have pro- ETS, cholesterol, tipidemias, and hypertension. The posed that increased blood-clotting capacity may ex- Framingham Offspring Study and a Kaiser Permanente plain the association of heavy smoking and .c.ardiovascu study report that cigarette smoking was not correlated lar events. However, the Framingham...study_reports an with Lipoprotein(a) levels (Jenner et al:, 1993; Selby et absence of cardiovascular risks for smokers of 1-10 cig- al., 1994). Between 1981 and 1990, an evaluation of arettes,/day, consonant with an immaterial _change in 11,199 randomly selected subjects in New England mean fibrinogen (about 1%) in smokers of less than one found no association of smoking and dyslipidemic hy- pack of cigarettes/day (Kannel et al., .1987). It has beenn pertension (Eaton et al., 1994). The British Regional suggested that two eicosanoids, prostacyclin and throm Heart Study, a large prospective study of 7735 men aged boxane A2, are altered in smokers, but studies indicate 40-59, reported on alcohol drinking, smoking, and cho- no change in smokers of less than 10-15 cigarettes/day lesterol levels, concluding that "current smokers w o (Wennmalm et al., 1991). The large Kuopio prospective were heavy drinkers or nondrinkers had the lowest mean study in Finland found a correlation of plasma fibrino- cholesterol levels" (Wannamethee and Shaper, 1992). A gen levels with several psychosocial and socioeconomic Japanese study also found no difference in plasma cho- variables, but not with smoking (Wilson et al., 1993). lesterol between smokers and nonsmokers (Imaizumi et Thrombomodulin levels were not found to be elevated al., 1991). A study of 51,723 participants of community- after smoking (Kubisz et al., 1994). Vicari et al (:1988) . based cholesterol screening clinics in 10 U.S. cities foun measured several thrombotic factors and found no no association between active smoking and plasma cho- differences due to active smoking. Haire et al. (1988) lesterol levels in men and women over age 60 and no ele- found no changes in fibrinolytic activity after active vation of plasma cholesterol for younger subjects smok- smoking. Yamashita et al. (1988) found no effect of ing less than.l0 cigarettes/day (Muscat et al., 1991). smoking on platelet aggregation. Handley and Teather The Cardiovascular Health Study Collaborative Re- (1974) and Barbash et al. (1993, 1994) give indication search Group reports that in a cohort of 5201 men and that smoking may protect against thromboembolic com- women over 65 years of age, cigarette smoking was a neg- plications after myocardial infarction and surgery. Sim- ative predictor of blood pressure, confirming a number ilar results were reported by Pollack and Evans (1978). of prior reports (Tell et al,,._1994). A.recent collaborative The Atherosclerosis Risk in Communities Study of study of the Center for Disease Control of the U.S. De- 15,800 men and women in the United States found that partment of Health and Human Services found that cor- smoking was positively associated with levels of Anti- onary occlusion was inversely correlated with levels of thrombin III, a major anticoagulant factor (Cn onlan et al., high-density lipoproteins (HDL), but unrelated to 1994). smoking (Freedman et al., 1994). Sonie studies that suggest differently have been flawed The massive WHO-MONICA study actually "showed or misinterpreted. A study by Burghuber et al. (1986) a strong negative association between regular smoking may be irrelevant because the exposure and ADP level and high cholesterol in the male populations and a

290 GIO BATTA GORI . .... . . . . . ....... . . ..... stzong negative association between regular smoking tomographyy study of patients with interstitial lung dis- and high blood pressure in female populations" (MON-_ ease reported no correlation of smoking with either the ICA, 1994). Well known is the so called "French para- disease or.emphysema (McDonagh et al., 1994). dox," whereby the French population shows 55% less CHDs.that the rest of the populations surveyed in the . ETS and carbon monoxide. The possible effects of MONICA project, despite experiencing high ievels.of carbon monoxide (CO) on the cardiovascular system are cigarette smoking, hypertension, and total cholesterol largely based on inferences from dated claims (Aronow, (Renaud and de Lo 1978). In general, such claims hold that 1-2~~'Q CO con- Study 1993). The Helsin ki _ Agem g ... .. .. Study reports a slight inverse association of active smok- centratlons trigger angina episodes, even though they ing and aortic valve degeneratlon:_in the elderly (Lin- are close to typical background ambient concentrations. droos et al., 1994). In China, coronary mortaiitv is some Such claims are difficult to accept because most epide- miologic studies of active smoking report no-effect 10 times less frequent than in Germany, although the ~ thresholds for angina at well over. 10 cigarettes smoked prevalence of smoking is 70:c: ~ in China versus 37 ~-c m . Germany. Total cholesterol,_however, is much higher in per day, as noted above. Over the years these claims have German than Chinese subjects (Stehle et al., 1991). Yet, been criticized even by the U.S. EPA as suffering from a also in Germany, the prevalence of most CVD risk fac- number of technical and design problems (Sheps et al., tors increased considerably during a period of substan- 1987). The National Research Council had this to say of ,.., . tial mortality declines (Hoffmeister et al., I994). Anom- these claims: alies such as these have led prominent clinicians to con There were some subjective elements in the evaluation of these clude that total and low-density lipoprotein (LDL) patients, and the physician conducting these tests was aware of cholesterol may be the only demonstrable CVD risk fac the test conditions, i.e., smoking or not and ventilated or not. tor.(Roberts, 1989). . ........ :,;....... Consequenly, the findings of this. study, in the absence of a true These considerations tell that the active smoking of double-blind approach, require verification b_v other research less than 10 cigarettes f day or ETS exposures are un- workers (NRC, 1986): likely to adversely influence lipidemic CHD risk factors. As such, they support the notion of NOAELs AELs fo..r actlve..::: In regard to the possible aggravation of angina by CO, smoking and cardiovascular diseases. In fact, the Na- the National Academy report on ETS as well as EPA tional Cholesterol Education Program only lists smok- regulations consider that such effects may occur at car- ing of over 10 cigarettes/day as a possible CHD risk fac- boxyhemoglobin (COHb) blood concentrations above tor (NCEP, 1988). 3io,(NAS, 1986, p: 261; USEPA, 1984).. Because this is a risk management figure with an ample safety margin, - ETS and ventilaton, function. Compromised venti- the true threshold would obviously be substantially latory function has been suggested_as a possible risk fac- . higher. In any event, a 3% COHb level is not reached tor for CHDs; however, there is evidence for_r..espiratory even at high ETS exposure (NAS, 1986; IARC, 1986; .NOAELs in the reported associations with active smok- Jarvis et aG, .1983...). Moreover, the first Surgeon Gener- ing. Early reports of forced respiratory volume (FEV,) . al's: report found no adverse cardiovascular effects in and forced vital capacity (FVC) deficlts in,smo..kers are moderate pipe and cigar smokers despite CO exposures still compatible with NOAEL thresholds below 10 ciga- far exceeding typical ETS exposures (USSG, 1964). rettes/day (Dockery et al., 1988; Sorlle et al. 198?). The Multiple Risk Factor Intervention Trial (MRFIT) EXPERIMENTAL STUDIES IN ANIMALS tested 6347 males randomly distributed in a usual care group (control) and a special intervention group that in- Animal studies mostly refer to MS exposures, are gen- cluded an intensive and substantially successful smok- erally set.up for maximum effect, and hence push dos- ing cessation program; During a 6- to 7-year follow-up, ages to the extremes :compatible with survival. Some FEV; measures were similar in the two groups, a finding studies, for instance, cause COHb blood levels as high as that confirms reports froin previous studies (Browner et .. 40 o in treated animals, a condition that finds no close parallel in human smokers. Also, experimental studies al., 1992). A threshold effect at some 10 cigarettes/day was also have utilized animals whose reactions to high doses of reported for emphysema in.asthmatic patients (Kon..doh MS, ETS, or their components are ostensibly of un- et at., 1990). Elastin peptide concentration in circulation ` known relevance to human responses, given profound has been suggested as a marker.of l...u ng elastin degrada- differences in smoke generation and exposure condi- tion and thus of emphysematous changes. A recent.stu.dy tions, and in anatomy, physiology, and metabolism. In found no correlation between elastin peptide concentra- the end, there is no doubt about potential pathogenic tion and smoking (Frette et at., 1992) Another study effects of cigarette smoke administered at extremely found that severity of microscopic emphysema did not high doses to animals. At the same time, it is equally increase with daily cigarette consumption (Gillooly and clear that the interpretation of such effects in equivalent Lamb, 1993). More recently, a high-resolution computed human terms requires inferences that are unwarranted

I I I I I I ~ I PASST'VE SMOKING A.*ID `ORONARy HEART SYNDROMES . 291 by the profoundly different conditions under which they concentrations two to three orders of magnitude higher are obtained. that for typical ETS exposures, reaching COHb levels Olson (1985) reported increased ornithine decarbox- near 9% in plasma and plasma nicotine levels two to ylase activity in tracheas of rats exposed to sidestream three times higher than in average active smokers. Base- _e cigarette smoke. However, this study failed to show in- line hematologic values were substantially different creased activity in the lungs, and no erfect`whatsoever among different groups. Paradoxically, the exposure sig- was noted at the 10% sidestream smoke dose, despite nificantly reduced total plasma cholesterol. The animals considerable exposure documented by COHb levels .. were also implanted with a snare occiuder to the left an- around 6%. ...:. .terior descending coronary artery, which was eventually Because they could both inactivate and activate xeno- activated to cause left ventricle infarction. No controls biotic to electrophiles of concern, a possible pathogno- were provided to account for the obviously severe gen- monic meaning is still a matter of conjecture in regard to eral and cardiac stress. Studies in cockerels produced the active-smoking induction of such polymorphic enzy- weaker results and are even less interpretable in relative matic systems as cytosolic glutthione S-transferase, human terms {Penn and Snyder, 1993; Penn et al., 1994). P450 cytochromes for aromatic hydrocarbons and de ...O _ ne of the largest, best-planned, and rigorously con- brisoquine-aryl hydrocarbon hydroxylases (AHH) or ducted inhalation studies of cigarette smoke ever per- microsomal monooxygenases (MMO) -and the arylam- formed was a 2-year massive inhalation study sponsored ine acetylators. The interpretation of their significance and directed by the National Cancer Institute (NCI) and is further complicated by a multitude of..genetic controls the National Heart, Lung, and Blood Institute and determinants in specific phenotypes (Ketterer et al., (NHLBI). It was performed between February 1978 and 1992; Anttila et al., 1992; Bartsch et al.,,. i992;,. Caporaso March 1980 on 220 purebred beagles fed a 5% cholesterol et al., 1992; Vineis and Ronco, 1992). For instance, Gair- diet and exposed by tracheostomy to mainstream ciga- ola (1987) found thatt such induction happened in mice rette smoke variously spiked to provide excesses of nico- . and rats but not in guinea pigs, the latter actually expe- tine or CO or of both (Hazleton Laboratories America riencing a small reduction of activity; a finding con- Inc., 1980). The study was designe d. and monitored by firmed by other studies and in other animal species (Bil- a group of top experts specifically assembled by NCI- imoria et al., 1977; Lubawy and Isaac, 1980). In any NHLBI. After 2 years of exposure, the unexpected re- event, HHA-MMO activity may be necessary to both ac- - suits gave unequivocal indication that increasing levels tivate and deactivate xenobiotics, and therefore the sig- of cigarette smoke, CO, and nicotine reduced the sever- nificance of such changes remains moot. ity of atherosclerotic lesions. The final report concluded The studies by Zhu et al. (1993, 1994) are equally re- . that "jtJhese results appear more indicative of a possible markable for their lack of adverse findings. The rabbits protective effect from cigarette smoking and/or CO in- of the first study were exposed to extreme concentra- halation than of an atherogenic effect." The study could tions of what amounts to freshly d...il..ute.d. sidestream . not find incidental lesions of the respiratory system nor smoke, with particulate concentrations 100 times and significant change.s in _blood lipids among the exposed 1000 times higher than for typical ETS at the low and groups. . ... .high doses, respectively, when considering that typical Thus, the data from laboratory and animal experi- ETS particulate concentrations listed by the EPA A are ments offer no plausible argument to classify ETS as a in the order of, 30 µg/m3 (USEPA, 1992c). There were human CHD risk, they do not contradict or rather sup- _.:. conflicting effects on serum triglycerides, ch .olesterol, port the evidence of NOAELs for active smoking and and high-density lipoprotein cholesterol. Bleeding time CHD, and therefore contribute to falsify the hypothesis decreased in the treated groups, but paradoxically plate- o.f CHD risks from ETS. let aggregation and platelet count also decreased in con- ;: tioned in trol . trol animals. The at .herosclerotic changes men .: the paper actually amounted to somewhat increased lipid deposits in the arterial walls, although the study does not provide the micrographs necessary for an inde- pendent evaluation. The pathognomonic significance of such changes is unknown, given that..they were similar in unexposed control animals and that no report is given of the likely effects of the stress induced by smoke exposure. Also, the results presented and the presence of lesions in control animals must be interpreted in the context of the natural predisposition of New Zealand rabbits to atherosclerosis, especially when fed a high- cholesterol diet as in this study. The second Zhu et aL study also entailed exposures to si..d.estream .smoke_ at CONCLUSION Plausible ETS doses are thousands of time less than MS doses that appear to have no adverse CHD effects in active smokers. Such determination precludes the infer- ence that ETS is a CHD risk, unless we are prepared to forgo all we have learned since Paracelsus about phar- macodynamic and kinetic discontinuities at low doses. By far the majority of experimental reports in man or ~ animals either do not contradict or support this conclu- 0 sion and together indicate that epidemiologic studies ~ have been chasing an absent CHD effect-a conclusion ~ sustained by the generally equivocal or null reports from ~ epidemiologic studies of ETS. The instability of data ~~ I

292 .' GIO BATTA GORI t I I ~ fzom most epidemiologic. studies, the heterogeneity in let aensitivity to prostacyclin in. smokers and non-smokers. Chesc study design, data collection, and evaluation methods, 90, a4-aa precludes a metaanalysis numerical summation that is Bush. T. LandComstock,G.W.(1983).Smokingandcardiovascular scientifically justifiable. The evidence favoring the ETS- mortality in women Am• J. Ept•demiol. 118, 480-488. $utler, T. L 11988) The relationship of passive smo ing to various CHD association remains conjectural, while the evi- health outcomes among Seven Day Adventists in California. Ph.D. dence against the association is suitably documented. Dissert'atxon,tiniversitv of California, Los Angeles. According to the scientific method, the only justifiable Caporaso, N. E., $hields, P. G., Landi, M. T.. et at. (1992). The de- conclusion is that available data continue to falsify the bri5oquine metabolic phenotype and DNA-based assavs: Implica- hypothesis that ETS is a CHD risk factor. tions'of misclassification for the association of lung cancer and the ;. _ debrisoquine metabolic phenotype. Environ. Health Perspect. 98, 101-105. REFERENCES Conlan, M G.,.Folsom, A. R., Finch, A., et al. (1994). Antithrombin III: Associations with age, race, sex and cardiovascular disease risk Ames, B. N.. and Gold. L. S. (1991). Endogenous mutagens and the factors. Thromb. H.aemostasis 72, 551-556. causes of aging and cancer, Mutat. Res 250,.3 16 .. : Cress, R. D.• Holly, E. Fl„ Aston, D.A., et al. 119941: Characteristics of Angerer, -1•, Heinzow, B., and Reimann. D. O. (1992). internal women nonsmokers exposed to passive smoke. Pree. Med. 23, 40- exposure to organic substances in a municipal waste incinerator. 47. Int. Arch. Occup. Enuiron: Health 64, '165=273. Cuckle, H. S., Alberman, E., Wald, N. J., et al. (1990). Maternal smok- Anttila, S., Vainio, H., Hietanen.,.E., et al 11..9921 immunohistochem- inQ habits and Down'ss syndrome. Prenatal Diagn. 10, 561-56 7. ical detection ofpulmonary Cytochrome P450IA1 and metabolic ac- Davts, J. W., Sheldon, L., Watanabe, I., et al. (1989). Passive sinoking tivities associated with P4501A I and P450IA2 isozymes in lung can- . effects on endothelium and platelets. Arch. Int. Med. 149, 386-389. cer patients. Environ. Health Perspect. 98, 179-182. - Dobson, A J Alexander, H M., Heller, R. F., et al. (1991). Passive Aronow, W. S. (1978). Effect of passive smoking on angina pectoris. smoking and the risk of heart attack or coronary.death. Med. J. N. Engl. J. Med. 299, 21-24. .4ust. 154 93-797. Ashbv, J., and Tennant. R. W. (1991). DeFinittve relationships be- Dockery..D W., Speizer, F. E., Ferris, B. G., et al. (1988). Cumulative tween chemical structure, carcinogenicitv and mutagenicity for 301 and reversible effects of lifetime smoking on, simple tests of lung : chemicals tested by the U.S. NTP. Mutat. Res. 257, 229-306• .- , function in adults. Am. Rev. Respir. Dis. 137, 286-292. . Baker, R. R, and Proctor, C. J. i 199.0)..The origms and properties of Doll, R., Gray, R.. Hafner, B., et al. (1980). Mortality in relation to environmental tobacco smoke. Environ. Int: 16, 231-245. smoking: 22 years' observations on female British doctors. Br. Med. Barbash. G. I., Reiner, •J. S., Granger. C. B., et ai. (1994). Smoking and J• 280, 967.-971. acute mvocardial infarction survival: Demographic and quantitative Doll R., and Peto R(1976). Mortality: in relation to sqioking: 20 .................................... . ......... . ........ angiographic foundations for the "Smoker Paradox" tAbstract). years' observations on male British doctors. Br. Med. J. 2, 1525- Circulation 90, 221. 1.5.36. Barbash, G. I., White, H. D., and Modan, M. (1993). Significance of Eaton, C. B.. Feldman, H. A., Assaf, A. R., et al. (1994). Prevalence of smoking on patients receiving thrombolytic therapy for acute inyo hypertension:' dyslipidemia, and dyslipidemic hypertension. J. ,... cardial infarction. Experience gleaned from the internationai tissue Fami1. Pract. 38, 17-23. plasminogen activatorJstreptokinase mortality trial. Circulation Emmons, K. M., Abrams, D. B., Marshall, R. J., et al. (1992). Exposure 87, 53-58. .. to environmental tobacco smoke in naturalistic settings. .4m. J. Bartsch, H., Petruzzeili, S., De Flora, S.- et al. (1992). Carcinogen me- Pubiic_Health 82, `~4=28. ' tabolism in human lung tissues and.the effect of tobacco smoking: Fontham, E.. T. H.,.Correa, P., Reynolds, P., et al. (1994). Environ- Results from a case-control multicenter study on lung cancer Pa mental tobacco smoke and lung cancer in nonsmoking women. A tients. Environ. Health Perspect: 98, 119-124. multicenter study. JAMA 271, 1752-1759. Bilimoria, M. H., Johnson, J., Hogg, J. C., et al. (1977). Pulmonary Foster, W M•, Langenbach, E. G., and Bergofsky, E. H. (1985). Disso- arvt hydrocarbon hydroxilase: Tobacco smoke-exposed guinea pigs. ciation in the mucociliary function ofcentral.and peripheral airways .... Toxicol. Appl. Pharmacol. 41, 433-•#40. of asymptomatic smokers. Am. Rev. Respir. Dis. 132, 633-639. Billen, D. (1990). Spontaneous DNA damage and its significance for Freedman, D. S,, Croft, J. B., Anderson, A. J., et al. (1994). The rela- the "Negligible Dose° controversy in radiation protection. Radtat tioti of documented coronary artery disease to ievels of total choles- Res. 124, 242-245. teiol and high-density lipoprotein cholesterol. Epidemiology 5, 80-. Breslow, N. E., and Day, N, E- (1980). Statistical Methods in Cancer .:.: 87• Research. Volume 1. The Analysis of Case-Control Studies, Pubiica- Frette, C. Wei, S. M., Neukirk, F., et al. (1992). Relation of serum tion No. 32. International Agency for Research on Cancer, Lyon. elastin peptide concentration to age, FEVi, smoking habits, alcohol 'c Brown, C. A., Crombie, I. K., and Tunstall-Pedoe, H. (1994). Failure consumption, and.Protease inhibitor phenotype: An ePtdemioto~ of cigarette smoking to explain international differences in inortal- study in working men. Thorax 47, 937-942. ity from chronic obstructive pulmonary disease. J. EpidemioL Com.- Freund, K. M., Belanger, A. J., D'Agostino, R. B., et al. (1993). The munity Health 48, 134-139. ° health risks of smoking. The Framingham study: 34 years of follow- Browner, W. S., Du C.hene, A G, an...d Hulley, S. B. (1992). Effects up Ann. Epidemtol. 3, 417-424. of the Multiple Risk Factors Intervention Trial smoking cessation Froggatt, P. (1988). Fourth Report of the Independent Committee on program on pulmonary function. A randomiied controlled trial. .. Smoking & Health. Her Majesty's Stationery Office, London. West. J. Med. 157, 534-538. Gairola, C G(1987) Pulmonary aryl hydrocarbon hydroxylase activ-. Brugnone, F., Perbellini, L., Maranelli,_G, et.a..l. (1992) Reference val ity of mice, rats and guinea pigs following long term exposure to ues for blood benzene in the occupationally unexposed general pop- . mainstream and sidestream cigarette smoke. Toxicology 45,. 177- ulation. Int. Arch. Occup. Environ. Health 64, 179-184. 184. Burghuber, O. C., Punzengruber, C., Sinzinger, H., et al. (1986). Plate- Gallagher, J. E., Vine, M. F., Schramm, M. M., et al. (1993). a2P-post-

PASSIVE SMOKING AND CORONARY HEART SYNDROMES . 293 label'~ng analys'ks ot D:3A addajcts inhuman sperm celts.from scnok- Hoffineister, H., Mensink. G. B. M., and Stolzenberg• H, ( 199-1,). \'a- ers and nonsmokers. Cancer Epidemiol: Biomarkers Prel 2, 5$1- tional trends in risk factor for cardiovascular disease in Germany. 585. . Preat Med. 23. 197-205. Garland. C.. Barrett-Connor, E.• Suarez. L.. etaf. (1985). Effect of pas- ' Hole. D. J., Gillis, C. R., Chopra, C.• et al. (1989). Passive smoking I I I I I sive smoking on ischemic heart disease mortality in nonsmokers. Am. J. EpidemioL 121, 645-650. Gerde. P_• Medinskv, :y1..A..,, and Bond, J. A. (1991). The retention of polycyclic aroinatic hydrocarbons in the bronchial airways and in the alveolar region=A theoretical comparison. Toricol. .-lppl. Phar- macol. 107, 239-2-52. .... ...... and cardiorespiratorv.heal.th in a general population in the west of Scotland. Br. Med. J. 299 423=-I27. . Hopki.ns. P, N., and Williams. R. R. (1991). A survey of 246 suggested coronary risk 1"actors. Atherosclerosis 40, 1-52. H....umble•.....C.• Croft, J.. :G..erber,.A.•_et al. (1990). Passive smoking and 20 vear cardiovascular disease mortality among nonsmoking wives• Gillis. C. R., Hole. D. •J., Hav.•thorne, V. :vl., et at. (1984). The et£ect of Evans Countv; Georgia. Am. J. Pub(ic Health 80, 599-601. ..:..:. environmenkal tobacco smoke (n two urban communi es in the west IARC (1986). Tobacco. smoking: Monographs on the Evaluation of the i?fScotland. Eur. J. Respir. Dis 65(Suppl. 1331, 1`?1-1`?6. Carc.i....a.dgenk tc..R..i.sk of Chemicals to Humans. Vol. 38. International Gillooly.M.-and Lamb. D. ! 1993)• Microscopic emp)r. >ema .n relatuon Agency For Research on Cancer-World Health Organization, to age and smoking habit. 7horas 4$, 491-495. Ly.o..n.. ... Gillum. R. F. (1994). Prevalence oFcardiovascular and pulmonary dis- IARC (1987). Overall Evaluation of Carcinogenicitv: An C pdating of eases and risk factors by region and urbaniiat on in the linited IARC MonoCraphs Vofumes I to 42. Supplement 7. International States. J. tiatl. Med. Assoc. 86, 105-1.1'?. , Agency for Research on Cancer. Lyon. Glanz. S. A.• and Parmley. W. W. 1199.1. t Passive smoking and heart ,- Imaizumi. T., Yoshida,.H.. Kawamura. Y.. et al. (1991). Effect of ciga- disease: Epidemiology'• physiology. and hiochemistry. C rcutauon rette smoking on the levels of platelet-activating factor-like lipid(s) 83, 1-1'3•. . in plasma.lipoproteins. Atherosclerosis 87, 47-55. Gorgels, W. •I., van Poppel. G., •Jarvis, M. J., et al• (1992). Passive Jackson. R. T. (1989). Passive sinoking. Ph.D. Dissertation, Auckland smoking and sister-ehromatid exchanges in lymphoc~ tes. Mutat. Heart Survey, Auckland, New Zealand. Re5.279.233-'?38. . Gori. G. B. (19.,6).. _L..ow risk cigarettes: A prescription. Science 194, Jarvis M. J., Russell, M• A. H•, and Feveraben ,•(1983). A sorption 1243-1246. of nicotine and carbon monoxide from passive smoking unde;: natu- ral conditions ofexposure. Thorax 38, 829-833. Gori• G. B.• and Mantel•. N. (1501). Mainstream and environmental . tobacco smoke. Regu1..Toxico! Pharmacol. 14, 38 10~. Jenkins, C. D.. Rosenman, R. H., and Zvzanski, S. J. (1068). Cigarette "' relationship to coronary heart disease and related risk smoking The Gravelle. J. G.. and Zimmermann. D. (March 8. 1994). Cigarette taxes to fund health care reform: An economic analysis. Congressional " factors in the Western Collaborative Group Study. Circulation 38, 1140-1155. Research Services-The Library of Congress Washington, DC. J Grzy bowska• E.. Hemminki, K.. 5zeliga, •J., et al. (1993). Seasonal vari- Jenner. J. L.,.Ordovas, J. M•• Lamon-Fava. S:, et al. (1993)• Effects of ation ofaromatic DNA adducts in human lym hocytes and granu- age• sex and menopausal status on plasma Lipoprotein(a) levels. locvices. Carcinogenes[s 14, 2523-25?6. p Circulation 87, 1135-1141. - Jones, N. J'VIcGregor, A. D., and Waters, R. (1993). Detection of Guerin. M. R.. Higgins. C. E.• and .)enkins• R. t1 t lyn7). Measuring DNA adducts in human oral tissue: Correlation of adduct levels with environmental emissions from tobacco combustion: Sides(ream cie- tobacco smoking and differential enhancement of adducts using the arette smoke literature review. Atmos Environ. 21, 291-?9 . Itutanol extraction and nuclease P1 version of32P postlabeling. Can- Haire. «•. D.. Goldsmith• J. C.• and Rasmussen, J. (1988). Abnormal cerRes. 53, 1522-1528. ' Rbrinolvsis in healthy..male cigarette smokers: Role of plasminogen activator inhibitors. ~lematology 31, 3n 40. Kannel. W. B.• D'Agostino, R. B., and Belanger, A. J. (1987). Fibrino- gen, cigarette smoking; Handle}. A. L• and Teather. D. (1974). Induence of smokingg on and risk of cardiovascular disease: Insights deep from the Framingham study. Am. Heart J. 1.13, 1006-1010. vein thrombus after myocardial infarction. Br. Med. J. 3, 230-23L Hazleton Laboratories America Inc.(1981). Final Report Inhalation Kennedy, S. M., Elwood, R. H., Wiggs, B. J., et al. (1984). Increase atrways mucosal permeability in smokers. Relation to airway reac- Bioassay of Cigdrette Smoke in Dogs. E/fects of Nicotine and Carbon trvity: Am Reu. Respir: Dis. 129, 143-148. Monoxide on ~ltherogenesis. Project ~lo. 976-904 Contract No. NCI- -- SHP• r5-112, the National Cancer Institute, Smokiag and Health Ketterer; B., Harris, J. M., Talaska, G., et al. (1992). The human Glu- tathione S-Transferase supergene family, its polymorphism, and its Program. Hazleton Laboratories America Inc., Vienna, VA. effects on susceptibility to lung cancer. Enuiron. Health Perspect. He. Y., Lam, T. H., Li, L. S., et al. (1994). Passive smoking at work as a risk factor for coronary heart disease in Chinese women who have 98, 87-94. :: never smoked. Br. Med. J. 308, 380-384. Keys, A. (1980). Seven Countries. A Multluariate Analysis of Death and He, Y., Li. L. S., Wan, Z• H., et al. (1989). Women's passive smoking Coronary Heart Disease. Harvard Univ. Press, Cambridge, MA. and coronary heart disease. Chin. J. Prev. Med. 23, 19-22. King, M. M., Hollingsworth A:, Cuzick, J., et al. (1994). The detection Helsing, K. J., Saddler, D. P., Comstock, G. W., et al. (1988). Heart of adducts in human cervix tissue DNA using "P-postlabelling-• A .. study of the relationship with smoking history and oral contracep- disease mortality in nonsmokers living with smokers. Am. J. Epide- . C is 15 1097-1100 enes n o miol. 127, 915-922. . g ... ,,....,, . Crve use.. .arcc. . Hill, A. B. (1965). The environment and disease: Association or Cau- sation? sation? Proc. R. Soc.. Med. 58, 295-300. Kline, J., Levin, B., Shrout, P., et al• .(1993), Maternal smoking and trisomy among spontaneous aborted conceptions. Am. J.. Human 421-43 1. t 35 Gene . . , , Hirayama, T. (1981). Non-smoking wives of heavy smokers have a higher risk of lung cancer: A study from Japan. Br J. Cancer 282., Kondob, Y., Taniguchi, H., Yokoyama, S., et al. (1990). Emphysema- 183-1$5. tous changes in chronic asthma in relation to cigarette smoking• C est 97, 845- . -- - - - Hirayama, T. (1984). Cancer mortality in nonsmoking women with smoking husbands based on a large cohort study in Japan. Prev. - Kubisz, P., Marculjak, I., and Cronberg, S. (1994). Thrombomodulin Med. 13, 680-690. and smoking. Atherosclerosis 107, 259-260. Hirayama, T. (1990). Passive smoking. N, Z. M...ed J.103,154. __. Lapidus, L. (1985). Ischemic heart disease, stroke and total mortality ~

294 .......... GIO BATTA GORI in women-Resulcs Erom a prospective population study in Go-.:. initial hazard assessment of chemicals. Organization for Economic thenborg, Sweden. Acta Med. Scand. Suppl. 705, 1-42. Cooperation and Development, Paris. LaVecchia, C., D`Avanzo, B., Franzosi, M. G., et al. (1993). Passive Oesch, F., Hengstler, J. G., and Fuchs, J. (1994). Cigarette smoking smoking and the risk of acute myocardial infarction. Lancet 341, protects mononuclear blood cells of carcinogen exposed workers 505-506. from additional work exposure-induced DNA single strand breaks. Layard, M. W. (1995). Ischemic heart disease, lung cancer, and spou- -ltutat. Res. 321, 175-185. sal smoking in the National Mortality Followback Survey. Regul. Oliver, M. F i 1974i. Ischaemic heart disease in young women. Br. Toxicol. Pharmacol. 21, 177 180 ' Med. J. 4, 253-259. - Lavard, M. W., and Viren, J.-R. (19891. .4ssessing the validity of a Olson, J. W i 1985). Chronic cigarette sidestream smoke exposure in- Japanese cohort study. In Present and future of indoor atr quaGit•. creasess rat trachea ornithine decarboxylase activity. Lije'Sci. 37, (C. J. Bieva. F. Courtois, and M. Govaerts, Eds.), pp. 17 7-180. Ex- 2166-?171. cepta Medica Int. Congr. Series No. 860, London. Ong, C. N.. Lee, B. L.. Shi. H. Y., et al. (1994). Elevated levels of ben- Lee, P. N. (1992). Environmental Tobacco'" Smoke and 1Lfortahh zene-related compounds in the urine of cigarette smokers. Int. J. Cancer 59, 17 7-180. Karger, Basel, London, New York. Lee, P. N. (1993). An estimate ofadult mortality in the United States Penn. A., Chen, L. C., and Snyder, C. .A., (1994). Inhalation ofsteady- from passive smoking. Environ. Int. 19, 91-100, state sidestream smoke from one cigarette promotes arteriosclerotic piaque development. Circulation 90, 1363-1367. .. Lee, P. N., Chamberlatn, J., and Alderson A R 11986 ), Relationship Penn, A.; and Snyder, C.. A. (1993). Inhalation of sidestream ci arette of passive smoking to risk of lung cancer and other smoking-associ smoke accelerates development of atherosclerotic plaques. Circula- ated diseases. Br. J. Cancer 54; 97-105. tton 88, 1820-1825. LeVois, M. E., and Layard, M. W. t1995). Publication bias in the en- Perbellini, L., Faccrni G. B., Pasini, F.. et al. (1988). Environmental vironmentai tobacco smoke/coronary heart dtsease":epidemiologic :: .. and occupational exposure to benzene by analysis of breath and literature. Regul. Toxicol. Pharmacol. 21, 181-188. ..` .. blood. Br. J. Ind. Med. 45, 345.-352, Lindroos. M., Kupari, M., Valvanne, J., et at. .(1994). Factors associ_ pollack, A V., and Evans, M. (1978). Cigarette smoking and postop- ated with carcific aortic valve degeneration in the elderly. Eur. Heart erative deep vein thrombosis. Br. Med. J. 3, 637-639. J. 15, 865-870. . . . . Pooling Project Research Group (PPRG) (1978). Relationship of Lubawy, W. C., and Isaac, R. S. (.1980). Acute tobacco smoke exposure . blood pressure, serum cholesterol, smoking habit, relative weight, alters the profile of metabolites produced from benzo[aJpyrene by and ECG abnormalit`t.es to incidence of major coronary events: Final the isolated perfused rabbit lung Toxicolo,ei~ 18, 3 7--)?. report of the Pooling Project. J..Chronic Dis. 31, 20I-306. Margetts, B. M., and Jackson. A A(1993) Interactions between peo- . "' - Renaud, S., and de Lorgeril, M. (I993). The French paradox: Dietary pie's diet and their smoking habits: The dietarv and nutritional sur- factors and cigarette smoking-related health risks, Ann. N. Y. Acad. vey of British adults. Br. Med. J. 307, 1381-1384 Sci. 686, 299-309. Martin. M. J.. Hunt, S. C., and \3 iItiams. R. R._(1986). Increased mci Richter, E., Falter, B., Scherer, G., et al. (1994). Regional differences dence of heart attacks in nonsmoking women married to smokers. .: . in hemoglobin adducts from aromatic amines (Abstract). Can. J. Paper presented at the 114th annual meeting of the American Pub- Physiol. Pharmacol...72(1), .596. lic Health Association. Roberts, W C. (1.989)...Atherosclerotic trisk factors-Are there ten or McDonagh. J.. Graeves, /b1., Wright, A. R., et al. (1994). High resolu- __ is there only one? Am. J. Cardiol. 64, 552-554. tion computed tomography of the lungs in patients with rheumatoid . Rosenberg, L., Kaufman; D. W., Helmrich, S. P., et al. (1985). Myo- arthritis and interstitial lung disease. Br. J. Rheumatol. 33, 118- cardiai infarction and cigarette smoking in women younger than 50 122. years_of age. JAMA 253, 296~-2969. Mercer, R. R., and Crapo, J. D. t1993). Three-dimensional analvsis of Rothman, K. J. (1986). Modern Epidemiology. Little, Brown & Co., lung structure and its application to pulmonary dosimetry modetsBoston. In Toxicology of the Lung. (D. E. Gardner et al., Eds.). Raven Press, New York. Rothman, K. J. (1982). Causation and causal inference. In Cancer Ep- iderniology and Prevention (D. Schottenfeld and J. F. Fraumeni, Mohtashamipur, E., Norpoth, K., and Lieder, F. (1985). Isolation of. . Eds.), pp. 15-22. Saunders, Philadelphia. frameshift mutagens from smoker's urine: Experiences with three Samet, J. M. (1992). Environmental tobacco smoke. In Environmental concentration methods. Carcinogenesis 6, 783 r8$ Toxicants (M. Lippmann, Ed.). Van Nostrand & Reinhold, New MONICA (1994). Ecological analysis of the association between mor- . York. tality and major risk factors for cardiovascuIar disease. Int. J. Epi- Sandler, D. P.. Comstock, G. W., Helsing,: K. J., et al., (1989). Deaths demiol. 23, 505-516. from all causes in nonsmokers who lived with smokers. Am. J. Pub- Muscat, J. E., Harris, R. E., Haley, N. J., et al. (1991). Cigarette smok- tic Health 79, 163-167. ing and plasma cholesterol. Arn. Heart J. 121, 141-147. Schmeltz, J., Tosk,. J., Jacobs, G., et al. (1977). Redox potential and National Academy of Sciences (NAS) (1983). Risk .dssessmerzt in the... .,, .: quin"one content of cigarette smoke. Anal. Chem. 49, 1924-1929. Federal Government: Managing the Process,. National Academy Selby, J. V., Austin, M. A., Sandholzer, C., et al. (1994). Environmen- Press, Washington, DC. tal and behavioral influences on plasma lipoprotein(a) concentra- National Research Council (NAS), National Academy of Sciences .. :. tion in womentwtns:Prev. Med. 23, 345-353. (1986). Environmental Tobacco Smoke: Measuring Exposures and Sheps, D. S., Adams, K. F., and Bromberg, P. A. (1987). Lack of effect Assessing Health Effects. National Academy. Pr..ess, _Washington, of low levels of carboxyhemoglobin on cardiovascular function in DC. patients with ischemic heart disease. Arch. Environ. Health 42, NCEP (1988). Report of the National Cholesterol Education Program 108-116. Expert Panel on detection, evaluation, and treatment of high blood Shields, P. G., Bowman, E. D., Harrington, A. M., et al. (1993). Poly- cholesterol in adults. Arch. Int. Med. 148, 36-69. . -. - cyclic aromatic hydrocarbon-DNA adducts in human lung and can- OECD (1984). The international expert group for hazard assessment cer susceptibility genes. Cancer Res. 53, 3486-3492. of the OECD Chemical Program. Data interpretation guidelines for Sonnenfeld, G., and Wilson, D. M. (1987). The effect ofsmoke age and

PASSIVE StiIOKING .aNU COEZON1.RY HEART S1'NDtZONfES , tlution on t e cytotoxictty. a sidestream {passivel smoke. ToxicoL._ ment of Health. Education, and Welfare. PublLc Hi Lett. 35, 89-94. Publication No. 1103, Washingtoin: DC, ' Soriie, P., Lakatos, E., Kannel, W. B., at a1. (1987). Influence of ciga- vati Doorn, R., Bos. R. P., Leijdekkers, C. M., at aL (197. rette smoking on lung function at baseline and at follow-up in 14 concentration and mutagenicity of urine from cigaret ,. _ years: The Framingham study. J. Chronic Dis. 40, 849-856. " Int. Arch. C7ccup. Environ: Health 43, 159-166. Sorsa. M.. Husgafvel-Pursiainen, K.. Jaerventaus, H.. at al (19f391 . van Schooten..F. J., Hillebrand. M. J., van_Leeuwen, F. E., ea Cytogenetic elfects of tobacco smoke exposure'among intioluntary Poivcvclic aromatic hydrocarbons-DNA adducts in wh, smokers. Mutat. Res. 222, 111-116. . cells irom lung cancer patients: Itio correlation with adduct it Steenland, K. (1992). Passive smoking and the risk of heart disease, lung Carcinogeiiesis 13, 987-993. _. ... . ,. J.41fA 267, 94-99.. ". _ Varttainen; E., Puska, P., Pekkanen; J., et al: (1994). Coconarv ht S'tehle. G., Hinohara, S.. Cremer. P., at aL. (19911. Differences in risk disease in FinIand. Br. Med. J. 309, 23-?7. factor patterns for coronary heart disease in China Japan, and Ger- --L astag, E.. :Ylattis, H., Kohler,.D,, at a(. (1985). Mucociliary clearance many. Klin. Wochenschr. 69, 629-632 and airway obstruction in smokers, ex-smokers, and normal sub- Svendsen. K. H.. Kuller, L. H.. Martin. M. •J.. et aL. 1193"). Edects 01 ;ects who never smoked. Eur. J. Respir. Dis. 139, 93-100. passive smoking in the multiple risk lactor :niervention trtaL Am Vicari A. M.• 'blargonato, A.• iVlacagni. A., at al. (1988). Effects of J. Epidemiol. 126, 783 795. . acute smoking on the hemostatic.system in humans. L'Iin. Cardiol,. 11, 538-540. Tanigawa, T„ Yoshikawa, T., Takahashi, S., et al. ~ 1994?. Spin trap- _ ping of superoxide in aqueous solutions of fresh and aged cigarette ?ineis: P., aiid Ronco, G. (1992). [nterindividual variation ion carcin- smoke., Fr2e Radicals Biol. .'Lfed. 17, 361-365. ogen metabolism and bladder cancer risk. Eixciron. Health Perspect. Tavlor, A. E., and .Johnson, D. C. (1992). Environmental tobacco 98, 9a-99. .. • _ _ smoke and cardiovascular disease: Circulatiorz $6, 699 02. tL'ald, V. •J . Idle. M L, Boreham, J.J at at. 1.1981) . Serum cotinine levels Tell, G. S.. Rutan. G. H., Kronrnal, R. A.. at al. 119941. Correlates of in pipe sr,aokers: Evidence against nicotine as cause of coronary biood pressure in`community-dweliing older a.dults. H\pertension . heart disease. Lancet 10, f fo-Lf7. .. 23, 59-62. '..' _Wannamethee, G„ and Shaper, A. G. (1992).Blood lipids: The rela- _ -. Tennant, R. W. (1988). Relationship between in vitro genet:c toxtcity tionship wtth alcohol intake, snzoking, and body weight. J. Epide- miol: Community Health 46, 197-202. and carcinogenicity studies in animals. Ann. N, Y. Acad. Sct 534, 127-132. . .':- e s, A. .(1994 ). Passtve smoking as a cause of heart disease. J. Am. U.S. Em•ironmental Protection Agency. IUSEP?.) i 1992a). Science Coll Cardiol 24. 546.,-5a4,. Advisorv Board, Environmental Tobacco Smoke Review Panel. Wennmalm, A., Benthin. G.• Granstrom: E. F., at al. (1991). Relation July 21 _>2; 1992. Crystal City Hoiiday Inn, Arltngton, VA 22202. between tobacco use and urinary ezcretion of Thromboxane AZ and Transcript from Barrera Assoctates Inc., 733 15th Street NW, Suite Prostacyclin metabolites in young men. Circulation 83, 1698-1704. 1120, Li•ashin6non, DC. . --- - -- Weston,. A., Caporaso, N. E., Taghizadeh. K„ at al. (1991). Measure- t;..S. Environmental Protection Agency (USEPA) (1992b),...Respira- .. ment of 4-aminobiphenvi•hemoglobin adducts in lung cancer casess tory health etfects of passive smoking. Lung cancer and otherdisor- and controls, Cancer Res. 51, 5'219-5223. - ders. May 1992. Office of Research and Development, Washington Wilson, T, W., Kaplan, G. A., Kauhanen, .J., at al. (1993). Association DC. between plasma fibrinogen concentration and five socioeconomic U.S. Environmental Protection Agency (USEPA) t1992c1. Respira- irrdices in the Kuopio ischemic heart disease risk factor study. Am. tory health etfects ofpassive smoking. Lung cancer and other discr J . Epi emiol. 137, 292-300. ders. December 1992, O.ffice of Rescarch mr d i~e : clnoment, « a=h W : nder. E. L. (1987). Workshop on: guidelines.to the epidemiology of itigton. DC. u eak assoctations. Prim. Med. 16, 139-141. U.S. Ern•ironmental Protectton.Agency it1SEP.a1 i;9a6) r,ulde in; , `,•,mashtta. F., Kobatashi, S„ Yamaguchi, S.. at aL. (1988). Effect of for carcinogen risk assessment. Fed. Regwc. 51 i looi, 33yys-,~4uUa: smoking on regional cerebral blood How in normai aged volunteers. September 24. Gerontology 34,199-204. t`.S. Envirohmental Protection Agency (USEPA) (1984). Revised Zayas, J. G., Man, G. C., and King, IvL (1990). Tracheal mucus rheol- evaluation of the health effects associated with carbon monoxide ogy in patients undergoing diagnostic bronchoscopy: Interrelation exposure: An addendum to the 1979 EPA air quality criteria docu- with smoking and cancer. Am. Rec. Respir. Dis. 141, 1107-1113. ment 1`or carbon rnonoxtde. EPA 60oj8-83-033F. U.S. Government .....,Zeiger, E., Hase.man, J. K., Shelby, M. D., at al. (1990). Evaluation of Printing Oflice, Washington, DC. ,. four iii vitro genetic toxicity tests for predicturig rodent carcinoge- L'.S. Occupational Safety and Health Administration '(USOSHA) nicity: ConBrmation of earlier results with 41 additional chemicals. (1994). Indoor air qual.ity; Proposed Rule. Fed. Regist. 59(65), . Enuirorz. Mol. Mutagen. Suppl. 18, 1-14. 15969-16039; April 5. Zhu, B., Sun, Y., Sievers, R. E., at al. (1994). Exposure to environmen- L`SSG (1986). The health consequences of involuntary smoking; a re- " tal tobacco smoke increases rriyocardial infarct size in rats. Circula- port of the Surgeon General. U.S. Public Health Service, Rockville, tion 89, 1282-1290. MD• Zhu, B. Q., Sun, Y. P., Sievers, R. E., at a!. (1993). Passive smoking USSG (1964). Smoking an.d.health. Report of the advisory committee increases experimeritalatherosclerosis in cholesterol-fed rabbits. J. 225-232. to the Surgeon General ofthe Public Heal,th .S.ervice. Lt.S. Depart- Am. Coll. Cardiot 21.