Tobacco Documents Online

ENVIRONMENTAL TOBACCO BXO~z A COMPENDIUM OF TECHNICAL INTO~TZON May 1991 DRAFT Q /// This document is a preliminary draft. DO not cite or quote. The contents represent only those views of the individual chapter authors. It should not be construed as representing the views or policies of the participating organizations. /

Draft - Do not cite or quote This compendium of technical perspectives on Environmental Tobacco Smoke (ETS) is intended to be a useful resource document for a diverse audience, including: decislon-makers such as labor and management officials concerned with workplace exposures, public health officials and corporate medical directors who are concerned with making health policy reco~mendations, educators, industrial hygienists and safety officers, ETS researchers, indoor air pollution investigators, and legislators who are considering legislation to restrict smoking in workplaces, restaurants, and public access buildings. Although the technical level varies, even the more technical treatments do not require a specialist's knowledge for understanding. There are eleven chapters in this compilation, including health effects of active smoking in adults a~d passive smoking in children and adults, ETS exposure and dosimetry, comfort aspects, ventilation and ETS, public beliefs about the harm of ETS and attitudes toward controlsr and effective workplace smoking policies, each of which is aimed at a somewhat different audience. Although not all chapters will appeal equally to such a varied group, it is hoped that the technical information in this document, written by experts in the field, will provide information necessary to allow the public, corporations, government agencies, and legislators to make well-lnformed choices regarding exposure to ETS. This perspective on ETS reflects the viewpoints and expertise of authors who were selected based upon their publications and recognition as experts on various aspects of ETS. Accordingly, the opinions expressed do not necessarily represent the official policies of the sponsoring agencies. This document is the result of a coordinated effort jointly sponsored and produced by the Environmental Protection Agency (EPA) (chapters 2,3,4,6,7, and 8), the National Cancer Institute (NCI) (chapters 1,5), the Office on Smoking and Health (centers for Disease Control) (chapter 9), the National Heart, Lung, and Blood Institute (chapter i0), and the office of Disease Prevention and Health Promotion (Department of Health and Human Services) (chapter Ii). The editors acknowledge with gratitude the following distinguished scientists, physicians, and others who lent their support to the development of this document by contributing critical reviews of the various manuscripts, by coordinating manuscript preparation, or assisting in other ways. Mr. Robert Axelrad, H.S. Environmental Protection Agency, 2

Draf~ - Do no~ cite or quote Washington, DC (Sponsor) Dr. Lois Biener, Miriam Hospital, Brown University, Providence, RI Ronald Davis, M.D. Office on Smoking & Health, Centers for Disease Control, Rcckville, MD (Sponsor) James W. Davis, M.D., Veterans' Administration Hospital, Kansas City, MO MS. Hildy Dillon, American Lung Association, New York, NY Dr. Cedric Garland, Dept. of Community Medicine, University of California, San Diego, CA Dr. Stanton A. Glantz, Department of Cardiology, University of California Medical School, San Franscisco, CA Dr. Lawrence Garfinkel, American cancer Society, New York, NY Dr. Katherine Hammond, Dept. of Family & Community Medicine University of Massachusetts Medical Center, Worcester, M-A Dr. Marvin Rristein, State University of New York, Stony Brook, NY State Univ. of New York, Stony Brook Dr. Joellen Lewtas, Office of Research & Development, U.S. Environmental Protection Agency, Research Triangle Park, NC Dr. Alfred H. Lowrey, Laboratory for the Structure of Matter, Naval Research Laboratory, Washington, DC Henry HcIntosh, M.D., A~erican College of Cardiology, Washington, DC Dr. Michael McGinnis, Office of Disease Prevention and Health Promotion, Public Health SerVice, Washington, DC (sponsor) Matthew Meyer, Esq., Coalition on Smoking or Health, Washington, DC D~. Gregory Morosco, Health Education Branch, National Heart, Lung, and Blood Institute, Bethesda, MD (Sponsor) Dr. Demetrios Moschandreas, Illinois Institute of Technology Research Ynstitute, Chicago, IL Dr. David Mudarri, U.S° Environmental Protection Agency, Washington, DC Dr. Terry Pechacek, Smeking, Tobacco, and Cancer Program, National Cancer Institute, Bethesda, MD (Sponsor)

Draft - Do not ¢itm or quota Mr. James Hepace, U.S. Environmental Protection Agency, Washington, DC (Editor) Mr. Donald Shcpland, National Cancer Institute, Bethesda, MD (Editor) John Slade, M.D., Dept, of Medicine, St. Peter's Medical Center, Rutgers University, New Brunswick, NJ Dimitri Trichopoulos, M.D., DrPN, HarVard School of Public Health, Bostcnl HA The editors also acknowledge the comments of the tobacco industry. Mr. Samuel D. Chilcote, Jr., President~ The Tobacco Institute, Washington, DC Dr. Thomas Borelli, Phillip Morris USA, Richmond VA

Draft - Do not cite or quote TABLR OF CONTENTS Chapter I. Effects of Smoking on Smokers. Donald Shopland .............................................. 9 Chapter 2. Environmental Concentrations of ETS. Jo~ McCarthy, Elizabeth Miesner, and John D. Spengler ..................................................... 16 Chapter 3. Measuring Exposure to Environmental Tobacco Smoke. Brian P. Leaderer ............................ ~ ............. 31 Chapter 4. Absorption of Smoke constituents hy Nonsmokers. Dietrich Hoffmann, Klaus D. Brunnemann, and Nancy J. Haley ........................................... 43 Chapter 5. Environmental Tobacco Smoke and Cancer. Jonathan M. Samet ........................................... 67 Chapter 6. Passive Smoking and Heart Disease. S~anton A. Glantz and William W. Parmley ..................... 81 Chapter 7. Exposure Assessment in Passive Smoking. James L. Repace ........................................... 112 Chapter 8. The Odor and Irritation of Environmental Tobacco Smoke. William S. cain ................... • ......................... 137 Chapter 9. Passive Smoking -- Eeliefs, Attitudes, and EXpOsures in the United States. Thomas E. Novotny ........................................... 152 Chapter I0. Passive Smoking and Daycare. Glen L. Bennett ............. ~ .............................. 180 Chapter Ii. NO Smoking Policies at the Worksite: A look at what companies are doing today. Ruth Behrens ............................................... 19T Chapter ii Appendix: Economic Justification for Worksite Smoking Policies. Ruth Behrens ............................................... 219 5

Draft - Do not cite or quote INTRODUCTZON In 1986, the Surgeon General and the National Research Council, the latter under contrac~ to EPA, examined the health effects of the breathing of Environmental Tobacco Smoke (ETS) by nonsmokers (also known as involuntary or passive smoking). They agreed that passive smoking caused lung cancer in ncnsmoking adults, caused increased rates of respiratory infections in children, caused acute noxious effects in many nonsmokers, and was a major contributor to indoor air pollution. Subsequent to the publication of these documents, smoking restrictions began to proliferate. However, a number of diverse technical questions arose concerning public attitudes toward smoking restrictions, health and comfor~ effects, factors affecting exposure, measuring environmental concentrations of ETS, effects of ventilation on ETS and indoor air quality, nonsmokers' u~take of tobacco combustion products, and corporate experience in effective smoking policy, all comprise chapters in this compendium. ~n the interest cf providing answers to this complex of questions, this technical compendium was commissioned. A brief summary of each chapter follows. Chapter 1 demonstrates that high dose exposures to tobacco smoke, i.e., the effects of smoking on smokers, are very toxic, causing cancers, cardiovascular diseases, and respiratory diseases. It is graphically illustrated why cigarette smoking is now recognized as the Nation's. single largest cause of premature death and disability. Chapter 2 reviews studies of the concentrations of certain ETS constituents observed in homes, offices, and other locations by personal exposure monitors. It is concluded that £TS is the primary contaminant contributing to respirable particulate air pollution, and contributes substantially to other indoor contamininants such as benzene, carbon monoxide, and others. Even in low doses, tobacco s~oke contains a wide variety of toxins, including many carcinogens. Chapter 3 treats the methods of assessing nonsmoker's exposure to environmental tobacco smoke by atmospheric markers, and the measurement of these marker substances in indoor air. It is concluded that atmospheric monitoring for respirable particles or nicotine from ETS is critical for assessing exposures and control efforts, and that a number of reliable methods are available for such monitoring. chapter 4 provides a detailed treatment of the absorption and metabolism of tobacco combustion products by nonsmokers. It shows that absorption has been conclusively demonstrated by studies of nicotine and its metabolite, cotinine, in the body fuids of nonsmokers, and that such biomarkers represent a reliable specific method for assaying the level of uptake of ETS. This exemplifies

Draft - Do not cite or quote that low dome exposure to tobacco smoke leads to the absorption of toxins from the smoke in amounts sufficient to potentially cause disease. Chapter B discusses the evidence that low dose exposure to tobacco smoke has been observed to increase the risk of lung cancer i~ nonsmokers, and discusses conclusions of the World Health Organization, the National Research Council, and the U.S. Surgeon General that ETS exposure increases lung cancer incidence in ~cnsmokers. Chapter 6 discusses the evidmnce that low dose exposure to tobacco smoke has been observed to increase the risk of hearu diseases in nonsmokers, and discusses the epidemiological, biochemical, and biological bases for this inference. It is concluded that the combined epidemiological and physlological evidence suggests that ETS exposure is a cause of heart disease in nonsmokers. Chapter 7 investigates the assessment of nonsmokers' exposures tc ETS by mathematical modeling, atmospheric indicators, and % blomarkers in body fluids. Exposures assessed by these various methods produce consistent results. Because of the large source strengtm of tobacco-burning products, exposure to environmental tobacco smoke is inadequately controlled by measures short cf physical separation of smokers and nonsmokers on different ventilation systems, making ETS a significant indoor pollutant of buildings. Chapter 8 explores the effects of ventilation on the perception of odor and irritation from ETS in both nonsmokers and smokers, and shows that attempts to control the odor and irritation of ETS through ventilation and air cleaning have significant limitations. Chapter 9 shows through national surveys of trends in public attitudes, that the general public, including both smokers and nonsmokers, believe that tobacco smoke polluted air is harmful and a large majority find it irritating. There is widespread support for restrictions against smoking, particularly in the workplace. Chapter IC discusses the evidence that smoking both at home and in daycare centers harms children and infants from tobacco- smoke polluted air. This has direct implications for public education of both parents and daycare providers, as well as for state policies and regulations affecting facilities which offer daycare. Chapter II points out the COrydon solution to the problem of ETS ls source control, and examines features of corporate smoking policies in the workplace, with attention to benefits, incentives, employee and union involvement, and education. Case histories are

Draft - Do n¢~ cite or quote ! discussed inv:iving several major corporations, detailing problems encountered and successes. It ks concluded that smoke free workplaces have been achieved in a variety of se~tings. If thoughtfully implemented, they enjoy widespread support.

Draft - Do not cite or quote CHAPTZR 1 BYYECTB OF SMOKING ON SMOKERB Donald 8hopland coordinator Smoking and Tobacco Control Program National cancer ~nstitutes Bethesda, MD Cigarette smoking is the nationts leading cause of premature death and disability. In 1988, smoking caused approximately 390,000 deaths in the United States (Figure i). By 1991, this number had increased to 440,000. In addition, tens of millions of people suffer from chronic disabling diseases and conditions caused or aggrevated by smoking. Every medical authority and organization who has objectively examined the evidence linking smoking to early death and disability has reached a similar conclusion. The 8vidence that smoking is a major health threat is staggering: over 50,000 citations from dozens of cultures are in the scientific literature. Smoking causes or is associated with cancers of the lung and bronchus, larynx, lip and oral cavity, bladder, pancreas, kidney, stomach and cervix, coronary artery disease, cerebrovascular disease (stroke), atherosclerotic aortic aneurysm, atherosclerotic peripheral vascular disease, chronic bronchitis, emphysema, low birth weight babies, and unsuccessful pregnancy. This chapter concentrates on the relationship between active smoking and three diseases caused by ETS -- lung cancer, heart disease, and nonmalignant lung disease. While there are qualitative differences between the mainstream smoke inhaled by the smoker and the ETS nonsmokers inhale, both forms of tobacco smoke contain the same carcinogens, irritants, and other toxins. The effects of high doses of smoke on smokers thus provide an indication of what effects low dose exposures of ETS would be expected to have on nonsmokers. This connection is particularly important because the diseases active smoking causes exhibit dose- response relationships, with higher doses producing greater effects. Because no threshold ham been demonstrated for the carcinogenic and other effects of tobacco smoke on the body, the existence of a dose-response relationship suggests that ETS would provide similar, but smaller, dangers than active smoking. Cancer Most estimates in the scientific literature indicate that nearly one-third of all U.S. cancer deaths result from cigarette smoking. Of the approximately 136,000 cancer deaths which occurred in 1985 because of smoking, 106,00D are of the lung (Figure I). Lung cancer alone is responsible for fully one-quarter of all

Draft - Do not cite or quote cancer mortality; were it not for the increasing number of deaths from lung cancer produced by smoking, we would be experiencing a substantial decline in the cancer death rate in the United StBtes. Approximately 85 to 90 percent of all lung cancer deaths are smoking related. The evidence linking smcking and excess cancer moz~ality is so strong that only the tobacco lobby continues to claim that no causative role has been established. An examination of the association between cigarette smoking and lung cancer graphically illustrates smokingls role in ~he causation of neoplastic diseases. Tobacco smoke contains at least 43 }~no~nn or suspected human carcinogens (Table !), several of which are regulated by the federal government as environmental toxins. There is no known threshold for the carcinogenic effects of these agents. A host of epidemiological studies published during the last two decades provides an abundance of data which demonstrate that exposure to these carcinogens because of smoking leads to an increase in cancer deaths. In particular are the major prospective studies on smoking and health. These studies, conducted in the United States, Canada, England, Japan and Sweden represent some of the largest population based studies ever undertaken by medical science (Table 2). They involved enrolling healthy men and women into a study design and then followed these individual over time. Numerous factor about them were recorded including where they lived, their occupations, dietary habits, whether they used tobacco, access to health care, and many other factors. As a group, these eight studies in the United States, the U.S. Veteran's Study and the American Cancer Society (ACS) 2S-state Study contained cohorts of iS0,000 and I million persons respectively. The Veteran's Study continues to this day and this cohort has been followed prospectively for 26 years. These studies convincingly demonstrate that smoking causes cancer. Lung Cancer Lung canceE mortality rates are strongly influenced by the total dose of cigarette smoke received. If one smokes ~ore cigarettes per day, inhales deeply, if they started smoking at an early age had has smoked for many years, the risk for lung cancer .is increased dramatically. The most often used measure to gauge lung cancer mortality is the number of cigarettes consumed daily. In the ACS 2S-state study, for example, among males smoking less than 1/2 pack per day their lung cancer rate was nearly 5 times greater than that of a nonsmoker. With each increase ix the number of cigarettes consumed daily, a corresponding increase in lung cancer mortality is observed (Figure 2), For those smokers consuming two or more packs daily, their lung cancer mortality is about 24 times greater than i0

Draft - Do not cite or quote that of the nonsmoker. At the other extreme, even light smokers, who consume only 1-9 cigarettes per day, see a quadrupling of the risk of lung cancer. An inverse dose-response relationship exists between an early age of regular smoking and lung cancer mortality. In the U.S. Veterans Study, those smokers who started smoking in their early teens had substantially higher lung cancer death rates than those who star~ed in their late teens or twenties (Figure 3). Those who began smoking before age 15 experienced a 19-fold greater lung cancer mortality, compared to a slightly greater than 5-fold excess risk for those who initiated their behavior after age 25. These results demonstrate that a dose-response relationship exists for exposure to the carcinogens in cigarette smoke and the risk of death from lung c~ncer: the greater the lifetime exposure to tobacco smeke, the greater the risk. Further evidence for the existence of a dose-response relationship comes from follow-up of people who stop smoking and so remove the exposure from the carcinogenic agents in mainstream 4moke. When an individual stops smoking, his or her lung cancer risk declines relative to the continuing smoker. After about 15 years off cigarettes the former smokerrs lung cancer risk approaches that of the life-long nonsmoker. However, it appears that some excess risk may be carried throughout life. This residual risk is strongly influenced by the indivldual's total lifetime exposure to the agent and the total number of years of smoking cessation. The presence of a dose-response relationship between smoking and lung cancer, combined with the fact that there are significant elevations in risk ~ssociated with even the lowest levels of smoking, demonstrates that there is no threshold for the carcinogenic effects of cigarette smoke. This result from active smokers is consistent with the observed elevations of lung cancer risk among nonsmokers exposed to ETS. Corona=y Heart Disease In contrast to cancer, in which smoking produces the disease through the cumulative effects of long term exposure to the carcinogens and co-carcinogens in the smoke, smoking effects the cardiovascular system immediately as well as over the long term. The carbon monoxide in the smoke reduces the oxygen carrying capacity of the blood by binding to hemoglobin competitively with oxygen. Nicotine is a vasoconstrictor, which increases blood pressure and narrows coronary arteries. Smoking causes release of catecholamine, which increase blood pressure and heart rate. Smoking also increases platelet aggregation and adhesion, which contributes to the development of atherosclerosis. All these ii

Draft - Do no~ cite or quote effects occur i~ediately upon smoking and resolve relatively quickly after stopping smoklng. As a resul~, one year after stopping smoking, the excess risk of death from hear~ disease falls by half; the sane drop in risk for lung cancer takes 10 years. As with cancer, these effects exhibit a dose-rmsponse relationship, with greater mere smokinq and smoking in combination with other heart disease risk factors, increaslng ~he risk of death from coronary heart disease. As with cancer, there Is no threshol~ for these effects, $o ~he effects of active smoking on the heart and cardiovascular system support the blological plausibili~y of the observed effects of ETS on ~he heart. Coronary heart disease (¢HD) continues to be this nation's leading cause of death, and for nearly 20 years, medical research has shown that smoking is one of tha major Independen~ risk factors or causes of CHD (along with high blood pressure and high cholesterol levels). In tha final report of the Pooling Project, an interaction between smoking and other ~isk factors was observed (Figure 4). Each independent risk factor contributed about the same increased level of-risk, ~owever, whQ~ ~o Or ~orQ factors were present, ~he risk of a major CHD event was increased beyond ~he sum of ~he independ~n~ risk -- thus, synergistic effect was crea~ed when two or mere risk factors wer~ present° Over~ll, smokers have a 70% greater CHD death rate, a two- to fourfold grmater incidence of CHD, and a two- t~ fourfold ~reater risk for sudden death than nonsmokers. Dose-response relationships between ~igarette smoking and CHD mortality have been demonstrated for several measures of exposur~ to cigarettes, includln~ the number of ~igaret~es smoked per day, the depth of inhalation, age at which smoking began, and ~he number of years of smoking. Smoking cigarettes with reduced yields of ~ar and nicotine does no~ reduce CHD risk, probably becaus~ ~hese cigarettes do not have reduced yields of carbon mcnoxidl and ether combustion produc~s which afflct the cardiovascular system. The independent risk of CHD fo~ smoking is greater a~ ~he younger age groups al~hough ~he greatest number of excess CHD deaths due to smoking actually ~ccurs in the older age groups (Figure 5). Smoking has also been shown to increase the risk for other cardlov~scular diseases, including peripheral vascular disease, cerebrov~scular dlsease (at younger age gr~up~), a~d aortic a~urysms. For women, s~oki~g "ca~ i~teract wi~h oral contraceptives ~o greatly i~crease ~he risk fac~o~ f~r ~a~al and nonfatal myocardial infarction and suhar~chnoid hemorrhage. Smokers exhibi~ ~ore a~herosclerosis, bo~h in the aorta and coronary ar~erles° Cigarette smokers who continue ~o smoke following ~ransluminal coronary angioplasty appear more likely ~o require repeat angioplasty than nonsmokers, suggesting tha~ ~he effects of smoking on atherosclerosis occur quickly. The polycyclic aromatic hydrocarbons which rmsult from the combustion 12

Draft - DO not cite or quote of the smoking materials contribute to these effects• The increase in platelet adhesion observed in smokers also contributes to the development of atherosclmrotic plaque. Cigarette smoking aggravates the conditions of people with CHD. Smokers have a more difficult course following coronary artery bypass surgery. Smokers who experience angina pectoris have a higher risk of death than nonsmokers, a poorer prognosis following non-fatal myocardial infarction, and a greater risk of sudden death. Smoking increases the risk of silent iscbemia in patients with stable angina. Many public health estimates place the total number of excess cardiovascular disease (including stroke) deaths due to smoking to be greater than those due to cancer (Figure i). Up to 30 percent of all CHD deaths may be due to cigarette smoking and its interaction with other risk factors. These effects all exhibit a dose-response relationship with no threshold in active smokers, with detectable damage even among light smekers. These facts support the biological plausability of ~he evidence linking ETS with heart disease in nonsmokers. • Nonmalignan~ Respiratory Diseases In addition to causing lung cancer, smoking causes or aggravates several related nonmalignant respiratory diseases, including emphysema, asthma, chronic bronchitis, and chronic obstructive pulmonary disease (COPD). While the number of s~oking-i~duced deaths classified due to chro~ic obstructive pulmonary disease (COPD) is smaller than for cancer or cardiovascular disease (Figure i), COPD afflicts about 12 million Americans. Even if not fatal, COPD and related disorders such as emphysema severely debilitate the victim and represent a substantial number of people who become disabled due to their condition, unable to work or even seek employment. For many years cigarette smoking has been known to increase the risk of developing and dying from COPD. Even the first Surgeon General's Report issued in 1964 identified a causative role between smoking and chronic bronchitis. AS with lung cancer, the risk of contracting and dying from COPD is substantially elevated among smokers (Figure 6) and this risk increases with an increased dose of cigarette smoke received; as with the other smoking-induced diseases discussed in this chapter, there is a positive dose- response relationship. Mortality rations for COPD in smokers versus nonsmokers are very high, exceeding 30 to 1 for heavy smokers (Figure 7). Smoking also has a dramatic effect on lung function. The normal rate of lung function decline with increasing age is accelerated in cigarette smokers (Figure 8). These effects 13

Draft - Do not mite or quote probably reflect damage to the small airways of the lungs as well as a thickening and increased reactivity of the airways in response to chronic exposure to the irritants in cigarette smoke. The volume an individual and exhale in one second of forced expiration (?EVl) is a measure of small airway function. Figure 9 shows that FEVI falls in a dose-dependent manner as the amount of smoking increases. There is no safe level of exposure: there is a measurable decrement in pulmonary function even among light smokers. Stopping smoking partially reverses the nonmalignant effects of the respiratory system (Figure 8). When one stops smoking, the decline in lung function with age resembles that of a nonsmoker, but a permanent decrement in lung function remains, indicating some permanent damage. The amount of this permanent deficit depends on the duration and intensity of smoking. ETS exposure produces similar, but more modest nonmalignant pulmonary effects. FEVI is reduced in passive smokers among both children and adults to levels similar to that observed in light smokers. Children of parents who smoke develop more asthma, bronchitis and other respiratory problems. The rate of lung ~evelopment in children exposed to ETS is smaller than that of unexposed children. These effects of ETS are what one would expect based on the effects of active smoking. Conclusions This chapter has reviewed the effects of active smoking in on these cancers, heart disease, and nonmalignant pulmonary diseases which have also been identified with passive smoking. In each case, cigarette smoking significantly increased the risk of disease in smokers in a dose-dependent manner. There is no evidence of a threshold level for adverse effects. Because ETS is similar to (bu~ more toxic than) mainstream smoke, these effects on the smoker help provide evidence for the biological plausibility for the epidemiological evidence linking ETS with lung cancer, heart disease, and nonmalignant respiratory disorders, after accounting for the lower dose the involuntary smoker receives. I. There is a dose-response relationship between exposure to tobacco smoke and the diseases of smoking. 2. There are no discernable thresholds of exposure for the diseases of smoking. 3. Adverse health effects observed in smokers provide biological plausibility for the occurrence of those diseases in nonsmokers. 14

Draf~ - Do no~ cite cr quote TABLES ~ F~GURES+ ~PT~R ~r

¥~QURR ~, US Deaths Attributed to Smoking in 1985 Source: US Surgeon General, 1989 CVD CHD 28000 115000 Cancer, lung 106000 Cancer, ether 30000 COPD 57000 Other 54000 H 11 t D O O it 43 p. D O II 0 m

Draf~ - DO not cite cr quo~e ~ ~ Men -f irm 1~ zcdo,s

Draft - Do not cite or quote 2O 18 10 8- 0 FIGUR~ 3. (1989 SURGEON GENERALIS RE~ORT, p. 49)

FIGUR2 40 Lung canceP mortafity ratios for males, by age began smoking -- U.S. Veterans' Study 20 •. 18.7 15 lO 5 1,0 Nonsmoker 9.5, 25+ 20-24 15-19 <15 Age began smoking (in years) t~ I 0 fl p. m 0 i-I 0

YZGURX 5o Major risk factor combinations, lO-year incidence of first major coronary events, men age 30-59 at entry, Pooling project 0 0 0 L- G) 0. (1) n¢ 200 180 160 140 120 100 80 60 4O 20 23 54 189 103 None SM C or H SM & C C & H All 3 ol3 Only Only or . (No $M) SM&H Risk Factor Status at Entry SM = smoker. C = choleslerol, H = hypertension -n -4, dr I 0 0 ,p- 0 0 0

Coronary heart disease edeaths, smokers vs;a~* onsmokers Deaths per 100,000 men - ~o not: ¢il:e or ~0%~ Nonsmokers ~ Smokers YZ GUI~I~ 6,

COLD deaths on._ oo oo: o~° or ~o~. smokers vs. nonsmokers Deaths per 100,000 persons 500 400 300 200 .100 0 Smoker/- ~Nonsmokers 35-44 45-54 55-64 65-74 75-84 Age group YZGUI~I~ 7.

0 o 4J .rl U 0 0 ! t) KI 4O 30 k,- 20 10 FIGURE 8* COLD mortality ratios for men and women, by number of cigarettes smoked per day, British Physicians' Study[llll Ma,. m Fernale 38.0 32.0 26.0 17.0 1.0 Nonsmoker 1-14 15-24 25+ Cigarettes per day

Dra:~ - Do not c£Re or c;uo~e }'ZGURE 10. --P~rcen( dls|nbucion of" ~redlc[."d vaiu~s o1" Por¢~.d expiratory volum~ in l -~.c IFEV i I ia su bje~[s with ~r~ing pack-~,e~r~ otsmok- SotJ]tc~ S~aO~l~ ~t L LYe7): ~k~ ~ ~J. L I ~~

Draft - Do not cite cr quote ~--US~ ~t,log Im ~IBB

(~Gf 1989; p, B6-87) ~Tum~ig©.l¢ mgenls ~ Ioblcco |rid I~ $m~e &r i~ml, llc mmlne$ 2 1 .luidm~ ~2(X~ nS Sul~;c,u In,a JeqLmle Napl~hylamin© 1-22 .| S.l~icl~l S.l~:~ic.'~ 4.Am.~,bil)l,~l,yl 2 $~| Sulr"cxm $ulfN:~m /*ldr;~ ) del A~clal~ch~,l¢" ~ 4~14roll IB I.~11~ rag" Sulhc~m NA Cump~bu~.d$ B©nrcac I~R~| Sull~C'~,~ Sul flcl~m I. I I ).wl~hylhydr ui~ hl~ 142 ~l~ .~m I[Ic i~ m ~A 2 Nitlopr~ar~ ~73~1 21 pl~ S.l~c~l HA IEihykacl.am~¢ ]10-)75 n| ~ll r,| Solf~ ~,~ N.•. Vin~ chloride ~16q Sul~cltml Su(Fu.~ I,~r ganl¢ ¢ompou.)dl HKk¢I ~VNI ~1[)1) ~,E @ b(ll~ .~| Sul~l~.~nl I I,I]ned I~,h.~,u.,.2 ~O O 2-1 2 iK'i ()113 I II I~i HA NA • ~k, I:pu~, I~¢p~ ,~1 ib~ ~m~-~l~ n~ ~l¢~,r,c ~ ~l~q~c .~ ~ $,,~k~,| ~ II~l~k~ i i'~#n~ p.~l,..,~d ~d~ ~ ,~ Ev~dew~ I~ IARC cv,am M&l.it icm i~ ¢1.~ Ir~cnKII ~ Fleccsscd t~,~c~ ~¢ U (t t U o 0 rt (I It J 0 0 0

TABL~ 3. Outline of Eight Major Prospective Studies Aulhora po~p¢lallon idJo 40,QO0 I.IDOO.G~tl ~.0¢0 Fimlkli ,.OOB tdl2.Q1 | • I% Age R4~Oe ~e~¢IS ~, I 36-44 ;).%04 Yurol ll~i 1940 1054 emoumeel I¢4tl Yull oi 20-22 IollOwup k,,~r o i2 Vl~rl; IlIvw8 i,pmlad Uumb~ ol i;,~M liO.OOO Iol.loo d~he P~n ye~4J of 0~.0~ |,000.a(~ ),GG4J.C~XI ~ts,(x)o O|,000 II1.0~l ll,~ K.O~l 142,~ 14,OG~ 27.~1~4) 40 30-11O 60"40 3344 lille and up 1G04 1966 11N2 11NS4 1043 I:l 1~.* II tr~uo 4"fa~ ~ 10yN~ 30.1~ 11.000 12.000 4.700 4.64~ 3,~oo.~ ~(I.~o 470.000 4110.0~4 160.~ I O cT (1 i~. (7 ID O '1 O

Draft - D~ not cite or quote Ind Ofd~ A~.~dimj to 5,~i~ 8ts~ ~ ~ Enr~Jm~t

Draft - Do not cite or quotg CXAPTZR 2 EXPOSURI8 TO INDOOR PARTXCULATI AIR POLLUTANTS John McCarthy ~hD, Elizabeth MiesneE PhD, John Spengler PhD Department of Environmental science and Physiology Harvard S~hool of public ~ealth Boston, Massachusetts 02115 Throughout our lives, we are exposed to gaseous and particulate contaminants in the air. For some airborne contaminants, our exposure is dominated By their Occurrence in outdoor air and ~he time we spend outdoors. However, even for the pollutants that have only outdoor sources, ~he air that ventilates our homes, offices, and vehicles originate outdoors. Considering chronic exposure or protection from acute episodic outdoor ~ollution events, the time we spend indoors and the protection these indoor environments provide are important considerations. In the presence of indoor sources of contaminants such as unvented combustion, evaporation of solvents, and dispersion of microbiological organisms among others, the time-activity patterns of people in their use of these indoor environments become important considerations in determining exposures. People can have very different exposures to indoor contaminants depending on social,• demographic and economic differences in the population, as well as the physical differences that exist across indoor environments. These differences are characterized by the use of the structure, its volume air flow and air exchange, the efficiency Of contaminant removal and, most importantly, the generation rate of the source itself. Thus, concentrations of air pollutants can and do vary depending on location. Outdoor pollutant levels may differ from indoo~ levels. Different indoor locations llke homes, schools or workplace can also register varying pollutant levels. An individual's total exposure to air pollutants therefore depends on the time spent in each of these microenvironments and the various concentrations of air pollutants. Time-Activity Patterns The activity patterns of people deter-mine the duration of exposure and, at times, the intensity of exposure to airborne 16

D=aft - DO not cite or quota contaminants. The amount ~f time a person spends in different nicroenvironments is influenced by age, sex, occupation, social class, and season. Letz et el. (1984) studied the time-activity patterns of 332 residents of Roane County, Tennessee. The results of study showed that these individuals spent 75% of their time in the hone. This fi~p/re was higher (84.9%) for housewives, unemployed and retired persons. Overall 10.8% of the participants time was spent at "work". Full-time employed individuals worked between 21-24% of the time. Of the remaining time, 8.5% was spent in public places, 9% in travel, and 2.8% in various other locations. Quakenboss et el. (1982) studied the time allocation for 66 family members from 19 homes in Portage, WI. Individuals were put into one of five general subgroups which are shown in Table i. Despite wide variations, each group spent most of the time at home. For all participants, total time spent indoors was 85%. More recently, Quakenbcss and his colleagues analyzed time- activity data for over ~0O individuals in the Portage, WI area. Participamts were categorized into three groups: workers, nonworkers, and students. Activity data was collected from both summer and winter seasons and is summarized in Table 2. Again all groups spent the largest percentage of their time in the home. Tin~ spent outdoors decreased from summer to winter. Infants+ because they are essentially immobile, spend most of their time in the bedroom according to a recent study by Harlos et al. [1987). The rest of their time is usually spent in the living room, kitchen, Or in travel as illustrated in Figure I. Knowing an individualTs or a population's activity patterns is not sufficient in itself to determine exposure to contaminants. Outdoor pollutants do penetrate indoors and can undergo rsactions. Indoor contaminant concentrations vary according to the source rate, air exchange and air flow, and reactions. Characterizing sources indoors will not always lead to accurate estimates of concentrations or exposures. Therefore, depending on the distribution of sources indoors and the degree of mixing, there may be considerable differences in pollutant concentrations across indoor environments. Lebret (1985) examined the respirable suspended particulate (RSP) levels in rooms while participants were smoking or within one-half hour of smoking. He found significant variation between the living room kitchen and bedroom. Ju and Spengler (1981), who studied 24-hour average concentrations of respirable particulates, also found statistically significant variation between some rooms although the absolute differences were relatively small. Monitoring 17

D~af~ - DQ not cita o~ ~ote There are a number of different Instruments available to monitor air pollutants. Often the type of instrumen~ used depends on the exposur8 of interest. ~mmediate exposures are most important when studying irritant and acute allergic responses. For this type of exposure, instruments which take short-term or instantaneous readings ate often used: the piezobalance or nephelometer are both used to measure particulates, the ecolyzer is used tD measurs carbon ~onoxide. One advantage to these types of instruments is their ability tc detect peak pollutant levels. For acute effects such as upper or lower respiratory infections, the exposures of interest range from hours to days. For increased prevalence of even a lifetime.(?) To measure these exposures, integrated or time-averaglng methods are used. These methods ihclude filters which are used to collect particles over long time periods. EXPOSURE TO AIRBORNE PARTICLZS ~ize Distribution and Composition of Particulates The distribution of particulates is essentially trimodal with peak diameters at approximately 0.0~ #m, 0.5 ~m and i0 ~m as shown in Figure 2. These size modes reflect the origins of the particles and the physical chemical processes affecting ~hem. The ultrafine fractions are typically fresh combustion emissions of aiken nuclei and condensing vapors. The submicron size (0.i-i ~m) has been c~lled the accumulation mode. Again, incomplete combustion adds particles to this size range; however, the oxidation of gases such as SO2 and NO2 tc form sulfates and nitrates are predominantly found in this range. Particles larger than I ~m can be of biological origin--fiber fragments, spores, pollens, and bacteria. Bursting bubbles and sea spr~ay can generate condensation nuclei. But it is mostly abrasion and/or erosion that generate larger particles. The fine particle fraction, or <2.5 ~m, is produced by combustion or condensation of vapors. At leash 75% of the sulfur, zinc, bromide and lead are found in this size range (Dzubay and Stevens, 1975). ~articles <2.5 ~m are very important for health reasons since these particles can reach the alveolar regions of the lungs. Particles greater than 2.5 ~m in diameter, or coarse particles, are usually formed by mechanical processes like grinding, crushing, and abrasion. At least 75% of the silicon, calcium a~d iron, elements commonly found i~ soil, appear in this size fraction (D~ubay and Stevens, 197~). particles from 2.5-10 ~m can be inhaled and can become deposited i~ the tracheobronchial 18

Draft - Do not cite or quote regions. Environmental Tobacco Smoke Environmental tobacco smoke (ETS) is a mixture of exhaled mainstream smoke and sides~ream smoke. Sidestream smoke is the smoke that is formed by smoldering between puffs of a tobacco product and is the major source of ETS. Approximately half the tobacco in a cigarette is burned in the sidestream mode. The complex mixture that the smoker inhales with each puff Of a cigarette, cigar, Or pipe is called mainstream smoke. The portion of mainstream smoke that the smoker exhales and the small amount of vapor diffusing through the wrapping of the cigar or cigarette add little %o ETS. ETS consists of fresh and aged sidestream and mainstream smoke. The particle sizes which make up ETS vary due to coagulation (the process where two or more particles col!ide and combine to form a larger particle), evaporation, and the adhesion of par%icles to surfaces. The size distribution of particles is also affected by air dilution, relative humidity and temperature. . Under controlled conditions, several researchers have measured the particle size distribution of sidestream smoke (Keith and Derrick, 1960; Porstendorfer and Schraub, 1972; Hiller et el., 1882; Leaderer st el., 1984; rngebrethsen and Sears, 1886). Based on these studies, the mass median diameter of sidestream smoke can be estimated to be between 0.2 Bm and 0.4 ~m. The mass median diameter is the diameter which divides the mass distribution in half, i.e. one half of the mass is contributed by particles larger than this diameter and one half by particles smaller. Because much of the time the tobacco is burning at substcichiometric conditions, particles are produced in the accumulation size mode. As ETS ages, the processes of coagulation cause particles to grow. This offsets mass loss due to evaporation. Composition of ETS Environmental tobacco smoke is made up of several thousand different chemical compounds. These compounds may be in the gaseous or solid phase or both. The chemical composition of sidestream smoke differs from that of mainstream smoke. Over 2,000 compounds have been measured in sidestream and mainstream smoke. Some of the constituents in nhe mainstream smoke of nonfilter cigarettes are listed in Table 3. Also given are ratios of these substances in sidestream smoke compared to mainstream smoke. A ratio of greater than 1.0 means the constituent is found in higher concentrations in sidestream smoke than mainstream smoke. Nicotine, a substantial component of tobacco combustion, is produced mainly in the particulate phase. However, as the ETS mixture dilutes and ages, the nicotine rapidly shifts to vapor phase. Chamber studies by McCarthy (1987) and others have 18

Draft - Do not cite or quote demonstrated that the half-life decay of nicotine is more than twice that of the particulate phase. A number Of the constituents listed are carcinogens or suspected carcinogens according to the International Agency for Research cn Cancer (~ARC). Measurement of ETS The large number of constituents in ETS make it impossible to assess overall exposure based on measurement of each one. ~nstead most researchers have measured one or more COmpounds and have used those to estimate the total exposure to ETS. Changes in ETS composition over time and exposure conditions limit the accuracy of this method. This chapter will discuss in detail only a few of the possible measures of ETS: panicles, nicotine, cadmium and nitrosamine. Most of the data presented will be from studies involving cigarette smoke since this is a major source of indoor ETS. Littla work has beer done on pipe or cigar smoke. EXposures to Environmental T~baccc Sm0ka According to ~he U.S. Department of Commerce (1985) about 30% of adults in the U.S. are smokers. 40% of homes nationwide have at leas~ one s.mcker. In a survey of over i0,000 children in six U.S. cities, the percentage of children living with one or more smoking adults varied from a low of 60% to a high of 75% (Perris et al., 1979). Lebowitz and Burrows (1976) reported 54% of children in a study in Tucson had at least one smoker in the home. These data indicate that the potential for exposure to ETS in the home is greater than that inferred from national statistics. ~n part, this reflects the demographics of smoking where it is adults in their child-raising years that are more likely to be smokers than the overall average. Surveying a new cohort of elementary- agm children in six U.S. ci~ies reveals that on average, parental smoking has decreased between 10% to 15% over a decade (mid 1970's to mid 1980's). Smoking between different demographic groups can vary widely and this will modify the exposure of nonsmokers to ETS° Overall, ETS exposure will depend on the proximity of an individual to the source of smoke. Patterns of smoking will be influenced by time, location, and type of activity. MICROENVIRONMENTAL MEASUREMENTS OF CONCENTRATIONS Concentrations of Particles and ETS Numerous studies have been conducted using respirable suspended particulates (RSP) as markers for ZTS. Both continuous and integrated measurements methods have been used, Although RSP 2O

Draft - Do not cite or ~ote is not specific for the presence of smokers in the home and other indoor locations, the number of cigarettes smoked have shown to correlate well with RSP. Particulate Concentrations in Homes Spengler et el. (1981) measured 24-hour respirable particulate levels in 55 homes in six U.S. Cities. The mean monthly ~oncentrati~n across cities is presented in Figure 3, wi~h indoor particulate levels similar to the outdoor levels. Table 4 shows the respirable particulate levels in the homes as a function of the number of smokers. The actual amount of smoking in the home was not reported. The researchers concluded that the major source of indoor particulates in smoking homes was cigarette smoke. Each smoker 'n the home raised the mean respirable particulate level by 20 .g/m}. Further analysis of the data by Dockery and Spengler (1881) showed that each cigarette smoked in the home increased the mean respirable particulate levels by 0.88 Mq/m3. In air conditione~ homes, the respirable particulate levels increased by 2.11 ~g/m per cigarette per day. This increase was probably caused by recirculation of indoor air which reduced the cigarette smoke dilution. More recently Spengler and colleagues (1986) analyzed RSP data from over 200 homes in WatertoD~n,, MA. Homes with smokers had RSP concentrations of 30 to 35 ~g/~higher than nonsmoking homes. RSP concentration and the number of cigarettes smoked per week were highly correlated. Models based on this data predict a contribution of 0.77 ~g/m5 per cigarette per day. This would mean a pack of cigarettes would increase the indoor RSP concentration by 15.5 ~g/m3. Particulate Concentration in Offices Using a piezobalance, Weber and Fischer (1980) monitored 44 workrooms at seven different companies in Switzerland. The workrooms had varying levels of smoking. A number of samples were taken in each room over a two-day period. After subtracting the particulate levels found in an unoccupied room, the mean particulate level for the 492 samples taken was i~3 ~g/m3 with a standard deviation ~f 130 ~g/m~. The maximum concentration measured was 962 ~g/m'. Quant et al. (1982) used a piezobalance to monitor three offices. The offices were divided into cubicles with half-wall partitions and contained both smoking and nonsmoking areas. Offices were monitored continuously for one work week. Figure 4 shows the results of continuous monitoring in one of the offices. For the three offices, the ten-hour day averages ranged from 37~g/m~ to 89 ~g/m3. 21

Draft - DO not cite or quote Miesner et el. (1988) used both continuous and integEated methods to monitor in five office buildings in metropolitan Boston. Both filters and nephelometer were used to measure in 12 offices, one conference room, and a designated smoking room of a large nonsmoking office. In offices without smoking, concentrations typically ranged from 15 to i0 ~g/m~. In offices with stoking, concentrations were higher, ranging from 20 to 80 ~g/m~. ~n designated smoking areas, concentrations were I00 to 500 ~g/m-. Shore-term concentrations measured with the portable MIN~Pu~M exceeded I000 ~g/m~ in one of the designated smoking areas. Particulate Concentration in Offices Repace and Lowry (1980) measured particulate levels in various indoor public facilities both in the absence and presence of smoking. For nonsmoking locations such as restaurants, libraries, a church, and a bakery, the mean indoor RSP level was less than 60 ~g/m3, Measurements taken in public facilities in the presence of smoking ~re shown off Table 5. Measurements range from 86~g/m3 to 187 ~g/m for restaurants and cafes that pez-mit smoking. O~her areas where there are likely to be more smokers per area than in restaurants had much higher~oflcentrations of particulars mat:or, ~anging from 2C0 to 700 ~g/m~. Besides monitoring in offices, Miesner et el. (1988) also took continuous and integrated HSP measurements in numerous public facilities incl~ding a library, museum, school, subway, bars, and restaurants. They found that for most public buildings where no smoking was~resent the particulate levels were low usually less than 30 ~g/m~. Levels in transportation facilities such as the subway and bus stations were slightly higher with a mean integrated measurement of 63 ~g/m3. Higher concentrations were found in smoking areas such as bars, rmstaurants and a public smoking room with a mean integrated measurement of 79 ~g/m3 and a standard deviation of 44 ~g/m3. Concentration of Other Components of ETS Numerous researchers have looked at other tracers for ETS. Because of its high specificity for tobacco smoke and its presence in high concentration, nicotine is a promising choice. McCarthy et el. (1987) measured indoor nicotine levels in smoking and nonsmoking homes. The home nicotine values ranged from a average cf 0.1 ~g/m3 in the nonsmoking households to 4.2 ~g/m~ in the smoking households. The presence of low nicotine values in some of the nonsmokinq households can be accounted for by visitors to the home who weEe smokers. A number of studies have used integrated readings to determine nicotine levels in offices and public buildings. A selection of these studies are presented in Table 6. 22

D~aft - E~ not =its or ~ote Cigarettes are also known to be a source of cad~mium. Lebret etal. (1987) considered cadmium as a useful tracer for ETS. They monitored twenty homes and one outdoor site for fine particulates in Watertown, MA. Particles were analyzed for elemental composition using x-ray fluorescence. At the outdoor site and in homes without smokers, cadmium levels were below the detectable limit. In homes with smokers, indoor cadmium levels were highly correlated with indoor fine particulate measurements. Nitrosamines, some of which have been listed as animal carcinogens by the IARC, have been studied in public facilities and homes (Brunnemann et al., 1978). Using continuous measurements they found mean levels of nitrosamines in public facilities which ranged from 0.01 to 0.24 ng/L, Bo~h homes monitored had levels of less than 0.005 ng/L. Wallace et al. (1987) measured the personal exposure and breath levels of benzene and other aromatics in 200 smokers and 322 nonsmokers in New Jersey and california. Benzene is listed as a human carcinogen by the IARC (1986). They found a significant increase in breath concentration with the number of cigarettes smoked. Smokers were found to have up to i0 times the breath concentration of benzene compared to nonsmokers. Nonsmokers who reported smoke exposure at work showed elevated levels for fall and winter but not for spring and summer. The authors concluded that cigarettes were the major source of benzene for about 50 million U.S. smokers. No single constituent of ETS is sufficient to completely characterize an individual's exposure to ETS., Research on ways tc relate these measurements to specific health effects continues tc be done. The most prudent course is to measure several of these components in exposure studies. Markers specific to the class of ETS components, or health outcome of interest, could be utilized in epidemiologic studies to enhance precision of the exposure. Personal Exposures Personal monitoring studies have many of the same problems that area monitoring has, such as trying to measure ETS exposure based on one or more markers. However, personal exposure monitoring has the advantage of including spatial and temporal dimensions to the measurements. It is also possible to use time- activity diaries to link exposure with location and activity. The results of a personal monitoring study by McCarthy etal. (1987) show that the exposurm of children to RSP was much higher than that of children from nonsmoking households. The average personal RSP value increased from 29 ~g/m3 for children from nonsmoking families to 56 ~g/m3 for children from smoking families. The average personal nicotine concentration increased from 0.3 23

Draft - Do no~ Cite or ~¢te ~g/m3 to 2.5~g/m3 for children from nonsmoking and smoking families respectively. A child's personal nicotine is highly correlated with the consumption of cigarettes in the home while ~he personal RSP was not. This implies that although there are multiple sources of RSP, the majority of ETS exposure is from the child's home. Spengler et el. (1985) had i01 nonsmoking volunteers from Kingston/Marriman, Tennessee wear personal resplrahle suspended particulate monitors for 3 days. Nonsmokers were divided in two groups: those who lived with a smoker and those who did not. Outdoor and indoor particulate levels were taken for comparison. Results showed that personal exposure was not correlated with outdoor concentrations but that ETS significantly increased an individual's personal concentration profile. In Spengler and Tosteson (1981), 45 nonsmoking adults were monitored for RSP for 18 days. They were also divided into two groups: those exposed to ETS and those who werm not. Area monitors were also placed inside and outside. Personal exposure was higher than both indoor and outdoor measurements. On average, the individual exposure was increased by 20 ~g/m3 among those who reported exposure to ETS. Cctinine i~ a major metabolite of nicotine. McCarthy etal. (1987) measured cotinine levels in the urine and saliva of 81 nonsmoking children. Nicotine levels in the air were also monitored as was RSP. They found a high correlation between personal nicotine levels and cotinine indicating a quantitative relationship may exist. They did however find high variability. Coultas et al. (1987) measured cotinine in the saliva Of 1360 nonsmoking children and adults. They found an increase with the number of smokers in the home at all ages. However, household variability was wide and even 30% of the nonsmokers living in a nonsmcking home had detectable cotinine levels. Summary I. Environmental tobacco smoke is the primary contaminant causing elevated RSP levels in enclosed spaces. 2. Environmental tobacco smoke can be a substantial contributor to the level of indoor air pollution concentration of benzene, acrolein, N-nitrosamine, pyrene and carbon mcnoxlde. 3. Measured exposures to respirahle suspended particulates are higher for nonsmokers who report exposure to ETS. 24

~aft - Do not cite or ~ote Y~¥E~NCEE BADRE, R.r GUILLERM, R*t ABRAN, N., BOUTDIN, M., DUM~S, C. Pollution atmospberique pat la ~umee de tabac (Atmospheric pollution by smoking). Ann,leE pharmaceuti~es Franc,lEes 3S(9/10):443-452, 1979. SRUNNEMA~'Ht K.R.e AD~LMEs J.Dt ROt D.R.Goe HOFYM~, D° The influence of tobacco smoke on indoor atmospheres: 2. Volatile and tobacco-specific nitrosamines in main- and sidestream smoke and their contribution to indoor pollution. Proceedings of the Fourth Joint Conference on Sensing fo Environmental Pollutants, New Orleans, 1977. American Chemical socle~y, 1979, pp. 876-880. CANOe J.P.w OATALINt ~.t BADRZ, R°r DD~AE t C*o VIALAs A.t GUILLRRME, R., Determination de la nicotine par chromatographie en phase gazeuse: 2. Applications (Determination of nicotine by gas- phase chromatography: 2. Applications). Ann,leE Pha~maceutiuues 29 (II):663-640, 1970/ COULTAE, D.B., HOWARD, CoA., REAKE, G.To, SKIPPER, B.J,, SAMET, ~oM. Salivary cotinine levels and involuntary tobacco smoke exposure in children and adults in New Mexico. American Rev. ResD. Dis. 136:305-309, 1997. CUDDEBAORt J.R., DONOVAN, J.R., BURG, W.R. occupational aspects of passive smoking. American ~ndustrial Hvuiene Association Journal 37(5) :263-26~, 1976. ROOKERY, D.W~r EPENGLER, respirable sulfates and 15(3):325-343, 1981. J.D. Indoor-outdoor relationships of particles. DUSCEAY, T.G., STEVENS, R.R. Ambient air analysis with dichotomous sampler and x-ray fluorescence spectrometer. Environmental sc~e~Se and Technolocv. 9(7):663-668, 1975. SLLIOTT, L.R., ROWE, D.R. Air q~ality during public gatherings. Journal of the Air Pollution Control Association 25(6):635-636, 1975. YIRET, M*W. Environmental tobacco smoke ~easurements: Retrospect and prospect. EUroDean Journal of ResPiratorY Di~@a~es S5(Supp. 133):369-376, 1984. HARLOB, DoP., MAR~URY, M., SAMET, J't SPENGLER, I.D. Relating indoor NO2 levels to infant personal exposures. E/IY-I~, 21(2) :369-37s, 1987. HARMSEN, N., EYPZNBERGER, E. Tabakrauch in verkehrmitteln, wohn un arbeitsraumen (Tobacco smoke in transportation vehicles, living and working rooms). Archiv fu~ bvfiene un~ bakterie~o~ 141(5):383- 25

Draft - DO not =i~a or quo~m 400, 1982. RILLRR, F.C*, MCXUSKER, K.T., M,%ZUMDER, M.K., WILSON, J.D., BONE, R.C. Deposition of sldestream cigarette smoke in the human respiratory tract. A/~erican Review of Resoirat~1-~ Disease 125(4):406-409, 1982. HINDS, W.C., TIRETs M.M. Concentrations of nicotine and tobacco smoke in public places. New Enaland Journal of Medicine 292(16):844-845, 1975. HOFFM~t D.0 HALEYt N.J.+ BRD~M-~Nn K.D.t ADAMS, J.D., WYNDER, E.L. Ciaarette Sidestream Smoke: Formation. Analysis and Model Studies on the uptake by Nonsmokers. Paper presented at the U.S.- Japan meeting on the new etiology of lung cancer, Honolulu, March 21-23, 1983. INGEBNETHSEN, B.J., SEARS, S.E. Particle Size Distribution of the Sidestream Smoke. Paper presented at the 39th Tobacco Chemists' Research Conference, Montreal, october 2-5, 1986. JU, C.t SPENGLER, J.D. Room-to-room variations in concentrations ~f respirable particles in residences. Environmental sciences an~ "~ 15(5) :592-596, 1981. JUST, J-r BOSKOWSKA, M., MASIARKA, 9. Zanieczyszcenie dymen ty~onlwym powietrza kawlarn Warszawskich (Tobacco smoke in the air of Warsaw coffee rooms). Roczniki Pantstwoweuo Zakladu Hvuienv 23(2):129-135, 1972, KEITM, C.H.e DERRICK, J.O. Measurement of the particles size distribution and concentration of cigarette smoke by the confuge. Journal of Colloid Science 15:340-356, 1960. LUS, J., XUHN, H. Verteilung verschiedener tabakrauchbestandteile auf haupt-und nebenstromrauch (eine ubersicht) (Distribution of various tobacco smoke components among mainstream and sidestream smoke (a survey). Bietraae zur Tabakforschun~ Tntern~ohal 11(5):229--265, 1992. LEADERERe B.P., CAIN, Mo$., ISSEROFT, R. Ventilation retirements in buildings: 2. Particulate matter and carbon monoxide from cigarette smoking. Atmospheric ENvironment 18(1):99-106, 1984. LRBRET, E. Air Pollution in Dutch Homes+ Ph.D. Thesis, Wageningen Agricultural University, The Netherlands, 1985. LERRET, Z., McCARTHY, J., SPENGLER, J.D. A survey of time-activity patterns in Kingston/MarrimaN. Methods end Suooort for Modelled Data. Presented at Quality Assurance in Air Pollution Measurements Conferences, Boulder, Colorado, October 14-18, 1984. 26

Draft - DO Bot cite or ~uote MCCARTHY, J. Physical and Biological Markers to Assess Exposure to Environmental Tobacco Smoke. Ph.D. Dissertation, School of Public Health, Harvard University, 1987. MCCarTHY, J., BBBNGLBR, J., CR~NGs B. A personal monitoring study to assess exposure to environmental tobacco smoke° the 4th International conference on Indoor Air Oualitv and Cllm~, Berlin [West), 17-21 Au~us~ 1987. MSESNE~, E°A., RUDNICX, S.N.t PRELL~R, L., BU, F.C., SPENGLER, J.Doa OZKAYNAK~ E., NELSON, W° Report to the U.S. Environmental Protection Agency, Cooperative Agreement NO° CR-813526-01-0, Harvard School of Public Health, 665 Huntington Avenue, Boston, MA (1988). NU~U~L~TBU, N., ~MURA, S,, O3L~D~ T,, TOMITA, X, Estimation of personal exposure ~o tobacco smoke with a newly developed nicotine personal monitor. ~ 35(i):218-227, 1984. NEAL~ A°D°t WADDEN, R.A.B., NOBBNBERG, S.H. Evaluation cf indoor particulate concentration for an urban hospital. Industrial Hvolene Association Journal 39(7) :578-582, 1978. PORSTENDORFER, J°s SCHRAUE~ A. Concentration and mean particle size of the main and side stream of cigarette smoke. ~. 32~33- 35, i~72. QU~CKENBOSS, J.~., KANAREE, MoS., SPENGLER, J.D., LETS, R, Personal monitoring for nitrogen dioxide exposure: Methodological considerations for a community study. Environmental International 8(i-6) ~249-258, 1982. QUAC~ENBOSSt J°J., SPENGLER, J.D., KANARZEo M.E°s LET~, R., DU?PY, C°P. Personal exposure to nitrogen dioxide: relationship to indoor/outdoor air quality and activity patterns° 20:775-783, 1985. QUANT, Y°R°, NELSON, P.A* SE, ~.J. Experimental measurements of aerosol concentration in offices. Environment International 8:223-227, 1982. RZP~CE~ J°L°, LOWREy, A.H. Indoo~ air pollu~ion, tobacco smoke, and public health. Science 208:464-472, 1980. RZPAOE, J°L°s LOWREYs A.M. Tobacco smoke, ventilation, and i~door air quality. American Socletv of Neatinu. Refri=eratina. and Air- Condltionina Enaineers. Inc.. Transactions 88 (part I):895-914, 1982° SAKUMA, N°, EUSAMA, M°s MUNKAKATA, S° OESUME# T., SUGAWAR~, S. The distribution of cigarette smoke components between mainstream and sidestream smoke: i. Acidic components° 27

Tabakforschunc 12(2):63-71~ 1983. SAKUMA, ~., KUSAM~, M., Y~GUCHI, The distribution of olgare~e smoke and sidestream smoke:2. Bases. 12(4):199-209, 1984. Draft - DO not ciue or quote K., M~TSUKI, T., SUGAWAR~, S. components between mainstream Beltraae zur Tabakforschuna SAKU~, H., KUaJ~L%, M.s YA~L%GUCHI, H., SUGAr, T. The distribution of cigarmtte smoke components between mainstream and sidestream smoke:3. Middle and higher boiling componenus. zur Tabakforschun~ 12(5):251-258, 1984. SCR34ELTZ, I° dePAOLIS, A., KOFFM~, D. Phytcsterols in tobacco: Quantitative analysis and fate in tobacco combustion. Beitraae zur Tabakforschuna 8(4):211-218, 1975. 8PEHGLER, J.D., DOC~[ER¥, D.H°t TURNER, H.A°s WOLFOOH, J°M°, FE~RIS, B.C. JR. Long-term measurements of respirable sulfates and particles inside and outside homes, onment 15(i):23-30, 1981. ~PEHGLER, ~.D., REED, H.P., LZBRZT, Z., CHANG, B°~., WARE, J.H., S~EIZER, Y.E°, FERRIB, B.G. JR. Harvardls indoor air ~ollution/health study. Paper presented at the 79th Annual Meeting of the Air Pollution Control Association, Minneapolis, Minnesota, June 22-27, 1986. 8PEHGLEHt J.D.t TREITMAN, H.D°, TOSTESON, T.Do, MAGE, D.T., BOCZEK, M.L. Personal exposures to respirable particulates and implications for air pollution epidemiology. Environmental Science and Technoloav 19(8): 700-707, 1985. SPENGLEH, J.D., TOSTZSON, T.D. Statistical Models for Pe~so~a~ Exposures Data. Paper presented at Environmetrics 81, Conference ~f th8 society for Industrial and Applied Mathematics, Alexandria, Virginia, April 1901. W~LLAdE, L., PELL~ZZARI, Z., HARTWELL, T.D., PERRITT0 Rt ZZ~GENFUS, R° Exposures to benzene and other volatile compounds from active and passive smoking° Archives of Environmental Health 42(5):271- 279, 1907. WEPZR, A., ~I0CHER, T. Passive smcklng at work. Archives af¸ ~ccuDational and Environmental Health 47C3):209-221, 1980. WHITBY, X.T. The physical characteristics of sulfur aerosols. Atmospheric Envlronment 12:135-159, 1978) U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES° The Health ConseQ~/ences of Znvoluntarv Smoklna. A Report Of the Surgeon General. DHHS Publication No. (CDC) 87-8398, 1986. 28

Draft - Do n=~ cite or qu=~e U.8° DEPARTMENT OF TRANSPORTATION AND U.S. DEPARTMENT OF HEALTH, EDUCATION, ~ND rELYk~E. Health Aspects of Snokina in ta ~a ~. U.S. Department of Health, Education, and Welfare, National Institute for Occupational Safety and Health, December 1971. 29

Draf~ - Do no~ =i~e or quote FIGUP.ES AND TABLES, CHA~ER 2 30

• I r l , F , I , ~ , f i I J I i I i I i 2 • e 4 10 I| 14 ~I ~i ~ ~Q HC;URZ 1, -Tl~e ~oca~:~.on Pa~:e~ns ~or 46 ~n£ant:s Source: Ha~Zos e~ aL. (~987~

~IGURE 2. Draf~ - D~ nc~ cite or quote l~l[xl ~L£I ~l--..~ Ig~lTI~--~ ~C~EC~LT K~UI(| ~ JErl~x ~CL£1 ~1¢4 Schematic of an atmospheric aerosol surface area d£~nrlbutlcn showing the thre~ ~odes, main source of ma~ for e&¢h mode, ~he prlncelpal ~roc~s~ involved inser~inE mass into ~ach mo~e. ~nd ~h~ $our=~: '~i=by (1978)

Draf~ - DO ~o~ cite Or ¢~o~e 120. hw.~ *** . , --: ~':'~'~--" / 'll~.ltPlll~B,l[J,i MonthLy Kean Mass ~espl=able ?sr~lculs~e Concenceacions (p8/=~) Across $1= ¢£~les Source: SpenKle: et al. (1981) A gP

---

D=a~t - Do no~ cite or quote TABL~ 1. ~o~ ~ IOJI ~T; 11174 57~ ¢4~I Out~kSI ~ &Ill l~JIl %47 I0~I l~l ~ok~l ~Jll ~ ~00 ~.~ OJk2 1:24 /

Draf% - Do no~ cite or quote TABLE 2, ~eAn Pe==enc Time ~penc ~n Var~ou~ L~ca~ions for Three Po~ulaclon Groups phi, ~mmlt WigLet popuJI(iOh ~O~D l~tiOm wor ke~ ~nwor ker| Ituden~ ~omb~ne4 to~Js home {$D) 59.3 (11.9) ?&2 (12.1) 6&3 (19,3) 654 (I~3) mJtalds (SD) ]2.2 I])]] ' ]2.J 19.9i 1~.0 (J.3] 13.'2 {9.4) moor vehicle (SD) S,I[ (4,2) 4.4 (2,'i) ~.314.3) 4.4 (4~) woPk/sel~w~ ISD) 15..5 (10 9) 0.2 (OJI 4,4 (7.$) 8.4 (10,6) other [nd(:¢~ (SD) 70 (6.4] 712 (6.4) 9.0 (9.1S) 8.I (8.2) ,VL37 32 ].r? 346 home {SD) 66,1 (1 ].4) 83,3 (&4) 8S, l (101) 67~ I11.5) OuLlide (~D) 3.2 ($.~,5) iJ (2~0) 319 (3.2) 3.5 (4.2) moLOr vehicJe ($D) ~,6 (5,6) 4.2 (2.5} ~,3 (2,6) 42 /4,1) '*'otk/~h~l ($~3) 18.~ [10.4) 3,0 ('7.11 ]9.5 ('T..~) 1?.9 19.7) ol~er indoors ($3~) 6.4 (6+0) 715 (5,3) ?.3 (8.3) ?.0 ~6+1) 2~127 28 17~ ~39 ~ource: Quackenboss..e~ al+ (198~)

~ABL~ 3. DisCriSu:ion of Con$ClC~e~= in Ma~sCrea= Smoke (MS) and :he ~cio of $Ldescream Smvkl (SS) co HS of Nonfilclc C£ga:ec~6J

Draf~ - D= ~o~ cite cr quote Resplrable Farti~u~a~e L~vels a~ • Function of N~ber Of $mcker~ ,,..

TA 3 T.,Z 5. Dzaft - Dc no~ tit8 o~ ~uote Par~¢ula~Is Measurad u~de~ ~mil~s~i¢ C~nd£t£onJ SG~ECE: U,S. Department o£ ~eal:h an4 Human Servlccs (IQ86>

TABL~ 6. Draft - Do nQu c£~m or quote ~co~ine ~masurad Under Realistic Conditions ~y ~drw m aL 19 *mekA~ N~ Ci~,q* ~ g~ mmp,b /m SOURCe: ~.5. Depar~men~ of Heal~h and H~an Se~ices (1986)

Drmft - Do not cite o5 quote CKAPTER 4 ABSORPTION OF SMOKE CONSTITUENTS BY NONSMOKERS Dietrich Hoffma~ PhD, Xlaus D. Bl-/nneman~ MSc, Man~y J. Haley PhD American Health Foundation V~lhalla, New York 10995 INTRODUCTION Exposure to environmental tobacco smoke (ETS) occurs at the worksite, in public places, and in private homes. ETS is a composite of effluents generated in various ways during the burning of tobacco products. The major source for ETS is sidestream smoke (SS) which is formed during smouldering cf cigarettes, cigars and pipes between the taking of puffs. Minor contributions to ETS are made by those pollutants of the mainstream smoke (MS) that are exhaled after inhalation of each puff by the active smoker. The smoke escaping into the air from the burning cone and from ~he mouthpiece of a tobacco product during and after puff-drawing is another minor contributor, in addition there is some diffusion of MS gas phase components through the cigarette paper into the environment. More information is needed on the relative sources of smoke in the complex mixture of ETS generated from different cigarettes under varying conditions. In the laboratory, MS and SS are generated under standardized conditions by machine smoking (1,2). While these conditions enable us to compare the yields of individual smoke constituents from various brands of cigarettes, cigars and pipe tobacco, they do not fully reflect the patterns of smoking by humans (3,4). The consumer's intensity of puff-drawing and inhaling of the smoke is profoundly influenced by the nicotine content of the MS (4,5), and smoking intensity is highest when cigarettes with perforated filter tips a~e being smoked [6). The SS release is governed by the velocity of air currents around the burning cone; ~hus, higher air flow generates higher yields of most SS components. Even though a major reduction of mainstream smoke yields Of the sales-weighted average cigarettes has occurred during the last three decades, (U.S. cigarettes declined from 35.5 mg tar in 1954 to 12 mg tar in 1982; (7)), the SS emissions of smoke constituents were not significantly reduced (5,9). The data in Table 1 emphasize this with a comparison of the yields of a select group of toxic compounds in the MS and SO of four types of U.S. cigarettes. These cigarettes were machine- smoked under identical conditions. Since the consumer of the low- yield filter cigarettes is likely to smoke more intensely, a 43

Draft - Do not cite or quote larger portion of the tobaczc column is burned during smoking of this type of cigarette than is burned during smoking of nonfilter cigarettes. Therefore, a somewhat lower yield of SS is expected from the low-yield cigarette smoked by the consumer than is obtained by its standardized machine smoking. The exposure of nonsmokers to the effluents of burning tobacco products usually occurs after considerable dilution of these air pollutants. This is well substantiated by analyses of the air in enclosed spaces polluted by tobacco smoke (i0,ii) . A. Biolocical Markers in Phvsiclocical Fluids The exposure of nonsmokers to ETS can be assessed with the help of q~estionnaires, by estimating the dose from the chemical analysis of smoke-polluted air, by personal monitoring of ETS components and/or by measuring the uptake of individual smoke components in physiological fluids of individuals during or after exposure. The last and most promising method will be discussed in this chapter. The degree of exposure to ETS depends on several factors, including length of time spent in a smoke-polluted area, the number of smokers within this area, the size and nature of the space, the degree of ventilation and the respiratDry rate of the exposed individual. Thus, optimal assessment of ETS exposure is achieved by analysis of physiological fluids of exposed individuals as well as by analysis of the respiratory environment. New biochemical methods enable us to quantify exposure to ETS by determining the uptake of certain smoke contstituents (or their metabolites) in biological fluids. An primary requirement for such biochemical measurements is the availability of highly sensitive and specific methods. I. Nicotine and Cotinine. Disregarding accidental or occupational exposure to tobacco (12,13), or the use of nicotine-containing chewing gum or nicotine aerosol rods as aids for smoking cessation (14), the presence of nicotine and of its major metabolites in physiological fluids is entirely due to the exposure to tobacco~ tobacco smoke, or ETS. LOW - levels of nicotine have been found in other members of the solanaceous variety of plants (14A) but could not be expected to make an impact on the body burden of nicotine which is obtained from tobacco sources. Nicotine and its major metabolite, cotinine, in saliva, blood or urine of active smokers and of passively exposed nonsmokers are primarily determined by gas chromatography (GC) with a nitrogen-sensitive detector, and by radioimmunoassay (RIA) (15-17). An HPLC method which has been developed for quantitation of cotinine in plasma or saliva of smokers (18) has not been applied to urine analysis even though the analysis of this biological fluid appears to have the greatest pouential for 44

Draft - Do not cite or quote evaluation of nicotine uptake by nonsmokers. A problem with this HPLC method seems to be an unusually high background of ootinine in persons reporting no exposure to ETS. The possibi!e co- migration of caffeine with cotinine in this system needs to be excluded. (18A) A recently published, highly sensitive method for determining nicotine in plasma by HPLC with dual electrochemical detection (2 ng/ml) has not as yet been applied to physiological samples of involuntary smokers (19). Another emerging analytical method for the determination of nicotine Qr cotinine is the enzyme- linked immunosorbent assay (ELISA; 20). Trans-3'-hydroxycotinine has been found to be the most abundant nicotine metabclite in the urine of active smokers (21), however, it is difficult tc quantitate. Since the compound is not readily soluble it has to be transformed into a heptafluoro derivative prior to GC detection (22). The levels of 3'- hydroxycotinine in plasma have teen found to be much lower than those of cotinine in the same smokers although the renal excretion of 3t-hydroxycotinine has been reported to be greater (23). Despite its abundance in urine of smokers, this compound has not yet been applied to the analysis of ETS uptake by nonsmokers. The GC and RIA mezhods are most widely used for assaying nicotine and cotinine in active as well as in passive smokers, primarily because of their specificity and sensitivity, and because the needed instrumentation is available in most modern laboratories. Chromatographic methods, especially those using GC with nitrogen-phosphorus detectors (detection limit 0.i ng/ ml fluid; 16), or a mass-spectral detection system, offer greatest specificity and high sensitivity; however, they require expensive instrumentation and technical expertise and they are rather labor intensive. Since the air as well as glassware in laboratories may contain traces of nicotine, the chromatographic methods require the utmost precautions to avoid contamination of samples. The RIA techniques are operationally simpler, less expensive and require smaller samples (detection limit 0.95 ng/sample; 17). More than 50 nicotine metabolites and structurally-related • olecules have been tested as inhibitors of nicotine and cotinine antigen-antibody reections; few of them interfere with the RIA (24). Nevertheless, the potential for cross-reactivity with unknown endogenous components exists. ' The fact that, upon analysis, thousands of samples obtained from nonsmokers in the US and UK have been found to be negative, indicates that diets and drugs commonly used in these two countries do not pose problems of interference. There is good Correlation between results obtained by GC and RIA analysis for plasma cotinine concentrations (r~0.99; 25). A potential problem in RIA analysis can come from extrapolation to values below the linear range of the standard curve. Care must always be taken to insure proportionality of response. 45

Draft - Do not cite or quote An interlaboratory comparison of data from ii laboratories in 6 countries has demonstrated that GC and RIA techniques can reliably quantitata nicotine and cotinine in urine and plasma samples. A good correlation of laboratory methods was observed in plasma samples and in urine samples to which cotinine had been added as a tracer. However, in urine samples without tracer, several RIA values for cotlnine were found to be slightly higher than those observed by GC. This could be due to a cross reaction of the antibody ~ith another ~ompound present in urine, or the discrepancy could arise from a loss of urinary cotinine during GC extraction. The former explanation is more likely to apply here although conventional GC extraction techniques have been reported to result in the loss of conjugated metabolites of nicotine. The quantitation of these conjugated compounds by GC methods has recently been reported by Curvallet el. (25a). In addition cross reactivity of various cotinine antibodies with drans- 3.hydroxycotinlne has been reported to range from 2% (J.J. Langone, pets. co~/~.) to 30%~ (25b)) All ir~munoassay methods have led, however, do perfect distinction between nonsmokers and active smokers (26). Table 2 presents data from model studies on the uptake of ETS by nonsmokers ~nder acute exposure conditions (27-30). The main purpose of these assays was to develop the methodology for field studies and to compare the uptake of nicotine from environments with various degrees of pollution and different types of pollutants under controlled conditions. It has been shown that the equilibrium of nicotine between vapor phase and particulate phase of ETS depends greatly on the concentration and pH of the emitted smokestream (31) and, thus, influences the uptake of nicotine by inhalation. After repeated exposure to ETS under controlled conditions, such ss ~wice daily 80-minute exposure on 3 consecudive days to the diluted pollutants of 4 CONCUrreNtly smoked cigarettes (221, the me~urements in 4 nonsmokers have shown that except for nicotine in the saliva, the physiological fluids do not reflect maximal concentrations of nicotine and cotinine until at least 24 hours later. This observation has led to comparisons of the elimination of cotinine in smokers and nonsmokers exposed to ETS (32). The elimination half-life (tl/~) of ootinine from the urine of smokers took 21.S hours and 22.7 hours for nonsmokers. In a second assay, five cigarette smokers were asked to abstain from tobacco use for 5 days and were then given nicotine gum for three days. After another 8 days of abstinence from nicotine, the volunteers were exposed to sidestream smoke (SS). At this point, the cotinine elimination (tl/2) from urine (ng/ml) by smokers took 15.4 hours, by nicotine gum users 18.2 hours, by 8-day exsmokers 27.5 hours, and by the never-smokers 25.6 hours (32). These findings suggest that the residence ~imes of nicotine, cotinine and other tobacco alkaloids, are likely related to the route of nicotine uptake as well as to possible differences in metabolism between smokers and 46

Draft - Do not cite or quo~e nonsmokers. The longer elimination time for cotinine in nonsmokers has been oonfil-~ed by other study groups (35-37), however, the observation has also been challenged (38,39). A longer residence time of nicotine metabolites in nonsmokers could conceivably increase the possibility of endogenous formation of carcinogenic N-nitrosamines (40). Most importantly, differences in the elimination times of cotinine from urine preclude a direct extrapolation to "cigarette equivalents of smoke uptake" by comparing the levels of cotinine excreted by active and passive smokers. This has been discussed by some investigators (i0). Table 3 includes comparisons of nicotine and cotinine in physiological fluids of nonsmokers with or without ETS exposure, and of active cigarette smokers in England (41). Data on the uptake of nicotine by involuntary smokers from additional studies are summarized in Table 4 (29,42-54). Most of these studies demonstrate that nicotine and cotinine levels in physiological fluids of involuntary smokers generally amount to 1 percent and neach maximally a few percent of the amounts determined in active cigarette smokers. Data by Matsukura et el. from Japan on the other hand, show exceptionally high levels of cotinine in the urine of passive smokers. This may be due to several factors including differences in the design of studies and measurement methods (26), Aside from differences in methodology one cannot rule out that differences in the uptake and metabolism of nicotine which have been observed in various population groups, and diet may be partially responsible for the exceptional data reported in the Japanese study (47). A recent finding indicates that the urinary excretion rates of Japanese smokers were significantly different from those determined in adult cigarette smokers in Europe and North America (55). Additionally, a large epidemiological szudy in the U.S. has demonstrated significant differences in serum cotinine levels between Black and ~ite smokers after adjustment for cigarettes smoked per day and daily nicotine availability (55a).These differences in nicotine metabolism require further thorough investigation. Survey data on exposure at home, in the workplace and on social occasions were collected from 319 employed subjects and were correlated with levels of cotinine in a random urine sample. Mean urine/cotinine/creatinine levels were higher for women than for men possibly due to differences in creatinine excretion between the sexes. It is also noteworthy that 94% of the women were employed indoors. Higher levels of urinary cotinine were noted in bo~h men and women who lived with a smoker than in those subjects who did not report living with a smoker (13.3±2.4 vs 5.1~0.4 in men and 13.9±1.9 vs. 5.6~0.6 in women). Differences in the prevalence of exposure at home existed between sexes (males 13.5% vs. females 29.2%). Levels of cotinine found across different exposures indicate that home exposure has a more pronounced effect on urine 47

Draft ° Do not cite or quote cetinine than does workplace exposure (Table 5; 55b). The nicotine uptake by infants due tD ETS exposure, caused by smoking mothers or caretakers, appears to be higher than that observed in adult passive smokers. The amount of cotinine excreted in the urine of the infants was correlated with the number of cigarettes smoked by the mother, oE caretaker or other persons, . during the 24 hours preceding the measurement (33). The primary determinant of urinary cotinine levels has been found to be the smoking behavior of the mother. The finding of relatively high uptake of ETS, as determined by nicotine/cobinine concentrations in the urine of infants, is in line with the observation that infants of smokers have higher rates cf respiratory infections than infants in nonsmokers' homes (56)° Analytical data on nicotine and cotinine in physiological fluids of nonsmokers can be misleading in a fmw cases. These pertain to the very light smokers and those nonsmokers who either chew tobacco or use oral snuff. It is possible, though rare, that the very light smoker shows nicctine/cotinine levels approaching those of passive smokers with extremely high ETS exposure. When ~sed in co~ination with cotlnine measurements, COMb analyses can help to differentiate between the two groups. In regular consumers of snuff or chewing tobacco, cotinine levels are comparable to those found in cigarette smokers while thiocyanate levels and COHb values remain low (57). The determination of nicotine and cotinine in hair has been tried in an attempt to differentiate between active and passive smokers [58). This determination revealed higher nicotine ~oncmntrations in the hair of smokers than in the hair of ETS- exposed nonsmokers and documented the absence of cotinine~ the major metabolite of nicotine, within the hairshaft of nonsmokers. Hair sampling for determining ETS-exposure of nonsmokers deserves more thorough investigation. In summary, in the hands of experienced biochemists, the determination of nicotine and, especially, of cotinine in saliva, serum and~or urine in involuntary smokers represents a reliable, specific method for assaying the level of uptake of ETS by nonsmokers. The choice of biological fluid for the quantita~ion ef cotinine depends upon the question asked. For the evaluation of changes in smoking behavior, sez~/m or urine are preferred while saliva is sufficient to determine whether or not a subject is a smoker (59). For studies of ETS exposure, it is often impractical to collect serum by venipuncture, and since nicotine concentration in saliva can be extremely high i~mediately following ETS exposure, several hours must pass before the concentration of cotinine in saliva is stabilized (30). Also, when large numbers of subjects are to be evaluated, it is preferable to avoid invasive procedures which might discourage participation and possibly bias the results. 45

Draft - Do not cite or ~ote Measurements of cotinine in urine and saliva have been successfully used to quanzitate ETS exposure in large populations. Cotinine excretion in urine is independent of pM, while nicotine excretion is greatly influenced by it. At values above pM 6.0, resorption of nicotine from the urine occurs especially during longer residence time in the bladder. Cotinine is not subject to resorption and, as far as it has been investigated, 3'- hydroxycotinine, a second major nicotine metabolite, is also not affected (60). Quantitation of cotinine in random urine samples can have methodological problems relative to the vol%1~e of urine excreted in any given time period as well as dilution effects. The ideal standard for evaluation of cobinine excretion in brine would be the analysis of a 24-hour urine sample. Since this is impractical in epidemiological studies, random urine samples are usually collected at the time a questionnaire is administered. En this case~ the ratio of cotinine to creatinine in a given sample is often used t~ allow for differences in urine dilution. Urinary creatinine excretion is usually constant from day to day for a given individual, but it does vary among individuals. As a reflection of muscle mass it is generally excreted at about i g per day (men, 1.1 to 3.2 g/day; women, 0.9 to 2.5 g/day). En older persons, the excretion of creatinine ~ay decrease to 0.5 g/day. Low levels of creatinine may also be found in dehydrated infants; this necessitates caution in the expression of ng cotinine/mg creatinine in a random sample (35). However, a recent study with pro-school children has shown that cotinine/crea£inine ratios in passively exposed children 'trackr over several weeks and reflect questionnaire data on exposure (61). Epidemloloqical studies in adults have also shown good correlations between self-reported indices of exposure and cotinine/creatinine ratios when data for men and women are analyzed separately.(55b) 2. Carbon Monoxide. Carbon monoxide (CO) is formed during the combustion of organic matter including the burning of a tobacco product. It is also produced in vivo during metabolic processes. Endogenous CO results primarily from the breakdown of home- containing proteins such as hemsglobin. In nonsmokers who are not exposed to industrial pyrolysis products or vehicle emissions, the baseline levels of CO, present in the bloodstream as .carboxyhemoglobin (COMb), are generally below 1.5% of the total hemoglobin. Persons exposed to heavy vehicle emissions can have COMb levels up ~0 about 2.5%. In cigarette smokers, COHb levels were found to average 5.7% in a study of 450 smokers (62) with litsle difference being noted between smokers of high- or low-yield products. This value is similar to that of 4.7% found in middle aged men in a study by Wa!d et el. (63). Carboxyhemoglobin levels are not good indicators of ETS 49

Draft - Do not cite or quote uptake, due to the fact that CO exposure is not limited to tobacco s~oke; in addition, the measurement of COHb is relatively insensitive. A study in England did not find significant differences in COHb levels in subjects reporting no exposure, some exposure, or a let of exposure (64}. This was confirmed by others (65) and also by a controlled chamber assay (61). One study in which significant elevations of COHb were found used controlled exposure to t~bacco smoke at a level of 25 ppm CO for 8 hours. This intense exposure resulted in an average increase of COHb levels by 2.5% (85). However, such results are not applicable to free-living situations in field studies (67). 3. ~. Hydrogen cyanide, absorbed from tobacco smoke is detoxified in the liver to thiocyanate (SON-). Measurement of SCN- has been used to differentiate smokers from nonsmokers or, as mentioned earlier, in combination with nicotine-cotinine assays to distinguish smokers from chewers of tobacco. Thiocyanate can also be derived from the diet, cruciferous vegetables being an excellent source [68). The specificity of SON as a marker of tobacco smoke ihhalation is poor and it is generally difficult to distinguish i~ght smokers from nonsmokers. This lack of specificity makes SCN- unsuitable for the evaluation of ETS uptake by nonsmoking subjects. 4. MvdroxvDroline. ~apanese investigators have studied the excretion of hydroxyproline in persons exposed to ETS as well as in active smokers and in persons exposed to high levels cf air pollutants (69). The rationale for these studies is that the inhalation of nitrogen dioxide causes degradation of lung collagen and elastin which results in urinary excretion of hydroxyproline. The investigations of the Japanese group suggested an elevated excretion of hydroxyproline by children of smoking parents as well as by wives of smoking husbands, active smokers, and individuals exposed to vehicle emissions. Since NOX levels in ETS are relatively low by comparison to mainstream smoke or vehicle emissions [56,70,71), such increased elimination of hydrcxyproline in passively exposed persons seemed surprising. In fact, another group of investigators has been unable to confirm this finding C72). 5. ~-N~trose-Amino Acids. The occurrence of endogenous nitrosation reactions in cigarette smokers has been demonstrated in several studies. This phenomenon entails the risk of endogenous formation of carclncgenic N-nltrosamines. Endogenous formation of H-nitrosamines has been proven by urinary excretion of the noncarcinogenic N-nitrosoproline (NPRO), N-nitrosothioproline (NTPRO), and N-nltrosomethylthloproline (NMTpRC). Whereas the average excretion of NPRO in nonsmokers amounted to 2.0±i.5 ug/24 hrs, cigarette smokers excreted an average of 7.0±4.0 ug/24 hrs [73-77). In the case of NTPRO, the average urinary excretion by nonsmokers (ug/24 hrs) was 5.9~ that by cigarette smokers 8.7 and that of NMTPRO was 5.6 and 8.$, respectively (75). Only two studies have explored the possibility that endogenous formation of 5O

Draft - DO not cite or quote uptake, due to the fact that CO exposure is not limited to tobacco smoke; in addition, the measurement of COHb is relatively insensitive. A study in England did not find significant differences in COHb levels in subjects reporting nc exposure, some exposure, or a lot Of exposure (64). This was confirmed by others (65) and also by a controlled chamber assay (61). One study in which significant elevations of COHb were found used controlled exposure to tobacco smoke at a level Of 25 ppm CO for 6 hours. This intense exposure resulted in an average increase of COHb levels by 2.5% (85). However, such results are not applicable to free-living situations in field studies (67). 3. ~. Hydrogen cyanide, absorbed from tobacco smoke is detoxified in the liver to thiocyanate (SCN-). Measurement of SCN- has been used to differentiate smokers from nonsmokers or, as mentioned earlier, in combination with nicotine-cotinine assays to distinguish smokers from chewers of tobaccQ. Thiocyanate can also be derived from the diet, crucifercus vegetables being an excellent source (66). The specificity of SON as a marker of tobacco smoke ihhalation is poor and it is generally difficult to distinguish l~ght smokers from nonsmokers. This lack of specificity makes SCN- unsuitable for the evaluation Of ETS uptake by nonsmoking subjects. 4. Hvdroxv~roline. Japanese investigators have studied the excretion of hydroxyproline in persons exposed to ETS as well as in active smokers and in persons exposed to high levels of air pollutants (69). The rationale for these studies is that the inhalation of nitrogen dioxide causes degradation of lung collagen and elastin which results in urinary excretion Of hydroxyproline. The investigations of the Japanese group suggested an elevated excretion of hydroxyproline by children of smoking parents as well as by wives of smoking husbands, active smokers, and individuals exposed to vehicle emissions. Since NOX levels in ETS are relatively low by comparison to mainstream smoke or vehicle emissions (56,70,71), such increased elimination Of hydroxyproline in passively exposed persons seemed surprising. ~n fact, another group of investigators has been unable to confirm this finding (72). 5. N-NitrDsc-Amino Acids. The occurrence of endogenous nitrosation reactions in cigarette smokers has been demonstrated in several studies. This phenomenon entails the risk of endogenous formation of carcinogenic N-nitrosamines. Endogenous formation of N-nitrosamines has been proven by urinary excretion of the noncarcinogenic N-nitrosoproline (NPRO), N-nitrosothioproline (NTPRO), and N-nitrcsomethylthioproline (NMTPRO). Whereas the average excretion of NPRO in nonsmokers amounted to 2.0~1.5 ug/24 hrs, cigarette smokers excreted an average of 7.0±4.0 ug/24 hrs (73-77). In the case of NTPRO, the average urinary excretion by nonsmokers (ug/24 hrs) was 5.9, that by cigarette smokers 8.7 and that of NMTPRO was 5.6 and 8.5, respectively (75). Only two studies have explored the possibility that endogenous formation of 5O

Draft - Do not cite or q~ote require confirmation hut they appear to indicate that the thioether analysis of the urine will most likely not be suitable for the determination of the ETS uptake by involuntary smokers due to varying background levels across subjects. B. Genetoxicitv of Phvsioloaical Piui~ Several studies have explored the possibility that physiological fluids of cigarette smokers contain significantly higher amounts of genotoxic agents than those of nonsmokers (81). The most extensive data base in this field has shown significantly higher mutagenicity in the Salmonella thvDhimurium assay of urine of cigarette smokers compared to those of nonsmokers. Since the original study by Yamasaki and Ames in 1977 (83) at least 20 investigations have shown that the urine of cigarette smokers is significantly more mutagenic than the urine of nonsmokers who are not exposed to genotoxlo agents in occupational environments. But it has also been shown that the mutagenicity of the urine cf smokers can be effected by diet (84). It has further been surmised that exposure of nonsmokers to ETS may lead tc increased urinary excretion of mutaqens. Of the 6 published studies in which the urine of passive smokers was tested for mutagenicity with the Ames test, 3 showed increased activity and 3 showed no increase or, at the most an insignifican~ increase in mutagenic activity (81,85- 87). C. Adduct Fo.~mation of Carcinouens in Passive Smokers. . Measurements in physiological fluids of nicotine and its major metabolite, cotinine, have been shown to be objective indicators of the uptake of ETS. However, these assays will not reflect an individual's response to specific ETS carcinogens. That information is best obtained by assessing levels of macromoleoular adducts with carcinogens or their metabolites. Development of such assays is based on examining the mechanisms of metabolic activation an~'detoxification of tobacco smoke carcinogens. I. Benzo(a)Dvrene. In the case of active smokers, adducts of at least 4 types of tobacco carcinogens or procarcinogens have been studied. These adduots are formed by reaction of specific metabolites of tobacco smoke constituents with DNA and/or hemoglobin. Benzo(a)pyrene (BaP), a carcinogenic representative of the polynuclear aromatic hydrocarbons in tobacco smoke is known to be metabolized to bay region diol epoxides (e.g. 7,8- dihydroxy-9,10-epoxy-7,S,9,10-tetrahydroBaP). Such diol epoxides can bind to DNA in human tissues and lymphocytes. Antibodies developed against the major BPDE-DNA adduct have been used to assess its presence in surgical specimens of lung tissue, in human placenta, and in peripheral blood lymphocytes (89-91). Evidence for the presence of such adducts in samples from smokers has been ascertained but significant differences between smokers and nonsmokers have not been observed. 52

Draft - DO not cite or quote require confirmazion but they appear tc indicate that the thicether analysis of the urine will most likely not be suitable for the determination of the ETS uptake by involuntary smokers due to varying background levels across suDjects. E. Genotoxicitv of Physiological Fluids Several studies have explored the possibility that physiological fluids of cigarette smokers contain significantly higher amounts of genctoxic agents than those of nonsmokers C81). The most extensive data base in this field has shown significantly higher mutagenicity in the Salmonella thvDhimurium assay of urine of cigarette smokers compared to those of nonsmokers. Since the original study by Yamasaki and Ames in 1977 (83) at least 20 investigations have shown that the urine of cigarette smokers is significantly more mutagenic than the urine of nonsmokers who ~re not exposed to genotoxic agents in occupational environments. But it has also been shown that the mutagenicity of the urine of smokers can be effected by diet (84). It has further been surmised that exposure of nonsmokers to ETS may lead to increased uriDary excretion of mutagens. Of the 6 published studies in which the urine of passive smokers was tested for mutagenicity with the Ames test, 3 showed ~ncreased activity and 3 showed no increase or, at the most an insignificant increase in mutagenic activity (81,85- 87). C. Adduct Fo.~mation of Carcinoaens in Passive Smokers. . Measurements in physiological fluids of nicotine and its • ajor ~etabolite, co~inine, have been shown to be objective indicators of the uptake of ETS. However, these assays will not reflect an individual,s response to specific ETS carcinogens. That information is best obtained by assessing levels of macromolecular adducts with carcinogens or their metabolites. Development of such assays is based on examining the mechanisms of metabolic activation an~'detoxification of tobacco smoke carcinogens. i. Benzo(alDvrene. In the case of active smokers, adducts of at least 4 types of tobacco carcinogens or procarcinoge~s have been studied. These adducts arm formed by reaction of specific metabolites of tobacco smoke constituents with DNA and/or hemoglobin. Benzo(a)pyrene (BaP), a carcinogenic representative of the polynuclear a~omatic hydrocarbons in tobacco smoke is known to be metabolized to bay region diol epoxides (e.g. 7,8- dihydroxy-9,10-epoxy-7,8,9,!O-tetrahydroBaP). Such diol epoxides can bind to DNA in human tissues and lymphccytes. Antibodies developed against the major BPDE-DNA adduct have been used to assess its presence in surgical specimens of lung tissue, in human placenta, and in peripheral blood lymphocytes (88-91) . Evidence for the presence of such adducts in samples from smokers has been ascertained but significant differences between smokers and nonsmokers have not been observed. 52

Draft - Do not cita or uote 2. Aromatic Amines. 4-Aminobiphenyl and 2-naphthylamine are the know~ tobacco smoke constituents which are most likely to contribute to the increased risk of bladder cancer of cigarette smokers. The mechanisms by which these compounds are metabolically activated and produce DNA adduots in the bladder epithelium have been extensively studied (92). These studies have shown that the corresponding hydroxylamlnes are key intermediates in DNA and protein modification. The hydrox~lamines also react with hemoglobin, in the case Of 4-amlnobiphenyl, a sulfinic acid amide of the beta-cysteine (93-95). This adduct readily releases 4- amincbiphenyl upon treatment with dilute acid. A method was developed to analyze the released 4-aminobiphenyl by gas chromatography with detection by negative ion chemical ionization mass spectrometry (95). Application of this method to smokers sh~wed that adduct levels were higher than in nonsmokers, and decreased upon smoking cessation. The method may be further refined for assessing the uptake of carcinogenic aromatic amines from ETS by nonsmokers. 3. Ethylene. This volatile unsaturated hydrocarbon is present in both mainstream smoke (200-400 ug/cigarette) and sidestream smoke of cigarettes (96). cigarette smoke contains also traces of the carcinogenic ethylene oxide (7.0 ug/cigarette; 97,98). Upon absorption, ethylene is ~etabolized to the reactive ethylene oxide. The latter binds to cellular macromolecules and to hemoglobin. The alkylated reline is cleaved off of the isolated hemoglobin and the derivatized hydroxyethylvaline is analyzed by GC-MS. cigarette smokers showed significantly higher hydroxyethylvaline levels (389~108 pg/g hemoglobin) than nonsmokers (58Z25 pg/g; 99). So far the method has not been applied to estimates of exposure of involuntary smokers to the procarclnogen ethylene. 4. Tobacco-Specific N-Nitrosamlnes. During tobacco processing and during smoking tobacco alkaloids give rise to tobacco-specific N-nitrosamines (TSNA). The nicotine-derived N- nitrosamines N'-nitrosonornicotine (NNN) and 4-(methylnitrosamino)- 1-13-pyridyl)-l-butanone (NNK) are powerful carcinogens. They occur in relatively high concentrations in cigarette mainstream smoke (NNN, 0.12-3.7 ug/cigarette; NNK, 0.08-0,77 ug/clgarette)and sidestream smoke (NNN, 0.15-1.7 ug/clgarette~ NNK, 0.2-1.4 ug/cigarette; 40). These agents are metabolically activated by alpha-hydroxylation, leading to a highly reactive intermediate which forms DNA adducts and protein adducts (Fig, I). Metabolic activation of NNN and NNK also leads to the formation of hemoglobin adducts. Acid or base hydrolysis of these releases a keto alcohol (compound 5; Fig. I; i00). A highly sensitive GC-MS method has been developed to facilitate the detection of a derivative of compound 5. Refinement towards further increased sensitivity of the method should lead to a dosimetry assay allowing determination of the ~ptake of the carcinogenic TSNA by passive smokers. 50

Draft - Do not cite or quote FUTURE NEEDS The absorption of tobacco-specific smoke constituents from ETS has been demonstrated through analyses of nicotine and its major metabolite, cotinine in ~hl body fluids of exposed nonsmokers. Less tobacoc-speciflc markers have also been measured in exposed populations; however/ the results were a~big%/ous in regard to the ~/antitative uptake of ETS. There is a need to provide information about the uptake and disposition of carcinogenic constituents by individuals exposed to ETS in acute and chronic situations. Analyses to be fully developed and applied to passive smokers will include measurements of adducts of genotoxic smoke constituents covalently bound to DNA or hemoglobin. These techniques have been developed for benzo(a)pyrene, 4- aminobiphenyl, euhylene, and tobacco-specific N- nitr~samines. It is not known whether or not all of these methods can be made sufficiently sensitive to monitor the uptake of tobacco-specific components from ~TS. Nicotine in ETS is prRdominantly present in the ¢apor phase of the smoke rather than bound to the aerosol particles. In order to measure the uptake of carcinogens and toxins residing in the particulate phase of ETS, deposition studies must be developed with specific markers. Particulate phase constituents which could be q~antitated include tobacco-specific N- nitrosamines, polyphencls, such as the immunoactive compound turin, or the tobacco-specific solanesol. (101) However, the levels of these compounds are expected to be low so that development of suitable methodology calls for highly sensitive detection methods. SUMMARY I. The absorbtion of tobacco-specific smoke constituents from ZTS has been demonstrated through analyses of nicotine and its major metabolite, cotinine in the body fluids of exposed nonsmokers. 2. The determination of nicotine or cotinine, in the saliva, serum, or urine of involuntary smokers represents a reliable, specific method for assaying the level of uptake of ETS by nonsmokers. 3. Although cotinine levels in physiological fluids of involuntary smokers generally are of the order of few percent of those of active smokers, differences in the elimination times of these compounds in active and involuntary smokers preclude a direct extrapolation to "cigarette equivalents of smoke uptake." 4. There is a further need to quantitate uptake and fate of carcinogenic constitutents of ETS-exposed nonsmokers, particularly the measurements of adducts of genotoxic smoke components attached to DNA or hemoglobin. 54

Draft - Do not cite or quote AC~CWLEDGEMENTS We thank Ilse Hoffmann and Bertha Stadler for editorial assistance. Our studies are supported by Grants Nc. CA-29580, CA-44377 and CA- 32617 from the National Cancer ~nstit~te. 55

O Draft - Do not =its or quote REFERENCES I. Pillsbury, H.C., Bright, C.C., O'Connor, K.J., and Irish, F.W. Tar and nicotine in cigarette smoke. J. Assoc. Offic. Anal. Chem. 52:458-462, 1969. 2. Dube, M.F. and Green, for analytical purposes. 1982. C.R. Methods of col!ecticn of smoke Recent Advan. Tobacco Sci. ~: 42-102, 3. Herning, H.I., Jones, R.T., Bachman, J., and Mines, A.R. Puff volume increases when low-nicotine cigarettes are smoked. Brit. Med. J. 283: 187-189, 1981. 4. Haley, N.J., Sepkovic, D.W., Hoffmann, D., and Wynder, 5. L. Cigarette smoking as Risk for Cardiovascular Disease. Part ~[. Compensation with nicotine availability as a single variable. Clin. Pharmacol. Thor. 38: 164-170, 1985. T.W. Nicotine and Med. Hypotheses 17: Kozlowski, L.T., Freoker, R.C., Khouro, p., and 5. Chamberlain, A.T. and Higgenbottam, cigarette smoking: An alternative hypothesis. 285-297, 1985. 6. Pope,M.A. The misuse of "less hazardous" cigarettes and its detection: Bole-blocking of ventilated filters. Am. J. Publ. Health 70: 1202-1203, 1980. 7. Federal Trade Commission. Report of tar and nicotine con- tent of the smoke of 208 varieties of domestic cigarettes, 1954-1983. U.S. Govt. Printing Office, Washington, DO, 1983. 8. Great Britain Laboratory of the Government Chemist. Report of the Government Chemist, 1981. Her Majesty's Stationery Office, London, p. 109, 1982. " 9. Toxic and carcinogenic agents in undiluted mainstrea# smoke and sidestream smoke of different types of cigarettes. Carcinogenesis 8: 729-731, 1987. I0. National Research Council. "Environmental Tobacco Smoke. Measuring Exposures and Assessing Health Effects." National Academy Press, Washington, DC, 1986. 337 pp. 11. U.5. Surgeon General. "The Health Consequences of Involuntary Smoking." U.S. Dept. Health and Human Services. DHHS (CDC) 87-8398, 1987, 359 pp. 56

Draft - DO nat cite cr quote 12. Saxena, K. and Scheman, A. A suicide plan by nicotine poisoning: A review of nicotine toxicity. Vet. Hum. Toxicsl. 27: 485-497, 1985. 18. Gehlbach, S.H., Williams, W.A., Perry, L.D., Freeman, J.H., Langone, J.J., Peta, L.V., and Van Vunakis, H, Nicotine absoz~ticn by workers hal-vesting green tobacco. Lancet I: 478- 480, 1975. 14. Pomerleau, O.P. and Pomerleau, C.S. "Nicotine Replacement - A Critical Evaluation". Progr. Cli~. Biol. ~es. 261: 1-317, 1988. 14a. Castro, A, Monyi, N. Dietary nicotine and its significance in studies On tobacco smoking. Biochem. Arch. 2: 91-97, 1986, 15. Feyerabend, C. Determination of nicotine in physiological fluids by gas chromatography. IARC Sci. Publ. 81: 299307, 1987. 16. Feyerabend, C. and Bryant, A.E. Determination in physiological fluids by gas chromatography. IARC Sci. Publ. 81: 209-316, 1967. 17. Van Vunakis, N., Gjika, H.B., and Langcne, J.J. Radioimmun~assay for nicotine and cotinine. IARC Sci. Publ. 81: 817-330, 1987. 18. Machacek, D.A~ and Jiang, N. Quantification of cctinine in plasma and saliva by liquid chromatography. Clin. Chem. 32: 979- 982, 1986. 18a. Thuon, N.T.L., MegUeres, M.L., Rocke, D. Elimination of caffeine interference in HPLC determination of urinary nicotine and cobinine. Clin. Chem. 99: 1456-1459, 1989. 19. Chien, C-Y., Diana, J.N., and Crooks, P.A. Determination cf nicotine in plasma by high performance liquid chromatography with electrochemical detection. LC-GC 6: 58-95, 1988. 20. Bjercke, R.J., Cock, G., Rychlik, N., Gjika, H.B., Van Vunakis, H., and Langcne, J.J. Stereospecific monoclonal antibodies to nicobine and cotinine and their use in enzyme- linked immunosorbent assays. J. Immunol. Methods 90: 202-213, 1986. 21. Ne~rath, G.E., Duenger, M., Crib, D., and Pein, F.G. t_t_t_t_ta/-~D~-3'-hydroxycctinine as a main metabolite in urine cf smokers. Internatl. Arch. Occup. Environ. Health 59:199-201, 1987. 22. Neurath, G.B., Pein, F.G. Gas chromatographic determination Of ~rans-3'-hydroxycotinine, a major metabolite of nicotine in smokers. J. Chrcmatcg. Biomed. Appl. 415: 400-406, 1987. 57

Draft - Dc not cite or quote 23. Adlkcfer, F., Scherer, G., Jarczyk, L., Heller, W.D., and Neurath, G.B. Phar~acckinetics cf 3-hydroxycotinine. In: The Pharmacology of Nicotine. M.J. Rand and K. Thurau, eds. IRL Press, Washington, DC 1988, pp. 25-28. 24. Langone, J.J. and Van Vunakis, H. Radioimmunoassay of nicotine, cot!nine, and gamma-(3-pyridyl)-gamma-cxo-N- methylbutyramide. Methods Enzy~ol. 84: 628-~40, 1982. 25, Gritz, E.R., Baer-Weiss, V., Benowitz, N.L., Van Vunakis, H., and JarVik, M.E. Plasma nicotine and cotinine concentrations in habitual smokeless tobacco users. Olin. Pharmacol. Ther. 30: 201- 205, 1981. 25a. Cu~¢all, M., VAIa, E.K., Englund, G., Enzell, C.R. Urinary excretion of nicotine and major metabolites. Presented at 43rd Tobacco Chemists' Res. CoNf., Richmond, VA., Oct 2-4, 1989. 25b. Schepers, G., Walk, R.A. Co%inine determination by ~mmunoassay may be influenced by other nicotine metabolites. Arch. Toxicol. 62: 395-397, 1988. 26. Biber, A.,°Scherer, G.I Hoepfnern I., Adlkofer, F., Heller, W.-D., Haddow, J.B., and Knight, G.J. Dete~ination of nicotine and cotinine in human serum and urine: an interlaboratory study. Toxicol. Left. 35: 45-92, 1987. 27. Harke, H.P. Zum.Problem des Passiv-Rauchens. Muench. Mad. Woohenschr. 112: 2828-1834, 1970. 28. Cano, J.P., Ca~alin, J., Badre, R., Duma, C., Viala, A., and Guillerme, R. Determ!nation de la nicQtine par chromatographie en phase gazeuse. I~. Appl. Ann. PhaZ-m. France 28: 683-840, 1970. 29~ Russell, M.A.H. and Feyerabend, C. Blood and urinary nicotine in nonsmokers. Lancet I: 179-181, 1975. 30. Hoffmann, D., Naley, N.J., Adams, ~.D., and Brunnemann, K.D. Tobacco sidestream smoke. Uptake by nonsmokers. Brev. Med. 13: 688-617,1984. 31. Eudy, L.W., Thome, B.A., Heavner, D.L., GreeN, C.R., and Zngebrethsen, B.$. Studies on the vapor-particulate phase distribution of environmental nicotine by selective trapping and detection methods. Proc. 79th Ann. Mtg. Air Pollution Control Association, Minneapolis, June 22-27, 14 p., 1988. 32. Hoffmann, Brunnemann, K.D., Haley, N.J., Sepkovic, D.W., and Adams, J.D. Nicotine uptake by nonsmokers exposed to passive smoking under controlled conditions and the elimination of cotinine. Proc. 4th InternBtional Conference on Indoor Air Quality and Climate, Berlin, "Indoor Air '87", Volume 2: 13- 58

Draft - Do not cite or quote I?,1987. 39. Greenberg, R.A., Maley, N.J., Etzel, R.A, and Lode, F.A. Measuring the exposure of infants to tobacco smoke. New Engl. J. Med. 310: 1075-1078, 1984. 34. Haley, N.J., Sepk~vlc, D.W., Louis, E.T°, and Boffmann, D. Abso~ti~n and eliminatioz of nicotine by smokers, nonsmokers and chewers of nicotine gum. In: The Pharmacology of Nicotine, Rand, M°J. and Thurau, K., eds., IP~L Press, Washington, DC, 1998, pp. 2C- 21. ~5. Goldstmin, G.M., Collier, A., Etzel, R., Lewtas, J., and Haley, N.J. Elimination of urinary cotizine in children exposed to known levels of sidestream cigarette smoke. Proc. 4th International Conference on I~dOQ~ Air ~uality and Climate, Berlin, ,Indoor Air 187", Volume 2: 61-67, 1987. 36. Etzel, R.A., Greenberg, R.A., Haley, N.J., and Lode, ~°A. Urinary Coti~ine excretion in neonates exposed to tobacco smoke products in utero. ~. Pediatr. 107: 146-149, 1989. 37. Scherer, G., Westphal, K°, Sorsa~ M., and Adlkofer, F. Quantitative and qualitative differences in tobacco smok~ uptake between ~ctive and passive smoking. In: "Indoor and Am~bien£ Air %uality~, R. Pe~ry and P.W. Kirk~ eds., L~ndon, 1988. pp 189-194. 38. Benowibz, M.L., Kuyt, F., Sacob, P., J=nes~ R.T. Ill., and Osman, A.-L° cotiniDe disposition and effect, clin. Pharmacol. Ther. 14: 604-611, 1983. 39. Jarvis, M.J., Russell, M.A.H., Benowitz, N.L., and Feyerabend, C. Elimination of c3Zinine from body fluids. Am. J. Publ. Health 78: 696-698, 1988. 40. Becht, E.S. and Koff~ann, Do Tobacco-s~ecific nitrosamines, an important group of carcinogens i~ tobacco and tobac=c smoke. Carcinogenesis 9: 87~-884, 1988° 41. Jarvis, M.J., Tunstall-Pedoe, H., Feyerabendl C., Vessey, C., and Saloojee, Y. Biochemical markers of smoke absorption and self- reported exposure to passive smoking. J. Epldemiol° Com~. Health 38: 335-339, 1984. 42. Fsyerabend, Higgenbottam, and Russell, M.A.H. Nicotine c~ncmntrations in urine and saliva of smokers and nonsmokers. Brit° Med°J. 284: I002-i004, 1982. 43. Foliart, D., Bencwitz, N.L., and Becket, C°E. ~assi~e absorption of nicotine in airline flight attendants. (Lettmr) New Engl. J. Med, 309: 1105. 1983. 99

Draft - Do not cit~ or quote 44. 3arvis, M.J., Russell, M.A.H., and Feyerabend, C. Absorp- under natural conditions of exposure. Thorax 28: 829-838, 1983. 45. Weld, N.J., Eoreham, A., Bailey, A., Ritchie, C., Haddow, J.E., and Knight, G. Urinary cotini~e as marker for breathing other peoples tobacco smoke. Lancet i: 230-231, 1984. 46. Weld, N.J. and Hitchie, C. Validation of studies on lung cancer in nonsmokers married to smokers. Lancet i: 1507, 1984. 47. Mats~kura, S., Tominatot T., Ritanor H., Seino, Y., Hamada, H., Uchihashi, M., Nakajima, H., and Hirota, Y. Effects of environmental tobacco smoke of urinary c~tinine excretion in nonsmokers. New Engl. J. Med. 311: 828-832, 1984. 48. Jarvis, M.J., Russell, M.A.H., Feyerabend, Eiser, J.R., Morgan, P., Gammage, P., and Gray, E.M. Passive exposure to tobacco smoke: saliva cotinine concentrations in a representative population sample of nonsmoking school children. Brit. Med. J. 191: 927-929, 1985. 49. Luck, W. and Na~, H. Nicotine and cotinine concentrations i~ serum and urine of infants exposed via passive smuking or milk from smoking mothers. J. Pedriatr. 107: 816-820, 1985. 50. Pabtishall, E.N., Strope, G.L., Etzel, R.A., Helms, R.W~, Haley, N.J., and Denny, F.W. Serum cotini~e as a measure of tobacco smoke exposure in children. Am. J. Dis. Children 139:1101- 1104, 1985. 51. Schwartz-Bickenbach, Schulte-Hobein, Abt, Plum, C., and Nau, H. Smoking and passive smoking during pregnancy and early infancy: effects on birth weight, lactation period, and cotinine concentrations in mother's milk and infant's urine. Toxicol. Lett. 39: 73-81, 1987. 52. Sepkovic, D.W., Axelrad, C.M., Colosimo, S.G., and Maley, N.J. Measuring tobacco smoke exposure: clinical applications and passive smoking. P~esented at the 80th Ann. Mtg. Air Pollution Control Association 1987, New York, MY, Abstr. 87-80-2, 1987. 53. Jarvis, M.J., McNeill, A.D., Russell, M.A.H., H4est, R.J., Bryant, A. and Feyerabend, C. Passive smoking in adolescents: One year stability of exposure in the home. Lancet i: 1324-132S, 1987. 54. C~ultas, D.B., Howard, C.A., Peake, G.T. Salivary cotinine levels and involuntary tobacco smoke exposure in children and adults in New Mexico. Am. Hey. Resp. Dis. 136: 305-309, 1987. 55. Muranka, H., Higashi, E., Itani, S., and Shimiza, ¥. Evaluation of nicotine, cotinine, thiocyanate, carboxyhemoglobin, and expired carbon monoxide as biochemical tobacco smoke uptake 6O

Draft - Do not cite or quota parameters. Int. Arch. Occup. Environ. Health 60: 37-41, 1988. 5sa. Wagenknecht, L., Cutler, G., Snook, C., Haley, N.J. Black/White differences in cotinine levels among smokers in the United States. In: The Pharmacology of Nicotine, (W. Thurou and M. Rond, eds.), IRL Press, Washington, DC 1988. S5b. Haley, N.J., Colosimo, S,G., Axelrad, C.M., Harris, R., Sepkovic, D.W. Biochemical validation of self-reported exposure to environmental tobacco smoke. Environ. Res, 49: 127-135, 1989. 56. U.S. Department of Health and Human Services. "The Health Consequences of Involuntary Smoking". A report of the Surgeon General. DHHS (CDC) 87-8398, 1986, 359 p. 57. Palladino, G., AdaMs, J.D., Brunnema~n, N.D., Maley, N.J., Hoffmann, D. Snuff-dipping in college students: a clinical profile. Mi!it. Med. 151: 342-346, 1986. 58. Haley, N.J. and Hoffmann, D. Analysis of nicotine and cotinine in hair to determine cigarette smoker status. Clin. Chem. 21: 1598-1600, 1985. 59. Sepkovic, D.W. and Baley, N.J. Biomedical applications of ¢otinine quantitation in smoking related research. Am. $. Public Health 75: 669-564, 1985. 60. U.S. Department of Health and Human Services. "The Health Consequences of Smoking - Nicotine Addiotion". A report of the Surgeon General, DHHS (CDC) 88-8406, 1988, 618 p. 61. Mumford, J.L., Forehand, L., Burton, R., Lewtas, J., Hammond, S.K., and Haley, N.J. Serum and urine cotinine as quantitative measures of passive tobacco smoke exposure in young children. Proc. 4th Internatinal Conference on Indoor Air Quality and Climate, BerliN, "Indoor Air '87", Volume 2: 18-21, 1987. 62. Hill, P., Haley, N.J., and Wynder, E.L. cigarette smoking: oarboxyhemoglohin, plasma nicotine, cotinine and thiocyanate vs. self-reported smoking data and cardiovascular disease. $. Chron. Dis. 36: 439-449, 1983. 63. Wald, N., Idle, M., Smith, P.G., and Bailey, A. Carboxyhemoglobin levels in smokers of filtered and plain cigarettes. Lancet l: 110-112, 1977. 94. Jarvis, M.J. and Russell, M.A,H. Measurement and estimation of smoke dosage to nonsmokers from environmental tobacco smoke. Eur. J. Hespirat. Dis. (Suppl) 133: 68-75, IS84. SS. larvis, M.J. Uptake of environmental ~obacco smoke. IARC Sci. Publ. 81: 43-58, 1987. 61

Draf~ - Do not cita or quote 66. Hoffmann, D., Brunnemann, K.D., Adams, J.D., and Haley, N.J. Indoor air pollution by tobacco smoke: model studies on the uptake by nonsmokers. PrQc. 3rd international conference on Indoor Air Quality and Climate, Stockholm, "Indoor Air", VolUme 2: 313-338, 1984. 67. Scherer, G., Westphal, K., Hoepfner, I., Adlkofer, F., and Scrsa, M. Biomonitoring of exposure to potentially mutagenic substances from environmental tobacco smoke. Proc. of the 4th International Conference on Indoor Air Quality and Climata, Berlin, "Indmcr Air '87", Volume 2: 109-114, 1997. 68. Haley, N.J. Axelrad, C.M., and Tilton, K.A. Validation of self-reported smoking behavior: biochemical analysis of cotinine and thiocyanate. Am. J. Publ. Health 73: 1204-1207, 1983. 69. Easuga, B., Matsukit H.l Osaka, F.I and Inoue, M. The study on the relationship between urinary hydroxyproline and creatinine ratio from the viewpoint of public health. Tokai J. E~p. clin. Bed. 4: 343-351, 1979. 70. Guerin, M.R. Formation and physico-chemical nature of sidestream smoke. IARC Sci. P~bl. "81: 11-24, 1987. 71. Repace, J.L., Indoor concentrations of environmental tobacco smoke: field studies. IARC Sci. P~hl. 81: 141-162, 1987. 72. Adlkcfer, F., Scherer, G., and Holler, W.D. Hydroxyproline excretion in urine of smokers and passive smokers. Prey. Med. 13: 670-679, 1984. 73. Hoffmann, D. and Br~nnemann, K.D. Endogenous formation of N- nitroscproline in cigarette smokers. Cancer Res. 49: 5570- 5574, 1983. 74. Ladd, K.F., Newmark, H.L., and Archer, M.C. N-nitrosation in smokers and nonsmokers. J. Natl. Cancer Inst. 73: 83-87, 1984, 79. Tsuda, M., Nutsume, J., Sato, S., Mirayama, F, Makizoe, T. and Sugimura, T. Increase in the levels of H-nitrosoproline, N- nitrosothioproline, and N-nitroso-2-methylthioproline in human urine by cigarette smoking. Cancer Left. 30: i17-124, 1986. 7~. Lu, S~H., Ohshima~ H~l FUA H~M., Tian. LiL F,M., Blettner, M-± Wahrendorf, J., and Bartsoh, H. Urinary excretion of N- nitrosamino acids and nitrate by inhabitants of high- and low-risk areas for esophageal cancer in Northern China: endogenous formation of nitrosoproline and its inhibition by vitamin C. Cancer Res. 45: 1485-1491, 1985. 77. Scherer , G. and Adlkofer, P. Endogenous formation of N- 62

Draft - D~ not cita o~ ~ote nitrosoproline in smokers and nonsmokers. Banbury Rpt. 23: 137- 147, 1986o 78. Brunnemann, K.D., Scott, S.C., Haley, N.J., and Hoffmann, D. Endogenous formation of N-nitrosoproline upon cigarette smoke inhalation. IA~C Sci. Pl%bl. 57: 819-82a, 1984. 79. ~atrianakos, C. and Hoffmann, D. Chemical studies on tobacco smoke LXIV. On the analysis of aromatic amines in Cigarette smoke. J. Anal. Toxicol. 3: 150-154, 1979. 80. EI-Bayoumy, K., Donahue, J.M., He,hi, S.S., and Hoffmann, D. Identiflcation and quantitative determination of aniline and toluidine in human u~in~, cancer Res. 46~ 606460~7, 19@6, 81. International A~ency for Researoh on Cancer. "Tobacco Smoking, ~A/~C Monogr. 38: 1986, 421 ~. 82° Heinonen, T~, Kytoniemi, v., Sorsa, M., and Vhinio, H. tArinary excretion of thioethers among low-tar and medium-tar cigarette smokers. Internatl. Arch. Occup. Environ. Health 52: 11-16, 1983. 83. Yamasaki, E. a~d Ames, B.N. Concentration of m~tagens ~rom urine by adsorption with thQ nonpolar resin EAD-2: cigarette smokers have mutagenio urine. Proc. Natl. Aoad. Sol. U.S.A. 74: 3555-3559, 1977. 84. Sasson, I.M., Coleman, D.T., LaVole, E.~., Hoffmann, D., a~d Wynder E.L. Mutagens in hu~an urine. Effeots of cigarette smoking and diet. Murat. Res. 158: 149-159, 1985. 85. Scherer, G., Westphal, g., Biber, A., Hoepfnerp I., and Adlkofer, F. Urinary mutagenloity after controlled exposure to environmental tobacco smoke (ETS). Toxical. I~tt° 35: 135-140, 1987. 86. Mohtashamipur, E.l Mueller, G., Norpoth, K., Endrikat, M., a~d Stuecksr, W. Urinary excretion of mutagens in passive smokers. Toxicol. Letters 35: 141-146, 1987.. 87. Husgafvel-Purslainen, N., Sorsa, M., Engstromf K., and Einistoe, P. Passive smoking at work: bioehemlcal and biological measures of exposure to environmental tobacco smoke. I~t. Arch. Oocup. Environ. Health 59: ~37-345, 1987. 88° Ling, P.II, Lofroth, G., and Lewtas, ~. Mutagenlc determination of passive smoking. Toxicol. LetZ. 35: 147-151, 1987. 89. Harris, C.C., Vahakangas, K., Newman, M.S., Trivets, G.E., Shamsuddi~, A., Sinapoli, N., Mann, D., and Wright, W.E° Detection 63

Draft - Do not cite or quote 66. Hoffmann, D., Brunnemann, K.D., Adams, J.D., and Haley, N.J. Indoor air pollution by tobacco smoke: model studies on the uptake by nonsmokers. Proc. 3rd International Conference on Indoor Air Quality and Climate, Stockholm, "indoor Air", Volume 2: 313-3S8, 1984. 67. Soberer, G., Westphal, K., Hoepfner, ~., Adlkofer, F., and Sorsa, M. Biomonitoring of exposure to potentially mutagenio substances from environmental tobacco smoke. Proc. of the 4th International Conference on Indoor Air Quality and Climate, Berlin, "Tndoor Air '87", Volume 2: 109-114, 1987. 68. Haley, N.J. Axelrad, C.M., and Tilton, K.A. Validation of self-reported smoking behavior: biochemical analysis of cctinine and thiocyanate. Am. J. Publ. Health 73: 1204-1207, 1983. 69. Nasuga, H., Matsuki, H., Osaka, F., and Inoue, M. The study on the relationship between urinary hydroxyproline and creazinine ratio from the viewpoint Of public health. Tokai J. EIp. Clin. Med. 4: 343-351, 1979. 70. Guerin, M.R. Formation and physico-chemical nature of sidestream smoke. IARC Sci. PUbl. '81: 11-24, 1987. 71. Repace, J.L., Indoor concentrations of environmental tobacco smoke: field studies. IA~C Sci. Publ. 81: 141-162, 1987. 72. Adlkofer, F., Schermr, G., and Holler, W.D. Hydroxyprcline excretlcn in urine of smokers and passive smokers. Prey. Med° 19: 670-679, 1984. 73. Hoffmann, D. and Brunnemann; K.D. Endogenous formation of N- nitrosoproline in cigarette smokers. Cancer Res. 43: 5579- 5574, 1983. 74. Ladd, K.F., Newmark, H,L., and Archer, M.C. N-nitrosation in smokers and nonsmokers. J. Natl. Cancer Inst. 73: 83-87, 1984, 75. Tsuda, M., Nutsume, J., Satol S., Hiraya~a, F, Kakizoe, T. and Sugimura, T. Increase in the levels of N-nitrosoprcline, N- nitrosothioproline, and N-nitroso-2-methylthioproline in human urine by cigarette smoking. Cancer Left. 20: i17-124, 1986. 76° Lu, S.H., Ohshima, H., Fu, H.M., Tian, Li, F.M., Blettner, M., Wahrendorf, $,, and Bartsoh, H. Urinary excretion of N- nitrosamino acids and nitrate by inhabitants of high- and low-risk areas for esophageal cancer in Northern China: endogenous formation of nitrosoproline and its inhibition by vitamin C. Cancer Res. 45: 1485-1491, 1986. 77. Schermr , G. and Adlkofer, F. Endogenous formation of N- 62

Draf~ - Do no~ cite or q~cte of benzo(a)pyrene diol epoxlde-DNA adducts in peripheral blood !ymphocytes and antibodies t~ the adducts in serum from coke oven workers. Prec. Natl. Acad. SCi. U.S.A. 82: 6672-6676, 1985. 90. Everson, R.E., Randera~h, E., Santella, S.A., Cefale, R.C.~ Avitts, T.A., and Randeratb, K. Detection of smoking-related covalent DNA adducts in human placenta. Science 231: 54-57, 1986. 91. Perera~ F.P. t Poirier, M.C., Yuspa, S.H., Nakayama, J., Saretzki, A., Curnen, M.M., Kn@wles, D.M., and Weinstein, I.E. A pilot project in molecular cancer !pidem!ology: determination cf benzo(a)pyrene-DNA adduces in animal and human tissues by immuncassays. Carcinogenesis S: 1405-1410, 1982. 92. Seland, F.A. and Eadlubar, F.F. Factors involved in the induction of urinary bladder cancer by aromatic amines. Banbury Rpt. 23: 315-326, 1986. 93. Neumann, H.G. Analysis of hemoglobin as a dose monitor for alkyla~ing and arylating agents. Arch. Toxic@l. 56: I-6, 1984. 94. Green, L.C~," Skipper, P.L., Juresky, R.J., Bryant, M.S., and Tannenbaum, S.R. In vivo dcsime~ry of 4-aminobiphenyl in raus via a cysteine adduct in hemoglobin. Cancer Res. 44: 4254-4259, 1984. 95. Bryant, M.S., Skipper, P.L., Tannenbaum, S.R., and Maclure, M. Hemoglobin adducts of 4-aminobiphenyl in smokers and nonsmokers. Cancer Res. 47: 602-608, 1987. 96. Wynder, E,L. ~nd Hoffmann, D. "Tobacco and Tobacco Smoke. Studies in Experimental Tobacco Carcinogenesis." Academic Press, New Yc~k, R¥, 1967, 730 p. 97. Binder, H. and Lindner, W. Bestimmung yon Aethylenoxyd im Rash garantiert unbegas~er Zigaretten. Fachliche Mitt. Oesterr. Tabakregie 13: 215-220, 1972. 98. International Agency for Research on Cancer. '.overall Evaluations of Carcinogenicity: An Updating of IARC Monographs, Volume 1-42." IARC Monogr. Suppl. 7: 1987, 440 p. 99. Tcrnqvist, M., Oste~an-Golkars, S., ~autiainen, A., Jensen, S., Fainter, P.S., and Ehrenberg, L. Tissue doses of ethylene oxide in cigarette smokers determined from adduct levels in hemoglobin. Carcinogenesis 7: 1519-1521, 1986. I00. Hecht, S.S., Carmella, S.G., Trushin, N., Spratt, T.E., Foiles, P.G., and Hoffmann, D. Approaches to the development of assays fer interaction of tobacco-specific nitrosamines with hemoglobin and DNA. IARC Sol. Publ. 89: 121- 128, 1988. I01. Benner, C.L., Bayona, J.M., Caka, F.M., Tang, R., Lewis, L., 64

Draf~ - DO nc~ cite or quote Crawford, J., Lamb, J.D., Lee, M.L., Lewis, E.A., Hansen, L.D., and Ea~ough, D.J. Chemical Composition of Tobacco Smoke. 2. Particulate Phase Compounds. Environ. Sci° Technol. 23: 688-699, 1989. 65

Oraf~ - Do no~ cite or q1.1o~e Fi~ure~ ~nd Tables for Chapter 4 66

Draft - DO not cite OF quote ,M-Q oi"o //2 c :.o.l glob~ adduct I I PI*~ "OH 0 F Figure Z.

i " Table 4 continued ... ?, Number of N0nSln0 ke r Group NoNslnoker s Results Reference Children and adults 529 males Cotinine/Saliva (ng/ml) Coultas 768 ~emales smokers in family eta[. (53) NOn~ ode > two a] <5 years old 0.0 (0.0-2.5} 3.8 {0.0-6.1) 5.4 (3.2-7.7) b) "6-12 years old O.o (0.0-2.]) 2.0 (0.0-3.8) 5.2 (1.5-7.0) c) 3-17 years old O.0 (0.0-2.0} 2.9 (0.0-4.9) 4.1 (2.7 7.6) d) 18-29 years old 0.0 (0.0-2.6) 0.0 (0.0-5.8} 0.0 (0.0-4.4) e) 30-64 years old 0.0 (0.0-2.7} 1.9 (0.0-4.5} 4.4 (|.8-]1.0) f) > 65 years old 0.0 (0.0-2.&) 3.6 (0.0-6.5) 0.0 *Numbers in parenthesis median values. U H, ! 0 0 ¢t CJ 0 0 m

Table 4 continued ... Number of Nonsmoker Group Nonsmokers Results Reference Huntclpal workers Cotiine/Brine {n~/m~ creatinine) Sepkovic et el., I. ET$ exposure in the ~2T-- workplace a) no exposure 25 4.5~0.6 b) llght expsoure 126 6.6i0.6 C) moderato exposure 84 7.2~0.8 d) heavy exposure 32 8.4~1.3 II. ETS exposure in the home a} no exposure 77 6.|i0.8 b) light exposure 83 6.7±0.6 e) moderate exposure 71 7.8"1.1 d) heavy exposure 34 7.6~1.3 School ig~~ (11-16 yrs) ETS exposure in the home a) neither parent smokes b) father smoke~ only c) mother smokes only d) both parents smoke 104 1.1~0.5 76 2.0"0.6 40 3.2*0.8 110 5.0~1.0 Jarvis et al., (53-~ continued ... u I O 0 O O D o O

Table 4 continued ... #, .. i Number of Nonsmoker Group Nonsmokers Resutts Reference Cotinlne/urine (n~/mj creatinine) Neonahes and infants No. . No. Schwartz a) Mother smokes, exp'd I exp[d If -- Bicken- bach et. breastfeeds 20 12 (|756) 0 -3520 8 (9]5) 488-2440 dl., ~|) b) Mother smokes, feeds bottle 16 4 (47) 0 - 160 ~2 (107) 0- 341 c) Father smokes 18 10 (0) 8 (0) 0- 308 d} No exposure in the home 15 9 (0) 6 (0) {* I 0 {* 0 0 {t

Tah~e 4 co~tL~ued ... Number o~ Nonsmoker GrOUp Nonsmokers ResoIts Refere,ce Neonates and infants a) No exposure (4-8 days old) b) Exposure via breast fee-~ng (3-8 days old) c) Passive smoking (2.5-6 months old) d% E~posure ~i~ breast fee~ng and passive smoking (I-12 months old) NJcotlne (ng/mg creatinine) cotlnine 10 6~ 0 - 14 (0) 0- 56 19 (14) 5 -110 (100) 10-555 tO (35) 4.7-218 (327) 117-780 9 (12) 3.0~ 42 (550) 225-870 Luck and Nau, {491 Infants (age 3-15 mouths) ¢otlnine/ Serum (n~/ml) Pattishall exposure in the et al., home ~0~-- Black infants a) no exposure 9 1.0 (1.87~2.38) Pattishall et al., 1985 b) passive smoking 15 4.0 (5.27~3.50) (51) White infants a) no exposure 9 0.0 (0.22~0.44) ! b) passive smoking 5 0.4 (0.90~1.30] .............................................. 0 r~ o ¢-t m o 0 m

l Table 4~continued ... Nonsmoker Group Number of Nonsmokers Results Refe[enc~ Husbands of CotinlnetlU[ine (ng/ml[ Wald and a) nonsmokers 101 8.5" 1.3 Ritchie, b) smokers 20 25.2±14o8 (46) Nonsmokers a) nonsmokers at home 200 b) smokers at home 272 Cigarettes smoked day in home of nonsmokers; 1- 9 25 10-19 57 20-29 99 30-39 38 > 40 28 unknown 25 Cotinine/Urine [9~.d/m9 creatinine) 0.5 ~0.09 0.79~0.1 Cotinine/Urine (p~/m@) creatinine) 0.31~0.08 0.42~0.l 0.~7"0.19 1.03~0.25 1.56£0.57 0.56~0.16 Matsnkura et al.0 (47) Infants (<10 months, Nicotlne/Urlne cotinine/Urine Greenber9 not breas~fed) ~mg creatinine--~ (,@/m~) et al.; ~) not exposed to ETS 18 ~5~ 4 (0-125) ~3~ b) exposed to ETS 2B 53 (0-370) 351 (41-T,Sf15) School children (11-16 yrs) a) Nei%her parant smoked 269 b) Only father smoked 96 c) Only mother smoked 76 d) Both parents smoked ]28 Cotinine/Saliva (ng/ml)" 0.44~006B darvis eL 1.31±1.21 a~., (4~ 1.95"1.71 3.38"2.45 l g O N Continued ... ,~ o ct Ib

Table 4, ,, Uptake oE nicotine by nonsmokers exposed to ETS under daily llfe conditions Honsmoker Group Number of Nonsmokers Results Reference Hospital personnel (?8 min in smoke- filled room) Nicotine/Urine [ng/ml) 14 12.4~16.9 Russell and Feyerabend ]3 8.9~9.1 (29} Hospital personnel and outpatients a) no exposure to ETS b) exposed to ETS Nicotine/Saliva (rig/m1) 26 5.9 7.5 Feyerabend 30 10.1 21.6 et al. (42) Flight attendants Nicotine/Serum (n~/ml) 6 pre flight: 1.6~0.8 Follart et al. pest flight~ 3.2~1.g (43) m Office workers 7 Content/m1 Nicotine (n~] Cotinlne (rig) Oarvls et al. a) I]:30 a.m. sample sa~a a}].90 b)43~63 a)l.SO b)8.04 (44) b) 7:45 p.m. sample serum 0.76 2.49 1.07 7.33 i after 2 hr stay urine 10.57 92.63 4.80 12.94 in pub Hospital staff and Cotinlne/Urine (ng/ml) Wald et al. outpations (45) ---- a) no exposure to ETS 22 2.0 (0.0 - 9.3 b) exposed to ETS 190 6.0 (1.4 -22.0) 0 Continued ... m • "

Table 3. Apl~roximate Relation¢ of NicotiDe as a Par~eter Between Nonsmokers, Passive Smokers. and Active SmokerBa (41) Nicotine/Cotinine Nonsmokers without Nonsmokers with ETS Exposure ETS Exposure ME. - 46 NO. = 54 Mean Smokers' Mean SmokersI Value Value Value Value Active Smokers NO. = 94 Mean Value Nicotine (ng/ml) in p~asma 1.0 7 0.8 5.5 14.B in saliva 1.8 0.6 5.5 0.8 673 in urine 3.9 0.2 12.1' 0.7 1,750 Cotinlne (ng/ml) in plasma 0.8 0.] 2.0* 0.7 275 in saliva 0.7 0.2 2.5"* 0.8 310 in urine 1.6 0.1 7.7** 0.6 1,390 I o 0 aDi[ferences ~tween ,onsmokers exposed to ~TS compare~ with nonsmokews without exposure: *p<0.01~ ** p<0.0Ol. 0 Q 0 o m

No. of ETS-Conditions Passive Smokers Table 2 continued. Results Investigator8 Room - 16 m3 6 4 cigarettes con- currently and con- tinuously smoked for 80 min; 6 air exch./hr. (200 g nlcotine/m3; 20 ppm CO) Time durln@ exposure Nicotine Cotinlne O Saliva 3 1.0 Plasma 0.2 0.9 SEine |7 |4 80 min. Saliva 730 1.4 Plasma 0.5 1.3 Urine 84 28 Time following exposure 30 min. Saliva 148 1.7 Plasma 0.4 1.8 150 " Saliva 17 3.1 Plas~ 0.7 2.9 Urine 100 45 300 " Saliva 7 3.5 Plasma 0.6 3.2 Urine 48 55 Hoffmann et al., (30) 1984 U M t *Nicotine and cotfnlne were measured in urine as ng/mg creatinine. 0 0 0 0 ct D 0 =1 o ~r

| Table 2. Uptake of nicotine by nonsmokers exposed to E'PS under controlled Conditions No. of ETS-Conditions Passive Results Investigator(s) Smokers '" Room - 170 m3 (11 smokers) 100 cigarettes were 7 smoked during 2 hrs; no ventilation (30 ppm CO) (b) same conditions as above 7 (a) but with ventilation (5 ppm CO) Room - 66 m3 (4 cigarette smokers) (a) Day 1, nonsmokin9 2 2, 98 cig's smoked 3, ~21 4, 98 5, 88 (b) Day 1, 97 2 2, 06 u 3, 04 " 4, 103 Room - 43 m3 9 smokers Consumed 12 00 cigarettes + 2 cigars no ventilation (38 ppm CO) • Urinary excretion H~coLine: 10~6.8 pg/6 hrs. CoLini.e: 35~34.5 pq/6 hrs. Nicotine: 18±7 p9/6 hrs. Contininez 19~9.4 pg/6 hrs. NicoLine/tlrlne (P9/24 hrs) Cano et al. (28) 0 - 0 35 - 44 50 - 61 62.5 - 70 47 - 50 23 - 34 22.5 - 58 47.5 - 69 I 32 - 65 Nicotine~Plasma (pg/ml) Russell and Before exposure: 0.73£1.6 Feyerabend d After 78 miu. exposu[e: 0.9~ 0.29 (29) n Nicotine/Urine (n~/ml) w 15 min. after expose: 0B.0~50.7 O continued .., " O D

Draf~ - Do not cita or ~uo~e Table I. Toxic ~d tumorigenic agents in MS and SS Ci@arette Smoke Smoke Constituent streama A (NF) B (F) C (F) D (PF) Tar (mg) Nicotine (mg) Co (mg) Catechol (ug) SaP (ng) Ammonia (ug) NDMA (rig) NPYR (mg) NNN (ng) NNK Ing) MS 20.1 15~6 6.8 0.9 SS 22.6 24.4 20.0 14~I MS 2.04 1.50 0.81 0.15 SS 4.62 4.14 0.54 3.16 MS 13.2 13.7 9.5 1.8 SS 28.3 36.6 03.2 26.8 MS 41.9 71.2 26.9 9.1 SS 58.2 89.9 69.5 117 MS 26.2 17.8 12.2 2.2 88 67.0 45.7 51.7 44.8 MS 76.0 19.4 34.0 40.4 88 524 893 213 236 MS 31.1 4.3 12.] 4.1 SS 735 597 611 685 -_-_ ......... _._ ................. ____.___E___ MS 64.5 10.2 32.7 13.2 SS 117 139 233 234 MS 1007 488 273 66.3 SS 857 307 185 338 MS 425 160 56.2 17.3 SS 1444 752 430 386 a Abbreviations: NF, nonfilter cigarette; F, filter ciga- rette~ PF, cigarette wi~h perforated fil~er tip; B~P, benzo- (a)pyrene; NDMA, N-nitrosodimethylamine; NPYR, N-nitrosopyr- rolidine; NNN, N'-nitrosonornicotine~ NNK, 4-(methylnitros- amino)-1-(3-pyridyl)-1-butanone.

Draf~ - Do not cite or ~ote CHAPTER $ ENVIRONMENTAL TOBACCO SMOKE ~d~D CANCER Jonathan M. Samet, M.D. pulmonary Division Department of Medicine University of Hew Mexico Albuquerque, NM 87131 Introduction L~ng cancer, an uncommon malignancy at the start of the century, has become the leading cause of cancer death in the United States (U.S. DHHS 1982). The American Cancer Society estimates that approximately 157,000 lung cancer cases will occur in the United States in 1990. Most cases are rapidly fatal and only a small proportion are cured by surgery or chemotherapy; five-year survival following diagnosis is less than IQ percent. Most lung cancers arise in the larger airways of the lung, the predominant site of deposition of inhaled particles in the size range of 0.5 to 3.0 microns in aercdynamic diameter. Primary cancer of the lung occurs in multiple hietopathological patterns that are generally distinct and classifiable by conventional light microscopy. The principal types of lung cancer are squamous cell carcinoma, small cell carcinoma, adenccarcinoma, and large cell caroincma~ in the general population, these four types account for approximately 30 percent, 20 percent, 25 percent, and 15 percent, respectively, of all lung cancers [Butler et el. 1987). Bronchioloalveolar cell carcinoma represents about 5 percent of all lung cancers. The cellular origins of the various cell types have not been established, ~nd controversy remains concerning the specificity of associations between certain cell types and specific etiologic agents. However, in nonsmokers, adenocarcinoma is the predominant type and small cell cancers occur only rarely. The epidemic rise of lung cancer during this century, stimulated laboratory and epidemiological investigation of its causes. Most of the early epidemiologioal evidence indicated that tobacco smoke was e potent respiratory carcinogen, and in 1964 the Advisory Committee to the Surgeon General of the U.S. Public Health Service concluded that cigarette smoking is a cause of lung cancer (U.S. PHS 1964). The numerous investigations performed subsequently have been consistent with this conclusion. The sssociaticn of lung cancer with cigarette smoking is strongest for squamous cell and small cell cancers, but the other major ~ypes are also caused by cigarette smoking. In active cigarette smokers, the risk of lung cancer increases with both the amount s~oked on a daily basis and 67

Draf~ - Do ~¢~ cita or ~ote with the duration of smoking (U.S. DHHS 1992; Doll and Peto 1979; Pathak etal. 1986). A threshold level of smoking that must be exceeded to cause lung cancer has never been demonstrated; any cigarette smoking is considered to increase lung cancer risk beyond that of the lifelong nonsmoker. In for~ner smokers, %he relative risk of lung cancer declines exponentially in comparison with those who continue to smoke. Agents other than tobacco smoke may also cause lung cancer, and cases Occur in lifelong nonsmokers. A recent study in ~ew Mexico showed that the lifetime risks of lung cancer were 0.5 percent and i.i percent in female and male nonsmokers, respectively (Same% e% el. 1988). Occupational exposures to arsenic, asbestos, chloromethyl ethers, chromium, coke oven fumes, nickel, and radon daughters have been linked to increased lung cancer risk, and many other occupational agmn%s are suspect respiratory carcinogens. A family history of lung cancer is also associated with increased lung cancer risk, although a clear pattern of genetic susceptibility to lung cancer has not been demonstrated. Outdoor air pollution may contain carcinogens and indoor air may have high levels of radon, which causes cancer in exposed underground mimers. A~imal and human studies suggest that low consumption of vitamin A or its precursor, beta-taro%one, may also increase lung cancer risk. While studies linking active smoking %o lung cancer were first published in the late 1940s and early 19909 (U.S. PHS 1964), involuntary exposure of nonsmokers to tobacco smoMe was not considered as a cause of lung cancer in nonsmokers until 1991, when the first two scientific papers on this subject were published. Subsequently, many additional reports have addressed involuntary smoking as a cause of lung cancer in nonsmokers. The World Health Organization (1986), %he U.S. Surgeon General (U.S. DHHS 1996), and the National Research Council (1986) have reviewed the evidence on involuntary smoking and lung cancer from human populations and judged it sufficient to support the conclusion that involuntary inhalation of tobacco smoke hy nonsmokers causes cancer. This chapter reviews that evidence and the conclusions of the research organizations. The chap%mr also addressess the more limited evidence on involuntary smoking and cancer at sites other than the lung. The EDidemiolouical APProach Epidemiology is %he scientific method used to describe the occurrence of disease in. human populations and to dete~ine the causes of disease by studying populations. Descriptive measures of disease occurrence include the incidence rate, which is %he rate at which new cases of disease develop~ the mortality rate, or rate of death; and the prevalence rate, which is the proportion of the popula%ion with disease. TO identify the causes of disease, epidemiologists generally perform either cohort or case-control 68

Draft - DO not clte Or quote studies. Each type ef study provides an estimate ~f relative risk as a measure of the association bet-~een exposure and disease. The relative risk describes the comparative occurrence of disease in exposed compared with nonexposed persons. In a cohort study, the subjects are selected on the basis of their exposure history and followed over time for the development of disease. For example, a study of involuntary smoking and lung cancer might be performed by enrolling nonsmokers married to smokers and another group of nonsmokers married to nonsmokers. The lung cancer risk associated with marriage to a smoker would be estimated by comparing incidence of or mortality from lung cancer in the two groups. ~n a case-control study, cases with the disease cf interest and controls without the disease are identified and their past exposures to factors of interest are assessed, often by interview. For example, a case-control study of lung cancer and involuntary smoking might be conducted by identifying nonsmokers with lung cancer and a suitable control group, and then interviewing the subjects concerning the smoking habits of their spouses, other household members, and colleagues at work. 2- Each type of study has advantages and disadvantages, and the results of both types may be distorted by bias. Misclassification of exposure is of particular concern in studying lung cancer and involuntary smoking. Misclassificaticn of exposure refers to the incorrect categorization of actually exposed subjects as nonexposed and cf ncnexposed as exposed. When mlsclassification occurs randomly, it tends to bias studies towards no association, that is showing negative results; if nonrandom, it may exaggerate or reduce the apparent effect of an exposure. With regard to involuntary smoking and lung cancer, two types of misclassification are of concern. Subjects classified as nonsmckQrs may have actually been active smokers and the degree of exposure of nonsmokers to the smoking of others may not be accurately classified. Misclassification of both types is discussed below in relation to specific studies. The diagnosis of lung cancer is also subject to misclas- sification; a cancer that originated at another primary site and ~ then spread to the lung may be incorrectly diagnosed as a primary cancer of the lung. For example, in the case-control study reported by Garfink~l and colleagues (Garfinkel etal. 1985), 13 percent of cases originally diagnosed as lung cancer were reclas- sified to other sites after histological review~ With regard to exposure misclassification in this study, 40 percent of the oases initially classified as nonsmokers on the basis of info~ation in medical charts were found to be smokers on interview. Confounding refers to bias that occurs when the effect of another risk factor is mixed with the effect of the exposure of interest; thus a confounding factor is a risk factor for disease that is associated 69

Draf~ - Do not cite or quote with the exposure under investigation. For lung cancer in nonsmokers, potential confounding factors include indoor air pollution by radon and combustion products other than environmental tobacco smoke, ambient air pollution, and occupational exposures. Although confounding always merits consideration as an explanation for association, the diversity of the populations in which passive smoking and lung cancer have been associated argues strongly against confounding as the source of the association. Although individual studies may be affected by one or more biases, the totality of the epidemiological evidence as well as other relevant research are considered in judging whether an exposure adversely affects health. A bias potentially important in one study may be unimportant or adequately controlled in another. Thus review of all pertinent literature may show that bias cannot satisfactorily explain an association between exposure and disease. EDidemiolo~ical Evidence on Involuntary Smokina and Lunm Cancer Evidence concerning involuntary smoking and lung cancer has been sought indirectly in descriptive data on mortality rates and directly with case-control and cohort studies. Time trends of lung cancer mortality across this century in nonsmokers have been examined with the rationale that temporally increasing exposure to environmental tobacco smoke should be paralleled by increasing mortality rates (Enstrom 1979; Garflnk$1 1981). These data can only provide indirect evidence on the lung cancer risk associated with involuntary exposure to.tobacco smoke. Enstrom (1979) cal- culated lung cancer mortality ratQs from various nationwide sources for the period 1914-1968 and concluded that'a real increase had occurred among nonsmoking males after 1935. In contrast, Garfinkel (1981) found no hime trends of lung cancer mortality in ncnsmoking participants in two cohort studies, the Dorn Study of U.S. veterans, 1954-1969, and the American Cancer Society study, 1960- 1972. Most of the case-control and the cohort studies indicate in- creased lung cancer risk in nonsmokers married to smokers, but these studies do not uniformly show increased risk for scurces of exposure other than smoking by the spouse [Tables 1 and 2). The first two major epidemiological studies were reported in 19el by Hirayama and Trichopoulos and colleagues (Tables 1 and 2). Mirayama conducted a cohort study of 91,540 nonsmoking women in Japan. Mortality in these women was assessed over a 14-year follow-up period. The ratio of the observed to expected numbers of lung cancer deaths increased in a statistically significant pattern with the amount smoked by the husbands. The findings could nob be explained by other factors, such as age and occupation of the husband, and were unchanged when the follow-up was extended by several years (Mirayama 1984). After its publication, the report of this study received intensive scrutiny, and correspondence in the British Medical Journal, which had published it, raised concern 7O

Draft - DO not =its or quote abou~ various aspects of the study,s methods and findings. In his responses to the correspondence, Hirayama satisfactorily answered most of the criticisms, although he could not eliminate the possibility of unreported smoking by women classified as nonsmokers. If self-reported nonsmokers married to smokers were actually more likely to he smokers, then the resulting bias would tend to indicate an increased risk from marriage to a smoker. Based on the same population, Hirayama has also reported significantly increased risk of lung cancer for nonsmoking married men whose wives smoke (Mirayama 1984). In 1981, Trichopoulos and coworkers (1981) also reported in- creased lung cancer risk in nonsmmking women married to cigarette smokers (Table 2). These investigators conducted a case-control study in Athens, Greece, that included selected histological types of lung cancer and ~ontrol subjects ascertained at a hospital for orthopedic disorders. The finding of increased risk was unchanged when the case and control series were enlarged (Triohopoulos et el. 1983). The results of subsequently reported case-control studies have also demonstrate~" significantly increased risk of lung can=or in nonsmokers exposed to environmental tobacco smoke (Table 2). The findings from the more recent reports based on studies throughout the world greatly strengthen the evldencm from the earlier studies. several of the newer studies included relatively large numbers of nonsmokers (Garfinkel etal. 1985; Akiba etal. 1986t Dalager et el. 1986t Lamet el. 1987; Gao etal. 1987). Furthermore, in most of the newer studies, involuntary smoking was assessed in greater detail than in the earlier reports. The results of two other investigations have also been interpreted as showing an increased lung cancer risk associated with involuntary smoking, but both of these studies have limitations. Enoth and cowcrkers (1983), in Go,any, described 59 lung cancer cases in females of whom 39 were nonsmokers. Based on census data, these investigators projected that a much greater than expected proportion of the nonsmokers had lived in households with smokers. In another report, Gillis etal. (1984) described the preliminary results of a cohort study of 16,171 males and females in western Sc~t!and (Table I) ; exposure to tobacco smoke in the home increased the lung cancer risk for nonsmoking men but not for nonsmoking women. This observation was based on only 16 oases of lung cancer in nonsmokers, however. Other investigations indicate lesser or no effects of exposure to environmental tobacco smoke on lung cancer risk (Tables 1 and 2). In these studies, however, the statistical uncertainty is large because of the relatively small numbers of subjects; ac- cordingly, the apparently negative findings are statistically compatible with the findings of those studies judged as positive. Two separate case-control studies in Hong Kong, where lung cancer 71

Draft - Do not cite OE quote incidence rates in females are particularly high, did not indicate excess risk from involuntary smoking (Chan et el. 1979; Chan and Fung 1982; Koo et el. 1984; 1985; 1987). ~n the more recent of the two studies, the investigators comprehensively assessed cumulative exposure from home and workplace sources, but misclassification of exposure may have biased towards the negative results. A subsequent study in Hong Kong did find a significant association of spouse smoking and lung cancer risk (Lamet el. 1997). Lee and coworkers (Lee et el. 1986) in England reported a small case- control study with negative findings, but the statistical power of that study is limited. Another recent hospltal-based case-control study, conducted in Japan, also failed to show an association be- tween lung cancer risk and spouse smoking (Shimizu etal. 1988). The results of the American Cancer society's cohort study of luhg cancer mortality in 176,139 ncnsmoking women have also been considered by many ~s not Bhowing an increased risk in those par- ticipants married to smokers IGarfinkel 1991). However, the risks for the ncnsmoking women with smoking husbands were increased somewhat, but the increase was not statistically significant. Misclassifioati~n of exposure from active and involuntary smoking may have affected the results of this study. Preliminary results from a nationwide case-control study also did not demonstrate increased lung cancer risk from domestic exposure to tobacco smoke (Mabat and Wynder 1984), but the auger of subjects was small. Two case-control studies of nonsmokers and smokers with selected histological types of lung cancer did not provide strong evidence for increased risk from involuntary smoking (wu et al. 1985; Brownson etal. 1987). However, both studies included only small numbers of nonsmokers. Conclusions on Involuntary Smokinu and Lung Cancer Scientists draw cn a wide range of evidence in judging whether an agent, such as environmental tobacco smoke, causes disease, in addition to epidemiological data, the findings of laboratory studies involving in-vitro systems and of animal studies involving exposure to the agent are often relevant. Criteria have been developed for guidance in making judgments on the causality of exposure-disease relationships, but these criteria only provide guidelines, not strict rules of evidence (U.S. PHS 1964; Rothman 1986). Interpretation of the evidence on particular exposure- disease relationships often requires review by multidisciplinary panels of scientists who are instructed to reach a consensus, often In a setting of substantial uncertainty. For example, the World Health Organization regularly convenes panels of scientists to address the carcinogenicity OZ environmental agents. For environmental tobacco smoke and lung cancer, the evidence has been considered by scientists convened by the International 72

Draft ° Do not cita or ~te Agency for Research on Cancer of the World Health Organization, the National Research Council, and the U.S. surgeon General (Table 3). All three groups concluded that environmental tobacco smoke causes lung cancer among nonsmokers, although the approach used by each group was different. Consensus among the three groups, in spite of differing methodology, strengthens the determination that involuntary smoking causes lung cancer. For all three types, the biological plausibility of this association was supported by the evidence on active smoking and lung cancer, knowledge of the constituents of environmental tobacco smoke, and data demonstrating the uptake of tobacco smoke by nonsmokers. The International Agency for Research on Cancer of the World Health Organization (1986) reviewed the evidence available through the end of 1984. It reached its conclusion concerning involuntary smoking and lung cancer largely on the basis of biological plausibility. The agency cited the characteristics cf sidestream and mainstream smoke, the absorption of tobacco smoke mater±als during involuntary smoking, and the nature of dose-response relationships for carcinogenesis, which project some risk for any ~evel of exposure. In reaching its conclusion, the National Research Council comities considered the biological plausibility of an association between environmental tobacco smoke exposure and lung cancer and the supporting epidemiological evidence, available through mid- 1986. The committee carefully considered the sources of bias that may have affected the spidemiological studies and concluded that the association documented in the studies could not be attributed solely to bias. Based on a pooled analysis of the epidemlologlcal data and adjustment for bias, the report's authors concluded tha~ the best estimate for the excess risk of lung cancer in nonsmokers married to smokers was 25%. The 1986 report of the U.S. Surgeon General also characterized involuntary smoking as a cause of lung cancer in nonsmokers. This conclusion was based on the extensive information already available on the carcinogenicity of active smoking, on ~he qualltativs similarities between environmental tobacco smoke and mainstream smoke, and on the epidemiologic data on involuntary smoking. The extent of the lung cancer hazard associated with involuntary smoking in the United states has appeared uncertain. (U.S. DHHS 1986; Weiss 1986). The epidemiological studies p;ovide varying and imprecise measures of the risk (Tables 1 and 2), and exposures to environmental tobacco smoke have not been characterized for large and representative population samples. Thus, any risk assessments for involuntary smoking and lung cancer are subject to substantial uncertainty. Nevertheless, risk assessment can provide insight into the magnitude of the lung cancer problem posed by involuntary smoking. 73

Draft - Do not cite or c~aote Repace and Lowrey [1995) used data on lung cancer mortality in Seventh Day Adventists, a nonsmoking group, tm estimate the effect of exposure to environmental tobacco smoke in increasing lung cancer risk. Their analysis led to an estimate of 4,666 lung cancer deaths per year attributable to environmental tobacco smoke exposure. A later estimate gave 3,450 female lung cancer deaths and 1,440 male lung cancer deaths per year.(Repace and Lowrey, 1986) An appendix to the National Research Council's 1986 report provides estimates of the numbers of lung cancer deaths attributable tc passive smoking. For the year 1985, the risk assessment projects approximately 1,000 lung canoer deaths in males and 2,000 tc 2,000 lung cancer deaths in females attributable to environmental tobacco smoke. Wells (1988) attributed 3,000 lung cancer cases annually in the U.S. to involuntary smoking. A recent review of S published risk assessments of environmental tobacco smoke and lung cancer found they averaged about 6,500 ~ 2,800 lung cancer deaths per year (Repace & Lowrey, 1990). Further epidemiclogical studies of involuntary smoking and lung cancer are in progress. These studies should refine cur understanding of exposure-response relationships for lung cancer and exposure to environmental tobacco smoke. Other investigations are addressing the characteristics and toxicity of environmental tobacco smoke and patterns of exposure to environmental tobacco smoke. While the results cf these new studies will provide needed information for scientific purposes, the available data and the conclusions of the scientific community already provide a compelling rationale for reducing involuntary exposure to environmental tobacco smoke. Involuntary Smoking and Cancer at Sites Other Than the Lunu Several reports have suggested that exposure to environmental tobacco smoke may increase risk of cancer at sites other than the lung. One study found that in children, maternal exposure to environmental tobacco smoke during pregnancy was associated with increased risk of brain tumors (Preston-Martin et al. 1982), and in another study paternal but not maternal smoking increased the risk of childhood rhabdomyosarcoma, a cancer of the soft tissues (Grufferman etal. 1982). In adults, involuntary smoking has been linked to a generally increased risk of malignancy (Miller 1984). Several studies have examined excess risk at specific sites. Sandlot and colleagues (Sandlot, Everson, and Wilcox 1985a; 1985b; Sandlot, Wilcox, and Everson 1985) conducted a case-control study on the effects of exposures to environmental tobacco smoke during childhood and adulthood on the risk of cancer. The cases included cancers of all types other than usual forms of skin cancer. For all sites combined, a statistically significant increase in risk was found for exposure to smoking by a parent (crude relative risk = 1.6) and 74

Draft - Do not cite or quotl by a spouse (crude relative risk - 1.5); the effects of ~hese ~wo so~rces Of exposure were independent ISandler, Wilcox, and Everson 1985). Statistically significant associations were also found for same individual sites. These provocative findings will require replication in additional studies. In a case-control study, such as reported by Sandler and colleag~es, biased information on exposure to environmental tobacco smoke is of particular concern. In the cohort study in Japan, Hirayama (1984) found significantly inoreesed mortality from nasal sinus cancers and from br~in tumors in nonsmoking women married to smokers. In a case-control study of bladder cancer, involuntary smoking at home and at work did not increase risk (Karat et el. 1986). cervical cancer, which has been linked to active smoking, was associatld with duration of involuntary smoking in a case-control study in Utah (Slattery et al. 1989) This unconfirmed finding needs additional investigation. These associations of involuntary smoking with cancer at diverse sites other than the lung cannob be readily supported with arguments for biological plausibility based on evidence from active smokers. Increased risks at so~e of the sites, e.g., cancer of the nasal sinus and female breast cancer, h~ve not been found in active • mokers (U.S. DHHS 1982). In fact, the International Agency for Research on Cancer (WHO 1986) has concluded that effects would not be produced in involuntary smokers that would not be produced to a larger extentin active smokers. SUMMARY I. For exposure to environmental tobacco smoke and lung cancer, the evidence has been considered by scientists convened b7 the International Agency for Research on Cancer of the World Health Organization, the National Research Council, and the U.S. Surgeon General. All three groups concluded that environmental tobacco smoke causes lung cancer among nonsmokers. 2. Further research in involuntary smoking and lung cancer will refine our understanding and are scientifically necessary~ however, existing scientific concl~slon$ already provide a compelling rationale for reducing involuntary exposure to environmental tebacco smoke. Akiba S, Kato H, Blot WJ. Passive smoking and lung cancer among Japanese women. Cancer Res 1986; 46:4804-7. Brownson RC, Reif JS, Keefe TJ, Ferguson SW, Pritzl JA. Risk factors for adenocarcinoma of the lung. Am J Epidemiol 1987: 125:25-34. 75

Draf~ - Do no~ ¢ita or ~ote Butler C, Samet J, Humble CG, Sweeney RS. The histopathology of lung cancer in New Mexico, 1970-1972 and 1980-1951. J Natl Cancer inst 1997; 78:85-90. Chan WC, Colhourne MJ, Fung SC, HO HC. Bronchial cancer in Hong Kong 1976-1977. Br ~ Cancer 1979; 39:192-192. Chan WC, Fung SC. Lung Cancer in nonsmokers in Hong Kong. In: GrUndmann E, ed. Cancer Campaign. Vol. 6, Cancer Epideni~iogy. Stuttgart: Gustave Fischer Verlag, 1989; 6:199-202. Correa P, Fickle LW, ;ontham E, Lin ¥, Haenszel W. Passive smoking and lung cancer. Lancet 1983; 2:595-7. Dalager NA, Pickle LW, Mason TJ, Correa P, Fontham E, Stemhagen A, Buffler PA, Zieg!er RG, Fraumeni JF Jr. The relation of passive smoking to lung cancer. Cancer Res 1985; 45:4808-11. Doll R, Pete R. Cigarette smoking and bronchial carcinoma: Dose and time relationships among regular smokers and lifelong non- smokers. ~ Epidemiol Community Health 1978; 32:309-13. Enstrom JZ. Rising lung cancer mortality among non-smokers. J Nail Cancer Inst 1979; 62:755-60. Gao YT, Blot WJ, Zheng W, Ershow AG, HSU CW, Lavin LI, Zhang R, Fraumeni JF Jr. Lung cancsr among Chinese women. Int J Cancer 1987; 40:504-9. Garfinkel L. Time trends in lung cancer mortality among nonsmokers and a note on passive smoking. J Natl Cancer Inst 1981: 66:1061- 6. Garfinkel L, Auerbach O, $oubert L. Involuntary smoking and lung cancer: a case-control study. J Natl Cancer Inst 1995; 75:463- 9, Geng GY, Liang ZH, Zhang AY, Wu GL. smoking and female lung cancer. In: S, eds. Smoking and Health 1987. 1988; 483-6. on the relationship between Aoki M, Nisamichi S, Teminaga Amsterdam: Excerpta Medi=a, Glllis CR, Hole DJ, Hawthorne VM, Boyle P. The effect of environmental tobacco smoke in two urban communities in the west of scotland. Eur J Respir Dis 1984; 65(suppl 133): 121-6. Grufferman S, Wang HH, DeLong ER, Ximm SY, Delzell ES, Falletta JM. Environmental factors in the etiology of rhabdomyosarcoma in childhood. J Natl Cancer ~nst 1982; 68:107-13. Hirayama T. Non-smoking wives of heavy smokers have a higher risk 76

Draft - Do not cite or quote of lung cancer: a study from Japan. Br Med J 1981¢ 282:199- 9. Hirayama T. Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan. Prey Med 1984; 18:686-90. Humble CG, Samet JM, Pathak DR. Marriage to a smoker and lung cancer risk in New Mexico. Am J Public Health 1987; 71:598-602. ~noue R, Hirayama T. Passive smoking and lung cancer in women. In: Aoki M, Hisamiohi S, Tominaga S, eds. Smoking and Health 1987. Amsterdam: Excerpta Medica, 1988; 283-5. Kabat GC, Wydner EL. L~ng cancer in nonsmokers. Cancer 1984; 53:1214-21. Habat GC, Dieck GS, Wynder EL. Bladder cancer in nonsmokers. Cancer 1985; 2:362-7. Mno~h A, Bohn H, Schmidt F. P~ssivrauchen als lung enkrehsursache ~ei nichtraucherianen. Med Klin 1983; 2:86-9. KOO LC, HO JH, Lee H. An analysis of some risk factors for lung cancer in Hong Hong. Int J Cancer 1985; 35:149-85. Moo LC, HO JH, Saw D. Is passive smoking an added risk factor for lung cancer in Chinese women? J Exp Clin Cancer Res 1984; 3:277- 8~. Kco LC, Mo $H, Saw D, Ho C. Measurements of passive smoking and estimates of lung cancer risk among nonsmoking Chinese females. Int J Cancar 1987; 39:162-9. La~'TH, Hung ITM, Wang CM, Lam WE, Hleevens JWL, saw D, Hsu C, Seneviratne S, Lam $Y, LC KE, Chan WC. Smoking, passive smoking and histological types in lung cancer in Hong Nong Chinese women. Br J Cancer 1987; 56:673-9. Lee PN, Chamberlain J, Alderson MR. Relationship of passive smoking to risk of lung cancer and other smoklng-associated diseases. Br J Cancer 1986; 54:97-105. Miller GN. Cancer, passive smoking and wives. West J Med 1984; 140:632-5. nonemployed and employed National Research Council, Committee on Passive Smoking. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. Washington, D.C.: National Academy Press, 1986. Pathak DR, Samet JM, Humble CG, Skipper BJ. Determinants of lung 77

D~aft - DO not cita or ~o~e cancer risk in oigarette smokers in New Mexico. J Nail Cancer Inst 1998~ 76:5S7-604, Pershagen G, Hruheo Z, Svensson C. Passive smoking and lung cancer in Swedish women. Am J Epidemiol 1997; 125:17-24. Preson-Martin S, YU MC, Benton B, Henderson BE. N-nitroso compounds and childhood brain t~mors: a case-control study. Cancer Res 1982: 42:5240-5. Repace, JL, Lowrey AH. A quantitative estimate of nonsmokers' lung cancer risk from passive smoking. Environ Int 1985; 11:3-22. Repace, JL, Lowrey AM. A rebuttal to criticism of the phenomenological model of nonsmokers' lung cancer risk from passive smoking. Environ. Carcino. Revs. IJ. Environ. Sci. Health) 1986; C4(2) :225-35. Repace, JL, Lowrey AM. Risk assessment methodologies for passive- smoking-induced lung ~ancer. Risk Analyis 1990; 10:27-27. Rothman KJ, Modern Epidemiology. Boston, Little, Brown and Company, 1988, pp 7-21. Samet JM, Wiggins CL, Humble CG, Pathak DR. Cigarette smoking and lung cancer in New Mexico. ~ Rmv Respir Dis 1988; 137:1110-13. Sandier DP, Wilcox AJ, Everso~ RB. Cumulative effects of lifetime passive smoking on cancer risk. Lancet 1985a; I: 312-15. Sandlot DP, Everson RB, Wilcox AJ, Browder JP. Cancer risk in adulthood from early llfe exposure to parents' smoking. Am J Public Health 1985b; 75: 487-92. Sandler DP, Everscn RB, Wilcox AJ. Passive smoking in adulthood and cancer risk. Am J Epidemiol 1995a; 121:37-48. Shimizu H, Morishita M, Mizuno K, Masuda T, Ogura Y, Santo M, et el. A case-control study of lung cancer in nonsmoking women. Tokohu J Exp Med 1999; 154:889-97. Slattery ML, Rohison LM, Schuman KL, French TK, Abbott TM, Overall JC, Gardner JW. Cigarette smoking and exposure to passive smoking are risk factors for cervical cancer. JAMA 1989; 261: 1593-98. Trichopoulos D, Kalandidi A, Sparros L. Lung cancer and passive smoking: conclusion of Greek study (Letter). Lancet 1993; 2:677- 9. Trichopoulos D, Kalandidi A, Sparros L, MacMahon B. Lung cancer and passive smoking. Int J Cancer 1981; 27:1-4. 78

Draft - Do not cite cr quota United States Department of Health and Human SerVices, Public Health SerVice. The health consequences of smoking. A report of the Surgeon General. Washington, D.C.: U.S. Government Printing Office, 1982. DHHS (PHS) publication no. 82-50179. United States Department of Health and Human SerVices, Public Health Service. The Health consequences of involuntary smoking. Washington, D.C.: U.S. Government Printing Office, 1986. DHHS (PHS) publication no. (CDC) 87-8398. Unitmd States Public Health Service. Smoking and health. Report of the Advisory Committee to the Surgeon General. Washington D.C.: U.S. Government Printing Office, 1964. PHS publication no. 1103. Wells AJ. An estimate of adult mortality in the United States from passive smoking. Environ Int 1988; 14:249-65. Weiss ST. Passive smoking and lung cancmr. What is the risk? (editorial). Am Rev Hesplr Dis 1986, 133:1-3. World Health Organization. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Tobacco smoking, Vol. 38. Lyon, Prance, World Health Organization, IARC, 1986. Wu AH, Henderson BE, Pike MC, ¥u MC. Smoking and other risk factors for lung cancer in women. J Nail Cancer Inst 1985; 4:747- 51. 79

Draft - Do no~ cite or quote P~GURES ~ND TABLES. C~APT~R S

Draf~ - DO not cite cr quot~ TABLZ 1 Cchoz~ Studies of Inveluntary Smoking and Lung Cancer Study "~ Findings Comments 91,540 nonsmoking females, 1966-1981, ~apan (Hirayama 1881). 176,139 nonsmoklng females, 1960-1972, U.S. (Garfinkel 1881). 8,128 males and females, 1972-1982, Scotland (Gillls et el. 1984). Age-o~cupation adjust- ed RR by husbands' smoking: Nonsmokers - 1.0a÷ Exsmokers - 1.36 ~rrent smoklr8 < 20/day - 1.45 20/day - l. Sl Age-adjusted RR hy husbands' smoki~g~ Nonsmokers - 1.00 ~u~rent S~okers ~ < 20/day - 1.27 ~ 20/day - 1.10 Age-adjusted i~ for ex'posure to a tobacco smoker in the home: Males - 3.25 ~e~ales - 1.00 Trend statistically significant. All histological types of lung cancer. All histologies. Zffec~ of husbands' smoking not s~at- Istically signifi- cant. Preliminary, small numbers of cases. *P-R - relative risk, as estimated by the ra~io of observed to expected number of lung cancer deaths. ÷referdnce category, risk arbitrarily se~ to unity as the reference point f3r comparison.

TABLE 2 Draft - Do no= cite or quote Case-control Studies of Involuntary Smoking and Lung cancer Study Findings comments 40 nonsmoklng female cesest 149 controls, 1978-1980, Greece (Trichopoulos at el. 1991). 84 female cases and 139 controls, 1976- 1977, Hong Kong (¢han st al. 1979~ Chart and Tung 1982). 2~ female and $ =ale nonsmoking casms, 133 female and 180 male controls, U.S° (¢~rrea SC el° 1953). 19 male and 94 female nonsmokihg cosset and 110 male and 270 female non- smoking controls, Japan (A~lba et el. 1996). 99 nonsmoking cases a~d ?36 controls, Louisiana, Texas, ~ew Jersey (Dalager et al 19~6). 28 nonsmoking controls, New Mexico (Humble e~ el. 1987). 77 nonsmoklnq caeese 2 matched control series, Sweden (pershagen st el. 1987). RR* by husband ~oking: Nonsmoklro - 1°0 Exsmokers- 1*8 Currant smokers < 20/day - 2.4 E 20/day - 3.4 RR of 0.7S associated vi~h emoklng spouse, ¢o~pared ~ i.0 foe a nonsmoking spouse, RR by spouse smokinq: t ~ons~okere - 2.00 < 40 pack years - 1.48 41 pack years - 3.11 For females, RR of 1.5 if husband smoked; for males, RR Of 2.9 if wife smoked. RR for marriage ~o a smoking spouse was 1.5 RR for =arriege to S smoking spouse was 9.2 NO effect £n a~cive slokd~Ts° RR for marriage to a smoker was 3.3 for squa:oue small cell caEcinolas. Trend etatlstlcally slgniflcant. His- tol~Ise other ~han adenocsr¢Inoma and hronchloloalveolar ceroln~a. All hletologlos. Two re o~s are inconsistent on the exposu:s variable. Signiflcant increase for ~ 41 pack years. Bronchloloalveolar carcinoma excluded. Cllnlcsl or radio- logical diagnosis for 43t. All types of lung cancer. Nearly i00% hlstc- logical confi~a- tlon. All types of l~u1q cancer° All types other than bronchlolo- alveolar carcinoma. No effect of expo- sure for other types. Study based within a cohort.

(ocmlinu~) ell - Do not cite or quota T~Z~ Case-control S~udies of Znvol~a~ Smoking ~d Lung Cancer S=udy_ 102 adenoca~cinoma =sees, 50 ~ales and re=ales, and 131 controls, Colorado (Brovnson e~ al. 1987). 25 malt and 53 female nonsmoMing cases with matched controls, 1971-1980, U,S. (Karat and Wynder 1984). 8S~nonsmokin~ female oa~ee~ 1981-1982,- Hong Kong (Xoo et el., 1984, and 1985). 31 normmoking and 139 smoking female cases, U.S. (wu et el. 1965). 134 nons=oklng female cases,.U.S. (Garfinkal et el. 1965). 15 male and 32 female nonamoktng CaSeS, a~d 90 male and 66 female non- smoklnq controls. O England (Lee etal., 1986). Findings No effe~ ~ entire group. In nonamokAng women, P.R Of l.? for exposure E 4 hrs/da~, versus 1.0 for <3 hzs/day. • R no~ s~iflcantly Increased tot ~u~Trent exposure at home: Mallu| - 1.26 Females - 0.92 RR of 1,24 (not stat- IsticalIy elgnlf~oant) for combined home and vork~laoe exposure vet- sue 1.0 fo~ mon~xposqd. No association wASh cumulative hours of exposure. No si~ificant effects of exposure from par- ante, spouse, Or work- p~ace in alokara and nonsmokers, Nonsignificant RR o~ do~emts Zmvmluntary smoking effeot not slgnlfl- cant in homemaking women, hut only 19 8umh cease included. All ty~ee, Findings negative for spouse smoking variable as ~oll. All types of lung cancer. Adenooarcinoma and s~uamous oell carci- noma only. All ty~es of lung 1.22 if husband Smoked, Significantly increased RR of 2011 if husband smoked 20 or more oiq- arettes daily at home. Sig~Ifioan~ trend of RR wi~ number of cigarettes smoked at home by the husband. Overall RR for spouse Hospital-based smoking of I.i. study. oanceE. C&ref~l ex- ~lualon Of em~keEs from the oase group. *RR - relative risk as estimated by the odds ratio.

Pratt - Do noc cite o~ quote TABLE 3 Conclusions of the World Health Organiza~ton, National Research Council and U.B° Surgeon General on Involun~azT Snokinq and Lung ~ance~ World Health O~anizatio~ "Knowledge of ~he nature of aldsa~raam and malns~raam smoke, of the materials absorbed during "passLve" smmklngs and of the quantitative relationships between dose and,effect ~hat are commonly observed from e~sure to carcinogens, however, leads to the conclusion that pasalveaoklng gives rlms to some rlak Of cancar~w National Research Council "The weight of svlder~e derlved;from apldemlol~i¢ studles shows an association between ET$ a~aura of nonsmokerl and lung cancer that, taken as s whole, is unllkel¥ to ~ due to chance¸ on syltmti¢ bias. The observed estlmata mf increased risk is 34%, largily for spouses of mkare compared with spouses of ~ns~okerSo" U.S. Surgeon General "Involuntary smoking can cause lung cancer in nonsmokers." "The absence of a ~hrRshold for respiratory carcinogenesis in active smoking, ~hs presence Of ~he same carcin~ens in mainstream and sldestrmam smoke, the demons~TaEed uptake of to~¢co ammke constituents by involuntax-f a~kers, and "¢h6 Gemonstration of an increased lung cancer risk in iomm populations wi~ exposures to ETS leads to the conclusion ~hat involuntary smoking is a cause of lung cancer."

D~aft - DO not ci~s or ~ote PABBZV2 BMOXZNG AND S:E;4P.T DZBEXSE: EPZDEMIO?.OGYs Fa~'BZOLOGTw AND BZOCHEXZSTR~[I Bt~ton A. Glaatu PhD W£11£~Wo PaZ~L~S¥ M.D. P~visLon of Oard~oZo~ B~hoo1 of Medioins nnivs~sit~ of Califorale Ban Frnsisoo, CA 94143 The first disease linked to active smoking was lung canosr. IZ is, therefore, not surprising that ~he first disease linked to passive smoking was also lung cancer (USPHS, 1886). Before the advent of mass marketed cigarettes, lung cancRr was a rare disease. The fact that smoking is the major identifiable cause of lung cancer made identifying this llnk -- for both active and passive smoking -- relatively straightforward. This situation contrasts with heal-~ disease, which has many risk factors, so it is not surprising that it took longer for the scientific community to conclude that active smoking caused heart disease (USPHS, 1983). Once the llnk between smoking and heart disease was established, it became clear that smoking a~ounted for mors heart disease deaths than lung (and other) cancers because of the high prevalence of heart dissasL. Similarly, smoking is the most important preventable cause of coronary disease. Given this history, it is not surprising that exposure to environmental tobacco smoke (ETS) has now been linked to heart disease in nonsmokers (Wells, 1988; Kristensen, 1989) and may result in a substantial number of unnecessary coronary hea~ disease deaths in nonsmokers. Most of the evidence linking ETS and coronary heart disease has appeared since the US Burgeon Go,oral (USPHS, 1986) and National Academy of Sciences (NRC, 1956) last reviewed ~he evidence on ths health effects of ETS. Based on !the information available as of early 1986, both these reports concluded that the svidsnce linking ETS and heart disease was equlvooal and that more research was • necessary before any definitive statements could be made. These conclusions were reasonable at the tlml they were made. Zn the four years since these reports were written, considerable information on both the epldemlology and bloloqical mechanisms h7 which ETS may cause heart disease has accuzulated from several areas of scientific investigation. In fact, most of the results IThis chapter is an adaptation of a peer-revlewed manuscript of the sams tltle (Glantz and Parley, 1990). 8O

D~af~ - Do not olte or quote ~PTIR 6 PASSIVE BMOXZNG AND HEART DZSZXSZ: BPIDSMIOLOGY, PHYSIOLOGY, AND SZO~STRTI Btantom A. GlSnt| ~hD Wllllaa W. Pa:aley M.D. Division o~ cerd4olocF/ S~hool of Medictine DAivs~s~ty of ~allfol'n~a Ben FESnOlsoo~ CA 945.43 The first disease linked to active smoking was lung cancer. It is, therefore, not surprising that the first disease linked to passive smoking was also lung cancer (USPHS, 1986). Before the advent of mass marketed cigarettes, Iug cancer was a rare dlsess@. The fact that smoking is the major identifiable cause of lung cancer made identifying this link -- for both active and passive smoking -- relatively straightforward. This situation contrasts with hear~ disease, which has many risk factors, so it is not surprising that it took longer fop the scientific community to conclude that active smoking caused heart disease (UBPHS, 1983). Once the link between smoking and heart disease was established, it became clear that smoking a~ountsd for more hsa~ disease deaths than lung (and other) cancers because of the high prevalence of heart diseasL. Similarly, smoking is the most important preventable cause of coronary disease. Given this blstory, iris not surprising that exposure to environmental tobacco smoke (ETS) has now been linked to heart disease in nonsmokers (Wells, 1988; Kristensen, 1989) and may result in a substantial number of unnecessary coronary heart disease deaths in nonsmokers. Most of the evidence linking ETS and coronary heart disease has appeared since the US Surgeon General (USPHS, 1986) and National Academy of Sciences (NRC, 1986) last reviewed the evidence on the health effeats of ETS. Based on the information available as of early 1986, both these reports concluded that the evidence llnklng BTS and heart disease was equivocal and that acre research was • necessary before any definitive statements could be made. These conclusions were reasonable at the tlml they were made. In the four years since these reports were ~Itten, considerable information on both the epldemlology and bioloqlcal mechanisms by which ETS may cause heart disease has ac~nlmulated from several areas of scientific investigation. In fact, most of the results IThls chapter is an adaptation of a peer-revlewed manuscript of the same title (Glantz and Parmley, 1990). 8O

Draf~ - Do not cite or ~uote presented in this chapter were published after the 1986 Surgeon General and National Academy of Science rlpor~s. First, there are now Ii epidemiological studies on ~he relationship bef~een exposure to environmental tobacco smoke in the home and the risk of heart disease in the nonsmoking spouse of a smoker. All but one of these studies yielded a relative risk • greater than 1.0. There are several lines of biologic evidence which mak~ this association plausible. There is evidence that exposure to ETS reduces exercise tolerance of healthy individuals as well as people with existing coronary artery disease. Such reduced exercise capability is one of the landmarks of acute compromises to the coronary circulation. There is evidence, from both human and animal studies, that exposure to tobacco smoke, including passive smoking, increases aggregation of blood platelets. Such increases in platelet aggregation are a~ important stmp in the genesis of athercsclerosis. In addition, increasing platelet aggregation contributes to coronary thrombosis, the cause of acute myocardial infarction. Finally, carcinogenic agents in ETS, including benzo[a]pyrene have been shown to produce injuries to the endothelial cells which line arteries. Such injuries are the first step in the development of atherosclerosis. Thus, &xposure to ETS can contribute to both short term and long term insults to the coronary circulation and the heart. Effects of Primary Smokin= Before reviewing the evidence linking ~TS with coronary artery disease, it is worth summarizing the evidence linking active smoking with coronary artery disease. This evidence was summarized in the 1983 Surgeon Generalts Report, which was devoted entirely to cardiovascular disease (USPHS, 1983); it concluded: In 1980, diseases of the circulatory system were responsible for approximately one-half of the total U.S. mortality. CHD was the si~le host important cause of death, accounting for approximately 30 percent of all U.S. deaths. Cigarette smoking is one of the three major independent CND risk factors. The magnitude of the risk associated with cigarette smoking is similar to that associated with the other two major CND risk factors, hypertension and hypercholesterolemia~ however, because cigarette smoking is present in a larger percentage of the U.S. population than either hypertension or hypercholesterolemia, cigarette smoking ranks as the largest preventable cause of CHD in the United States. Cigarette smoking also acts synergistically with the other major risk fautors to greatly increase the risk for CHD. Arteriosclerosis is the predominant underlying cause of cardiovascular disease, and atherosclerosis is the form of 81

D~af~ - DO Not olte or quote arteriosclerosis that most frequently causes clinically significant disease, including CHDe atherothrombic brain infarction, athero$clerotic aortic dlsease, and atherosclerctlc peripheral vascular disease. Cigarette smoking contributes bo~h to the development of atherosclerotic lesion~ and to the clinical manifestations of a~herosclerotlo vascular dIsoasi# including sudden death. Although the precise pa~hophys~l~icbasis of these clinloal • manifestations is not understood, it may be rsZated to several deleterlous cardiovalcular effects of olgaratte smoking, including pr~uction of an imbalance between myocardial oxygen supply and demand, a decrees, in threshold for ventricular fibrillation, and an Increase in plahaloh aggregation. Nicotine end caEbon monoxide are the tobacco smoke constituents most closelyasaoclated with these adverse effects; other cigarette smoke constituents such as hydrogen cyanide, oxides of nitrogen, and carbon disulfide are being studied for possible pathoZoglc cardiovascular affec~s. Since 1983, evidence has also mounted that the polycyclic aromatic hydrocarbons in cigarette smoke can injure the arterial endothelium and initiate the atherosclerotic process. All the compounds implicated as damaging to the cardiovascular system of smokers have been Identiflad in ETS (USPHS, 1986; NRC, 1986}. EDidemiolcoicsl StudleB "on ETS anf Mean Disease Since 1984, the epidemlologlcal evldence linking exposure to ETS with heartdlsease has rapidly accumulated. The results oft he eleven published studies are summarized in Table I and Figure ~i; four studies present date On men, nine OD women, add one on both sexes combined. Despite minor differences in methodology or end points (some used death from ischemlc hear~ disease of any origin and some were limited to death from myocardial infarction), the results of these studies are remarkably consistent. All the studies on men yielded relative risks of death from heart diseasm exceeding i for ncnsmcking men married to smokers, with a median risk of 1.2. All but one of the studies on women (Lee at el, 1986) yielded ~elative risks exceeding I, with a median relative risker • 1.4. Several studies allo suggested an increase in ~he risk of nonfatal coronary symptoms (Svendeen et ~i, 1987; Palmer et el, 1988~ Hole et el, 1989; Dobson et el, 1990) t quantitative results in Table i ¸only reflect risk of death, not coronary symptoms. Consistency of an observation acrols different studies increases the confidence one can have in the belief that an association is causal, unless all studies have the same bias. When interpreting the results of such epidemiologlcal studies, it is always important to consider biological plausibility and potentiml confounding variables which could explain the resul~s. .82

Drat~ - Do no~ cite Or quota ar~eriosc~erosis tha~ ~o$~ frequently causes clinically significant ~isease, inoluding CHD, atherotbrom~£c brain infarc~ion, atherosclerotl¢ aortic disease, an~ atherosclerotic peripheral vascular disease. Cigare~e smoking contributes bo~h ~o ~he development of atherosclerotlc lmaions and ~o the o11nical manifestations of a~heros~lerotic vascular ~lseas~, £nolu~in~ sudden death. Although t~e p~e~ise pa~hophys~ol~ia basis Of ~heae Clinical • manifes~a~ions il not undershot, it may be relatQ~ ~ several dele~erious cardiovascular effec~ of cl~arQt~e smok±~, lncludtn~ pro~uc~ion of an in~alance between myocardial oxygen supply and demand, a die,ease in ~hreshold for ven~ricular fibrillation, an~ an increase in platele~ agg~ega~i~n. Nicotine and carbon monoxide a~e the ~obac¢o smoke constituents mos~ closel~ asso¢ia~ed wi~h these adverse effects: o~her cigarette smoke cons~i~uents such as hydrogen cyanide, oxides of n~t~c~en, ar~ ca~on dilulfide a~e ~ein~ studied for possible pa~holo~l~ cardiovascular effe¢%$. Since 1983, evidence has also ~oun~ed~ha~ ~he polycycllc aromatic hydrocarbons in cigarette smoke can injure the arterial endo~heliu~ and initiate ~he a~herosclero~ic pro~ess. A11 the compounds implicated am damaging ~ ~he cardiovascular sys~e~ of snoke~s have been l~en~lfie~ in ET$ (USPHS, 1986; NRC, 1986). ~oidemiolo~ical S~udi~B "~n ~S and Heart Disease Since 1984, the epi~emiol~ical evidence linking exposure ~ ETS with hear~ disease has rapidly acoumula~ed. The ~esul~s of ~h~ ~leven pu~lishe~ $~udiem are summarized in Table 1 and ?tgurel; four $~udies presen~ da~a on ~e~, nl~e On womQ~ and one On bo~h $~xes co~binad. Despi~e minor differenoe$ in methodology ~r end points (some used death from i~chemic hear~ ~isease of any origin and so=e were limited to death from myooardial infarction), ~hl results of these s~udies are remarkably consistent. All ~he studias on men yleldedrela~ive risks of death from heart disease exceeding 1 for nonsm~kin~ men married ~o smokers, wi~h a median risk of 1.2. All b~t one of the s~udies on women (Lee Qt al, 1986) yielded relative risks exce~ding 1, wi~h a median relative riskof • 1.4. Several studies also suggested an increase in ~he ~isk of nonfa~aI coronary s~p~oms (svendsen e~ ~1, 1987~ Palmer et al, 1~88; Holl e~ al, 1989; Dobson e~ al, 1~90)~ quan~i~a~ive results in Table 1 only reflec~ ~lsk of ~eath, no~ ¢or~na~-~ E~ptou° C~nsis~ency of an observation across different studies incr~ases the oonffdence one can bare in ~he belief that an association is causal, u~less all s~die8 have ~e same biaB. When interpreting ~he results 0f such epidemiological studies, it is always impo~c~n~ ~o consider biolo~ical plausibill~ and pc~en~ial confounding variables which could explain ~he re~u!~s. 82

Draft - Do not cite or Aside fro~ noting that the =ompounds in mainstream smoke that have been implicated in heart disease are in ETS, we will defer the discussion of biological plausibility until later in this chapter, when we discuss the effects of ETS on platelets and the atherogenic agents in ETS. ~ For now, we will concentrate on potential confounding variables. These are particularly important in a disease llke heart disease, because it is known to be caused by multiple risk factors. All of the studies controlled for the most important confounding variable, age, and several (Garland etal, 1985; Svedsen et el, 1987; He, 1989; Hole et el, 1989; Humble et el, 1990) controlled for several known risk factors for coronary artery disease, in particular levels of cholesterol, blood pressure and weight (or body mass or body mass index). Most of the studies also included one or more measures of socioeconomic status, such as the nature of the housing or amount of education. Lee (1988, 1989, 1990) has suggested that the elevated risk of heart (and other) disease with passive smoking could be due to misclassificaticn of nonsmokers who are really smokers. In • ddition, Wald (1986) has noted that some people who say they live N~th nonsmokers have detectable levels of the nicotine metabolite cotinine in their blood, indicating that ~hey are actually exposed to ETS, either at work or at home. The former type of misclassification will tend to lead to an overestimate of the risks associated with ETS and the letter will lead to an underestimate of the risk. Careful analysis of the question of misclasslfication -- which applies generally to studies of ETS and not just heart disease -- have demonstrated that the observed risks cannot be 6xplained by this technical problem (Weld, 1996; Wells, 1986, 1988, 1990; Kawaohi End Pearce, 1989; Reinken, 1989). In addition, both the Surgeon General (USPHS, 1986) and the National Academy of Sciences (NRC, 1986) were presented with the argument that misclassificaticn errors accounted for the link between ETS and lung cancer and concluded that ETS caused lung cancer in healthy nonsmokers. To date, no compelling case has been made that this technical error explains consistent findings linking ETS with heart (or lung) disease. Indeed, the net effect of these two types of misclassificaticn errors is to lead to an underestimate of the effects of passive smoking for lung cancer. There is always the pcsslbility that there is some other confounding variable relating to cultural factors, such as the nature of housing or employment or the nature of time spent outside the home. Most studies look only at a crude measure of exposure--spouse smoking--and it is possible that this is an indicator variable for other things, such as poor diet, risky lifestyle, or stress. The fact that results are similar from all ov~r the world in widely varying cultural settings -- including several regions in the United States, the United Kingdom, Japan, and China -- argues against this concern. 93 uote

D~a~ - Do not cite or ¢~ote Several authors also observed a dose-response relationship (Table I) between increasing amounts ot smoking by the spouse and the risk of heart disease in the nonsmoklng spouse (Helslng et el, 1998 (statistically significant in women, but not men); Hole et el, 1989; Garland et el, 1985 (alath.ough not statlstioally slgniEl¢ant); Humble etal, 1990; He, 198g, Hirayama, 1994). The presence of such dose-response effects across multiple studlel, done in different locatlons wi~hdifferent criteria supports the hypothesis that the epidemiolo~y is revealing a real effect of ETS on heart disease in nonsmokers. While all but one of the studles in Table 1 and Figure 1 ylelded relatlve risks greater than 1, the fact remains that 3 of the studies in men and 4 of the studies in women had 95% confidence intervals for the relat$ve risk of passive smoking for heart disease that fell below 1.0, meaning ~hat ~he risk was not statistically significant17 elevated above 1.0 (with P<.05)° It is important to nots that the 95% aonfidence intervals do not lle symnetrically about 1.0, but rather are skewed towards higher risks. To avoid false negative conclusions, Hothman (1978) ~uggested examining the confidence interval, as we have done, in concluding the exposure to ZTS elevates thR risk of heart dimease. One can assess fo~ally how confident one =an be in reaching a negative conclusion by computing the power of the study to detect an effect of specified size (¥risdman et ai, 1978)o Table i shows estimates of the power of each of the studies to detec~ a 20~ increase in risk -of heart disease (i.e., a relative risk of 1.2) with the available samples. The power was computed as described in Huh: and 01shan (1989), using a two-slded test for the relative risk with a Type I risk of 5% (i.m., requiring the 95% confidence interval for the relative risk t~exclude i°0 ¸before concluding a statistically significant elevatlon in risk in an individual study). Host of the studiem have low to moderate power. The two CHelsing st el, 1988; Hole et al~ 1989) that have power above the desirable level of 80% both identified significant ingresses of heart disease risk with ZTS exposure. IntereBtingly~ the mtudy by Lee (1986) which was the only one with a relative rlsk below 1, also had the lowest power to detect an effect, only 3%. It is possible to combine th~ results of these studies in a formal analysis to derive a global estimate Of the relatlve risk and associated 95% confidence Interval. By combining the studies, the sample size and so the power to dete~t an effeat increases° Pooling the studies in Table 1 yields an estimate of the relative risk of death from heart disease of 1.3 (95% Cl 1.1-1.8) for men and 1.3 (95% cI 1.2-1.4) for wo~n. These resul~s are consistent with those reported by Wells (19~S) who used the studies by Sillis et al (1984), Lee et 81 (1986), and Helsing et 81 (1997) to compute a pooled relative risk of 1.3 (with a ~5~ confidence interval from 1.1 to 1.6) for men and the studies by Hirayama (1984), Gilliset 84

Draft - DO not cite or quo~'.e al (1984), Garland et al (1985), Helalng et al (i~88), Lee st al (1986), and Martin et al (1986) to compute a pooled relative risk of 1.2 (with a 95% confidence Intezwal from i.I to 1.4) for women. Exposure to ETS significantly (p < 0.001) increases the risk of death from hear~ disease in nonsmokers. Finally, it is worth noting that all these studies are based on the smoking habits of the nons~oksr's s~ouse0 and so exposure to ETS at home. Household exposures to ETS at home are generally much smaller than exposures at work, where ~he density of smokers is generally higher (Repace and Low~ey, 1985,1987). AS a result, these studies will generally ~the risk and attendant public health burden due to ETS-Induced heaz~ disease if a substantial proportion of the controls are exposed at work. Kawachl et al (IS89) have adjusted Wells, (1988) relative risks to account for workplace exposures to ETS and found that the relative risks increase to 2.3 (95% CZ 1.4 - :.4) for men and 1.9 (95% Cl 1.4 - 2.5) for women. In addition, Wells (1988) and Kawachi et al (1989~ indicate that the number of heart disease deaths due to passive smoking is an order of magnitude greater than the number of lung cancer deaths due to passive smoking. These epldemlologioal studies demonstrate a connection between ET5 exposure and death from hea~ disease. We now turn our attention to possible physicloglcal and biochemical mechanisms which could explain these observations. Acute Effects of ETS E~eure Chronic exposure to ETS exerts carcinogenic effects by increasing the cumulative risk of a molecule of one of the carcinogens in the ETS damaging the DHA in a call and initiating or promoting the carci~ogenlc process. TO date, no on! has identified any affects of acute exposure to ETS (or, for that matter, any other carcinogen) on cancer. The situation with bea~t disease is different. In heart disease there are both important chronic changes (i.e., the development of stherosclerotlc lesions) and acute changes. The latter include an In¢rease in myocardial oxygen demand which may outstrip the oxygen supply and produce isohemla, and increased platelet aggregation which can lead to coronary thrombosis and acute myocardial infarction. When the coronary circulation cannot provide enough oxygen to the myocardlum to meet the demand, the result is Ischemia which can be silent or result in anginal chest pain. Earlier onset of angina or hypotension during exercise is a reflection of more severe heart disease, oxygen supply can be reduced by athercsclerotic narrowlng or vasoconstriction of the coronaries or by reducing the oxygen carrying capacity of the blood by forming carboxyhemoglobin. Whalfen and Klochkov (1987) confirmed earlier work by Aronow (1978) demonstrating that exposure to ETS significantly reduced exercise ability in patients with coronary artery disease and the rate 85

Draft - Do not cite or pressure product (hea~ rate ti~es systolic blood pressure). In bo~ studies, patients were exposed to realistic levels of~s simply sitting in a waiting rosa while so~eone was smoking. These effects were present in both smokers and nonsmokers (~G~alfen and Kloc~ov, 1987) and regardless of vhether or no~ the room was ventilated {Arcnow, 1978; ~alfenand Klochkov, 1987). Exposure to ~S also increased resting bea~l-ata and systolic and diastolic blood press~e, and resulted in a~l~er heart rate at the onset of angina (~onow, 1978). Blood carboxyhel(Wlobin was increased a~ut 1% after exposure to ~S (~onow, 1978). Sheps et al (1987) found no change in cardtovascular~function in sub~eots wl~h angina in response to mild elevation in blood carbon monoxide similar to ~hat experienced in passive s~okers when ~ey exposed ~heir s~Jects to pure carbon monoxide. Zn contrast, Allred et al (1989) found a significant dose-response relation between carboxyhemoglobin level and the change in the leng~ of time to both electrocardiographic and sylpton manlfeetation in men with angina pectoris exercising after exposure to CO. Even a small increase in the carboxyhemoglohin level, representing a seemingly minor reduction in ~he oxygen-car~ylnq capacity of hemoglobin, was assoclated with the statistically significant effects. Acute exposure to ETS Iiads ~o an imbalance between myocardial oxygen supply and demand during exercise in patients with coronary artery disease. While thls discussion has concentrated on the carbon monoxide An ETS as the active agent, it As likely that some other component of the ETS is also contributing to this effect. The effects of ETS on cardiac performance are, An fact, severe enough to affect exercise perforlance An young healthy subjects with no evidence Of heart disease. McM~rray et al (1985) blindly exposed young healthy women to pure air and air contaminated with ETS while they exercised on s treadmill. The results were similar to those observed in patients with coronary artery disease. Resting heart rate was increased during exposure to ETS, which increased blood carboxThemoglobin by about i~. Exposure to ETS significantly reduced maximum oxygen uptake Cby 0.251~men and time ~o exhaustion (by 2.1 man). Exposure to ETS also increased the perceived level of exertion during exercise, maximum heart rate, and COt output. It also significantly increased levels of lactate in venous blood (from a mean of 5.5 mM during control period to 6.8 mM after exposure to ETS). This greater lactate at a lower Oxygen consumption during the passive smoking trials indicates a greater reliance on anaerobic metabolism. The combined effec~ of the reduced oxygen oarrying capacity and increased lactate resul~ed An a reduction in maximal aerobic powe~ and "the duration of exercise. Thus, even in heal~hy subjects, exposure to ETS adversely affects exercise perfo~ance. The acute effects of CO and direct tobacco smoke on exercise performance are well doCUmented in the lltarature. Exposure to CO (or CO in tobacco smoke), and the subse~ent elevation of blood carboxyhemoglcbin levels to ca. 3%~ has been shown to decrease q~otl 86

D~af~ - Do not cite or quote pressure product (hea~ rate tizes systolic blood pressure). In bo~h studies, patients were exposed to realis~ic levels of¸ ETS by simply sitting in a waiting room vhile someone was smoking. These effects were present in ~=th smokers and nonsmokers (~alfen and Klochk~v, 1987) and regardless of whether or not~ ~e room was ventilated (A=onow, 1978; ~alfen and KlOChkov, 1987). EXposure to ZT$ also increased restinqheaz~ rata and symbolic and diastolic blood pressu=e, and resulted in a lower hea~ ~ate at the m~aet of a~gina [~ronow, 1978). Blood ¢arboxyhe:Oqlobin yes increased by about 1% after exposure to ~S (~onow, 1978). Shape et al C1987) found no change in oardiovascular~functlon in subjects with angina in response to mild eleva~i=n in blood carbon monoxide s~ilar to that experienced in passive sackers when they exposed their subjects to pure carhon aonoxide. Zn contralt, Allrsd et al C1989) found a sig~ificant dose-response relation between carbox~bemoglobin level and the change in the !enOch of ~ime to both electrocardiographic and s~ptom manifestation in men with angina pectoris exercising after exposure to CO. Even a small increase in the uarboxyhemoglobln level, representing a seemingly minor reduction in the oxygen-carrying capacity of hemoglobin, was associated with the statistically si~ifioant effects. Acute ~xposure to ETS leads to an imbalance between myocardial oxygen ~upply and demand during exercise in patients wi~h coronary artery disease. While this discussion has concentrated on the carbon monoxide in ETS as the active agent, it is llkeiy that some other component of the ETS is also contributing to this effect. The effects of ETS on cardiac perfor~Qance are, in fact, severe enough to affect exercise performance in young healthy subJeots with no evidence of heart disease. MCMUrray et al (1985) bllndly exposed young healthy women to pure air and air contaminated with ETa while they exercised on a treadmill. The results were similar to those observed in patients wlth coronary artery disease. Resting heart rate was increased during exposure to ETS, which increased blood carhoxyhem~Iobln by about i%. Exposure to ETS significantly reduced maximum oxygen uptake (by 0°2$1/min and time to exhaustion (by 2.1 min). Exposure to ETS also increased the perceived level of exertion during exercise, msxi~um heart rate, and CO2 output° It also significantly increased levels of lactate in venous blood (from a mean of 5.5 ~ during control period to 6.8 mM after exposure to ETS). This greater lactate at a lower oxygen consumption during the passive smoking trials indicates a "greater reliance on anaerobic metabolism, The combined effect of the reduced oxygen carrying oapaoltyand increased lactate resulted in a reduction in maximal aerobic power and the duration 9f exercise. Thus, even in healthy sub~ects, +exposure to ETs adversely affects exercise performance° The acute effects of CO and di~eot tobacco smoke on exercise performance are well documented in the literature. Exposure to CO (or CO in tobacco smoke), and the subsequent elevation of blood carboxyhemoglobin levels to ~. 3t, has been shown to decrease 8~

Draft - Do not site or ~ots exercise duration in patients wi~h ischemic heaz~c disease and decrease short-term maximal exercise duration in young healthy men. ~t is conceivable, ~herefore, that elevations i~ COHb due to ETS could have similar effects. While the association between active smoking and cardiovascular disease is well known (USPHS, 1983)~ little is known about the relative importance of each component of tobacco smoke that may he responsible for this relationship. Most experts agree, however, that both CO and nicotine are important, and other constituents of the smoke may play a role as well. Active smoking clearly aggravates the decrease in 02 capacity induced by CO through an increase in the 02 demand of the heart (Deanfield et el, 1986). Passive smoking exposes an individual to all components in the cigarette smoke, but the CO component dominates heavily because only 1% or less of the nicotine is absorbed from passive smoking compared to 100% in an a~tive smoker (Wall et el, 1988; Jarvis, 1987). Currently available information indicates that acute exposure (I to 2 h) to passive smoke will increase a nonsmoker's COHb level by about 1% (Jarvis, 1987). This small incremental increase in COHb due to ETS alone may not be enough to trigger acute cardiovascular effects unless combined with other sources of CO or with other components of ETS having a s.imilar effect (e.g., nicotine). Lamb (1984) has suggested that at maximal exertion levels, up to 90% of the oxygen carrying capacity of the blood may be needed. Because of the carbon monoxide, and perhaps other constituents, ETS reduces this capacity, so the muscle cannot maintain its high rate of aerobic metabolism unless cardiac output is further increased; people with heart disease and reduced ventricular reserve have difficulty meeting this demand. In sum, exposure to ZTS increases the demands on the heart during exercise and reduces the capacity of the heart to respond. This imbalance increases the ischemic stress of exercise in patients with existing coronary artery disease and can acutely precipitate symptoms. Moskowitz et al (1990) found evidence that adolescent children of parents who smoked may suffer from chronic tissue hypoxia such as that observed in anemia, chronic pulmonary disease, cyanotic heart disease or high altitude. These children had significantly elevated levels of 2,3-diphosphoglyoerate (DPG), which suggests that the body is attempting to compensate for hypoxia by increasing DPG level in blood to meet tissue oxygen requirements, even after correcting for age, weight, height and sex. These changes were dose-dependentl the greater the exposure to ET$ (measured both in terms of parental smoking and serum thlocyanate in the children), the greater the increase in DPG. There is also evidence that acute exposure to ETS directly affects the myocardial muscle at a cellular level. GvozdJakova et al (1984) exposed rabbits in a 50 liter child's incubator to the smoke of three burning cigarettes smoked over a 30 minute period and measured several variables related to the metabolism of cardiac 87

Draft - ~ not cite or ~ote Jit~ondria. (Mitochondrta &~e the s~cellular elements tha~ control cellular respiration~ ~ey aonve~ o~gen into usable ene~ in the form of ATP.) Theyhad three groups of rabbits: one group exposed to a single dose of ETS, one group exposed to 30 min of ~S twice daily for ~o wee~, and one group exposed to 30 minutes of ~s twice a day for eight wea~. They meas~ed mitochondrlel respiration (QO2) as the consumption of oxygen after addi~ ADP to a vessel containing mltochondrlal fragments. Using pyruvate as a substrata, mitochondrlal respiration QO2was reduced slgniflcantly compared to control (pure air) for all doses of ETS, even a single exposure (Figure 2), to aboUt half the control value. The oxidative phosphorylation rate was also reduced significantly at all exposures by about one-thlrd. There were no significant changes in the coefficient of oxidative phosphorylation (ADP:O2) with ETS exposure. GvozdJakov& et al concluded that pyl~vate as a substrata was a sensitive indicator of the toxic action of the ETS cn the oxidative pr~ess. Later, to further isolate where in the process of mltochondrial respiration, the ETS acted, GvozdJ~k et al (1995, 1987) reported data on succinate, NADH, and cy~oohrom@ oxldase activity in the mltochondria in the four groups of rabbits. Figure 2 shows the results of exposure to ETS on the a~ivity of NADH oxidase, succinate oxldase and oy~ochrome oxldase of myocardial mitochondria. The activity of the first two oxidesos exhibited no changes compared with the control group -- neither after a single exposure to ETS or following exposures up to 2 weeks. Cy~cohrome oxldase activity decreased both after a single exposure to ETS and over time, with increaslng effec~ as the duration of exposure to ETS is extended° The observation that cy~ochrome oxidaee and not NADH or succlnate oxldase activity was affected by ETS suggests that the deleterious effects of ETS on myocardial mltochondrlal respiration occur at the terminal seglu~t of the mitochondrial respiration process. Prolonged exposure to carbon monoxide has been shown in some studies to induce ultrastructuralchanges in mycoardlum (KJeldsen et al, 1974; Thomsen and K~eldsen, 1974; Lough, 1978). Later, KJeldsen and co-workers (HugOd! et el, 1978), using a blind technique and ~he same criteria tO assess morpholoqlcal myo=ardlel damage found no significant changes in the coronary arteries or aorta in normocholesteroleml¢ rabbits exposed to CO at concentrations from 200 to 4000 ppm fo~ up to 12 weeks. They suggested that the positive results obtained earlier were due to the non-blind evaluation technlques end the smell Dumber of animals used in these studies. Later~ Huged (1981) : confirled these negative .results using electron~ microscopy. ID addition, the earlier studies were conducted a~ wmoderate, levels of CO (I00 to 150 ppm) which are considerably higher than levels of CO found in smoke-polluted environments (reported to be as high as 40-S0 ppm, hu~ more typically are around I0 ppm) (NRC, 1986). These negative studies only argue against an effect of CO in inducing coronary 88

Draf~ - Do not cit~ or c;tlote atherosclerosis, and not a direct effect of Co on myocardial oxygen supply and demand (Deanfield et el, 1986). Acute exposure to ETS not only increases the demand and compromises the supply of oxygen to the heart as a whole, but also reduces the myocardium's ability to use this oxygen to create ATP to provide energy to support the heart's pumping activity. This effect probably results from several of the compounds in ETS acting simultaneously on the cardiovascular system. EL~ The action of ETS to increase platelet aggregation is another way in which ETS can acutely increase the risk of a Coronary event. When blood platelets aggregate inappropriately and form a thrombus (blood clot), this clot can form in a fissured plaque in the coronary circulation and precipitate a myocardial infarction. Platmlets are important for the nol-mal body process of hemostasis, to prevent blood loss after an injury. Hemostasis depends on complex interactions among the dynamics of blood flow, components of the vessel wall, blood platelets and plasma proteins. A thrombus can be considered as an inappropriate form of hemostasis and is composed of a mass of cellular material held together by a network of fribrin. Definitive evidence has confirmed that platelets play a major role in thrcm~us formation and embolization, especially in the arterial system. In addition, increasing evidence has shown that platelet deposition and thrombus f~rmation can contribute to the growth and progression of atherosclerotic plaques (Fuster and Chesebr~, 1981; Ross, 1986). An arterial thrombus appears to develop in three phases: platelet adhesion, platelet aggregation, and activating of clotting mechanisms. Passive smoking increases platelet aggregation and so increases the likelihood of thrombus formation and myocardial infarction. Table 2 summarizes the results of three studies (Davis et el, 1885a, 1986, 1989) on the effects of cigarette smoke on platelet ag~egation and damage to the arterial endothelium (lining). (we will discuss the effects on the endothelium below.) Davis et al (1999) also measured platelet aggregate ratios and endothelial cell counts in nonsmokers before and after being exposed to 20 minutes of ETS while sitting in a hospital atrium. Mean values before and after passive smoking were 0.87 and 0.78 (P-.002) for platelet aggregate ratios and 2.8 and 2.7 (P~.OO2) for counts of anuclear endothelial cell carcasses in venous blood. These changes are in between the effects observed after nonsmokers smoked two tobacco cigarettes and the effects observed after smoking two non-tobacco cigarettes (Davis et el, 1985a) and similar to the values observed in nonsmokers who smoked two cigarettes while trying not to inhale (Davis et el, 1986). These effects were not correlated with the level of nicotine in the blood of the experimental subjects in any of these or other (Davis et el, 1985, 1987), related studies on how drugs modify platelet aggregation and endothelial cell counts. In 89

Draft - DO not cite or quote mltochondrla. (Ml~ochondrla are the su~cellular elements tha~ control cellular resplratlon; ~hey conver~ oxygen into usable energy in the for= of ATPo) They had three groups of rabbits: one g~oup exposed to a slngls dose of ETS, one group exposed to 30 mln of ZTS twice daily for ~"~o week, s, and one group exposed to 30 minutes of ETS twice a day for eight week, s. They measured aitochondrlal respiratlon (QOz) as the consumption of oxygen after adding ADP to a vessel containing mitochondrlal fragments, using pyruvate as a substrata, mitochondrial respiration QO2 was reduced significantly compared to control (pure air) for all doses of ETS, even a single exposure (FigUre 2), to about half T,he control value. The oxidative phosphorylation rate was also reduced significantly at all exposures by about one-thlrd. There were no significant changes in the coefficient of oxidative phospborylatlon (ADP:02) wlth ETS exposure. GvozdJakov~ @tal coDcluded that pyl-avate as a substrata was a sensitive indicator of the toxic action of the ZTS on the oxidative process. Later, to further isolate where in the process of mltochondrial respiration, the ETS acted, Gvozd~k et al (1985, 1987) reported data on succinate, NADH, and cy~ochrome oxidase activity in the mitochondria in the four groups of rabbits. Figure 2 shows the ~esults Of exposure to ETS on the activity of NADH oxidase, succlnate oxldase and c~ochrome oxidase of myocardial mltochondrla. The activity of the first two oxidases exhibited no changes compared with the control group -- neither after a single exposure to ETS or following exposures up to 2 weeks. Cytochrome oxidase activity decreased both after a single exposure to ETS and over time, with increaslng effects as the durat£on of exposure to ETS is extended. The observation ~hat Cytoohrome oxldase and not NADH or suocinate oxidase activity was affehted by ETS suggests that the deleterious effects of ETS on myocardial mitochondrial respiration occur at the terminal segment of the mltochondrial respiration process. Prolonged exposure to carbon monoxide has been shown in some studies to induce ultrastructural changes in myocardlum (KJeldssn et el, 1974; Thomsen and KJeldsen, 1974; Lough, 1978). Later, KJeldsen and co-workers (Hugodl et el, 1978), using a blind technique and the same criteria tO assess morphological myocardial damage found no significant changes in the coronary arteries or aorta in normocholestsrolemlc rabbits exposed to CO at concentrations from 200 to 4000 ppm for" up to 12 weeks. They suggested that the positive results obtained ,earlier were due to the non-bllnd evaluation techniques and the small number of animals used in these studies. Later~ Hugod (1981) : confirmed these neqatlve .results using electron~ microscopy. In addition, the earlier studies were conducted at "m~erate" levels of CO (i00 to 150 ppm) which are considerably higher than levels Of CO found in smoke-polluted environments (reported to be as high as 40-50 ppm, but more typically are around 10 ppm) (NRC, 1986). These negative studies only arque against an effect of C0 in inducing coronary 85

Draft - Do not cite or ~o~Q particular, ~he effects observed in Donsmokers smoking wi~ou~ inhaling were similar to the ~ffsc~s on smokers smoking ~wo cigarettes, despite the fact that the plasma nicotine levels in the nonsmokers were a factor of 5 smaller than those observed in the smokers (Davis et el, 1986). O~her work in the same laboratory comparing smoking with snuff use revealed similar changes in platelet function in response to these two forms of tobacco use (Davis etal, 1990). This result, combined with the finding that smoking non-tobacco clgarsttes (Davis st el, 1985a) fa~lad to produce changes in platelet function as large as observed wi~% tobacco cigarettes, suggests that nicotine .is an i~por~ant active agent, since non-tobacco cigarettes also affected platalet aggregation somewhat, however, it is possible that carbon monoxide or other combustion products are also influencing the platelets. Sinzinger and Kefalides (1982) measured platele% sensitivity to antiaggregatoryprostaglandlns (El, 12, and Dz) before, during and after 15 minutes of exposure to ETS in healthy nonsmokers and smokers (Table 3). Passive smoking reduced platelet sensitivity to the antiaggregatory prostaglandins Z2 and E2 significantly (P<.01) by a factor of about 2 by the end of 15 minutes exposure to ETS among nonsmokers. This effect persisted at 20 minutes after the end of exposure, and was gone by 40 minutes. Platelet response to prostaglandln D~ changed modestly in a similar pattern, but did not reach statistical signlfloance. Among smokers, the control level of platelet aggregation was higher (P<.01) and the prostaglandins had no significant effects on platelet aggregation over time during or following exposure to ETS. Sinzinger and • Virgollni (1989) also showed tha~ repeated exposure to ETS for one hour per day for ten days produced lasting changes An platelet function in nonsmokers similar to that observed in smokers. Thus, nonsmokerst platelets seem much more sensitive to a single exposure to ETS than do smokers' platelats, with platelet sensitivity to dlsaggregating prostaglandins having similar effects in nonsmokers acutely exposed to ETS as It does on the chronic levels of platmlet aggregation observed in long-ten smokers. Further evidence from the same laboratory that passive smoking increases platelet aggregation comes from work by Burghuber at al (1986), who had smokers and nonsmokers smoke two olgare~tes and also exposed a different group of smokers and nonsmokers to ETS in an 18mJ room in which 30 cigarettes had been smoked Just before. the nonsmokers. They measured ~he of exposing sensitivity platelets to the disaggregatlng substance prostaglandin 12 (PGI2), which is released by endotheliumand inhibits platslet aggregatian. (PGI2 is also called prostacyclin.) Figure 3 shows the results of this experiment. Zn smokers, neither smoking nor passive smoking affected the sensitivity of the plateleta to the dlsaggregatlng effect of prostaglandln Zz. The sensitivity of platelets in smokers was also significantly lower than nonsmokers. In contrast, platelets were more sensitive to prostaglandln 12 in nonsmokers, with both smoking and passive smoking producing slmilar reduction 90

Draf~ - DO no~ cite or quo~e in platelet sensi~ivi~y to prostaqlandin 12. These results suggest that the platelets of smokers are already desensitized to the anti-aggregatory substance prostaglandin I , so that no further decrease in aggregation is seen. The slgn~ficant decrease in platelet sensitivity to PGI~ after acute exposure to ETS suggests that after ETS exposure platelete are more likely to aggregate, wi~h the adverse ¢onsequenues de~rl]~ above° Earlier work by $a~a and Mason (1975) also indlcated that nicotine increased a variety of measures of platele~ aggregation in nonsmokers and smokers. While the effects of nicotlne on platelets from smokers was greeter than in non, q~okers,~/~e effect generally did no~ vary wi~h dose (between 2xl0"~and 2x10~ molar), suggesting that the effects of nicotine on pletslets occur a~ low doses and that the syste~ saturates ~/ic~y. This observation may explain why passive and active smoking have such similar effects on platelets (Sinzlnger and Kefalldem, 1982; Burghu~er et el, 1986; Davis et el, 1989). The probable link between nicotine ~nd adverse physiologic effects is nicotine-induced release of ca~echolamlnes. cateoholamines are then responsible for increased platelet aggregation. This reasoning suggests that beta blockers might provide some protection in smokers. This premise i$ borne out by the MAPHY trial --a trial oomparin~ the effects of the beta blocker metoprolol to a thiazlda dlureti~ in the control of moderate hypertension (Wilkstrand, et el, 1988). For the same reduction in blood pressure, the metoprolol trseted~oup had a lower mortality than the thiazide treated group. VIz~ually all of this reduction in mortality, however, was seen in smokers, and not non-smokers. This study provides evidence tha~ blocking th~ effects of cateoholamines (released by nicotine) was the cause of the reduced mortality in smokers who were receiving metoprolol. In sum, passive smoking has significant effe~s on platelet aggregation, of a magnitude similar to tha~ observed in active smoking. Moreover, the response of nonsmokers to both active and passive smoking appears to be different from smokers, with nonsmokers being more sensitive to low exposures to cigarette smoke than smokers. This observation suggests that the pharmacology of ETS in nonsmokers may be different than in smokers, with nonsmokers being more sensitive to low doses of ETS. ~n particular, it invalldabes attempts to estimate "clgarstte equivalent- doses of ETS in nonsmokers or extrapolatlng from risks of smoking in smokers to effects of ETS on nonsmokers. The ruultlnq increase in platelet aggregation can contribute to acute thrombus formation and myocardial infarction. In addition to the role of platslets in acute thro~us formation, platelets are also important in the development of atherosclerosis (Ross, 1986). Once there is damage to the a~erial endothelium, either through mechanloal or chemical factors, 91

Draf'~ - Do no~. ci~:e or ~o~:e pla~elets interac~ with or a~hlre ~o subendo~hellal connective ~$ssue and inltla~e a sequence which leads ~o atherosclero~ic plaque. When pla~ele~s interact wi~h or adhere ~o subendocar~lal connective ~issue, they are stlmula~ed ~o release ~heir granule conten~So Endothelial cells normally p~even~ pla~ele~ adherence because of the nonthrombogenic character oft,belt lurflce and ~helr oapaolty ~O for= antithromboti¢ substances such as pros~acycZin° once the endothellal ceZls havebsen damaqed, ths platelets can stlck to them. Once the pla~elets are bound ~0 the endothellum, they releasa mltc~ens such as platelet-derlved ~row~h factor (PDGF), which encourage =igration and prollfara~ion of smoo~h muscla cells in ~he region of ~he endo~heZiaZ injury (FOX and DLCorle~o, 1984). If pla~elet aqgre~a~£on Ls inc~easedbecausl of exposure to ETS, ~he chances Of platelets buildin~ up at an endothelial inJuz7 wL11 aZso be ~ncreased. Thus, in addition ~o contributin9 to acute effects ~h~ough Incr~asln~ ~he likellhcod of ~hrombu$ formation, the effects Of ETS on pla~elets also increase the chances tha~ endoth~llal injury wiZl Zead ~o arterial plaqua. ETS also plays a role in causln~ damage ~o the endotheliumand inltia~in~ the a~herosclerotic procl~s. As discussed above, Davis et al (1989, 1986, 1985, 1987, 1985b, 1990) found ~hat acute exposure to ETS (1989), like active smoking (19~G, 1985a, 198~, 1985b) and use of ¢hewin~ tobacco (1990), lead to a si~nifican~ increase (P<.002) in ~he appearance of anu~lear endothelial cell carcamses in ~he blood of peopll expoBed ~o ETS (or t~bacco or tobacco smoke) cons~i~uln~So The appearance of thesQ cell ~arcasses Indicates damage to tha ~ndothelium, whLch is ~he initla~ing s~ep in ~he a~heroscl~roti~ process. As noted above, in nonsmokers ~he a~pearance of endothelial cells ~ollowin~ passive smoking is aZmost as great as followin~ primary slokLnq (Ta~le 2). The process by which endo~ellal Injury leads to ~he davelopmen~ of an a~heroscler~tic plaque, ~nclu~in~ ~he role of platele~s, is described in ~$gure 4. ~ased on ~he information presented so f~r, ex~osur~ ~o ETS appears ~o ~r~ducl Injuries similar ~o ~hose observed with exposure to prlma~y, smoke and also affects pla~ele~s in a way ~ha~ increases ~he chances that ~hey wit1 bind to the injured area and promote ~row~h of smoo~h ~uscle cells. The R~le of the P~Ivcvcllc ~roma~ic Hvd~ocarbanS in ETS Many a~herosclero~ic plaquRs in humans are lither monoclonal or possess a predomlnantly ~onoclonal component (Bendltt and Bending, 1973), which indicates tha~ the smoo~h ~um~le cells of each plaque have a predominant c~II ~yp~. Several animal s~udies have also shown ~ha~ injections of polycyclic aromatic hydrocarbons (PAHs), in particular 7,12-dlme~hylben~(a,h)an~hracene (DM~A), benzo(a)pyrene (A1~ert at al, 1977; Revis, et al 198~; Penn e~ al, 1981; Penn e~ al, 19B61 MaJesky e~ al, 1983) accelerate ~he dev~lopmen~ of a~heros~lerosis. O~hers (Rogers ~t al, 1980, 1988) failed to find an effect of active smokin9 or ~he ex~mnt of fatty 92

0raft - DO no~ cite or ucte deposits in the coronary arteries of baboons. (There was a significant effect on the carotid arteries.) Benzo(a)pyrene is an important element in ETS (USPHS, 1985). The effects of pAHs or other carcinogenic or mutagenic elements in ETS (Remmer, 1987) relates directly to the response to injury theory of atherogenesis discussed above (Ross, 1986). Changes in the underlying snooth muscle stimulated by these agents could then initiate the "injury. that leads to platelet aggregation and plaque formation. Thus, chronic exposure to ET9 could have effects on plaque formation through mechanisms similar to that by which long term exposures produce cancer in other organs. Alher~ et al (1977) gavm chickens weekly intramuscular injections of DBMA and benzo(a)pyrene for up to 22 weeks, then killed the chickens at various times beginning after 13 weeks and measured the plaque volume in the chickens' aortas. They found that both DBMA and benzo(a)pyrene significantly increased the volUme of plaque compared to control chickens who had just received injections of the solvent used to carry these agents. This study provided the first evidence that known carcinogenic chemicals could be atherogenic as well. Penn et al (1981) extended this result in a similar experiment by showing that the effects of DBMA on the extent of plaque buildup in chickens was dose-dependent. The median cross-sectional area of plaques on individual aortic segments and the plague volume index (an approximate measure of the total volume of plaque per aorta) increased in a nearly linear fashion with DBMA dose. In contras~ to the marked increase in plaque area in the DBMA-treated animals, there was only a slight increase in the percentage of aortic sections with plaques in carclncgen-treated animals than in controls. Plaques with a small cross sectional area were present in all animals. Lesions of widely differing cross sectional areas appeared to be similar histologically under the light microscope. Together, these data suggest that a major effect of chronic DBMA exposure is to increase the size of spontaneous aortic lesions. Rather than inducing some sort of cancer-llke change in an individual ceil that begins the process which ultimately leads to formation of a plaque, Penn et al suggested that chronic DBMA exposure causes preferential division of individual cells or patches of cells within the preexisting spontaneous lesions. From this perspective, DBMA and other exogenous compounds would be acting as a mitogen, similar to that released by activated platelets, to stimulate division of aortic smooth muscle. Revis et al (1984) found similar results in White Carneau pigeons injected with DMBA and benzo(a)pyrene weekly for 6 months, beginning when the pigeons were 3 months old. Compared with the work described above, they fo~d a greater effect ON atherogenesis of benzoIa)pyrene than DBMA, and also failed to observe a dose- response relationship between the dose given and the amount of aortic plaque. These differences from the work just described may 93

Draft - DO no~ cite or quote be related to species differences, differences in the car~ier used to inject the pAHs (DMSO in the previous studies vs. corn ell in this one) or differences in the age of the pigeons or dosing schedule. They also found an increase in aortic plaques in pigeons treated with the PAH 3-methyloholanthrene, hut not the carcinogen 2,4,6-trichlorophenol or the PAH benzoCe)pyrene, whi=h is not considered a carcinogen. This result sugges~ that carolnogenio PAH$, rather than carcinogens or pARs in gaberal, are Implicated in the a~heroeclerotic process. Reviset al also studied the ~istrihution of these compounds after they had been radiolabelled. Foray-eight hours after the injection of PAHs, radloactlvlty in the liver, aorta and lung accounted for ?dr of the injected ~osa, whereas in ablmals injected with 2,4,6-trlohlorophenol, radloactiVlt¥ in the liver end kidney accounted for 80% of the dose. ~ In addition, 80% of the radioactivity observed in the~ pia$=a iuediatsly following injection =f radiolabelled PAHS wJa associated with the LDL and HDL cholesterol fractions,, compared with only 24% of the 2,3,6 trichlorophenol, suggesting that plasma lipoproteina are an important vehicle for transporting PAHm to their sites of activation in the arteries. There is also some evidence that ETS directly affects plasma lipoprotains. Moskowitz et al ~(1990) showed that adolescent children whose parents smoked had elevated levels of cholesterol and depressed levels of HDL, even ~ftsr correcting for age, weight, height and eex. These effects were dose depend6~t; the greater the exposure to ETS, the greater the changes in these variables. High cholesterol and low HDL are i:portant for the development of plaque. Data on cholesterol and HDL from adults married to smokers¸, however, do not show similar differences (Garland et el, 1985; Svendsen et el, 1987). To further elucidate the possible =eohanisms by which PAHs induce atherosclerotio changes, MaJesky st al (1953)¸ gave White carnesu and Show Racer pigeons a slngle injection of benzo(a)pyrene, then looked for :etaholites of the benzo(a)pyrene in aortic and hepatic tissues 48 hours later. White carneau pigeons develop severe atheroaclerosia by the tlmethe¥ are 3 years old, whereas Show Racer pigeons are relatively resistant to aortic atherosolerosls. Aortic preparations of the white Carnsau strain exhibited a much g=eater inducibillty of the ai¢=osomsl monooxygenase system than did those of the Show Racer strain, particularly in young pigeons. Aortic tissues from White Carneau pigeons aged 6-12 months exhibitsd a 3-12 fold inducibility whereas aortic tissues from the same shraln at 2-5 years of age exhlbi~ed only =inor Cmaxlmum of 3.3 fold) and, for the =0st part, statistically insignificant increases. No age differences in inducibility could be detected in the Show Racer strain. Interestingly, the differences in induclbillty :anlfest in aortic tissues were greater in aortic tlssues than in hepatic tissues from 94

Draft - Do not cite or c~uote ~he same birds. Thus, the PAHs seem to accelerate any preexisting tendency to develop atheros=lerosis. Regardless of the ultimate mechanism by whiah PAHs exhibit atherogenic effects, it seems logical tc suppose that the reactive intermediary metabolites of these ohemlcals are the proximate atherogenic or co-a~her~enl= agents slnoetheparent compounds are relatively iner~ both chemically and biol~Ically. Thus, bioactivatlon and inactivation (and regulatory ccntrol of these processes) may be presumed to play extremely important roles in their atherc~enic prope~les. Bioac~ivated chemicals vary in their stabillty/reactlvlty accordinq t~ four general categories. (i) those which are extremely unstable and persist only a~ the immediate site (enzyme) of bioa~clvatlon, (ll) those which persist only within cells in which bioac~Ivation O~curs, (ill) thcse which persist primarily only within tissues in which blcactlvation occurs, and (Iv] those capable of being transferred in the circulation from one organ to a~Other. For the first three of these four categories, biotraneTc~latlon in the ao1-~a per se (target tissue activation) would be of prime interest and importance. Thus, it appears that PAHs could ha playing either a mutagenic or mitogenic role in beginning the atherosclerctio process in suscpptible cells or individuals, depending on how the PAHs in ETS are metabolized in the aoz-~ao The flndingthat enzymes that metabolizeDMBAand benzo(a)pyrene are in the artery wall led Penn et al (1986) to search for specific molecular events in plaque calla that would lead to D~A changes similar to those previously found in t~ors. Identification of such processes would be supportive of the monoclonal hypothesis of atheroqeneeis. They obtained human DNA samples from coronary artery plaques as well as DNA from normal sectlons of the coronary arteries at surgery to remove ~he plaque. These DNA samples were tested with the NIH 3T3 cell transection assay. Foci arose in cells transfected with each of the DNA samples obtalned from the hu~n coronary plaque, with an efficiency (number of fool per ~g of DNA) ranging from 0.016 to 0.060 (mean 0.036). The transfectlon efflclencies for DNA~S from normal coronary a~cery, liver, spleen, lung, kidney and trachea were all below 0.008. The transformed cells were also injected into the scalps of nude mice, where they developed tumors. These results provide direct evidence for similarities on the molecular level in the development of plaques and tumors. Human coronary artery plaque DNA contains sequences capable of transforming NIH 3T3 cells and these transformed cells can cause tumors after injection into nude . mice. Control experiments refilled that the transforming cells did indeed conthin human DNA and that the ttmorlgeni= (or transforming) activity was not due to the rasoncogene famil~. Although t hiss results clearly demonstrate that human plaque DNA has transforming ability, the temporal expression of this activity ~ is not known. The plaques were taken from adult patlents in late stages of vascular disease. Thus, we cannot determine from these samples whether the 95

Draft - Do not ci~e or quota manifestation of transformation~is a relatively late event in plaque development or an early but stable event, oncogene activation and expression is an important early event in transformation and tumor genesis. These results identify specific molecular events that may underlie ~he prollf~ration of smoo~h muscle cells that is a helllark oZ etherosclsrotlc plaque development and demonstrates that plaque cell~ sxhibit molecular alterations that had previously only been thought to Me present in cancer-cell transformation and tumorigenesis. These results provide direct suppor~ for the monoclonal hypothesis. Randera~h et al (1988) also demonstrated that constituents of cigarette "tar," including benso(a)pyrens, ere preferentially attracted to the heart and damge DNA there. They studied molecular mechanisms of smoklng-r01atedcarcinogenesis by examining the induction and distribution of covalent DNA damage in internal organs of the mouse following topic application of cigarette smoke condensate daily for i, 3, or S days then killed 24 hr later. DNA samples were obtained from skin, lung, hsar~, kidney liver, and spleen. Adducts containing ben~o(a)pyrene-derived moieties were identified, together with others. At all three times, the number of adduots in hear~ and lung DNA was about five times higher than that in liver and slightly higher than that in skin. Covalen~ DNA damage was estimated to he 5.2, 5.7, 3.9, and 1,9 times higher, respectively, in lung, heart, 8kln and kidney than in liver, ranging from approxlm~tely, l adduct in 5.4xio- DNA nucle0tides in lung to 1 adduc~ in 3.3xi0" DNA nucleotides in liver. Spleen DNA was vi~ually adduct free. Whilethe DNA adduct profiles resembled each other qualitatively among the different tissues, there were maJor quantitatlve differences between the different .tissues, with the highest DNA binding occurring in the lung and heart. The reasons for the high incidence of DNA adducts in the heart are not known, hut may be related to the role of plasma liplds in transporting PAHs such as benzo(a}pyrene and preferential binding of these llpids to cardiac tissue, as discussed earlier. In sum, there is a grawing body of evidence at a molecular level supporting the monoclonal hypothesis of atherogenesis, with compounds in tobacco smoke and ETS strongly implicated as agents which stimulate the development of coronary lesions. Regardless of whether the monoclonal hypothesis proves to he true Cot, more llkely, one of several initiators of theatherosclerotlo process), the fact is that there is clear evidence that components of ETS, in particular PAHs such as benzo(a)pyrene, initiate or accelerate the development of plaque. These biochemical findings ere consistent with the epldemlological finding that chimney sweeps, which are exposed ~o high levels of PAHs in soot, have an Increased risk of heart disease (as well as cancer) and tend to develop these diseases younger than "Other, co.arable, occupations which avoid exposure to PAHs (Hanssn, 1983). summa~ 96

Draft - Do no~ cite or quoue There are eleven epidemiological studies, done in a variety of locations, which reflect a 30% increase in risk of death from ischemio hear~ disease or myocardial infarction among nonsmokers living with smokmrs. The larger studies also demonstrate a statistically slgnlficant doss~rssp~nsa offer, with larger exposure to ETS being associated wi~h greater risks of death from heart disease. These epldemiclogical studies are complemented by a variety cf physiological and biochemical data from htt~a~studles which suggest that ETS may adversely affect platelet funotlon and damage arterial endothelium in a way that increases the risk of heart disease. Moreover, ETS, in realistic exposures, also exacts significant effects on exercise capability of both normal people and people with hear~ disease by affecting the body's ability to deliver and utilize oxygen. In animal axperlmsnts, ETS also depresses cellular respiration at the level of mitochondria. The polycycllc aromatic hydrocarbons in ETS also accelerate, and may initiate, the development of atherosolerotio plaque. It is also important to note that the cardiovascular effects of ETS appear to he different in nonsmokers and smokers. Nonsmokers appear to be more sensitive to ETs than smokers, perhaps because some of the 9fleeted systems ar~ sensitive to low doses of the compounds in ETS, then saturate and also perhaps because of physlological adaptions smokers ttndergo as a result of chronic exposure to the toxins in cigarette smoke. In any event, these findings indicate that, in terms of cardiovascular disease, It is unreliable to compute "cigarette aquivalentm- for passive exposure to ETS, then try to extrapolate ~he offeots of this exposure on nonsmokers from the effects of direct smoking on smokers. These results combine tc suggest that heart disease is an important consequence of exposure to ~TS. The combination of epidemlologlcal studies with demonstration of physiologloal changes with exposure to ETS, bogeyer with biochemical evidence that elements of ETS have slgnificanb effects on the cardiovascular system, lead tothe conclusion that ETs causes heart disease. This increase in risk translates into about 10 times as many deaths from ETS-Induced hea~ disease as l~mg cancer, a~d contributes 37,000 to the estimated 53,000 deaths annually from passive smoking (Wells, 1988). This toll makes passive smoking the third leading preventable cause of death in the United States today, behind active smoking and alcohol. 97

Draft - Do not cite or quote e Alhe~ R, Vanderlaan F, Nishizuml M (1977) Effect of ¢arclnogens on chicken atherosclerosis. ~ 37: 2232-2235. AllredEN, BleeckerER, CA1aitmanBR, Dahm~TE, Gottlleb SO, Hackney JD, Pagano M, EelvesterP/4, Walden SM, Warren J (1989) Shor~-tez~ effects of carbon monoxide exposure on the exercise performance of subjects with coronary a~ery disease. ~ 321: 1426- 32. Aronow W (1978) Effect of passive smoking on angina pe=tori$. N. 299: 21-24. Bardford-Hill A (1984) A Sho~: TsXtb~k of Medical Statistics London: Hodder and Stoughton (fled.) Barrett T, GaJdusek C, Schwartz 8, McDOugall J, Bendltt E (1984) Expression of the sis gene by endothelial cells in culture and in vlvo. ~ 81! 6772-6774. Benditt E, Benditt J (1973) Evidence for s monoclonal origin of human atherosclerotlc plaques. ~ 70: 17S3- 1756. B~rghuber O, Punzengz~Iber c, Slnzinger H, Eaber P, Silberbauer (1986) Platelet sensitivity to prosta~yclin in smokers and non- smokers, chest 90: 34-38. Butler T (1990) The relationship of passive smoking to various health outcomes among Seventh-Day Adventists in California. Seventh World Conference on Tobacco and Health, 316 (abstract). Davis J, Hartman C, Lewis H ~r, Shelton L, Eigenberg D, Hassanein K, Hlgnite c, Ruttlnger E (198Eh), Cigarette smoking-induced enhancement of platelet function: Lack of prevention by aspirin in men with coronary artery disease. ~ 105: 479- 483. Davis J, Sheltcn L, Eigenberg D, Hignlte C, Watanabe I (1985) Effects of tobacco and non-~obacco clgerette smoking on endothelium and pla~sls~s. ~ 37: 529-533. Davis J, 5helton L, Eigenberg D, Hignlte C (1957) Lack of effect of aspirin on cigarette smokl-lnduced increase in circulating endothelial cells. ~ 7~ 66-69. Davis J, Sheltcn L, Earta=an C, Eigenberg D, Ruttinge~ E (1986) Smoking-induced changes in endothellum and platelets are not affected by hydroxyethylrutosides. ~ 67: 765-771. Davis J, Shelton L, Watanabe I, Arnold J (1989) Passive smoking 9S

Draft - Do no~ ci~e or quote affects @ndothell~Lm and pla~mlets. ~ 149~ 386- 389. Davis J, ShQlt~n L, Zucker M (1990) A comparison of some acutl effQc~s of smoking and Bmokells8 tobacco on pla~elitu and en~othelium. (submitted) Deanflal~ /E, Shea MJw wilson RA, ~orlock p, de LandshQrro CM, $Qlwyn AP (1986) Diroct effects of smokln~ on tho heart~ silQn~ Ischemic disturbances of coronary ~low. ~ 57~ I005- 1009. Dobson A, Heller R, Alexander H, Lloyd D (~990) Pammiva smokln~ and the risk Of hear~ at~cko SavAnth Wo~Id Canferencm on Toba~Q ~, 102 (abstract). FOX P, DiCorle~o P (1984) Ragula~ion of produatlon of a ~atelet- ~riv~d g~ow~h fac~or-lik~ protai~ ~y cul~urQ~ bovine aortic endo~helia~ cells. 121~ 298-208. Friedman J, Chalme~s T, Smith ~ J~, Keubler R (1978) Tha ~mportan~e of beta, ~he Type II error, and sample size in the design an~ Int~rp~ation of the ran~omlz~ ~on~o11~ t~ia1~ Survey of 71 "negative" trials. ~ 299: 690-694. Fuster V, chesebro J (i~81) Ant~thrombotlc therapy~ Role of plat~let-inhlbitor drugs~ I. Curr~nt Concepts of ThrombogenesIi: Role of Pla~ele~s. ~ 56~ 102-112. Ga~land C, Barret~-Connor E, Suaroz L, Crlqui M, wi~gard D (1985) EffQcts of passive smoki~ on Ischem~c heaz~ diseasa mortality of nonsmokers. ~ 121~ 645-650. Gillis C, Hole D, Hawthorne V, Boyle ~ (1984) ThQ effect of envlronmen~al ~obacco smoke in ~wo urban communities i~ ~he w~t of Scotland. ~ 65 (suppl 133): 121-12~. Glantz and Parmley (1990) Passive smoking an~ hoar~ d~sea$~: Qpide~iology, physi~1ogyand blochemls~ry. ~ (in p~ss). GvozdJ~k J, GvozdJ~kov~ A, Kucharsk~, Ba~a V (~987) The e~fec~ of ° smoking on ~yocar~ial metabol~Smo ~ I0~ 47-53. Gvozdj~kov& A, Bada V, S~ny L, Kucharmka J, Kru~ F, Bo ek, Trltanks~ L, Gvozdj~k J (I~84) Smoke cardiomyopa~hy~ Disturbance of oxidative process in ~yocardial mi~ochon~r~a. 18; 229-232. GvozdJ~kov~ A, Kucharsk~ J, S~n~ L, Bada v, Bo ek, Gvoz~J~k J (198~) Effect of smoking on the cytochro~ and ~xidase ~ys~m of the my~cardlum. ~ 8~: 10-15. 99

D~aft - DO not cite or quote Hansen E (1983) Mortality from cancer end Ischemi¢ heart disease in Danish chimney sweeps: A flve-year follow-up. 117: 160-164. Hartman P (1983) Mutagerm: Soml possible health impacts beyond carcinogenesis° ~ 5s 139-152. He • (1969) Women's passive nok£ng and coronary hear~ dlseaae. Chuna-Hua-Yu-Fana-I-Hsu~h-Tsa-Chln 23: 19-22. (English traDslatlon of entire article.) Helsing K, Sandler D, Comstock G, Chee E (1988) Hear~ disease mortality in nonsmokers living with smokers. 127: 915-922. H~rayama T (1984) Lung cancer in Japan= Effects of nutrition and passlvm smaking. In Luna Canca~ Caugel and Pr.vantlon M° Mizell and P Cortes eds, pp 175-195o Verlag Chemie Internatlonal, New York. Hole D, Gillis C, Chopra C, Hawthorns V (1989) Passive smoking and cardiorespirat~ry health in a general po~lation in the west o~ Scotland. ~ 299~ 423-427. Hugod C, Hawkins ZH, KJeldsen K, ThomsenHK, Astrup P (1978) Effect of carbon monoxide on ao¢~¢ic and coronary intlmal morphology in the rabbit. ~ 30: 333-342. Hugod C (1981) Myocardial morphology in rabbits exposed to various gas-phase constituents of tobacco smoke. ~ 40: 181- 190° Humble C0 Croft ~, Gerber A, Casper M, Hames C, Tyroler H (1990) Passive smoking and twenty yea~ cardiovascular disease mozcelity among nons~oklng wives In Evans county, Georgia° 80: 599-601° ~arvis M~ (1987) Uptake of environmental tobacco smoke. In Environmental cercinooenf! methods of analvmi~ end e~coolu~ lea~ement~ Vol 9! Paslive Smoklna OINeill I~ Bro~e~ar1~ KDt Dodet B, Hoffman D (eds), Lyon France= ~ARC, pp. 43-58. Kawachl Z, Pearce N (1989) Passive gmokln~ in NQW Zealand (letter). ~ 102J 479. Kawachi I, Pearce N, Jackson R (1989) Deaths from lung cancer and ischaemic hear~ disease due to passive smoking in New Zealand. ~. I02:337-340o Khalfen E, Klochkov V (1987) Effect of passive smoking on physics1 tolerance of Ischemic heer~ disease patients. ~ 59: 112- 115. 100

Draft - Do not oite or quote KJeldsen K, Thomsen H, Ast~up P (1974) Effects of carbon monoxide on mycoardlum: Ultrastructural changes in rabbits after m~erata, chronic exposurs, ~ 34: 399-348. Kristensen T (1989) Cardiovamcular diseases and ~he work environment: A critical review of the epidemioloqio literature on chemical factors. Stand J Work Environ Neal~h 15:245-254. Lanb D (1984) Phvsloloav of exe~ise| Responses and adaptation. MacMillan Publishing co.: New York. Lee P (1988) Misclassiflcation~ of smokino habits and nassive sm~kln~. A review of ~he evldenee. Intex~natlonal Archives of Occupational and Environmental Heal~h. Berlin: Springer-Verlaq. Lee p (1989) Deaths from l~g cancer and ischaemic heart dlseame due to passive smoking in New Zealand (letter), i02: 448. Lee P (1990) An estimate of adult mortality in the United Steres from passive sm?king: A response (letter). ~ 16:179- 181. Lee P, chamberlain J, Alderson M (1986) Relationship o~ passive smoking to risk of lung cancer and o~her smoklng-associated diseases. ~ 54: 97-105. Lough ~ (1978) Cardiomyopathy produced by cigarette smoke. Arch. 102: 377-380. Majesky M, Yang H, Benditt E (1983) Carcinogenesis and acherogenesis: Differences in monoxygenase inducibility and hioactivation of henzo[a]pyrene in aortic and hepatic tissues of atherosolerosis-susceptlble versus resistant plgecns. 4: 647-652. Martin M, Hunt S, Williams R (1986) Increased incidence of heart attacks in nonsmoklng women married to smokers. Paper presented at annual meeting of APHA, October. McMurray R, Hicks L, Thompson D (1985) The effects of passive inhalation of cigarette smoke on exeroise performance. 54: 195-200o Moskowltz W, Mosteller M, Schieken R, Bossano R, Hewitt J, Bodur~ha 3, SegreSt J (1990) Lipoprotein and oxygen transpo~ alterations in passive smoking preadolescent children: The MCV twin study. 81: 586-592. Muhm J, Olshan A (1989) A program to calculate sample size, power, and least detectable relative risk using a programmable calculator. 101

Draft - Do not cite or quote 129: 209-11. NRC (1986) Environmental Tobacco Smokax Measurina EXPosure and ~U~. Washin~on DC: National Academy Press. Palmer ~, Rosenberg L, Shapiro S (1988) Passive smoking and myoGardial infarction. ~ 43: 29, 1988 (abstract) Penn A, Batastlni G, Soloman J, Burns F, Albe~ R (1981) Dose- dependent size increases of a01~¢ic lesions following chronic exposure to 7,12-Dimethylbenz(a)anthracene. ~ 41: 988- 592. Psnn A, Gaffe S, Warren L, Nests D, Mindlch B (1986) Transforming gene is human atheroscZsrotic plaque DNA. 83: 7951-7955. Pittilo R, Mackie I, Rowles P, Machin S, Woolf N (i982) Effects of cigarette smoking on the ultrashructure of rat thoracic aorta and its ability to produce prostecyclln. 48: 173-176. Randerath E, Mittal D, Randersth K (1988) Tissue distribution of covalent DNA damage in mice treated dermally with cigarette 'tar': preference for lung and hear~ DNA. ~ 9: 75-80. Relnken J (1989) Passive smoking in New Zealand (letter). N. Zeal. I02: 515. Rammer H (1987) P~ssively inhaled tobacco smoke: A challenge to toxicology and preventive medicine. ~ 51: 89-104. Repace J, Lowrey A (1985) A quantitative estimate of nonsmokers' lung cancer risk from passive smoking. ~ ii: 3-22. Repaoe J, Lowrey A (1987) Predicting ~he lung cancer risk of domestic passive smoking. ~ 135: 1308. Revis N, Bull R, Laurie D, schiller C (1984) The effectiveness of chemical carcinogens to induGe atherosclerosis in the white carneau pigeon. ~ 92: 215-227. Rogers WR, Bass RL III, Johnson DE, Kroskl AW, McMahan CA, Montiel ~, Moth GE, wilbur RL, McGill RC Jr (1990) Atherosclerosls-related response to cigarette smoking In the baboon. ~Sl: 1198- 1193. Rogers WR, Carey KD, McMahan CA, Montlel ~, Moth GE, Wigodsky HS, McGill HC Jr (1988) cigarette smoking, dletaryhyperllpldemla, and experimental atherosclarosls in the baboon. ~48: 135-151. 102

D~f~ - DO no~ cite or quc~e Ross R (1986) The pathology of athezosclerosis -- An Update. N. 314: 488-500. Rothman K (I~7~) A show of confidence. ~ 299: 1362-1363. S~a S, Mason R (1975) Some efZects of nlootlne on p1atelets. ~ ?: 819-824. SarabJlt-slngh C, Bend J, Phiipo~ R (1985) cytochr~me P-450 monooxygenasa sys~mm: Locallza~on i~ smooth muscle of rabbi~ aoz~a. ~ 28: 72-79. Sheps D, Adams K Jr, Brombaz~ P, Goldmte~n G, O'Neil J, Horstman D (1987) Lmck of effect of low levels of car~oxyhemoqlobln on cardiovascular function in pat$ents with ischemlo hea~ disease. 42: 108-116. Slnzlnger H, EefalidRs A (19~2) Passive smoking severely deoreases pla~I1et sensitivity ~o an~iaggrea~ory prosta~landlns. Lancst 2(8294): 392-393. Sinzlnger H, vir~olin± I (1989) Are pasBive smokers a~ ~reater risk of thrombosis? Wi~ne~ kllnlsmhe W~eh~nmohrift 20: 694-698. Sven~sen K, Kuller L, Nar~in M, Ockene J (1987) E~fects of Passive Smoking in the Mu1~Iple Risk Ya~tor Znterv~ntion Trial. 126: 783-795. Thomsen H, KJ~Idsen K (1974) Thremho1~ lim~ for carbon monoxide- induced myocardial damage. ~ 29: 73-78. USPHS (1983) The ~malth Conseamencms of Smokln~: Cardiovascular A rm~oz~c of the Surgeon G~nerB1. DHHS(~S) 84-50204. USeS (1986) The Health CDnsmcmencms of Involuntary Smokin~:~ A rmDort of the Sur~emn Gan~rm1. DHS(cDC) 87-8398. Wald N (1986) Does breathing ot~er peop1e~m ~o~aoco smoke cause lunq cancer? ~ 293: 1217-1222. Wall MA, Johnson J, Jacob P, Benowi~z NL (1988) Co~$nine in the serum, sal~va, and urine of ~onsmoke~s, passive amokers, a~ active smokers. ~ 78: 699-701. Wells A (1986) Misclasslflcatlon as a factor in passive smoking ~isk. Lancsk ii: 638. Wells A (1988) An estimat~ of adult moz~ali~y in the United S~a~es from passive smoking. ~ 14: 249-265. 103

Draft - Do no~ cite or ~uote Wells A {1990) An estimate of adult mor~allty in the Unite~ States from passive smokln~: A response to critloism. ~ 16: 187-193. Wikstrand J, warnold I, Olsson G, , Tvomilehto J, Elmfeldt D, Berglund G (1988) Primary prevantlon wlth metoprolol in patients wi~h hypertenslon: Mor~allty Eesults from the MAPHY trial. Jk~t~ 259: 1976-1982. 104

Draf~ - ~ no~ cite ur ~ote FTGURES AND TABLES, CHAPTER 6 105

Draf~ - DO not cite or quot8 Rmb¢ Smo~q aad Hart D~ [p~em~o~ ~ ~d B~m7 CR~0213R3 TABLZ i. 32 &l 13 370 1~ Q.7-2.1 1.2 O.i-2., 2.1 0.7-6.3 1.3 1.1-1,6 1.3 1.1-1., SX m e~l. Ill~tl| stttu~ e(¢~o( t.2.1,&

Draft - Do not ¢±t~a or~r crucot~e 4

D~a~"~ - Do no~ cJ/ce or ~ol:a SmokJat ~nd ~M DLs~s~ ~idemJo~ ph~o~, and B~ ~.c~02~R-1 table 2 Ef~*~: of ~essLve et~ Ac:Iv~ Sa~Iclng on P|4teLet A~IgrtIMtiw~ ord |~theIIs( CeLL Dm PLeteLet A~r~te IltlO ~c~theLSeL Call C4~n~ lefc~e Aftl~ ~nge ~ After Cher4e Pmiw ~tr4 .U ,78 -,~ |J 3.7 0.9 (ncrmak~) n 10 ZO 17

~a~ - ~o no~ cite or c~.to~e pr~it Ntardln I! m l! Tn{S pta~sie: rdr4ft~v|t~ to Pr~t~Lsndfml h~.s ~d &ftsr PI~|,~ ~f~ (rig N/~ DIItltet P~ch p~lm) 1.;Mr0.11 18o7S.3.1 ~.7do8 |. 16b'@.21" ~.h6A" 55oid.3 N.h6.1° 51 .]/,t.z 6~ Itlt~ 1 ..'~r0. lY 1.7510.26 |.08~.16 2.06~.1S l.Srh0.23 Z?.S*Z.3 30.~o$ 31.0~,+1 29.1~.9 'P~.01 ~ to control. IINU|~I IP" ~ +i sl~4o tl~jr.Na Slr~ll'~Ir Ir, d l~Ifltldll ¢1M~) -l~.° . c~[antz~an~n~p~,dc~

Draft - DO not cite or quote Figure i: Relative risk in epidemiological studies of the risk of death from coronary heart disease or myocardial infarction among nonsmokers living with smokers co~pared wlthnonsmokers living with nonsmokers. Lines indicate 95% Confidence intervals. (Note that two studies have upper hounds to confidence interval off the scale of the graph.) Figure 2: Effect of passive smoking on myocardial mitochondrial respiration. QO2(S3) - oxygen consumption in mltochondria in the presence oft he su~strate and ADP; QOz(S~ - capacity of respiratory chain without ADP added; OPR - ox£dation phosphorylation rate; ADP:O - oxidation phosphorylation coefficient; RCI - respiratory control index. Source: GvozdJak et al (1987) Figures i and 2. Figure 3: Effect of active (left) and passive (right) smoking on platelet aggregation in smokers and nonsmokers. The sensitivity index, S~zz, is defined as the inverse of the concentration of prostaglandin ~ necessary to inhibit ADP-lnduoed platelet aggregation by 50%. Lower values of Sz ] indicate greater platelet aggregation. Source: Burghuber et a~ 11986) Figures 3 and 4. Figure 4: Advanced intimal lesions of atherosclerosis may occur by at least two pathways. The pathway demonstrated by the clockwise (long) arrows to the right has been observed in experimentally induced hypercholesterolemia. Injury to the endothellu~ (A) may induce growth factor secretion (short arrow). Monooytes attach to endothelium (B), which may continue to secrete growth factors (shor~ arrow). Subendothelial migration of monocytes (C) may lead to fatty-streak formation and release of growth factors such as platelet-derived groWth factor (PDGF) (shor~ arrow). Fatty streaks may become directly converted to fibrous plaques (long arrow from C to F) through release of growth factors from macrophages or endothelial cells or both. Maorophages may also stimulate or injure the overlying endoEheli~t~. In some oases, macrophagel may loose lose their endothelial cover and platelet attachment may occur (D), providing three possible sources of growth factors -- platelets, macrophages, and endothellum (sho~ arrows). Some of the s~ooth-muscle cells in the possible lesion itself (F) may form and secrete groWth factors such as PDGF (shor~ arrows). An alternative pathway for the development of advanced lesions of atherosclerosis is shown by the arrows from A to E to F. In this case, the endothelium may be injured but remain intact. Increased endothelial turnover may result in growth-factor formation by endothelial cells (A). This may stimulate migration of smooth- m~scle as well as growth factor secretion from the "injured" endothelial cells (R). These interactions could then lead to fibrous-plaque formation and further lesion progression (F). The PAHs in ETS probably act by the second pathway, source: Ross (1986) Figure 6. ii0

Draft - Do no~ cita or quote P~GURR~ ~ TABLRg. C~IPT~R 6 111

Draf~ - ~o no~ cite or quote 4686~) elOH I ! (~6 L) PuE2IJI~"-J (lw~6 ~- ) ~J!ll!~ (Z.g6 L ) uespue^s I {,686'.) eH -'4'-- ('~86L) ~u!$1eH ~t- [886 ~1 ee7 I 1 (88e~.) eu~sl~-~ --.k-,I I , (886L) ee'l I i ('¢8e~-) sti[1~ I i I I i i i I L['% '~" 03 C'J "- o% X U,, 0 N'$!~ e^!~.gle~l

Dra~ ~I ~i~--~I'~ ~---~' °~

Draft - Do not cite or ~ote

Deaths from Passive Smoking Total Deaths: 53,000 Heart Disease 37000 Lung Cancer 3700

D~.-aft - Do not cite or quc~e O CHAPTER 7 ZXI~OSURE I%SSZSSMZNT ZN P&SeZVE 8MOKZNG James Lo Repate, MSO Zndoo¢ ~tl~ Did, Isles office of ]k.iz & Rad~.abian U.S. Environmental ~rotemtion &qency Weahingtom, DC 20460 This chapter will discuss some of ~he factors involved in the assessment of exposure to indoor air pollution from tobacco smoke. Exposures to environmental tobacco smoke (ETS) have been assessed by questionnaires, personal air contaminant monitoring of ETS constituents, modeling of concentrations, and biological markers. {NRC, 1986) Most of the epidemiol~ical studies of passive smoking and disease have relied on questionnalres relating to the presence or absence of a smoker at home, and have assumed that this is a good surr~ate for total exposure to ETS. To the extent that nonsmokers heavily exposed outside the home, e.g., the wor loce, will not differentiate exposure variable surrogate well between a more exposed and less exposed group, and tend tO cause epidemlolc~ic studies of pamslve smoking and disease to find no effect or to lack statistical signlfioanoe.(Repmme & Lowrey, 1990) For this reason, several workers, in assessing risk from passive smoking, have attempted to ¢orre~ for exposures outside the home by adjusting for the finite urinary cotinlne concentrations in those who have reposed "no exposure" tc ETS. (NRC, 1986; Wells, 1990) NO studies have yet been performed which yield a national probability sample of exposures to ETS. Thus, all attempts to assess exposure in individual epidemlologioal studies and o~herwlse have relied on some aesuJed paradigm of exposure. This chapter will discuss the evidence for exposure, and emphasize the insights which derive from a modeling approach. Exposure to ETS occurs when an individual occupies a microenvlronment which possesses an ETS concentration. A dose is said to occur when the individual breathes the conoentration. An individual's total exposure to ETS is ~he tlme-welghted sum of the individual microenvironmental ET8 exposures encountered during the day's activities.(Repace et el., 1980). The dose of ETS will be affected by the individual~s respiration rate during the exposures. The dose of various ETS constituents to the body will be determined by their relative rates of absorption and removal. The amount of ETS inhaled is given by the product of the individual's respiration ra~e during exposure, the ETS concentration in the building, and the duration of the individual.s stay in that microenvironment. In i12

Draft - Do not cita or ~ote equilibrium, the ETS concentration is directly proportional to the product of n~er of smokers, smoking rate, and emissions per cigarette, pipe, or cigar, and I$ inversely proportional to the produc~ of space volune and removal rate. (Rape=e, IARC, 1967) In the epidemlolo~ic studies of passive smoking and lung oancer, e~sures are ty~ically estlmated on the basis of a questionnaire which assesses smokir~ s~atus, a~ ty~i=ally ask silple questions of the sort: "if you are a nons~ker, do you live with, or work witht or have regular contact with persons who are nonsmokers?" (NRC, 1986) Some studies assess ~st e~sure history and spouses' smoking rate as well. Thls kind of question, though useful, is not likely to be fully reliable or precise, particularly for non- domestic exposures. (NRC, 1966~ IARC, 1997} On the other hand, it has been shc~n tha~ those nonsmokers who repo~ exposure to ETS at home tend to have higher non-domestic exposures as WelI.(NRC, 19S6; Weld, 1986) Those individuals who have exposures both at home and at work appear to have higher exposures than those who are exposed at home only or at work only, as reflected by their urinary ootinine excretion. (Riboli, et el., 1990} Riboll et el. (1990) repor~ data from a ten-country study of 1369 wo~en, showing that when appropriately questioned, nonsmoklng women can provide a ~asonably accurate description of ETS exposure. Ideally, the health effete of ETS might be assessed by quantifying the tlme-dependent exTposures for each of the several thousand compounds in tobacco smoke add defining dose-response r.lationships for these compounde in producing disease, both as isolated co~pounds and in various combinations. However~ the en0z~ity of this task has led to simpler approaches which attempt to use ~easures of exposure to Indlvidual smoke oonstituents as estimates of whole smoke exposure. For this rlason, exposures to ETS are often assessed using markers of the vapor phase o~ particulate phase. Although biological markers show promise as measures of exposure (and dose}, ~hey also have limitations. Another consideration is duration of exposure. For ohronic diseases s~ch as cancer, average exposures occuring over a year or lifetime are of greater importance ~han short-tars exposures. (SG, 1986} The two mos~ promising aTJmospherlo markers for ETS are respirable suspended partlcles in the size range ~3 um (RSP)and " nicotlne.(NRC, 1966~ SG,1996~ IARC,19S7) A majority of field studies have used RSP as an indicator of exposure to ETS because of the substantial e~isslon of RSP in indoor spaces from ~ohacco combustion. ETS is the dominant contributor to the indoor levels of RSP. The total RSP, as measured by personal monlt~rs, has been found to be ~ubstantially elevated for those who repor~ exposure to ETS relative ~o those who report no exposure. Both air monitoring and modeling olearly indicate ~hat RSP ~oncen~ra~ions will be elevated over background levels in indoor spaces when evln low smoking rates occur. (NRC, 1966) Although lacking specificity 113

Draft - Do not cite or quote for toba¢co smoke, the prevalent• and number of smokers correlates well with RSp levels in homes and other enclosed areas.(sG,1986) RSP is the single largest component of ETS by weight, and RSp is CUrrently the best an~ most-utilized general category 9f air Contaminants to represent ETS.(NRC,1986) A recent study of week- long averages of RSP and nicotine in about i00 homes with smokers in New York Stats showed an RSP-to-n~c0tlne ratio of about II:I, above a background of about 20 ug/m'. (Lead•rat, 1990) Similar results were obtained in • surv~ of 21 ¢onerclal buildings by Miesnsr, et el. (1989), who found a RSP-tc-nicctine ratio of abput i0:i for workday averages, above • background of about 29 ug/m~ Biological markers in body fluids have also been used for validating self-reports of exposures to ETS. For example, Haley et ai.(1989) and cummings et •i. (1990) found that cotinlns levels in the urine of ~hose who reported exposure to ETS were more than twlce as high as those who denled having been exposed. Nicotine and its mstabolite cotlnine, which ~erlve exclusively from tobacco products, are the most Important markers° Almolt all nicohlne shifts from the particulate phase in mainstream and ~resh sldestream smoke to the vapor phas• in ETS. Nicotine and Cotlnlns can be quantified in saliva, blood, and urine. Generally, the mean concentrations of niootlns and ~cotlnlns in plasma or urine of nonsmokers exposed to ETS under natural conditions is about 1 percent of the mean values in smokers, (NRC,1986) reflecting the fac~that smokers are present in nearly all environments, including most workplace•, restaurants, and even ID ~anyvehlclen, making it almost imposslble for nonsmokers to avoid exposure to ETS.(SG,1986) A. Sources of ZT8 In 1996, an estimated 50 million US smokers aged ~17 yrs smoked about 584 billion cigarettes annually. (NRC, 1986: Tobacco Institute, 1987) They consumed an additional 3.2 billion cigars, as well as an estimated 24.4 million pound• of tobacco for pipes and hand-rolled olgarettes.(NRC, 1986) The average US clgartttl smoker smokes 32 clgarettes per day at a rats of 2 cigarettes per hour and emits about 22 mg of RSP per clgarette.(Repace, IARC, 1987) Since the average person spends about 90% of the time indoors, an estlmated 12,00o me~rlc tons of RSP are eRihted into US indoor mioroenvlronments eachyear from cigarette smoking al~ne. Assuming cigars produce 3 time• as mu~h RSp a• cigarettes and that pipes produce as much RSP as a cigarette, (Repace and Lowrey, 1982) where pipes and hand-rolled cigarette• are assumed to contain 1 g of tobacco, then all cigars are estimated to contribute as ~uchRSP indoors as 11 billion clgarettesi while all plpes and hand-rolled cigarettes are estimated to ¢o~tribute as ~¢h as 15 billion regular cigarettes. This increases the estimated total RSp generated in US indoor microenvlronments from all cigarettes, pipes, and cigars to nearly 19,000 metric tons per year. As exemplified by data from EPA's TEAM study, ETS predominates over 114

Draft - Do not cite or c~o~e other sources of RSp indoors (see Fig, 1.) Although the percentage of the population that smokes has declined from nearly 50t in ~le 1960's to about 30t presently (OSN,1988), ~he percent of slokers who are heavy smokers has increased steadily over the pamt 30 years, thus although ~he percentage of s=okers has gone down, the Increase in smoking rate =ay tend to offset ~hat trend towards lowering nonsmoker expesure to ETS.(~C,19S6)° B. Indoor sin transpo~ of ~T8 Nonsmokers are exposed to_~ET5 in indoor spaces. The determinants of these enclosed space exposures Include smoking occupancyt source air-contamlnant e~i|llon charac~eristlcst source use, building oharactBrlstics, space volume, infiltration or ventilation rates, efficiency of air mixing, Burface sorptlon, chemical transformation, and ~he efflclency~ of air clRaning equipment. The interaction of thele variables in determining the resultant concentrations of ETS has been evaluatld In both ¢ontrollnd laboratory settings and in field studies within the theoretical framework of the main-balance ~quation. The mass- balance equation-may be applied to tobacco smoke either as an equilibrium mod~l (tlms-lndependmnt) or as a dynamic model (time- dependent). Dynamic and QquillbrlUmmodels ere useful in laboratory studies; equilibrium models are beet ~ulted to evaluating and predicting ETS ¢oncenhratlons In field studies, particularly when average concentrations over a per~c~ of a workday or longer are of interest. (NRC,1986) • Laboratory and fleld studies typically utilize some form of • single-compartment equilibrium model to evaluate the input parameters of the mass-balance squatlon, to evaluate field study data, and to project RSP concentrations from ETS indoors. These studies ~ave reduced the general slngle-compartment mass-balance equation to the followlng simplified fore: C~ m G[~(Nv + Nt)V]"I {i}, where C~ is the equilibrium concentration of ETS-generated RSP in a space, expressed in units of micrograms per cubic meter (ug/m3), G is the RSP generation rate from tobacco combustion in units of mlcrograms per hour (ug/hr), Nv in the ventilation or infiltration rate in units of alrchanges per hour (ach}, N, is the loss rate of RSP due to surface removal in a Ipace in sir changes per hour, V is the volume of the space in cubic meters (m), and m is the mixing rate (Repace, IARC, 1987) expressed as a fraction. The above model assumes no air-cleanlng devlces~ e~ther in the space or reoirculated air; Leaderer (1984) has given a detailed ~eview of thls model° Under laboratory conditions, these input parameters can be controlled end evaluated. In conducting field studies or in estimating past RSP levels Ind~rs, the val~es on the right side 115

Draft - DO not ci~e Or ~o~e of eq.l have to be detez~ined from available data. This equation assumes equilbriun conditions, and to the extent that any of the generation or removal terms are intermittent (e.g. smoking rate) or variable (e.g. ventilation rate), errors are introduced. (NRC,1986) According to ~he National Research Council (1986) the most extensive use of the mass-balanceequatlon for assessing RSp levels due to ETS in occupied spaces has!been hy Repace and LowTey (1980), who proposed and applied in field observations a condensed version of the mass-balance equation for estimating RSP exposures due to ETS in a variety of indoor mlcroenvlronmsnts. Their model is. C~ - 650 DJNv (2}, where C~ is the equilibrium concBntration Of RSP due to ETS in units o~ micrograms per Cubic meter, D is the density of ctlve smokers [burning cigarettes) observed ~n a space per lOS m~ over the sampling time, and Nv is the ventilatlon or infiltration ratm in ach.(NRC,1986) The constant~ts~m (650) is calculated from a standard set of assumed conditions for smoking fetes, RSP emission rates, mixing factors, ventilation rates, and sink rates. These standard sets of conditions are derived largely from experimental data and building standards. In applying equilhrlum mass balance models such as eq.(2), gathering data on easil~ measured input parameters such as smoking rates or volume can substantially reduce the variablity of the estimated RSP levels. (NRC,1986) Eg. 2 was v,lidated under controlled experimental conditions in real world i.ttlngs, and was found to predict the equilibrium values of ETS within a high degree Of accuracy in exposure chambers using real smokers.(Repace & Lowrey,1980, 1982, Repace, IARC, 1987) ~rther, the predictions of the model were round to be consistent with RSP levels from ETS measured in the field. However, the NRC stated that additional field testing of eq.(2) as well as a better understanding of the variability Of the input parameters was needed.(NRc, 1986) In 1987, Ricker~ st ai.(1987) tested a key theoretical assumption regarding the ratio ~etween the effective and ventilatory air exchange rates in the constant term in eg. 2 (Repace, IARC, 1987) and found that the model explained 87% of the variation between observed and predicted values for RSP concentrations from ETS in their mxperimsnts. More recently, Repace(IARC,19$7) has published a derlvatlv~ of eq.2 which incorporate advances in %understaNdlng. Eg.{2} assumes a steady generation of tobacco smoke, which is generally only valid when 3 or more smokers are present in a space. For less than 3 smokers, it represents an upper bound. It is also l lmited for modeling purposes by being based on the room denslty of active smokers. The derivative equation is based on the room density of habitual smokers (number of habitual smokers per unit space volume). Thus, the presence of an archetypical "habitual" smoker (i.e., one who is assumed to smoke at an average rate of 2 116

Draft - Do not cite or quot~ cigarettes per hour at i0 minutes per cigarette, with an emission factor of about 22 mg of S$ RSp per cigarette) is the modeling parameter rather than the room density of bur~Ing cigarlttss.(IARC, 1987) This derivative equation is given as follows: C~ - 217 D~/Nv {3}, where C~ is the equilibrium concentration in units of micr~rams per cubi~ meter, Dh, is the number of habitual smokers per i00 cubic meters of space volume, and Nv is the number of space air changes per hour. Eq. 3 assumes that tobacco smoke concentration is in equilibrium, which occurs when the rate of generation equals the rate of removal, and the concentration Ii in a steady state. This assumption presumes three or more smokers, since the average smoker smokes three cigarettes per hour and takes ten minutes to smoke a cigarette. This means that with three smokers in a room, a cigarette will always be burning.. (During growth, e~. {3) becomes A,- Ceg {l-eXp-tN~); during decay,A~(tET} - AI exp""v, where T is the smoking duration. (Repace, 1987) A more ~etalled description of the derivation and validation of eg. 3 and the uses and limitations of these models is glv~ by Ropace, (IARC, 1987, chapter 3).) If the number of habitual smokers being modeled is only 2 or I, steady-state conditions no longer apply, and other simple approximations have been suggested, (Repaoe (IARC, 1987) in lleu of using exact time dependent growth and decay models. A one- smoker-approxlmation model proposed by Repacm (IARC, 1987) agrees very well with the instantaneous predictions of an exact computer simulation performed by Rioke~ (1988), hut significantly underestimates newly available fleld data, which represent time- averaged concentrations. On the other hand, by contrast, eq.{3) for less than 3 smokers represents an upper bound to the ETS concentration, and as is illustrated below, produces reasonable agreement with, and provides useful insights into, the analysis of field data. For example: As part of the the HarVard 6-city study of indoor and outdoor air quality, Spengler and collla~ues (1981) collected RSP samples in 55 homes in 6 olties between May 1977 and April 1979. The number of smokers living in each home was recorded. The quantity of tobacco smoked was not reported, nor were the number of hours in the each smoker spent home or air exchange rates measured. The daily average "background," or mean indoor R~P level in the homes of nonsmokers was found to he about 24 ug/~; uslng regression analysis, the authors estimated that~he average i~paot of a single smoker (a composite averaged OVer both sexes) o~ 24-hr average RSP levels from ETS in a rmslde~oj was about 20 ug/m~ above backgroundo On average, two such habltual imokers would make about 40 ug/m~ above background (~4-hr average}, and so fo~h. Added to a background of 24 ug/m~, this yields a daily average RSP concentration for one smoker homes of 44 ug/m3, and for two-smoker 117

Draft - Do no~ cite or guote @ homes of 64 ug/m3. What does eq, 3 predlct? Although the air exchange rates were unknown, Fig. 2 shows a histogram of the frequency of ooourrence of various air exhange rates (called infiltration rates) during the heating season for typical mlddle income housing (R. Grot, personal oommunioation~ for 266 homes in 14 cities around the US in 19?4 (SG,1986) o Ocoupants were asked to keep windows and doors closed duri~ the tests. Under these conditions, the mean air exohange¸ rate found was i.i ~ 0.9 ach (Grot & Clark, 1979), this value~x llkely to he sonewhat lower than a full seasonal average with no restrictions on door and window openings. The s=oker density for a one-smoker home of volume 340 ~ has been given by Repace and Lowrey (1995) as D~ - 0.29 habitual smokers per i00 m~. Similarly, for a two-smoker home, Dhs - 0.54. In the absence of Infor~atlon on air exchange ra~es, let us assume, from Gro~ and Clark (1979), a ra~e, Nv m i.I ~h. Then eq. 3 predicts a value of C~ = 217 x 0.29/;.1 = 57 ug/m" during smoking, for a one smoker home, and 114 ug/m= for s two-smoker home. From table A1 for time budget studies in Repace and Lowrey (1995), averaged over employed men and woRen, and homelakers, the average amount of time spent awake in the home, (allowing for 8 hrs of sleep per day) is about 7.9 hours per day. converting our calculations to a 24-hr average and adding a background of 24 ug/m yields an estimated 57 x 7.9/24 + 24 ~43 ug/m for a one- smoker home, and 114 x 7.9/24 + 24 - 62 ug/m f~ the two s~oker home, in good agreement wlththe values of 44 ug/m and 64 ug/m from ~he 6-clty study above. A oomparlson oft he predictions of sg. 3 with 17 months of RSP data on 55 homes in 6 Cities (Spenqler, et el., 1981) is given in fig. 3. This example illustrates the utility of models in estimating nonsmokers' domestic exposures to ETS. As a second exampl f conslds~ the measured aerosol mass concentration in a ?00 m~ (25000 f~ floor area) office with one smoker (smoking rate not reported), and a measured air exchange rate of 1 aoh (see Fig. 4); the large impact on the office aerosol oonoentratlon caused by smoking is apparent by comparing the daytime and evening RSP oongentrations.(IARC, 1947) The predictions of eq. 3 for ~$ - 1 smoker per 7 hundred ~ublo meters and Nv - 1 aoh yields C - 31 ug/m=; with an 19 ug/m= haokgrgund added, the predicted RSP~level is 49 ug/m~ in good agreemen~ with observations (fig 4). It is clear!from flg; 4 end also from models, that ETS can be very persistent in indoor environments: at an air exchange rats of 1 ach, it takes 3 hrs for 95% of the smoke from i cigarette to be rlmoved (Repacs, IARC, 1987) A third example, where more information is available on the smoking rates, home volumes and air exchange rates, is provided by the data obtained in the NYSERDA study of weekly average residential aerosol concentrations in 141 homes with 118

D~f~ - DO not cite or ~n~o~e smokers.(Leaderer, et el., 1990) The measured weekly average RSP levels were 43 ~g/m3 in the smoking homem, the background levels averaged 16 ug/m in the nonsmoklng homes, the average air exchange rates were 0.54 ach, the average house volu~es 353 m3, the average number of cigarettes smoked per week was 99.3, and the average number of hours of smoking per day is calculated at 7.1 hrs/day assuming 2 cig/smoker-hr (Repace and Lowrey, 1980). From~his data, using eq. 3 as before, we calculate R - 217 x 0.28/0.54 x 7.1/24 + is - 49 ~/m3, in reasonable aqrlemmnb with the observed average of 43 ug/m-~, but higher, as expected. Thus, although eq. 3 for less than 3 smokers represents a simplified upper~bound approximation, it has utility in producing estimatBs which are reasonably consistent with field data, and has the advantage that it is simple to use. However, since none Of these studies were specifically designed for model validation, further comparisons with field data are impotent as new data sets become available. A sampling of whole-building elf-exchange rates in $ large federal office buildings in 7 states with different climates is shown in Fig. 5, and these generally approximate the ASHRAE 62-81 ventilation standards for offices (20 cfm per occupant, equivalent • ~ 0.84 ach, for smoking buildings), on average, although there are some buildings significantly lower. A recent EPA study of air exchange rates in 6 buildings ~(3 new, 3 old) did not show significant differences between the new (.5 ach) and old (.5 ach) buildings' alrchange rates, although for a given building nighttime measurements tended be be lower. ~(Sheldon, st ai.,1987) Recent research has revealed several interesting factors in large office building air exhange. There are many pathways for floor-to-floor air communication, particularly return air shafts, where the existence of such pathways can cause a huildlng~s air exchange characteristics to closely approximate those of a single large open space; it does not require unusual numbers or sizes of openings to create these conditions, (A.Persily, personal communication; Persily and Grot,1986) a condition for which eq. 3 was designed. This implies that ETS may diffuse throughoub a large office building, exposing nonsmokers even in private offices. Nicotine measurements in office buildings support this observation. (Williams, et el., 1985; Vaughn and Hammond, 1989) In summary, limited field te~ts of the general equilibrium - model, in which some of ~he input parameters are measured and others are estimated from eltherche~ber studies or building codes, have predicted RSP levels reasonably well over a wide range of values of input parameters° It is clear that both models and observations based on personal monitoring or area monitors in various microenvironments yield consistent results: RSP levels when smoking is allowed will result in substantial increases over RSP levels in nonsmoking occupancy.(NRC,1985) Co Tobacco Smoke and Ventilation 119

Draft - Do not cite or quote Because of wldesprsad public smokings many buildings are contaminated with tobacco smoke combustion products. Many building owners and managers have assumed that ventilation is a viable control mechanism for the clouds~ of smoke that are generated by cigarettes, pipes, and cigars. ~n particular, the ventilation standardsnz roposed by ASHRAE (the American Society of Heating, Refrigerating, and Air Conditioning Engineers) ere ~ften thought to afford adequate control of tobaccosmoke. However, ASHRAE standards are not health-based s~andardedestgned to limit cancer risk or eye irritation to levels acceptable to nonsmokers. are designed only to limit dissatisfaction with tobacco smoke odor to a maximum of 20% for .visitors. (a test panel consisting of 50% smokers and 50% nonsmokers) to a building where smoking occurs. Currently ASHRAE recommends 20 cublo feet per minute par occupant (cfm/occ) for this purpose. P~oviding ventilation adequate to control cancer risk has been estimated to rsquirs 5400 ofm/occ, an unrealistic ventilation rete.(Repace & Lowrsy, 1985) Air cleaners have three fundamental problems, one, most air cleaners do not scrub gases from the air, end many of the harmful tars appear in the gases. (nrltohard, 1990) Two, air cleaners cannot remove smoke which encounters the nonsmoker before it reaches the air cleaner. Three, even air cleaners which are close to 10o% effective in removing pertlolss which reach them must process hundreds of room air volumes per hour to reduce cancer risk to an acceptable level.(Repaoe & Lowrsy, ISSSr Rspace, i%89a) Separation of smokers from nonsmokers within a space will expose nonsmokers to smoke which diffuses from the smoking area. Separation on the same ventilation system will reduce peak concentrations to which nonsmokers are exposed, but will expose nonsmokers to smoke recirouleted by the ventilation system.(Repace, 1989; Repace and L~wrey, 1987) The foregoing considerations demonstrate that source control, i.e., removing the smoking from the air volu~e containing nonsmokers, is the only viable ~ontrol option, source control adequate to protect nonsmokers takes two forms: separation of smokers from nonsmokers on separate Ventilation systems, or restriction of smoking from the hu~ldlng. {USEPA, 1989) Separation of smokers in a designated smoking area exposes them to much higher levels of exposure to ETS, and may significantly increase an already considerable risk from a~tive "smoklng.(Repace, igssb) D. Measured =onoentzatlons of RTS =onstltuents: RSP: Both field studies (Table l) as well as chamber studies have demonstrated that ,t~bac=o combustion has a major impact on the mass of suspended patriciate matter in occupied spaces in the size range <2.5 um, defined here as RSP. RSP is a major component of ETS. Even under conditions of low smoking rates~ easily measurable increases in RSp have been recordRd abovm background levels. The term RSP, however, encompasses a broad 120

Draft - DO not cite or quo'c:e range of particulates of varying chemical composition and aizl emanating from a number of echoes (outdoors, cooklng indoorS, kerosene heaters, eto.)(NRC, 1986) The apportionment of RSP indoors depends primarily on the presence of these other sources. However, in western society, there are few Indoor sources generating ¢oncentatlons which approach in strength those due ~o ZTS. There appears to he littli varlability between brands cf cigarettes or tobaccos for RSP e~.asions, although cigars wii~ produce greater emissions than cigarettes. Thus, it may be inferred from Table 1 that from a comparison of smoking and nonemoklng buildings, the bulk of the RSp found in buildings where ~here is smoking is due to ZTS. For example, by comparison of ~he data of First (1984), Leaderer, et al (1986) and Repaoa and Lowrey (1980,1982) for a total of 42 ~cklng buildings and 21 nonamoking buildings, ~e weighted average RSP level in the smoking buildings is 262 ug/m , while in the nonsmoking buildings it is 36 ug/m', suggesting that about 85% of the indoor RSP levels in those buildings is due to ETS. Most of the buildings involved were pUblic access buildings. Hammond at el. (1988) measured personal exposures to RSP i~ several htt~dred railroad workers. Mean calculated ETS-derlved RSP exposures for railroad office workers averaged over 90 ug/m3; by comparison, all other sources of RSP for these dieseloexh&ust exposed workers averaged 39 ug/m3, The U.S. Department of TSanspor~atlon (1990) measured concentrations of RSP in the smoking section of a random sample of 69 smoking and 23 nonsmoklng flights. Nonmmoklng flight attendants must work in the smoking sections on aircraft. Levels of RSP on the smoking flights • averaged 175 ug/m3, whereas measurements In the same section of t~e aircraft on nonsmoking fllgb~s averaged 35 to 40 ug/m3. ° Figure 6, (IARC, 1987) a plot of the data of Repaoe and Lowrey (1980, 1982) illustrates the large:impact of smoking on RSP levels; smoking data points *A' thru ~T' encompassed a wide variety building mioroenvironments,inoludlng I0 restaurants, 3 cocktail lounges, 3 bingo games, 2 dinner-danoe halls, 1 bowling alley, 1 spor~s arena, i hospital waiting room, and a residence during a dl~er party. Studies of the dispersion of RSP from ETS in US homes showed at most a factor of 2 difference among various rooms in resldenoes~ averaged over 24 hrs. In a setting such as a work 8nvlronment, where the average exposure is several hours, ETS would be expected to disseminate throughout the airspace where smoking is occurlng.(SG, 19B6) Although most people spend approximately 90 percent of their time in Just bwo mloroenvlronments (home and work), important exposures CaD also be encountered in other mioroenvlronents, e.g., in transit, whloh accounts for 0.5 to 1.5 hrs per day for most people..(SG, 1986) Exposures on aircraft can also be considerable. (Repaoe and Low~ey, 19881 USDOT, 1989) BENZENe= Wallace (1989) in reporting the results of EPA.s TEA~ study with respect to the benzene exposure of the population, found that ETS was a significant source of population benzene exposure, accounting for about 5% of total nationwide exposure. Wallace 121

D~aft - Do not cite or ~t~ots reported that workplace exposures for nonsmokers not exposed to ETS at home, but who report being exposed to ETS more than 80% of the tlme at work showed significantly higher benzene concentrations than those who report exposure to ETS less than 50% of the time. Wallace estimates that ETS-benzsne mxposurss are about equal at home and at work. NICOTZNE~ Leaderer and Hammond (in press) measured vapor phase nicotine and RSP concentrations in 96 residences, Vapor phase nicotine measurements were found ~o be closely related to number of cigarettes smoked and highly ~predlctlve of RSP generated by tobacco combustion. The mean RSp background in the absence of measurable nicotine was found to be 15.2 ~ 7 ug/m3. The mean RSp value in the presence of nicotine was 44.1 Z 30 ug/m3. Weekly mean nicotine concentrations in the residences was 1.1 ug/m3. stillman et el. (in press) measured weekly average nicotine concentrations using the method of Leaderer and Hammond (above), in 9 (Y. Stillman, pars. comm.) univsrsltyofflcss. Concentrations averaged 2.1 ug/m3. After a smoking policy was implemented, the nicotine levels decreased by 95%. Vaughn and Hammond (1990) measured weekly average nicotine concentrations in offices in a modern office buildlng using both active and passive samplers. Before.the smoking control policy, nicotine vapor concentrations at nonsmokers, desks were about 2 ug/m3, and were reduced by 95% after a smoking policy was implemented, in good agreement with the findings of Stillman, above. Hammond et el. (1988) measured nicotine and RSP in two employee smoking lounges at the University of Massachusetts. RSP levels varied between 220 add 350 ug/m3 during smoking, with associated nicotine levels from 40 to 70 ug/m3. After charcoal-filter air cleaners were installed, nicotine levels were virtually unchanged, add RSP levels varied between i00 and 310 ug/m3. A study of personal exposures to airborne nicotine in 4 US office workers showed about a 0.i mg me4n exposure (mean personal nicotine concentration of 15 ~ 9 ug/m~, daily workday average, times a 0.8 m~/hr inhalation ra~e times an 8-hr workday). The nonsmokers were exposed to the smoke of a co-worker who smoked 9 cigarettes per workshift, about half the rate of the average US smoker. (Hammond et ai.,1887) I. Zxposure of nonemoklng populations to ETS In the general population (both sexes) aged E 17 years in 1980 (160,798,000 persons), a majority has smoked at some time: 32.6% were current smokers, and 21.3% were sxs~okers, while 46.1% had never smoked (see Table 2). Among current 1980 smokers, 53% were male, and 47% were female, with~some race- and gender-speciflc differences : whi~e males, 35.9%, black males, 42.0%; white females, 29.3%, black females, 29.7%. (R. Wilson, NCHS, personal communication) In terms of the population at risk, both lifelong 122

Draft - Do not cite or ~.~ote nonsmokers and former smokers, 66.7% of the adult population and the overwhelming majority of children are potentially at risk from involuntary exposure to ETS (in 1970, Bonham and Wilson (1970) found in a national probability sample of children, ~hat 62% of US homes with children contained 1 or more smokers). EXposure to various subpopulatlons or individuals, however, may vary considerably. For example, the prevalence of smoking among su~oups of ~he population who ~oeorlbe smoking on religious grounds (such as Moz~ons and seventh-DayAdventists) is much lower: for example, in 1980, only 1.78 Of Seventh Day Adventist men and 0.5% of Seventh Day AdVentist women reported current smoking, although 35% of the total were exsmokers. The in=idence of lung cancer -- a disease for which ~he majority of cases occur in smokers -- among SDAs is 218 of ~hat in the general population; Thus, SDA homes would be, in general, e~1~e~ed to be E~-free. The mlcroenvlronments of Importance for exposure to ETS will be those where the population spends the bulk of its time. As Table 3 (Oft, 1981) (based on 1972 data} shows, employed men spend an estimated 56% of their time at home, and 28% of their time at work, for a total of 84% of the time at home and at work; employed women spend 64% of their time at home, and 22% of the time at work, for total of 86% of the time at home and at work; while homemakers spend 85% of the time at home. When time spent in other peoplest homes and in non-work places of business are added in, the • population averages about 88% of its time in homes and workplaoes. These sites, therefore, must, on average, predominate as potential sites for exposure to ETS for the~ general nonsmoking population. A UK study of exposure to ETS in 20 uonsmoking men whose wives smoked showed that 78% of the men's reported hours exposure came from outside the homel by contrast, %0% of the ETS exposure of i01 nonsmoking men whose wives did not smoke was reported to come from non-domestic microenvironments. (Table 4). Since the second largest source of time spent by men is in the workplace (Table 3), this suggests the workplace may be the major source of exposure for nonsmoking men. Cummings et al (1990) studied the prevalence of exposure to ETs in 663 never- and ex-smokers who attended a cancer clinic in Buffalo, N.Y. in 1986, by questionairre and urinary cotlnine level. 76% reported exposure to ETS in the 4 days preceding the interview, while 918 had detectable urinary ootinine levels. Reported exposure locations in order of frequency were workplace (28%), home (278), restaurants (16%), private social gatherings (Ii%), in transit (I08), and in p~lic buildings (88), (total of 1008). 77% of subjects reported being exposed to tobacco smoke at work, while 22% of the subjects lived with a smoker. In a second study, Cummings et el. (1989) reported on 380 never smokers from the same study: A total of 878 reported exposure to tobacco smoke at work. 24.3% of the men reported spousal smoking, whereas 84.7% reported workplace exposure; significantly, when asked to rate the severlny of exposure, on a scale where spouJal smoking was normalized to i, 123

D~aft - Do not cite or quote severity of workplace exposure was rated S. 66% of women reported spousal smoking, whereas 83.5% reported being exposed to smoking in the workplace. The women studied rated theseverity of spousal smoking at about 30% higher than workplace exposurls. Coultas et el. (1990), in a pilot study of 15 nonsmokers in Albuquerque, N.M., exposure q~estlonnaires and saliva~ urine, and personal air samples were obtained pro- and post wcrkshlft. Nicotine and ootinine levels were quantified, as were atmospheric nicotine and RSP samples. Statistically significant correlations were obtained between RSP and nicotine and total reported hours of exposure; between nicotine and total number ofsmokers, total hours of exposure, and postshift urlnazy ootlnlne~ between urinary cotlnine and total hours of exposure; and between salivary cotinlne and total number of smokers. Objective evidence of exposure to ZTS was obtained in various workpleces. Spengler et al.(19B5) and Sexton st el. (1984] demonstratedby personal monitoring of RSP and the use of time-activity questlonDalres that exposures to ETS both at home and at work are significant contributors to personal exposures. A survey of exposure to ETS in a California population subscribing to a health-maintainanoe plan indicated that 63% of nonsmokers surveyed reported exposures to tobacco smoke (Yriedman, 1983); this occurred despite the fact that in the 19S0's California has has been in the forefront of restrlatlons on smoking in public, with 44% of its population currently living in communities that have enauted workplace smoking restriotlons.(sG, 1986) Garfinkel [1981), in a study of 176,000 nonsmoking US women (1960-1972), found 72% had smoking husbands. Kabat end Wynder (1996), in a recent study of 219 sixty-year-old US women nonsmokers, found that 65% were exposed at home and 67% reported exposure at work, averaged over adulthood. Studies~of the concentration cf nicotine and cotinine in the body fluids of nonsmokers report similar rlsults (Table 4); Jarvis & Russell, (1994) showed that in a study of about 100 UX nonsmokers, only 12% of the subjects had undetectable cotinine levels. Moreover, in the latter study, surprisingly, nearly 50% reported no exposure, suggesting that ETS permeates indoor atmospheres to such an sxtent that many nonsmokers are unwittingly exposed. This is borne out by a study of 46 US infants, 40% of whom were reported by their mothers to be unexposed to ETS, but only 20% had undetectable urinary cotinine "levels.(Greenberg, 1986) In a third UK study (Weld, 1986) of urinary cotinine in 221 nonsmokers, the 20% who reported DO exposure had mean urinary cotlnine levels which were 21% of the remainder of the group who reported exposure. The foregoing illustrates that exposure to ETS is very widespread in the population, even among those nonsmokers who believe themselves to be unexposed, however it tends to be greater in those who say they are exposed at home, possibly Indicatlng a greater tolerance for ETS among men with nonsmoklng spouses. 124

~aft - ~ noC cite or ~ots Although there have been numerous measurements of ETS concentrations in various indoor settings, these data do no~ represent a comprehensive description of the actual distribution of ETS exposures in the US population, However, additional data on the distribution of smokers in ~he nonsmokers' environment as well as ~he distribution of ETS levels in that environment, are needed in order to characterize ~he actual ETS exposure of the population. In ~hs absence of such data, population exposures can be estimated by models or by extrapolation from biological markers from existing studies. (SG, 1986; IARC, 1987) In summary~ exposures to ETS can he assessed by personal air contaminant monitoring, modeling of concentrations based upon air sampling, time-actlvity patterns, and questionnaires, or upon biological markers. The two best ~ethods at present are based upon ~he hlological markers, nicotine and its metabolite, cotinine, which are present in the saliva, plasma, and urine of active and passive smokers, and upon atmospheric markers su¢h as nlco~ine in the vapor phase and RSp from the particulate phase of ETS, the latter of which has been used in J any field studies because of the substantlal emission of RSP from tobacco combustion. In US, ETS ~s generated by 50 million smokers, who smoke the equivalent (including pipes and cigars) of 610 billion cigarettes annually. Although the number of smokers has been declining, the percentage of heavy smokers has been increasing. There are nodels in use, based on the mass-balance equation, and validated boch under laboratory and limited field conditions, which can predict the concentrations of RSP fronETS to a reasonable degree of accuracy. Application of such models, together with field studies of RSP concentrations and sociological studies, has suggested that exposure to ETS is very widespread in the population. Environmental tobacco smoke is not readily controlled by either ventilation or air cleaning. Field studies of RSp in buildings where smoking ocCUrs suggest that RSP from ETS contributes 80 to 90 percent of the particulate load during the period of smoking, and that it persists for long periods after smoking ends at typical building air exchange rates, thus prolonging nonsmokerst exposures. Available data suggest the workplace as a significant site of exposure to ETS. F. Integrated exposure analysis Exposure to ETS can be quantlfied either by atmospheric or biological markers. Of the lat~er, expired carbon monoxide, carboxyhemoglobin, plasma thiocyanate, plasma, urinary or salivary nicotine, and plasma, urinary, or salivary ootlnine have been used to evaluate exposure to ETS. However, successful attempts to quantify the degree of exposure have been limited largely to measurements of nicotine and cotlnine. Urinary nicotine is a sensitive indicator of recent ETS exposure, while cotlnlne appears to be the short-term marker of choice for epidemlologic studies. Nicotine and cotinine are the best markers currently available. 125

Draft - Do not cite or ~ote Levels in body fluids may be elevated i0 or more times in the most heavily eA~osed groups compared with the least exposed g~oups. Mean levels of urinary nicotine and Ootinine in body fluids increase with an increasing self-reported ETS exposure and with an increasing number of cigarettes smoked per day by active smokers. (SG, 1986) Coghlin, Hammond, and GarLq (1989), in assessing current weekly ETS exposure in 53 normmoking volunteers by personal ni=otlnemonltors, diaries, and q~estionneires, found that~he best predicto= of total nicotine exposuro was given by the formula hap: the number of hours of exposure (h), times the number of smokers (s), times the proximity of those smokers (p), accounting for 83% of total exposure. A significant finding was that exposures derived from social situations (e.g. restaurants, bingo games, bars, and bowling alleys) (which are workpiaces for some persons) may contribute significantly (34%) to to~al exposure. Nicotine is found in measurable concentrations in the saliva and urine of most urban nonsmokers, and is present in higher concentrations in those with some recent exposure. Estimating the magnitude of the passive smoking dose is difficult, and it is of doubtful validity to extrapolatei~rom the uptake of one marker to another. Over a period when one cigarette equivalent of carbon monoxide is absorbed, the dose of nicotine appears to be only between i/I0 and i/3 of a cigarette equivalent. Similarly, under extreme conditions of indoor polletion, it has been calculated that a nonsmoker would inhale volatile nltrosamlnes equivalent to i0 nonfilter cigarettes or 35 filter ulgarettes.(Hoffmann, IARC, 1987) The average concentration of =otinine ~in the blood of habitual smokers is about 300 ng/ml, and is =aloulatad to represent the consumption of about 36 mg of nicotine per day. Onthis basis, and on the assumption that formation of cotlnine from nicotine and clearance from the body does not differ substantially from smokers to nonsmokers, present data suggest that average urban nonsmokers (in the UK) take in 0.2 mg of nicotine per day.(IARC, 1987) [.2 mg represents .6% of the smokers~ dose] The highest plasma cotinine oonoentrationobserved in a nonsmoker corresponds to an approximate maximum dose of 2.5 mg of nicotine per day, i0 times higher, and 7~ of the average smoker's dose. Recent studies of salivary Cotlnine in schoolchildren in the UK showed, in the case where both parents smoked, average concentrations Just over 1% of the levels seen in heavy cigarette smokers.(IARC, 1987) Although the ratio of nicotine to other tobacco smoke oonstituents differs in MS and SS smoke, nicotine uptake may still be a valid marker for total ETS exposure. Nicotine uptake in nonsmokers has been estimated in terms of cigarette equivalents from various studies to vary between i~6 to I/3 of a cigarette per day. The NRC repoz%$ various estimates of cigarette equivalents based upon cotlnine in nonsmokers vanglng from 0.1 to I cigarette per day, and utilizes a ratio of urinary cotlnlne in ETS-expcsed nonsmokers (25.2 ng/ml) to that in active smokers (1826 ng/ml) 126

Draft - Do not cite Qr %-uote passive smokers, reasonably consistent with estimates based on doses from nicotine and co~inine, above. Table 6 gives estimates of the probabillty-weighted exposures to ETS for US nonsmoking adults at home and at work, the two most- frequented microevlronments.(Repace and Lowrey, 1985) Table 6 is derived from RSP Qoncentrstion m~nlln~ based upon Eg.'s 2-5, and from assessments of exposure prohahil~ty based on a limited national survey of top management and health officials concerning prevalenoe of smoking in the workplace in 3000 US corporations, large, medium, and small (29% response), and a national probability sample of the prevalence of smoklnq in homes wi~h children (used as a surrogate for all homes). ExPosure probabilitles ware a weighted average taken over the number of workers in white-collar and blue-collar occupations, and including the different exposure probabilities for white and blue Collar workers. Air exchange rates and building occupancies were" taken from ASHRAE Standard ventilation rata tables for whi~e-oollar workplaoes (which were used as surrogates for blue-collar workplacas). . Table 6 estimates average the workplaoeETS exposure probability at 63%, and the average estimated domestic ETS exposure probability at 62%, where the focus was on estimation of ETS exposures in the 1950~s to mld-1970~s~ since these exposures were held to be of primary signifioanue for the studies of passive smoklng and lung cancer, given the long latency for lung cancer. Comparison of these exposure probability estimates to adult llfe ~TS exposure histories ~aken by Kabat and Wynder (1986) for 215 60-yr old female nonsmokers, 65% at home and 67% st work, shows good agreement. Table 6 estimates a 0.45 mg/day RSP exposure for nonsmokers at home, (weighted for male and female tlme-actlvlty parts differences, and tot respiration rate) corresponding~o a 19 ug/m~ 24-hr average, in good agreement with results (19 ug/m~ per smoker) published in the 6-City study.(Repace and Lowrey, 1985) Table 6 also estimates a 1.82 mg/day RSP exposure for workers," (again weighted for male-female time-actlvlty p~tterns and for respiration rate) corresponding to about a 230 ug/m= workplace concentration, using ASHRAE Standard 62-73 for workplace occupancy and performing a weighted average workday for the different hours worked by men and women (Rapace and Lowrey, 1985)° This is In3good agreement with the weighted average concentrations (262 ug/m ) reported for ETS in public access buildings. Riholi et el. (1990) in a 10-country oollabQrative study of exposure of nonsmoking women to ETS, examined the relationship between smoking by spouse and urinary cotinlne levels as an indicator of exposure to ETS. Riboll et ai.(1990) found that ootlnlne values were significantly higher for women exposed to ETS from the husband than from other sources~ they also found that questionairres in epidemlological studies based upon self-reports of spousal smoking in fact identified a most-exposed populatlcn. A clear increase in urinary cotinlne levels was found from the 128

Draf~ - Do no~ cite or ~ote women who were exposed neither at work or at home, to women who were exposed bo~-~ at work and at home, as suggested by ~he work of Repace and Lowrey, above. In StL~a~, exposures to ~S can be assessed by persQnal air or area contaminant monitoring, modeling of exposurem, or by biological markers of ET5 ¢ontamlnants in body fluids. Using either the biological markers such as ootinine or the athospheric markers such as RSP produces a consistent assessment of ETS exposure, i.e., of the order of 1% of that in smokers. The most- exposed individuals appear to have levels about ten times higher. Based upon limited data, the typical nonsmoker appears to carry a daily body burden of about 0.2 milligrams (ng) of nicotine. The cotinine-based estimates have the advantage that they reflect actual dose of an ETS constituent. They have the disadvantage that they do not reflect a wide distribution of target populations; are based mostly on UK ETS exposures, and may substantially underestimate exposures to other constituents of ETS. The RSP- based estimates have the advantage that they are model-based, can be used to estimate exposures in a variety of mlcroenvironments, represent the great bulk of ETS carcinogens, and can he compared wi~h atmospheric measurements of RSP. They have the disadvantage that they do not represent whole smoke exposure, and do not reflect absorbed dose. The greatest source of uncertalnty is that neither cotinine nor RSp measurements are based on a national probability sample, and on an absolute scale, represent a limited amount of data. Nevertheless, the NRC(1986), the SG(1986), and IARC(1987). have found this data base acceptable for exposure assessment purposes. Estimates of the adult nonsmoking population's exposure to RSP from ETS suggest that the range of exposure is from 0 to 14 mg per day, with the population average put at 1.5 mg pe~ day, where the peak-to-mean ratio is about a factor of i0, consistent with the biomarker-based findings. Summary I. Nonsmokers~ exposures may be assessed by mathematical modeling, as well as by biomarkers such as nicotine or ootinine in body fluids or atmospheric indicators such as nicotine or RSP. 2. Despite limitations of the data base, mathematical models, hlological and atmospheric markers have produoed~ reasonably consistent assessments of nonsmokerst ETS e~pos~te. 3. E~osure to environmental tobacco smoke is inadequately controlled by ventilation, air cleaning, special separation withi~ a space, or on the same ventilation system. 4. Data indicate that ETS is a significant indoor pollutant of buildings, typically representing 80 to 90% of particulate indoor air pollution during smoking, and that nearly all nonsmokers carry a significant burden of tobacco combustion products in their body 129

fluids. Draf~ - Do nc~ ¢it8 or quo~8

I~aft - DO not cite or quote Bonham, G; Wilson, RW (1981). Childran.s health in families with cigarette smokers. A~er. J. Public Health ?I: 290-293. Cogblin J, SK Hammond, PM Gann..Development of spldemiologic tools for measuring environmental tobacco smoke exposure.. Amer. J. Epidemlol. 130:606-704 (1989). Coultas DB, JM Samet, JF McCarthy, ~D Spengler. "A personal monitoring study to assess workplace exposure to environmental tobacco smoke." AJPH 80:988-990 (1990). Cummings ~M, SJ Markello, MC Mehoney, JR Marshall. "Measurement of lifetime exposure to passive smoke." Am. J. Epldemiol. 130: 122- 132 [1989). ~ings ~, M Mahoney, AK Bhargava, PD McZlroy, JR Marshall. "Measurement of current exposure to snvIEonmental tobacco smoke." Arch. Env. Heal~h 45:74-79. ~irst, MW (1984). EnvironmentaI tobacco smoke measurement: retrospect and prospect. Eur. J.Respir. Dis. 5(Suppl.):9-16. Garfinkel, L (1981). Time trends in lung cancer mor~allty and a note on passive smoking. J. Natl. Cancer Inst. 66~i061-i060 Greenbur~, RA; Ealey, NJ; Etzel, RA; Lode, FA. Measuring the exposure of infants to tobacco Smoke: Nicotine and cctinine in urine and saliva. New England J. Mad. 310: 1075-1070. Grct, RA, & Clark, RE (1986). 170-194. Measured air infiltration and ventilation rates in 9 large office buildings. ASTMSpec. Pub. 904, Ed. M. Trechsel & P. Lagus, ASTM, Philadelphia 151-103. Haley, NJ; Colosimo, SG; Axelrad, CN; Harris, R; Sepkmvic, DW. Biochemical validation of self-reported exposure to environmental tobacco smoke. Environment~l Research 49:127-139 (1989). Hammond, SK; Leaderer, BP; & Roche, A. [1987). Collection and analysis of nicotine as a marker for envSronmental tobacco smoke in personal samples. Atmos. Env. Hammond SK, TJ Smith, SR Woskie, BP Leaderer, & N Eettinger. • Markers of exposure to diesel exhaust and cigarette smoke in railroad workers. Am. Ind. Hyg. Assoc. J. 49:516-522 (1988). Jarvis, M,7; Russell, MAH; Yeyerabend~ C; Eiser, JR; Morgan, M; et al (1985). Passive exposure to tobacco smoke: saliva cotlnine concentrations in a representative population sample of nonsmoking school children. Br. Mad. J. 291: 927-929. 131

Draft - Do not cite or ~ocs Leaderer, SP; Cain, WS; Isseroff, G; Berglund, LG. (1984). Ventilation requirements in buildings. II Particulate matter and carbon monoxide from cigarette smoking. Atmos. Environ. iS: 98-i06. Leaderer B, Koutrakls p, Sriggs S, Rizzuto J. Impact of indoor sources on residential aerosol oonoentratloneo Proc 5th Znt Conf on Indoor Air Quality & Climate, V 2, 269-274. Toronto 29 July- 3 August 1990. Leaderer, BP (1990). Risk Analysis i0:19-26 (1980). Leaderer, SP, and Hammond, SK. "An evaluation of vapor-phase nicotine and resplrable suspended particle mass as markers for environmental tobacco smoke." Environmental Science & Technology, in press. Lofroth G, Burton RM, Forehand L, Ha~ond SX, Sella RL, Zweldinger RB, and Law,as J. Environ. Sol. TeCho 23:610-814 (1988) . Matsukura S., et el. Effects of environmental tobacco smoke on urinary cotinine excretion in nonsmokers -- evldenoe for passive uoking. New England J. Mad. 311:828-832 (1984). Miesner EA, Rudnick SN, Prelle~ L, Nelson W. Particulate and nicotine sampling in public facilities and oEflces. JAPCA 39; 1577-1582 (1989). National Research Counci~ (1986). Environmental tobacco smoke -- measuring exposures and assessing health effects. National Academy Press, Washlng~on, DO. office on Smoking and Health (1988). Estimates of the mortality from smoking. Centers for Disease Control, Washington, DO. Oft, WR. HUman activity patterns: A review of the literature for estimation of exposure to air pollution. U.S. Environmental Protection Agency, Washington, DO. Pritchard at el. (1990). Repots, JL (1989a) Smoking in the workplace: Ventilation. S~oking Policy Questions and A~$wers, |8, NatiOnal Cancer Znstitute, Bethesda, MD. Repaca, JL (1989b) Protecting workers from the threat of secondary smoke. Indoor Pollution Law Report, Cadwallader, Wiokersham, & Taft, Leader Publlcations, Washington, DO. Repace, JL, and Lowrey, AH (1980). Indoor air pollution, tobacco smoke, and public health. Science 208: 464-472. Repace, JL, and Lowrey, AH (1982). Tobacco smoke, ventilation~ and 132

Draft - Do not =its or quote indoor air quality. ASHRAZ Trans. 88: 894-914. Repace, JL, and Lcw~ey, AH (1985). A quantitatlve estimate of nonsmokersj lung cancer risk fro~ passive smoking. Environment Inte~atlonal Ii: 3-22. Repace, JL, and L~w~ay, AH (1990). Rebuttal to Lee/Katzenztein comentary on passive smoklng° EnvirorL~ent International, in pregB. Riboli E, Preston-Martln S, Sara=ci R, Haley NJ, Trichopoulos D~ Bather H, Butch JD, Fon~ham ETH, Gao YTa Jindal SX, KOO LC, LeMarchand L, Segnan H, Shimizu H, Stanta G, Wu-Willlams AH, and Zatonski W. Exposure of nonsmoking women to environmental tobacco smoke, a ten-country collaborat~ve study. Cancer Causes and control, 1:243-252 (1990). Eiokert, WE; Robinson, JC; Colllshaw, NE (1987). A study of the gro~h and decay of cigarette smoke NOx in ambient a%r under controlled conditions. Environ. Internato iSz 399-408. ~icke~t, WE; and Labstat, In=. (1988). Some considerations when estimating exposure to ETS with particular reference to the home environment. Canadian J. Publ.Hea1~h 79:E33-S37. SeX, on, gl Spengler, JD~ Triet~an, RD (1984). Personal exposure to respirable particulates: a case-study in Waterbury, Vermont. Atmos. Environ. 18: 1385-1398.- Surgeon General (ISaS), The Health Consequences of involuntary smoking. U.S. Dept. of Health & Human Services, WAshington, DC. Spengler, JD; Treitman, RD; Tosteson TD~ Mage DT~ and Soczek ML (1985). Personal exposures to respirable particulates and implications for air pollution epldemiology. Environ. Sci. & Te~nol. 19:700-707. Stillman FA, DM Seeker, RT Swank, et el. "Ending smoking at the • ohns Hopkins medical institutions: an evaluation of smoking prevalence and indoor air pollution. JAMA (1990), in press. Tobacco Institute (1987). Tobacco industry profill, ~987. Washington, DC. U.S. Department of Transportation. U.S.: Department of Transportation Study of Airliner Cabin Air Quality (1990). u.s. Environmental Protection Agency, Indoor Air Facts #5, Environmental Tobacco Smoke, Washington, DC, 1989. Vaughn WM & SK Hammond. "impact of 'Designated Smoking Area' Policy on Nicotine vapor and particle concentrations in a modern office 133

DEaf~ - DO no~ cite Or q~ots building. ~.Air Waste Management ASSOC. 40:I012-i017 (1990) Wald, NJ; Nanohanai, Kr Thompson, SG; Cuckle, HS (1986). British Msd J. 293:1217-1222. Wallace LA. "Major sources of benzane exposure." Environ. Health Persp. 82:165-169 (1989). Williams DO, Whltaker JR, Jennings WG. Measurement of nicotine in building air as an indicator of toha==o smoke levels. Environ Heal~h Persp 60:405-410 (1985) 134

Dra~'c - Do nc~ :~:e or quc:e FTGURZS ANI) T~BLZS FOR CH~I~Z~ 6 135

T~,BLZ 1 (h'lt¢, 2SlE)=ega~ - Do n=¢ e~¢e Qr ~o~I Particulate Levels Measure;:1 in Indoor ]~nv~ronn~nts, Including Smoking ;lad Nonsmoking Occupancy Cm¢~m Type ~ Velame. vtadhrrim Mmltee;q Ma,t I;'ee~), s.~ hwmm Oer~.n~ # .ryee/e.m ~ MI/,a~ Cemw. B~tmdt ;~e~q at~ 4 n~i4w.e~ N$ -- N/-- (]/2am ~S (20-qO) TSP, r~eal melu~ m lle~, l~2 0.2 ml ? ~ S s I ~ HI-- G/2 ~ t2~ {6~.2.,(0) TSP se~mwit y 14 ve~dt.ae~ S ~ 2 -- N'-- O;2 m ]~2 {(~-340) TSP ~sitb*tW I ~niele~ $ = 3 -- HI-- GZ2 me 33S {--) TSP lea ~ivJ~ O,~I~ -- -- -- G/2 m -- i41-D) | t Ir*et~s $ " $-~ -- N,M/I..~ Idt G~h 44~ (~.13"q~b) T~P vlmiiit iota N$ a 5-260 mimt~d T " ]0-3~O BI~o4 and 3 arrm, N$ -- -- G/~4 h 55 (42-92) "13P 14,2"?? Fret. 19~4 I tchoel NS M;-- P/-- ~0(--) TSP S pabik $ N,M/-- Pt-- ~4 (40-6(~) TSP b~ildlnrt Hlw~tm.l~ IIr~*i(L~s HS 1~-:~74 M/O.I~'O.~ OCMII.S-ISmin g"40(--) ~P.*incL~*/$em'ncr- I~E4 It ;'uidt.'nc:~ r45 l.~4)-~74 M/0.26- I .~ Q~t41/~- IS mm 12-4b limP, wlnt er/tumm~ (m~ 6 h) ~rees" 2 iva~du.tes $ I.~*-674 MI0,]~oI.4? I~MII~- I~ rain ~-106 RSP, ~nwr/~mm~-- (~s,~ e h) aevnm" + cq, L~md~.ef ' 3p~b[k NS 1~.*1,~6 MIO-~?-$.~ Gl~t-]lh ]TJ~(9,l.32.~ "r~P*mlpntmetsttm' a d., bulidtelp ill ",m~'. ~t,~ul 7 ~lbl~ L7,-4,37t i b8.~O) M 10,T7o?,,$3~ G/2-24 h ;Z~* 1 (~t-4~2) . M4ISUlId ~tlimU111. l~iidtlt~l T I 2,,~ (l~,0 pelk] Cuddeback el tl.. 1976 Me~fll~dreas Ou I~c~rs -- G/24h ]7,~ {--) P~P, TSP also t~ aL I~1 Immn~ 2 etfk'~ -- -- -- G/24h le.8-20.2 RSP, TSP aim (53 ~lk) m~ed $ ft*~eiK~ N5 -- N/0.~-1.3 I~h G/24b Tq,4-4,01 pSp, "rsp aim T a, 2-6 1111~9 p~k) mu~'d t~dinces 5 -- N/D.~1.1 mh G/24 h 36,q-~9,9 RSP. TSP ilte Nh~l~e Out~ -- -- -- G/IMh 11.3 =~ ~.0 (t-~) RSP • II,. Iqe3 L9 rr.ldem~l NS 31~-).021 NI-- G/IM h I~.l) :~ 22.~(b-~) RSP, ~ m~Hm'et. mft~ mix" II ~ S 2~-~ N/-- G/IM h ~9.2 ~ 2~.$110-1a4b RSp. repell mel~ur~, i~1 ft.e mitr

D~'att - Do no~c ¢:Lte or ~1~o~.o TJUJ[.]I i. ¢on~,.d. (~R¢, 19~16) C0nxm:aed Cm at YeJg~. V*~t ~tti~t ~ Maa (ranp). Iq~4 TI3 n~c'a(~s $ J L.2 -- K/0.2.O.~ ~Jt D/24 h 10-46 (--) "]'~ .D T " 3-4 Itmwe ~ O~deau a P/2 n 42.9 (22-63) RSP. aw:ra ~ Loeb. 2-ram ~mpJ~ I~0. 19~2 ~ Publk= 0* L1-3..S4j -- )4/~ p/2 mi~ 2"~1 (86-1.140) R.SP. awq~p ~r ~lee ~t ,d.. Oulcl~n -- -- G;24 h 1";.0 :~ t .b (6-23) RSP. rrpell ump~e~ 24 ~ NS* -- N/-- G,'24 h ~.~.0 ± 1.0~]3-b3~ US~ ~repb~'es 5¢~11 ~" O~ I~ors -- -- ~ G.~4 h 21.| 2:11.9 (~) R.SP, ;'~'~11 n:m,.u m • 1~.. IMI 3Stem i<S ~ N/-- G/24h ~4.t :~ II.b(~) RSP,~almr~ I$ i $ ~ ~ ~ ~/~ G/24 h ~6.~ ~ 14.$ (~) PAP. r~t mm $ residences $ ~" 2 -- ~-- 0/24 h ~O.4 =: 42.9 (o) RSP, ~l~aT m®~urn S~n|l~ Out~ers -- -- N/-- G;24 h 18 ~ 2.] (--) RSp, ~.eal me;~u r~ • ~IS ~:sidcnc~ $ -- G/24 It 74 ~ ~.~ (~1 RSp, ~peal a~a~ure; S~C~nl~ and I ~f~ $ w. -- -- C,4 ?)~ -- ]5.~ { ]S-3(~) TSP ~. 1~.3 22 ed~,aet $ w GI?~/-- .11.7 I--) TSP U.$. Dep~nmenc S demes~k $ ~ Ml-- G/1°1/4. I~ i~.n (--) ~SP ,ram. Iq';I 2D m31~ $ -- M/-- G/b-'; h <~10-12~ (~) "~3P pflmes T i Zt6.21q web~ and 44 ~ S ~ N.M/~ P/2 O 1.33 -~ 130 RSP. mh~us F'e.cn~r. IqS0 • i~(~) (962 ?eakl ~krr~nd k.~el

Table 2. Number of persons 17 years and over, by olgaret~h smoking status, race, Bex, and ago: All races' Both Sexes 17 yrs. 17-24 years 25-44 years 45-64 years 65 years & eve: Male ~ 17 years 17-24 years 25-44 years 45-64 years 65 year & over Female_> 17 years 17-24 years 26-44 years 45-64 years 65 years & oyez White> 17yrs. 17-24 years 25-64 years 45-64 years 65 years & owe: Male Female slack> 17yrs. 17-24 years 25-44 years 45-64 years 65 years & ovel Male Female Total 160798 32176 61042 41556 24024 75970 15699 23549 20810 9891 84820 16472 22725 31472 14133 139036 2709.5 51889 36470 21635 65941 '73095 16767 4094 6584 4071 2018 7465 9302 All smoker 86611 13286 35258 27170 10896 49048 6640 19696 16238 6473 37563 6646 15562 10933 4423 76041 11525 10336 24160 10019 43124 32917 8314 1451 3656 2471 736 4512 3802 re ers Regular Regular and/or smoker ~ o.tv Num bets in 52442 51770 10069 9827 22916' 22656 15336 15236 4121 4050 27751 27445 5018 4866 12591 12503 8402 8357 1760" 1718 24690 24325 5051 4961 10345 10152 6933 6880 2361 2333 45090 44515 8582 " 8400 19723 19211 13383 13284 3676 3619 23654 23435 21416 21079 5903 5831 1264 1226 2657 2624 1636 1636 346 346 3136 3078 2767 2753 136 Regular Regular and/or smoker Thousands 33130 30731 3009 2632 11985 11167 11474 10541 6664 6341 20672 19297 1504 1303 6666 6393 7612 7113 4670 4488 12452 11434 1504 1379 5099 4774 3862 3423 1993 1853 30197 27976 2758 2463 10728 9955 10480 9622 6231 5936 19025 17732 11171 10244 2209 2045 158 142 874 829 785 711 391 369 1258 1192 950 653 Nsver ..oknd 74086 18890 28754 16330 13112 26906 9039 9837 4593 3413 47180 9832 15917 11737 9694 62910 15565 21471 14269 11600 22802 40108 8439 2643 2928 1586 1282 2953 5486 All ocoaslonal 3486 652 1269 1144 421 1994 405 692 645 253 1492 248 577 499 168 3045 566 1063 O 1036 : 380 1735 n 1310 ! 35o * 56 159 O "108 O * 26 0 184 166 D O q 0 m

~.~'¢ - DO ;~o~. ~-s ~ qua~-a. T~r~ 3. (~epa=a and L~=~oy, 1S85) ° . Tim¢ S~fl( ~n vMtOu.S m~fOqlRVlro~mt~l:s ~y ~.rf~flJ in 4.4 ~) l ~~ ~t~. e~ pr~ i~ averllle ~ours ~ dly. (Orl, ;n or~s: ~RC. J98h $la~. Z972), Em~y~l Em~ay~ Mmrn"d Mcn, wom~n, HCu4r~m, Microcnvironrneaz AJI Da:n AI| O|ys ~dl Days [nsi~ one's ~ome 13.4 15.4 20J Just outside one's home 012 0,0 O. I At one's ~o~kplace/ 6,? S,2, [n cran|at1,6 I,.I LO In ocher peoole's homes 0,$ O,T 0.8 In places of business 0,? 0.9 1.2 In res|auran~s Ln¢l i~&s 0.4 0.2 O. I tn all o~her Iocal~or~ OJ 0.3 0.3 Tot,tl 24.0 24.0 - 24.0

Urinary Cotinine Concentration and Number of Repo~ed Hours of Exposure to Other P~pie's Tobacco Smoke Within the Past 7 Days in Nonsmokin$ Married Men According "to Smoking Habits of Their Wives E=posu~ to O~h~ People's Smoka in UrinatT Cotinine PrecedinI Weik, h Cod'~nt ration, Smokini NO. ng/ml TOlal Ou ~side Home CatelotT cat of Wife Men Metn(SE) Median Mean(St) Median Mesn~(SE) Median Nonsmoker lot 8.-~L3)~ $.0 11.0(1.2)~ 6.5 ~ ~(1.2~r 6.0 Smoker 20 2~.3(14.8) 9.0 ~.2(4.1) 21.1 1o.4(3,3) 10,7 NOTE: Diff£mnc'~ (nonsmokinll wife versus smokin$ wi/e): "p < 0,0~; tp < 0,001; "p < 0.06 ~Wi~xi~ rt~xk s~m x~L SOURCE: W~Jd tnd Ritcl~i~ (19~).

?u,~,z s, (~I, 19~,:), ~a~ - Do not ¢ita or c~uo~i ~e m I,, Itoumnokerl st+ql~ 4m'r~,pOuznlnl~Ll l~Okl .~ Lad eomg~r~o~ w'It.,,h d~C~ve zmolctuI I1 I IO-II ~ PII~I / A m -- IIJ -- I1~.1 -- |10' -- m -- I -- I -- I

T~LZ 5.. OOn~'do ~ I ~ gh[~ I H ~B~hN ! O [C,.~p m ~ ~d I-L~ Nolml ~ • ~ -- I . ~ 1|¸

TA/jT.I 6. ES~ed average ~ons~lokers, ax'L:osu~es Co ~p f~o1~ ~I"S a¢ hone and a¢ vorX.(Re~¢e a~td Lot~.ey, 1985) The ¢oncan~ra~ions are ¢alcula~d for aodel home and vor~place mio¢~envi~o~ents and are veigbt~5 by average roslpiratlon ra~es and tlae budge~-s~les foe percent ~f. tlme spent at hone and a¢ york by sale and feIale nommokeEs..The t~p£oal nonsmoker is ant/Iat~l to ~i exp~Ied CO f~oI 0 to 14 I~ of RSP f~oa ZTS pe= day, with an average expo|l?aze of 1.$ ~/~y. f Lt/eslyie: D~iy Aver;li P~OhllbliiZy Of ~inI E;{pOI~ t~ound¢ Value) A~ work and at home: r, 63 x 62 = 39 Nclthe~at¥,orknoralhome:~ ' yt x 38 • 14 AI home bu~ not al ~,ork: ~0 62 x 3"f m 23 At work ~t not al home: % 63 x 38 i 24 Tou~: ~ I~0 Ex~sure (mI) 0.~ 010 0.45 0.]0 I .$2 0.4'/ 1,43 -- The average nonexcluslve pr0~lll~/ of a nonsmoker being exposed ~o ~'S at vor~ IS aspirated as 63t; ~.he probabil~ty of not being exposed at york is 37t~ ~he nonexclus~ve prob~illty of being e~sq~ to ETS a~ hOM LI ts~iEate~ El 62t; t~@ pcoba~tlity of not ~e~ng e~s~d at ~ome ii 38%.

D~af~ - Do no~. ei~..e or cF;ocj" 40- ,* h r't*t ~Ok C~'$ • PrWSenL ~ ~ °~1 ~ °~tmgQ~°Q! ~ln ~O~$~king ~ I // f/ // I~l~4~ =a~!:=~j:e~ ~° 1o~ ~,

nOnSl~ke=$' e~lU~lS ~o ~o ~ a~ u~qe ~e8 ~ htqheri ex'p=euzu. Shorn ~s a bim:o~-ms o~ tnf!It=ation" ~aZuQs In a qF ua~ple of 2~6 o~e~ ~S mi~41~ =~asl homes a~oun~ ~ne ~ou~. Averaqee hes~ing sea~on value~ sr~ ~¢n. The u~ian of *.he distr~ution is 0.9 ach and the alan is 1.1 ± 0.9 ach.(Gro= and 1979) (b~RC, ~9~6) - . . . . ALL 14 CITIES ~lR ~C~GE R~TE I~R" I

NOv. Dee. Jan. Feb. P, qN'. Apr. M~ Jm, JuL Aug. Sep. O~ Nov. Dec, ,Era. Feb. V~. AW. |976 1977 1979 FIOURH 3. -- Comparison oS prediotions of RSP model for smoking in single-family homes with fiel4 4ata (spengler eL al., 2S82) for monthly mean RSP oonoentrations for 55 homes in 6 Cities.

Mitmea~llm office buJ.ld£z~. ~e ~nt~'as~ bl~'vem~ day%J.ml ~p " levels, vhan sm©k~q o~rad, amd nlqh~-tlma P41P leve~s, ~-b~ it • did no*:, i~ =arid. Aerosol r~ass ~onca~trstuon ~ a ?00,m3 offi¢# with 0he Smoke! (Nais0~ #t dl. ~982! ~e sm0kiN~ trlstrLJmerl| ~IrlCl smokirl~ rite war• rio| $~4~ifie0. Howevsr, the |Jr 4xch&n~e ralll lot th• SDaC• may be ~q¢~la~e~ t)y organ• o( •guaticm 2. For t~a d~.sy ~( [T$ on ThurSc~a¥, 4~ty 9. • non-flnear regression ana~ys~s of the RSP Jay•is, with an 18 ~J~m= ~ackgrou~¢l level $uD|racflon, yie~Os Cem ~0 aCh (~0.9~), Thi• value Is ci~e to the ASHRAE-recomme~Oe~ vent~l~tio~ ¢a~• Io¢ ~ffl¢• s~e.¢a. ~eo ,~ ;[| | t,, 4[ 4~ v~ ~) 144 vl ql | i r! ! i r i i n • " ! I j I , I . I ! • . MeN WED ¥~ ~RI SAT

D~12~ - DO no~ ~i~a o~ c~r~o~a . Nodotk FIG. l~oca:~a of :ht tig*: /e~eva/ o/J~ce &u(/dl~#s. ~:;XLDDIG DGO[~SZOM8 (100 a2 ~ 1000 A'V.2) Occupiable FloOr Volume, ~,,~..--~;.~D~L.~" ~ocJ,~n A~i. w~ m~ {a~) An~hO~ale ~ 174000: 0.82 ~- .3S An. A~b~r 4900 S~ ?00 1.04 4. ,69 Columbia 24";'00 I$90~0 0.8S ~- .23 Fayet~r~ilJe 3 400 2] 3¢0 0.37 ~- , 09 Hu~ b 420 2? SOO 0.32 4. .16 l'~offoJk ]?300 b03~O 0.79 + ,19 Pittsfield I ?30 8520 O.'/0 _+ .~.9

~af~ - Do no~ ci~m or ~otm Yicj~rt 4, Zndoor ai~ pol2ution £~om I~S ae~oSO~o Xndoo~ 14rvels of rosplra~1o pa~i=1os in ~ui2di=qs vho~o tobaaao is smoXed {dais polnts ~-T)) g~oatIy exceed those In whlch s=oklng is p~obi~ited (~n1~:elod) n and ex=eed the iove~s ~o~ hea1~h-based U°S. a~lent ai~ c/~a~i~ stanGa~ds. 1200 1100 '~00{ go( ~' 80( !: =: 400 300 200 100 T NAAQS 24.H AV SIGNIFICANT HARM LEVEL FOR TSP NAAQS 24.H AV. AIR POLLUTION EMERGENCY LEVEL FOR TSP o~,~, • • ° ° ° • * • SMOKER DENSITY Eb"rlMATED .~ - • MEASURED DATA O CALCUt.~,TED EQUlUSRIUM LEVEl. mB eC DO eE eG ............... ?A~.a.~.:4:~ ~RI._~.A~.~ ~E.~_~ ~0~ T~.~ .... He OU • I° Jo " K@ NA£QS 24N PRIMARY t.EVEL PO# '(~NIO) M N L° $O O oe 1~R ep gO NAAGS ANNUAL PAFMARY LEVE~. FOR P~(10 =I m . m . d . ~ Sin m I. I *ll m • # ~ # E a : ~ 33 DATA pOINTS I I I I I I I ,~ 1.0 '1.5 2.0 2.5 3.0 3.5 Active smoker density (100 x bgl~lJrlci ~geereltes ¢)et m='~

D~-af~ - Do no= ¢:i~e or c~uo~s

D1"lf~ - DO no~ ni~a or ~o~e C~ER 8 DZSCOMFORT ASSOCIATZD W~TX ZlqVXRONM~NT~aL TOBXCCO ~MOK~ Willi~ s. cain Pho •ohn B. Pierce LLbo=ator~ and Tale Unlverslt7 New H&ve~, CT 08319 The atmosphere Inside buildings contslns many chemicals generated by the presence and activities of people. People's bodies give off small qusntlties of organic materials in the brea~h and from the skin and alimentary tract. Al~hougb a chemical analysis may reveal hun~eds or even ~housands of materials, we usually perosivs them in the aggregate as what we call occupancy odor. We often notice it consciously when we enter hot, muggy room. ~evertheless, 9ccupancy odor exists in occupied spaces at essentially all other tlmes, but remains at a low level because of ventilation with outside air (¥aglou, Riley,and coggins, 1936). When engineers and public health specialists began to study ventilation requirements for buildingsquantitatlvei¥, ~hey started with the smell of occupancy (Cain, 1979). The fresh-air requirements so derived exceeded those based on metabolically-relevant gases (oxygen, carbon dioxide) several-fold. In general, occupancy odor poses a mild challenge to the HVAC engineer. {HVAC refers to heating, ventilating, and air-conditioning.) This odor constitutes the baseline case. ADything else that people do in the space will increase ventilation requirements. This would include cooking, painting, operating machines (e.g., photocopier), woodworking, smoki~q, and so on. of these various activities, smokinghas traditionally been the most common. In a questionnaire study of odor problems in such spaces, Leonardos and Kendall (1971) stated, "Tobacco smoke is by far the most important odor contributor in enclosed space as indicated by the consistent agreement of che panel [principally experts in HVAC], and by their ranklngs. Also, it is considered a problem in virtually all (i! of 14) of the enclosed spaces" (p. i01). Tobacco s~oke has accordingly received considerable attention hlstorically in studies of odor control via ventilation or filtration (e.g., Yaglou, 1955; Kerka and Humphroys, 1956; Weber, Jermini, and Gra~dJean, 1976). As he has with occupancy odor, the HVAC engineer has confronted environmental tobacco smoke (ETS) via its sensory characteristics, i.e., its odor add irritation, rather than via Its chemical or physical complexity. The chemical complexity of ETS likely exceeds that of emissions from bodies and chemical 137

D]:aft - DO not cite or quote analysis of ETa-containing air has offered little of practical significance regarding specific chemicals responsible for its odor or irritation. Specification of the relevant chemicals might, however, assist in the mitigation of offending characteristics (National Research Council, 1986). In what follows, we shall review how human beings perceive ETS. We shall ask: How much ventilation air must he introduced into a space in order to satisfy visitors to ~hat space? Will the amount of air required by smokers difZsr from ~hau required by nonsmokers? Does ETS-odor decay spontaneously after smoking ceases? Do occupants accustomed tot he envlror~ment impose less stringent criteria for ventilation than visitors fresh from a nonsmoklng space? Does the odor and irritation of ETS come from the smoke particles Or from ~he vapors ~hat accompany the particles? Does filtration offer oppcr~unltles for control? Ventilation Retirements Based on Responses of the 'Visitor' A customary setting to explore how indoor contaminants affect the sshSes is a climate-controlled ~nvironmental chamber with relatively inert s~rfaces, e.g., aluminum or stainless steel, and variable ventilation. Such • model environment offers control over the physical and chemical characteristics at the expense of what we may call ecological realism, i.e., an everyday seth!rig. For the study of occupancy odor, human beings Occupy the chamber in order to generate the odor of interest, judges may enter the chamber briefly or may place their faces into i box fed with the atmosphere of the chamber. (In so sampling the atmosphere, the Judges essentially y_~ the space.) The odor Judgment may comprise a mark on an annotated rating scale (e.g., 'no odor' to 1overpowering odor') or the choice of a matching odor intensity. The latter Judgment generally entails the use of a device called an olfactometer that delivers the vapor of some standard odorant, such as n-butyl alcohol (1-butanol), at various concentrations. A matching odor has the advantage Of reproducibility from lab to lab. Many mode~ investigations also obtain Judgments of scceptabilihy in order to ,¢alibrate, intensity Judgments. Acceptability Judgments address the question~ How many people will .object to any given level of odor (or irritation)? The answer will depend on individual differences in clfact=ry sensitivity and on esthetic criteria. Whereas we can expect average intensity judgments to remain constant through the decades for any fixed stimulus, we can expect acceptability Judgments to shift somewhat wi~h prevailing standards. Three or more decades ago, when approximately half ~he adult population smoked and when restrictions cn smoking were relatively few, people seemed more tolerant of tobacco smoke odor than today (see cain,1979). Figure 1 depicts how occupancy odor varied with 138

Draft - Do not cite or quote ventilation rate per occupant under nonsmoking occupancy in a study conducted in a 1200-ft~ climate chamber (Cain, Leaderer, Isserof~, Eerglund, Huey, Lipsitt, and Perlman, 1983). Visitors made Judgments of air circulated through an outside sampling-box and were ~herefore naive to the conditions of occupancy. The scale refers to the concentration of I-butanol match~ to ~he occupancy odor present after one hour of occupancy. Just as odor level decreased with increases in ventilation rate, so also did dissatisfaction, i.e., Judqments that the odor was unacceptable. The point of 20% dissatisfaction holds special interest. The ventilation standard of the American Society of Heating, Refrigerating, and Air-Condltioning Engineers (ASHRAE) (1989) reoo~ends a maximum of 20% ~issatisfaction among visitors to a space. By this criterion, ~he data from the investigation imply the need for 17 cfm per oc¢upant. The ASHRAE standard s~ggests 15 cfm or more per occupant for most spaces, e.g., 15 ofm for classrooms, libraries, auditoriums, do~itoriee; 20 ¢fm for offices, conference rooms, dininq rooms, lobbies; 25 cfm for discos, beauty shops; 30 cfm for bars, casinos; 60 cfm for smoking lounges (see Fig. 2). Hence, practice coincides with the experimental data about as well as could be expected regarding the baseline case. When cigarettes were smoked in the climate chamber, odor level increased markedly. Fibre 3 dlaplays ET$ odor for various conditions of smoking: intermittent (4 clg per hr) or continuous (8 or 16 cig per hr). As Tig. 4 shows, the degree of dlssat~sfaotlon mirrored the higher odor level. Based on the rule of 20% maximum dissatisfaction, the ventilation r~te required per cigarette during active smoking exceeded 4,000ft~, ~n order to convert ventilation per cigarette into ventilation rate per person for typical conditions of occupancy in a 'smoklng-permltted' space, it was assumed that 10% of occupants would be smoking at any given time (see Repace and Lowrey, 1980). The resulting ventilation rate equalled 53 cfm, three times that for nonsmoklng occupancy. (The average smoking rate will of course va~y and the estimate of 10% may be high for 1990. The assumption of a lower rate of smoking would entail a proportional change in rate of ventilatlon.) Does the higher ventilation rate for smoking imply that the Judges in the investigation showed a special aversion to the odor of cigarettes? Apparently not. The Judges, one-thlrd of whom were smokers and two-thirds of whom were nott seemed to base their dissatisfaction strlctly on odor intensity. Degree of dissatisfaction varied with odor ~ntensity in the same way for both occupancy odor and tobacco smoke odor (Fig. 5). Stronger odors meant greater dissatisfaction irrespective of odor type. How well does the higher rate implied by the investigation compare wi~h the ASHRAE standard? As indicated above, the standard recommended 60 cfm per occupant in a smoking 139

Be t - lou g., where pres bly most or all Z i .quct" If 100% rather than i0~ werl smoking simultaneously, then the rate would need exceed an unachievahle 500 cfm per occupant. If 50% were smoking, perhaps a more realistic expectation, then the rate would need to exceed a still unachievable 250 cfm per occupant. (The maximum achievable rata for typical design occupancy in a mechanically-ventilated space will usually equal about 60 ofm per occupant, though as discussed below a generous allotment of space per person can increase that value.) Fortunately, however, the smoker seems less concerned about the odor of ETS than the nonsmoker. As it turns out, smokers as a group seem satisfied with about one quarter the ventilation slr of a mixed group containing a typical proportion of smokers ahd nonsmokers. Hence, a rate of 60 cfm per occupant may actually almost meet the customary ASHRAE criterion of a maximum of 20% dissatisfaction. How about nonsmokers? Just as a group of smokers will hold a less stringent criterion than the mixed group, a group of nonsmokers will hold a more stringent criterion. The data from the investigation suggest that with 10% smoking at any given time, nonsmokers would need over 100 cfm per occupant to hold dissatisfaction at only 20% . At the present time, we do not know whether the difference between smokers and nonsmokers derives from -~Ifactory sensitivity to ETS or to esthetic criteria. Clausen (1986) confirmed differences in tolerance of ETS odor between smokers and~nonsmokers. For any given level of odor (expressed as concentration of butanol), a group of nonsmokers expressed much more dissatisfaction than smokers (Fig. 6). ~thgroups exhibited a lawful relation between odor intensity and dissatisfaction, but the difference between the groups grew as odor level increased. At the point where 20% of smokers expressed dissatisfaction, almost half of nonsmokers did so. As ETS enters the atmosphere, its many chamlcal constituents react with each other and with surrounding materials both chemically and physically. D~es this behavior change the nature of the contaminant over t~me? Yes and no. Irrespective of whatever chemical changes occur, the odor of ETS behaves in the short run llka a stable contaminant. After the source has been removed, ETS odor decays in a manner entirely predictable from ventilation rate (clausan, Fanger, Cain, and Laadsrer, 1985). In this respect, it differs from ocoupan¢y~or which has a half-life of 55 mln, presumably dictated by slow oxidation of its chemical constituents into less odorous products (Clsusen, Fanger, Cain, and Leaderer, 1986). ETS odor offers no such easy benefit to the engineer. Indeed, when ventilation fails to, eliminate the contaminant entirely, ETS carrles a penalty derived from its physical interaction with surfaces. Because the ETS aerosol adsorbs s~rongly to walls, fabrics, add so on, it becomes a source of odor later. The background odor of the emitted products tattles its own demands for ventilation, predictable in part from the 140

Dz'af~ - Do no~ oite o~ ~ote typical amount of smoking in a space (Clausen, M~ller, Yanger, Leaderer, and Dietz, 1986). In a laboratory situation where o~her Sources Of combustion can be eliminated, carbon ~onoxide can offer a gross index of level of ~TS. Fig~ro 6Jhows that Clsusen could relate dissatisfaction to concentration of carbon monoxide ,in ETS as well as to matched level of butanol. Thl8 occurrsd because of a strong correlation (r>O.90) between odor intensity and iin=remental carbon monoxide due to smoking. Such a relationship makes it possibls, within limits imposed by brand-to-brand variability in emitted carbon monoxide, to compare one Itudy to another. We can ask, at what concentration of carbon monoxide will ETS roach a given level of dissatisfaction in one or another group? As Fig. 6 revealed, ~he concentration at which 20% of nonlmokers expressed dissatisfaction fell about eight times below that at which 20% of smokers expressed dissatisfaction. Up to this point, we have concerned ourselves only with the reactions of visitors. Standards for ventilation ~ave ~ocused on the reactions of the visitor, rather than those of the occupant, because the visitor will have a Rote sensitive, and hence more critical, nose than the person adapted to the contaminant. On the other hand, a focus on the visitor sidesteps another important tlme-dependent sensory response of the occupant, irritation. Whereas air containing an irritant may seem only barely irritating at first, it maybeoome intolerably so over time. Figure 7 illustrates the time-course of eye irritation experienced by occupants expoled to ETS at constant concentrations of 2 or 5 ppJ carbon monoxide, used here as a tracer in the manner mentioned abovl (Cain, Tosun, See, end Leaderer, 1987). The lower concentration led to slight, though statistically si~ificant, irritation above pro-smoking baseline. The higher concentration led to irritation that incrRasedover time in lensory magnitude and caused an increasing degree of dissatisfaction. Whereas essentially none of the occupants found the irritation objectionable at first, by the end of an hour about 30% found it so. In an extension, clausen, Nielsen, Sahin, and Fanger (1987) found that an asymptotic level of 20% dissatisfaction would occur at a concentration of 3.8 ppm carbon monoxide. ,A comparison wi~h the odor judgments of visitors in ~ig. 8 reveals that only smokers would find such a level tolerable by the '20% rule., clausen It el. estimated that the ventilation rate necessary to control irritation of occupants to a dlseatlsfautlon of 20~ would equal only one-tenth of that needed to control odor perceived by visitors to the same level of dissatisfaction. Although Clausen et el. did not argus in favor of 141

Dl-ef~ - DO not cite o~ ~ote lowering ventilation to meet only~e dissatisfaction of occupants, there could exist some temptation to do so (see Wi~eke, Plischke, Roscovanu, and $chlipkoeter, 1984). Cain et el. (1987) cautioned against ~e temptation to see irritation and odor in the same light: Apart from ~i issue of Yhether viaitors or co,pants are more sensitive, ~ere exists • q~estlon regarding whether ~e .20~ rule' should govern dissatisfaction based on irritation just as it governs dissatisfaction based on odor alone. ~ereas odor may be inte~retablm naz~owly on grounds of comfoz~, irritation would seem inte~ret~d~le on groundl of h • a 1 t h. Some people may find themsslvee quiteneutral with ~to one or another odor, but no one could plausibly argue neu~rality with respect ~o burning eyes. It could be argued, therefore, that any consistent irritation above baseline should be deemed unaccept~le. [p.352] A~licabilitv of Chamber Studie~ The data presented above may raise two issues of ~oncern: 1) Should chamber studies influence ventilation policy in "view of thei~ remoteness f~om real-world circu~stances? and 2) Would small errors in the results lead to large differences' in reco~endedpolicies? ~e firs~iesue ham no simple answer. ~n the real world, people engage in such a wide variety of activities that any single field study, even assuming accurate execution, • would itself have very limited generality. Only a set of field studies wlth a .variety of scenarios could even approach the generality desired. Such field studies have not been done, A group of subjects sitting in a chamber with no task other than to focus attention on odors might seem likely to behave very conservatively, i,m., to Judge even weak odors unacceptable, which would in tu~ imply the need for high ventilation rates. We can neither confirm nor deny ~his tendency, though circumstantial evidence runs against it. AS already noted, visitors in Cain et el.*s (1983) study found ETS odor no more objectionable than occupancy odor~at the same perceived intensity. Could this just mean that sub, sots treat each odor equally conse~atively? Unlikely, slnoethe recoamended ventilation rates for occupancy odor from that study converge with a great dea~ of other lab and field evidence regarding ~he need for about 15 to 20 cfm of ventilation per occupant. Even if the chamber experiment happened to encourage conservatism, persons who choose to pa1~iclpate in it and hence to expose themselves to potentially aversive environmental odors may represent a less reactive fraction of the population. Persons who find ETS odor aversive, for example, would seem unlikely to accept such work. Concern about these matters might, however, stimulate some productive research into the demographic factors 142

~a~ - Do not eits or gucts that govern reactivity to indoor odors. Chamber experiments on ETS can be criticized because they have explored levels that largely exceed those of everyday life. The tendency to explore high levels derives in part from a desire to cower a wide range of con6~tlons and in part from crude estimates of levels of smoking in the c~un~T1es and during the eras when the experiments were performed. Even Just ten years ago, smoking in the U.S.A. occurred m~Ie commonly and at higher levels than today. In countries such as Denmark, the l~:ation of some recent studies, smoking occurs with a higher frequency a~d with fewer restrictions than IN the U.S.A. Some recent field surveys have found surprisingly low levels of ETS in common spac6m, e.g.~ offices (Kirk, Hunter, Back, Lector, and Perry, 1988; Oldeker, 1989). In order to understand how to relate the chamber studies with such field data, we need to factor in the ventilation rates in the field (see Nystrom and Green, 1986, for a die~sslon of variables relevant to the evaluation of ETS). Although a building code may specify a ventilation rate of, say, 20 ofm per occupant, the actual rate will depend on the number of occupants actually in the space. If a space typically contains only one-third the design number of occupants, the ventilation rate will equal 60 cfm per person. This situation occurs frequently since the design occupancy listed in a standard commonly comes from fire regulations regarding maximum density of occupancy. Aooordinglys one cannot argue, as has been done, that~ a putative low frequency of complaints in field settings offers evldenoe against the recommendations of chamber studies and in favor of lower rates. Field data, if collected in spaces occupied well below design levels and if reported without actual per-occupant ventilation rates, can give the illusion that rates of ventilation suitable for occupancy odor can lead to adequate control of ETS odor. When normalized to a ventilation rate and hence when seen without assumptions regarding occupancy, chamber studies have probably yielded quite valid data, irrespective of the levels of smoking explored. Regarding the sedond concern mentioned at the beginning of this section, small errors in the estimate of dissatisfaction could in fact lead-to large errors in recommended rate of ventilation since ~e relation between percent dissatisfaction and ventilation rate for ETS has a rather low slope (Fig. 4). Merely on general grounds, it would seem advisable to replicate this relation with new participants in order to check its stability and valldity. Alternatives to Ventilation It might seem intuitively reasonable that the odor of ETS should come from its vapor phase and the irritation from 143

0 D~raft - DO not cite Qr q~ote its particulate phase. At one time this seemed likely, but recent investigations that have employed electrostatlo ai= cleaning have shown clgarly that the gas phase ~coounts for the majority of odor and irritation (of. Hugod, 19841 Weber, 1984). Comparison of the right and left sldRs of Fig. 7 will reveal that ellminatlon of the particulate phase had only a trivial effect on ~he eye irritation caused ETS at 2 and 5 ppB carbon monox~dQ (cain, Tomun, See, and Lead&rat, 1987). The same held true for ~ud~ents of odor and of nose and throat irritation. Claulen, Ni&Imans Sahib, and Fanger (1987) confirmed these results° fin finding that par~ioles played essentially no role in odor, ho~h investigations also confirmed Clausen at al.'s (1985) earlier experlments wlthvlsitors. Hence, particle filtration holds no promise for immediate oliminatlonof the dlscomfo~ of ETS. The major advantage of such air cleaning will derive from reduction of haze and coil&orlon of ~tar' that would otherwise adsorb elsewhere in ~he space. Although both ~he odo~ and irritation of ETS come from the vapor phase, the chemi=al constituents that give rise to the one probably do not give rise to the other. Undoubtedly, the odor comes from a very large number of constltuents. The sense of ~mell rill respond to almost all alrborne organic materials present ~in sufficient concentration CCaln, 1988). For One s~$tance, however, a tsufficient concentration' may fall a millionfold below that of another.. Furthenore, individual constituents will combine perceptually in mixtures in complicated, nonlinear ways. Although one or a few materials could in principle dominate the odor, it see=s unlikely. Many fewer materials can cause irritation at the concentrations present in ET$ and its Irrltatlon could realistically arise from a ~ew or perhapB even one constituent. Little is known about how irritants combine with each other perceptually though it is known that odor and irritation interact (cain and Murphy, 1980). Zrritation can suppress the perception of odor and vice versa (cain, See, and Tosun, 198~). In so far as irritation =ay have a less complex origin than odor, it =ay offer easier opportunities ~or control through filtration. As yet, however, experiments on the orlg~ of ETS have told more about what ~ails to cause irritation than about what ~auses it (Weber, Jermlni, and GrandJean, 1976~ Weher-Tschopp, Fischer, and Grandjean, 1977~ Weher-Tschopp, Fischer, Glarer, and GrandJean, • 1977~ Hugod, Hawkins, and Ast~up~ 1978). The complexity of ETS ~ore or less guarantees that almost sny means of ai~ oleanlng will eliminate part of it, even though no simple procedure will &li=Inate all of it. ThrOugh the use of air washing that presumably eli=in&ted some water-soluble constituents, Clausen, M~llsr, and Fanger (1987) achieved some reduction in level o~ dlssatisfaotlon though not in the perceived intensity of ETS. The air-washed ETS smelled fresher. The results offered little encouragement for the use elf-washing alone, hut 144

D~af~ - Do no~ c~a or ~o~e shoved that the odor character of ETS can play some role in degree OE acct~t~ncQ° Undoubtedly, a ao~bina~ion of pa~iculate ai~ cleanin~ and vapor-phase cleaning via adsorption on activated carbon or via chemisor]~ion on oxidant-~pre~nated alumina can control bo~.h ~e irri~a~ion and o¢~or of ETS ~o some degree. Unfo~cuna~ely, there exis~ no standa~ls to assess ~he efficacy of vapor-phase flltra~ion ~edia. ~2~e ins~,alla~on of such media occurs more c~nonly in s~ecial environments, Q.g., l£~rRries and co~ut~r facilities, unde~ exl~e~-t~uidance than ~ spactl designed for ~eneral Qc~upancy. I~ ~.~e ~ve~he~ming maJori~ of cases, attempts ~o control ETS ~ly on ven~ila~$on (dilu~i~n). As ve havl seen, however, v~n~ila~ion has i1:41 ll~i~a~ions. 1. A~ an average smokin~ ~ate of 10~ Imokin~ a~ any one ~ime, nonsmokers would need in excess of 100 ¢fm/occupan~ ~o hold dissa~isfac~ion ~ ~he~:~Zcrite~i~n of 20t. cx~or acceptl~ili~y. 2. Exposure ~o ETS generates odor and ir~i~a~ion ~n ~oth nonsmokers and smokers° Nonsmokers as a group are le~s ~ole~an~ of ETS ~ha~ s~okeEs~ ~° The ±r~i~ation an~ odor from L~S a~p~a~ ~o ~lsida in ~e vapor phase. The control o~ ETS irri~a~ion and odor by ventila~ion or air cleanin~ can p~ovide only li~t~d rssul~s. 145

Draft - Do no¢ ¢ite o1: quotm American Society of Heating, Refrigerating, and Air-Conditionlng Engineers (ASh~RAE) (1989). Ventliatlon for Acoe~tmble Zndoor ~IT Oualltv. ANSI/ASHRAE 62w1989. A~la~ta: ASHR~. Cain, w. S. (1979)o Ventilation and odor control: prospects for energy savings. ~, 8S(I~, 784-792. Cain, w. S. (1988). Olfac~io~. In R. C. Atkinson, R. J. Herrnstein, G. Lindzey, and R. D. Lute (Eds.), of E~erimsntal P~vchol~. Vol. I~ Pe~eDtlon and Motivation, ray. ed. Hew York: Wiley• Pp. 409-459. Cain, W. S. and Mushy, C. L. (1980). Zntsraction between chemoreceptlve modalities of odour and irrlta~ion. Nature, 284, 255-25?. Cain, W. S., See, L•-C., and TosUn, T. (1986). Irritation and ndor from fo~aldehyde: chamber studies. In , • ~ndoor Air for Health and Enerav conservation. Atlanta: ASHRAE. Pp. 126--137• Cain, W. S., Tosun, T., See, L.-C., and Leaderer, B. (1987). Environmental tobacco smoke: sensory reaotlons of occupants. ~, 21, 347-353. Cain, W. S., Leaderer, H. P., Isseroff, R., Berglund, L. G•, Huey, R. J., Lipsitt, E. D., and Perlman, D. (1983). Ventilatlon requirenen~s in buildings - I• Control of occupancy odor and tobacco smoke odor. ~, 17, 1183-1197. Clausen, G, H• (1986). Tobaksrog - lugtgenar og ventilatlonsbehov. Doctoral thesis, Technical University of Der~mark° Clausen, G. H•, Fanger, P. O. Cain, W. S. and Leaderer B. P. (1985). The Influence of aglng,~ar~i¢le filtration and humldlty on tobacco smoke Odor. In P. O. Fanger (Ed.), • ' . Copenhagen: ~VVH Kongres - WS Masse. Pp. 345-349. Clausen, G. H., Fanger, P. O., Cain, W. S., and I~adersr, H. P. (1986). Stability of body odor in enclosed spaces. ~, 12, 201-205. Clausen, G. H•, M~llsr, S. B., FangeE, P. O., Leaderer, B. P., and Dietz, R. (1988). Background odor caused by previous tobacco smoking. In IAO 186: Manaaino Indoor Air for Realth and Enerav ~. Atlanta: ASHRAE. Pp. 119-125. 146

Draft - Do not cite or ~uo=e Clausen, G. H., M~ller, S. S., and Fanger, P. O. (1987). The impact of air washing on environmental tobacco smoke odor. In S. Seife~, N. Ksdcr~, M. Fischer, H. R den, and J. Wagner (Eds.), t ? V . Berlin: Institute for Water, Soil and Air Hygiene. Pp. 47-51o Clausen, G. H., Nielsen, K. S., Sahln, F., and Fanger, P. O. (1987). Sensory irritation trom exposure to environmental tobacco smoke. In B. Seifsrt, N. Esdorn, M. Fischer, H. R den, and 3. wagner (Eds.), ' 7 V . Berlin: Institute for Watert Soil and Air Hygiene. Pp. 52-56. Hugod, C. {1984). Indoor air pollutlon with smoke constituents - an experimental investigation. ~L~, i~, 582-588. Hugod, c., Hawkins, L. B., end strup, P. (1978). Exposes of passive smokers to tobacco smoke constituents. Archives of Occupational and EnviTo~ental ~ealth~ 4~, 21-29. Kerka, w. F. and Humphreys, C. M. (1S56). Temperature and h~idity effect on odor perception, HeatinG. Pimina. and A}~ ~, 28, 128-135. Kirk, P. W. W.,'Hunter, M., 5ask, S. 0.~ Lester, G. N., and Perry, R. (1988). Environmental tobacco smoke i~ indoor air. In R. Perry and P. W. W. Kirk (Eds.), Indoor and ~u~blent AlE Oualitv. London: Selpero Pp. 99-112. Leonardos, O. and Kendall, D. A. (1971). Questionnaire study on o~or problems of enclosed spaoe. ~, 77 (1),101-112. Leopold, C. S. (1945). Tobacco smoke control - A preliminary study. ~, 51, 255-270. Na~onal Research Council (1986). Environmental Tobacco Smoke - Measurlna Exnosures and Assassins Health Effects. Washington: National Academy Press. Hystrom, C. w. and Green, C.R. (198S). Assessing the impac~ of environmental tobacco smoke on indoor air quality. In Manaain~ the IndooE AlE foe Health an~ Ene~v C~nsel-vation. Atlanta: American Society of Heating, Refrigerating, and Air-Conditioning Engineers. Pp. 213-233. 01daker, G. S. (198%). Environmental tobacco smoke (ETS): How much is in the air? Presented at the International Tobacco Conference MinisvmDosium on Environmental Tohscco Smoke and Scientific Affairs, Winston-Salem, NO. Repace, J. L. and Lowry, A. H. (1980). Indoor air pollution, tobacco smoke, and public health. Science, 208, 464-472. 147

Draf~ - Do not cite or quota weber, A. (1984). Annoyance and irritation by passive smoking. ~, 13, 618-625. weber, A., Jermini, C., GrandJean, E. (1976). Irritating effects on man of air pollution due to ¢i~arettl smoke. ~, ~, 672-676° Weber-TsChopp, A., Fischer, T., GrandJean, E. (1977). Reizwlrkungen des Formaldehyds (HC~O) auf den Menschen. (~rritating selects of formaldehyde on men.) Archives of Occuoational and Envlronmental Health. 39, 207-218. Weber-Tschopp, A., Fischer, T., Gierer~ R., and GrandJean, E. (1977). EXperlmentelle Reizwlrkungen yon Akroleln auf den Menschen. (Experimentally indu¢~ irritating effects of acrolein on men.) Archives of OccuDatlonal and Envfronmental Health, 40, I17-130. Winneke, G., Plischke, K., Roscovanu, A., and Schlipkoeter, H.-W. ~1984). Patterns and determinants of reaction to tobacco smoke in ~h experimental exposure setting, in B. Berglund, T. Lindvall, and J. Sundell {Eds.), ~ Stockholm: Swedish Council for ~uilding Research. Pp. 351-356. Yaglou, C. P. (1955). Vsntilatlon requirements for cigarette smoke. ~, 61, 25-32. ¥aglou, C. P.," Riley, E. C., and C~ins, E. ~. (1936). Ventilation requirements. , 42, 133-162. 148

~a~t - DO flo~ cite or c~.ote Figure 1. showing the relation between level of occupancy odor (indicated hy concentration of 1-butanol matched to the odor) and ventilation rate per .co,pant when 4 to 12 persons occupied a climate chamber for enhou~ (filled sguares). Judgments of odor were made by visitors who sampled the air of the chamber at a remote sampling box. Also shown (unfilled squares) is the frequency of dissatisfaotlon expressed by,he visitors in response to the question, Is the air acceptable or unacceptable ? Dashed llne shows ventilation rate that led to 20% dissatlsfaction. Data from caln etal. (1983). Figure 2. Frequency distribution of ventilation rates recommended for various types of spaces (e.g., offices, auditoriums, ticket booths, waiting rooms) By the ASHRAE standard on ventilation and indoor air guality. Figure 3. Showing the intensity of ETS odor perceived by visitors to the sampling box during and after ~ntermittsnt (4 clg/hr) or continuous (8 and 15 cig/hr) smoking in the climate chamber. Results are expressed relative to level of butanol matched to odor during presmoking occupancy, The open squares in the left panel show a fuLnction for nonsmoklng occupancy for comparison. Ventilation rate per occupant under smoking conditions refers to smokers, who were the only occupants in the chamber. From Cain etal. (1983), Figure 4. Percent dissatisfaction among visitors vs ventilation during the last 15 min of smoking in the experiment shown in Fig. 3. Ventilation rate per cigarette based an 7.5-min smoking time per cigarette. Ventilation rate per occupant adjusted to conditions of smoking occupancy that assumed 10% of occupants will be smoking at any give time. Modified from Cain et al. (1%83). Figure 5. Percent dissatisfaotlon vs odor intensity (graphic rating) for occupancy odor and for ETS odor. Data from Cain etal. (1983). Figure 6. Left: Percent dissatisfaction vs odor intensity (matched level of butanol) judged by smokers and nonsmokers. Right: Percent dlssatisfactlon vs increment in concentration of airborne carbon monoxide. Modified from Clausen [1986). Figure 7. Perceived magnitude of eye irritation and degree of dissatisfaction expressed by occupants exposed to ETS for an hour. Concentrations of carbon monoxide were held constant throughout the exposures and indicate severity of exposure. Filtration refers to elimination of particles via 149

Draf~ - Do not cite or quota electrostatic preclpita~ion. Filtration had little effec~ on irri~ation. From cain at al. (1987~. 150

TI%BL2S kND FIGURRS FOR CHAPTZR R Draft - Do no~ eita or quota 151

~a~t - Do not ¢ita or quote / fJ = ¢7 Figure 2

,,1 ~t D Bulanol (ppm) o~.. o o o~ '// ~ o m | , i I~ o Percent OiSSallsfled 0 ~r ,.) 0 0 rt

=E .m .. • -~-~--,-~.~ - ~-- . ~ ~ ~ .P- ~,~ .0 . ' ~|1 ..... ' III!iIi-1- _/ o° o° ~ o° ~ ~ ~ ~ ~ 0 , ' 0 +,

01"a~c - Do not C:~a or ~uoCo ~8C 20 0 5 Ventilafio~ Aateper Occupant (cfm) 20 50 kO0 5 I0 ..... :,. ,. ~0 00 • 0 0 0 0 0 0 I 2~ Ventilation Rati I~r Oig~rette ~ft3~ ?lEers 4

Ora~¢ - Do hoe ¢t¢1 o= ~;o~s 80, ~ .... ~f Occupancy Odor o~Z ~/ ETS Odor / • (,Oe : +iI I • 0 I I ] I I 2 3 4 5 678 Odor Intensity (cm)

D1"lt~ - DO no~ cite o~ c~o~e ~ St~;llcef ,i~ n • 0 !.. ze BUl0n01 (ppm) 1 ACO (pprn) \, 7~S 6

? ,i ,/ -4 ! PlaNT DI55AIISFICO Z 0 t. +, PI~I~IEW1ED IHTEHSITY C I Ia "iI 0 0 I1 0 it

~?ZR 9 Dr~tft - Do not cite or quoge PASSZ~FE SMOKING--SELIZFS, ~TT~TUDSS, AND ~XPOSURES IN T~ UNITED STATES Thomas z. NovoZnT, X.D. Offlae on Smoking and Xealth Center for Chronic Disease Prevention Lnd Eealth Promotion Centers for Disease Control The relationship among public attitudes, policies, and exposure to certain health hazards is complex. With bhe release of the IBth Surgeon Generalls repor~ on smoking e~d health, Consemuences cf ~nvoluntarv Smok~na (PHS 1986), public attention on the issue of environmental tobacco smoke (ETS) was morsstrongly focused ~han ever before. For ma~y years, however, pollsters, the tobacco industry,~and the health promotlon community have hays been surveying the public concerning ettltudes toward ETS and toward restrictions an smoking in pu~li~ places. The Surgeon General~s Report described data from several of those surveys as well as results from evaluations of worksi~e and local polloy changes. Additional detailed date on public beliefs and attitudes toward smoking in general are found in the 1989 Burgeon General*s Report: Reducln= the ~ealth Co~se~ences of Smoklno -- 2~ yeats of Pro~rea~ (PHS 1989). Recently, surveys have also included questions on beliefs about the harmfulness cruETS to the nonsmoker and on respondents' reported exposure to ETS. In addition to s~ch measures of individual exposure to ETS, surveys Of worksites and of personnel managers have provided information about restrlctions on.~moking in the workplace. Because changes in public attitudes toward ETS usually precede laws or policies regarding ETS exposure (PHS 1986), examining brends in these data over time is useful. This chapter will summarize the most important findings from several different nationally based data sources. Some of this information was included in the 1989 Report of the Surgeon General (PHS 1989). Data Sources and Methodolo~ Several surveys of public beliefs, attitudes, and reported exposures to ETS are available (Table i). Although these surveys may repoz~ discrepant results, most dlscrepancle8 can be explained by the differences in methodology, especially iN the ways questions are worded. To describe the effect of increasing numbers and strength of laws and policies against smoking in public places, 152

ft - Do not cita or ~ctm national surreys of workaites were also ca~ed out in the 1980s IPHS 1986). These su~eys indicate the degree to,which workers ~ay be prc~ec~ed from ~S exposure. The 1987 National Heal~ Inte~iew Survey of Cancer Epidemiolo~y and Control also collectRd lnfo~ati~n abou~ respondents~ actions in response to ETS exposure. i. Roper surveys: The Roper Organization conducted slx blennial national opinion surveys for The Tobacco institute between 1966 and 1978. The 1974, 1976, and 1978 surveys focused on the passive snoklng/nonsmokerts rights issue (Rope~ 1978)~ whereas all six surveys dealt with pu~llc attitudes toward the smoking and health issue in general, toward the tobacco industry itself, and toward government regulation of tobacco. The surveyl were cross- sectional, population-based telephone interviews. The sample Included over 2,000 adults, abed i? years or older~ other information about the exact methodology and response rates is unavailable. The 1974 1978 Rope~ surveys permit comparisons of data collected for the tobacco in~ustrywith similar data collected in the 1970s by the office on S~oklng and Health (OSH~ formerly known as the National Clearlnghouse for Smoking and Health). 2. Hamilton, Frederick, and Sch~elders: In December 1988, the Tobacco Institute sponsored a telephone-based national adult survey of 1,500 adults (401 smokers and 1,099 nonsmokers) which was conducted by Hamilton, Frederick, and Schneiders (Hamilton, Frederick and Schnelders, 1989). This survey asked about various public policy issues and was designed to measure levels of support for governmental "policy on smoking. The respondents were asked what they thought about restrictions on smoking in restaurants and worksites. Neither the responserates nor the results by smoking status of the respondents were reported. Cthe~ Public o=inion Surveys 1. Gallup Surveys: The Gallup Organization has published Gallup Poll results monthly since 1965. Surveys are either personal interviews cr by telephone and have a population-based sample of at least 1000 adults, aged 18 years or older. The sampling error for overall responses is reported to be no more than ~39 (Gallup Report 1987). In addition, Gall~p surveys may be commissioned by a variety cf organizations. The surveys reported here were commissioned by the A~erlcan Lt~g Association (1983, 1985, 1987, and 1989) and the American Cancer Society (1998) to .describe both the prevalence of smoking and public opinions regarding smoking issues. An additional Gallup Survey was commissioned by the National Restaurant Association (1987) to obtain public opinion on smoking in restaurants. The 1989 Gallup Survey sponsored by the American Lung Association did not ask respondents about their smoking status.

Dra~ - Do not cite or q~ote 2. Harris Poll: Louis Harris and Associates have performed eight national surveys covering smoking between 1974 and 1987 using probability samples of adults aged 18 years and older. These surveys were conducted on behalf of various organizatlons~ including ~ magazine and Pacific Mutual Life Insurance Co., to ascertain heal~h practices of Americans. Zn 1987, 1,250 persons were also asked about regulating smoking in public places. Gover~B~t S~o~c~ed Surveys I. Adult Use of Tobacco Surveys (AUTS): The Office on Smoking and Health commissioned surveys of adult smoking behavior, attitudes, and beliefs in 1964, 1965, 1970, 1975, and 1986. These surveys oversampled persons who had ever smoked, but final results were weighted to represent the United States resident population aged 21 years and older (1864, 1966r 1870, 1975). The 1989 AUTS oversampled ever smokers but collected data from persons aged 17 and older. The final data in this survey (overall response rate, 74.89) were weighted to reflect the educational, regional, racial, and age distribution of the U.S. population on the basis of the 1986 Current Population Survey of the U.S. Bureau of ~he Census, The 1988 AUTS collected detailed info~matlon on attitudes, beliefs, and exposure regarding BTSo 2. National Health Interview Survey: The National Health Interview Survey of Cancer Epidamiology and Control (NHIS-CEC) collected data in-person from 22,000 adulta aged 18 years and older in households throughout the United States. The data were weighted to reflect the adult U.S. population, and the overall response rate for NHIS-CEC was 82%. Respondents were asked about the harmfulness of ETS and about attitudes toward passive smoking. Questions included items on perceived annoyance and whether smoking should occur inside public places. Nonsmokers ware asked how they would act in response to smokers~ lighting up in ~helr presence. Other SurVeys 1. Bureau of Natlonal Affairs: The Bureau of National Affairs (%NA) and the American Society for Personnel Administration (ASPA) condu=ted a mail-ln questionnaire sudsy of ASPA members, and 823 respondents reported on activitles related to smoking in the workplace. The response rate was 54%. A slmilar survey was carried out by the BNA in 1986 on 662 businesses. 2. office of Disease Prevention and Health Promotion: In 1988, the office of Disease Prevention and Health Promotion (ODPHP) of the United States Public Health Servi=e reported on worksite health promotion activities, including~ smoking control. The survey, carried out in 1985 on a sample drawn from the Dun and Bradstreet 154

D=aft - Do not cite or ~ote list o~ businesses, used telephone inte~iewing. To develop a probabiiity sample based on geographic region, size of firm, and indus~ ~ype, 320 worksites vith 50-80 employees and 1,038 worksites with leo or more employees were suI-~eyed. Questions about smoking restrictions were ~ncluded. 8. American Board of Family P~ac~ice: In Dece~er 1984, the American Board of Family Practice (ABFP) sponsored a national telephone survey conducted by R0seerch and Forecasts, Inc., of 1,007 adults aged 18 years and older and of 303 family physicians. Questions were asked regarding beliefs about the harmfulness =f ETS, the rights of smokers and nonsmokers, a~ whether smoking restrictions are effective in stopping or discouraging smoking. The final sample response rates were 41% for the general public and 37% far physicians. Data for ~he general public portion of this survey were weighted to reflect the estimated 1985 U.S. population. The physicians surveyed represented a random sample of U.8. family physicians. The results were pvblished in a report entitled, Riuhts and ResDonslbilities: ~ealthcare Options (ABFF 1985). i. Perceived Hal-mfulness of Environmental Tobacco Smoke The Roper Surveys asked questions regarding harm and annoyance caused by ETS. All AUT surveys asked about annoyance caused by ETS, but only the 1986-AUTS asked.if respondents believed that ETS is harmful to th~ nonsmoklr. The 1983 and 1988 Gallup Surveys asked if respondents believed that smoking is hazardous to the health of nonsmokers. The 1978 Roper SurVey, the 1986 AUTS, and the 1988 Gallup survey provide intmresting information on the change over the last several years in public beliefs about the harmfulness of ETS to nonsmokers. The 1985 ABFP Survey asked both adults and physicians if they believed nonsmokers are harmed by breathing in the smoke of others in the same room. Questions regarding har~ caused by ETS showed that between 1974 and 1986, the percentage of smokers who believed that ETS is harmful to the health of the nonsmoker more than doubled (Table 2). In 1974, most nonsmokers believed that ETS is harmful to ~health in general, and the percentage of those who held this belief increased substantially over time. ~n answering an additional question on the 1986 AUTS, 69% of nonsmokers felt that ETS is bagful to their own personal health. The results of the 1969 Gallup poll suggest that there is even stronger belief by respondents (smokers and nonsmokers) in the harm of ETS to pregnant women and children. These data show that there has been a major change in the perception of ETS as a health hazard over the last decade. 155

Draf~ - Do no~ oi=e or quote 2. Annoyance Caused bv Environmental Tobacco amo~e The AUT surveys show an lnoreasinq trend in the percentage of respondents who are annoyed by ETS (Table 3a). Data regarding annoyance to ETS from Roper Surveys other than the 1978 survey are not available. However, the results of bo~h the 1978 Roper Survey and the AUTS suggest that increasing numbers of Americans are annoyed hy ETS exposure. The results of the 1987 NHIS-CEC also indicate increased annoyance from ETS. ~n this sudsy, a smaller percentage of current smokers reported a~oyance than on the 1998 AUTS, but this difference may be due to different methodologies. The NHIS-CEC also collected information about what nonsmokers did in response to exposure to ETS (Table 3b). About half of respondents moved away from the exposure source, 40% did nothing, 3% did something else, and only 4% asked the person not to smoke. Despite their high positive responses to perceived harm caused by ETS and annoyance from ETS, most nonsmokers remain rather passive in their behavior toward smokers (Davis et el., 1990). 3. Limltina or Bannlna Smoklna in Public Places The majority of respondents to the 1978 Roper SurVey felt that smokers should at least be segregated in all the public places cited (Table 4). After being asked about segregation of smokers and nonsmokers, respondents were then asked if smoking sh6uld be banned outright in selected publlc places. The majority of respondents favored smoking hans in retell stores, physlclans' or dentists' waiting rooms, and elevators (Table 5a). The narrative description of the surVey results pointed out that after recognizing the option to segregate smokers, respondents were probably less likely to be in favor Of a total ban (Roper 1978)o The two most important reasons given by Roper Survey respondents before 1978 as to why smoking should be restricted had to do with dangers to others, specifically, cigarette smoking as a fire hazzrd and ETS as a health hazard to nonsmokers, i~ 1978, the chief reason respondents gave in favor of public laws against smoking was that ~he "health of nonsmokers is hamed by other people smoking in their presence.. - In 1983, 1985, 1997, and 1989, the Gallup Organization conducted telephone surveys for the American Lung Association (ALA) that asked if smokers should refrain from smoking in the presence of nonsmokers. Overall, the percentage of respondents to these surveys who agree that smokers should not smoke in the presence of nonsmokers has increased from 69% in 1993 to 82% in 1989 (Table 6a). This trend holds true for both smokers and nonsmokers. Unfortunately, the 1989 survey did no~ differentiate between smokers and nonsmokers. 156

D~aft - Do not ci~@ or quota The ALA Gallup SurVeys also Included questions on wherm smoklnq should ~e restric%ed or banned. With regard tQ smoking in hotels, motels, and restaurants, ~he maJorlty of respondents in 1983, 1987, and 1989 fel~ tha~ cer%ain areas should be $e~ aside for smoking (Table 7). Completa bans were less favorQ~, Rspeclally by persons who currently smoked. In addltlon, respondents to thl ALA Gallup Surveys were askad in 1983, 1985, and 1987 if companles ~hould have a policy on s~oklng at work. By 1985, almos~ 90% of all rRspondentm, Includlng 80% of smokers and 89% of nQnsmokers, fel~ that smoklnq should be assigned to certain areas ~f the worksi~a or that i~ mhould bQ totally banned at work (Tab1~ 7). In 1987, the monthly Gallup POI!s (no~ commlsslonQd ~y the ALA) asked if respondents favored or O~pooed a ¢ompl~te ban on Bmoklng in all pu~li~ places. The resu1~s of ~hes~ polls are much more strongly in favor of total bans on smoklng in public places. These resul~s contrast shar~l~ with ~he~Roper ~esul~s of al~ost a decade ago an~ are even morm in favor of IncrQased restrlc~ions on smoking in public places than the ALA-sponsored survoys in thQ samQ year. ~n the Gallup survey c~nducted ~or thQ National Restaurant Association in 1987, 61% of adults repor~e~ %ha~ %hey preferred no-~moking sections in restaurant~. Theso included 20% of smokers, ~S% of ~ormer smokers, an~ 83% of never smoker~ (Gallup 1987). Theme re~ul~s are similar ~o ~hose of the A~Y~S on preferences concernin~ no-smoking sections d~mc~ibed la~er in ~his chapter. The 1987 NHIS-C~C. asked a gligh~iy ~fferent question ~han either t~e Gallup SurVeys ~r the AUTS. This que~tlon re$%rlc~ed t~e respondent t~ consider indoor public pla~el. The p~rcsn~age ~f all respondents, especially former smokers, agreeing ~hat smoking should not be allowed inside public places, was higher on this survey than ~n the 1987 Gallup survey (TablQ 6b). The Gallup question applied to a general statement about re~raining from smoking in the presence of nonsmokers. Interestingly, the Tobacco Institu~e-~ponmor~d survey by ~amil~on, Frederick, and Schneide~s in 1988,~howe~ even s%ronge~ preferences for restaurant and worksite restrictions than the ALA surveys mentioned prevlously (Table ~b). For ~ach of th~ sites, ~he question rmfarred to ~h~ "c~r~en~ ~ollcy" ~B a response cholce; for res%auran~s, %he .current poli~y"mea~ ~ha~ customers mus~ sele~% smoking vs. nonsmokln~ sections; for worksites, employers and employees should de~i~e ~n worksita rm~t~Ictlve policies. In this survey, 2% of respondents favored no restrictions on smoking ~n restaurants compared with 8% i~ the ~I~ survey, and 3% favored ~o r~strlctions on smoking in worksites compared with 10% in the ALA survmy. Between I~64 and 197~, the percentage of r~spondents to the AUT 157

Draft - Do not cite or quote su~eys who favored restrictions on smoking in public places increased from 52~ to 70% (strongly agree and mildly ¸agree) (Table 8). However, the ~estion asked in 1966 was quite different from the ~estions asked in ~e earlier,su~eye (Table 9). Between 1964 and 1975, AUTS respondents who favored more restrictions in,reamed by 18 percentage points. &bout half of respondents in 1966 felt ~at rest~ictions against s~oking were adequate, perhaps because uany more restrictions were in place by 1966. In 1966 and 1975, respondents were also asked if employers have a right to regulate smoking in ~heir places of business. In 1986, 92% felt that the "employer has a right to tell a person when or where he can smoke while on the Job," whereas in 1975, 769 felt that "management should have the right to nroh±bi~smoklng in their places of business." These are vezy different questions: the first concerns management's right to regulate employees, and the second concerns management's right to regulate customers, visitors, and employees. In 1987, respondents to the Harris Poll that was performed for magazine were asked if laws should prohibit smoking in public places or require separate smoking and nonemoking sections, or should smoking in public places not be regulated by law. Among 611 respondents,'29% felt that laws should prohibit smoking in public places, 619 felt that laws should require separate smoking and nonsmoking sections, and only 139 felt that laws should not regulate smoking in public places st ell (3% were unsure). Again, more than 80% of respondents, smokers and nonsmokers, favored restrictions against smoking in public places. 4. Public O~inicn cn Restrictions After Enactment of Law~ Few evaluations of the acceptability of laws banning smoking in public places have been performed. New York City enacted a ban on smoking in most public places, including restaurants, in April 19~. Three months after the ban took effect, a telephone poll of 976 randomly sampled Pew Yorkers (~WCBS-TV poll) revealed that 73% of respondents approved of the law, including 84% of nonsmokers and 43% of smokers (~, July 5, 1988). The 1986 AUTS asked respondents if they would select nonsmoking sections in airplanes, restaurants, and other public places if given a choice, overall, 619 choose noasmoking seating, including 82% of never smokers, 69% of fo~er smokers, and even 149 of current smokers (CDC 1988). Finally, a clean-indoor-elf ordinance that took effect in March 1987 in Cambridge, Massachusetts was evaluated by researchers at Harvard University after three months of implementation. This evaluation study revealed that 78% of Cambridge residents favored 156

the restriction, including 41% (Rigo~ti 1988). ~af~ - DO not cite or quote of smokers and 90% of nonsmokers Many studies have demonstrated ~he blochemlcall¥ measurable exposure of nonsmokers to ETS (PHS 1986). However, only ~ha 1986 &UTS has asked a nationally repreBentatlve sample of residents about exposure to ETS. A subaample of 8,600 working respondents from the AUTS was analyzed with respect to ~eported exposure at the workslte and reported polloles restricting¸ smoking at their worksltes (Table 9). ~if~y-three percent of rlspondent5 who worked in environments with restrictive smokln~ p~llcles still re~orted exposure to ETS. of these, 11% reported that their worksite is "very smoky." Even amonqths 2.5% of roapondents reporting a t~tal ban on smoking in the workplace, 21% reported still being at least somewhat exposed to ETS at work. Theme data help confi~iR ~he notion that workslte restrictions decrease but do not eliminate reported exposure to ETS at the workslte. 6. The Increasina Number of ~olIcia~ILawg Reut~ictin~ Smokino at I~ 198~ 54~ Qf respcndents to the BNA~ASPA survey ~ep~rt~d that tbei~ w~ksite~ had ~e~tri~tive smoking p~l~c~e~ ~p fr~ ~ i~ ~8~ ~ 1~. Th~ 1~8~ figure ~a~ ~ the ~me am tbe p~tag~ of i~di~i~l ~rke~ ~p~rti~g th~ p~ese~e ~ ~h p~lic~s ~n t~e 1~ ~UT~° Among r~spo~nt~ tc the ~5 ~PHp ~rk~te ~urv~ ~.~ o~ w~rk~it~ re~r~d offering s~oki~g ~tr~l ~cti~iti~s~ In~uding ~a~e~ i~rm~tio~ sp~ial e~e~ts~ o~ ~tes~ ~f those ~pe~ie~ ~ ~so ha~ f~rma~ ~o~ing p~l~i~s ~str~oti~ or ~rohi~iti~ ~ ~ddition to f~e~ue~t~ ~it~d be~efi~s-~h ~$ imp~o~ed e~pioy~ m~a~e~ i~p~ ~pl~yee h~lth°-~espo~de~ts ~epo~ted ~e~e~ ~ir and w~rk e~ir~r~m~t~ fewer s~ck~r~ i~ the ~orkf~ an~ fewer compl~int~ ~ n~smok~r~ ~M~ 1~. ~° p~r~ed Fu~u~ ~f~e~t of ~t~i~e ~n~ ~ T~e ~ationa~ $~rv~ of ~eai~h~are Opinions sponsored by the ~ e~d ~ried ~t h~ ~e~r~h and ~r~st~ ~nCo~ i~ ~8~ aske~ ad~lt~ and family physicians ~f r~tr~ction~ o~ s~oki~g i~ medlcai f~i~t~es ~r ~ ~h~ ~oh wo~d h~ e~fe~tlve ~ ~t~pping o~ discouraging smoking. Among ~he nonphysi¢ians, 57% felt that restrictions in medi=al facilities would be effective, and 40% felt that restrictions by employers against smoking on the Job would be effective. Among physicians, 83% felt that much restr$ctions would be effe=tive in health care facilities, and 67% Eelt that restrictions would be effective on the job. These responses should be differentiated from ~hosa ~n o~her surveys that ask ahou~ support for restrictive smokingpoli¢les. The ABFP survey tried 159

the restriction, including 41% (Rigotti 1988). Draft - ~ not cite or quote of smokers and 90% of nonsmokers Many studies have demonstrated the biochemloally measurable exposure of nonsmokers to ETS (PHS 1986). However, only the 1986 AUTS has asked a nationally representative sample of residents about exposure to ETS. A suhsample of 8,600 working respOndents from the AUTS was analyzed with respect to reported exposure at the wcrksite and reported policlas restricting smoking at their worksites (Table 9). Fifty-three percent of respondents who worked in environments with restrictive smoking policies still reported exposure to ETS. Of these, 11% reported that thelr worksite is "very smoky." Even among the 2.5~ of respondents reporting a total ban on smoking in the workplace, 21% reported still being at least somewhat exposed to ~TS at work. These data help confirm the notion that worksite restrictions decrease but do not eliminate reported exposure to ETS at the worksite. 8. The Zncreamino Number of Pollcles/Laws Remtrlctin~ Smokies st the Worksite In 1987, 548 of respondents to the BNA/ASpA survey reported that their worksites had restrictive Smoklnq policies, up from 36% in 1986 (BNA 1987). The 1986 figure was nearly the same as the percentage of individual workers reportln~ the presence of such policies in the 1988 AUTS. Among respondents to the 1985 ODPHP Wcrksite Survey, 35.6% of worksites reported offering smoking control activities, including classes, information, special events, or contests, of those companies, 76.5% also had formal |moking policies (restriction or prohibitlo~). In addition to frequently cited benefits--such as improved employee morale, improved employee health--respondents reported cleaner air and work environments, fewer smokers in the workforce, and fewer complalnts from nonsmokers (RHS 1988). 7. ~slved Future Effect of Remt~ict~ve Smokies Policlea The National Survey of Heslthcare Opinions sponsored by the ABFP and carried out by Research and Fore=eats, Inc., iD 1985 asked adults and family physicians if restrictions on smoking in medical facilities or on the job would be effective in stopping or discouraging smoking. Among the nonphysicians, 57% felt that restrictions in medical facilities would be effective, and 40% felt that restrictions by employers a~inst smoking on the Job would be effective. Among physicians, 83%felt that such restrictions would be effective in health care facilities, and 87% felt that restrictions would be effmctive on the job. These responses should be differentiated from those in other surveys that ask about support for restrictive smoking pollcles. The ABFP survey tried 159

Draft - Do not oitQ or acts to ascertain if respondents thought policies were an effective intervention for smokers to refrain from using tobacco, whereas ~he Gallup surveys tried to ascertain what people want in terms of protecting the nonsmoker from exposure to ETS. Few studies have actually been able to assess the effect of restrictive smoking policies on smokers' behavior, but some studies from individual worksites show decreased numbers of cigarettes smoked per day without a change in the prevalence of smoking (Peterson etal., 1987, Rosenstock et el., 1986). These data indicate an important shift in public beliefs and attitudes toward ETS over ~he last decade or more. The majority of U.S. citizens have recognized that cigarette smoking directly harms the health of smokers (89% of men and 90.9% of women An 1975 [AUTS 1975]; 92% of men and 91.8% of women An 1986 [AUTS 1986]). Moreover, the percentage of survey respondents who believe that ETS also harms the health of nonsmokers has increased dramatically (46% overall An 1974 [Roper 1978] to 81% overall in 1986 [AUTS 1986, Gallup 19881). Even more Americans agree that ETS harms vulnerable populations such as pregnant women and children. Many laws and local ordinances that were put into place during the last decade undoubtedly increased public awareness of ETS issues (PMS 1989). The National Academy of Sciences Report and the Surgeon General's Report on involuntary smoking were released in late 1986. However, not all of the change An belief about harmfulness to ETS can be attributed to the publication of these reports, even though they received enormous media attention) most of the 1986 AUTS had been completed by late 1986. Therefore, ~he increase in reported beliefs about the harmfulness of ETS likely reflects a growing and sustained awareness among U.S. rmsidents rather than merely a public response to the highly visible Surgeon General's Report. This report may have convinced more persons about the harmful effects of ETS, as evidenced by the results of the 1989 Gallup Survey. The slightly discrepant results on attitudes toward laws regarding restricting smoking in public places found in the 1988 AUTS and the 1988 Harris Poll may be explained by the differences in the way the question was asked in this survey. Many laws were put into place by 1986, and respondents may have felt less concerned about increasing regulations than they did in earlier surveys, before these laws were in effect. These laws have been evaluated directly by researchers in some jurisdictions and Indlrec~ly by surveys, and they are apparently widely accepted by both smokers and nonsmokers. There appears to be a trend towards limiting smoking in workplaces. It is unclear whether laws and regulations restricting smoking in 160

D:af~ - Do not cite or ~o~e public places (vhich became widesprea~ in ~he late 1970s) were the s~imuli for policies reshrictlng smoking in the workplace (which are mostly a phenomenon of ~he 19808) or whether simply~e concern for ~e heal~ of nonsmokers is ~e s~imulus for this ~rend. The 1986 A~S ~esul~s, which show ~a~~ even withe total ban on smoking in ~he workplace some workers are expomed to ETS, suggest that there is incomplete enforcement of restri¢tionl. In worksltes where smokers and nonsmokers are segregated, exposure to ETS may result from the inefficiency of separating Bmokers and nonsmokers within the same airspace. The 1986 Surgeon General's Report con¢luded that this level of reetriction wag inadequate to protect the nonsmoker from ETS (PHS 1986)• The 1990 Health OhJectlves for the Nation, which were endorsed by ~he U.S. Public Heal~h Service, race,end that all 50 states have laws by 1990 that both prohihlt smoking in enclosed publlc places and require separate smoking areas in ~he workplace and in dining establishments (PHS 1990). The number and strength of these ,clean indoor air" laws oontlnues to increase at both the state and i=cal level. (PHS, 1989) As of late 1988, 31 states had laws restricting smoklng in publlc ~orksi~es, 13 had laws restricting smoking in private worksltas, and 26 had laws restricting smoklng in restaurants (PHS 1989). Continuing to assess public knowla~qe and beliBfa regarding tobacco use remains important as new information becomes available. These survey results assist publi~ health pr~vlders in measuring the success of policies to control health hazards such as ETS. In addition, these data emphasizl the ¢hanqe in the social mililu surrounding ~ohaccc use. The shlft in public attitudes away from the s~cial acceptability of smoking may increase the pressure for smokers to quit and for potential smokers to avoid smoking. ~olicy-makers ~ay also find it easier to address tobacco issues more directly if they understand ~he public opinions expressed through these surveys. • SUMMARy 1. The majority (81%) of U.S. citizens have recognized that cigarette smoking harms the health of nonsmokers. 2. As of late 1988, 91 states had laws restricting smokin~ in public worksites, i~ had laws restricting smokin~ in private • worksites, and 26 had laws ~estr~cting smoking in restaurants. 9. There appears to be a t~en~ towards limiting smoking in workplace$ ; however, ther~ are indication~ of incomplete enforcement of restrictions. 191

O~:aft - I~ not ¢~te or quote Burlau of National Affairs. Where there's smoke: problems and policies concerning smoking in the~workplace. A BNA special report 2nd ed. Ro=kville, Maryland: B~reau of National Affairs, 1987. CDC. Clgarettm smoking in ~he United States, 1988. MMWR 1987;36(3B):Bs1-ss5. CDC. Passive smoking: Beliefs, attitudes, and exposures--Unlted States, 1988. MMWR 1988;37(15):239-241. Davis RM, Boyd GM, Schoenborn CA. ~co~on Couresty~ and the elimination of passive smoking. Results of ~he 1987 National Health Interview Survey. JAMA IBB0l 253: 2208-2210. Gallup. Survey of attitudes toward smoking. Conducted for the American LUng Association. Princeton, New Jersey: Gallup Organization, July 1985. Gallup. Attitudes toward smoking in remtaurants and fast food establishments. Conducted for the National Restaurant Association. Princeton, New Jersey: Gallup Organization, February 1987. Gallup. Majority backs ben on smoking in public places. Gallup Report HO. 258. Princeton, New Jersy: Gallup Organization, March 1%87. Gallup. On-the-goAmericans prefer smoke-free air. Am J Pub Health 1988;78(5):563. Gallup. A telephone sur~ey of 1549 adults conducted in 1985 for the American Cancer Society. The Gallup Report 1988, No. 268. Princeton, New Jersey: Gallup Organization, September 1986. Gallup. Survey of attitudes toward smoking. Conducted for the American Lung Association. Princeton, New Jersey: Gallup Organization, AugUst 1989. Harris, Louis and Associates. Prevention in America V: steps people take or fall to take for better health, 1987. SUrvey performed for - Prevention Magazine. May 13, 196S~ APPendix B:page 8~ Hamilton, Frederick, and Sohneiders. National Survey of American's Attitudes on Various Publio Policies and Practices. Conducted for The Tobacco Institute, December 1988. National Center for Health Statistics. Smoking and other tobacco use: United States, 1967. Hyattsville, Maryland: National center for Health Statisitics. DHHS Pub. NO. 89-1597. NCHS Series 10, # 169. 162

National Clearlnghouse for Smoking and Health. Adult use of tobacco 1970. Rockville, Maryland: US Department of Health, Education, and welfare. Public Health SerVice. ~une 1973. National Clearinghouse for Smoking and Health. Adult use of tobacco 1975. Rockvilla, MarC!and: US Department of Health, Education, and Welfare. Public Heal~h service, canter for Disease control, J~a 1977. National Clearinghouse for Smoking and Health. Usl of tobaoco: practices, attitudes, knowledge, end beliefs, United states--Fall 1964 and Spring 1966. Washington 0.C.: U.S.Dapartment of Health, Education, and Welfare. Public Health Service July 1969. office of Health Promotion and Disease Prevention. National SurVey of Workslte Health Promotion Activities. Washington, D.C.: U,S. Department of Health and Human Servlues. Public Health SerVice. summer 1987. Paterson LR, Halgerson SD, Gibbons C~4, Calhoun OR, Ciacco EH, and ~itchford XC. Employees smoking behavior changes and attitudes ~ollowing a restrictive policy on workslte smoking in a large company. P~bllc Health Rap 1988;103(2)~i15-120. Public Health Service. Promotlng health/praventlng disease; ohjectlves for the nation. Waahlngton, D.C.I US Department of Health and N~an SerVices, Publi~ Health Service, 1980. Public Health Service. The health consequences of involuntary smoking: a report of the Surgeon General. Ro~kville, Maryland: US Department of Health and Human ~ervices, Public Health SerVios~ Centers for Disease Control, 1985; DHHS publicatlon no. (CDC) 87-8398. Public Health Service. Reducing the Health ConSequences of Smoking--25 Years of Progress. A Report of the Surgeon General. Rockville, Maryland: U.S. Department of Health and Human Servlcs, Public Health Service, Centers for Disease Control, 1989; DHHS publlcatlon no. (CDC) 89-8411. Public Health Service. Major local smoking ordinances in the Dnlted States. A detailed matrix of the provisions of workplace, restaurant, and public places smoking ordinances. Bethesda, MD: U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, 1989. DHHS Puhl. # (NIH) 90-479. Research and Forecasts, Inc. Rights and responslhilities -- a national survey of health Care opinions sponsored by ~he American Hoard of Family Practice. Lexlngton, Kantuoky: American Board of Family practice, 1985. 153

D~a~. - Do not cite or ~ota Rigcttl NA, Stoto M-%, Kleiman M, Schelling TC. Implementation and impact of a cambridge, Massachusetts, ordinance restricting smoking in public places and the workplace. In Aoki et el., ads° Smoking and Hsal~h 1987. Proceedings of the 6th World Conference on Smoking and Wealth, Tokyo, 9-12 November 1987. 3~msterdam: Excerpta Medi=a, 1988. Roper Organization. A study Of public attitudes toward cigarette smoking and the tobacco industry in 1978. New York: Roper Organization, May 1978. Rosenstock IM, Stergachis A, Heaney C. Evaluation of smoking prohibition policy in a health malntalnanca organization. Am J Public Health 1986;76(8):I014-I01S. Anonymous. Support for smoking ban. New York Times, July 5, 1988:B2. 164

Draf~ - Do not tits or quote ?IGUP~S ~ ~XBLZS fOR C~I~Za g • ° ~65

D1"aft - Do not cite or %'uo~e Table i. Surveys With Info~ation on Beliefs, Attitudes, and Zxposures ~o Environmental Tobacco Smoke 1S£: ~mln~ AdUlt Use of Tobacoo '54,'66,'?0, g75,'86 Offi¢a On Smoking and Heal~h Roper organization Gallup SurVey Research & Forecasts '85 Gallup Survey '8? Harris Poll ,87 Gallup Survey '88 Halilton, Frederick '$8 & Schnelders '74,176,t78 183, t85,187ti89 Tobacco Institute American Lung Association American Academy of Family Physicians National Restaurant Asso¢lation Prevention Magazine American Cancer Society Tobac¢o Institute 166

Dratl: - Do not cite or cj[uo~.l Table 2. Beliefs About Harmfulness of Environmental Tobacco Smoke to Nnonsmokers (% of Respondents) by Smoking Status Roper Roper '76 Roper '78 Gallup °S3 Research ForecaststB5 Year Currant Former Nonsmokers Smokers SmoksTs t74 30 57 38 61 40 69 64 80 AUTS '86 69 82 85 NNIS-CEC 187 67 84 89 Ga~lup 'B8 64 86 89 Gallup 'B9 Harmful to adults Harmful to pregnant women Smokers Respondents 46 52 58 84 77 (Physicians-ST) 87 81 82 81 86 88 Harmful to children SS Source: Roper Organization 1978; Gallup Surveys 1983, 1988; Adult Use of Tobacco survey 1986, Research and Forecasts 1985 167

D~-if~ * Do hoe ¢ita or cFto~e Table 3a. Annoyance Caused by Environmental Tobacco Smoke (% of Respondents Reporting Annoyance) by Smoking Status ~urvav ~Un~U~ ~ ~ ~*wr Ell s~ok~rg smokers ~U~t ~ AUTS 1964 20 49 64 69 46 AUTS 1968 26 52 70 48 AUTS 1970 34 63 73 78 59 AUTS "1975 35 72 79 79 63 ROPER 1978 5 60 AUTS 1986 42 73 80 83 88 - N~IS-CEC 1987 34 75 88 69 Source: Adult Use of Tobacco Surveys 1984, 1966, 1970, 1975, 1986; Raper Organization 1978~ NHIS-OEC 1987. 188

D~a£t - Do no~ =i~e or quote Table 3b. Reactions to Secondhand Smoke in Publlc Places, 1987. E~mtx ~ ~iI gm~uum ~ Ask person not to smoke 4 5 4 Move away 52 46 52 Do nothing 40 47 40 DO 8omething else S 3 3 *No~ asked of current smokers Source: 1987 NHIS-CEC (Davis et al.y 1990) 169

Draft - Do not ci~,e or quol:a Table 4. Public Opinion (% of Respondents Who Agree) on Separating Smokers and Nonmmokers in Selected Public Places, 1978 In trains, airplanes, and buses In theaters In eatlnq places At indoor sporting events At public meetings In train, plane, bus stations In work places or-offices In barber or beauty shops Smokin= should be ~ermitted: An~here 91 7 83 11 73 25 73 22 67 28 62 34 61 34 53 42 Source: Roper Organization 1978 170

D~af~ - Do no~ cite Or q-.1ot8 Table 5a. Public Opinion (% of Respondents Agreeing) on Banning Smoking in Selac~ced Public Places, 1978 In elevators In doctors' or dentists' waiting rooms Zn retail stores In theaters At Indoor sporting events • At~ubllc meetings 3n city, stats, or federal buildings In taxis 3n trains, planes, buses In eating places In barber or beauty shops In work places or offices In train, plane, bus stations I 86 69 55 44 34 32 32 32 26 23 21 17 16 l 12 27 41 47 57 58 63 64 65 68 70 73 75 Source: Roper Organization 1978 171

Draft - Do no~ Cite or ~o'ce Table 5b. Public opinion (% of Respondents Agreeing) on Prohibiting Smoking or Retaining Current Policies In Restaurants 74 In Worksites 76 Source: in Selected Public Places, 1988 24 2 20 3 ~amilton~ Frederiokt and S~hnei~eri 1S8S 172

D~a£t ° Do no~ cite or ~ote Table 6a. Should Smokers Refrain from Smoking in the Presence of Nonsmokers? (% of Respondents) by Cigarett~ Smoking Status, 1985, 1985, 1987, and 1989 ~orQB Sm~kina Status Current Smokers 55 62 64 39 37 31 6 1 5 Po=mer Smokers 70 78 76 22 22 19 8 0 5 Nonsmokers 32 85 88 14 15 10 4 * 4 All Respondents 69 7S 77 82 2S 24 19 15 6 1 4 2 DO~Jt XnOW *Less than 0.5% Source: Gallup Surveys 1983, 1985, ~987, 1989 173

Drift - DO no~ cite or quote Table 6b. "If People Wan~ ~O Smoke, They ShOUld Not Do So Inside Public Places where it Might Disturb O~hers" (% of ~espondents Agreeing) 1987 Current smokers 67 22 9 Former smokers 90 1O @ Never smokers 89 5 S All respondents 91 Ii 7 Source: NHIS-CEC 1997 (Davis et el., 1990) 174

Q 0 Q ~llotels and Notels Set Aside O Certain Areas 19/ '"7 '8__29 ! Current Smokers 49 61 ~Former SmokerB 54 72 Non~mokerm 60 68 D Table 7. Opinions Regarding Smoktnq in Selected Public Places and Worksltes, (k o~ Respondents with the opinion) by Smoklnq Status, 1983, 1985, 1987, 1989 Totally Ban Smoking No Restrictions ~nttKnow '8__/3 '87 '89 '83 '87 '8 'e9 /9/ /9/ /ga 7 6 42 30 2 13 9 27 16 6 3 15 14 20 15 S 3 All Respondents Bestaurant_~ Current Smokers Former Smokers Monslmokerg 54 67 63 12 lg 12 30 20 18 4 3 6 Set Aside Totally Ban ~ po Re~trictioms ~ttKnow '8 's~ /~Z /ga /9/ I~Z .tsa' '83 '87 's9 Zg/ '_9/ 'e9 74 79 12 7 13 13 1 1 71 74 19 19 9 6 1 1 65 71 26 23 7 5 2 1 All Respondents Current Smokers Former Smokers Nonsmokers All Respondents Source: 69 74 66 19 17 23 10 8 8 2 1 3 Set Aside Totally Ban • Certain Areas Smo~ki ~o Restrictions ~n'tKnow 'e5 /9/'89 ~ /9~ 'e7 '89 /~/ '85 /9/'89 /9/ ~ '87 '89 64 76 72 ll 4 8 21 19 18 4 l 2 68 80 73 14 12 16 14 6 8 4 2 3 63 80 67 24 9 23 9 10 8 4 1 2 64 79 70 65 17 8 17 21 15 12 11 10 4 1 2 4 Gallop Surveys 1983, 1985, 1987, 1989 175

Dra~ - Do not: ci~l Or ¢~uots Table 8, Restrictions on Smoking in ~11¢ Places (% of respondents favoring increase) by Smoking Status, 1964, 1966, 1970, 1975, and 1986 Current smokers • - 34 35 42 51 23 Fo~-=er smokers 56 58 61 77 53 Never smokers 681 6? 68 82 63 All Respondents 52 52 $7 70 SO *The question for the first four surveys read "The smoking of ciga/ettes should be allowed in flwer places than it is now." The question in 1986 read "There are already enough restrictions on where people can smoke." Source: Adult Use of Tobacco Surveys 1964, 1966, 1970, 1975, and 1986. 176

D~'a~ - ~ not cite or ~ota Tablm 9. Reported Worksi~a Smoking Policies and Wdrksi~e E~osure to Environmental Tobacco Smoke (% of Respondents), 1986 No~ Res~rlcted 95.4 Restrictive 42.1 Total Ban 2.5 % RiDaT~in= R~gUTe to ETS 64.8 53.2 21.1 Source: Adult Use of Tobacco SurVey 1986 177

BLANK PAGE Draft - DO not olte or quota 178

BLANK PAGE 179

Draft - Do ;lot ei~e or gl:ote THE EFTECTS OF PASS~VZ SMOKING AND DAY CAR~ ON ~LESPI~ATORY ILLN2SSES Glen Beneath MRE Office of Pzeventlon# Rduoatlon, and ¢ont=ol National Heart, Lung, and Blood Institute Betbesda, MD 20892 I. INTRODUCTION Reports of the Surgeon General (43) and the National Research Council (27) concluded that children of parents who smoke have more lower respiratory diseases and otltls media. Other reviews (i, 17) showed that children in day care have more upper respiratory illnesses, especially oriels medls. The overlap in ~hese findings ~ises a new concern. Does passive smoking and day care attendance interact to increase the rate of respiratory diseases in infants and young children? This chapter examines the data to determine if evidence exist to support this concern. The chapter begins with a review of the day-care market to show its complexity. Ignoring the diversity of day care might lead to faulty ¢oncluslons and recommendations. 2. DAY C3LRE IN TMH'U. 8. 2.1. G~N2RAL CHARACTHHISTICS In 1982, 6 million mothers (48.2%) with a child under the age of 5 were in the civilian work force.. (285 The mos~ drastic change has been the return of parents to work while their children are infants. (26) These children get care in three basic types of day care delivery systems. They are in-home oars, family day care, and group day care. Parents, relatives, or non-relatives provide in-home care in the home of the child. They also give family day care (day care homesl in a private home other than ~he child's. (2, 50) Day care centers, including nurseries, provide care HOE 12 or more children in nonresidential huildlngs. (17, 49) This sector is almost always subject to government regulation and is the smallest of the 3 sectors. (17) However, centers are the fastest growing segment in the day oars market. (i, 24, 28) 180

Draf~ - DO nob cite or quo~e Table 1 lists the percent distribution of the type of child day care used by the age of the child. Nine percent (9%) of working mothers were able to care for their children while working. Almost one-~hird (30.5%) arranged for in-home care¸ Df their children. However~ day oar~ homes were the predominate source of oareti.e., 48.2%. Table 2 describes the percent distribution of care-givers by ~he age of the child. Relatives provid~lchild care to 29~ and non-relatives provided 27.5~ of all day care to chil~sn of working mothers. Data in Table 3 show that 22t of all children and almost 25% of infants and toddlers got care in ~he home of a non-relative. (28) 2.2. RltgULATZONB The U.S. does not have a national policy on child care and efforts to develop one have reached a stalemate. (35) Sponsors have withdrawn the 1980 Federal Znteragency Day Care Requirements. However, they continue to serve as a guideline for minimum standards. (49) Each state regulates its Own day care faollitles. They have written very tough requirements but enforcement is poor. (35) .. All states have passed regulations which contain some provisions for health and safety. However, they are not consistent. (17) Licensing practices also vary from state to state. ~24) Forty-four (44) states now regulate family day care homes. (49) However, children oared for in their own home are beyond the reach of federal and state policy (17). 2.3. PREVALENC~ OF SMOKING In a 1980 survey, 28.9% of female child care workers smoked cigarettes. This is less than females in general. However, their rates are much higher than those for female elementary school teachers (19.8%) and higher than secondary school teachers (24.8%). (44) 3. RESPZR~TOR¥ ZNFECTIO~8 3.1, MAGN~TUDB OF TZ2 PROBL2M Upper respiratory infections are the most common diseases affecting children under 5 years of age. They are important causes of chilShcod illness and their treatment consumes a largo portion of health care resources. (8, 14, 17) Infants average 7-8 acute respiratory infections per year. Older ¢hildren~ 1-5 years of age, • average one or two fewer infections than infants. (17) Acute otitis media (AOM) is the most common complication of upper respiratory diseases in infants and young children. (16, 1~, 31, 47) AOM is the largest single cause of morbidity with po~slble sequelae in children. (47) Recurrent episodes are also very common iSl

Draf~ - Do not cite or ~o~a in children during ~e first years of life. (22) AOH accouflt for one-third of pediatric office visits (31) end ~ree-fou~hs of follow-up visits. (16) Nearly all children have at l~ast one episode with effusion (0~) during their fira~ 6 years of life. Some develop chronic O~ or chronic otitia Jedia with perforation and discharge. (45) Repeated episodes of O~ in early life may lead to transient or permanent hearing lose and L=peded speech. These conditions may then lead ~o developmental or educational delays. (17, 47) Sronohioli~ia is the most conon manifestation of lower reapirato~ infections in infants and small children. The t~e incidence is unknown. However, about i0 per 1,000 infant are hospitalized with bronchiolitia. The mean age for respiratory $yncytial virus (RSV) bronchiolitls is 7.8 months and the peak age is 2 months. Half of children hospitalized for the condition are under 3 months of age. (42) 3.2, DAY-CARE AND RESPIRATORy I~ECT~ONS Respiratory see are the moat common ailments affecting dhildren in disaSters.day (1) TOday, infants and pre--school age children get infections at earlier ages and are spending more time outside ~he home. A common factor in this changing pattern is the increasing popularity of day-care canters. (24] Day care centers with many children in the same place create favorable conditions for respiratory epldemloa. (30) However, the total burden of respiratory dlaeesea seems no greater for the day care child. They simply occur at younger ages. (i) The association of day care and respiratory diseases began in the 1920~s. (17) In the 1970's, Scandinavian researchers (19, 23, 31, 22, 39, 41, 45) found an increased rate of otltis media among children in day care. Children in centers had the highest rate. Those in family day care held an Inter~edlata position between centers and in-home care. Moreover, home-reared children with abnormal findings at first testing were significantly more likely to have normal results at subsequent testings. There are obvious dlfficultiea in transferring the results from studies conducted in Scandinavian countries. However, Haakins (17) concluded that the high quality of these studies lake the findings "worthy of careful attention. The~,show that children in day care are at 2-3 times the risk of otit~ ledla as chose reared at home. Two American research teams (14, 42) ocnflrmed the Scandin~vlan results. Visscher and colleagues (47) studied patients in a large pediatrics group practice in Minneapolis. They collected data on every child attending the clinic during a 2b-~week period in February, 1982. Cases were patient presenting with AOM on a study day. Controls had a diagnosis other than AOM and no prior history of otltis. Attending a day care facility was the second ~ost 182

D~'aft - Do not cite or guo~;e important risk factor. The risklncreased with the number of other children at the facility. Exposure to smokers was not a risk factor. Fleming and colleagues (14) also studied childhood infections in Atlanta. They found that c~ildren attending day care were slgnificantly morB likely to have an upper resplratoz~ tract infections durinq a 2-week. Maternal maoking also increased the risk. The effects of attendlnq day cars and smoking mothers were independent. Aqe and livinq In ¸crowded conditions were also risk factors. The researcherl estimated that 31t of upper respiratory infections can be attributed to day care attendance. Most studies of bronchlolitis focused on children in hospitals. CamprQhensive studies of this disease in ambulatory patients or day-cars centers are lacking. (42) However, s Chapel Hill, NC study compared the rates of bronchiolitis in • day care center and a pediatric practice. The rats was much higher in the day care oen~er for childran 6 Bon~hs Qf age or youn~er. However, the proportion of cases requiring medical treatment and hospitalization ._ Sas less among day care children. (I0) Reviewers (i, i?, 18, 19) have Identified problems which limits the generallzation of the findings from these studies. They are: ~ i. Control groups were less than satisfactory. Researchers observed children in day care more fre~ently than those in home care. 2. Some studies reported symptoms while others used diagnostic categories. "3. The ages of children studied and the manner of reporting illnesses by age category differed widely. 4. The reliability of case-controlled and ochor~ studies depends on the accurate quantification of disease occurrence. This raises the questions of whether day care parents seek a physician fo~ their children,s illnesses more frequently. When a day care provider suggests taking a child to a physician this might have important effects on parents. 5, MOSt studies did not control for other factors that probably influence the incidence: of respiratory 111nesses. These factors include housing, humidity, ventilation, passive smoking, and other air pollution. Nonetheless, reviewers conclud~ that most studies have shown an increase in respiratory diseases among children in day care. There is stronger evidence for initial and recurrent otltls ~edls. (17) The rate of otitis is greater in large group day care centers and 183

Dz~af~ - Do not cite or guote probably smaller day care homes. (2, 17, 18, 31) Studies also show reduced rates of ho~h symptoms and acute illnesses with increasing age in all sites. However, there is very little evldence of an excessive rate of illness in day care children for the more serious respiratory diseases. (17) Aaymptomatio children do not have higher levels of respiratory !tract pathogens or even different pathogenl than children reared at home. (i, 17, 40) However, The true incidence of infectious disease in family day care is unknown since most of it is unlicensed. ~2) 3.3. pAHSX~ SMOKZHC ~ DIHZAHHH IN XHF]~TB AND CHZLDHZN A number of studies have demonstrated a positive association between passive smoking and lowe~ resplratory symptoms (4, S, IS, 36) and diseases. (ii, 12, 22, 29, 40) The effect was stronger in infants. Maternal smoking, when measured, showed a high correlation. However, paternal ~oking was rarely significant. Studies on the relationship of passive smoking to the development of bronchlolitis are less clear, Two studies ~32, 38) showed a positive association wlthmaternal smoking. However, another study (29) did not find a relationship. .Otitis media is the only upper respiratory disease reported in the literature as being associated with passive smoking. Five studies (3, 20, 21, 30, 34) showed an inorassed incidence of otltis media with maternal smoking. However, in five other studies (14, 39, 45, 46, 47) parental smoking was not significant. However, the study by Fleming and colleagues (14) included only 34 cases among the 575 children wi~h upper respiratory illnesses. Pukander and oolleag~es (30) also suggested that day care attendance may mask the effect of parental smoking. Two comprehenslve reviews (27, 43) concluded that lower respiratory diseases and otltis media occur more frequently in children with mothers who smoke. TWO researchers (29, 48} offered explanations for the association with only maternal smoking. They argued that children are more likely to be with their mothers at the times smoking occur. Some mothers also remain at home with the child. This suggests ~hat the duration of expos~tre to smoke rather than Just the presence of a smoker is the more Important factor. Both repo~cs (27, 43) emphasized the need for caution in the interpretation of these studies, Zndependent risk factors, such as age and sex, were not always taken into acuount. The use of questionnaires to collect info~etlon on symptoms are prone to recall bias. Most studies examined only the effects of exposure to parental smoking, excluding exposures outside the immediate family. Future studies must control for potential confounding variables. /" 184

Draf~ - Do no~ cite or quote 3.4. D~ ~ AS ~ CONFOUNDING V~RZ~BL~ Passive smoking increases¸ ~he risk of upper and lower respiratory diseases in infants. Oay care attendance also increases the occurrence of uppe~ respiratory infections and perhaps some lower resplratory illnesses in infants and toddlers. However, studlas~focusing primarily on the effects of passive smoking did hOt,control for day care attendance. Many of ~he studies on day~care~ Infections did no~ consider parental smoking as a possible confounding variable. Moreover, none of ~he a~udlas in either area considered the smoking habits of day care workers* Seven of the day cars studies (14, 19, 30, 39, 45, 46, 47) did consider parental smoking. Two of these studies (14, 30) found an independent effect for both day care attendance and maternal smoking. The effect of day care was strongest in both cases. The remaining studies showed a statistical si~iflcancl for day care attendance only. It is unfortunate that researchers have ignored the smoking h~blts .~mf day care givers. Especially since the duration of exposure is portent. (29,.4~) Smoking by day care workers exposes the child to smoke. Th~ Section on ~llergy of ~he Canadian Pediatric Asscclation (37) provided support for thla premise. They reported that infants admitted to hospitals for chest problems bad significantly more day care givers who smoke than did control infants. The smoking practices of workers in day care homes deserve special attention. This sector includes more children and is especially popular with mothers of infants and toddlers. Day care providers who smoke probably spend as much time wi~h ~hese children as their mothers. Thus, the smoking habits of these workers potentially confound the results of studies of the effect of parental smoking. 4."4 RECOMMENDATIONS 4.1. REGULATIONS Existing day care regulations clearly are deflclent in mandating a safe and healthy day care envlr0nment. Federal regulation, while desirable, is not possible now. The prevailing attitude today is away from federal intervention and toward state and personal responsibility. (49) The regulation of day care homes, which contain ~he most children, is an especially" delicate issue. Increased regulation of homes might have the effect of actually decreasing the availability of this mode of child care. (2) Moreover, the sheer number of providlrs and the small size of these units would make effective oversight difficult. (17) 185

D~aft - Do not cite or quote The most impotent interIR steps are to give parents better information and improve state regulations that ~elate to health. [17) Requiring all facilities to have written policies on heal~h and give copies to parents is a etar~. Parents should also be informed about ~he potential interaction between passive smoking and day care on the risk of early childhood infections. (14) 4.2. llttS~ The sparse data available regarding family day care make it imp0r~ant to pay more attention toY J lie mode. SiNce day care homes includes more children, partlcule~ly infants and toddlers, it is importan~ to understsnd the dlsease experlsn=es In these homes. (2) Surveys are needed to determine the smoking patterns of day care workers. Data from the National Heslth Interview Surveys, 1978-1980 put the prevalence of smoking among female child care workers at 28.9%. However, these data excluded private household child care workers. 4.3. EDCCAT~ON ~resently, Parents must Judge for themselves the quality of care given ~o their children. However, most parents do not know what to look for in a day care setting and there are no federal standards. (35, 50) There are, however, guidelines that the child development community supports. (35) There is also s checklist that =an dlffersntiate between centers of high and low quality. The ¢heokllst includes one ite~ on smoking: "Adults do not smoke in rooms where children are." (7, 35) Ea~cation sffo~s to disseminate this information are needed. Low-cost materials must also he available to day care providers. (i) 5. CONCLUSION The children of woc~ing parents ere receiving day care primarily in their own home, family day oars homes, and day care centers. Family day care is ~he largest of the three sectors hut day care centers represent the fastest growing segment. S~udies, mostly in Scandinavian countries, have demonstrated that children attending day care have more respiratory inf@ctlone. The effectwas stronger among infants and toddlers. Another group of studies have llnked parental smoklng, primarily maternal smoking, with an Increase Ln respiratory dleeases among infants. However, most of these studies did' not control for atte~in~ day care. Th~ few studies that cont~olle~ for parental smoking and day cars showed a consistent and positive association for day care. Parental emoklng was less cleor. None of the studies, however, controlled for the exposure to smoke from day care workers. 186

Studies controlling for potential confou.di a c o needed in this area. The smoking practices of day care workers, particularly day carg homes, may have been a major uncontrolled factor in past studies. In the interim, parents must be educated. They must know about the bagful effects of parental smoking and the pot~tial for added exposure from day care Workers. ~ Some S8% of Dn@ sample attended full-tlme day care, i.e., 40 hours o~ more per week. (14) Thus, children of nonsmoking parents are not wlthout risk. Staying with day care smokers may increase their exposure to smoke similar to that with smoking parents. Children of smoking parents may face as much as twice the exposure. This has special implications for day care homes. First, the chil~Ten are yo~ger. They also spend most of the time in a smaller environment ,with other children and the day care worker. If smoking occurs, the expos~tre should not be materially different from that found in the home. Day care providers must also know a~ut the posslhle interaction of passive smoking and day car~ attendance. Those in day care homes, particularly, should not smoke in the presents of hhe children. Since strict regulatlons of this sector in not possible, the parents must insist upon this practice. Day care centers, while providing a different environment, should adhere to the same principle. They are similar to the school system where teachers can not mmoke in the classroom. State regulatory agencies should also include this provision in the licensing of day care facil±tles. / SUMMARY 1. Studies have linked both parental smoking and day care attendance with increased respiratory infemtlons. Smoking by daycare workers may have been "a major uncontrolled confounding factor in studies of infections caused by maternal passive smoking. 2. Parents and daycare providers should be educated to know about the harmful effects of parental smoking and the potential for added exposure from day care workersa which could double total ETS exposure. 3. State regulatory agencies should include prohlbltlons against smoking in daycare as they do in Classrooms. 187

Draf~ - Do not cite or quote 6. REPE~NCZB 7. 8. Aronson, SS; Osterholm, M. "Znfectlous Diseases in Child care: Management and Pre~entlon. Su:=ary of the Symposium and Recommendations;" Revlevof Infectious Diseases; 1986, July-Aug; 8(4): 672-679 * 2. Ba~lett, AV; O~on, P; ~er, M "Day Care Homes: The Silent Majority of child Care;" Review of Infectious Diseases; 1986, July-AugUst; 6(4): 663-671 3. BlaCk, W. "'l~e Aetioiogy OE Glue Ear: A Case-Control Study," International Journal of Pediatric Otorhinola:~rngology, 9(2); 121-133; July, 1885 4. BurChfiel, CM; Higglns, MW; Keller, JB; Butler, WJ; Howatt. WF; Higgins, ITT; "Passive Smoking in Childhood: Respiratory conditions and Pulmonary Function in Tecumseh, Michigan," American Revlew of RespiratoEy Disease, 133 (6); 966-873, June, 1986 S. Charlton, A. -Children,s Coughs Related to Parental Smoking," British Medical ~ouz~al, 288(6431): 1647-1649; J~e 2, 1984 S. Cherish, A and Feldman, w° Personal uommuni¢atlons reported in: Section Allergy, Canadian Pediatric Association; • secondhand Smoke Worsens Symptoms in Children Wlth Asthna;" canadian Medical Association Journal; 1988, August 2~ 135(4): 521o323 Clarke-Stewar~, A. Daycera, Cambridge, MA: Harvard University Press, 1982 Cypress, 9K; "Pattern of A~ulatory Care in Pediatrics: The National A:bulatoryMedical Care Survey: U.8o, January 1980 - December 1981," in Vital Health Statistics, Series 13, No. 75; U.S. Departhent of Health and Human Services; Publication No. 94-1798; Government Printlng Office, 1983 -9. Denny, FW~ -Childhood Acute Respiratory Tract Infections Deserve Our Attention;" Alerican Journal of P~bllc Health; 1988, January; 78(1): 18-17 I0. Denny, FW; collier, AM; Henderson, FW; Clyde, WA; "The m Epidemlology of Bronchiolitls, Pediatric Research, II: 234-236, 1977 ii. Evans, D; Levison, M; Feldmen, C; Clark, N; Wasilewlskl, Y; Levln, B; Hellins~ R. "The Impact of Passive SMoking on 182

Emergency Room Visits of Urban C~'~ ~i~t~ quota American Review of Respiratory Diseases; 1987; 13S: 567-572 12. Fer~sson, DM; Holm~ood~ L~ Shannon, FT; Taylor, R. "Parental Smoking and Lower respiratory Illness in the First Three Years of Life," Journal of Zpidemiology and Conunity Health, 35(3): 150-154; Sept~er, 1981 13. Ferris, 8G~ Warep 0~4; Berkey, CS~ ~kory, DW; Splro Ill, A; speizer, FE; "Effects of PasslvO Smoklng on Health of Children," Envlron~ental Health Rerlpectives, 62: 289-295~ 1985 14. Tleming, DW; Cochl, SL~ Hightower, AW~ Broome, CV; ,,Childhood Upper Respiratory Tra~ Infections: To What Degree is Incidence Affected By Day Care At~endance~" Pediatrics; 1987, ~anuaryl 79(i): 55-60 15. ¥osbur~, S~ Family Day Ca~Q In The United $~ates: Summary of Tindings~ Government P~inting office, 1981 16. Giebink, GS; -Epidemloloqy and Natural History of otitls Media;" in Lim, DJ~ e~ el; Recan~ Advances in Otltls Media with Effusion; 1984; 5-$ 17. Hasklns, R; "Day Care and Illness: Evidence, Costs, and Public Policy;" P~iatrics! 1986~ ~?| 951-982 18. Henderson, FW~ Gieblnk, GS~ "O~i~Is Media Among Children in Day care: ~ Epidemlology an~ Pa~enesls;" Review of ~nfectlous Diseases~ 1986, July-Au~ust~ 8(4): 533-5~8 19. Ingvarsson, L; Lundgren, K; Olofsson, B~ "Epidemiolo~y of Acute Otitis Media in childr~n-A Cohor~ Study in an Urban Population;n in Lim, OJ~ et el; Recent Advances in Otitls Media With ~ffuslon~ Philadelphla: B C Decker~ 1984; 19-22 20. Iverson, M; Birch, L; Lundqvlst, G; Elbrond, 0. "Middle Ea~ Effusion in Children and the Indoor Environment: An Epldemiol0glcal S~u~y," Archives of Environmental Health 40(2): 74-79~ March-April, 1955 21. Kraemer, M~; "Risk Factor for Perslsten~ Middle Ear Effusions;- Journal of American Medlcal Associatlon~ 1983, February 25~ 249(8): 1022-1025 22. Leeder, SR; Corkhill, RT~ I~W~g, I2~; Holland, WW. "~nfluence of Family ~ac~o~s on the ¸Incidence of Lower Respiratory Illness During the First Year of Life," British Journal of 189

23. 24. 25. 25. 2"7. 28. 29. 30. 31. 32. D'CaI"~ - DO no1; ¢ita or quote Preventive and Social Hedlclne, 30(4) : 203-212, Decesher, 1976 Lundgren, X; Ing~arsson, L; Olofsmon, B; "Epidemiological Aspect in Children with Recurrent Acute Otltls Media;" in Llm, DJ; et al; Recent Advan¢s$ in Otlti8 Media Wi~ Effusion; Philadelphia: B C Decker; 1984; 22-25 MarvicJ¢, C; simmons, K; "Changing Childhood Disease Pattern Linked with Day-Cars Boom;" Journal of American Medical Association; 1954, March 9; 251(10): 1245-1247, 1250-1251 McConnochie, K; Mall, C; Barker, W; "Lower Respiratozy Tract I11ness in the First Two Years of Life: Epldemlologic Patterns and Costa in a Suburban Pediatric Practical" American 3ournal of Public Health; 1988, January; 78(i) : 34-39 Morgan, G; Stevenson, C~ Fiene, R; Stephens, K; "Gaps and Excesses in ~he Regulation of Child Cars: Repo~ of a Panel;" Revlew of Infeotlou8 Diseases; 1986, July-August; 8(4): 634-643 Natlonal Research council; ~nvlror~ental Tobacco Smoke - Measuring Exposure and Assessing Health Effects; Washington, DC: National Academy Presst ~86 OTCor~sll, M; Rogs~I, CC; "Child Care Arra~gsme~ts of Working Mothers: June 1982;" Currant Populatio~ Reports (Bureau of Census)~ 1982~ ¸Special $~udies P-23~ NO. 129 Pedrelra, F; Guandolo, V; Feroli, E; Mella, G; weiss, I; ,Involuntary smoking and Incidence of Respiratory lllness During the First Year of Live," Pediatrics, 1985; 75: 594-597. Factors Affecting the Occu~renoe of O~iti~ Medla Among 2-3 Year Old Urban Children;- Acts Otola~yngology [Stockholm]: 1985, September-October; I00(3-4): 260-265 Pukander, J~ Sipira, M; Kaz=a, P; • "Occurrence of and Risk Factors in Aoute Otlti~ Media;" in Lim, DJ~ ~t al~ Reoen~ Advances in Otitis Media With Effumion; Philadelphia: B ¢ Ducker~ 1984; 9-13 ~llan, CR; Hey, EN. "Wheezing, AsCJ1ma, ~nd Pulmonary Dysfunction 10 Years After Xnfoction Wi~h Respiratory syncytial virus in Infancy,¸" British Journal of Medicine, 284(5330): 1665-i~s9, June 5, 1982 190

33. 34. 38. 36° 37. 38. 39. 40. 41.~ 42. 43. 44. Draft- ~eciuote Center: summary Findings and their Implications; Cambridge, MA: Abt Books: 1979 Said, G; Zalokar, J; Lellouch, J; Patois, E: "Parental Smoking Related TO Adenoidectomy and Tonsillectomy in Children,- Journal of Epidemiology and community Health, 38(2)| 97-101; June, 1978 Scaz-~, B; 1984 Mother Care, O~her care; New York: Basic Books; Schenker, MB; Samet, J'M; Spli|er, FZ "Risk Factors for Childhood Respiratory Disease: The Effect of Host Factors and Home Environmental Exposure," Amerlcan Respiratory Disease, 128: 1038-1043; 1983 Section Allergy, Canadian Pediatric Assooiatlon; "Secondhand S=oke Worsens Symptoms in Children Wlth Asthma:" Canadian Medical Association Journal; 1986, Au~p~st 2; 135(4): 321-323 Sims, DG;. DoWnham, M; Gardner, PS; ,ebb, Jr Welghtman, D. "study o~ 8-Year*Old Children With A History of Respiratory 9yncytlal Virus Bronchiolitis in Infancy,. British Gournal of Medioine, 1(6104}: 11-14, January 7, 1978 Stahlberg, MR; "The Influence of Form Day Care on the Occurrence of Acute Respiratory Tract Infections Among Children;" Acta Paedlatri= Scandinavla [Supplement]; 1980; 292: 1-87 " Strengert, K: Carlstrom, G; Jeansson, S; Nord, DE; "Infsc~lons in Preschool Children In Group Day Care," Acta Paediatric Scandinavia, 69: 455-463, 1976 Strangert, K; "Respiratory Illness in Preschool Children With Different Forms of Day Care," Pediatrics, 57(2): 191-196; February, 1976 Task Force on Epldemiology of Respiratory Diseases; Epidemlology of Respiratory Diseases: Division of Lung Diseases, National Heart, Lung ~& Blood Instltute; Nove~er, 1991 Public Health Se~ice, The Health Consequences of Involuntary Smoking: A Report of the Surgeon General, U.S. Department of Health and Human Services, Rockv111e, MD: Government Printing Offloe, 1986 Public Health SerVice, The Health Consequences of Smoking: Cancer and Chronic Lung Disease in the Workplace, U.S. 191

45. 46. 4?. 48. 50. Draf~ - Do no~ =i~e or quota 0epartmen~ of Heal~h and Human Servlcas, DHH$ (PHS) 85-50207, 198S Van Cauwenber~e, PB; Kluyskens, PM; "Soml PTmdlsposing Factors in Otltis Media wi~h Effusions" in ~im, DJ; e~ al~ Recen~ Advances in Otitls Media Wi~hEffuslon; Philadelphia: B C Decker~ 1984r 28-32 Vin~her, B~ Elbrond, CB~ "APopulation study of Otitis Media ~n Childhood,- Aa~a O~ola~yngo1~, [$tockholmJ Supplement 360: 13S-13T~ 1979. Visacher~ W; Mandel, JS~ Ba~alden, PBt Rusa, JN; Gieb~k; GS; "A Case-Control S~udy Exploring Posslbll Risk Fmctors ~or Childhood O~i~Is Media~" in Li~, ~J; e~ al~ Recen~ Advances In O~i~is Media With Effusion; Philade1~ia: B C Decker~ 1984; I~-15 Ware, JH; Dockery, D; Sp~ro, A~ Spelze~r F~ Fe~rls, B. "Passive Smoking, Gas Cooking an~ RespSra~ory Heal~h of Children Living in ~ Cl~ies~" American Review of Respiratory Diseases; 1984, March: 129(3): 366-374 ¥~ung, KT and Zi~ler, E; "~nfan~ and Toddler Day care: Regulations and Pollcy Impllca~ion," American Journal of Orthopsychia~ry, 1986, January; 56(I): 43-55 Zigler, E~ Muenchow~ S; mI~fec~ious Di~easQs i~ Day Ca~e: ~arallels Between Psychologically and Physically Hea1~hy Care;" Review of Znfec~ous Diseases~ 1986, July-August; $(4)~ 514-520 192

7 • T~ T.~ 8 Owaf~. - DO no1= c:L'~s or quots TYPE O1 C~EZLD CARl BY AGZ OF CHZZ,D TLI=lo I AGE < 1 Year IN-HOME CARE DAYCAR~ HOME GROUP CARE MOTHER TOTAL Source : 1-2 Years 3-4 Years 34,3% 33.3% 42.?% 43.0% 5.3% 11,7% 9.2% 8*6% 91.5% 96.6% O'Connell and Rogers, 1982 (28) TOTAL 24.6% 30,5% 35.4% 40.2% 25,8% 14.8% 9.9% %,1% 95,7% 94.6% 193

Dz'a~t - Do no1: o:L~e oz' q'ao~o AGE < ~ Year GZ'V~R BY &GZ 03' CHZLD TabZo 2 ~-2 Yearm 3-4 Year8 TOTAL FA"~":.~ 13.9% 15.8% MOTHER 9.2% 8.6% GHANDP.~J~'I' 22.4% 16.$% OTHER RELATZVZ 11.3% 12.3% NOt~ELATIVE 29.4% 31.4% GROUP CARE 5.3% 11.7% TOTAL SOUrCe: 91.5% 96,6% O'Connell and Rogez's, 1982 (28) 11.0% 1~.9% 9.9% 9.1% Z4.6% 17.2% 12.8% ~2.1% 21.6% 27.5% 25.8% 14.8% 95.7% 94.6% 194

Drsf~ - Do no~ Cite or q~.~ota CAREGZVZRS BY T~PE OF CHILD CARB AND AGE OF CXILD Table 3 AGE < 1 Year I-2 Years 3-4 Years TOTAL Z Grandparent Other Relative Non-relative • AY CARE HOME Grandparent other Relative Non-relative GROUP CARE Nursery Day care Center MOTHER 9.2% 15.9% 11.0% 13.99 8.9% 6.3% 3.6% 5.9% 5.1% 5.0% 5.7% 5.2% 9.4% 6.2% 4.39 5.5% 13.59 10.59 11.09 11.39 S.29 7,3% 7.19 6.9% 23.0% 25.2t 17.99 22.0"I; 1.79 3.2% 11.79 5.5% 3.6% 8.5% 14.1% 9.2% 8.69 9.9% 9.19 TOTAL 91.59 96.6% 95.7% 94.6% Source: O'Connell and Rogers, 1982 (28) 195

D~a2"c - DO ncY¢ ¢4~I or q'doe.o 7ZGURE~I ]~qD 2'~BL]~B ]?OR C'~ZI~ 10 196

Dra~ - Do not oits or quote CHAPTER 11 NO SMOKING POLZCZZ9 HT T~ WOP~V~TZ A Look at What Companies ~re Doing Today ~uhh hehrense Washington Buminem| Gzoup on Xealth The move~nent of businesses to develop and impla:ent smoking control policies appears to be strong, and may even by gaining momentum. A national survey released in 1987 by ~he Office of Disease Prevention and Health Promotion, U. S. Depa~snt of Health and Human Services, found that 27 percent Of all U.S. companies with 50 or mort employees had a formal smoking policy. Of ~hese, 40 percent reported the policy was in place to protect nonsmokers; 40 percent reported ~he policy was designed to comply .with regulations; 13. peccent reported a need to protect equipment, and ~'percent advised that ~he policy was designed to protect employees at high risk for health problams.I A more recent study that looked "only at large and medium-slzed companies, the 42nd annual Northwestern University Lindqulst- Endicott Report, found that 70 percent had restricted, or were prepared to limit, sm~klng in the workplace. The study was released in early" 1988. Another study oE 916 large and mid-sized U.S. companies, conducted in 1989 by Hay/Huggins, a management consulting firm, found that Sl percent of surveyed firms wlth revenues of $i billion or more re~trict smoking; the p~rcent dropped to ~5 percent for companies grossing less than $200 million per year. A 1989 survey by the Gallup Organization commissioned by the American Lung Association found that 21 percent of individuals surveyed supported a total ban on smoking at the work.sits, with an additional 85 percent in favor of smoking only in designated areas.& The development and implementaticnof a no smoking policy within a business is a multl-faceted proGees. Experiences of the growing number of companies that have developsd written statements spelling cut how smoking will he limited or prohibited i~luBtrates vividly that the process Invalves many individuals and groups, and that deliberations often are emotionally charged. 197

DTaft - Do not cite oE quote This chapter contains a series of case studies outlining how several companies have successfully approached this process. But before discussing company-speciflc examples, there is merit in examining some of the key issues that must be looked at by any company considering developing a no smoking policy. Among the questions to be answered are~ o What kind of specific smoking restrictions are beet for the company? o What h.neflts can be realistically anticipated from the policy? o How should employees he Involved? o How should unions be involved? o What kind of education should be offered and to whom? o What kind of incentives should be offered? o How should the policy be enforced? Further details about each of these steps arm contained within the "Case Studies" section of this chapter. OBtlons for Smokina Restrletlons Restrictions on smoking in the wo~kslte are not new. For years--even decades--businesses have had polloles that banned smoking in specific areas such as elevators, hallways, auditoriums, sections of cafeterias, laboratories, rooms with delicate equipments, etc. In many instances, these restriotions were imposed because of laws or ordlnancse requiring them or to protect property. Bmfors the 1980s, they were seldom implemented for health reasons. The assumption wee, of course, the entire company is considered a Smoking Permltted~rea unless otherwise specified. Another type of ~ollcy began appearing with regularlty in the mid end late 19S0s. It banned smoklng throughout the company except in designated areas. While many of these policies did not necessarily put greater limits on smoklng--often allowing offices and work areas, special lounges,; large pacts of the cafeterias, etc, to be designated as Smoking Permitted areas~ they did set the precedent that the company is Smoke Free except in specified areas, While the difference between these two types of policies may seem subtle at first glance, there is a strikingly different corporate 198

~aft - Do no~ mite or q~ote philosophy underlying the c~o approaches. And in the late 1980s, i~ was this latter approach--establlshlng a smoke free company, possibly with a few, carefully selected areas ~bat permit smoking- -that appeared to be se~ing the pattern for worksite smoking policies. ACCording to a spokesperson for Texas Instruments, Inc. (TI), determining what approach to take in limiting smoking was the most difficult aspect of developing and implementing its policy. AS TI and many o~her companies have found, designating even a few smoking areas within a company can still pose serious health hazards for employees. Smoke from lounges, caf~terlss, hallways, and enclosed offices, gets into the ventilation system and is circulated throughout the building, including into no smoking sections. (Sea "case Studies: Pacific Northwest Bell.") As an interim step in a phased-in nonemoking work environment, Pacific Mutual Life Insurance Company, Wew~or~Beach, California, installed electronic filters in the temporary smoking area of its cafaterla. TI chose to avoid this problem by eliminating smoking from the worksite except for designated smoking areas which were, to the extent possible, separately ventilated. Similarly, the ~eadqua~ers complex of General Telephone Of California prohibited smoking in all areas except a small portion of the oafeterla that has its own ventilation system. For others like Amsrlcan Family Insurance Group, Madison, wisconsin, Pennsylvania Blue Shield, and UNUM Life Insurance Company, Portland, Maine, the choice to provide separate ventilation was either too expensive or physically impossible, so they chose to ban smoking completely at the worksi~e. On October I, 1987, Ralston Purina's headquarters in St. Louis, Missouri, became the first Fortune 500 company to completely ban smoking in its facilities. Clearly, more and more companies are banning smoking or severely limiting it. Most require that all visitors abide by the company's requlatlons. Some will not allow smoking on company property, including grounds and parking 10he. GroUps like MIohlgan ~ell, which has a large number of motor vehicles, are expanding their bans to all company-owned vehicles. However, others are voiding a ban in company cars and trucksbecause they believe enforcement will be vlrtually impossible. A few companies have gone even .fllrther, and may be bellwethers for a future tend. These companies require that all hew employees sign a statement that they are nonsmokers, even on their own time. Company pollcles prohibiting the'hirlng of smokers got natlonwida publicity when Acoustical Products Companyi~ a subsidiary of Chicago-based US Gypsum Corporation, announced that because of exposures to fibers that could have adverse health effects, all present workers were required to quit smoking or face termination, 199

Draft - Do not ¢itl or q~ote and in the future, only nonsmokers would be hired. The Non-Smokers Inn in Dallas, Texas, provides on~nonsmoking rooms and hires only individuals who do not smoke. At Cardinal Industries, columbus, Ohio, new employees must state on the application fog whether or not they smoke, and only nonsmokers will be hired; but the company does not make any sffor~ to validate applicants' statements. Loulsimna Pacific Corporation, s~Por~land, Oregon-based national timber company with 15,000 employies,Ldces not hl~s smokers in any of its plants or in oorporsts Offices "because of ~he =edical costs, absenteeis~, envi~or~snt of smoke ~nthe yes.lace, tha fire problems in the =ills, and lung cancer."° The vast maJori~ of companies s~ill do not require ~hat new employees be nonsmokers. But many companies wl~h strict bans are seeing fewer smokers apply. -why¸ would i s~oksr want to work for us," one company spokesman said, "when we deprive his of his habit for eight hours every vorkday?" Dgn~fits of NO Smokina ~elieles Developing and implementing a vorksits no smoking policy may not be easy and may cause some disc~mfor~ for smokers and managsmsh~ alike. So vhy do companies do it? What benefits do they receive? In a recent national survey of all types of worksitss with 50 or more employees, the Office of Disease Prevention and Health Promotion, U. S. Depart:snt of HQslth and Human Services, asked ~hoss wi~h s~oking p=ograms what benefit8 they perceived. o 41 percent said smoking control policies and programs improved employeesr health; o 16 percent said they increased employees' productivity; o 9 percent said they improved morale, and o 8 percent said smoking control ac~ivities reduced costs.1 Some companies have conducted evaluations of the ~esults from their smoking con~rol errors. Several Of ~hese studies, along with some anecdotal findings, are reposed in this chaptQr's Appendix, "The Econ~mi~ ~ustiftcstion for NO $~oking Policies." To many Co,panics, a rlduction in the numbe~ or percent of e=ployees who smoke is benefit ~nough from a policy. Smokers d~opped from 21 percent of the workforce to 1~ percent in two¸ years at U~ ~ife ~nsurance Company. In addition, 87 pe=cen~ of the smokers reported they were smoking leSS after the policy was 6 i~ple=sn~sd. At Pacific No~hwlst Bell, smokers dropped from 28 percent at the time the ban was implemented to Just 20 peroent of employees two years later. 200

Draf~ - Do not cite or quote V Veto Some companies, especially those wanting quick results that can he controlled, develop smoking polioles at thl top executive or managemen~ level and announce them to the employees° But a more frequently seen pattern today involves employees in the p~ocess of toz~ulating and implementing a policy from the outset but with varying degrees of direc~ion frDm~anagement. Companies have found ~hat involvement of employees,¸¸ including smokers, facilitates compliance with the resulting policy. In some companies, the involvement takes the form of responding to a charge. For example, an employee committee might he asked to examine the issues and problems related to smoking at the worksite and to present to management within three months recommendations for a policy and implementation plan to deal with them. In o~hers, management may decide that smoking is a serious health hazard to its employees and that smoking is to he eliminated in 12 months. This organization's charge to employees might be to review how other companies have successfully moved to a smoke free workplace ~nd to present reco~endatlons for steps the company should nder~ake during the next 12 months to make ~hat transition both smooth and as painless for smokers as possible. Regardless of the approach, if employees are to be involved, it is important that their contributions have meanin~ and he llste~ed to objectively by management, r When tracing the hlstory of smoking policies in organizations, it is not unusual to find that the initial push to limit or eliminate smoking came not from management, but frc~ the employees, themselves. At UAqJML~fe Insurance Company, employeesI complaints, coupled with a Maine law requiring employers to reduce smoking, resulted in the company-wide ban. Pacific Northwest Bell emphasizes that no company officer or executive advocated its move to implement a smoking policy. Rather, the impetus came from employees. A grass roots group conducted • survey of workers and eventually recommended that pNW Bell ban smoking. The E~ployee Advisory . Council at Cardinal Industries' ~anford, Florida, plant initiated the idea of a tough no smoking stand. (See "Case Studies--Cardinal Industries and Pacific Northwest~Bell") At Holiday Corporation, Memphis, Tennessee, a task force of employees developed a Clean Air Policy covering its headquarters offices. The task force was originally set up as a Wellness Ccmmlttee a full year before wmrkhegan on the zmoking policy. The employee group researched various aspects of the smoking problem by gathering data, talking to other companies that had already done 201

Draft - DO not olte or ~ota it, and working with the local cancer society and lung association. A su~sey was conducted of all employees to identify their habits and attitudes related to smoking. The bask force, itself, created the phased-in process that resulted in Holiday Corporation headquarters and several of its subsidiary groups going smoke free on January i, 1987. But the a~ of involving amployUm el not always as easy as it might seem, according to Charles Niellonof Texas Instruments. Zt is important to involve employees in theprocess of developing a policy as early as possible, and a survey of their habits and attitudes provides invaluable data to management, says Nielson. Becauls TZ has so many Iocatlons,~ however, timing of an employee survey was sometimes very difficultr in some locations, the policy had already been set hy corporste~eadquarters before the attitude survey could be conducted. As a result, some employees felt they were being manipulated. "Data is vital to planning, hut timing is also Important so that the company maintains its credibility,. cautions Nielson. U~io~ I~volveme~t ~h any unionlzed organization, consideration must:be given at ~e outset to h~ and when unions will be involved. Popular thinking Just a few years ego was that unlons would block most company- sponsored wellness efforts, partlcalarlythosethat interfered with individual lifestyle choices, su0h as smoking. But through the work of several ploneerlngunlonl ~uch as the Amalg~ted Clothing and Tax, ill Workers Union, the United Steelworkers Union, and the Unltsd Auto Worker$~ as well as the effor~l of national groupl including the Workplace Health Fund, more and more labor groups are willing to cooperate with management in reducing ~moking if they are approached properly--and early in the process. Unions also recognize ~hat their membership reflects closely the national averages, therefore the vast majority of their members do not smoke. AS a result, many unions are reoslvlng increased pressure from their membership to help control smoking in the workplace. After having been involved in all aspects of policy development, the Communications Workers of America sent a memo to its 8,000 members at Pacific Northwelt Bull acknowledging that the company was implementing a smoki n~ban, b~t stating that CWA would not oppose it because of the possibility that nonsmoking members would sue the union--and probably win. (See "Case Studiel: Pacific Northwest Bell.") In late 198S, the Workplace Health Fund, in cooperation with the Office of Disease Prevention and ~ealth Promotion, (US DHHS), held a conference of union people to discuss the ~rits and value of health promotion. One of the outcomes of the meeting was a set of criteria for union involvement in worksite wellness efforts. 2O2

Draf~ - Do not cite or quote Among ~he ~eoo~endations were ~wo that placed heavy emphasis on the need to have a good working relationship between ~he union and management before attempting to implement any kind of wellness program. "Worksttes in which labo~ and management are not cooperating ~o bring health and safety hazards under control should not be sites for bealth promotion a¢~ivitie|." In addi~ion, "where the worksite is not undercontrol~or the employer ls uncoopRrative, and wherlthe union has ee~ablished the need for health promotion, ~he proqrams should he oonductedoutside the worksl~e."" But for any unionized company considering a smoking policy, the first step must be to look carefully at its union contracts, particularly for any ¥ordlngthat~mlghtguarantea members ~he right to smoke. If such agreements exist, ~he likelihood of the union supporting a no smoking policy is sllm. Sue Pisha, area director of the northwest region of the Communication Workers of Amerioei believes that with motivational infor~atlon and education, therl is the potential for unions to eventually become a proectlve force for nonsmoklng policies. "Policies seem to eliminate in-flghting," she says. "without a ~olicy, the issua, is messy and polarizing."~ A companion element of virtually every successful workplace no smoking policy is an educational program designed to inform employees about the new r~les and ~o provide opportunities for smokers to kick the habit. While behavior modification programs are the most commonly presented, some companies have offered innovative approaches such as acupuncture, hypnosis, self-help materials, hot lines, incentives for nonsmoklng employees ~o encourage and assist their co-workers to quit, and multi-day intensive programs for hard-core smokers. Now that the nicotine in tobacco is widely recognized as an addictive substance, in much the same way that alcohol and drugs are considered addictive, other education techniques also have come into use. They include aversion techniques such as satiation and rapid smoking, relaxation training, uoping skills training, stimulus control, and nicotine fading.9 In addition to on-slte opportunities, businesses have gotten good results by encouraging participation in commun~y-sponsored stop smoking classes merely by providing lists of sessions available through reputable groups such am cancer and lung associations, hospitals, ~'s, and for-proflt organizations. Because quitting can be very difficult and often is greatly enhanced by peer and family support, many companies make cessation 2O3

Draft - Do n=t cite or quote opportunities available to spouses and other lm~ediate family members, as well. (See "~noentives.") UNU~ Life Insurance Company offers classes for e usually over- looked group, nonsmokers. The classes are designed to help those who do not smoke understand the problems faced by smokers tz~zing to quit and to urge them to encourage fellow workers to quit or to refrain from smoking. At~inierBank, Seattlel Washin~on, stress managuent olasses were offered to help swokers adJult to the polioy as well as to assist those who were trying to quit. While no national data are available on work~ite quit rates, strong worksite programs claim anywhere from 20 to 50 percent quit rates after one year. However, most published s~udies repol-~ six-month • abstinenoe rates of 30 pe~oent or lese.~ As a result of an intensive smoking cessation campaign, Johnmon ~ Johnson, New B~nswicks New Jersey, reports a two-year SUCCess rate Of 23 percent of all smokers in the company, not~ypt 23 percent of those who went through a program or completed it.-" Hany companies go a step beyond offering cessation classes by providing incentives for smokers to quit. Some also have devised rewards for non-smokers. The most widely used incentives for smokers are monetary, often tied to completing a cessation program and/or stopping smoking. Many uompanies offer cessation classes free to employees and their families, often during company time, or reimburse them for the cost of taking a communlty-bas~d class. Others, llke ~he Utah State Department of Health, reward smokers who actuallyquit. The "Mealthy~ah" programs pays $25 to smokers who quit at the end of three smoke-free Jonths, aDother ~25 after six months, and $50 at the end of a year of not smoking. Honmonetary incentives, too, can be appealing. Employees who participated in a 24-hour "Cold Turkey" stop smoking day at MSI insuranoe, Arden Hills, Minnesota, beoame eligible for a drawing f~r a frozen turkey. Those who quit for six monthm were eligible for a drawing for a free YMCAmsmbershlp, and anyone who stayed o~I cigarettes for a full year was ellgible for a weekend vacation. Some companies also have gotten creative in finding ways to reward employees who are nonsmokers or who quit before a policy goes into effect. Employees who take a health risk appralsal at Westlake Co~unlty Hospital, Melrose Park, Illlnols, reoelve a $50 "bounty" for par~Iclpating plus several "good health bonuses" including $25 for not smoking. Weekly paychecks at Speedcall Corporation, HayWard, California, include an extra $7.00 for those who do not • smoke at work. Backsliders who light up one week and lose their reward are encouraged to get back quickly to Dot smok~pg; so the next week without smoking earns the $7.00 bonus again." 204

Draft - Do not cite or g~ote Businesses also offer nonsmokers discounts on llfe and health insurance, a very visible and tangible incentive he stop smoking and improve health. Smokim~ Poli~ znf~asment Without a doubt, one of ~he moat difficult questions asked by companies considering a smoking policy is -How can a no smoking policy be enforced?" The response from most businesses that have moved to a ban is that the company must first demonstrate to employees that it is serious about eliminating smoking in all or parts of the building. Second, it must handle violations in the same way that infractions of all other personnel policies are dealt with. cardinal Industries had a highly visible and dramatic way of demonstrating its oonitment. Its president, Austin Gurlingar, a cigar smoker, stated to all employees that ha would refrain from smoking at the workplace. ~aklng certain that each employee receives a copy of the policy in advance of its implementation and posting signs clearly delineating where workers may and may not smoke are small steps that can help show a company's commitment to smoking controls and increase compliance, as well. Some companies are enforclngthelr no smoking policies by referring employees who are unable to quit be¢ause they are addicted to nicotine to an Employee Assistance program. These companies may apply the same enforcement guidelines to addicted smokers as they do to users of alcohol or dr~gs, requiring that they overcome the habit in order to remain with the company. Most companies say that no employees have quit their Jobs because of the new rules. However, most also point out that a few have "tested" the policy, with some pushing it all the way to probation. According to Dick Backer, employee services representative for American Family Insurance Group, "Some employees tested the waters, sneaking cigarettes in the rest rooms. SupervisoEslmt it be known that smoking would be treated lime ~ny other vlolatlon of policy, for example, inappropriate dress," Holiday c6rporatlon follows its usual procedurm for violation of any company rule--first a verbal warning, then a written warnlng, followed by a "final" warning, and if necessary, termination. But all agree, termination is not the objective. Everythlng possible should be done to encourage employees to comply, and mos~ feel that peer pressure is the beat policing mechanism. However, when an employee continues to break the r~les, h8 or she must be 205

Draft - Do not cite or quote disciplined appropriately, or the entire policy will crumble. (see • Case Studies.") The following case studies illustrate how four widely different companies approached the development and implementation of a policy to reduce or eliminate smoking within their organizations. CARDINAL ZNDUSTRIZBt ZN~* Overall Policy: A total ban on smoking on any company property exists; all new employees must attest to belng nonsmokers. Beginning January i, 1987, the 8,650 employees Of cardinal Industries were assured of a totally smoke free work environment. One year prior to ~he ban, Cardinal, the nation's largest manufacturer of modular homes, had taken an even more dramatic step by ins~itutlng a multi-faceted policy that included hiring only nonsmokers as new employees. Bsnoflts ~tloi~ated Although insurance carriers are saying it will take 12 to 18 months tO e@e any deorease in insurance ratie~ Cardinal's management expects to significantly lower operating costs, increase productivity, reduce absenteeism, and eventually pay lower insurance premiums as a result of the new policy. Even more importantly, It expects to improve ~he health of its key asset*- it~ human resources. But employees at Cardinal~e Sanford, Florida, locatlon--one of four regional sites throughout the country--are convinced they would have gone smoke free even without the corporate edict. Why? Because employees wanted it, and because management recognized the negative impact of smoking on employees' health and prc~uotivity. The passage of Florida's Clean ~door Air A~ in October, 1985, focused attention on the plant's effo~s end established it as one • of the most progressive workslt8 no-smoking policies in the state, stimulating a letter of commendation from the go~ernor. Because of the nature of materials used at Cardlnal, the second largest residential builder in the country, the company had a long- standing policy prohibiting smoking in its five manufacturing plants. But at the Sanford location, the real push for a ~ough 2O6

Draft - Do not Cite or quote policy~hat extended beyond the manufacturing faeillty came~hrough its Employee Advisory Counell ~n late 1985. Made up of employees elected by the workers within eath department, the Council meets reqularly with top management. Based on employees' suggestions, a three-phase policy wee developed and implemented Jan. l, 1986, that gradmelly eliBinated smOking in meeting rooms, the cafeteria, and other common areas over the nex~ 12 months. AS par~ of the policy, which was designed to ~ake cardinal smoke-free by the end of the year, the company began hiring only non-smoklng personnel, current employeesj smoking privileges (in designated areas) were g~andfathered for the remainder of the year. But before the policy was implemented, it required approval by top management, including the company,s 33 year-old founder and president who was a cigar emoker--a situation that has stopped many other companies with good intentions. "We had been looking for ways to reduce our health cape costs and at the same time improve efficiency and productivity," said a company spokesperson, ,and the evidence about the health consequences of smoking were too powerful to ignore. When you add the fact that cardinal pays for i00~ of ~mployeest health insurance, thedeclsion seemed inevitable.,, In many ways, the fact that the chief executive at sanford was a smoker aided in convincing employees that ~e plant was serious. The announcement that only nonsmokers would beh~red and that there would be no exceptions to the rule--even the president--helped overcome one of company's blgqest obstacles to successful implementation...convinclng employees that the company is serious about the ban. A second advantage Cardinal has in tlrmsof enforcement is a highly desirable work environment. It pays top benefits and offers excellent working conditions. An employee must balance sacrlfiolng hls/her smoking habit for eight hours each day wi~h sa=rlficlng a job at Cardinal. So far, cardinal has won every time. Not only has no one quit, but the ban has not even been tested. "They know we are serious, end if they test us, they must be willing'to llve with the consequences." Management also believes %hat the iong- standing positive environment among employees and management has contributed to the easy transltien. During the 12 months between the announcement and the implementation, various company-pald educational programs and cessation classes were offered. Zn addition to requler stDp 207

Draf~ - Do aot =its or ~ote smoking seminars provided after business hours, employees and their families also were offered a hypnosis program, and for those who felt they were addicted to smoking, an intensive two-day, off-site treatment program was provided. FDA approved pharmaceuticals also were offered as quitting aides. During the period preceding the ban, smoking areas within the locatlons were gradually restricted until, on January I, 1987, ~he entlre company became smoke free. Although no survey has been taken to determine how many employees have quit smoking, a survey taken before the ban was implemented revealed that more white collar employees ~han blue collar workers were smokers. At the Florida location, for example, some 40 to 45 percent of employees could be classified as "blue collar." But partly because the manufacturing plants were always nonsmoking, there has been no partiCUlar problem in implementation. In a Position Paper discussing its policy, Cardinal Industries states, "The program only concerns itself with smoking in the workplace add not what employees do on their own personal time. Cardinal Industries never has, and never will try to regulate the activities of its employees on their OWn personal time." Thus, while Cardinal's application fqrm asks prospective employees whether or not they smoke, and while its policy prohibits the hiring of smokers, no attempt is m~de to test ,employees or to check on their off-work habits. TEXAS INSTRUMENTS Overall Policy: Smoking is prohlbitsd in all owned and leased faoillties except in specific locations in each facility that are designated as smoking areas and, to ~he extent possible, are separately ventilated. In late 1985 and early 1986, several of ~he 37 ma~or sites of Texas • Instruments began implementing their own smoking policies as a result of employee complaints and local" Clean Air legislation. Rather than be faced with 37 dlffar~t policies to implement, TI made a declslon to Implement a siogle corporate-wide policy. Emolovee Habits and ~ttltudes Before embarking on policy developmenh, TX surveyed its employees to learn how many smoked and how they viewed worksi~e smoking restrictions. About half of the more than 50,000 workers surveyed 208

Draft - DO not cite or quote took the time to respond, revealing that ?7 percent of TI employees were nonsmokers or exsmokers, and that Just 23 percent were ¢~rrent smokers. Of those who smoked, o~er 40 percent said they wanted to quit. moE~o~ata o~4e=tlvee and Poli~ Before designing the Clean Air Policy, top corporate management agreed on three objectives tha~ would form its underlying philosophy. There were to: o provide a healthf~l and safe working envlronlent; o ensure high quality in all ~TI productst and o initiate the companyms clean air approach rather than be forced to react to Isgislatlon (including ~he possibility of legislation from many dlffbrent states and municipalities). From these objectives grew Tits Clean Air Policy. I" .'It is the ~al of Texas Instruments to provide for its employees a healthful and safe working environment. In accord with this goal, Texas Instruments w~ll prohlblt smoking in all TI owned and leased facilltiss~;except for specific locatlonm in each facility which are designated as smoking areas." TO underscore the importance of :the new policy, the eight-month, phased-ln implementation process took a top-down track, with the President and CEC Jerry Junkins working directly with a key operating manager and the personnel director from each location th~ughout the organization. During the session, Mr. Junkins emphasized the organlzation*s complete commitment to the new personnel policy and each individual manager's responsibility for its successful implementation. These teams then headed up similar training programs in their own lOCations. Training sessions were conducted for selected managers hslng a centrally prepared manual to ensure consistency among the3Y locations. Specially developed brief video tapes offered all ,~mplOyees an innrodu=tion to the policy (3 minutes), briefed managers ~and supervlsors on issues related to smoking (i0 minutes), and assisted managers and supervisors in learning techniques for resolving smoking-related problems at the worksite (16 minutes). Making every effort to assist smQklng employees to prepare for the new policy, TI provided company-pald smoking cessation programs on company time durlnq the initial phase-ln of the clean air program. Classes were scheduled to accommodate workers on all three shifts, 209

Draf~ - I;o flmt cite or and "maintenance sessions,, were affered to provide additional suppo~. More than 4,700 employees signed up for cessation classes, representing 40 percent of the company's smokers--almost exactly the percent that said they wanted to ~it in the employee survey. Of the group, 3,235 completed all the required classes (including maintenance classes), with 1887, or 58 percent, reporting they had ~it by the end of the program. As a further aid, a Tip Sheet, "How to Make Life Easier Until the Next Cigarette Break" provided "some practical suggestions to help you when you need to change your regular smoking routine." A "Wrap Sheet: Daily Cigarette Count," designed to be wrapped around a pack of clgarettms, offered an easy place to keep track of how much was smoked, when and why, in the hope that the information would assist the smoker in altering hls/her habits. However, all communication was not downward! Employees were given opportunities to ask questions and voice concerns during educational programs. Special attention was p~id to employees ¢oncerns and complaints in In-house communication vehicles, as well. Facilities Modification Because TI chose to designate a limited number of areas in each building as smoking areas rather than to completely ban tobacco, it faced the problems of recirculating contaminated air. Thus, where necessary and possible, facilities were modified to provide separate ventilation. In addition, all =igarette machines were removed from TI facilities and a decision was made that no new ones would be installed. TI made it clear from the beginning that a new personnel policy had been established that would be monitored and enforced in the same way as all other policies, such as attendance. Thus, anyone found smoking in non-designated areas would be given an oral warning. If there w~re no further problemst no further action would be taken. However, with subsequent smoking incidents, the employee would be given written guidance, followed by probation for additional infractions, with termination as a final step. But TI stressed to all supervisors that they should make every effort to educate smokers about the importance of the policy, rather than to be heavy-handed. After nine months, "two or three oases" have gone to probation, but no one has been terminated because of smoking. Considering that 50,000 to 60,000 employees are covered by the 210 [ ~uote

D~aft - DO not cite o~ ~ote policy, this is an excellent record, says Charles Nielson, Vice President and Manager of U.S. Employee Relations. Advice: Keep Poliav,s Purpose in ~ersDective Hielson cautions other co=parties considering establishing a no smoking policy that one of the most difficult problems they face will be keeping the desire to ~iAmAnate the heal~h hazards o~ smoking at the workslts in proper perspe¢~ive. TZ made a corporate decision to eliminate smoking at ~he worksite except in designated areas. The decision wasa ~usiness decision, not a moral or a value Judgement. TI, which has 50 percent of its business in semiconductors, ~s facing intense competition, according to Nielson. Therefore, it must have productive employees. And that means it must have good relationships with all its employees. But s~oking is anemotlonal issue for many people, both smokers and nonsmokers. "I'm not sure ~hose of us in the personnel field have yet learne~ how to deal with ~his kind cf highly charged issue and still maintain sue productivity,- states Hiel$on. "It takes a lot of hard work to achieve the desired ~tmosphere of teamwork, rather than an advarsarlal relationship." RAINIER B~/RAXNZER BANCORI~RATXON overall Policy: Following a one year phase in period, smoking is banned in qll 200 domestic facilitles and in car pool vehicles. In September, 1985, the 5,800 employees of Rainier Banoorporation~s U.S. facilities received a communication from their Preslde~t, John D. Mangels stating that "we are committed to insuring a healthful and comfortable environment for all employees." As part of that commitment, he announced, the corporation would become smoke-free on october i, 1986. As part of a transition plan, beginning october 15, 1985, smoking would be restrlcted to designated areas, and +.he company would sponsor and,~ay foe smoking cessation classes to assist employees who choose to quit. Rainier Bancorporation is head~artered in Seattle, Washington, with 200 offices in Washlngton, ~plus Alaska, Oregon, California, Hawaii, Arizona, New York, and the Far East° Heaitb Threats. Emoiovee COROiaim~s~ LmOai conaerns PEOmDt POii=V According to Peter Broffman, personnel officer for Rainier Hank, the major subsidiary of Rainier Bancorporation, the policy resulted from three converging issues, the major one being a concern for 211

Draft - DO not clta or q~cte enployee heal~h and wellness. Additional factors ware an increasing number of employee complaints about smoke in the workplace, and the changing legal and rsgulatory cllmate. In July, 1985, the sta~s of Washington had adopted a Clean Indoo~ Air Ac~ that prohibited smoking in public places, including public areas of banks, and there was reason to bali@~s that unless employees actwd on their own initiative, there might elmo be legislation regarding private workspace° T~st, Ooupied with an Incrsaslng number of court cases upholding the right of e~ploysas to have a smoke free workplace, added impetus to,he development of a policy. communloatlons vital to Ymmlementatlon Once the decision was made to go forward wi~h a phased-ln ban, communications wi~h employees bscale a key link to successful i~plementation. Emphasis was placed on the fact that Rainier was prohibiting smoking at the workplace, not smokers. Pbase-ln Period During the transition period, managers were given discretion to determine the most appropriate way to make the transition. The company policy s~ated that "The needs and 'comfort lavel' of both smokers and non-smokers should be considered during this period." Guidelines for Phase I stated, in part: o All cOUCh areas, including lobbies, elevators, conference rooms, hallways, lih~aries, ~eBt r~u and computer rooms will he smoke free. o In open-office work environments, managers should use discretion in deciding whether those areas should be smoke free. Individual employees may, of course, designate their assigned immediate work space as a nc-smoklmg area. o Employees with enclosed offices may designate their area as a smoking or no-smoking area. However, the rights of non- smokers who must come into an enclosed office to conduct business should be respected. o Lunchroom and lounge areas will be divided into areas for smokers and non-smokers. Managers are given discretion to divide the rooms as appropriate for their locale. According to Broffman, there were relatively few difficulties in the initial phase of implementation. The few problem that did exist were due largely to differer~es in ~he ways various managers chose to implement and police their smoking restrictions. Occasionally disputes arose over what areas should he smoking and 212

Raft - ~ not cite or quote nonsmoking, especially in the smaller branch offices where there were few options for allocating space. For the most payS, Broffman says, the problems were minor and easily resolved when the total prohibition was enforced. However, in retrospect Rroffman believes that a shorter trsnsltlon period right have been more desirable.¸ "A ~hrse to six month phase-in period probably would¸ harm been adequate," he "Many smokers go~hrough s edJue~ent period. A few Indlcatsdsays" that putting off the inevltable ban for too long really Isnlt doing them a favor because it prolongs the period of anxiety and allows them to procrastinate in making the adjustment. Also, a shorter transition emphasizes the resolve of the company to become smoke free." In addition, any employees who want to defeat the policy will use the entire phase-ln period to rally support. A shorter transition period would shorten the debate and lessen the p~$sibility that the detractors will succeed. Bnforoln= the Ban Phase II, the total smoking ban, was introduced in a low-key, ~a~ter-of-fact manner: a slmple "remlnder" that smoking would be prohibited in all work areas. With the exception of minor, final protests by a few "die-hards", employees accepted the new policy. Rainier has received no formal complaints, has had no problem with recruiting, and no one has resigned. The only complaint Broffman is aware of is that a few employees who still smoke do so immediately outside company building during breaks, and some employees are concerned about the impression this gives to customers entering the bank. Because of its stance that Rainier is eliminating smoke, no~ smokers, the organization makes no attempt to discriminate against hiring smokers. Slooest Obstacle to Pollovt Fea~ Broffman acknowledges that when the policy was first proposed, there was concern on the par~ of a few senior manager of -What might happen." Although the majority and the leadership of senior management supported the policy, a few were initially concerned that there could be mass deletions, that dlsagrelments about smoking would cause major disruptions in work units, and that it could turn into an "employee rights" issue. However, these things did not happen at Rainier. "We had more complaints from nonsmokers before the policy was implemented than we got from smokers afteT it was enforced." 8roffman says. His advice to other companies considering s smoking ban? "Do it| You can make it work~" 213

Draft - Do not citl or qluote P~CIFIC NORT£W~ST SELL Overall Policy: Because of heal~h concerns relatld to smoke, PNB does not allow smoking in any company faoillty. On october 15, 1985, Pacific Nor~hwestBell became the first large company (15,000 employees in 750 buildings) to go completely smoke free. Its policy is simple: "To protect the heal~h of PNB employees, there will be no smoking in any company facility." o~tlons for ~mokina Restrletlo.e Prior to the establishment of ~he policy, PNB had allowed each work group to decide, itself, whether or not it would be a smoking area. Problems arose, however, when ed~acent work groups had differing approaches. Smoke would drift around barriers, waft across no- smoking desks, and generally infiltrate all areas of the building. Smokers assigned to no smoking areas would merily walk into work groups that permitted smoke, making ~he atmosphere even worse for nonsmokers in the area. Difficulties occurred even within¸ individual units that voted to ellmlnatelsmoking. ~f 80 percent voted to be a clean air area and 40 percent voted for smoking, the question arose as to whether the wishes of four-out-of-ten employees COUld really be ignored. While this kind of democratic approach had initially sounded llke an easy way to avoid forcing a company-wldQ policy, it was seen as unfair and inequitable by most workers. NO one was really satisfied and all the underlying problems still existed. Eventually both managers and employees began exerting pressure on PNB to develop a company-wide pol~cy. In ~anuary, 1983, a Smoking Issues Steering commlttee was established consisting of smokers, nonsmdkers, and. a group often forgotten, former smokers. Employees representing their unions and from the legal, health services, safety, and many operating departments were part of the task force. One of its first undertakings, an employee survey, brought an astonlmhlng74 percent response rate, attesting to the importance of the issue among workers. In addition to comments fro~ those who were randomly surveyed, 151 people who were not pa~ of the survey group made the effort to get copies from their ~rlends so they, too, could have their view heard. They included ~35 nonsmokers and 16 smokers. 214

Draft - Do not ¢i~I or quota Results of the 1983 qllestlonnai~e showed ~hat 28 percent of PNB employees smoked, but that the majority of employees were bothered at least occasionally by smoke at the workplace, and almost 80 percent said the company should he concerned about smoking at the worksite. Two-and-one-half years after its Inceptlon--followlng a great dlal of research and discussion by the task force, as well as involvement in the issue hy numerous QUality Of work Life teems and various ad hoc groups--~e employee committee recommended to the officers that smoking be elimlnated at PNB. O V At PHB, unions were instrumental in all phases cf policy development. Not only were they included In the employee committee making recommendations about a future policy, but leadtrs of both unions were part of the June 1985 presentation to the company prmsldent of the committee's recommondatlon. "They were ~ot there as advocates for a no smoking worksito," cautions Len Bell, ~irector of huma~ resources planning and employee involvement. "They were pre~ent, rather, to state that they had been involved in the process and what Shelf positions would bo on a strong policy. While they did not endorse the complete elimination of smoking in all buildings, they stated that their unions would not formally fight its implomentatlon, either." Bell adds that t~e union members on the co~ittee were "eXtremely helpful" in all aspects of pollcTdevelopment, and that while they never fought against the policy, they negotiated successfully for several compromises that proved to be fair and beneficlal for all employees. Initially, the company wanted to reimbursl employees for smoking cessation classes after sucnessful completlon. The union position was that PNB's goal was to assist and encourage emp'~oyees to live with the policy and comply with It--not necessary to get them to stop smoking. Therefore, they pressed, the company should relmburse totally for cessation classese whether or not the employee completed the series. On the issue of smoking in company vehicles, union representatives stressed the difficulty cf enforcement and potential probloms If algaretto butts were found in a company car or truck ash tray. On both issues, PNB wont with the unions' requests. All employees got full reimbursement for taking a cessation clnss and smoking in company vehicles is a matter of "common courtesy." The unions also urged that any policy be oonslstent throughout all company iccatlons and for all empl~yees. 215

D~aft - Do not cite or quots The day the policy was announced, which was 90 days before ~he policy was to go in effect, PNB provided two telephone hot lines to answer questions about the policy and provide Info~ation on free cessation programs for employees and their dependents.TM A wide range of quit opportunities were luade available, many on company time, with PNB paying all fees following completion. But PNB also garnered kudos from many employees by allowlng~hemto take classes outside the company and still get reimbursement. The ability to choose ~eir own quit method seemed to add to thelr coni~ent to succeed and helped encourage a friendlier attitude toward the pcllcy. Within the first two years, 1,738 people had gone through cessatlon programs--1,3$3 of them employees, 360 spouses, and 25 dependents- -receiving full reimbursement from ~he company for a cost of about $250,000. Is this investment woET~h it to PNB? "Yes," says Sell. "Zt is money well spent. This equals the cost of Just two or three ~ancers cases. And we would much rather pay for 1,738 to try to quit smoking than pay the results of their continued habit." PNB reports that there have been "no real problems" with enforcement. On the first day, there were reports that one or two people were smoking behind closed doors in several locations. But "word got around- and by the second day they were abiding hy the rules. Although several people threatened to contact lawyers and a few employees tried to organize a Smokers Rights day, nothing significant came from any of the attempts to block implementation. The bottom line: After two years, no one has quit because of the no smoking policy, there have been no grievances, and smokers at PNB have dropped from 28 percent to 20 percent in the two years since implementation. All in all, the company views its no smoking policy as an unqualified success. SD~ARy i. The movement of businesses tO develop ar~ implement smoking control policies appears to be strong, and gaining momentum. 2.. Employees and unions should be involved in the developement and implementation of workplace smoking policies. 3. Enforcement of smoke-free workplace policies has not proved to be a real problem for business. 216

Cr~aft - Do not cite or ~ruo~e REFERENCES i. office of Disease Prevention and Health Promotion, Survey of WarkRite Health ~omotlon Activitlas, U.S. Department of Health and Human Sa~vlces, 1987~ Washington, DO. 2. W , .Around ~he natlon: Many Eirms limit smoking," Washington, DO, February 23, 1988. 3. Hav/Hu~ains Benefits Re~o~, November 17, i989t Philadelphia. 4. American Lung Association, ,summary of results of the 1989 survey on public attitudes toward smoking," NOV. 1989, New York. 5. Tripp, J, "Tobacco smoke disappearing in workplace: Employers impose ban" ~, March 17, 1986. 6. Read, K. "Smoking bans: Corporate cold turkey," Co~orate Fitness! The Journal for Employee Wealth and Wellnes~ Pr~rams, Aug/Sept 1967. 7. Kaiser, J andBehrens, R, Health PTomotlon and the Labor ~, Washington Business Group on Health, July 1986. Washington, DO. s. Smoking Policy Instltutl, "Smoking policies and the unions." 1986. Seattle, WA. 9. U.So Department of Health and Human Services, Conseouences of Smokin=: Nicotine Addlction--A Re~o1~ of ~he , Office on Smoking and Heal~h, Rockville, Md i0. U.S. Department of Health and Human Services, Of Zice on Smoking and Health. The Health conseauences of Smokina-- Cancer: A Report of the Su~aeon Genes810 U. B. Government Printing Office, Washington, Oc° Secondary Source: office of Disease Prevention and Health Promotion, MakarIs Guide to Smokina &t the Work.ibm. 19SB. Ii. office of Disease Prevention and Health Promotion. DecisiDn Maker's Guide to Smokino at the Work,ire° U.S. Department of Health and Human services, 1985.. 12. Yenney, SL, Usina Incentlvs~ to Promote E~D~ovee Health, Washington Business Group on Heal~h, 1985. Washington, DO. 13. Behrens, R. Reducing Smoklno at the WorkD1aca, Washington Business Group on Heal~h, 1985. Washington, DO. 21~

D~af~ - Do no~ cite or quo~m 14. Bureau of National Affairs, WherQ There's Smoke: Problems and Policies Concernino Smoklna in the W~k~lacA~ Washington, DC. 218

~.'a~ - DO not cite or ~ote ~1~2R lit &PPB~ZX IS THERE ECONOMZC JUSTZFIC~TZON MR NO SMOKZ~G POLICZBS AT T~ WORKBZTZ? By Ruth BehBens* Washington Business Gzoup on Xmalth The health hazards of smoking--in¢ludlng smoking at the vorkplace- -have been well documented. Smoking greatly increases an individual's chances of contracting serlous illnesses, such as cancer, chronic bronchitis, emphysema, and coronary hear~disease, and of dying prematurely as a result cf these diseases. There is little doubt that smoking also has a significant economic impact. I~is estinated that businesses pay over SIOB billion per year in health care costs. A significant portion of this bill is ~he result of smoking, and is paid out through insurance premiu~s for employees, dependents, and retlreeswho smoke or breathe second- hand smoke, as well as for nonemployeea who smoke or breathe Others' smoke through programs supported by state and local taxes. ~n other words, smoking Is costing businesses a lot of money. HOW much does smoking cost U.S. businesseS? NO one knows exactly. But a growing llst of researchers are tackling the difficult Job of attempting to identify these coats. Costs o£ Bmoklna ~o the Nahlon At least three major studies have addressed the questlon of what smoking is costing the nation. ~n 1978, Luce and Bchweitzer estimated the economic coats of smoking in the United States to be $47.6 billion. They further broke this down to $811 per adult smoker, or $1.5~ per pack of cigarettes sold.I Zn 1985, the Office of Technology Assessment, U.S. Congress COTA), estimated that smoking costa the nation about $65 hi11ion per year in lost productivity and health care coatsal0ne, eTA estimates - that smoklng-~aused illness results in $43 billion in lost productivity annually (or $1.45 for each peck of cigarettes sold), expenses borne largely by empl~qers. Businesses also pay a significant portion of another $22 billion In smoking-related health care costs, since nearly two-thirds of the coats are incurred by those under 65. Acoordlng to the OTA, combined loa~ productivity and health costs related to smoking equal $2.17 per pack of ciqarettes sold z 219

Ora£~ - Do no= =t~e cr :~uota Most recently in 1986, a group of researchers, which included the former director of ~he government's National Center for Hea1~h Statistics, concluded tha~ smoking ~os~s the United States a~ least $59.7 billion each year in direct medical costs and salary losses alone. These cost estimates were calculated by comparing the health costs and income losses from smokers in excess of the same amounts incurred by nonsmoker. The study concluded that. smokers are sicker and re,Ire Bore medical care than nonsmokers." The components of the $53.7 price tag were broken out as follows. o Direct medical costs such as doomer hills, drugs, and hospital and nursing home expenses were $23.3 billion more for smokers than the average of nonsmokers. o A total of nearly $9.3 billion was lost in salsbies due to smokers being sick with smoking-related d~aeness in¢ludlng lung cancer, hea~ attacks, stroke, emphysema, and other respiratory illnesses. o In 1984, lifetime earnlnq losses from smoking related deaths were approximately $21.1 billion. The authors characterize their findings as "conservative" since they "did not take into account the adverse effects of passive smoking, risks of abortions, stillhIz~hs, and neonatal deaths, or deaths under age 20 that might he associated with smoking." In their paper published in , Rice et al translated all three of those studies to 1984 dollars. The result is three analyses oZ the economio impact of smoking on the nation that demonstrate enough similarity to underscore that smoking does, indeed cost our country a staggering amount:' o Lute & Schweitzer show a cost to the nation of $82.8 billion per year in 1984 dollars; o OTA, $82.2 billion in 1984 dollars; and o Rice etal, $53.7 billion in 1984 dollars.3 Differin~ Methodoloates Make Pinooint~n~ Workstte Costs Hard A number of researchers also have~a~smpt6d to assessthe specific costs of smoking to businesses. But many pr~lams arise when attemptin~ to identify one, or even a "best" methodology for arriving at these costs. A~ong the difficulties in conducing any study of the costs of smoking is the fact that smokers dt£fer f~om nonsmokers in several genetic, social, and economic characteristics that may contribute to disease. ¥or example, the prevalence of smoking varies by race (more blacks smoke that whites), education (fewer college graduates 220

D~aft - Do not cite oE quote smoke than persons with only some high school), income (males with lower income smoke more, while the opposite holds for women), and occupation (blue collar workers smoke more than professional or technical workers). If factors known to be related to health status and smoking habits are not controlled, the impaq~ of smoking on heal~h and the costs of smoking may be overstated/ When examining smoking in the worksi~e, specifically, other methodoloqio issues must be resolved. Marvin M. Kristein and William Weis both pUblished studies in the early 1980s identifying the cost to business of each smoking employee. Kristsin estimated the cost in 1980 dollars to be betwesn $336 and $601 ~er smoker annually,4 while weis placed the figure nearer $4500.s These findings are now outdatsd; in an article published in 1989, Kristein has stated that "...the ¸typical Boking employee in 1988 cost the typical employer at leash $I000 in excess oosts# compared to a similar nonsmoker.° However, a look s~ why their conclusions differed i0 fold dramatically illushrates two points: i) hhe difflculty of pinpointing the cos~ of smoking to businesses, and 2) the wide range of business costs that can be affected by enviro~L~enhal tobacco smoke. ~ucb of the ra~h~ stag~erinq disarepancy behwesn ~he two studies ~s attributable to their seleoh~on of diffmrent categorieI of coshs to inolude in the equetion~ the weight give~ eeoh oateg~ry~ a~ ~he s~l~ry sssi~d ~ tha avsra~e s~oker. ~c~ording h~ ~eis, business oosts i~ at ~eas~ ten srees e~ affeoted by smoking cr s~oking ~nhro~s~ including no smokin~ poli~ies~ ~ea~h iNs~ran~a~ incrs~eNtal ~hsenteeis~ iifs a~d ~is~bili~y ins~ran~e~ fire, liability an~ industrial ~iden~ ~n~an~e~ ~enti~atio~ and energy ~n~um~i~n f~r heating ~d si~ ~o~iti~niNg~ le~l ~iabii~ty~ ~sr~y da~a~a, depreciation a~d mai~teNa~e~ ~im~ lost to the smo~i~ ~it~sl~ empioyee morels, and o~rporate image°~ ~rist~n fe~h~ i~ heel~h ~nd life insura~e~ ~i~ losses~ wo~ker~~ oompensahion cna~s~ absent~eism~ pro~ti~i~y~ end ~co~pational hesl~h ¢~st~° ~n ~ ~99~ article~ Kristein looke~ ~ oniy sbor~cerm ~s~s a~ in~uded firs, • ooide~hs~ venhil~io~ ~le~ni~g~ prod~chivity, and o~pahional health risks°)~ • o hel~ illustrate the differs~ces heh~e~n Krishsin and ~eis~s to~ai smoking~rel~te~ ~sts~ one,can io~ at ~ow ea~ ~alouiatss the ccshs of absenb~eis~ ho sm~loyers ~e tn sm~ki~go ~si~ u~e~ go~ernme~ de~ tha~ ~h~ws • ~ok~r is ah~ent ~o~ ~eys ~er year more th~n a noNsmoker° ~sing $9~00 ~er em~ioyee aa the • ~erag~ annual wa~e a~d salary~ including frin~s and payroli ta~es~ the company ~ay~ appro~imahely $~0 p~ working day ~or every ~mp~oyee o~ the payrollo ~ss~ing a ~ percent return on ~ayroll ~ol~ars~ the ~ir~t cost to the employer is $~ ~r ahsenoe~ e~o~in~ the oost o~ ~e~orary ~eplace~sn~so ~o~ordin~ ho thi~ 22~

Draft - Do not cite or quota fonula, the total cost per smoker per year due to absenteeism is $310. A similar system is used by Wels in dete~ining costs in o~her categories.7 Kristein, on the other hand, used 1979 data showing smokers are absent 33 to 45 percent more ~han nonsmokers, or 2.0 days more par year, and assigns a daily salary of Just $40 per noker due to smoking (versus $150 for Weis). Thus Kriatain in¢ludes from $40 to $80 per smoker per year attributable to absentselsm in his total (versus $310 for Weis)." While Kristein's estimates are based on what he called "real numbers" drawn from Insurance companies, U.5. government statistics, and detailed academiastudies, he =autlons, "we lack meaningful 'case controlled, company comparisons of experience with smoking employees versus nonsmok~g employees .... In gtneral, the emphasis is on ~ the costs to business." E~onomi= I~Da~t of Smokers on ~he Workslte ~vldence also shows that, in addition to excess absences of two or more days per year, s~oksrs exe~ other types of economic impacts on businesses over their nonsmoking countexpal-~m. Studies have shown that: o smokers have twice as many Job related accidents as nonsnaokers. 10 o Smokers are 50 percent moreilikely to be hospitalized than those who do not smoke. o Employers have been held legally responsible for at least part of the disability cost for smoking employees who contracted smoking related illnesses, In addition to claims from nonsmcking ployees who were adversely affected by the smoke of others.~ o Companies with certain occupational hazards can expect greatly increased costs related to smoking. For example, an asbestos worker who smokes ~s ten tlmas more likely to die prematurely than his nonsmoking cowo~kerSo A smoking uranium miner has six times the rls~ of contracting lung cancer as a nonsmoker in the same johD In addition, many health consequences of smoking translate directly into increased health care costs, since elployers pay for a major portion of these costs for ~helr employees, dependents, and retirees. o HeaVy smokers (two or more packs a day) are iS to 25 times more likely to die of lung cancer than nonsmokers, and 222

o o o o o o D~aft - Do not cite or quota overall, smoke~s are ~ ~imss morm likely to die of l~ng cancer than nonsmokers.~ Eighty to 90 percent of such long term severe lung disease e as emphysema end chronic b~onchltis are related to smoking.~ It is estimated that 30 percent OZ all cancers are caused by smoking That means that 138,000 Americans died Of cancer in 1986 because of smoking.I] HeaVy smokers are thrls to four times more likely ~o die of cancer than nonsmokers and Overall, the risk to ~okers is two times greater than for those who don!t smoke." More than 550,000 Americans will die of coronary heart disease this year, and up to 30 perce~ of those deaths will be attributable to cigarette smoking~ HeaVy smokers have a 200 percent greater risk of dying from coronary heart dlseasethao nonsmokers, and Overall, the risk for all smokers Tegardlses of ?.he amount s~ked, is 70 percent greater than for those who dofl't smoke. 4 Evidence demonstrates that smoking during pregnancy has a significant adverse effect upon the well being of the fetus and the health of the newborn, including causing lower birth weight infants and increasing the risk of spontaneous abortion and neonatal deaths." Children of smoking parents have increased prevalence of respiratory symptoms and have an ino~ssd frequency of bronchitis and pneumonia early in life.~ Two studies relate smoking directly wi~ costly health-related events, stroke and automobile accidents. A study has concluded that smokers who qultcan decrease their risk of having a stroke by more than half when compared to those who careC°ntinUecosts1~t° moke, thus cuttlng dramatically their potentlalhealth A two-year study in Worcester Co~ty, Massachusetts, comparing the motor vehicle driving records ofsnokere wlthnonsmoksrs found that smokers had 50 percent more accidents than nonsmokers and 46 percent more traffic violations. The study Identlfled several reasons for the smokers: increased risk of being ~nvolved in costly accidents and violations, including o smokers' more frequent use of alcohol and dr~gs, o smokers' greeter rlsk-taklng behavior, and 223

D'L'aft - DO not ~ite or quota o smokers ' diminished attentlon to driving due to ~he 16 distractions associated with smoking. Indlvi~ual comnanles Document costs. Conseo~enaas of Smoklna While s~udies conducted by individual companies have varying degrees of validity, they do offer some further insights into the price businesses pay for their smoking employees. In a study of 40,000 employees at 27 locations of the control Data Corporation, CDC found that smokers cost the companysu~stantially more in health related costs than nonsmokers. The study, using health data collected from 1981 to 1994, found: o~ Smokers of one pack of cigarettes per day or more generate health claims 18 percent higher than nonsmokers. o Smokers of one cigarette to one pack per day accrue claims costs 10 percent higher than nonsmokers. o Heavy smokers have 2S percent more inpatient days than their counterparts who do not smoke. o Heavy smokers are 29 peroenE more likely to 1~ave health claims OYeZ $S0OOO thao thOSe who dO not ricks. one Los Angeles company estimates produotlon losses alone at $675 per smoker per year. Adding longeE te~oosts such as absenteeism, premature death, and illness wqpld raise the cost to at least $i,000 per year for each smoker~" Provident Indemnity Life Insurance Company charges its smoking employees the excess rate of their insurance coverage o~r that of nonsmokers, an amount in the vluinity of $800 per year. Smoking and the Bottom Line When viewed in the aggregate, these studlee may appear to make a compelllng case for the potential of smoking control programs and • policies to significantly cut long-terl business costs. However, a number of researchers, inc1~ding health promotion and smoking control advocates, point out that this oon¢luslon may not be Justified. In some cases, the s~udiee presented have significant methodological problems or their underlylng assumptions may be flawed. Equally important, the total costs of developing and implementing smoking control programs and polloies, coupled with the increased costs associated with longer life resulting from quitting smoking (pensions, retiree and dependent health care costs), may eliminate any financial gain for the company. 224

Draft - Do not olte or quote Following are several examples and studies illustrating how ~hese supposed cost savings may ~ot he what they Initlally appear to he. As is pointed out in this appendix, many of the costs associated with smoking can be attributed to characteristics of smokers (risk- taking style, alcohol and drug use, low socioeconomic status). However, it is unlikely ~hat these basic characteristics would change, even if ~he individual e~p1myee was induced to stop ssoking. A portion of the supposed economic penalty associated wi~h hiring smokers results from an increase in absenteeism seen in employees who smoke. Statistics indicate ~hat people who smoke are eight times more likely than nonsmokers to have alcoholism. Thus, helping current employees stop smoking might not have the expected effect on absenteeism, s~ce i~ some, alcoholism also is a root cause of the absenteeism.~ Some argue that smokers already are "paying their own way. through cigarette excise taxes. In examining the lifetime costs tha~ smokers impose on others through collectively financed health insurance, pensions, disability insurance, group life insurance, ~ires, motor-vehicle accldents,~and the criminal Justice system, "Willard G. Manning, et el, conclude that on balance, smokers probably pay for their own torts to society under the current level of excise tax on cigarettes.~ According to Kenneth E. Warner, Ph.D., a successful workplace smoking cessation program will reduce oertaln health care costs, llfe insurance co~ts, disability costs, and absentmeism, and it may increase productivity as well. "H~ever," he adds, "one thing that It is almost certain to do, by virtue of its success, is to extend the lives of a subset of employees well into retirement, implyin~ both pension and health care (and other) cost implications .... ";~ Warner concludes that when all costs are taken into account--such ass fo~ example, the increased costs of pensions, hesl~h caEe, and disability for retired workers who llve longer because they stopped smoklngf versus the decreased costs for workers who continued to smoke, die prematurely, and are replaced by a younger, less expensive employee--buslnesses might very well conclude that, from a purely economic point of view it may be cheaper to allow employees to continue smoking. Loulse Russell, Thomas 24 Schelling, and others have moms to similar conclusions based on cost savings alone. Individuals such as these, who debunk the idea that smoking cmntrol programs will result in most savings for businesses, do not, however, conclude that it is in heat the interest of businesses and society to advocate smoking or to shun smoking control pollcies. 225

Draft - Do not cite or quote There are obvious sho~ tea benefits of a smoke free workplace, over and above the health-related savings Warner lists above. They include reduced building and equipment cleaning and maintenance costs, reduced costs from fire damage and insurance, reduced energy consumption cost because of reduced ventilation needs, and reduced turnover. In addition, there erect he lass tangible benefits of a working snvironmen~ ~hat is pe~eived as being better by the overwhelming majority of e~ployees, as well as an Improved company image o But for many, the potential of he~ter heal~h for employees, and of eliminating or delaying ~he onset of degenerative or fatal diseases is the most compelling reason to implement a company-wlde smoking control policy. So the real bottom line for companies considering whether or not to implement a smoking control policy or a smoking ban may not be a simple dollars and cents formula. But rather, the ~ttom llne may be as pragmatic as the need to comply with local legislation, or the desire to improve productlvity, as pate1~nallstio as the desire tc have happy, loyal employees, or as alt~ulstic the desire "do the right thing" by providlng the most healthful environment for its employees. If costs savings follow, these companies may, themselves, have received a bonus. SUMMARY i. Smoking in the workplace increases business costs because the diseases of smoking increase absenteeism-and hospltalizatlon, and may increase insurance, disability and legal ~osts. However, these costs may be offset by the longer lifespan of employees who quit smoking as a result of workplace restrictions, increasing pension costs to employers. 2. The most compelling reason to restrict smoking in the workplace is the potential for better health Zor both nonsmoking and smoking employees, by eliminating or delaying the onset of degenerative or fatal diseases. 226

DEa£~ - DO not cite o: ¢~.oti BL and SO Schwei~zer, .,Smoking and AlCohol A~use: A Comparison of their Economic Consciences," 298, 569-571. 1978 2. office of Technology Assessment, U.S. Congress. "Smoking- Related Deaths and Financial Costs." (OTA S~aff Memorandum). 1985 flashing~on, DO. 3. Rice, DP, TA Hodgson, P Sinshelmer, W BroWner and AN Kopstein. "The Econo~i~ Costs of the Health Effects of Smoking, 1984" The Nilhank ~az~erly. Vol. 64 , No. 4, 1986. Cambridge University Press. 4. Kristein, MM, "flow Much Can Business EXpeot ~o Profit from Smoking cessation?" ~, 12, 358-381, 1983. 5. Helm, WL. "No Ifs, ands or Buts: Why Workplace smoking should he banned" ~, 339-44, Sept 9981. Krlstein,i RM, "Economic issues related ~o smoking in the workplace." N.Y. State J fledo 89:44-47 (1989). 7. Smoking Policy Institute, "The costs of Smoking in the Workplace," 1986, Seattle, WA. 8. Kristein,~ -Wanted: Smoking Policies for the Work Place, ~, flashing~on Bus~nsss Group on Health, NOV. 1984. Washington, DO. 9. B~rea~ of National Affairs, "Where Therets Smoke: Problems & Pclicles Concerning Smoking in the workplace," 1986, Washington, DO. 10. U.S. Department of Health, Education and Welfare, office on Smoking and Health. Smoklna and Mealth~ A ReDOE~ of the ~. U.S* Government Prl~ing Office, 1979, Washington, DO. 11. American Lung Association, "Smoking at the Workplace: The Changing Legal SituatloN. More Facts & Peatureg for M~I~. 1983. New York, New York. 12. U.S, Department of Health and Human Services, office on Smoking and Health. T~e ~eelth Consemuencem of Smoklna-- Cancer! A ReDOZ~ of the Suroeon General. U.S. Government Printing Office. 1982. Washlngton, DO. 13. U.S. Department of Health and Human servicesS office on Smoking and Health. Thm Health Cmnseauences of Smoklno-- 227

14. 15. 16. 17. 18. 19. 20. 21, 22. 23. 24. Ch~nic Obst~ve Luna Disease: ~ RaD~ ~f the Su~eon General. U.S. Gove~en~inting Office, 1984. Washington, DO. U.S. Depa~ment of Heal~ and Human Sluices, Office on Smoking and Health. The Maal~h conse~luencas of Smokina--~ ca~diovas~lar Dis~a~! A ~e~ ~ the Suroeon General. U. S. Government Printing Office, 1983. Washin~on, ~. Abbott, ~, et a., mRisk of StrOke in Male Cigarette Smokers," ~ew Enaland Jou~al of Medicine, Sept° 18, 1986 315:717o20. Di~ranza, JR, e~ al, "The relationship o~ s~king to motor vehicle accidents and ~raffio violations," ~, Sept. 1986. Milliman & Robe~son, I~o. Health Rigk and Behavior: The ~. 1~87, Brookfield, WI. Rice, DP and TA Hodgson, ""Economic Costs of Smo~ng: ~ Analysis of Data for ~e U.S.," presen~ a~ ~e Allied Social Science Association a~nual meeting, Sa~ Francisco, Dec. 28, 1983. Behre~s, ~. Reducin~ Sm~kin~ at ~m W~lace. Washington¸ \~ Business Group on Health. O~t 1985 Waehin~on, DO. WarneE, KE, WicMizer, TM~ Wolfe, RA, Schil~ro~h, JE, Samuel.on, ~. "Economic implications of workplace health promotion programs: review o£~he literature. J. Oct. Med. 30:106-112 (1988). Manning, WG, et al, "The Taxes of Sin: DO Smokers and Drinkers Pay Their Way," Journal of the American Medical VOI. 261, NO. II, March 17, 1989. warner, KE, "Selling Heal~h Pro~otlon to Corporate America: Uses and Abuses of the Ecoflo~c Ar~%tment," ~, Vol. 14, No. i, Spring 1987. Russell, LB, ~s Prevention Bettar than Cure? Brook~ngs Institute, 1986, Washlngtone DC. Schelling, TC, "Economics and Cigarettes," Medi~/ne, Vol. 15, 1986. * present address: 3026 East Marlette, Phoenix, AZ 85016 228